Smart Materials

Smart Materials

43722_C000.indd i

10/18/2008 7:26:39 AM

43722_C000.indd ii

10/18/2008 7:26:40 AM

Smart Materials Edited by

Mel Schwartz

Boca Raton London New York

CRC Press is an imprint of the Taylor & Francis Group, an informa business

43722_C000.indd iii

10/18/2008 7:26:40 AM

CRC Press Taylor & Francis Group 6000 Broken Sound Parkway NW, Suite 300 Boca Raton, FL 33487-2742 © 2009 by Taylor & Francis Group, LLC CRC Press is an imprint of Taylor & Francis Group, an Informa business No claim to original U.S. Government works Printed in the United States of America on acid-free paper 10 9 8 7 6 5 4 3 2 1 International Standard Book Number-13: 978-1-4200-4372-3 (Hardcover) This book contains information obtained from authentic and highly regarded sources. Reasonable efforts have been made to publish reliable data and information, but the author and publisher cannot assume responsibility for the validity of all materials or the consequences of their use. The authors and publishers have attempted to trace the copyright holders of all material reproduced in this publication and apologize to copyright holders if permission to publish in this form has not been obtained. If any copyright material has not been acknowledged please write and let us know so we may rectify in any future reprint. Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying, microfilming, and recording, or in any information storage or retrieval system, without written permission from the publishers. For permission to photocopy or use material electronically from this work, please access www.copyright.com (http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and registration for a variety of users. For organizations that have been granted a photocopy license by the CCC, a separate system of payment has been arranged. Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. Library of Congress Cataloging-in-Publication Data Smart materials / [edited by] Mel Schwartz. p. cm. Includes bibliographical references and index. ISBN 978-1-4200-4372-3 (alk. paper) 1. Smart materials. 2. Smart structures. I. Schwartz, Mel M. II. Title. TA418.9.S62S48 2008 620.1’1--dc22

2008018721

Visit the Taylor & Francis Web site at http://www.taylorandfrancis.com and the CRC Press Web site at http://www.crcpress.com

43722_C000.indd iv

10/18/2008 7:26:40 AM

Dedication To Carolyn and Anne-Marie who light up my life every day of the year

43722_C000.indd v

10/18/2008 7:26:40 AM

43722_C000.indd vi

10/18/2008 7:26:40 AM

Contents Preface ......................................................................................................................................................... xi Editor ........................................................................................................................................................... xv Contributors .............................................................................................................................................. xvii

1

Residual Stress in Thin Films ............................................................................................................ 1-1

2

Intelligent Synthesis of Smart Ceramic Materials ............................................................................ 2-1

3

A.G. Vedeshwar

Wojciech L. Suchanek and Richard E. Riman

Functionally Graded Polymer Blend ................................................................................................. 3-1

Yasuyuki Agari

4

Structural Application of Smart Materials ........................................................................................ 4-1

5

Composite Systems Modeling—Adaptive Structures: Modeling and Applications and Hybrid Composites ..................................................................................................................... 5-1 5.1 H ybrid Composites S. Padma Priya .........................................................................................................................5-1

R. Sreekala and K. Muthumani

5.2 Design of an Active Composite Wing Spar with Bending–Torsion Coupling Carlos Silva, Bruno Rocha, and Afzal Suleman .......................................................................................................... 5-7

6

Ferromagnetic Shape Memory Alloy Actuators ................................................................................ 6-1

7

Aircraft Applications of Smart Structures ........................................................................................ 7-1

8

Smart Battery Materials .....................................................................................................................8-1

9

Piezoelectric and Electrostrictive Ceramics Transducers and Actuators ........................................ 9-1

10

Yuanchang Liang and Minoru Taya Johannes Schweiger

Arumugam Manthiram

9.1 Smart Ferroelectric Ceramics for Transducer Applications A.L. Kholkin, D.A. Kiselev, L.A. Kholkine, and A. Safari ........................................................................................... 9-1 9.2 Smart Ceramics: Transducers, Sensors, and Actuators Kenji Uchino and Yukio Ito ................................................9-12 9.3 Noncontact Ultrasonic Testing and Analysis of Materials Mahesh C. Bhardwaj ...........................................................9-27

Chitosan-Based Gels and Hydrogels................................................................................................ 10-1 10.1 C hitosan-Based Gels Kang De Yao, Fang Lian Yao, Jun Jie Li, and Yu Ji Yin ......................................................... 10-1

10.2 Chitosan-Based Hydrogels in Biomedical and Pharmaceutical Sciences Claire Jarry and Matthew S. Shive ....................................................................................................................... 10-13

vii

43722_C000.indd vii

10/18/2008 7:26:40 AM

viii

11

Contents

Films, Coatings, Adhesives, Polymers, and Thermoelectric Materials .......................................... 11-1

11.1 S mart Adhesives James A. Harvey and Susan Williams ........................................................................................11-1 11.2 Oxides as Potential Ā er moelectric Materials S. Hébert and A. Maignan ............................................................ 11-4 11.3 Electrically Conductive Adhesives Yi Li, Myung Jin Yim, Kyoung-sik Moon, and C.P. Wong ....................................................................................................................... 11-12 11.4 Electrochromic Sol–Gel Coatings L.C. Klein ..................................................................................................... 11-25

12

Cure and Health Monitoring .......................................................................................................... 12-1

13

Drug Delivery Systems ..................................................................................................................... 13-1

14

Tatsuro Kosaka

13.1 Smart Drug Delivery Systems Il Keun Kwon, Sung Won Kim, Somali Chaterji, Kumar Vedantham, and Kinam Park ..................................................................................................................... 13-1 13.2 Drug Delivery Systems: Smart Polymers Joseph Kost ..........................................................................................13-10

Fiber Optic Systems: Optical Fiber Sensor Technology and Windows .......................................... 14-1

14.1 Introduction and Application of Fiber Optic Sensors Nezih Mrad and Henry C.H. Li ...........................................14-1 14.2 S mart Windows John Bell ...................................................................................................................................14-17

15

Flip-Chip Underfill: Materials, Process, and Reliability ................................................................ 15-1

16

Dielectric Cure Monitoring of Polymers, Composites, and Coatings: Synthesis, Cure, Fabrication, and Aging ......................................................................................... 16-1

Zhuqing Zhang and C.P. Wong

David Kranbuehl

17

Magnetorheological Fluids ...............................................................................................................17-1

18

Intelligent Processing of Materials .................................................................................................. 18-1

19

Magnets, Magnetic, and Magnetostrictive Materials ..................................................................... 19-1 19.1 M agnets, Organic/Polymer Joel S. Miller and Arthur J. Epstein ................................................................ 19-1

J. David Carlson

J.A. Güemes and J.M. Menéndez

19.2

20

Powder Metallurgy Used for a Giant Magnetostrictive Material Actuator Sensor Hiroshi Eda and Hirotaka Ojima .......................................................................................................................... 19-8

Shape-Memory Alloys and Effects: Types, Functions, Modeling, and Applications ..................... 20-1

20.1 Magnetically Controlled Shape Memory Alloys Ilkka Aaltio, Oleg Heczko, Outi Söderberg, and Simo-Pekka Hannula ..................................................................................................................................... 20-1 20.2 Shape Memory Alloys in Micro- and Nanoscale Engineering Applications Yves Bellouard ................................ 20-8 20.3 Mathematical Models for Shape Memory Materials Davide Bernardini and Ā omas J. Pence ............................20-17 20.4 Shape Memory Alloys Jan Van Humbeeck ........................................................................................................ 20-28 20.5 Smart Materials Modeling Manuel Laso ...........................................................................................................20-36 20.6 On the Microstructural Mechanisms of SMEs Monica Barney and Michelle Bartning....................................... 20-41

21

Current Developments in Electrorheological Materials ................................................................. 21-1

22

Carbon Microtubes and Conical Carbon Nanotubes ..................................................................... 22-1

23

“Smart” Corrosion Protective Coatings ..........................................................................................23-1

24

Smart Polymers for Biotechnology and Elastomers ....................................................................... 24-1 24.1 C onducting Polymers Gordon G. Wallace and C.O. Too .................................................................................... 24-1 24.2 Piezoelectricity in Polymers Aleksandra M. Vinogradov ................................................................................... 24-10 24.3 Polymers, Biotechnology, and Medical Applications Igor Yu Galaev and B. Mattiasson .....................................24-13

43722_C000.indd viii

Frank E. Filisko

Santoshrupa Dumpala, Gopinath Bhimarasetti, Suresh Gubbala, Praveen Meduri, Silpa Kona, and Mahendra K. Sunkara Patrick J. Kinlen and Martin Kendig

10/18/2008 7:26:40 AM

Contents

25

ix

Vibration Control for Smart Structures ......................................................................................... 25-1 25.1 Vib ration Control Seung-Bok Choi and Young-Min Han ................................................................................... 25-1

25.2

Smart Materials for Sound and Vibration Control Cai Chao, Lu Chun, Tan Xiaoming, and Zheng Hui .............................................................................................. 25-16

26

Active Truss Structures ................................................................................................................... 26-1

27

Application of Smart Materials and Smart Structures to the Study of Aquatic Animals ............ 27-1

28

Molecular Imprinting Technology ................................................................................................. 28-1

29

Biomedical Sensing ......................................................................................................................... 29-1

30

Intelligent Chemical Indicators ...................................................................................................... 30-1

31

Piezoelectric Polymer PVDF Microactuators ................................................................................. 31-1

32

Ultrasonic Nondestructive Testing and Materials Characterization ............................................ 32-1

33

Lipid Membranes on Highly Ordered Porous Alumina Substrates ............................................... 33-1

B. de Marneffe and André Preumont

Jesse E. Purdy and Alison Roberts Cohan David A. Spivak

Dora Klara Makai and Gabor Harsanyi Christopher O. Oriakhi

Yao Fu, Erol C. Harvey, and Muralidhar K. Ghantasala John A. Brunk

Andreas Janshoff and Claudia Steinem

Index ............................................................................................................................... .............................I-1

43722_C000.indd ix

10/18/2008 7:26:41 AM

43722_C000.indd x

10/18/2008 7:26:41 AM

Preface Many “ smart” m aterials w ere i nvented mo re t han 3 0 ye ars a go, b ut t heir de velopment a nd i mprovement o ver t he pa st t hree decades has led to new, more varied uses of these adaptable materials. Smart materials have properties that can be altered by temperature, moisture, electric or magnetic elds, pH, and stress. Ā ey can change shape and color, become stronger, or produce voltage as a result of external stimuli. Magnetostrictive materials—materials that expand when exposed to a magnetic eld—were discovered in the 1800s, but have uses today as varied as automotive sensors for collision avoidance and vibration dampening for surgical tools. Materials such as brass, nickel-titanium, and gold-cadmium are shape memory alloys (SMAs), which alter their shape in response to changes in temperature and then return to their original shape. In the past 30 years, SMAs have seen widespread use in applications such as miniature surgical tools that can twist and pull when a small amount of heat is applied, wires that expand and contract for use by dentists in straightening teeth, cell phone antennas that resist breaking, stents and bone plates that must fuse together or expand in order to stay in place, and exible eyeglass frames. “Smart m aterials,” w hich  nd widespread applications today, should be recognized along with the familiar metals, plastics, ceramics, composites, powder metals, a nd specialty-type a nd multifunctional materials (SMAs, microelectromechanical systems [MEMS], f unctionally g raded m aterials [FGMs], a nd na nomaterials). Ā ree de cades a go, i t app eared a s t hough sm art m aterials would be the next step in engineering design. By using smart materials instead of adding mass, engineers can endow structures with built-in responses to a myriad of contingencies. In their various forms, these materials can perform as actuators, which can adapt to their environments by changing characteristics such as shape and stiff ness, or as sensors, which provide the actuators with information about structural and environmental changes. All aircraft, whether military commercial, or privately owned, should be and are inspected and maintained on a re gular basis. Intelligent research programs benet t his segment of t he aerospace industry. Sensors introduced into current a nd f uture aircraft designs will help technicians and mechanics to provide a more sophisticated inspection technique for maintenance and repairs. Smart materials are beginning to play an important role in civil engineering designs for dams, bridges, highways, and buildings. An example of a smart materials project sponsored by the U.S. Navy and Army Corps of Engineers is to remove corrosion that has damaged a Navy pier and install sheets of composite materials containing sensors. Ā ese sensors embedded throughout a concrete and c omposite s tructure c an ac t l ike ner ves, s ensing w hen a reas o f t he s tructure b egin to de grade a nd a lerting m aintenance engineers to the need for repairs. Ā e automotive industry is also eager to i ncorporate intelligent materials technology. Currently, researchers are working on an industry-sponsored project to develop smart car seats that can identify primary occupants and adapt to their preferences for height, leg room, back support, and so forth. More profound changes are looming in automotive design based on smart materials. For instance, the technology to enable cars to tell owners how much air pressure tires have, when oil changes are needed, and other maintenance information exists as of today. However, this technology is expensive, but developing solid state and smart materials technologies will bring costs down. Ā ere is a shift in the culture toward consumers being given more information and taking more responsibility for knowing when maintenance and repairs are needed. Ā e future of stereophonic sound will be altered by another facet of smart materials research. Ā e development of u ltrahigh-delity stereo speakers using piezoelectric actuators, which expand and contract in thousandths of a s econd in response to applied voltage, is aimed at t urning whole house walls or car interiors into speakers by imbedding them with the tiny actuators. Ā us, 50 years from now, we would not need to install separate speakers in our homes and cars in an attempt to achieve maximum musical effects. Our cars and houses will offer built-in surround sound. A new technology for implants that may improve construction or repair of bones in the face, skull, and jaw has been developed by researchers from the American Dental Association Foundation and the National Institute of Standards and Technology. Ā e new

xi

43722_C000.indd xi

10/18/2008 7:26:41 AM

xii

Preface

technology provides a method for making scaffolds for bone tissue. Ā e scaffold is seeded with the patient’s own cells and is formed with a cement paste made of minerals also found in natural bone. Ā e paste is mixed with beads of a natural polymer (made from seaweed)  lled with bone cells. Ā e paste is shaped or injected into a bone cavity and then allowed to harden with the encapsulated cells dispersed throughout the structure. Ā e natural polymer beads gradually dissolve when exposed to the body’s uids, creating a scaffold that is  lled by the now released bone cells. Ā e cement, a c alcium phosphate material, is strengthened by adding chitosan, a b iopolymer extracted from crustacean shells. Ā e implant is further reinforced to about the same strength as spongy natural bone by covering it with several layers of a biodegradable ber mesh already used in clinical practice. Bone cells are very smart. Ā ey can tell the difference between materials that are bioactive compared to bioinert polymers. Ā e material is so designed to be similar to mineral in bone so that cells readily attach to the scaffold. In addition to creating pores in the hardened cement, the natural polymer beads protect the cells during the 30 min required for the cement to harden. Future experiments will develop methods for improving the material’s mechanical properties by using smaller encapsulating beads that biodegrade at a predictable rate. Other developments include a smart bandage, which has an in situ programmable medical device to treat wounds and burns, a smart pill for vitamin and nutraceuticals formulation that prevents cognitive decline in aging and amphiphilic polymeric materials where the polymers can be used in the development of physiologically stable, nonleaking, nonimmunogenic, safe, and efficacious targeted drug delivery systems. Recent advances in nondestructive evaluation (NDE) sensor technologies, health monitoring, and life prediction models are revolutionizing component inspections and life management, and will signicantly improve the reliability and airworthiness of aerospace s ystems. S everal ke y adv ances i n N DE, health, monitoring, a nd prognostics programs i nclude emerg ing s tructural health monitoring technologies for aerospace applications, high-temperature health monitoring, advances in NDE technologies for measurement of subsurface residual stresses, computational methods and advanced NDE techniques, and materials damage assessment techniques a nd p rognosis mo dels f or p rediction o f rem aining u seful l ife. Ā e adv ancements de scribed p rovide t he te chnologies required for lowering m aintenance c osts, i ntegrated d amage prognosis a nd l ife prediction, en hancing rel iability a nd s afety, a nd improving the performance and operational efficiency of current and future aerospace systems. Another area of focus is smart machines. For example, moving composite manufacturers out of hand layup and open molding into more environmentally acceptable closed-molding alternatives where a reliable source of preforms is a necessity for volume production in closed-mold processes, such as resin transfer molding (RTM), vacuum-assisted RTM (VARTM), and vacuum-assisted resin infusion molding (VARIM). Large preforms are facilitated by the large-scale preformer (LSP). Ā e LSP’s computerized spray-up process requires no human contact from the time the tooling enters the preform manufacturing cell until it is ready to be demolded: spray-up, compression, cure of the binder, cooling, and demold operations are all accomplished robotically according to i nstructions preprogrammed into the system’s soft ware. LSP is a smart system that tells the user where every part is in the process at all times. Ā e LSP’s inaugural application, and the largest preform produced thus far, was for fuel containment vessels, which were resin transfer molded. Ā ese vessels have been installed beneath gasoline pumps in gas stations to prevent contamination of soil if there is a leak in the pump-to-tank plumbing. Composites a re now used i n monitoring systems where t hey a re combined w ith ot her materials a nd sensors. A n example is a composite “shape sensing mat” for use with a metallic riser system. Ā e exible mat, which incorporates ber optics, wraps around a steel riser, and enables operators to monitor excessive bending and fatigue life during riser deployment on a dynamic positioned oil drilling ship moored in the Gulf of Mexico as well as smart downhole coiled tubing complete with power and data transmission capabilities for drilling or workover applications and oileld pipelines. Finally, machines can process simple commands, but they are not very good at guring out complex orders or unstated common sense. Command a machine to “paint the computer case before you box it,” or “provide power to the computer before you switch it on” and the machine may box the product before the case is dry, or plug and then unplug a c omputer before switching it on. Ā e meaning of the word “before” is quite different in these two cases. Ontologists, who study and understand the thought process, hope to end t he a ge o f s tupid m achines. O ntologists, w ho h ave c reated s ome o f t he mo st adv anced log ic s ystems, p lan to s hare t heir leading-edge concepts on such comprehensive ideas as time, space, and process. Ā e promise to cooperate eventually could lead to soft ware that will enable machines to interpret and act on commands with near human common sense. Efforts to equip machines with articial intelligence capacity have, up to now, been relatively rudimentary. Softwa re programs might, for instance, give machines used to m ake furniture considerable “understanding” of terms and frames of reference used in the furniture business. But such collected knowledge is of limited use, and human operation is necessary at virtually every step in the manufacturing process. A machine that incorporates expanded frames of reference of such “higher ontologies” as space

43722_C000.indd xii

10/18/2008 7:26:41 AM

Preface

xiii

and cost might be able to make design and shipping decisions virtually on its own. Ā e future is bright and with optimism will enable the leading ontologists throughout the world to continue this promising work. Ā e book contains many of the examples and of the aspects that I have discussed and the contributions and efforts of 60 experts in the various  elds of smart materials and smart material systems. I h ope the readers will appreciate the work of a m ultitude of scientists, educators, researchers, academia, and industry people who have made considerable innovative progress in bringing forth their endeavors. Mel Schwartz

43722_C000.indd xiii

10/18/2008 7:26:41 AM

43722_C000.indd xiv

10/18/2008 7:26:41 AM

Editor Mel S chwartz h as de grees i n me tallurgy a nd eng ineering m anagement a nd h as s tudied l aw, me tallurgical eng ineering, a nd education. His professional experience extends over 51 years. He has served as a metallurgist in the U.S. Bureau of Mines; as a metallurgist and producibility engineer in the U.S. Chemical Corps; as a technical manufacturing manager, chief R&D laboratory, research manufacturing engineering, and senior staff engineer in Martin-Marietta Corporation for 16 years; as a program director, manager and director of manufacturing for R&D, and chief metals researcher in Rohr Corp for 8 years; and as a staff engineer and specication specialist, chief metals and metal processes, and manager of manufacturing technology in Sikorsky Aircraft for 21 years. After retirement, Schwartz served as a consultant for many companies including Intel and Foster Wheeler, and is currently editor of SAMPE Journal of Advanced Materials. Schwartz’s professional awards and honors include Inventor Achievement Awards and Inventor of the Year at Martin-Marietta; C. Adams Award and Lecture and R.D. Ā omas Memorial Award from AWS;  rst recipient of the G. Lubin Award and an elected fellow from SAMPE; an elected fellow and Engineer of the Year in CT from ASM; and the Jud Hall Award from SME. Schwartz’s other professional activities include his appointment to ASM technical committees (joining, composites and technical books, ceramics); manuscript board of review, Journal of Metals Engineering as peer reviewer; the Institute of Metals as well as Welding Journal as peer reviewer; U.S. leader of International Institute of Welding (IIW) Commission I ( brazing and related processes) for 20 years and leader of IIW Commission IV (electron beam/laser and other specialized processes) for 18 years. Schwartz owns ve patents, the notable one being aluminum dip brazing paste commercially sold as Alumibraze. He has authored 16 books and over 100 technical papers and articles and is an internationally known lecturer in Europe, the Far East, and Canada. He has taught i n U.S. Universities (San Diego State University, Yale University), ASM i nstitutes, McGraw-Hill seminars, a nd i nhouse company courses.

xv

43722_C000.indd xv

10/18/2008 7:26:41 AM

43722_C000.indd xvi

10/18/2008 7:26:42 AM

Contributors Ilkka Aaltio Department of Materials Science and Engineering Helsinki University of Technology TKK, Finland

John A. Brunk National Nuclear Security Agency’s Kansas City Plant Kansas City, Missouri

Hiroshi Eda Intelligent Systems Engineering Department Ibaraki University Hitachi, Japan

Yasuyuki Agari Osaka Municipal Technical Research Institute Osaka, Japan

J. David Carlson Lord Corporation Cary, North Carolina

Monica Barney Nitinol Devices and Components Freemont, California

Cai Chao Institute of High Performance Computing Singapore

Arthur J. Epstein Department of Physics and Department of Chemistry Ohio State University Columbus, Ohio

Michelle Bartning Cordis Advanced Medical Ventures Freemont, California John Bell Queensland University of Technology Brisbane, Queensland, Australia

Somali Chaterji Weldon School of Biomedical Engineering Purdue University West Lafayette, Indiana

Frank E. Filisko Ā e University of Michigan Ann Arbor, Michigan Yao Fu Silverbrook Research Pty. Ltd. St. Balmain, Sydney, Australia Igor Yu Galaev Lund University Lund, Sweden

Yves Bellouard Mechanical Engineering Department Eindhoven University of Technology Eindhoven, Ā e Netherlands

Seung-Bok Choi Department of Mechanical Engineering Inha University Inchon, South Korea

Davide Bernardini Department of Structural and Geotechnical Engineering University of Rome Rome, Italy

Lu Chun Institute of High Performance Computing Singapore

Mahesh C. Bhardwaj Ā e Ultran Group State College, Pennsylvania

Alison Roberts Cohan Pacic Whale Foundation Maui, Hawaii

Suresh Gubbala Department of Chemical Engineering University of Louisville Louisville, Kentucky

Gopinath Bhimarasetti Department of Chemical Engineering University of Louisville Louisville, Kentucky

Santoshrupa Dumpala Department of Chemical Engineering University of Louisville Louisville, Kentucky

J.A. Güemes Department of Aeronautics Universidad Politécnica de Madrid Madrid, Spain

Muralidhar K. Ghantasala Department of Mechanical and Aeronautical Engineering Western Michigan University Kalamazoo, Michigan

xvii

43722_C000.indd xvii

10/18/2008 7:26:42 AM

xviii

Young-Min Han Department of Mechanical Engineering Inha University Inchon, South Korea Simo-Pekka Hannula Department of Materials Science and Engineering Helsinki University of Technology TKK, Finland Gabor Harsanyi Department of Electronics Technology Budapest University of Technology and Economics Budapest, Hungary Erol C. Harvey Industrial Research Institute Swinburne University of Technology Hawthorne, Melbourne, Australia James A. Harvey Under the Bridge Consulting Inc. Corvallis, Oregon S. Hébert Laboratoire Crismat CNRS ENSICaen Caen, France

Contributors

Claire Jarry Bio Syntech Canada, Inc. Laval, Quebec, Canada Martin Kendig Teledyne Scientic Company Ā ousand Oaks, California A.L. Kholkin Department of Ceramics and Glass Engineering and Center for Research in Ceramic and Composite Materials University of Aveiro Aveiro, Portugal L.A. Kholkine Department of Electrical and Computer Engineering University of Porto Porto, Portugal Sung Won Kim Department of Pharmaceutics Purdue University West Lafayette, Indiana Patrick J. Kinlen Crosslink Fenton, Missouri

Joseph Kost Department of Chemical Engineering Ben-Gurion University Beer Sheva, Israel David Kranbuehl Chemistry and Applied Science Departments College of William and Mary Williamsburg, Virginia Il Keun Kwon Department of Pharmaceutics Purdue University West Lafayette, Indiana Manuel Laso Laboratory of Non-Metallic Materials Universidad Politécnica de Madrid Madrid, Spain Henry C.H. Li School of Aerospace, Mechanical and Manufacturing Engineering RMIT University Fishermans Bend, Victoria, Australia Jun Jie Li School of Chemical Engineering and Technology Tianjin University Tianjin, China

Oleg Heczko Department of Materials Science and Engineering Helsinki University of Technology TKK, Finland

D.A. Kiselev Department of Ceramics and Glass Engineering and Center for Research in Ceramic and Composite Materials. University of Aveiro Aveiro, Portugal

Zheng Hui Institute of High Performance Computing Singapore

L.C. Klein Rutgers University Piscataway, New Jersey

Yuanchang Liang Department of Mechanical Engineering University of Washington Seattle, Washington

Silpa Kona Department of Electrical and Computer Engineering University of Louisville Louisville, Kentucky

A. Maignan Laboratoire Crismat CNRS ENSICaen Caen, France

Jan Van Humbeeck Catholic University of Leuven Leuven, Belgium Yukio Ito Ā e Pennsylvania State University University Park, Pennsylvania Andreas Janshoff Johannes Gutenberg Universität Mainz, Germany

43722_C000.indd xviii

Tatsuro Kosaka Graduate School of Engineering Osaka City University Osaka, Japan

Yi Li School of Materials Science and Engineering Georgia Institute of Technology Atlanta, Georgia

Dora Klara Makai Department of Electronics Technology Budapest University of Technology and Economics Budapest, Hungary

10/18/2008 7:26:42 AM

Contributors

Arumugam Manthiram Materials Science and Engineering Program University of Texas at Austin Austin, Texas B. de Marneffe Active Structures Laboratory Université Libre de Bruxelles Brussels, Belgium B. Mattiasson Lund University Lund, Sweden Praveen Meduri Department of Chemical Engineering University of Louisville Louisville, Kentucky J.M. Menéndez Composite Materials Technology Department Airbus Getafe (Madrid), Spain Joel S. Miller Department of Chemistry University of Utah Salt Lake City, Utah Kyoung-sik Moon School of Materials Science and Engineering Georgia Institute of Technology Atlanta, Georgia Nezih Mrad Department of National Defence National Defence Headquarters Ottowa, Ontario, Canada K. Muthumani Structural Dynamics Laboratory Structural Engineering Research Centre, CSIR Chennai, India Hirotaka Ojima Intelligent Systems Engineering Department Ibaraki University Hitachi, Japan Christopher O. Oriakhi Imaging and Printing Supplies Hewlett-Packard Company Corvallis, Orgeon

43722_C000.indd xix

xix

Kinam Park Department of Biomedical Engineering Purdue University West Lafayette, Indiana

David A. Spivak Department of Chemistry Louisiana State University Baton Rouge, Louisiana

Thomas J. Pence Department of Mechanical Engineering Michigan State University East Lansing, Michigan

R. Sreekala Structural Dynamics Laboratory Structural Engineering Research Centre, CSIR Chennai, India

André Preumont Active Structures Laboratory Université Libre de Bruxelles Brussels, Belgium

Claudia Steinem Georg-August Universität Göttingen, Germany

S. Padma Priya University of Mysore Mandya, India

Wojciech L. Suchanek Sawyer Technical Materials, LLC Eastlake, Ohio

Jesse E. Purdy Department of Psychology Southwestern University Georgetown, Texas

Afzal Suleman Department of Mechanical Engineering University of Victoria Victoria, British Columbia, Canada

Richard E. Riman Department of Materials Science and Engineering Rutgers University Piscataway, New Jersey Bruno Rocha Department of Mechanical Engineering Instituto Superior Técnico Lisbon, Portugal A. Safari Rutgers University Piscataway, New Jersey Johannes Schweiger German Aerospace Society Ban Heilbrunn, Germany Matthew S. Shive Bio Syntech Canada, Inc. Laval, Quebec, Canada Carlos Silva Laboratory of Aeronautics Portuguese Air Force Academy Sintra, Portugal Outi Söderberg Department of Materials Science and Engineering Helsinki University of Technology TKK, Finland

Mahendra K. Sunkara Department of Chemical Engineering University of Louisville Louisville, Kentucky Minoru Taya Department of Mechanical Engineering University of Washington Seattle, Washington C.O. Too Intelligent Polymer Research Institute University of Wollongong Wollongong, New South Wales, Australia Kenji Uchino Department of Electrical Engineering Ā e Pennsylvania State University University Park, Pennsylvania and Micromechatronics Inc State College, Pennsylvania Kumar Vedantham Department of Pharmaceutics Purdue University West Lafayette, Indiana

10/18/2008 7:26:42 AM

xx

Contributors

A.G. Vedeshwar Department of Physics and Astrophysics University of Delhi New Delhi, India

C.P. Wong School of Materials Science and Engineering Georgia Institute of Technology Atlanta, Georgia

Aleksandra M. Vinogradov Department of Mechanical and Industrial Engineering Montana State University Bozeman, Montana

Tan Xiaoming Institute of High Performance Computing Singapore

Gordon G. Wallace Intelligent Polymer Research Institute University of Wollongong Wollongong, New South Wales, Australia

Fang Lian Yao School of Chemical Engineering and Technology Tianjin University Tianjin, China

Susan Williams Hewlett Packard Company Corvallis, Oregon

Kang De Yao Research Institute of Polymeric Materials Tianjin University Tianjin, China

43722_C000.indd xx

Myung Jin Yim School of Materials Science and Engineering Georgia Institute of Technology Atlanta, Georgia Yu Ji Yin Research Institute of Polymeric Materials Tianjin University Tianjin, China Zhuqing Zhang Hewlett-Packard Company Corvallis, Oregon

10/18/2008 7:26:42 AM

1 Residual Stress in Thin Films ntroduction .............................................................................................................................. 1-1 Models and The oretical Background..................................................................................... 1-1 Experimental Methods for the Measurement of Residual Stress ...................................... 1-2 Residual Stress-Dependent Optical Properties of Some Layered Structured Semiconductors .........................................................................................................................1-4 1.5 Summary and Future Direction ........................................................................................... 1-12 References ........................................................................................................................................... 1-12

1.1 I 1.2 1.3 1.4

A.G. Vedeshwar University of Delhi

1.1 Introduction One o f t he mo st c ommon a nd ye t a n i mportant i ssue i n t hin  lms is the persistence of stress aĀer t he  lm growth, which is termed as t he internal residual st ress. A lthough t he attempt to measure a nd u nderstand t he s tress i n  lms s tarted a s e arly a s 1877 [1], the topic continues to be interesting and important even today with various innovative analyses and measurement techniques. A few early authoritative reviews [2,3] on this topic have been excellent sources of information, which has assisted in the further g rowth o f k nowledge. The re sidual s tress de pends o n various factors like t he method of growth, growth pa rameters, nature of substrates, and starting material and then aĀer growth  lm processing, etc. Development of stress in  lms c ould b e both adv antageous a nd d isadvantageous. A v ery w ell-known, advantageous effect is the formation of self-assembled quantum dots i n h eteroepitaxial g rowths c aused b y t he de velopment o f stress due to the lattice mismatch between lm and the substrate [4,5]. Thus, the quantum dot growth can be tailored or controlled in a de sirable w ay b y a p roper c hoice o f t he e xtent o f  lm– substrate lattice mismatch. This has already led to the successful development of various devices based on quantum dots. However, residual stress may be quite undesirable and disadvantageous in the f abrication o f a lmost a ll ot her p lanar t hin  lm electronic devices as it could lead to failure of the device, or else can modify the de vice p erformance u ndesirably, e .g., m icroelectronic o r microelectromechanical s ystems ( MEMS). Sub sequently, t he study of the residual stress effect and its elimination or minimization b ecomes ne cessary f or suc h app lications. M ost o f t he properties of the materials are affected by the stress, either externally applied or the internal residual stress. Therefore, the study of t he residual stress effect in thin  lms could be a nalogous to the e xternally app lied p ressure e ffect. This fac t faci litates t he

possibility of studying the pressure-dependent physical properties without any actual externally applied high-pressure experiments. As the band structure depends crucially on the structure, the electronic properties of the semiconductors can be expected to e xhibit i nteresting s tress–strain de pendence. The s tudy o f residual stress therefore seems to be of signicant importance, although t he f undamental me chanisms f or i ts o rigin a re f ar from being fully understood. Therefore, I will try to present the necessary and sufficient information on the topic in this limited review.

1.2

Models and Theoretical Background

Almost all lms have stress and conceptually the total stress can be thought of having three major contributions, external, thermal and the intrinsic. It can be represented as [6] σ = σ external + σ thermal + σ intrinsic (1

.1)

as o Āen d istinguished i n t he l iterature. I ntrinsic s tresses a re developed d uring t he de position p rocess o f t he  lms. Ther mal stress arises due to the difference between the thermal expansion coefficients o f t he sub strate a nd t he  lm m aterial, a nd o ccurs especially during the cooling phase. External stresses are mainly considered as due to, for instance, oxidation and incorporation of impurities etc. Most of the formulas used in experimental determinations of  lm stress sf are the modications of an equation rst derived by Stoney in 1909 [7], which is given by σf =

E Sd S2 (1 6R (1 − ν S )d f

.2) 1-1

43722_C001.indd 1

10/3/2008 8:42:10 PM

1-2

Smart Materials

where ES and νS are Young’s modulus and the Poisson ratio of the substrate, respectively d S and d f are the thickness of substrate and film, respectively R i s t he r adius o f c urvature o f t he b ending i n t he  lm– substrate system caused by the stress The me asurement of R w ill de termine sf. I n t he l iterature, t he (1 − νS) correction is frequently omitted. Most of the optical interferometric methods devised to measure R (or change in R) are based on the dependence of fringe-width on R. The origin of intrinsic stress is not yet fully explained by any universal mechanism. However, Buckel [8] suggested and classied the processes leading to the generation of internal stress into t he c ategories a s: ( 1) d ifferences i n t hermal e xpansion coefficient o f  lm a nd sub strate, (2) d ifferences i n t he l attice spacing of single crystal substrates a nd t he  lm during epitaxial growth, (3) variation of the interatomic spacing with crystal size, (4) adsorption or incorporation of atoms from residual gases o r c hemical re actions, ( 5) re crystallization p rocesses, (6) microscopic voids and special arrangements of dislocation, and (7) phase transformations. Large i ntrinsic ten sile s tresses ob served i n me tal  lms has been explained by a model [9], which suggests that the annealing and shrinkage of the layer buried behind the advancing surface of t he g rowing  lm le ads to s tress. Hoff man [10] t ried to e stimate the intrinsic residual stress using a model in which the isolated g rains c oalesce a nd f orm g rain b oundaries. Gr ains o f radius R are assumed to be separated by a distance d (constrained relaxation leng th), o f t he o rder o f i nteratomic d istances. This model u ses a g rain-boundary p otential ( similar to M oorse potential) having a minimum at equilibrium atomic separation. The tensile or compressive strain developed in the  lm is attributed to the situation of growth whether the interatomic distance is smaller or greater than the equilibrium distance, respectively. The intrinsic residual stress deduced by this model is given by σ intrinsic =

Efδ P (1 (1 − ν f )2R

.3)

where P is the packing density of the lm Ef an d νf a re Young’s mo dulus a nd P oisson r atio o f t he lm, respectively It should be noted that the term Ef /(1 − νf ) replaces Ef because of the assumed biaxial nature of the stress. The parameter d can be determined from the interaction potential between the concerned atoms, which is oĀen not so easy. Recently, Nix and Clemens [11] attempted to i mprove t his mo del o f c oalescence me chanism using a thermodynamical approach. They derived a relationship for the maximum value of the stress as 1/ 2

⎡ E f (2γ g − γ gb ) ⎤ σ intrinsic (max) = ⎢ ⎥ (1 ⎣ (1 − ν f )R ⎦

43722_C001.indd 2

.4)

where gg and ggb are the surface tensions per unit area of the isolated g rains a nd t he g rain b oundary, re spectively. A gain, h ere too, the determination of the parameter ggb is quite difficult in this improved model. Another very recent report [12] attempted to explain t he i ntrinsic residual stress i n metal  lms using a mo del based on t he size-dependent phase t ransformations of t he na nograins via the size dependence of the melting temperature of nanoparticles assumed to be present at the early stage of lm deposition. Similarly, a model for compressive stress generation in polycrystalline  lms during t hin  lm growth is reported [13] in which the driving force is an increase in the surface chemical potential caused by the deposition of atoms from the vapor. The increase in surface chemical potential induces atoms to  ow into the grain boundary and hence c reates a c ompressive s tress i n t he  lm. A n umber of other s tress mo dels [14–18] der ived f rom t hese ide as h ave b een invoked in t he literature to e xplain both tensile a nd compressive stress in thin lms, which will not be discussed further. The magnitude of intrinsic stresses in lms may also be related to the microstructure of lms, i.e., morphology, texture, grain size, etc. Ther mal effects can also contribute to  lm stress signicantly. Films grown at elevated substrate temperatures and then cooled to a mbient tem perature w ill de velop t hermal s tress. Si milarly, the lms either thermally cycled or cooled to cryogenic temperatures will also develop thermal stress. A biaxial stress can appear in lms grown on substrates having different thermal expansion coefficients th an th e  lm, at a tem perature T higher or lower than t he subst rate o r de position tem perature TS. The strain eT developed under such conditions is given by [3] ε T = (α f − α s )∆T (1

.5)

where ∆T = (T – TS), a s, and a f are the thermal expansion coefficients of the substrate and  lm, respectively. Using Hooke’s law, the relation between thermal stress and the elastic strain in the absence of any plastic deformation in t he  lm–substrate structure during temperature change can be obtained as σ f (T ) =

E f (α s − α f )∆T (1 (1 − ν f )

.6)

A t hird k ind, ter med a s e xtrinsic s tress, c an b e d istinguished apart f rom t hermal s tress a nd i ntrinsic s tress i n t he o verall residual s tresses i n  lms. Various mole cules c an p enetrate t he open voids and pores present in a not so fully dense lm and are adsorbed on pore walls. The interaction forces between adsorbed species, e specially b etween p olar s pecies suc h a s w ater mole cules, can modify residual stresses. Hirsch [19] proposed a model to explain the origin of extrinsic stress based on the adsorption of polar species on pore walls.

1.3 Experimental Methods for the Measurement of Residual Stress A wide variety of methods have been explored for measuring the stress in thin  lms [6,20]. They may broadly be categorized on

10/3/2008 8:42:12 PM

1-3

Residual Stress in Thin Films

the basis of the physical phenomena used as the technique such as diff raction (x-ray or electron diff raction), optical interferometry, electrical or electromechanical etc. However, a c onvenient and app ropriate me thod m ay b e c hosen de pending o n t he requirements like measurements in situ or ex situ, types of lm– substrate m aterial, le vel o f ac curacy e tc. The interferometric methods are normally used to measure the extent of bending or deformation of a thin substrate (cantilever) from its equilibrium position caused by the  lm h aving s tress de posited o n i t. The fringe width is proportional to the extent of bending or deformation and hence the stress can be determined. The nature of the biaxial stress such as balanced, unbalanced, or the different type of stress components, etc. can also be analyzed by the shape of the f ringe pat tern a s has been reported for sputtered Mo  lms [21]. A tensile stress will bend the substrate so that the lm surface is concave and the opposite (convex) happens with the compressive stress. A thin cantilever is most widely used as the substrate. A v ariety of i nterferometric techniques such as laser reective interferometry [22–24], laser spot scanning interferometry [25], optical leverage with a laser beam [26,27], etc. have been developed. However, it should be noted that the stresses measured by cantilever techniques and x-ray diffraction (XRD) need not ne cessarily the same because X RD gives the strain in the crystal lattice of the grain (intragranular) while the cantilever me thod de termines t he s tress d ue to g rain b oundaries between the grains (intergranular). There are a few other innovative modern techniques developed to meet specic requirements such a s i ndentation [ 28,29] a nd l aser s pallation [ 30], s train gauges [31–33], Raman measurements [34–36], and so on. X-ray or electron diff raction i s one of t he si mplest ye t mo st powerful me thods f or m icrostructural c haracterization. Therefore, more emphasis will be given to this technique here. The details can be found from the standard and widely referred books [37,38]. XRD data can also be used to determine either uniform or nonuniform residual stress in the lms. The effect of strain, both uniform and nonuniform, on the direction of x-ray reection causes the diff raction peak in a denite way. The uniform strain shiĀs the peak on either side of the unstrained 2q to the lower side for tensile and to the higher side for compression. If a g rain is given a u niform tensile strain at r ight angles to t he reecting p lanes ( parallel to sub strate p lane), t heir s pacing becomes larger than unstrained equilibrium spacing d 0 and the corresponding diffraction line shiĀs to lower angles but does not otherwise change. This line shiĀ for the given (hkl) can be used to calculate the strain e z present on (hkl) along the z-axis in the lm as εz =

0 ∆d hkl d hkl − d hkl (1 = 0 0 d hkl d hkl

.7)

The r ight h and side o f E quation 1.7 w ill b e p ositive for ten sile stress and negative for a c ompressive stress. However, it should be noted that normally the d0 value for the desired (hkl) is taken from powder data (ASTM or JCPDS) of the concerned material because f or a go od app roximation, p owder m ay b e t reated a s

43722_C001.indd 3

stress free or having a negligible stress. Thus, strain for all (hkl) peaks o f c onsiderable i ntensity i n t he d iffractogram c an b e determined and their average ∆d/d can be calculated for a r andomly o riented p olycrystalline  lm. The m acrostress t hus c an be de termined b y m ultiplying t he a verage s train ∆d/d by t he elastic constant of the material. However, quite oĀen, there will be a single most intense peak (hkl) in the diff ractogram for (hkl)oriented parallel to the substrate plane and the determination of strain is only for that (hkl). Similarly, using electron diffraction data, the d-spacings can be determined [39] and hence the strain. For a biaxial stress along the x- and y-axes (sx ≠ 0, sy ≠ 0), sz = 0 we have strain normal to the  lm surface given by [37,38] εz = −

νf (σ x + σ y ) (1 Ef

.8)

The above equation represents a contraction if sx and sy are tensile. Therefore, a p ositive e z determined by Equation 1.7 would imply t he c ompressive b iaxial s tress. A c omponent o f b iaxial stress in any desired direction on the xy-plane can be measured by tilting the sample by an angle y, the details of which may be found in Refs. [37,38]. This method is known as the sin2y method in t he mo dern l iterature. A no nuniform m icrostrain c auses a broadening of t he corresponding diffraction line. The relationship between t he broadening produced and t he nonuniformity of the strain can be obtained by differentiating Bragg’s law. We therefore obtain b = ∆2q = −2

∆d tan q (1 d

.9)

where b is the extra broadening, over and above the instrumental breadth of the line, due to a fractional variation in plane spacing, ∆d/d. This equation a llows the variation in strain, ∆d/d, to be c alculated f rom t he ob served b roadening. The maximum strain thus found can be multiplied by the elastic modulus Ef to give the maximum stress present. Although considerable progress has taken place both in theory and in experimental methods, the residual stress data in the literature show wide variations for various materials measured by different me thods. Bro ad t rends a re v isible i n t he p ublished results and can be summarized as 1. Most o f t he me tal  lms e xhibit i nvariably ten sile s tress with a l arge magnitude (200–1100 MPa or 106 N/m 2) and less dependence on the nature of the substrate. 2. Refractory metals and metals having high melting points generally e xhibit h igher s tresses ( 600–1200 MPa) t han soĀer and lower melting point metals (20–100 MPa). 3. Normally, b oth c ompressive a nd ten sile s tresses a re observed i n no nmetallic  lms f requently w ith sm aller magnitudes (7–700 MPa). The crossover between the two types of stress is also observed. Data on some metals can be found in Refs. [3,9]. Similarly, early investigations on dielectric and optical coatings have been

10/3/2008 8:42:14 PM

1-4

Smart Materials

described in Refs. [3,40,41]. A re cent report shows the relationship b etween t he p referred o rientations a nd s tresses f or s ome metals [42]. The stress behavior in a very important technological material, GaN, has been studied [43–45] and ot her nitrides also have been investigated [46–48]. These are only a few informative examples. However, there are a large numbers of reports on v arious m aterials, s pecic s tudies, e xperimental me thods, analyses e tc., w hich a re b eyond t he s cope a nd p urpose o f t his chapter.

1.4

Residual Stress-Dependent Optical Properties of Some Layered Structured Semiconductors

It may be i nteresting to u nderstand t he origin a nd behavior of residual stress i n  lms; however, it is more important to know how the residual stress affects the various properties of the lm. In this section, some selected data will be presented for illustrating some of the points so far discussed. The layered structured semiconductors (like metal dihalides, chalcogenides, etc.) were chosen for t his purpose b ecause of t heir a nisotropy a long a nd across the layer. In a layered structure, the constituent atoms are bound by a strong covalent bonding within the ab-plane forming

a sheet-like structure (layer) and these sheets are stacked along the t hird d irection (c-axis) b y a w eak v an der W aals b onding between t hese s heets to f orm a t hree-dimensional s olid, t hus leading to a nisotropy a long pa rallel a nd p erpendicular to t he c-axis. A ll t he  lms p resented i n t his s ection w ere g rown b y thermal evaporation using a molybdenum boat onto a glass substrate at room temperature (RT) or liquid nitrogen temperature (LNT) at a v acuum of about 10−6 Torr a nd t heir stoichiometry was well characterized and conrmed by XPS and EDAX. XRD in Bragg–Brentano focusing geometry was used for structural and residual stress analyses. Apart f rom t he i nherent a nisotropy, m any me tal io dides prefer certain (hkl)-oriented growth and a great affin ity for crystallinity (even at LNT) in thin lms. A typical case of lead iodide is illustrated in Figure 1.1 where the structural and strain analyses are su mmarized. Without a ny exception, PbI2  lms g row w ith (00l)-orientations (that is (00l) planes parallel or the c-axis perpendicular to the substrate plane) under any growth conditions including the low temperature. Th is is manifested by the appearance of only (00l) peaks in XRD as shown in Figure 1.1 even for ultrathin  lms. Films thinner than 20 nm were analyzed by transmission electron microscopy (TEM). Only representative micrographs are shown in the insets (a) and (b) for a 5 nm thick lm. The d-spacings (with t he M iller i ndices i ndicated o n t he

400

(106) (105)

(00l)-Oriented PbI2 Films

(001)

Film thickness 61 nm 93 nm

(201) (111) COUNTS (arb. units)

Counts

300

200

12

100

(b) (001)

12.5

13

13.5

2θ (degrees) 334 nm

(002)

(003)

(004)

0

(c)

Counts (arb. units)

254 nm 334 nm

5 nm thick film (a)

133 nm 173 nm

10

20

30

40

50

60

2q (⬚) 15 nm

(003)

(004)

30 nm 10

20

30

40

50

60

2q (⬚)

FIGURE 1.1 Typical XRD scan of (00l)-oriented ultrathin PbI2  lms for two  lm thicknesses as labeled in the gure. The inset (a) shows TEM for a 5 nm thick  lm and (b) is its electron diff raction. Inset (c) displays the XRD for very thick  lm and the inset within shows the shiĀ of (001) peak from unstrained value (ASTM data indicated by the broken line) for various  lm thicknesses.

43722_C001.indd 4

10/3/2008 8:42:16 PM

1-5

Residual Stress in Thin Films

0.0075

18

Grain size (nm)

Ultrathin PbI2 films

0.005

0.0025

by TEM 13

by XRD

8 3

∆d/d

0

5

10

15

20

25

30

Film thickness (nm)

0

−0.0025

by TEM by XRD

−0.005

2

6

10

14 18 22 Film thickness (nm)

26

30

FIGURE 1.2 Development of s train e z in ultrathin PbI2  lms as a f unction of  lm thickness. Inset shows a l inear grain size growth with  lm thickness. The method of strain and grain size determination is identied in the gure.

micrograph) calculated from the radius of the rings match quite well with ASTM card No. 07-0235 for 2H polytype of PbI2. The unstrained value of dhkl is taken from the above ASTM data for determining the strain. The (001) peak for thicker  lms are displayed i n t he i nset w ithin i nset (c) de picting t he s train, ei ther tensile or compressive, in comparison with the ASTM value indicated by the broken vertical line. Strain is almost negligible in  lms thinner than 12 nm as depicted in Figure 1.2, which shows the development of strain with  lm thickness. The functional dependence of strain on grain size can be expected to be similar to that on lm thickness because of the linear relationship b etween t he t wo a s shown i n t he i nset. Q ualitatively, t his seems to b e c onsistent w ith t he c oalescence mo del men tioned earlier a s g rains a re well s eparated i n u ltrathin  lms (T EM i n inset (a)) and start coalescing in thicker lms, causing the biaxial stress responsible for the observed strain. However, quantitative estimation of t he s tress re quires more i nformation. Ther efore, the development of biaxial stress can be expected, in general, for any  lm t hicker t han 1 0 nm. The i mportant c onsequence a nd necessity o f t he s train a nalysis i n u ltrathin Pb I2  lms i s h ighlighted and demonstrated [49], especially while determining the quantum connement contribution to the change in optical features. Results could lead to w rong conclusions if one ignores or overlooks t he e ffect of strain on t he opt ical properties. Re sults for t hicker Pb I2  lms a re d isplayed i n F igure 1 .3 f or v arious experimental parameters. The data shown are for lms grown at RT and LNT. Interestingly, the behavior of strain is exactly opposite to e ach ot her i n t he t wo c ases, most l ikely due to t he different thermal stress contribution to the total stress. However, the variation of strain w ith  lm thickness, annealing temperature, a nd t ime a s s hown i n F igure 1 .3a, b , a nd c , re spectively, may not be explained in a simple manner. Again, the qualitative explanations may have to b e der ived f rom t he coalescence a nd grain boundary models with required modications and inputs.

43722_C001.indd 5

For i nstance, t he t hickness de pendence o f s train re quires t he knowledge of g rain si ze, g rain den sity, a nd de fect den sity, e tc. for understanding the grain boundary structure. Switching from one type of biaxial stress to t he other with  lm thickness could possibly b e a n e xplanation f rom t heir m icrostructure. That is, different sx and sy could lead to this kind of strain variation with  lm thickness. For example, strain is negligible for a 63 nm thick  lm, which indicates very small biaxial stress, possibly resulting from the opposite type of the two components. Similarly, it may be p ossible to ac count f or t he ob served s train b y t he h elp o f microstructural analyses. Annealing is normally carried out to facilitate t he rel axation o f t he re sidual s tress. H owever, i t m ay not ne cessarily re duce t he s tress b ecause t he c hanging g rain boundary, microstructure, defect density, etc. could increase the stress as well in some cases. Figure 1.3b displays the effect of annealing temperature on the strain. The wavy nature is difficult to explain just like the one with annealing duration as shown in Figure 1.3c. Nevertheless, data need to be analyzed more carefully u sing ot her p ossible m acrostructural de tails ob tained o n the s ame s ample f or a b etter u nderstanding o f suc h b ehavior. The above discussion illustrates the t ypical intricacies inherent in the analyses of the residual stress origin. Such a detailed study is difficult to nd in the literature. The s tress b ehavior diff ers f or d ifferent m aterials. A nother material, t he z inc io dide h as a d ifferent st rain de pendence o n  lm thickness as shown in Figure 1.4. The strain in d104 is quite large c ompared to f ew ot her p eaks o f c onsiderable i ntensity using XRD. ZnI2  lms are also not well oriented. Ther efore, the average of strains of a few intense peaks is determined and plotted along with that of (104) in the gure. Although the comparative qualitative behavior of the two materials with lm thickness is not very much different, t heir e ffect o n opt ical p roperties i s entirely different [50], which we will see little later. The increasing and s aturating s train w ith t he  lm t hickness c ould si mply b e

10/3/2008 8:42:17 PM

1-6

Smart Materials 0.006 (00l)-Oriented PbI2 films

Grown at LNT

0.004 0.002

∆d/d

0 −0.002 −0.004 −0.006 −0.008

Grown at RT 0

100

200

300

400

Film thickness (nm)

(a)

93 nm Thick Pbl2 film annealed in vacuum for 2h

0.006 0.004

Grown at RT

∆d/d

0.002 0 −0.002 −0.004 −0.006 Grown at LNT −0.008 20

70

120 170 220 Annealing temperature (⬚C)

(b)

270

320

0.008 174 nm Thick Pbl2 film annealed at 200⬚C in vacuum

0.006

Grown at LNT

0.004

∆d/d

0.002 0 −0.002 −0.004 Grown at RT

−0.006 −0.008 0 (c)

1

2 3 Annealing time (h)

4

FIGURE 1.3 Residual strain in thicker PbI2  lms grown at R T and LNT as the function of (a)  lm thickness, (b) annealing temperature, and (c) annealing time.

43722_C001.indd 6

10/3/2008 8:42:17 PM

1-7

Residual Stress in Thin Films

Znl2 films

0.018 Znl2 films

0.028

∆d/d

∆d/d

0.022

0.014

0.024

(Average) (104)

0.02 0

100

200

Treatment temperature (⬚C)

0.01 200

300

400 500 Film thickness (nm)

600

700

FIGURE 1.4 Average strain (average of fe w i ntense peaks of X RD) a nd strain i n d104 a s t he f unction of t hickness for Z nI 2 lms. The inset shows the dependence of these quantities on annealing temperature (annealed in vacuum for 1 h).

understood to mimic the grain density behavior with lm thickness [51,52]. The inset of Figure 1.4 shows the effect of vacuum annealing or heat treatment temperatures on the strain. Strain is found to i ncrease w ith t reatment i n b oth t he qu antities, h owever, more dramatically for strain in (104). This increase could be due to t he ad sorption o f re sidual ga ses p resent i n t he v acuum chamber maintained in the range 10−4–10−5 Torr during annealing. The uctuation i n t he v acuum le vel i s d ue to de gassing caused by the prolonged heating in the chamber. Another different type of strain dependence on  lm t hickness i s s hown i n Figure 1 .5 f or a ( 00l)-oriented C dI2  lm as w ell as (102) a nd

(002)-oriented HgI2 lms. The t ypical t hickness dependence of strain in CdI 2 i s not su rprising a nd c an b e e xplained b y t he increasing grain density with  lm thickness. Even the decreasing strain in CdI2  lms with annealing temperature as shown in the inset is very much like that one can ideally expect [53]. This opposite behavior compared to that of ZnI2 in the inset of Figure 1.4 is mainly responsible for t heir basic microstructural d ifference, t hat i s, C dI2  lms a re c ompletely oriented a s opp osed to randomly or iented Z nI2  lms. H owever, t he in creasing s train with substrate temperature as depicted in the inset is hard to explain by thermal stress alone because the substrate temperature

0.0035 Cdl2 films

∆d/d

0.002

Substrate

0.001

Hgl2 films (002)

Annealing

∆d/d

0 100

0

0.0025

200

Temperature (⬚C)

Cdl2 films Hgl2 films (102)

0.0015 0

500

1000 1500 Film thickness (nm)

2000

2500

FIGURE 1.5 Film thickness dependence of strain for (00l)-oriented CdI2  lms and (102), (002)-oriented HgI2 lms. The inset shows the dependence of strain on annealing and substrate temperatures for (00l)-oriented CdI2 lms.

43722_C001.indd 7

10/3/2008 8:42:17 PM

1-8

Smart Materials 0.0025

0.03 Znl2 films

∆d/d

∆d/d

(104) 0.0015

0.02

Average 0.01 50

100 150 200 Grain size (nm)

0.0005 0

4

8

12

Grain size (µm)

FIGURE 1.6 Dependence of strain on grain size for (00l)-oriented CdI2 lms. The inset shows the average strain and strain in d104 as a function of grain size for ZnI2 lms.

is known to affect the grain boundary structure as well. Another example i n t he  gure, HgI2, h as a si mple l inear s train de pendence on lm thickness, although it differs substantially both in magnitude a nd slope for t he t wo d ifferent g rowth or ientations [54]. It will still be difficult to speculate or predict the origin and behavior of the stress with certainty in either completely or randomly o riented  lms o f a ny m aterial, e specially, t he d ielectric and compound semiconductors etc. Nevertheless, the strain data will still be very useful in understanding its effect on the various physical properties of t he m aterial. The e xplicit d ependence of strain on the grain size is illustrated for CdI2  lms in Figure 1.6 and for Z nI2  lms i n t he i nset of t he  gure. The strain dependence of C dI2 on t he g rain si ze s eems to b e qu alitatively ide al and as expected according to the model of increasing grain density, especially in oriented lms [53]. Normally, in such cases, the grain s ize v aries lin early w ith  lm t hickness a nd t herefore a similar strain dependence on  lm t hickness c an b e e xpected. Actually, t his t rend i s i ndicated i n t he l imited re gion o f  lm thickness o f C dI2 in Figure 1.5. However, the grain size data shown i n F igure 1 .6 i ncludes t he re sults f rom  lm thickness, annealing, a nd sub strate h eating e xperiments. The mo re p uzzling and strange result is the strain dependence of ZnI2 shown in the inset where data for both average and the (104) plane are shown a s d iscussed a bove. A ny re asonable e xplanation o f t his trend is difficult. The unexpected behavior could well originate from the random orientation of grains and the minimum may be occurring f or opt imal c lose-packing g rain si ze. I t i s o nly o ne more e xample i llustrating t he v ariation i n d iversied strain behavior. A s we have s een s o f ar f rom t hese l imited e xamples, the origin and behavior of stress is fairly simple in some materials a nd a re e qually d ifficult a nd u npredictable f or s ome ot her materials.

43722_C001.indd 8

We will see how much essential and useful the stress analysis is in t hin  lm s tudy f or u nderstanding t he a ffected physical properties. The a bove e xamples o f s train f or v arious m aterials will be extended to i nterpret the optical properties of a re spective material. Analysis becomes fairly simple for oriented  lms like P bI2. I n c-axis o riented  lms, t he s train a long t he z-axis resulting from the xy-plane biaxial stress can be treated as equivalent to the effect of uniaxial stress along the z-axis. This would certainly aff ect t he ba nd s tructure to a me asurable e xtent i n optical properties, at least in layer structured materials. Such an analysis is illustrated in Figure 1.7 for PbI2. The various known excitonic peaks of PbI2 observed in the optical absorption spectra a re l abeled i n t he i nset of t he  gure for c onvenience. B oth XRD a nd opt ical a bsorption me asurements w ere do ne o n t he same sample for t he reliability of t he analyses. Excitonic peaks show a reasonably good and consistent linear dependence on strain, either positive or negative. Slightly different slopes of the peaks can be attributed to the extent of distortion of the atoms in the u nit c ell, w hich a re re sponsible f or t he o rigin o f t he c oncerned p eak. The f urther de tails o f t he a nalysis i n v iew o f t he band structure of PbI2 can be found in the recent report [55]. The important p oint h ere i s t he s train i nduced energ y s hift o f the optical peaks and the magnitude of the shiĀ compares with that due to quantum connement. Therefore, the true picture of quantum connement effect in PbI2 is possible only if the strain contribution is separated [55]. Many such minute and important details could be obtained through a very careful stress analyses. Similar analyses for oriented CdI2 and HgI2 lms are shown in Figure 1.8. B oth t hese c ompounds d id not s how a ny e xcitonic peaks at R T a nd t herefore, t he opt ical ba nd gap Eg wa s de termined u sing s tandard me thod f rom t heir a bsorption s pectra [53,54]. The s train de pendence of Eg for C dI2 i s not l inear a nd

10/3/2008 8:42:18 PM

1-9

Residual Stress in Thin Films

Peak C

3

Pbl2 films Peak C

2.2

Peak B Peak B

Absorption

Exciton peaks’ energy (eV)

3.3

1.7 Peak A

1.2

2.7

93nm 133nm 174nm Film thickness

0.7 330

2.4 −0.016

430

Wavelength λ (nm)

−0.012

530

−0.008

Peak A

−0.004 ∆d/d

0

0.004

0.008

FIGURE 1.7 Strain dependence of three excitonic peaks of PbI2 lms. The three excitonic peaks are identied on the optical absorption spectra shown in the inset. The inset also depicts the shiĀ in peaks for few  lm thicknesses due to strain.

unusual as can be seen from the gure. It suggests an important fact that there exists a threshold strain up to which Eg is slightly affected a nd de creases s harply b eyond t he t hreshold [56]. This threshold lies on the side of tensile strain and it increases with a small slope on the compressive side [57]. The compressive strain was produced in the sample by energetic ion irradiation [57]. The effect o f s train o n (102) a nd (002) o riented Hg I2  lms i s qu ite opposite both in the magnitude and the slope [54]. A f ew more examples of oriented growths are illustrated by XRD in Figure 1.9 for BaI2 a nd SrI2  lms. As mentioned earlier, t he presence of a very intense single peak indicates the (hkl)-orientation parallel

to the substrate plane. Furthermore, such different orientations exhibit a different texture or morphology, which can be observed by scanning electron microscopy (SEM). Thus, bo th Ba I2 an d SrI2  lms tend to grow in three different orientations under certain g rowth c onditions a s s hown i n F igure 1 .9. F or t hese t wo materials too, the strain dependence of Eg of respective orientation differs signicantly from each other as shown in Figure 1.10. Here also, Eg was determined from their absorption spectra and no e xcitonic p eak w as v isible at R T. These c ould b e i mportant results in understanding the band structure and probably can be veried by the density functional (DF) calculations. Finally, the

3.6 Cdl2 films

E g (eV)

3

Hgl2 films (002)

2.4 Hgl2 films (102) 1.8 0.0005

0.0015

∆d/d

0.0025

0.0035

FIGURE 1.8 Strain-dependent optical band gap Eg for (00l)-oriented CdI2  lms and (102), (002)-oriented HgI2 lms.

43722_C001.indd 9

10/3/2008 8:42:18 PM

1-10

Smart Materials

(hkl)-Oriented Bal2 films

Counts (arb. units)

(015)

(113)

(301)

5

15

25 2q (⬚)

(a)

35

45

(022)

Counts (arb. units)

(hkl)-Oriented Srl2 films

(202)

(210)

10 (b)

15

20

25

30

35

2q (⬚)

FIGURE 1.9 XRD scans depicting the various (hkl)-oriented lm growth for (a) BaI2 lms and (b) SrI2 lms. The growth orientations are labeled on each scan.

example of randomly oriented polycrystalline ZnI2  lms is summarized in F igure 1.11. The a bsorption s pectra o f t he v arious heat-treated samples are displayed in Figure 1.11a, which denotes various optical features. Peaks B and C are identied as excitonic peaks a nd t he de termination o f Eg i n suc h a c ase i s de scribed

[50]. All these three quantities are to scale with the strain in d104 as shown in Figure 1.11b and do not show any systematic dependence on the average strain. In contrast, peak A does the opposite as shown in Figure 1.11c. This is the reason why the strain in (104) and the average strain are separately determined as shown

(hkl )-Oriented Bal2 films

(015) 3.5

Eg (eV)

(113)

(301)

3.3

(301) 3.1 (015) (113) 2.9 −0.008 (a)

FIGURE 1.10

43722_C001.indd 10

−0.006

−0.004

−0.002 ∆dhkl /dhkl

0

0.002

Strain dependence of optical band gap Eg for various (hkl)-orientations in (a) BaI2 lms

10/3/2008 8:42:18 PM

1-11

Residual Stress in Thin Films 3.8 (hkl)-Oriented Srl2 films

3.6

Eg (eV)

(202) 3.4 (210) (022) (210)

3.2

(022) (202) 3 −0.02

−0.015

−0.01

−0.005 ∆dhkl /dhkl

(b)

0.005

0.01

(b) SrI2  lms. The growth orientations are labeled in the gure.

Znl2 films B

A

C

250⬚C

a (cm−1)

150⬚C 200⬚C

Exc. peak B, exc. peak C & Eg (eV)

FIGURE 1.10 (continued)

0

4 Znl2 films 3.6 Exciton peak B Exciton peak C Eg 3.2

2.8 0.02

0.025 ∆d104 /d104

(b) 4.3

50⬚C

Eg RT

3.1

3.4 3.7 4.0 4.3 h ν (eV)

Peak A (eV)

100⬚C

2.5 2.8 (a)

0.03

4.2

4.1 0.01

(c)

0.014

0.018

0.022

∆d/d (average)

FIGURE 1.11 Summary of strain-dependent optical properties of ZnI2 lms: (a) Optical absorption spectra of 470 nm thick lm for various annealing (in vacuum) temperatures as labeled on e ach spectrum. The various optical features like excitonic peaks B a nd C, peak A, and the band gap Eg are identied on the spectra, (b) dependence of peaks B, C, and Eg on the strain in d104, and (c) the dependence of peak A on the average strain.

in Figure 1.4. Thus, the carefully carried out residual stress analysis a long w ith ot her c haracterizations c ould h elp i n b etter understanding the material. These f ew d iversied e xamples of opt ical prop erties a nalyzed in view of the stress demonstrate the importance of stress analyses in lms. In fact, a number of other physical properties also depend on stress and their stress dependence can reveal many microscopic details. For instance, the Hall resistance of

43722_C001.indd 11

semiconductors, critical temperature of superconducting lms, ferromagnetic re sonance i n f erromagnetic  lms, d ielectric constant o f d ielectric  lms, a nd t he l ikes c an b e e xpected to exhibit t he s train de pendence. C onversely, t he s hiĀ in t hese relevant quantities may be used to estimate the order of magnitude of the stress present in the  lm provided the exact stress dependence of the concerned quantity is known. The linear stress dependence of an easily and conveniently measurable

10/3/2008 8:42:19 PM

1-12

Smart Materials

physical quantity of any material can qualify it as a smart material for strain gauge applications. The main reason for the wide data range of various physical quantities reported in the literature for the lms of the same material could be due to ignoring the stress in the lm. One such example is the optical band gap of a p articular s emiconductor, which mostly d iffers f rom one report to t he ot her for t hick  lms a s much a s i n t he r ange of 1 eV where stress is normally ignored. Without the knowledge of t he in uence o f s tress, i t i s d ifficult to de termine t he t rue intrinsic value of t he ba nd gap w hich is normally used as t he reference f or ba nd s tructure c alculations. Si milar si tuations exist f or a f ew o f t he ot her ph ysical qu antities a s w ell a s f or which stress effect is quite signicant.

1.5

Summary and Future Direction

It i s a h ard t ruth t hat a ny t hin  lm i s not c ompletely f ree of residual s tress. This fact makes us learn to live with stress by understanding its nature and behavior. Therefore, we practically we c annot a fford to a void o r ig nore t he s tress c ompletely. A n appropriate, affordable, and convenient method of stress analysis can be adopted out of several available choices from t he literature. The i nformation a bout t he s tress i s a lready t here i n many o f t he ro utine s tructural c haracterization te chniques, which have to b e extracted by careful analyses as illustrated in this b rief ac count. I ndeed, a t ruly u niversal a nd qu antitative theory for the internal residual stress in  lms is perhaps yet to be developed if there could be any. The entire data in the literature and t heir d ifferent possible explanations make it doubtful that such a unique theory valid for different  lm materials and different methods of deposition will ever emerge. The main stumbling blocks f or suc h a n at tempt a re s ome o f t he i nherent p roblems and uncertainties of lms and their growth processes. Film stoichiometry ( or atom ic c omposition), s tructural d efects, g rainboundary st ructure,  lm–substrate i nterfacial p roperties e tc. are some of the difficulties in describing them in terms of macroscopic elastic models. Nevertheless, the study of residual stress in t hin  lms w ill c ontinue to b e i mportant a nd ne cessary f or exploiting them in various applications even if the origin cannot be traced convincingly. Sooner or later, the residual stress analyses in lms will draw more attention and be consolidated in nding further importance and priority in materials research.

References 1. E. J . Mills, On ele ctrostriction, Proc. Ro y. S oc. Lo ndon 26, 504–512, 1877. 2. R. W. Hoffman, The me chanical p roperties o f t hin  lms, in Ā e U se o f Ā in F ilms i n Ph ysical I nvestigations, J . C. Anderson (ed.), Academic Press, New York, p. 261, 1966. 3. R. W. Hoffman, M echanical p roperties o f t hin co ndensed lms, in Physics of Ā in Films, Vol. 3, G. Hass and R. E. Thun (eds.), Academic Press, New York, p. 211, 1966. 4. W. J . P arak, L. M anna, F . C. S immel, D . G erian, a nd P. Alivisatos, Quantum dots, in Nanoparticles: From Āeor y to

43722_C001.indd 12

5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18.

19. 20. 21.

22.

23.

Application, G. Schmid (ed.), Wiley-VCH Verlag, Weinheim, p. 27, 2004. R. Notzel, Self-organized growth of quantum-dot structures, Semicond. Sci. Technol. 11, 1365–1379, 1996. D. S. C ampbell, M echanical p roperties o f t hin  lms, in Handbook of Ā in Film Technology, L. I. Maissel and R. Glang (eds.), McGraw-Hill, New York, 1970. G. G. Stoney, The tension of metallic lms deposited by electrolysis, Proc. Roy. Soc. London A82, 172–175, 1909. W. Buckel, Internal stresses, J. Vac. Sci. Tech. 6, 606–609, 1969. E. Klokholm and B. S. Berry, Intrinsic stress in e vaporated metal lms, J. Electrochem. Soc. 115, 823–826, 1968. R. W. Hoffman, Stresses in thin lms: The relevance of grain boundaries a nd im purities, Ā in S olid Fi lms 34, 185–190, 1976. W. D . N ix a nd B . M. Clemen s, Cr ystallite coales cence: A mechanism for intrinsic tensile stresses in thin lms, J. Mater. Res. 14, 3467–3473, 1999. G. Guisbiers, O. Van Overschelde, and M. Wautelet, Nanoparticulate o rigin o f intrinsic r esidual st ress in t hin  lms, Acta Mater. 55, 3541–3546, 2007. E. Chason, B. W. Sheldon, L. B. Freund, J. A. Floro, and S. J. Hearne, Origin of compressive residual stress in p olycrystalline thin lms, Phys. Rev. Lett. 88, 156103(1–4), 2002. H. Windischmann, Intrinsic stress in sputter-deposited thin lms, Crit. Rev. Solid State Mater. Sci. 17, 547–596, 1992. Y. Pauleau, Generation and evolution of residual stresses in physical vapour-deposited thin lms, Vacuum 61, 175–181, 2001. C. A. Davis, A simple model for the formation of compressive st ress in t hin  lms b y io n b ombardment, Ā in Solid Films 226, 30–34, 1993. M. Itoh, M. Hori, and S. Nadahara, The origin of st ress in sputter-deposited tungsten lms for x-ray masks, J. Vac. Sci. Technol. B9, 149–153, 1991. J. -D . K amminga, T . H. de K eijser, R . D elhez, a nd E. J . Mittemeijer, A mo del f or st ress in thin la yers ind uced b y mistting particles: An origin for growth stress, Ā in Solid Films 317, 169–172, 1998. E. H. Hirsch, Stress in porous thin lms through adsorption of polar molecules, J. Phys. D13, 2081–2094, 1980. D. A. Glecker and S. I. Shah, Handbook of Ā in Film Process Technology, IOP, Bristol and Philadelphia, 1995. R. E. Cuthrell, D. M. Mattox, C. R. Peeples, P. L. Dreike, and K. P. Lamppa, Residual stress anisotropy, stress control, and resistivity in p ost ca thode ma gnetron sp utter dep osited molybdenum  lms, J. Vac. S ci. T echnol. A6, 2914–2920, 1988. C. H. Wu, W. H. Weber, T. J. Poter, and M. A. Tamor, Laser reective interferometry for in situ monitoring of diamond lm growth by chemical vapor deposition, J. Appl. Phys. 73, 2977–2982, 1993. M. Sternheim, W. van G elder, and A. W. Hartman, A las er interferometer system to mo nitor dry etching of patterned silicon, J. Electrochem. Soc. 130, 655–658, 1983.

10/3/2008 8:42:19 PM

Residual Stress in Thin Films

24. J. K empf, M. N onnenmacher, a nd H. H. Wagner, Ele ctron and ion beam induced heating effects in solids measured by laser interferometry, Appl. Phys. A56, 385–390, 1993. 25. J. Kempf, Optical in situ sputter rate measurements during ion sputtering, Surf. Interf. Anal. 4, 116–119, 1982. 26. G. J. Leusink, T. G. M. Oosterlaken, G. C. A. M. Janssen, and S. Radelaar, In situ sensitive measurement of stress in thin lms, Rev. Sci. Instrum. 63, 3143–3146, 1992. 27. T. Aoki, Y. Nishikawa, and S. Kato, An improved optical lever technique f or me asuring  lm stress, Jpn. J. Appl. Phys. 28, 299–300, 1989. 28. S. Suresh and A. E. Giannakopoulos, A new method for estimating residual stresses by instrumented sharp indentation, Acta Mater. 46, 5755–5767, 1998. 29. M. Herrmann, N. Schwarzer, F. Richter, S. Frühauf, and S.E. Schulz, Determination of Young’s modulus and yield stress of p orous low-k materials by na noindentation, Surf. C oat. Technol. 201, 4305–4310, 2006. 30. R. Ikeda, T. Uchiyama, H. Cho, T. Ogawa, and M. Takemoto, An advanced met hod for me asuring t he residual st ress of deposited  lm u tilizing las er sp allation te chnique, Sci. Technol. Adv. Mater. 7, 90–96, 2006. 31. K. T. V. G rattan, B . T. M eggitt (e ds.), Optical F iber S ensor Technology, Chapman and Hall, London, 1995. 32. S. M. M. Quintero, W. G. Quirino, A. L. C. Triques, L. C. G. Valente, A. M. B. Braga, C. A. Achete, and M. Cremona, Thin lm st ress me asurement b y  ber o ptic stra in ga ge, Ā in Solid Films 494, 141–145, 2006. 33. A. Rizzo, L. C apodieci, D. Rizzo, and U. Galietti, L ow cost technique f or me asuring in si tu st rain o f na nostructures, Mater. Sci. Eng. C 25, 820–825, 2005. 34. J. K. Shin, C. S. Lee, K. R. Lee, and K. Y. Eun, Effect of residual stress o n the R aman-spectrum a nalysis o f t etrahedral amorphous carbon lms, Appl. Phys. Lett. 78, 631–633, 2001. 35. A. C. F errari a nd J . Rob ertson, I nterpretation o f R aman spectra of disordered and amorphous carbon, Phys. Rev. B61, 14095–14107, 2000. 36. J. Zh u, J . H an, A. Li u, S. M eng, a nd C. J iang, M echanical properties and Raman characterization of amorphous diamond lms as a function of lm thickness, Surf. Coat. Technol. 201, 6667–6669, 2007. 37. B. D . C ullity, Elements o f X-r ay Di ffraction, 2nd e dn., Addison-Wesley, Reading, MA, 1978. 38. H. P. Klug and L. E. Alexander, X-ray Diffraction Procedures for Polycrystalline and Amorphous Materials, 2nd edn., John Wiley, New York, 1974. 39. D. Shindo and T. Oikawa, Analytical Electron Microscopy for Materials Science, Springer-Verlag, Tokyo, 2002. 40. A. E. Ennos, S tresses de veloped in o ptical  lm coatings, Appl. Opt. 5, 51–61, 1966. 41. H. K. Pulker, Coatings on Glass, Elsevier, Amsterdam, 1984.

43722_C001.indd 13

1-13

42. J. Zhang, K. Xu, and V. Ji, Dependence of stresses on grain orientations in thin p olycrystalline  lms on substrates: an explanation of the r elationship between preferred orientations and stresses, Appl. Surf. Sci. 180, 1–5, 2001. 43. C. Kisielowski, J. Kruger, S. Ruvimov, T. Suski, J. W. Ager III, E. J ones, Z. Lilien tal-Weber, M. R ubin, E. R . Weber, M. D. Bremser, and R. F. Davis, Strain-related phenomena in GaN thin lms, Phys. Rev. B54, 17745–17753, 1996. 44. R. F. D avis, S. Einfeldt, E. A. Preble, A. M. Roskowski, Z. J. Reitmeier, a nd P. Q. Mira glia, Galli um ni tride a nd r elated materials: c hallenges in m aterials p rocessing, Acta M ater. 51, 5961–5979, 2003. 45. E. Eiper, A. Hofmann, J. W. Gerlach, B. Rauschenbach, and J. Keckes, Anisotropic intrinsic and extrinsic stresses in epitaxial wurtzitic GaN thin  lm on γ-LiAlO2 (1 0 0), J. Cryst. Growth 284, 561–566, 2005. 46. C. Sarioglu, The effect of anisotropy on residual stress values and mo dication o f S erruys a pproach to r esidual st ress calculations for coatings such as T iN, Z rN and HfN, Surf. Coat. Technol. 201, 707–717, 2006. 47. M. A. Moram, Z. H. Barber, C. J. Humphreys, T. B. Joyce, and P. R. Chalker, Young’s modulus, Poisson’s ratio, and residual stress a nd stra in in (111)-o riented s candium ni tride thin lms on silicon, J. Appl. Phys. 100, 023514(1–6), 2006. 48. W. J. Zhang and S. Matsumoto, Investigations of cr ystallinity a nd r esidual st ress o f c ubic b oron ni tride  lms by Raman spectroscopy, Phys. Rev. B63, 073201(1–4), 2001. 49. V. G ulia, A. G. Vedeshwar, a nd N. C. M ehra, Qua ntum dot-like b ehavior o f ul trathin PbI 2 lms, Acta M ater. 54, 3899–3905, 2006. 50. A. G. Vedeshwar and P. Tyagi, Excitonic absorption in ZnI2 lms, J. Appl. Phys. 100, 083522(1–6), 2006. 51. P. Tyagi a nd A. G. Vedeshwar, Op tical p roperties o f Z nI2 lms, Phys. Rev. B64, 245406(1–7), 2001. 52. P. T yagi a nd A. G. Vedeshwar, Resid ual st ress dep endent optical properties of ZnI2 lms, Phys. Stat. Sol. (a) 191, 633– 642, 2002. 53. P. Tyagi and A. G. Vedeshwar, Grain size dependent optical properties of CdI2 lms, Eur. Phys. J. AP19, 3–13, 2002. 54. P. Tyagi and A. G. Vedeshwar, Anisotropic optical band gap of (102) and (002) oriented lms of red HgI2, Phys. Rev. B63, 245315(1–6), 2001. 55. V. G ulia a nd A. G. Vedeshwar, Op tical p roperties o f PbI 2 lms: Quantum connement and residual stress effect, Phys. Rev. B75, 045409(1–6), 2007. 56. P. Tyagi and A. G. Vedeshwar, Effect of residual stress on the Optical properties of CdI2 lms, Phys. Rev. B66, 075422(1–8), 2002. 57. R. S. R awat, P. Arun, A. G. Vedeshwar, P. L ee, a nd S. L ee, Effect o f ener getic io n irradia tion o n C dI2 lms, J. Ap pl. Phys. 95, 7725–7730, 2004.

10/3/2008 8:42:19 PM

43722_C001.indd 14

10/3/2008 8:42:20 PM

2 Intelligent Synthesis of Smart Ceramic Materials 2.1 I ntroduction ............................................................................................................................... 2-1 2.2 H ydrothermal Synthesis of Smart Ceramic Materials—An Overview ............................... 2-1

2.3

Wojciech L. Suchanek Sawyer Technical Materials

Richard E. Riman Rutgers University

2.1

Process De nition • Merits of Hydrothermal Synthesis of Ceramics • Chemical Compositions and Morphologies of Smart Ceramics • Hydrothermal Hybrid Techniques • Industrial Production of Smart Ceramic Materials

Intelligent Control of Phase Assemblage ................................................................................2-4 Construction of a Ā ermodynamic Model • Methodology for Generating Stability and Yield Diagrams • Utilization of Ā er modynamic Modeling

2.4 Intelligent Control of Crystal Size and Morphology............................................................. 2-7 Ā ermodynamic Variables • Non thermodynamic Variables

2.5 S ummary .....................................................................................................................................2-8 Acknowledgments ...............................................................................................................................2-8 References .............................................................................................................................................2-8

Introduction

Ceramic materials are being used in nearly all advanced materials applications either as pure ceramics or as ceramic components of composites or devices w ith metallic or organic constituents. Growing dem and f rom v arious i ndustries to s ynthesize sm art materials w ith mo re a nd mo re s ophisticated f eatures re quires the use of smart starting materials, w ith very well-dened and controlled specic properties, such as size, shape (morphology), chemical composition and defect structure, surface functionalization, e tc. Ā is in turn requires use of advanced synthetic methods, which, in addition to producing superior product, should b e a lso i nexpensive a nd en vironmentally f riendly. Ā e low-temperature hydrothermal synthesis meets all of the above requirements; mo reover, i t i s v ery w ell su ited to p roduce b oth smart ceramic starting materials, such as powders or bers, and already shaped products, such a s bulk ceramic pieces,  lms or coatings a nd s ingle cr ystals [ 1–6]. Ā e n umber o f s cientic papers on hydrothermal synthesis of materials has been steadily increasing since 1989 a nd is c urrently on the le vel of hundreds per year [2,3]. Ā e number of papers on hydrothermal synthesis of ceramic powders alone has nearly quadrupled between 2000 and 2004. Consequently, a l arge family of smart ceramics, primarily powders and coatings, has emerged that can be prepared under very mild hydrothermal conditions (T < 200°C, P < 1.5 MPa). Ā is has generated lots of c ommercial interest in hydrothermal technology [1–6].

2.2 2.2.1

Hydrothermal Synthesis of Smart Ceramic Materials—An Overview Process Defi nition

Hydrothermal synthesis is a process that utilizes single or heterogeneous phase reactions in aqueous media at elevated temperature (T > 25°C) and pressure (P > 100 kPa) to crystallize ceramic materials directly from solution [1,6]. Reactants used in hydrothermal s ynthesis a re u sually c alled p recursors, w hich a re administered in forms of solutions, gels, or suspensions. However, hydrothermal growth of single crystals requires in most cases use o f s olid n utrient, w hich re crystallizes d uring t he g rowth process. Mineralizers are organic or inorganic additives that are used to c ontrol pH or en hance solubility. Other add itives, a lso organic or inorganic, are used to s erve other functions, such as controlling crystal morphology, chemical composition, particle dispersion, etc. Syntheses are usually conducted at autogeneous pressure, which corresponds to t he saturated vapor pressure of the solution at the specied temperature and composition of the hydrothermal s olution. H igher p ressures up to o ver 5 00 MPa and temperatures over 1000°C [5] may be necessary to facilitate reactant d issolution a nd g rowth o f c ertain t ypes o f c eramic materials, a nd a re u sually app lied i n si ngle c rystal g rowth. Nevertheless, mild conditions are preferred for commercial processes where temperatures are less than 350°C and pressures less than approximately 50 MPa. Intensive research has led to a better 2-1

43722_C002.indd 1

10/3/2008 8:42:53 PM

2-2

Smart Materials

understanding o f h ydrothermal c hemistry, w hich h as sig nicantly reduced the reaction time, temperature, and pressure for hydrothermal c rystallization of c eramic m aterials, pre dominantly p owders a nd c oatings (T < 200°C, P < 1.5 MPa) [2,4,5,7]. Ā is b reakthrough h as m ade h ydrothermal s ynthesis mo re economical since processes can be engineered using cost-effective and p roven p ressure re actor te chnology a nd me thodologies already established by the chemical process industry.

2.2.2

Merits of Hydrothermal Synthesis of Ceramics

Hydrothermal s ynthesis o ffers ma ny adv antages o ver co nventional a nd no nconventional c eramic s ynthetic me thods. A ll forms of ceramics can be prepared with hydrothermal synthesis,

(a)

(c)

(e)

5 cm

5 µm

1 µm

such as powders,  bers, and single crystals, monolithic ceramic bodies, and coatings on metals, polymers, and ceramics (Figure 2.1). From the standpoint of ceramic powder production, there are far fewer time- and energy-consuming processing steps since hightemperature c alcination, m ixing, a nd m illing s teps a re ei ther not necessary or minimized. Moreover, the ability to precipitate already crystallized powders directly from solution regulates the rate a nd u niformity o f n ucleation, g rowth, a nd a ging, w hich results i n i mproved control of si ze a nd morphology of crystallites and signicantly reduced aggregation levels, that is not possible with many other synthesis processes [8]. Figure 2.2 shows several e xamples o f t he v arieties o f mo rphologies a nd pa rticle sizes p ossible w ith h ydrothermal p rocessing. Ā e elimination or re duction of a ggregates combined w ith na rrow pa rticle si ze distributions i n t he s tarting p owders le ads to opt imized a nd

(b)

(d)

(f)

10 nm

3 µm

6 µm

FIGURE 2 .1 Examples of v arious for ms of c eramic m aterials s ynthesized h ydrothermally: (a) α-quartz si ngle c rystal, ( b) c arbon n anotube, (c) PZT powder, (d) carbon bers, (e) epitaxial KNbO3  lm on SrTiO3 wafer, and (f) epitaxial KNbO3  lm on LiTaO3 wafer.

43722_C002.indd 2

10/3/2008 8:42:54 PM

2-3

Intelligent Synthesis of Smart Ceramic Materials

(a)

(b)

300 nm

(c)

200 nm

(d)

2 µm

5 µm

(e)

(f)

10 µm

10 µm

FIGURE 2 .2 Examples of v arious si zes a nd mor phologies of c eramic p owders s ynthesized h ydrothermally: ( a) e quiaxed n anosized Z nO, (b) nanosized hydroxyapatite needles, (c) LiMnO3 platelets, (d) carbon nanotubes, 50–100 nm in diameter, several microns in length, (e) hydroxyapatite whiskers, and (f) equiaxed α-Al2O3.

reproducible properties of ceramics because of better microstructure control. From the standpoint of thin lms (coatings), other methods such as physical vapor deposition, chemical vapor deposition, and sol–gel suffer from the disadvantage that they all require h igh-temperature processing to c rystallize t he ceramic phase. Ā ermally induced defects result, such as cracking, peeling, undesired reactions between the substrate and coating, and decomposition of the substrate material. In contrast, hydrothermal synthesis can be used to directly crystallize  lms on to substrate s urfaces a t l ow t emperatures, t hereby e nabling n ew combinations of materials such as ceramic coatings on polymer substrates. Hydrothermal processing can take place in a w ide variety of combinations o f aque ous a nd s olvent m ixture-based s ystems. Relative t o s olid-state p rocesses, li quids a ccelerate diff usion,

43722_C002.indd 3

adsorption, re action r ate, a nd c rystallization, e specially u nder hydrothermal conditions [3,7]. However, unlike many advanced methods that can prepare a large variety of forms and chemical compounds, such as chemical vapor-based methods, the respective costs for instrumentation, energy, and precursors are far less for h ydrothermal me thods. H ydrothermal me thods a re mo re environmentally b enign t han m any ot her s ynthesis me thods, which c an b e at tributed i n pa rt to energ y-conserving lo wprocessing temperatures, a bility to re cycle w aste, a nd s afe a nd convenient disposal of waste that cannot be recycled [3]. Ā e low reaction tem peratures a lso a void ot her p roblems en countered with high-temperature processes, particularly during single crystal growth, for example, poor stoichiometry control due to volatilization o f c omponents o r c rystal c racking d ue to ph ase transformations taking place during cooling.

10/3/2008 8:42:55 PM

2-4

Smart Materials

An important advantage of the hydrothermal synthesis is that the purity of hydrothermally synthesized materials signicantly exceeds the purity of the starting materials. Ā is is bec ause hydrothermal crystallization is a purication process in itself, in which the growing crystals or crystallites reject impurities present i n t he g rowth en vironment. M aterials s ynthesized u nder hydrothermal c onditions o ften e xhibit diff erences in p oint defects when compared to materials prepared by high-temperature s ynthesis me thods. F or i nstance, i n ba rium t itanate, hydroxyapatite, o r α-quartz, w ater-related l attice de fects a re among t he mo st c ommon i mpurities a nd t heir c oncentration determines essential properties of these materials. Ā e problem of water incorporation can be overcome by properly adjusting the synthesis c onditions, u se o f w ater-blocking add itives, o r e ven nonaqueous solvents (solvothermal processing). Another important technological advantage of the hydrothermal technique is its capability for continuous materials production, which can be particularly useful in continuous fabrication of ceramic powders [9].

2.2.3

Chemical Compositions and Morphologies of Smart Ceramics

A g reat v ariety o f c eramic m aterials h ave b een s ynthesized b y hydrothermal methods. Most common are oxide materials, both simple o xides, suc h a s Z rO2, T iO2, SiO 2, Z nO, F e2O3, A l2O3, CeO2, SnO2, Sb2O5, Co3O4, HfO2, etc., and complex oxides, such as B aTiO3, SrTiO3, PZ T, PbTiO3, K NbO3, K TaO3, L iNbO3, f errites, apatites, t ungstates, v anadates, molybdates, zeolites, e tc., some o f w hich a re me tastable c ompounds, w hich c annot b e obtained using classical synthesis methods at high temperatures. Hydrothermal synthesis of a variety of oxide solid solutions and doped compositions is common. Ā e hydrothermal technique is also well suited for nonoxides, such as pure elements (for example Si, Ge, Te, Ni, diamond, carbon nanotubes), selenides (CdSe, HgSe, CoSe2, NiSe2, CsCuSe4), tellurides (CdTe, Bi2Te3, Cu xTey, AgxTey), suldes (CuS, ZnS, CdS, PbS, PbSnS3),  uorides, nitrides (cubic BN, hexagonal BN), aresenides (InAs, GaAs), etc. [2,4,10–12]. Crystalline products with a specic chemical or ph ase composition c an b e u sually s ynthesized h ydrothermally i n s everal different forms, such as single crystals, coatings, ceramic monoliths, or powders. Among them, the powders exhibit the largest variety of morphologies, suc h a s e quiaxed (for e xample c ubes, spherical), elo ngated ( bers, w hiskers, na norods, na notubes), platelets, nanoribbons, nanobelts, etc., with sizes ranging from a few nanometers to tens of microns (Figure 2.2). Core–shell particles and composite powders consisting of a mix of at least two different p owders c an b e a lso p repared i n o ne s ynthesis s tep. Some of the powders can even adopt nonequilibrium morphologies (Figure 2.4c and d).

2.2.4 Hydrothermal Hybrid Techniques In o rder to add itionally en hance t he re action k inetics o r t he ability to m ake ne w m aterials, a g reat a mount o f w ork h as been do ne to h ybridize t he h ydrothermal te chnique w ith

43722_C002.indd 4

microwaves ( microwave–hydrothermal p rocessing), e lectrochemistry (hydrothermal–electrochemical synthesis), ultrasound (hydrothermal–sonochemical s ynthesis), m echanochemistry (mechanochemical–hydrothermal synthesis), optical radiation (hydrothermal–photochemical s ynthesis), a nd h ot-pressing (hydrothermal hot pressing), as reviewed elsewhere [1,3,5–7].

2.2.5

Industrial Production of Smart Ceramic Materials

Several h ydrothermal te chnologies, p rimarily f or t he p roduction o f si ngle c rystals, suc h a s α-quartz f or f requency c ontrol and opt ical app lications ( Sawyer T echnical M aterials, T okyo Denpa, N DK), Z nO f or U V- a nd bl ue l ight-emitting de vices (Tokyo Denpa), and KTiOPO4 for nonlinear optical applications (Northrop Grumman-Synoptics), have a lready b een developed that demonstrate the commercial potential of the hydrothermal method. Ā e volume of the hydrothermal production of α-quartz single crystals is estimated at 3000 tons/year [2]. However, the largest potential g rowth a rea for commercialization is ceramic powder production. Ā e widely used Bayer process uses hydrothermal methods to dissolve bauxite and subsequently precipitate a luminum h ydroxide, w hich i s l ater he at-treated at h igh temperature to crystallize as α-alumina. In 1989, the worldwide production rate was about 43 million tons/year. Ā e production of pe rovskite-based d ielectrics a nd zi rconia-based st ructural ceramics is a promising growth area for hydrothermal methods [9]. C orporations suc h a s C abot C orporation, S akai C hemical Company, M urata I ndustries, Fe rro C orporation, S awyer Technical M aterials, a nd ot hers h ave e stablished c ommercial hydrothermal pr oduction pr ocesses for pre paring c eramic powders.

2.3

Intelligent Control of Phase Assemblage

Ā ermodynamic modeling can be used to i ntelligently design a process to b e t hermodynamically f avored u sing f undamental principles i nstead o f t he t ime-consuming E disonian me thods [1,13]. For a given hydrothermal synthesis of ceramic, the effects of precursor or additive concentrations, temperature, and pressure can be modeled to dene the processing variable space over which t he ph ases o f i nterest a re s table. Ā e thermodynamic modeling c an b e ac complished u sing c ommercially a vailable OLI computer soft ware, which also contains a database of thermodynamic properties for many common systems [1,13].

2.3.1

Construction of a Thermodynamic Model

In order to construct a thermodynamic model, one has to know all possible species a nd i ndependent chemical reactions occurring in the system being investigated. Even a simple hydrothermal system such as Ba–Ti–K–H2O, which is used to synthesize BaTiO3

10/3/2008 8:42:57 PM

2-5

Intelligent Synthesis of Smart Ceramic Materials

ceramics, contains a signicant number of species and independent chemical reactions. Determining t he equilibrium concentrations f or e ach o f t he s pecies re quires t he u se o f a utomated computer-based s olution a lgorithms. St ability d iagrams c oncisely re present t he t hermodynamic s tate o f m ulticomponent,

log10 [m(Pb(acet)2) = 1.5m(NbCl5) = 3m(Mg(NO3)2)]

0

2

0

4

T = 200⬚C (473 K) 6 8

Pb3Nb4O13 stable

−0.8

12

0

−0.4

KNbO3 + Mg(OH)2 + Pb3Nb4O13

Mg(OH)2 stable Pb3Nb4O13 stable

KNbO3 stable Mg(OH)2 stable

All aqueous species

−1.2

10

Nb2O5 + Pb3Nb4O13

Nb2O5 −0.4

multiphase aqueous systems in wide ranges of temperature and reagent (precursor) concentrations (Figure 2.3). Yield d iagrams specify the synthesis conditions suitable for quantitative precipitation of the phase of interest (Figures 2.3b and 2.4a). Calculations of stability a nd y ield d iagrams a re ba sed on a t hermodynamic

−1.6

−2 14

−2 (a)

4

2

6 m (KOH)

8

10

T = 160⬚C (433 K)

log10 [m(TiO2) = m(Ba(OH2) = 10m (CO2)]

0

0

2

4

6

8

−1.6

−1.2

−1.6

Mg(OH)2

0

−0.8

10

12

Yield = 0.99 12

14 0

−1.6

BaCO3{s}

−3.2

−3.2 BaTiO3{s}

−4.8

−4.8

Aqueous species

−6.4

−8 (b)

−6.4

−8 0

2

4

6

8

10

12

14

pH

FIGURE 2.3 (a) Calculated stability diagram for t he Pb–Mg–Nb–K–H2O system at 2 00°C where input precursor concentration is plotted as a function of m ineralizer (KOH) c oncentration. Ā e s ymbols d enote e xperimentally o btained ph ase a ssemblages c orresponding to t he re action conditions specied by t he equilibrium diagram:
Mg(OH)2 + Pb 3Nb4O13,  KNbO3 + Mg(OH)2. (b) Calculated stability/yield diagram for t he Ba–Ti–CO2–H 2O system at 160°C using Ba(OH)2 and TiO2 (rutile) when the amount of CO2 in precursors is 0.1 times the amount of TiO2 and the Ba/Ti ratio is equal to 1. Ā e solid lines denote the incipient precipitation boundaries for BaTiO3 and BaCO3. Ā e shaded area corresponds to the BaTiO3 yield >99%.

43722_C002.indd 5

10/3/2008 8:42:57 PM

2-6

Smart Materials All aqueous species 0

PbO + PZT

12

Average particle size (µm)

PZT 70/30, yield > 99% -0.8 • log (Ti)

TiO2 precursor 6m KOH, 0.33m PZT, 150⬚C, 24h

11

PbO

−1.6

• 0% < Yield < 99%

−2.4

No PZT

10

ZTO precursor 1m TMAH, 0.11m PZT, 150⬚C, 24h

9 8 7 6 5 4 3 2 1

−3.2 (a)

0

0

2

4

6 8 10 [KOH] (mol/k g H2O)

(c)

12

14

0

(b)

5 µm

200

400

600

800

1000 1200 1400 1600 1800

Stirring speed (rpm)

(d)

5 µm

FIGURE 2.4 Example of intelligent control of chemical composition, size, and morphology of PZT crystallites using thermodynamic and nonthermodynamic variables. (a) Calculated stability/yield diagram in Pb–Ti–Zr–K–H 2O system at 150°C showing stability eld of the PZT phase with over 99% yield. (b) Control of crystallite size between 250 nm and 10 µm using simple agitation speed during hydrothermal synthesis and type of the precursor used. (c) and (d) Control of mor phology during hydrothermal synthesis at 1 50°C using concentration of t he TMAH mineralizer, which was (c) 0.5 m and (d) 1.0 m.

model t hat c ombines t he H elgeson–Kirkham–Flowers–Tanger (HKFT) equation of state for standard-state properties of aqueous species with a nonideal solution model based on the activity coefficient e xpressions de veloped b y Bro mley a nd P itzer, a nd modied b y Z emaitis e t a l. F or s olid s pecies, s tandard-state properties a re u sed i n c onjunction w ith ba sic t hermodynamic relationships. Fugacities of components in the gas phase are calculated f rom t he Re dlich–Kwong–Soave e quation o f s tate. A more detailed description of the thermodynamic model as well as citations t hat cover more of t he f undamentals c an be found elsewhere [1,13].

2.3.2 Methodology for Generating Stability and Yield Diagrams First, the desired product and components of the hydrothermal system have to be dened. Ā us, the identities of the precursors, mineralizers, a nd ot her add itives needed for t he s ynthesis of a

43722_C002.indd 6

required so lid p hase n eed t o be speci ed. Ā is i nformation is used a s i nput d ata, a long w ith t he r ange of re agent concentrations, temperature, and pressure specied by the user. Ā er efore, it is important that there is a data bank that is relevant for all the components in the system, which contains standard-state thermochemical p roperties a nd i ndependent re action s ets f or a ll species, HKFT equation of state parameters for aqueous species, and Redlich–Kwong–Soave equation of state parameters for gaseous species. Ā e OLI soft ware also stores binary parameters for ion–ion, ion–ne utral, a nd ne utral–neutral s pecies i nteractions. It should be noted, however, that the standard-state properties and parameters are frequently obtained by regressing numerous kinds o f t hermodynamic d ata. Ā ese d ata i nclude v apor p ressures, osmotic coefficients, activity coefficients, enthalpies, and heat capacities of solutions, and solubilities and heat capacities of solids. When data are not a vailable in the OLI data bank for the specied hydrothermal system, a private data bank must be constructed. Ā e literature can be consulted for data, as well as

10/3/2008 8:42:58 PM

2-7

Intelligent Synthesis of Smart Ceramic Materials

several methods for estimating thermochemical data. Ā e chemistry model generation step creates the species and reactions that are possible with the given components of the system. All possible combinations of ions, neutral complexes, and solids are considered i n t his s tep. Ā e c hemical s peciation mo del i s a s et o f equations, wh ich co ntains ch emical eq uilibrium eq uations, phase equilibrium equations, mass balance, and electroneutrality equations. O nce t he t hermodynamic c onditions a re s pecied, the chemical speciation model is solved. Equilibrium concentrations of all species are calculated as a function of variables such as mineralizer concentration. Ā is gives the equilibrium composition of a specic set of reaction conditions, but to u nderstand the o verall b ehavior o f t he s ystem c omputations m ust b e performed over a wide range of conditions. For this purpose, we specify t he p rocessing v ariables o f i nterest. Ā e O LI s oftwa re offers a  exible choice of independent variables as x- and y-axes of s tability a nd y ield d iagrams, w hich i nclude p recursor a nd mineralizer concentrations (Figures 2.3a and 2.4a), solution pH (Figure 2.3b) and temperature, in addition to the electrochemical potential, t he latter being used for t he simulation of corrosion. In the case of y ield diagrams, the y ield value (e.g., 99%, 99.9%, and 99.95%) of the desired material must be chosen. Ā e upper temperature l imit o f t he O LI s oft ware for creation of both stability and yield diagrams is around 300°C, which covers mild hydrothermal synthesis conditions for most ceramic materials.

2.3.3 Utilization of Thermodynamic Modeling Calculated phase diagrams can perform many functions during the c ourse o f i ntelligent s ynthesis o f c eramics. I n add ition to precisely dening the concentration–temperature–pressure processing variable space over which the phases of interest are stable, many different types of precursor systems and additives can be compared, and experiments can be designed to make materials that have never been previously prepared in hydrothermal solution [1,13]. Moreover, t he t hermodynamic modeling enables to compute chemical supersaturation in a g iven system. Ā u s, one can evaluate how the processing variables on the phase diagram inuence supersaturation within a ph ase stability region. Once the sup ersaturation i s k nown, b y u sing c onventional c rystal growth models, one can identify when the process is dominated by nucleation and when by the growth rate, which can be utilized in morphology and size control, as demonstrated in the following section.

2.4

Intelligent Control of Crystal Size and Morphology

We will discuss this aspect of intelligent synthesis of ceramics using mo stly p owders a nd c oatings a s e xamples. Ā ese two forms of ceramics constitute the vast majority of ceramic materials synthesized under mild hydrothermal conditions and their size or morphology control is governed by principally the same

43722_C002.indd 7

rules. Control of single crystal growth has been reviewed elsewhere [2]. With t he t hermodynamic v ariables p rocessing s pace w ell de ned for t he phase of i nterest, a r ange of c onditions c an b e then e xplored to c ontrol re action a nd c rystallization k inetics for the purpose of developing a process suitable to produce the desired form of ceramics (e.g., particular size, morphology, aggregation level for powders, and microstructure, i.e., size and morphology of constituting crystallites for coatings). Ā er modynamic p rocessing v ariables, suc h a s tem perature, p H, c oncentrations of reactants, and additives, determine not only the processing s pace f or a g iven m aterial b ut a lso i nuence both reaction a nd c rystallization k inetics. Ā e p henomena t hat underlie the size and morphology or microstructure control using t he t hermodynamic v ariables a re t he overall nucleation and growth rates, which control crystal size and the competitive growth r ates a long p rincipal c rystallographic d irections t hat control morphology. Size of the crystals can be thus controlled by varying temperature and concentration. Crystal morphology and size can be additionally affected by surfactants, which can adsorb o n spe cic c rystallographic f aces a nd s olvents, w hich adsorb similarly as well as regulate solubility. Unfortunately, changing the thermodynamic synthesis variables is constrained by the phase boundaries in a specic phase diagram. Ā er efore, it may or may not be possible to exploit all sizes, morphologies, and m icrostructures for a c ertain m aterial by mo difying only the th ermodynamic v ariables. H owever, n onthermodynamic variables are also extremely important when operating in thermodynamically limited processing variable space. For instance, changing the stirring speed during synthesis of lms or powder can change the crystallite size by orders of magnitude [14].

2.4.1 Thermodynamic Variables Ā e technique of using thermodynamic variables, such as temperature, c oncentrations o f re actant, v arious add itives o r s olvents, is the most widely used method to control size and morphology of c rystals u nder h ydrothermal c onditions. I t i s because w ith i ncreasing tem perature a nd c oncentrations ( i.e., supersaturation), both nucleation and growth rates signicantly increase. It is widely believed that size of the crystals increases with increasing temperature or concentrations. However, this is true only when t he g rowth r ate dominates over t he nucleation rate. If t he nucleation rate dominates over t he growth rate, t he relationship will be opposite, i.e., size of the crystals will decrease with increasing temperature or concentrations, because there are so many particles formed rapidly and they have little time to grow. In extreme cases of very high nucleation rate, even amorphous materials c an b e s ynthesized. N ucleation- a nd g rowth-ratecontrolled s ynthesis r anges d iffer f or e very m aterial a nd h ave to be calculated or determined experimentally in each particular case. Changing re actants c oncentration c an b e u sed not o nly to control si ze o f t he c rystallites b ut a lso to c ontrol t heir s hape. Fibers or cubes of lead zirconate titanate (PZT, chemical formula

10/3/2008 8:42:59 PM

2-8

Pb(ZrxTi1−x)O3, where x = 0 – 1) can form during the hydrothermal s ynthesis de pending up on t he c oncentration o f te tramethylammonium hydroxide (TMAH) mineralizer (Figur e 2 .4c and d) [15]. Ā e morphology of crystals can be furthermore controlled by the use of appropriate chemical additives or solvents present in the crystallization environment. Ā is approach utilizes preferred adsorption of the additives or solvents on particular faces of the growing crystals, thus preventing growth in certain directions. Consequently, crystals of t he same ceramic phase w ith various levels of aspect ratios, from platelets through whiskers, can be synthesized. A good example of using this approach is hydroxyapatite [16,17]. If chelating agents (lactic acid, EDTA, etc.) are used to prevent precipitation of amorphous precursor at ambient temperatures, precipitation at ele vated temperatures u sually y ields larger crystals [16]. Use o f tem plates, suc h a s s elf-assembled o rganic mole cules, emulsions, or nanotubes in order to direct the growth of the ceramic ph ase re sults i n a v ariety o f m aterials a rchitectures including me soporous m aterials, na norods, a nd na nosized monodispersed powders [10]. Ā ese approaches, however, require removal of the template after the synthesis, which often involves high-temperature calcination. Monodisperse c eramic pa rticulate s ystems w ith m inimized aggregation le vels c an a lso b e s ynthesized f rom aque ous s olutions without templates by precipitating  rst a small number of nuclei and subsequently reducing supersaturation level in order to prevent further nucleation and enable diff usion growth of the formed nuclei [18]. While the powder uniformity is excellent, it requires working with dilute reactant concentrations (e.g., ≈0.005 M) so that both supersaturation and nucleation are highly uniform and colloidal stability can be maintained [12].

2.4.2 Nonthermodynamic Variables Nonthermodynamic s ynthesis v ariables, suc h a s s tirring r ate and me thod o f p recursor p reparation, c an b e e ssential i n si ze and morphology control of crystals, while keeping the thermodynamic variables constant. Generally speaking, stirring during crystal growth leads to an increase in the probability of spontaneous nucleation, a decrease in supersaturation inhomogenities, and an increase of the growth rate. Ā e nal size of the crystals is t hen de termined b y a ba lance b etween t he n ucleation a nd growth r ates. A go od e xample o f a v ery s trong i nuence of nonthermodynamic v ariables o n c rystal si ze a nd mo rphology is PZT [14]. Concentration of the chemicals used in hydrothermal PZT synthesis was well within the calculated stability eld of the PZT phase, which is shown in Figure 2.4a. Increasing the stirring speed alone from 200 to 1700 rpm has reduced the particle size of PZT crystals almost two orders of magnitude, from over 10 µm to a submicron range (Figure 2.4b). In this particular case, the nucleation rate seems to dominate over the growth rate, therefore t he c rystal si ze de creased w ith i ncreasing s tirring speed. A t h igher s tirring s peeds, t he PZ T pa rticles e xhibited

43722_C002.indd 8

Smart Materials

better-dened cube habit and were more uniform in size, which is in line with reduction of supersaturation inhomogenities with increasing a gitation r ate. Use o f d ifferent precursors i mpacted the crystal size as well (Figure 2.4b). Stirring speed or velocity of the solution ow a long t he substrate can affect signicantly the thickness, morphology, and chemical c omposition o f  lms g rown u nder h ydrothermal conditions, as demonstrated using BaTiO3, SrTiO3, and PZT thin  lms [14]. Film t hickness a nd morphology i s a lso sig nicantly impacted by t he t ype a nd c rystallographic or ientation of t he substrate used (Figure 2.1e and f).

2.5

Summary

Hydrothermal s ynthesis i s en vironmentally f riendly, energ yconserving low-temperature crystallization of all forms of smart materials directly from aqueous solutions. Hydrothermal crystallization a ffords e xcellent c ontrol o f c hemical c omposition (oxides, nonoxides, native elements, dopants, solid solutions), as well as shape, size, and morphology of ceramics. Understanding the ph ysicochemical p rocesses o ccurring i n t he aque ous s olution is the key for intelligent engineering of hydrothermal synthesis of smart ceramics. Ā ermodynamic modeling i s u sed to predict formation conditions for phases of interest. Subsequently, size, mo rphology, a nd a gglomeration le vel o f t he s ynthesized ceramics can be controlled in wide ranges using both thermodynamic and nonthermodynamic (kinetic) variables. Such intelligent app roach ena bles to ac hieve c ost e ffective sc ale-up a nd commercial p roduction o f sm art c eramics u sing t he h ydrothermal technology.

Acknowledgments Ā e authors w ish to g ratefully acknowledge t he supp ort of t he Office of Naval Research, Ā e National Institute of Health, OLI Systems, I nc., a nd S awyer Technical M aterials, L LC, f or t heir generous support of the research cited in this manuscript.

References 1. Riman, R .E., S uchanek, W.L., a nd L encka, M.M., Hydrothermal cr ystallization o f cera mics, Ann. Ch im. S ci. Mat., 27, 15, 2002. 2. Byrappa, K. and Yoshimura, M., Handbook of Hydrothermal Technology, N oyes Pub lications/William Andrew Publishing LLC, Norwich, NY, 2001. 3. Yoshimura, M., Suchanek, W.L., and Byrappa, K., Soft solution p rocessing: A stra tegy f or o ne-step p rocessing o f advanced inorganic materials, MRS Bull., 25, 17, 2000. 4. Sômiya, S., Ed., Hydrothermal Reactions for Materials Science and Engineering. An Overview of Research in Japan, Elsevier Science, London, 1989. 5. Roy, R . Accelerating the kinetics o f low-temperature inorganic syntheses, J. Solid State Chem., 111, 11, 1994.

10/3/2008 8:42:59 PM

Intelligent Synthesis of Smart Ceramic Materials

6. Suchanek, W.L., L encka, M.M., a nd Rima n, R .E., H ydrothermal synthesis of ceramic materials, in Aqueous Systems at Elevated Temperatures and Pressures: Physical Chemistry in Water, Steam, and Hydrothermal Solutions, Palmer, D.A., Fernández-Prini, R ., a nd H arvey, A.H., E ds., Els evier, Amsterdam, 2004, Chapter 18. 7. Yoshimura, M. a nd S uchanek, W., I n si tu fa brication o f morphology-controlled advanced ceramic materials by soft solution processing, Solid State Ionics, 98, 197, 1997. 8. Riman, R .E., in High P erformance C eramics: S urface Chemistry in Processing Technology, Pugh, R. and Bergström, L., Eds., Marcel-Dekker, New York, 1993, p. 29. 9. Dawson, W.J., Hydrothermal synthesis of advanced ceramic powders, Ceram. Bull., 67, 1673, 1988. 10. Adair, J.H. and Suvaci, E., Morphological control of particles, Curr. Opin. Colloid Interf. Sci., 5, 160, 2000. 11. Niesen, T .P. a nd D eGuire, M.R ., Re view: D eposition o f ceramic t hin  lms a t lo w tem peratures f rom aq ueous solutions, J. Electroceramics, 6, 169, 2001.

43722_C002.indd 9

2-9

1 2. Sugimoto, T., Fine P articles, S ynthesis, Cha racterization, and Mechanisms of Growth, Marcel-Dekker, Inc., New York, 2000. 13. Lencka, M.M. a nd Rima n, R .E., in Encyclopedia of Smar t Materials, Vol. I, S chwartz, M., E d., John Wiley, New York, 2002, pp. 568–580. 14. Suchanek, W.L. et al ., H ydrothermal dep osition o f <001> oriented epitaxial Pb(Z r,Ti)O3 lms under va rying hydrodynamic conditions, Crystal Growth Design, 5, 1715, 2005. 15. Cho, S.-B ., Ole dzka, M., a nd Rima n, R .E., H ydrothermal synthesis of acicular lead zirconate titanate (PZT), J. Crystal Growth, 226, 313, 2001. 16. Suchanek, W. et al., Biocompatible whiskers with controlled morphology and stoichiometry, J. Mater. Res., 10, 521, 1995. 17. Riman, R .E. et al ., S olution syn thesis o f h ydroxyapatite designer particulates, Solid State Ionics, 151, 393, 2002. 18. Matijevic, E., M onodispersed met al (h ydrous) o xides—a fascinating  eld of colloid science, Acc. Chem. Res., 14, 22, 1981.

10/3/2008 8:42:59 PM

43722_C002.indd 10

10/3/2008 8:42:59 PM

3 Functionally Graded Polymer Blend 3.1 I ntroduction ............................................................................................................................... 3-1 3.2 Mechanism of the Preparation Method ..................................................................................3-2 3.3

Diff usion–Dissolution Method • P olymerization–Diffusion Method

Preparation and Characterization of Several Types of Functionally Graded Polymer Blends ............................................................................................................3-5 Amorphous Polymer/Amorphous Polymer Miscible Blend (Dissolution–Diffusion Method) • Amorphous Polymer/Crystalline Polymer Miscible Blend (Dissolution–Diff usion Method) • Amorphous Polymer/Amorphous Polymer Immiscible Blend (Dissolution–Diff usion Method) • Amorphous Polymer/Crystalline Polymer Immiscible Blend (Polymerization–Diffusion Method)

3.4 Functional and Smart Performances and the Prospect for Application .............................3-8

Yasuyuki Agari Osaka Municipal Technical Research Institute

3.1

Functional and Smart Performances of PVC/PMA Graded Blend • Functional and Smart Performances of PEO/PLLA Graded Blend • Functional and Smart Performances of PEO (or PEO/LiOCl4)/PBMA Graded Blend

3.5 Prospects for Application in the Functionally Graded Blends........................................... 3-13 References ...........................................................................................................................................3-13

Introduction

Many reports have been published on functionally graded materials made of metals and ceramics [1]. Ā ese g raded m aterials have characteristics of improved strength against thermal stress as well as excellent electromagnetic and optical properties. Ā er e have also been many reports on a functionally graded ceramic, which can be called a sm art material. Ā ese ceramics showed a strong t hermoelectric per formance, wh ich sh ifted w ith a n increase in temperature. As a re sult, the thermoelectric performance was very high in a wide temperature gradient. Ā ere have also been some reports on functionally graded polymeric materials [2–43]. Ā ese functionally graded polymeric materials c an be classied i nto four t ypes f rom c urrently u sed materials, as shown in Table 3.1. Ā e graded structures may be classied into six types. However, reports on functionally graded polymer blen ds a re s omewhat l imited [ 4–7,14–28], a lthough there h ave b een p ublished re ports o n v arious t ypes o f blen ds. Ā e f unctionally g raded polymer blend has been considered to have a structure as shown in Figure 3.1. Ā e blend has two different surfaces without an interface and can have both the advantages of a laminate and a homogenous blend.

We have devised a new method called the dissolution–diffusion method f or p reparing f unctionally g raded p olymer blen ds [4–7,24–28]. In t his method, g raded polymer blends a re classied into two types. Graded polymer blends are usually prepared by t hree me thods: su rface i nclination i n t he mel t s tate [14,15], surface inclination in the solution method [16,17], and diffusion in t he melt me thod [19–23]. Ā e dissolution–diffusion method devised by us is only one of the methods that can be used in preparing b oth t ypes o f g raded p olymer blen ds. O ur me thod h as the following advantages in comparison with the other methods. Ā e preparation time in our method is very short, and primarily is 1 00 t imes s horter t han t hat o f d iff usion i n t he mel t s tate method. Ā e opt imum c ondition c an b e e asily de termined because our method has many controllable factors. Furthermore, the chemical decomposition of molecules does not occur in the preparation by our method because the preparation is performed at a lower temperature. Ā erefore, we believe that our method is the most useful method. In further work, we recently found a new preparation method (polymerization–diff usion me thod) b y t he p olymerization o f a monomer during diff usion using a m acroazoinitiator [29,30]. 3-1

43722_C003.indd 1

10/3/2008 8:43:36 PM

3-2

Smart Materials TABLE 3.1

Various Types of Functionally Graded Polymeric Materials

Types of Used Materials

Structure

Metal (or ceramic)/polymer

Composites

Polymer–polymer

Immiscible polymer blend

Miscible polymer blend Atom–atom (intramolecules)

Copolymer (random) Copolymer (tapered) Density of cross-linking

High-order structure (same polymer) Crystal structure

Content of polymer A (%)

Polymer A

Polymer B

100

100

100

0

0

0

Position of measuring point Laminate

FIGURE 3.1

Functionally graded blend

Homogeneous blend

Schematic model of functionally graded blend.

It i s u seful f or t he p reparation o f a n i mmiscible g raded blen d and can form a graded structure of similar molecular weight. Ā is chapter presents a detailed description of the preparation mechanism us ed in diss olution–diffusion a nd p olymerization– diffusion metho ds t o p repare f unctionally grade d p olymer blends. Ā e chapter also explains how to determine the optimum

43722_C003.indd 2

Preparation Method • Laminate method • Electric eld method • Centrifugation method • Flame spraying method • Surface inclination in melt state method • Surface inclination in solution method • Dissolution–diffusion method • Diffusion in melt method • Dissolution–diffusion method • Diffusion method of monomer in polymer gel • Living anion or radical polymerization method • Changing method of cross-linking concentration • Changing method of cross-linking temperature • Injection mold method

Size of Dispersion Phase Big

Molecular order Atom order

Same atoms and molecules

condition f or the s everal typ es o f f unctionally grade d p olymer blends (polyvinyl chloride [PVC]/polymethacrylate, polymethyl methacrylate [PMMA], p olyhexyl methacr ylate [PHMA]) o r polycaprolactone (PCL), polyethylene oxide (PEO)/poly(l-lac tic acid) (PLLA), bisphenol A type polycarbonate (PC)/polystyrene (PS), PEO/polybutyl methacrylate (PBMA), and (PEO/LiClO 4)/ PBMA us ed in c haracterizing grade d str uctures f or b lends b y measuring the Fourier transform infrared (FTIR) spectra, confocal Raman spectra, thermal behaviors around the glass transition temperature ( Tg) b y Diff erential S canning C alorimetry (DSC) methods, b y S canning Ele ctro-Microscopy–Energy Disp ersive X-ray S pectrometry (S EM-EDX) obs ervation, a nd b y n uclear magnetic resonance (NMR), or Gel Permeated Chromatography (GPC) measur ement. F urthermore, s everal typ es o f f unctional properties are discussed especially for smart performance, which were ca used b y the grade d str ucture. Finall y, the p rospects o f functionally g raded p olymer blends a nd t heir a pplications a re discussed.

3.2

Mechanism of the Preparation Method

3.2.1 Diffusion–Dissolution Method Ā e mechanism of formation of a graded structure by the diffusion–dissolution method is considered to be the following [27]. After a polymer B solution is poured onto a polymer A  lm in a glass petri dish, polymer A begins to dissolve and diff use in the solution on t he a ir side ( Figure 3.2), but t he d iff usion is i nterrupted when all the solvent evaporates. Ā us, a blend lm is produced, which consists of a concentration of gradient of polymer A/polymer B in the thickness direction.

10/3/2008 8:43:36 PM

3-3

Functionally Graded Polymer Blend

Evaporation Polymer B solution Dissolution and diffusion

Polymer A film

FIGURE 3.2

rate of p olymer A i n p olymer B s olution, a nd (c) i nterruption time for the diff usion due to the completion of solvent evaporation. F actors f or c ontrolling t he a bove ph enomena a re (1) t he type of s olvent, (2) t he c asting temperature, (3) t he mole cular weight of polymer A, and (4) the amount of polymer B solution. Until polymer A completely dissolves or reaches the surface of polymer B solution, i.e., in the formation of the  rst and second types of structure, the diff usion of polymer A i n the polymer B solution is considered to obey Fick’s second law (Equation 3.1) by the a ssumption of ne glecting t he e vaporation of t he s olvent i n polymer B solution during the diffusion:

Schematic model of dissolution–diffusion method.

Following t he s teps o f d issolution a nd d iff usion o f p olymer A, the graded structures should be classied i nto t hree t ypes (Figure 3.3). First type: Polymer A begins to dissolve and then diffuses, but does not yet reach the air side surface of polymer B solution. Ā e blend has three phases (polymer A, polymer B, and a thin graded structure). Second type: Just when all of polymer A has dissolved, the diffusion f rontier reaches up to t he air side su rface of polymer B s olution. Ā e blen d h as o ne g raded ph ase from t he su rface to t he ot her o ne, w hile t he su rfaces are composed of polymer A only or polymer B only. Ā ir d type: After the dissolution and diff usion of polymer A reached up to the air side surface of polymer B solution, polymer A and polymer B molecules began to mix with each other and became miscibilized. Ā e concentration gradient then began to disappear. Ā e formation of a concentration gradient should depend on (a) dissolution rate of polymer A in polymer B solution, (b) diffusion

Graded structure 1 Polymer B Narrow graded phase Polymer A Graded structure 2

Wide graded blend

⎛ ∂2C A ⎞ ∂C A (3 = DAB ⎜ ∂t ⎝ ∂x 2 ⎟⎠

.1)

where CA is the concentration of polymer A t is the passed time x is the distance from the surface of polymer A sheet DAB is an apparent diffusion coefficient Ā e point where CA approaches 1 and shifts to t he petri glass side, thus carrying forward the dissolution of polymer A. Ā us, by c onsidering t his e ffect a nd re arranging m athematically Equation 3.2 is obtained from Equation 3.1: ⎛ (x − b) ⎞ C A = erfc ⎜ ⎟ ⎝ 2 DAB t ⎠ ∞

⎛ 2 ⎞ erfc(x ) = ⎜ exp −x 2 dx ⎝ p ⎟⎠

∫

( )

(3.2)

x

where b is the distance between the petri glass side su rface and the ot her side o f rem ainder o f p olymer A , w hich h as not d issolved yet. Ā erefore, the gradient pro le in the blend at t can be estimated by Equation 3.2. Adaptability of Equation 3.2 t o t he e xperimental data w as examined in the PVC/PMMA graded blend, which is given a detailed e xplanation in S ection 3 .3.1. Ā e e xperimental d ata agreed approximately with the ones predicted by Equation 3.2, as shown in Figure 3.7. DAB and b were obtained as 6.38 mm 2/s and 57 mm, respectively. Ā e DAB value was much larger t han the v alue ob tained b y d iff usion i n mel t s tate me thod, w hich implies that the dissolution–diff usion method is very useful. Further, a thicker and a more excellent graded blend lm was prepared by the m ultiple-step metho d, as i llustrated in Figur e 3.4. In this case, the graded blend was obtained by repeatedly preparing and changing the composition of the blend in the pouring solution.

Graded structure 3

3.2.2 Gentle graded blend

FIGURE 3.3 Schematic models of various types of graded structures.

43722_C003.indd 3

Polymerization–Diffusion Method

The mechanism of formation of a graded structure by polymerization during the diff usion method is as f ollows [30]. After the polymer A lm containing a macroazoinitiator, i.e., a radical type initiator, wi th o ligomeric p olymer A s egment is cast o n a  lter, the

10/3/2008 8:43:37 PM

3-4

Smart Materials

Two-step method

One-step method Evaporation of solvent First step

Polymer B solution Dissolution–diffusion

Evaporation of solvent

Evaporation of solvent

Polymer A/Polymer B (5/5) solution Dissolution–diffusion

Polymer A/Polymer B (7/3) solution Dissolution–diffusion Polymer A film

Polymer A film

Polymer A film

Evaporation of solvent Polymer B solution

Second step

Four-step method

Dissolution–diffusion Film formed in the first step

Evaporation of solvent Polymer A/Polymer B (5/5) solution Dissolution–diffusion Film formed in the first step Evaporation of solvent Polymer A/Polymer B (3/7) solution Dissolution–diffusion

Third step

Film formed in the second step Evaporation of solvent Polymer B solution

Fourth step

Dissolution–diffusion Film formed in the third step

FIGURE 3.4

Schematic models of multiple steps method.

resulting laminate is p ut on a mo nomer B s olution in a al uminum petri dish kept at a constant temperature (Figure 3.5). Ā e monomer B begins to diffuse into polymer A lm to the air side with polymerization. When the process is interrupted on the way by enough diffusion of the monomer B, a blend lm is produced, which consists of a concentration gradien t o f p olymer A/polymer B in the thic kness direction.

Ā e graded structures could be determined by the bala nce of polymerization a nd diff usion o f mo nomer B . Ā e temperature content o f a macr oazoinitiator, time , i .e., la rgely eff ected the balance. Ā e temperature and the time enha nced both the p olymerization and the diffusion, and the content of the macroazoinitiator enhanced only the polymerization. Ease of evaporation of monomer B enhanced only the diffusion.

Diffusion of Monomer B

Evaporation of Monomer B

Polymer A film

Polymerization

Filter Monomer B Heating

FIGURE 3.5 Schematic model of polymerization–diffusion method.

43722_C003.indd 4

10/3/2008 8:43:38 PM

3-5

Functionally Graded Polymer Blend

Preparation and Characterization of Several Types of Functionally Graded Polymer Blends

3.3.1

Amorphous Polymer/Amorphous Polymer Miscible Blend (Dissolution–Diffusion Method)

In the cas e of the PV C/PMMA system [6,7], s amples were prepared by changing the four controllable conditions: (1) the type of solvent, (2) the castin g temperature, (3) the mo lecular weight of the PVC, and (4) the amount of the PMMA solution. Ā e graded structures o f thes e s amples w ere c haracterized b y FTIR -ATR, Raman microscopic spectroscopy, and DSC methods. Here, FTIRATR was the spectroscopy, by which the IR spectra on the surface layer (a bout 1-10 mm o f thic kness) was me asured. Figur e 3.6 shows the graded structure of the samples in the direction of the thickness, measured by FTIR-ATR. In the similar blend with the graded structure 1, for a la minate, the PMMA co ntent increased at 60% of the distance/thickness and was conrmed to have a thin graded layer (about 10%–20% o f the dist ance/thickness). In the blend with the graded structure 3, the PMMA content was kept to about 50% in all dist ance ranges. However, in the b lend with the graded structure 2, the PMMA content gradually increased in the range f rom 0% t o 100% o f the dist ance/thickness. Ā us, i t was found that this blend had an excellent wide concentration gradient. In this case, the PMMA content was estimated from the ratios of the absorption band intensities at 1728 cm−1 (the stretching of the carbonyl group in PMA) and 615 cm−1 (the stretching of C–Cl bond in the PVC). Ā e change in the PMMA content in the thickness direction of the blended lm was estimated by measuring the FTIR-ATR spectra on a sliced layer of the blended lm. Ā e c hange i n t he PV C c ontent o f t he g raded blen d w as characterized b y t he c onfocal R aman s pectroscopy me thod, similar to t he F TIR-ATR me thod, a s s hown i n F igure 3.7. Ā e measurement of the Raman microscopic spectra was performed by measuring the Raman spectra at the focused point, which was

PMMA content (%)

100 Graded structure 1 Graded structure 2 Graded structure 3

80 60 40 20 0 0

20 40 60 80 100 (Distance from petri glass side)/(Sample thickness) (%)

FIGURE 3.6 Change i n PMMA c ontent i n t he t hickness d irection of several types of the PVC/PMMA graded blends.

43722_C003.indd 5

1.0 Focused area

0.9

FTIR method Raman method Predicted values

0.8

Change

0.7 PVC content

3.3

0.6 0.5 0.4 0.3 0.2 0.1 0.0

0

20 40 60 80 100 120 140 160 180 Distance from petri glass side (µm)

FIGURE 3.7 Change in the PVC content in the thickness direction of the PVC/PMMA blend.

shifted 1 0 b y 1 0 mm, f rom o ne su rface a rea to t he ot her o ne. It could be conrmed that the blend had a c omparatively thick layer of a g raded structure phase. Ā is method is considered to be signicantly useful because an easy and detailed estimation can be made for a graded pro le of a blend. Furthermore, t he g raded s tructure w as c haracterized b y the DSC method. The DSC curve of the blend having a widely graded s tructure ( graded s tructure 2 ), w hich s hows a mo re gradual s tep ar ound Tg t han t he ot hers, i s s hown i n F igure 3.8. Similarly, the structures of the samples, which were prepared in the several types of the conditions mentioned previously, w ere i nvestigated a nd i t w as f ound t hat t he t ypical optimum c ondition ( molecular w eight o f PV C: M n = 35,600, Mw = 60,400, t ype of s olvent: te trahydrofuran (T HF)/toluene (5/1), volume of solvent: 0.23 ml/cm 2 , tem perature: 3 33 K) could be obtained. In the case of PVC/PHMA system [7], the graded structure of the s ample c ould not b e e stimated b y t he F TIR-ATR a nd t he DSC me thods b ecause PHMA w as very s oft at ro om temperature. Ā us, the graded structure was measured by the SEM-EDX method (Figure 3.9). Ā e chlorine content in the sample increased gradually o n t he p etri g lass side a nd t hen i t w as c onrmed to have a wide graded structure. Finally, the structures of the samples, which were prepared in t he several t ypes of t he previously mentioned conditions, were investigated. The typical optimum condition (molecular weight of P VC: M n = 35,600, Mw = 60,400, t ype o f s olvent: MEK, volume of solvent: 0.37 ml/cm 2 , tem perature: 3 13 K) was obtained.

10/3/2008 8:43:39 PM

3-6

Smart Materials

PVC content (%)

Exotherm

Graded structure 1

Graded structure 2

Graded structure 3

100 J 90 E E 80 J 70 60 50 40 30 Solution volume 20 J 0.182 ml/cm2 10 E 0.364 ml/cm2 0 0 50 100

E J

E

EJ EJ E JE

E E E J

150

200

250

Distance from petri glass side (µm) 330

360 390 Temperature (K)

420

FIGURE 3.8 DSC curves of several types of PVC/PMMA blends.

FIGURE 3.10 Graded structures of the PVC/PCL graded blends measured by the FTIR-ATR method.

100 90 E 80 70 60 50 40 30 20 10 0 0

340 J

320 E

300

E J

E

E

280

J E

260

E

J Tg E PVC content

J

E J

240

E

50 100 150 200 Distance from petri glass side (µm)

J

Tg (K)

20 µm

PVC content (%)

Chlorine content

360

E E J

220 200 250

FIGURE 3.11 Graded structures of PVC/PCL graded blends characterized by the DSC method. Thickness direction

FIGURE 3 .9 Chlorine c ontent a long t he t hickness of PV C/PHMA graded blend (×750).

3.3.2

Amorphous Polymer/Crystalline Polymer Miscible Blend (Dissolution–Diffusion Method)

In the case of the PVC/PCL system [25], we obtained the optimum conditions when t he graded polymer blend was prepared w ith a wider compositional gradient, similar to the PVC/PMMA system. Figure 3.10 shows the PVC content of the samples for the direction of thickness, measured by the FTIR-ATR method. PVC decreased at about 70 mm from the petri glass side and decreased gradually until the surface of the air side, that is, about 240 mm apart from the petri glass side in both cases of the solution volume. Ā en, the change of Tg for the thickness direction of the blend  lm was characterized by the DSC method (Figure 3.11) in the case o f 0. 364 ml/cm 2 o f s olution v olume. Tg de creased w ith the increase of the distance from the petri glass side, similar to the PVC content. Ā us, the graded structure in the PVC content was conrmed by the graded prole in Tg.

43722_C003.indd 6

Furthermore, t he change of t he PCL crystalline content was evaluated from that of the amount of the heat diffusion effect on the PCL crystalline, measured by the DSC method. Ā e heat diffusion began to increase on the specimen after it w as i nitially kept at 0 a nd then about 130 mm of the distance to t he air side. Ā en, it i ncreased i mmediately a round a bout 180 mm. Ā u s, it was found that the graded structure in the PCL crystalline was formed i n a d istance o f 1 30–240 mm. Ā is m eans th at th e obtained graded PVC/PCL blend had both a gradient concentration of the PVC and a gradient content of the PCL crystalline, as shown in Figure 3.12. Ā e PCL content was about 30% for a distance o f 1 30 mm. Ā is result corresponded to that of the PCL crystalline i n a h omogeneous PVC/PCL blend, which emerged in the case of the larger than 30% of the PCL [44]. It was further considered that the amorphous phase was made up of a miscible amorphous PVC/amorphous PCL blend. Finally, the PCL crystalline phase decreased the closer it came to the surface of the air side again. Ā is phenomenon occurred because t he forming of the amorphous phase was more thermodynamically stable than that of the crystalline phase. Ā e graded structure of the PVC/PCL graded blend could be schematically illustrated in Figure 3.13.

10/3/2008 8:43:39 PM

3-7

160

100 E

140

90

Type of solution E PC -b-PS copol ymer/

80

J

J J

100

E J J

J PVC content E Heat of diffusion

80

J E

J

60

J J E

40 20 0

E

E

0

E

E

J J J

E

50 100 150 200 Distance from petri glass side (µm)

100 90 80 70 60 50 40 30 20 10 0 250

FIGURE 3 .12 Graded st ructure o f P VC/PCL g raded b lends ( PCL crystalline).

PS(1/ 9) blend PS onl y

70 PC content (%)

120

PVC content (%)

Heat of diffusion (J/g)

Functionally Graded Polymer Blend

60 50 40

J

J J E J JJ E E

30 20

E

10 0

J

E J 0 20 40 60 80 100 120 140 160 180 Distance from petri glass side (µm)

FIGURE 3 .14 Graded s tructure of P C/PS g raded ble nd w ith or without the PS-b-PC block copolymer.

(1) Effect of macro phase separations PCL crystalline phase PCL

PVC

FIGURE 3 .13 Schematic mo del of t he PV C/PCL g raded ble nd (graded structure).

3.3.3 Amorphous Polymer/Amorphous Polymer Immiscible Blend (Dissolution–Diffusion Method) We attempted to p repare a g raded PC/PS blend by the dissolution–diff usion method [24], similar to the PVC/PMMA system. In this case, the PS solution was poured onto the PC lm. However, we d id not ob tain a g raded structure, but obtained a homogeneous t wo-layer system, which was composed of about 50% and 0%–10% of PC, as shown in Figure 3.14. Next, the macrophase separation was observed in the former layer. Ā is re sult w as t hought to o ccur b ecause o f t he f ollowing; only three factors, dissolution rate, diff usion rate, and evaporation time, could affect the process in forming a graded structure of miscible blend. However, in forming the graded structure of an immiscible blend, three additional factors also played a role: macrophase s eparation, su rface ro ughness, a nd g ravimetry, a s shown i n F igure 3 .15. E specially, t he m acrophase s eparation may increase t he size of t he separated phases up to t he size of the thickness in the prepared lm, resulting in a possible break in the formation of a strongly graded structure, when its concentration becomes higher by the evaporation of the solvent. Ā erefore, we attempted to prevent macrophase separation during graded structure formation by adding PS-b-PC block copolymer [45]

43722_C003.indd 7

(2) Effect of surface inclination

(3) Effect of gravitation

FIGURE 3 .15 Other f actors t hat h ave a n e ffect on for ming g raded structure in an immiscible blend.

to the PS solution (PS-b-PC block copolymer/PS = 1/9). Āe copolymer may act as a compatibilizer by decreasing the interface energy of the phases. In this case, the PC segment content in the block copolymer was 46% (NMR measur ement). It was f ound that the wide grade d structure in the ob tained PC/PS b lend was f ormed in the dist ance range of 0–100 µ m from the petri glass side (Figure 3.14). Furthermore, we attempted to prepare a graded PC/PS blend by pouring the PC solution containing the block copolymer onto the PS  lm. Figure 3.16 shows the change of the PC content for the direction of the  lm thickness. A wide graded structure was formed and conrmed not only, in the furtherest distance from, but also in the distance closest to the petri glass side. Ā is result meant that surface roughness signicantly inuenced the forming of the graded structure. Ā erefore, we were able to obtain the graded immiscible PC/PS blend by adding the PC-b-PS block copolymer. Āe graded structure a ssumed to b e f ormed f or t he P C/PS g raded blen d could be schematically illustrated in Figure 3.17.

10/3/2008 8:43:40 PM

3-8

Smart Materials 100 90 70 E

60 50

E E

40 30 20 E

E

10

E

E

E

0 0

20

40 60 80 100 120 140 160 180 200 Distance from petri glass side (µm)

FIGURE 3.18 PEO content in the thickness direction of PEO/PBMA graded blend measured by NMR and confocal Raman spectroscopy.

Distance in refraction (abs.)

FIGURE 3.16 Graded s tructure of P C/PS g raded ble nds i n t he c ase that the PC solution was poured on a PS  lm.

PEO content (%)

PC content (%)

80

100 NMR measurement 90 No.1 Confocal Raman spectroscopic measurement 80 70 No.2 No.5 No.6 60 No.4 No.7 50 No.3 40 30 20 10 0 0 200 400 600 800 1000 Distance from the surface of the air side (mm)

Peak 1 (Block coplymer) No. 7 No. 5 No. 3 No. 1 Retention time abs. Large

PC

PS

Peak 2 PEO or MAI

Long Molecular weight abs.

FIGURE 3.19 GPC charts in several types of the points, measured by NMR, in PEO/PBMA graded blend.

FIGURE 3.17 Schematic model of PC/PS graded blend.

3.4 3.3.4

Amorphous Polymer/Crystalline Polymer Immiscible Blend (Polymerization–Diffusion Method)

In t he c ase o f t he i mmiscible g raded blen d o f t he P BMA/PEO system [30], we obtained the optimum conditions in the preparation by the polymerization–diff usion method. Figure 3.18 shows the PEO c ontent of t he s amples for t he d irection of t hickness, measured b y t he c onfocal R aman s pectroscopy a nd t he N MR methods. Ā e d ata obtained by t he N MR me thod were a lmost the same as those by the Raman method. It was found that the NMR method was available and so it was used. Figure 3.19 shows the GPC data of the layers around the thickness points (No. 1, 3, 5, 7 in Figure 3.18) as measured by NMR method. Ā e peak corresponding to t he blo ck c opolymer ( EO-b-BMA c opolymer) became larger when the number of the points was larger. Ā en, the mole cular w eight a nd t he c ontent o f t he blo ck c opolymer became larger as the number of points became larger, that is, the BMA content was larger (Table 3.2). It was considered to o ccur because the polymerization rate was larger with a larger amount of the BMA monomer.

43722_C003.indd 8

Functional and Smart Performances and the Prospect for Application

3.4.1 Functional and Smart Performances of PVC/PMA Graded Blend Our s tudy [6,7] found t hat g raded p olymer blends h ad s everal types of functional properties, inclusive of smart performance. Ā us, the functional properties of the PVC/PMA blend containing graded structure 2 (an extremely wide graded concentration) were explained by comparing t hem w ith t hose of a blen d containing graded structure 1 (similar to a laminate system), a perfectly miscible blend (5/5), of PVC and PMA only. 3.4.1.1

Tensile Properties

Ā e re sults o f t he ten sile p roperties o f PVC, P MMA, t he p erfectly miscible blend (5/5), and the blends with graded structure 1 a nd 2 ( blend t ype 1 a nd 2 ) f or t he v ertical d irection o f t he thickness are summarized in Table 3.3. Ā e tensile strength of the homogeneously miscible blend was the highest, followed by blend type 2, surpassing PVC, PMMA, and the blend type 1. Ā is phenomenon means that the formation

10/3/2008 8:43:41 PM

3-9

Functionally Graded Polymer Blend TABLE 3.2 Molecular Weight and Composition of Several Types of Areas in the PEO/PBMA Functionally Graded Alloy in the Ā ic kness Direction Peak 1 No. 1 2 3 4 5 6 7

Peak 2

Mw

Mw/Mn

Mw

Mw/Mn

Ratio of Peak 1 to Peak 2

1.39 × 106 1.40 × 106 1.80 × 106 1.69 × 106 1.79 × 106 1.58 × 106 5.66 × 105

2.10 2.06 2.22 2.21 2.16 2.39 2.29

4.33 × 104 4.61 × 104 3.93 × 104 3.69 × 104 3.80 × 104 3.84 × 104 3.34 × 104

1.15 1.16 1.13 1.17 1.12 1.13 1.11

0.38 0.37 0.94 1.04 1.15 1.27 2.68

of a graded structure not only suppresses a break at the interface but also gives superior properties compared to the source materials. Ā is was considered to occur because the blend phase with the c oncentration g radient h ad a su fficiently high tensile strength. T he elo ngation at b reak o f blen d t ype 2 app eared sufficiently good. Ā e tensile modulus of t he blend t ype 2 w as higher than that of the blend type 1. It was found that the break in t he ten sile s tress c ould b e suppressed by t he formation of a concentration gradient. 3.4.1.2

Thermal Shock Resistance

Ā ermal s hock re sistance te sts were p erformed b y mo ving t he specimens f rom o ne b ox to a nother ( kept at 25 3 a nd 3 73 K) repeatedly (ve times) every 30 min. Ā e s pecimens w ere t hen evaluated f or t heir t hermal s hock re sistance b y me asuring a maximum angle of warp, as illustrated in Figure 3.20, and adhesive strength in shear by tension loading.

Ā ermal s hock re sistance te sts o f t he blen d t ype 2 w ith t he graded structure 2 w ere performed a nd t he re sults (maximum value of w arp a ngle a nd ad hesive s trength i n s hear by ten sion loading) were compared w ith t hose of t he blend t ype 1 h aving the graded structure 1, as shown in Table 3.3. Ā e  lm of the blend type 1 was highly warped, while that of the blend type 2 did not show any warp. Ā e adhesive strength in shear by tension loading of the blend type 2 was higher than that of the blend type 1. Ā e reasoning for the above properties was as follows: Ā e differences i n t he e xpansion o f t he PV C ( rubber state) and the PMMA (glass state) at high temperature (395 K) concentrated the warp stress at t he interface and decreased the strength o f t he i nterface. H owever, i n blen d t ype 2 , t he ph ase containing a n excellent w ide concentration g radient prevented the wa rp st ress f rom co ncentrating. Ā us, t he t hermal s hock resistance of the blend (blend type 2) with excellent wide concentration gradient was found to be superior to that of the similar blend (blend type 1) to a l aminate lm. Ā e formation of an excellent w ide c oncentration g radient w as f ound to h ave improved the strength of the interface. 3.4.1.3 Smart Performance (DMA Properties) Ā e c hange i n ten sile s torage mo dulus a nd t an d of t he PVC/ PMMA blend type 2 with a wide concentration gradient around Tg w as c ompared w ith t he p erfectly m iscible blen d (5/5) u sing the DMA me asurement (temperature i ncreasing r ate: 1 K/min, frequency: 0.2 Hz). Ā e Tg width measurements of storage modulus and half temperature of Tg width of tan d were estimated as shown in Table 3.3. Ā e h alf w idth o f t he tem perature o f t an d f or t he f ormer (16 K) was signicantly larger than that of the latter (10 K). Ā us, the blend type 2 was conrmed to be a continuous phase, having a wide range of Tg.

TABLE 3.3 Properties of PVC/PMMA Functionally Graded Blends PVC/PMMA Blend Unit

Type 2 a

(kgf/mm2) (%) (kgf/mm2)

6.4 4.5 200

4.5c 2.8c 190

7.2 5.2 220

(K) (K)

20 16

8.6, 11c —

(degree) (kgf)

9 98

170d 71d

Properties Tensile properties Tensile strength Elongation at break Tensile modulus of elasticity DMA properties (tensile mode) Tg width of storage modulus Half-temperature width Tg in tan d Ā ermal shock resistance Maximum warp angle Adhesive strength in shear by tension loading

Type 1

PMTb

PVC Only

PMMA Only

5.7 3.9 230

6.1 3.1 230

11 10

— —

— —

— —

— —

— —

Blend containing graded structure 2. Perfectly miscible blend. c Prepared by the hot press method. d Blend containing graded structure 1. a

b

43722_C003.indd 9

10/3/2008 8:43:42 PM

3-10

Smart Materials

Maximum angle

FIGURE 3.20 Measurement method of maximum angle of warp.

As a re sult, t an d o f t he g raded blen ds o f PVC a nd s everal types of polyalkyl methacrylate (PMA) containing graded structure 2 w ere me asured, a s s hown i n F igure 3 .21. T an d of t he graded PV C/PHMA blen d h ad t he w idest tem perature r ange. Ā erefore, it was concluded that the wide temperature range was caused by the greater differences of Tg in the pair polymers of the graded PVC/PHMA blend. Finally, we i nvestigated t he opt imum c onditions for preparing t he g raded PVC/PHMA blend w ith t he w ider temperature range of tan d. We obtained a PVC/PHMA blend containing an excellent graded structure 2, which showed a peak of tan d in a Solution volume 0.455 ml/cm2

much wider temperature range in comparison with those of the blend containing t he graded structure 1 a nd perfectly miscible blend (5/5), as shown in Figure 3.22. In both systems of the PVC/PMMA and the PVC/PHMA blends, we found that the tensile storage modulus of the blend type 2 contained an excellent graded structure 2 t hat began to de crease at a lower temperature t han t hat of t he perfectly m iscible blend (5/5) and did not have a terrace, while that of a similar blend containing the graded structure 1 in a laminate had some terraces. Sandwich steel beams combining a polymer have been used as damping materials [46]. It is known that the damping efficiency is maximum i n t he temperature range at w hich t he used polymer has a peak of tan d. Ā erefore, it is expected that an excellent graded blend with a peak of tan d in a much wider temperature range i s u seful a s a d amping m aterial i n a l arge tem perature range. Ā is is what the graded polymer blend is expected to be for smart materials. Ā e following principle reects the reasoning for this condition for smart materials. An excellent graded blend was used for the polymer combined with the steel plates, as shown on the right of Figure 3.23. Tg of the graded blend decreases when shift ing occurs from the left to the right side. At the highest temperature, that is, the same temperature as the higher Tg of the pair polymers in the blend, the area in the farthest left side shows the highest and best damping performance. As the area shifts to t he r ight side , t here i s a decrease of temperature. Finally, at the lowest temperature, that is, the same temperature as the lower Tg of the pair polymers in the blend, t he a rea on t he farthest right side s hows t he highest

DMA

PVC/PHMA

PVC/PMMA

Graded structure 1

Increase of tan d

PVC/PBMA

Increase of strage modulus

Increase of tan d

Increase of storage modulus

PVC/PMMA

Graded structure 2

PVC/PBMA

Graded structure 3

PVC/PHMA

230

270

310

350

390

Temperature (K)

FIGURE 3.21 DMA data of PVC/PMA graded blends.

43722_C003.indd 10

200

240

280

320

360

Temperature (K)

FIGURE 3.22 DMA data of PVC/PHMA graded blends.

10/3/2008 8:43:42 PM

3-11

Functionally Graded Polymer Blend

Medium temp. Lower temp. Most superior area in damping property at each temperature

Most superior area in damping property

Weight loss of PLLA (wt%)

100

Steel plates

Temperature

Higher temp.

80 60 40 20 0 0

FIGURE 3.23 Schematic model of the so-called smart performance in the d amping prop erty of s teel pl ate c ombined w ith a f unctionally graded blend.

and b est e xcellent d amping p erformance. Ā erefore, t he a rea showing the high damping performance shifts with the changing temperature. Ā is performance is thus considered as one of the so-called “smart performances.”

3.4.2 Functional and Smart Performances of PEO/PLLA Graded Blend PLLA is known as biomass and biodegradable polymer and its biodegradability was enhanced by blending with PEO. Ā us, we prepared the PEO/PLL A graded blend, which was exp ected to have both hig h b iodegradability a nd t ensile str ength in the v ertical direction o f thic kness [26]. Ā e va rious typ es o f grade d PLL A/ PEO blends, homogeneous blend, and PLLA only were subjected to degrada tion b y a p roteinase K enzyme . Ā e biodegradability

2

4

6 8 Past time (day)

10

12

Graded blend type 1

Graded blend type 2

Graded blend type 3

PLLA only

Homogeneous blend

FIGURE 3.24 Changes of net weight loss of PLLA in various types of graded blends, homogeneous blend, and PLLA only, in the biodegradation test.

was evaluated by the net weight loss of PLLA calculated by difference of weightbefore and after testing. Ā e net weight loss in all of the g raded blends was hig her tha n thos e in the ho mogeneous blend a nd the PLL A o nly (Figur e 3.24). Ā e net w eight loss o f graded blend type 2 wi th the b est excellent graded structure was the highest. Ā us, the graded structure was co nsidered to la rgely increase biodegradability. It was conrmed by an SEM observation of thes e ma terials (Figur e 3.25). Ā e net w eight loss o f all the graded blends, w here a p orous str ucture formed b efore the b iodegradation test, did not result in increase to a great extent. It was

(a)

PLLA

(b)

Homogeneous blend

(c)

Graded blend I

(d)

Graded blend II

FIGURE 3.25 SEM photograph of the cross-section of the sample after biodegradation test for 11 days.

43722_C003.indd 11

10/3/2008 8:43:43 PM

3-12

Smart Materials TABLE 3.4 Tensile Strength of Several Types of Materials

Elastic modulus (mPa)

Materials

25

Tensile Strength (MPa)

Graded blend 1 Graded blend 2 Graded blend 3 Homogeneous blend PLLA only

2.46 3.05 3.20 1.78 2.88

15 10 5 0 0

considered as t he f ollowing: PEO, no t o nly diss olved in to wa ter resulting in increasing the surface area attacked by the enzyme, but also, absorbed the PLL A oligomers that acted as acid ca talysts of water decomposition and subsequently, it promoted the decomposition of PLLA. Furthermore, the strengths of all the graded blends were la rger tha n tha t o f the ho mogeneous b lend (T able 3.4). Ā erefore, it was concluded that the graded structure enhanced the biodegradability while maintaining its high tensile strength.

3.4.3 Functional and Smart Performances of PEO (or PEO/LiOCl4)/PBMA Graded Blend PEO/PBMA graded blend was prepared by the polymerization– diff usion method [30]. It was found that the water vapor permeability ac ross t he g raded blen d  lms w as d ifferent i n t he PEO Permeation direction in the graded blend PEO rich

Laminate (2 layers) Graded composite 1 Graded composite 2

20

10

20

30

40

50 60 D/d (%)

70

100

content on the side of the higher pressure of water vapor (humidity: 90%) at 313 K. Ā us, the permeability of the vapor from the PEO-rich side (direction 1) was always higher than that of the vapor f rom t he P BMA-rich side ( direction 2 ) ( Figure 3 .26). Ā erefore, i t w as f ound t hat t he g raded s tructure c ould c ause anisotropy in the permeability of the water vapor. (PEO/LiClO4)/PBMA g raded blen d w as p repared u sing EO-b-BMA blo ck c opolymer b y t he d iffusion–dissolution method [28]. Ā e change i n t he elastic modulus of t he g raded blend i n t he t hickness d irection w as e stimated a s s hown i n Figure 3.27. Ā e modulus of the graded blend decreased gradually around 45%, while that of the laminate began to decrease immediately (Fi gure 3. 28). Ā us, i t w as c oncluded t hat t he graded blend was prevented from its breakaway at the interface

PBMA rich

10−2 10−4

2

10−6 10−8

Test condition: Temp. 313 K, humidity 90%

Laminate (2 layers)

10−10 10−12 10−2

0.18

Direction 1

0.16

Direction 2

Electric conductivity (S/cm)

0.2

Permeability of water vapor (g/mh)

90

FIGURE 3.27 Tensile elastic modulus of several types of materials in the thickness direction (D/d: (Distance from petri glass side)/(thickness of sample) ).

1

0.14 0.12 0.1

10−4 10−6 10−8 Graded blend 1

10−10 10−12 10−2 10−4

0.08 10−6

0.06

10−8

0.04

0

Graded blend 2

10−10

0.02

10−12 0

0

5

10

15

20

25

30

Past time (h)

FIGURE 3 .26 Permeability of w ater v apor of P EO/PBMA g raded blend in both types of the thickness directions.

43722_C003.indd 12

80

20

40

60

80

100

D/d (%)

FIGURE 3 .28 Electric c onductivity of s everal t ypes of m aterials i n the thickness direction (D/d: (Distance from petri glass side)/(thickness of sample) ).

10/3/2008 8:43:44 PM

3-13

Functionally Graded Polymer Blend TABLE 3.5 Possibility of Applications of Functionally Graded Polymer Blend Expected Functional Property Relaxation of thermal stress

Prevention of vibration and sound Electromagnetic materials Photomaterials Medical materials Packing materials Chemical

Application Mechanical device for antiabrasion Sporting goods Construction materials Vibration and sound proof Electromagnetic shield Copy machine device Optic ber Lens Articial internal organs Articial blood vessels and organs Waterproof adhesive Chemical resistance materials

by its graded structure. However, all of the electric conductivities o f t he g raded blen ds a nd t he l aminate b egan to si milarly increase immediately around 40%.

3.5

Prospects for Application in the Functionally Graded Blends

Functionally graded polymer blends are expected to be used in the future as a substitute for laminates because of their superiority in their strength and thermal shock resistance. Ā e superiority, we believe, is caused by the lack of interface, which suppresses the break at t he interface and the thermal stress. Furthermore, the excellent wide compositional gradient for the graded structure results in improvement in several types of physical properties. Ā erefore, some new functional performances are expected, which could be caused by an improved physical property gradients, a nd c an b e app lied i n v arious t ypes o f app lications, a s shown in Table 3.5.

References 1. Society o f Functionally G raded Material, e d., Functionally Graded Materials, Kogyo Chyosakai, Tokyo, 1993. 2. T. Kitano, Kogyo Zairyo, 43(6),112, 1994. 3. M. T akayanagi, Ā e 23r d C olloquium o f S tructure a nd Property of Polymer, Tokyo, 1993. 4. Y. Agari, Func Grad.. Mater., 16(4), 32, 1996. 5. Y. Agari, Koubunshi Kako, 46(6), 251, 1997. 6. Y. Agari, M. S himada, A. U eda, a nd S. N agai, Macromol. Chem. Phys., 2017, 1996. 7. Y. Agari, M. S himada, A. U eda, T. Anan, R . N omura, a nd Y. Kawasaki, Func. Grad. Mater., 1996, 761, 1997. 8. Y. Agari, M. Shimada, M.Ueda, R. Nomura, and Y. Kawasaki, Polym. Preprints, 47, 701, 1998. 9. Y. Agari, M. S himada, H. S hirakawa, R . N omura, a nd Y. Kawasaki, Polym. Preprints, 48, 698, 1999.

43722_C003.indd 13

10. J. Z. Yu, C. L ei, a nd F. K. K o, Annu. Tech. C onf. S PE, 52, 2352, 1994. 11. M. Omori, A. Okubo, K. Gilhwan, and T. Hirai, Func. Grad. Mater., 1996, 764, 1997. 12. M. Funabashi, T. Kitano, Seni-Gakkaishi, 50(12), 573, 1994. 13. C. M. Ā ai, T . K ato, a nd A. Yoshizumu, J. Ā er mosetting Plastics, Japan, 16(3), 126, 1995. 14. X. M. Xie, M. Matsuoka, and K. Takemura, Polymer, 33(9), 1996, 1992. 15. S. Kanayama and T. Umemura, Seikei Kako, 2(4), 216, 1995 16. Y. Kano, S. Akiyama, H. Sano, and H. Yui, J. Electron Microsc., 44(5), 344, 1995. 17. S. Akiyama and Y. Kano, Kagaku To Kogyo, 71, 44, 1997. 18. S. M urayama, S. K uroda, a nd Z. O sawa, Polymer, 34(18), 3893, 1993. 19. E. J abbari a nd N. A. P eppas, Macromolecules, 26, 2175, 1993. 20. P. F. Nealey, R. E. Cohen, and S. Argon, Macromolecules, 27, 4193, 1994. 21. K. C. F arinas, L. D oh, S. Venkatraman, a nd R . O . P otts, Macromolecules, 27, 5220, 1994. 22. T. E. Shearmur, A. S. Clough, D. W. Drew, M. G. D. van der Grinten, and R. A. L. Jones, Macromolecules, 29, 7269, 1996. 23. M. A. Parker and D. Vesely, J. Polym. Sci., Part B Polym. Phys., 24, 1869, 1986. 24. Y. Agari, M. S himada, A. U eda, T. K oga, R . N omura, a nd Y. Kawasaki, Polym. Preprints Jpn., 45, 2241, 1996. 25. Y. Agari, M. S himada, A. U eda, T. K oga, R . N omura, a nd Y. Kawasaki, Polym. Preprints Jpn., 46, 657, 1997. 26. Y. Agari, Y. Kano, K. Sakai, and R. Nomura, Polym. Preprints, Jpn., 51, 1065, 2002. 27. Y. Agari, Y. Anan, R. Nomura, and Y. Kawasaki, Polymer, 48, 1139, 2006. 28. Y. Agari, T. Morita, M. Shimada, and R. Nomura, J. Jpn. Soc. Col. Mater., 75, 474, 2002. 29. Y. Agari, K.Ohishi , M. S himada, a nd R . N omura, Polym. Preprints Jpn., 51, 647, 2002. 30. Y. Agari, T. Yamamoto, a nd R . N omura, Kagaku t o Ko gyo (Japan), 80, 138, 2006. 31. M. Kryszewski and G. Czeremuszkin, Plaste u Kautchuk, 11, 605, 1980. 32. P. Milczarek and M. Kryszewski, Coll. Polym. Sci., 265, 481, 1987. 33. Y. K oike, H. H idaka, a nd Y. Oh tsuka, Appl. O pt., 22, 413, 1983. 34. Y. Koike, N. Tanio, E. Nihei, and Y. Ohtsuka, Polym. Eng. Sci., 29(17), 1200, 1989. 35. C. F. J asso a nd E. M endizabal, Annu. Tech. C onf. S PE, 50, 2352, 1992. 36. S. Ashai, Polym. Preprints, Jpn., 27, 18, 1978. 37. S. Ashai, Ā e 6th Symposium Functionally Graded Materials, 61, 1993. 38. S. Ashai, Kagaku To Kogyo, 71, 50, 1997. 39. Y. Tsukahara, N. N akamura, T. H ashimoto, a nd H. K awai, Polym. J., 12(12), 455, 1980.

10/3/2008 8:43:45 PM

3-14

40. D. G reszta, K. M atsuoka, a nd K. M atyaszewski, Polym. Reprints ACS, 37, 569, 1996. 41. M. Furukawa, T. Okazaki, and T. Yokoyama, Polym. Preprints, Jpn., 45, 2239, 1996. 42. G. B . P ark, M. H irata, Y. K agari, T. M atsunaga, J . G ong, Y. O sada, a nd D. C. L ee, Polym. Pr eprints, Jp n., 45, 1836, 1996.

43722_C003.indd 14

Smart Materials

43. Y. Ulcer, M. Cakmak, and C. M. Hsiung, J. Appl. Polym. Sci., 60(1), 125, 1996. 44. Y. Agari and A. Ueda, J. Polym. Sci., Part B Polym. Phys., 32, 59, 1994. 45. M. S himada, Y. Agari, a nd Y. M akimura, Polym. Pr eprints, Jpn., 45, 1958, 1996. 46. D. J. Mead, J. Sound Vib., 83, 363, 1982.

10/3/2008 8:43:45 PM

4 Structural Application of Smart Materials 4.1 I ntroduction ...............................................................................................................................4-1 4.2 M aterials and Application .........................................................................................................4-1 Materials

R. Sreekala Structural Engineering Research Centre, CSIR

K. Muthumani Structural Engineering Research Centre, CSIR

4.3 S tructural Uses ...........................................................................................................................4-2 Active Control of Structures • Passive Control of Structures • Hybrid Control • Smart Material Tag • R etrotting • S elf-Healing • S elf-Stressing for Active Control • S tructural Health Monitoring • Active Railway Track Support • Active Structural Control against Wind

4.4 C onclusion ..................................................................................................................................4-7 Acknowledgment .................................................................................................................................4-7 References .............................................................................................................................................4-7

4.1 Introduction Ā e development of durable and cost-effective high-performance construction m aterials a nd s ystems i s i mportant f or t he e conomic well-being of a c ountry mainly because t he cost of civil infrastructure constitutes a major portion of the national wealth. To add ress t he p roblems o f de teriorating c ivil i nfrastructure, research is very essential on smart materials. Ā e research a nd development pro jects a iming to appl y a dvanced t echnologies, such as new materials and new structural systems, can improve the performance of the buildings, reduce the expense of maintenance, and eventually ensure the future sustainability of the buildings. Ā is chapter highlights the use of smart materials for the optimal performance and safe design of buildings and other infrastructures particularly those under the threat of earthquake and ot her nat ural h azards. Sm art m aterials w ith emb edded desired f unctions suc h a s s ensing a nd p rocessing o r w ith improved s tructural p erformances suc h a s s trength, d uctility, usability, and low cost are the features that need to b e explored and applied in structures.

4.2

Materials and Application

4.2.1 Materials 4.2.1.1

Shape Memory Alloys

Ā e term shape memory alloys (SMA) refers to the ability of certain a lloys (Ni– Ti, Cu– Al–Zn, e tc.) to u ndergo l arge s trains,

while re covering t heir i nitial conguration at t he en d o f t he deformation process spontaneously or by heating without any residual d eformation. Ā e particular properties of SMAs are strictly a ssociated w ith a so lid–solid p hase t ransformation, which can be t hermal or stress-induced. Currently, SMAs are mainly applied i n me dical s ciences, ele ctrical, aerospace, a nd mechanical eng ineering a nd t he re cent s tudies i ndicate t hat they can open new applications in civil engineering specically in seismic protection of buildings. Properties, w hich ena ble Ni– Ti w ires f or c ivil eng ineering application, are as follows: 1. Repeated absorption of large amounts of strain energy under loading without permanent deformation 2. Possibility to ob tain a w ide r ange o f c yclic b ehavior from supplemental a nd f ully recentering to h ighly d issipating b y si mply v arying t he c haracteristics o f S MA components 3. Usable strain range of 70% 4. Extraordinary fatigue resistance under large strain cycles 4.2.1.1.1 Substitute for Steel? It is reported that the fatigue behavior of Cu–Zn–Al SMAs is comparable with steel [1]. If larger diameter rods can be manufactured, it has a p otential for use in civil eng ineering applications. Use of  ber-reinforced plastics with SMA reinforcements requires future experimental investigations. 4-1

43722_C004.indd 1

10/15/2008 4:11:11 PM

4-2

4.2.1.2

Smart Materials

Piezoelectric Materials

4.2.1.4 Electro- and Magnetorheological Fluids

Piezoelectricity was discovered as early as 1880 by Curie brothers [2]. When integrated into a s tructural member, a p iezoelectric material generates an electric charge or voltage in response to mechanical forces or stresses. Ā is p henomenon i s c alled direct piezoelectric effect. Ā is is useful for sensing. Ā e converse piezoelectric e ffect c an b e u sed f or c ontrol. A mong t he w ide variety of sensing and actuation devices, the advantages of using piezoelectric ac tuators a nd s ensors i nclude t heir e ffectiveness over a wide frequency range, simplicity, reliability, compactness, and lightweight. Even though lot of progress has been made in the laboratory level structural testing, their usage in large-scale civil eng ineering s tructures i s l imited d ue to h igh-voltage requirements.

Electro- and magnetorheological (ER/MR) uids have essential characteristics t hat cha nge f rom f ree-owing, l inear v iscous uid, to a s emisolid w ith a c ontrollable y ield strength in milliseconds when exposed to a n electric a nd magnetic  eld. Ā ese uids are variable contenders for the development of controllable devices. It is intended to develop a structure that controls its stiff ness and damping characteristic to behave adaptively against earthquake or w ind forces a nd ac hieve s afety a nd f unction by using E R/MR de vices w ith le sser energ y. A l arge-scale M R damper of 20 t capacity has been constructed and tested [3] and the experimental results indicate the applicability of MR damper in the semiactive control mode in the real-world applications.

4.2.1.3

Induced strain actuators (ISA) change their own shapes according to external electric and magnetic elds and vice versa. Ā ese materials have been widely used for small and precision machines because of some advantages such as small sizes, rapid reaction, high power, and high accuracy, etc. ISA materials act as sensors because they cause change in electric or magnetic elds under deformation. ISA m aterials c ould b e u sed to de velop sm art members to realize smart, comfortable, and safe structures. ISA materials c an b e a lso ut ilized f or v ibration mo de c ontrol o f structural memb ers a nd s ensor de velopment. Ā ey a re w idely used in active oor vibration control. Possible applications of ISA include

Engineered Cementitious Composites

Engineered cementitious composites (ECC) is mortar or concrete reinforced by chopped ber. Such composite materials have been m icrostructurally de signed u sing m icromechanical p rinciples. ECC exhibits strain hardening with large strain capacity and s hear d uctility, a nd go od d amage-tolerant me chanical behavior. Ā e use of high-performance cementitious structural elements a s energ y d issipation de vices a nd d amage-tolerant elements h elps to ac hieve a b etter p erforming a nd a d amagetolerant s tructural s ystem. D evelopment of c oncrete e ncased steel column elements also comes under this category. 4.2.1.3.1

Carbon Fiber Reinforced Concrete

Its ability to c onduct electricity and most importantly capacity to change its conductivity with mechanical stress makes a promising material for smart structures. It is evolved as a part of densied reinforced composites (DRC) technology. Ā e high density coupled w ith a c hoice o f  bers r anging f rom s tainless s teel to chopped carbon a nd kelvar, applied under high pressure, g ives the product outstanding qualities as per DRC technology. Ā is technology makes it p ossible to pro duce surfaces with strength and durability superior to metals and plastics. 4.2.1.3.2

Smart Concrete

A mere add ition of 0.5% specially treated carbon bers enables the increase of electrical conductivity of concrete. Putting a load on this concrete reduces the effectiveness of the contact between each ber and the surrounding matrix and thus slightly reduces its conductivity. On removing the load, the concrete regains its original c onductivity. B ecause o f t his p eculiar p roperty, t he product is called “smart concrete.” Ā e concrete could serve both as a structural material as well as a sensor. Ā e smart concrete could function as a traffic-sensing recorder when used as road pavements. It has got higher potential and could be exploited to make concrete reective to radio waves and thus suitable for use in electromagnetic shielding. Ā e smart concrete c an b e u sed to l ay sm art h ighways to g uide s elf-steering cars, which at present follow tracks of buried magnets. Āe strainsensitive concrete might even be used to detect earthquakes.

43722_C004.indd 2

4.2.1.5

Induced Strain Actuators

1. Long span structure 2. Axial force and friction control (for base isolator, including trigger application) 3. Active sound transparency (noise control) [4] 4. Wireless sensors of deformation

4.3 4.3.1

Structural Uses Active Control of Structures

Ā e concept of adaptive behavior has been an underlying theme of active control of structures, which are subjected to earthquake and other environmental type of loads. Ā e structure adapts its dynamic c haracteristics to me et t he p erformance objectives at any instant. A f uturistic smart bridge system (an artist’s rendition) is shown in Figure 4.1 [5]. A thermomechanical approach to develop a constitutive relation f or b ending o f a c omposite b eam w ith c ontinuous S MA bers embedded eccentric to neut ral a xis was used by Sun and Sun [6]. Ā e authors c oncluded t hat SMAs c an b e suc cessfully used for the active structural vibration control. Ā ompson et al. [7] also conducted an analytical investigation on the use of SMA wires to d ampen t he dy namic re sponse o f a c antilever b eam constrained by SMA wires. To date, active structural control has been successfully applied to over 20 commercial buildings and more than 10 bridges. One example of active control is the Kurusima bridge in Shikoku area

10/15/2008 4:11:12 PM

4-3

Structural Application of Smart Materials

Epoxy binders (matrices) Multidirectional carbon or other fibers Actuators Feedback control

Cable with sensors Data acquistion (from sensors) and processing

Dampers Pier sensors Wireless sensors

Protective coating Advanced composite materials

Sensors for intelligent vehicles Composite pipes for de-icing bridge deck (with geothermal energy)

Structural controls Stream sensors

Optical fiber sensors

FIGURE 4.1 Futuristic smart bridge system (an artist’s rendition). (Courtesy: USA Today, 3 March 1997.) J. Holnicki-szulc and J. Rodellar (eds.), Smart Structures, Requirements and potential applications in mechanical and civil engineering, NATO Science Series 3. High Technology, Vol. 65, Springer 1999.

in Japan. Ā e bridge was designed with the application of active vibration c ontrol a s i ntegrated s tructural c omponents. S everal modes of the bridge tower, which were anticipated to be excited by w ind vo rtex, w ere c arefully p rotected b y app ropriate controllers during the construction phase. It therefore made it possible for the tower of this bridge to be built much lighter and more s lender t han o ne f ollowing t raditional de sign. A ctive tuned mass dampers have been installed in the 11-story building, t he Kyobashi S eiwa building i n Tokyo (the  rst full-scale implementation of active control technology) a nd t he Nanjing Communication Tower in Nanjing, China. Ā ere a re t wo s erious c hallenges t hat rem ain b efore ac tive control c an ga in gener al ac ceptance b y eng ineering a nd construction pr ofessionals at l arge. Ā ey a re (1) Re duction o f capital c ost a nd m aintenance (2) I ncreasing s ystem rel iability and robustnessActive control systems consist a set of sensors, a controller, an active control system (actuators), and an external power supp ly. A s chematic s ketch o f ac tive c ontrol s ystem i s shown in Figure 4.2. Nowadays, m uch w ork o n s tructural c ontrol i s f ocused o n intelligent structures, developments of ac tuating materials a nd piezoceramics. D ue to i ts l imited f requency ba ndwidth, S MA has traditionally been used for passive strategies such as dampers or other types of energy dissipation devices instead of being used in active control strategies as actuators. On the other hand, it i s w ell k nown t hat a S MA ac tuator i s c apable o f p roducing relatively large control forces despite its slow response time. Ā is unique c haracteristic of SMA m akes it very at tractive for c ivil

43722_C004.indd 3

Controller

Sensors

Sensors

Power Active control

Excitation

Structure

Response

FIGURE 4.2 Schematic sketch of active control system.

engineering c ontrol app lications w here l arge f orces a nd lo wfrequency band width are mostly encountered. Ā ey conducted experimental s tudies o n ac tive c ontrol o f a  ve-story building model with SMA actuators [8]. It was proven that despite its slow response, it is feasible to use SMA for active control of civil engineering s tructures. But w hile s electing t he a lloy t ype, ut most care should be taken in specifying the temperature range or the transition temperature or the process in alloy-making so as to suit our requirements. Dynamics of SMA should be considered in control design.

4.3.2 Passive Control of Structures Two f amilies o f pa ssive s eismic c ontrol de vices e xploiting t he peculiar properties of SMA kernel components have been

10/15/2008 4:11:12 PM

4-4

Smart Materials

implemented a nd te sted w ithin t he M emory A lloys f or N ew Seismic Isolation a nd Energy Dissipation Devices (MANSIDE) project. Ā ey are special braces for framed structures and isolation devices for buildings and bridges.

4.3.3

Hybrid Control

Ā e term hybrid control generally refers to a combined passive and active control system. Since a portion of the control objective is accomplished by t he pa ssive system, less ac tive control effort, i mplying le ss p ower re source, i s re quired. Si milar control resource savings can be achieved using the semiactive control scheme where the control actuators do not add mechanical e nergy dir ectly t o t he s tructure, h ence b ounded-input/ bounded-output st ability i s g uaranteed. S emiactive co ntrol devices are often viewed as controllable passive devices. A side benet of hybrid and semiactive control systems is that, in the case of a p ower failure, the passive components of the control still offer some degree of protection unlike a fully active control system. Ā e materials described above can be used to make the h ybrid c ontrol s cheme w orkable. I n t he c ase o f E R/MR dampers, hybrid control scheme is a viable option for realistic structural control.

4.3.4 Smart Material Tag Ā ese sm art m aterial t ag c an b e u sed i n c omposite s tructures. Ā ese tags can be monitored externally throughout the life of the structure to r elate t he i nternal ma terial co ndition. S uch m easurements as stress, moisture, voids, cracks, and discontinuities may be interpreted via a remote sensor [6].

4.3.5 Retrofitting SMAs ca n us ed as s elf-stressing  bers a nd t hus t hey c an b e applied for retrotting. Self-stressing bers are the ones in which reinforcement is placed into the composite in a nonstressed state. A prestressing force i s i ntroduced i nto t he s ystem w ithout t he use o f l arge me chanical ac tuators b y p roviding S MAs. Ā ese materials d o n ot n eed specia lized e lectric eq uipments n or d o they create safety problems in the eld. Treatment can be applied at a ny t ime a fter hardening of t he matrix i nstead of during its curing and hardening. Long- or short-term prestressing is introduced by triggering the change in SMA shape using temperature or e lectricity. Ā ey ma ke ac tive la teral co nnement of b eams and columns a more practical solution. Self-stressing jackets can be manufactured for rehabilitation of existing infrastructure or for new construction. 4.3.5.1

Restoration of Cultural Heritage Structures Using SMA Devices

An innovative technique using superelastic SMA devices for the restoration o f a c ultural h eritage s tructure e specially m asonry buildings were implemented under the framework of the European Commission-funded IS TECH P roject. M asonry b uildings a re

43722_C004.indd 4

largely vulnerable to earthquakes because of their low resistance and ductility during earthquake ground motion. To enhance the seismic b ehavior o f c ultural h eritage s tructures, t he mo st common method traditionally used has been the introduction of localized reinforcements. Usually steel bars or cables served this purpose by increasing stability and ductility. But in many cases, these rei nforcement te chniques p rove i nadequate to p revent collapse. Ā e development of the connection technique was based on the ide a o f usin g t he uniq ue p roperties o f N i–Ti a lloys esp ecially i ts su per elastici ty a nd hig h r esistance t o co rrosion [9]. Ā e idea was to connect the external walls to the oors, the perpendicular walls o r the r oof with an SMA Device that should behave as follows: 1. Under lo w-intensity h orizontal ac tions ( wind, sm all intensity earthquakes), the device remains stiff, a s traditional s teel c onnections do, not a llowing sig nicant displacements. 2. Under h igher i ntensity h orizontal ac tions ( i.e., s trong earthquakes), the stiff ness of the device decreases, allowing “ controlled d isplacements,” w hich s hould re duce amplication of a ccelerations (as c ompared t o s tiff connections) and permit the masonry to dissipate part of the transmitted energy, mainly owing to elasticity exploitation and microcracks formation in the brick walls; consequently, the s tructure s hould b e a ble to su stain a h igh-intensity earthquake w ithout c ollapse, t hough u ndergoing s ome minor damage. 3. Under e xtraordinary h orizontal ac tions, t he s tiffness of the device increases and thus prevents instability. 4.3.5.2 SMA for Seismic Retrofit of Bridges Unseating o f supp orts w as t he m ajor c ause o f b ridge f ailures during earthquakes. Retrot measures to re duce the likelihood of collapse due to unseating at the supports have been in place for m any ye ars. Ā e d amage to b ridges i n t he re cent C hi-Chi, Kobe, and Northridge earthquakes indicate the need to provide better methods of reducing the damaging effects of earthquakes in b ridges. Ā e u se o f re strainer c ables a nd re strainer ba rs to limit t he rel ative h inge d isplacement b ecame p opular i n t he United States following the collapse of several bridges due to loss of supp ort d uring t he 1 971 S an F ernando e arthquake. Re cent earthquakes have demonstrated that restrainers were effective in some c ases. However, m any bridges w ith re strainers su stained serious d amage o r c ollapse. Br idges t hat h ad b een re trotted with restrainer cables failed in both the 1989 Loma Prieta and 1994 N orthridge e arthquakes. F ailure o f J apanese r estraining devices a lso o ccurred d uring t he 1 995 K obe e arthquake. Experimental tests of restrainer cables have shown that failure occurs i n th e c onnection e lements o r th e th rough-punching shear in the concrete diaphragm. In addition, restrainers do not dissipate any signicant amount of energy, since they are generally designed to remain elastic. Analytical studies of bridge and restrainer systems have demonstrated that a very large number

10/15/2008 4:11:14 PM

4-5

Structural Application of Smart Materials

SMA restrainer bar/damper

Real system

(a)

Stay cable

SMA damping device

(b)

FIGURE 4.3 SMA restrainer bar used in multispan simply supported bridge abutments and intermediate piers conguration.

of re strainers a re often re quired to l imit jo int mo vement to acceptable le vels, pa rticularly f or h igh s eismic lo ads. I n t hose cases, t he e xcessive n umber o f re strainers w ould i nduce l arge forces in other components of the bridge, such as bearings and columns. Ā e shortcomings of traditional restrainers can potentially b e add ressed w ith t he u se of SMA re strainers. Ā e SMA restrainers, in the superelastic phase, act as both restrainers and dampers (Figure 4.3). Energy dissipation and base isolation are found to be optimal candidates for structural control of structures. As far as bracing systems a re c oncerned, u ntil now a ll t he app lications a nd t he research studies on this technique were focused on the energ y dissipation c apability. F or s eismic re trotting p urposes, t he supplemental recentering devices (SRCD) are found to be useful as t hey provide forces to re cover t he u ndeformed shape of t he structure at the end of the action. In existing structures, in fact, particularly when they were designed without any seismic provision, the energy dissipation can turn out to b e insufficient to limit damage to structural elements. It would then be necessary to strengthen some elements to f ully achieve the design objectives. L ocal s trengthening w ould i mply e xpensive w ork, a lso involving nonstructural parts. Retrotting could turn out to be economically in convenient, and, yet some residual displacement c ould o ccur i n c ase el astoplastic de vices a re u sed. A n alternative s trategy c an b e p ursued b y u sing S MA de vices having s upplemental forc e t o re cover t he u ndeformed s tructural conguration, resulting in the elimination of any residual displacement, while accepting y ielding i n structural elements. A comparison of properties of Ni–Ti w ith steel is g iven i n t he Table 4.1.

43722_C004.indd 5

TABLE 4.1 Comparison of Properties of Ni–Ti SMA with Typical Structural Steel Property

Ni–Ti SMA

Steel

Recoverable elongation (%) Modulus of elasticity (MPa)

8 8.7 × 104 (A) 1.4 × 104 (M) 200–700 (A) 70–140 (M) 900 (f.a.) 2000 (w.h.) 25–50 (f.a.) 5–10 (w.h.) Excellent

2 2.07 × 105

Yield strength (MPa) Ultimate tensile strength (MPa) Elongation at failure (%) Corrosion performance

248–517 448–827 20 Fair

Note: f.a. denotes fully annealed and w.h. denotes work hardened, w hich a re tw o typ es o f tr eatment gi ven t o the allo y. A and M deno te the tw o phases of the allo y namely, austenite and martensite.

4.3.6

Self-Healing

Experimentally proved self-healing behavior [10], which can be applied at microlevel of a material, widens their spectrum of use. Here signicant deformation beyond the  rst crack can be fully recovered and cracks can be fully closed.

4.3.7

Self-Stressing for Active Control

Self-stressing for ac tive c ontrol c an b e u sed w ith c ementitious ber composites with some prestress, which impart self-stressing

10/15/2008 4:11:14 PM

4-6

Smart Materials

thus avoiding difficulties due to the provision of large actuators in ac tive c ontrol, w hich re quire c ontinuous m aintenance o f mechanical pa rts a nd r apid mo vement, w hich i n t urn c reated additional inertia forces. In addition to SMAs, some other materials such as polymers can also be temporarily frozen in a prestrained state that have a potential to be used for manufacturing of self-stressing cementitious composites [1].

4.3.8

4.3.9

Active Railway Track Support

Active control system for sleepers is adopted [5] to achieve speed improvements on existing bridges and to maintain the track in a straight and nondeformed conguration as the train passes. With the help of optimal control methodology, the train will pass the bridge w ith re duced t rack de ections a nd v ibrations a nd t hus velocity could be safely increased. Figure 4.5 shows various positions of the train with and without active railway track support.

Structural Health Monitoring 4.3.10

Use o f p iezotransducers, su rface b onded to t he s tructure o r embedded in the walls of the structure can be used for structural health monitoring and local damage detection. Problems of vibration and UPV testing can be avoided here. Jones et al. [11] applied neural networks to  nd t he magnitude a nd location of an impact on isotropic plates and experimented using an array of piezotransducers surface bonded to the plate. Figure 4.4 shows a typical health monitoring setup making use of optic bers.

Active Structural Control against Wind

Aerodynamic control devices to mitigate the bidirectional wind induced v ibrations i n t all b uildings a re energ y e fficient, since the e nergy i n th e  ow i s u sed to p roduce t he de sired c ontrol forces. Aerodynamic ap system (AFS) is an active system driven by a feedback control algorithm based on information obtained from t he v ibration s ensors [ 5]. Ā e a rea o f  aps a nd a ngular amplitude of rotation are the principal design parameters.

Seismometer Vibration meter Optic fibers

Optic fiber

Optic fibers

Damage detection of piles using optic fibers

Vibration meter Hitting

Cracks Damage detection after damage occured

FIGURE 4.4 Health monitoring setup.

43722_C004.indd 6

10/15/2008 4:11:14 PM

4-7

Structural Application of Smart Materials

References

Bridge response without active railway track support

Bridge response with active railway track support

FIGURE 4.5 Active railway track support concept.

4.4

Conclusion

Ā e technologies using smart materials are useful for both new and existing constructions. Of the many emerging technologies available, t he f ew de scribed i n t his c hapter ne ed f urther research to e volve t he de sign g uidelines o f s ystems. C odes, standards, and practices are of crucial importance for the further development.

Acknowledgment Ā e a uthors gra tefully ac knowledge the Dir ector, S tructural Engineering Res earch C entre (S ERC), CS IR, I ndia f or the co nstant encouragement and support rendered in preparation of this manuscript a nd als o gi ving p ermission t o p ublish i t. Ā e kind support a nd guida nce o f all the t eam mem bers o f S tructural Dynamics Laboratory deserve acknowledgment.

43722_C004.indd 7

1. N. Krstulovic-Opara and A.E. Naaman, Self Stressing Fiber Composites, ACI Structural Journal, 97, March–April 2000, 335–344. 2. W.G. Cady, Piezo Electricity, Dover, New York, 1964. 3. G. Yang, B.F. Spencer Jr., J.D. C arlson, and M.K. Sain, L arge scale MR uid dampers: Modeling and dynamic performance considerations, Engineering Structures, 24, 2002, 309–323. 4. A. Sampath and B. Balachandran, Active control of multiple tones in an endosure, Ā e Journal of the Acoustical Society of America, 106-1, July 1999, pp. 211–225. 5. J. H olnicki-szulc a nd J . Ro dellar (e ds.), Smart St ructures, Requirements and potential applications in mechanical and civil engineering, NATO Science Series., 3. High Technology, Vol. 65, Springer 1999. 6. G. Sun and C.T. Sun, Bending of shape memory alloy reinforced composite beam, Journal of Materials Science, 30(13), 1995, 5750–5754. 7. P. Ā omson, G.J. Balas, and P.H. Leo, Ā e use of shape memory Alloys f or pa ssive s tructural d amping, Smart M aterials Structure, 4, 1995, 36–42, IOP publishing limited. 8. J. Li, B. Samali, and C. Chapman, Experimental realisation of active co ntrol o f a  ve st orey b uilding mo del usin g S MA actuators, Advances in Mechanics of Structures and Materials, Loo, Chowdhury and Fragomeni (eds.), 699–704. 9. M.G. Castellano and G. Manos, Ā e ISTECH Project: Use of SMA in t he Seismic Protection of Monuments, Monument– 98 (W orkshop o n S eismic P erformation o f M onuments), Lisbon, 1998. 10. D.J. Hannant and J.G. Keer, Autogeneous healing of Ti based sheets, Cement and Concrete Research, 13, 1983, 357–365. 11. R.T. Jones, J.S. Sirkis, and E.J. Friebele, Detection of impact location a nd ma gnitude f or is otropic p lates usin g neural networks, Journal o f I ntelligent M aterial S ystems a nd Structures, 7, 1997, 90–99.

10/15/2008 4:11:15 PM

43722_C004.indd 8

10/15/2008 4:11:16 PM

5 Composite Systems Modeling—Adaptive Structures: Modeling and Applications and Hybrid Composites 5.1 H ybrid Composites.................................................................................................................... 5-1 Introduction • Fut ure Directions

Acknowledgment ................................................................................................................................. 5-7 5.2 Design of an Active Composite Wing Spar with Bending–Torsion Coupling ..................5-7 Introduction • M ulticell Cross-Section Spar Design • R esults • C oncluding Remarks

References ........................................................................................................................................... 5-16

5.1 Hybrid Composites S. Padma Priya 5.1.1

Introduction

Composites a re materials made by t he synergistic assembly of two or more constituting materials (matrix and reinforcement), engineered in such a way that they form a single component and yet can be distinguished on a macroscopic level. Matrix encases the rei nforcements a nd ac ts a s a b inder f or t he  bers. Additionally, t he m atrix h olds t he rei nforcements i n a  xed position a nd t he reinforcement enhances t he properties of t he composite s ystem b y i mparting i ts me chanical a nd ph ysical properties. Composite materials can be engineered with desired properties by choosing and varying the concentration of different t ypes o f m atrices a nd rei nforcements. The d urability o f composites is very high; they have very high strength to weight ratios, a re re sistant to en vironmental c orrosion, a nd p rovide ease of use concept. Composites a re m aterials k nown f rom a ncient d ays, w hen bricks were made (mud reinforced with straw). Until today, we have applications for components of aerospace materials. Mother Nature i s t he  rst m aker of c omposites a nd s ome e xamples of natural composites include wood, bone, etc. Composites are classied into several broad categories depending upon the type of matrices and reinforcements used. Depending on the types of matrices used they can be classied as

1. Metal matrix composites (MMC) 2. Ceramic matrix composites 3. Polymer matrix composites Depending up on t he rei nforcement t ypes t hey c an b e c lassied as 1. Fiber reinforced composites 2. Fabric reinforced composites 3. Particulate composites In o rder to f urther en hance t he p roperties o f t he c omposite materials, hy brid c omposites we re d eveloped. The b ehavior of hybrid composites appears to be simply a weighted sum of the individual components in which there exists a more favorable balance between the advantages and disadvantages. The hybrid composites o ffer a dvantages r egarding s tructural in tegrity a nd sustained load under crash and impact conditions. Hybrid composites a re i nuenced b y a l arge n umber o f m icrostructural parameters such as type of reinforcement, volume proportion of reinforcement, weave pattern of the fabric used as reinforcement, volume fraction of the reinforcement, etc. There are several types of hybrid composites and can be characterized as: (1) interply or tow-by-tow, i n w hich to ws o f t wo o r mo re c onstituent t ypes o f ber a re m ixed i n a re gular o r r andom m anner; ( 2) s andwich hybrids, a lso k nown a s c ore-shell, i n w hich o ne m aterial i s sandwiched between t wo layers of a nother; (3) i nterply or laminated, w here a lternate l ayers o f t he t wo (or mo re) m aterials a re stacked in a regular manner; (4) intimately mixed hybrids, where 5-1

43722_C005.indd 1

10/3/2008 8:44:44 PM

5-2

Smart Materials

the constituent bers are made to mix as randomly as possible so that no overconcentration of any one type is present in the material; (5) other kinds, such as those reinforced with ribs, pultruded wires, thin veils of ber, or combinations of the above. Hybridization can be done through many ways as follows:

is its high fracture toughness, which is 50–100 times higher than that of epoxies. Another important advantage of PEEK is its low water absorption, which is less than 0.5% at 23°C compared to 4%–5% for conventional aerospace e poxies. B eing s emicrystalline, it does not dissolve in common solvents.

1. 2.

5.1.1.1.2 Natural Fibers

Reinforcements Matrices 3. Both reinforcements and matrices

5.1.1.1

Reinforcement Hybridized Composites

These are the composites, which consist of two or more types of reinforcements embedded in a single matrix. The reinforcements can be of different types as follows. 5.1.1.1.1 Synthetic (Man-Made) Fibers Th is c omposite i s c omposed o f t wo t ypes o f rei nforcements and both of them are synthetic. Composites of such types have been developed in order to achieve superior mechanical properties a nd h igh h eat re sistance c apabilities. Various t ypes o f synthetic re inforcements h ave b een d eveloped s uch a s g lass ber, carbon ber, Polyetheretherketone (PEEK), rayon, nylon, Kevlar, polyester, acrylic, ole n, vinyl, aramid, etc. The following bers can be hybridized (e.g., carbon ber/PEEK hybrid fabric, p olyethylene [ PE]/glass f abric, c arbon  ber/aramid bers) in diff erent (weight or volume) ratios according to t he desired en d p roperty. F or i nstance, i f le ss w eight h owever, a strong c omposite i s re quired, a h ybrid f abric c an b e c onsidered. Th is would have one fabric composed of different types of bers in it and some examples for this type of reinforcement are carbon/aramid, aramid/glass, and carbon/glass. Of the materials mentioned above the PEEK material is very much in demand. PEEK is a linear aromatic thermoplastic based on the following repeating unit in its molecules. O C

O

Ketone

Ether

O n Ether

Polyether ether ketone

Continuous-carbon  ber rei nforced P EEK c omposites a re known i n t he i ndustry a s a romatic p olymer c omposite (APC). PEEK is a semicrystalline polymer, and amorphous PEEK is produced when the melt is quenched. The presence of  bers in PEEK composites tends to increase the crystallinity to a higher level because the bers act as nucleation sites for crystal formation. Increasing crystallinity increases both the modulus and the yield strength of PEEK but reduces its strain to failure. PEEK’s maximum continuous-use temperature is 250°C. PEEK i s t he f oremost t hermoplastic m atrix t hat m ay re place epoxies in many aerospace structures. Its outstanding property

43722_C005.indd 2

A c omposite i n w hich i ts rei nforcements c ome f rom nat ural resources is considered natural. Natural  bers, because t hey a re naturally rene wable, c an b e i mbedded i nto b iodegradable p olymeric materials, which can produce a new class of biocomposites. Recently, society has realized that unless the environment is protected, it w ill be t hreatened due to t he loss of nat ural resources. Conservation o f f orests a nd opt imal ut ilization o f a gricultural and other renewable resources like solar energy and seawater has become almost mandatory. In this view, many attempts have been made to replace high-strength synthetic bers with natural bers. A si ngle nat ural  ber m ay not i mpart a ll t he p roperties f or t he nal composite, thus it is hybridized with other types of natural bers. S ome o f t he k nown nat ural  bers a re w ool, c otton, si lk, linen, he mp, r amie, jut e, c oconut, pi newood, pi neapple, a ngora, mohair, a lpaca, k apok,  ax, s isal, b ast  ber, k enaf, w ood  ber, bamboo, ba nana, e tc. To t ailor t he p roperties o f a nat ural  ber reinforced composites, it is hybridized with the second type of natural ber. Researchers have conducted studies on thermal conductivity and specic heat of jute/cotton, sisal/cotton, and ramie/ cotton hybrid fabric-reinforced composites. Small inorganic particles have also been used as  llers to prepare pa rticulate c omposites. The mo st re levant re ason to u se ller in the composite is to re duce the overall cost of the material. F illers a re a lso u sed to i ncrease  ame re tardancy, su rface hardness, e sthetic app eal, a nd t he t hermal p roperties o f t he composites. Some of the types of  llers that have been used are y a sh, si lica, m ica, g ranite p owder, w ood  our, aluminum oxide, carbon black, etc. Using such  llers and natural bers or fabric, a new class of hybrid composites has been produced. One of the reinforcements is  ller while the other is the  ber or the fabric, a nd h ere t he  ller pa rticles a re emb edded b etween t he brous rei nforcement a nd t he matrix. Other research developments underway have used sisal/saw dust in conventional composites and they possess an ease of processing, are environment friendly, and are economically affordable. 5.1.1.1.3 Natural and Synthetic Fibers Th is type of composite consists both of natural and synthetic reinforcements i n t he c omposite s ystem. N atural  ber reinforced composites do not always ful ll all the properties required for some technical applications. In such cases, hybridization w ith sm all a mounts o f s ynthetic  bers a llows t hese natural ber composites to be more suitable for technical applications. The m ost co mmon s ynthetic  ber u sed t o h ybridize the natural ber as a rei nforcement is glass  ber. Some examples of these types of bers are sisal/glass, silk/glass, pineapple ber/glass, jute/glass, etc. With the addition of a small amount of g lass f abric to t he nat ural  ber r einforced co mposite, t he

10/3/2008 8:44:44 PM

5-3

Composite Systems Modeling—Adaptive Structures: Modeling and Applications and Hybrid Composites

mechanical strength and the chemical resistance of the composite has been shown to i ncrease considerably (see Table 5.1 and Figure 5.1). 5.1.1.1.4 Miscellaneous Reinforcements Miscellaneous reinforcement is also done to achieve required properties in the composites and some of them are synthetic ber–particulate (glass along with y ash) composites used to increase t he abrasion resistance a nd compression strength of

the composites, blending various levels of glass to wollastonite allows for tailored composites with high-strength properties or go od d imensional s tability. F iber me tal l aminates ( FML) offer signicant improvements over currently available materials f or a ircraft s tructures d ue to t heir e xcellent me chanical characteristics a nd relatively low density a nd some examples for t hese t ypes o f c omposites a re a luminum 2 024 a lloy; carbon ber/epoxy (Ep), and aluminum 2024 alloy/glass fiber/ Ep composites.

TABLE 5.1 Properties of Selected Commercial Reinforcing Fibers

Fiber

Typical Diameter (mm)a

Specic Gravity

Tensile Modulus (GPa) 72.4 86.9

Glass E-glass S-glass

10 10

2.54 2.49

PAN carbon T-300c AS-1d AS-4d T-40c IM-7d HMS-4d GY-70e

7 8 7 5.1 5 8 8.4

1.76 1.80 1.80 1.81 1.78 1.80 1.96

Pitch carbon P-55c P-100c

10 10 11.9 1.47

Aramid Kevlar 49f Kevlar 149f Technorag Extended-chain PE Spectra 900 Spectra 1000 Boron SiC Monolament Nicalon (multilament)h Al2O3 FiberFPf Al2O3–SiO2i Fiberfrax (discontinuous)

Tensile Strength (GPa)

Strain to Failure (%)

Coefficient of Thermal Expansion (10−6/°C)b

Poisson’s Ratio

3.45 4.30

4.8 5.0

5 2.9

0.2 0.22

231 228 248 290 301 345 483

3.65 3.10 4.07 5.65 5.31 2.48 1.52

1.4 1.32 1.65 1.8 1.81 0.7 0.38

−0.6, 7–12

0.2

2.0 2.15

380 758

1.90 2.41

0.5 0.32

−1.3 −1.45

1.45 1.79 1.39

131 3.45 70

3.62 1.9 3.0

2.8

−2, 59

4.4

−6

38 27 140

0.97 0.97 2.7

117 172 393

2.59 3.0 3.1

3.5 2.7 0.79

5

140 14.5

3.08 2.55

400 196

3.44 2.75

0.86 1.4

1.5

20

3.95

379

1.90

0.4

8.3

2–12

2.73

103

1.03–1.72

−1.72

−0.75

0.35

0.2

1 mm = 0.0000393 in. 1 m/m per °C = 0.556 in./in. per °F. c Amoco. d Hercules. e BASF. f DuPont. g Teijin. h Nippon carbon. i Carborundum. a

b

43722_C005.indd 3

10/3/2008 8:44:44 PM

5-4

Smart Materials

Specific stiffness (tensile strength/density, in ⫻ 106)

Goal

(Kevlar 29) (Kevlar 49) aramid aramid 10

S glass Polyethelene

6

Boron Carbon (P-100)

4 E glass SiC (ceramic fiber)

2

0

200

Carbon (cellon GY-70)

Alumina (fiber FP)

600 400 800 1000 1200 Specific stiffness (tensile modulus/density, in ⫻ 106)

1400

Specic properties of advanced bers.

Matrix Hybridized Composites

The m atrix i s a lso o ne o f t he i mportant c onstituents o f t he composites. H ybridizing a m atrix w ith a nother m atrix a lso means toughening of t he matrix. Suc h matrices w hen u sed to prepare t he c omposites a lso c an le ad to a h ybrid c omposite. Toughening of the matrix is carried out in order to re duce the brittleness of the binder matrix and as a consequence to improve the (mechanical and chemical resistance) properties of the composites. An unmodied matrix usually consists of single-phase materials, while the addition of modiers turns the toughened matrix into a multiphase system. W hen modier domains a re correctly d ispersed i n d iscrete f orms t hroughout t he m atrix, the f racture energ y o r to ughness c an b e g reatly i mproved. I n the c ase of me tals, d ifferent types of alloys have been used to produce t he c omposites. P olymer blen ds, b y de nition, are physical m ixtures of s tructurally d ifferent homop olymers or copolymers. In polymer blends or polymer alloys, the mixing of two or more polymers provides a new material with a modied array of properties. For polymers, there are as known there are two types of matrices: thermoplastic (polymethyl methacrylate, polycarbonate [PC], polybutylene terephthalate [PBT], polystyrene [PS], etc.) and thermosets (Ep, unsaturated polyester, etc). Toughening of thermoplastics is usually carried out by blending them with different types of thermoplastic materials and elastomers i n order to i mprove t heir mechanical properties a nd t he exibility of t he matrix. E xamples i nclude blends of PC a nd a

43722_C005.indd 4

Ordered polymer fiber

(Kevlar 49 (dry))

8

0

5.1.1.2

PBO Carbon (T-300)

Titanium Aluminum HSLA Steel

FIGURE 5.1

PBT

Resin impregnated strands

12

thermoplastic copolyether/ester/elastomers, which is shown to have a signicant effect on t he properties. Polyoxymethylene/ elastomer/ ller ter nary co mposites ha ve b een p repared, i n which a thermoplastic polyurethane and inorganic ller, CaCO3, were used to achieve balanced mechanical properties for polyoxymethylene. Two other thermoplastics have also been blended together to achieve composites with high-performance properties as well as improve the mechanical properties of the composite: p olyamide ( PA) d ispersed i n a P E t hermoplastic a s a matrix for glass bers. Researchers have also found that there is a signicant increase in the degradation properties of thermoplastic blen ds w ith t he add ition o f c eramic m aterials ( PE-coethyl ac rylate w ith a p olyisobutyl me thacrylate p olymer reinforced with ceramic oxide powders). Some other examples of h ybrid t hermoplastic m atrices a re P C/acrylonitrile-butadiene-styrene ABS blend, PS/PC, PVA/polymethyl methacrylate, etc. Though a variety of thermoset materials have been used to prepare m atrix h ybrid c omposites, to ughening o f t hese c omposites by using Ep and unsaturated polyester are probably the most widely used combination. Different kinds of modiers have been studied to improve the toughness o r d uctility o f c ured t hermoset re sins. The y can be classied as liquid rubbers, engineered thermoplastics, reactive diluents, a nd i norganic pa rticles. Some of t he following examples illustrate this type of toughening: Ep toughened with polymethylmethacrylate ( PMMA), Ep to ughened w ith P C, Ep toughened with PBT, Ep with reactive liquid rubber such as Ep

10/3/2008 8:44:45 PM

5-5

Composite Systems Modeling—Adaptive Structures: Modeling and Applications and Hybrid Composites

phenol l iquid w ith c ashew nut s hells, u nsaturated PE w ith Ep, unsaturated P E to ughened w ith P MMA, u nsaturated P E w ith PBT, etc. Signicant increases in some mechanical- and chemical-resistant properties were found for all of the above combinations of cured thermoset resins. 5.1.1.3 Reinforcement and Matrix Hybridized Composites Composites of this type consist of multiphase materials (four phases) i n which both t he matrix a nd t he rei nforcements have been hybridized. In one way, it can be said that it is the combination of t he above t wo t ypes. This a rea of research is a ne w a nd challenging c oncept b ecause a s t he i ncorporation o f m aterials increases, t here a re mo re i nterphases i n t he c omposite s ystem and if there are any differences in the interfaces of the materials, it could lead to the failure of the complete system. Care should be always taken to meet these future challenges in order to achieve a material with superior mechanical properties. It should also be noted t hat t he v iscosity of t he toughened matrix resin be sufficient enough to wet the hybrid reinforcements. One of the examples that has been developed is silk/glass fabric and used as a reinforcement i n t hermoplastic-toughened Ep re sin to ac hieve high mechanical properties. The potential for the usage of hybrid composites far outweighs any ne gative a spects si nce t here a re a v ast n umber o f p lastic/ plastic, p lastic/metal ar amid-reinforced a luminum l aminate (ARALL), metal/metal (MMC), metal/ceramic, ceramic/ceramic (space shuttle outer tile), and ceramic/plastic composite combination systems yet to be explored, investigated, developed, and in use. 5.1.1.4 Hybrid Composite Applications Hybrid composite materials (HCM) represent the newest of the various c omposite m aterials c urrently u nder de velopment. The hybrid composite category covers both the hybridizing of a composite material with other materials (either other composites or base unreinforced materials) and composites using multiple reinforcements. Fu rther, t his c ategory c overs t he u se o f multiple m aterials ( at le ast o ne o f w hich i s a c omposite) i n structural app lications a nd h ighlights t he m ultiple u ses a nd advantages of composite materials. Hybrid composites can be divided into  ve major subcategories: (1) HCM, (2) selective reinforcements, (3) thermal management, (4) sm art s kins a nd s tructures, a nd (5) u ltralightweight materials. 5.1.1.4.1 HCM HCM are dened as a c omposite material system derived from the integrating of dissimilar materials, at least one of which is a basic composite material. A typical example of a HCM is a reinforced polymer composite combined with a conventional unreinforced homo genous me tal. The H CM blen ds t he de sirable properties of two or more types of materials into a single material system, which displays the benecial characteristics of the separate constituents. An existing example of a hybrid composite is

43722_C005.indd 5

7075-T5 Aramid/epoxy

NOM. 1.3 mm

7075-T5

7075-T6 sheet 0.03 mm

Aramid/epoxy

FIGURE 5.2

ARALL is an example of an HCM in production.

ARALL, which consists of high-strength aluminum alloy sheets interleafed with layers of aramid ber-reinforced adhesive as illustrated in Figure 5.1. The A RALL h ybrid c omposite (Figure 5.2) is baselined on several secondary structural composites of  xed-wings ubsonica ircraft. A second example of a hybrid composite i s a c arbon–carbon c omposite (CCC) w ith a si ngle side application o f t he re fractory me tal rh enium. This carbon–carbon–rhenium material is being developed for thermal management heat pipes on space-based radiator systems. Other examples include i nterpenetrating p olymer ne tworks (I PNs), w hich a re hybrid resin matrices consisting of thermoset and thermoplastic resin combinations. Still another HCM concept involves multiple rei nforcement t ypes w ithin a c ommon m atrix suc h a s chopped bers and continuous bers within a polymer matrix. Other HCMs i nclude new composite materials such as na nocomposites, functionally gradient materials (FGMs), hybrid materials (Hymats), IPNs, microinltrated macrolaminated composites (MIMLCs), and liquid crystal polymers (LCPs), which may force development of previously uneconomical process routes if they offer the path to a technical solution for advanced system capability. These materials present opportunities for reducing the number of stages in turbine engines and in so doing may be economically benecial even at a higher material cost because the smaller number of stages leads to greater economy of use. HCM technology is in its infancy in comparison to that of the other types of composite materials. Whereas ARALL and IPNs have been used for the past decade, the other types of HCM are truly emb ryonic. F rom ne arly a ll a spects o f re search, t hese HCM offer great potential for structural applications; however, their widespread use remains a decade or more in the future. 5.1.1.4.2 Selective Reinforcement Selective reinforcement is the category of hybrid composites that provide reinforcement to a structural component in a local area or areas by means of adding a composite material. An example of this is t he use of superplastic forming-diff usion bonding (SPF/ DB) as a means of integrating a titanium-reinforced MMC into a base titanium structure. As part of the initial design approach, consideration must b e g iven to t he to oling re quired f or p lacement o f t he r einforcing m aterial w ithin t he s tructure. This is accomplished by building into t he form tooling areas in which the M MC m aterial i s placed t hat p ermit SPF e xpansion of t he

10/3/2008 8:44:45 PM

5-6

Smart Materials

base t itanium to t he M MC d uring p rocessing a nd a llowing diff usion b onding t o o ccur, t hus a ccomplishing in tegral r einforcing of the nal structure. The selective reinforcement design approach allows the aerospace de signer to ut ilize t he mo re c ostly, h igher p erformance materials o nly w here t hey a re re quired a nd t he le ss e xpensive materials in areas where they can perform the job. This approach leads to an optimization of both cost and performance in the most ideal case. Similarly, problems associated with use of the reinforcing c omposite, suc h a s lo w me chanical jo int s trength, can be eliminated by not rei nforcing the area where the mechanical joining occurs. In reality, while this approach can be very efficient, it requires the introduction of multiple materials to solve the design problem and may in turn result in increased fabrication costs and risks. 5.1.1.4.3 Thermal Management In t he  eld o f t hermal m anagement, c omposite m aterials a nd HCM c an b e i nnovatively c onstructed to e ffectively l imit t he maximum tem perature o f s tructural h ardware a nd r apidly transfer heat from hot areas to cool areas. This capability arises from t he u nmatched t hermal c onductivity o f g raphite  ber, which is higher (in the  ber direction) than that of oxygen-free high-conductivity copper (OFHC). The graphite bers act as heat paths and, by suitable arrangement in the structure of interest, can remove heat by transmitting it along its length. In contrast, the matrix materials can act as thermal insulators so that thermal conduction through the thickness is lower by orders of magnitude than in-plane. This allows designers to develop a structure that is a thermal insulator in some directions but a thermal conductor in others. 5.1.1.4.4

Smart Skins and Smart Structures

Smart skins and smart structures are related in that each contains embedded, no nstructural elemen ts. A sm art s tructure c ontains sensors that monitor the health of the structure itself, such as ber optics to de termine temperature and structural deformations or cracks. A sm art s kin c ontains c ircuitry a nd ele ctronic c omponents that enable the skin to double as part of the electronic system of the parent vehicle, be it an aircraft or a missile. Smart s tructure te chnology, l ike t hat o f sm art s kins, i s s till evolving. The present technology consists of the incorporation of sensors into structural elements in the material processing stage to better control cure (or consolidation), and so on. 5.1.1.4.5 Ultralightweight Materials The c ategory of u ltralightweight materials i ncludes t he emerging family of liquid crystal ordered polymers, which by virtue of their mole cular s tructure, e xhibit e xtremely h igh s pecic strength and specic stiffness. These highly directional materials are similar to c omposite materials in t hat t he long, ordered molecular c hains w ithin t he p olymer ac t very much l ike rei nforcing f ibers i n a c omposite m aterial. E xamples o f t hese materials a re p oly-p-phenylene b enzobisthiazole (P BZT) a nd

43722_C005.indd 6

poly-p-phenylene b enzobisoxazole (P BO). A nother e xample i s gel-spun P E. These po lymers ex hibit ex tremely h igh speci c strength a nd speci c stiff ness a nd c an p otentially b e u sed a s reinforcing bers in composite laminates, as rope or cable, or as a self-reinforced thin  lm structure. These ordered polymers, in the form of thin lms, nd use in shear webs and skin applications f or a ircraft. These t hin  lms c an a lso b e p rocessed i nto honeycomb for l ightweight s tructural applications. It h as b een estimated that a s hear web made of PBZT would be one-eighth the weight of a n a luminum web a nd one-sixth t he weight of a graphite Ep web.

5.1.2

Future Directions

A t ype o f me tallic s tructure, w hich h as o nly re cently re ceived attention is nanostructure, in which the microstructural dimensions a re in t he na nometer range (10−8 to 1 0−9 m). The se nanostructures can be tailored to be either basically equiaxed or layered in nature, depending on production conditions. These nanostructures offer novel combinations of strength, ductility, and stiff ness not possible using conventional processing te chniques. I n t he l ayered c onguration w hen the th ickness o f t he l ayers i s b elow app roximately 1 00 nm, s o-called Koehler strengthening occurs in conjunction with an apparent modulus i ncrease, op ening up t he p ossibility o f v arying a nd controlling both strength and stiff ness by varying layer thickness. P ractical te chniques to p roduce na nostructures f rom metals include the electron beam coevaporation process developed b y t he Ro yal A cademy o f E ngineering ( RAE), w hich i s capable o f 18 k g/h de position r ates a nd me chanical a lloying. The r ealms o f poss ibilities h ere see m a lmost bo undless, w ith the caveat of also including ceramic bers in an overall tailored structure. Aluminum-based a lloys h ave b een u sed to d evelop h ybrid engineered materials such as MMC, mechanically alloyed dispersion strengthened a luminum (DISPAL), and AR ALL. The latter two materials offer high-temperature capability and outstanding fatigue performance, respectively. A number of production approaches have been studied, evaluated, and used to produce MMCs including powder metallurgy (P/M), casting, an in situ (XD) precipitation technique, and plasma spraying. The result is increased stiffness (by a f actor of 3) a nd enhanced temperature capability. In the casting approach, a pa rticulate is dispersed uniformly in the matrix to give attractive levels of strength and stiffness, the l atter o f w hich i s c lose to qu asi-isotropic g raphite-Ep. The X D te chnique (X D i s a t rademark o f M artin-Marietta Corporation now Lockheed-Martin) is a process in which an in situ p recipitation te chnique i s u sed f or t he f ormation o f t he reinforcement dir ectly in t he m atrix o f in terest. U sing X D technology, it is possible to simultaneously produce a number of different reinforcements in a given matrix (Al, Ti, Cu, and intermetallics), thus allowing design of the microstructure to meet a specic ne ed, e .g.,  ne pa rticles f or s trength a nd w hiskers f or

10/3/2008 8:44:46 PM

Composite Systems Modeling—Adaptive Structures: Modeling and Applications and Hybrid Composites

Ultimate strength as a function of temperature

Strength ~ KSI

200

Silicon carbide/ aluminum

Silicon carbide/titanium

150 Unreinforced titanium 100 50 Unreinforced aluminum 0

200

400

600 800 1000 1200 1400 1600 Temperature (⬚F)

FIGURE 5.3 Ultimate tensi le st rength a s a f unction of temperature for monolithic and SiCf reinforced aluminum and titanium.

creep-rate reduction. Thus, XD technology is a very useful technique f or ac hieving de signer o r eng ineered m icrostructures. Such  exibility at t he m icrostructure le vel a llows c oncurrent design, i .e., t he si multaneous de sign o f t he m aterial a nd component. Titanium alloys reinforced with continuous ceramic bers such as SiC enhanced the strength (Figure 5.3) and modulus (by a factor of 2). It i s p rojected t hat r einforced ti tanium a luminides m ay b e useful up to 1000°C. The estimated strength value for unidirectional r einforced ti tanium a luminides w ith S iC  bers (SCS-6/ Ti xAlyM) is nearly twice that of Inconel 718, over twice that of Ti-6Al-4 V at 540°C, and double that of unreinforced titanium at 760°C.

Acknowledgment I thank my parents, brother, and dearest friend Dr. Ramakrishna for their support and encouragement to write this chapter.

5.2 Design of an Active Composite Wing Spar with Bending– Torsion Coupling Carlos Silva, Bruno Rocha, and Afzal Suleman 5.2.1

Introduction

Adaptive materials are nowadays a lready reliable for industrial applications a nd c an b e e asily i ntroduced i n s ystems ne wly designed, w ith t heir i ntroduction a lready t aken i nto c onsideration, or in older s ystems as a retrot. They have the advantage o f t heir app lication b eing e asily ac hieved, i ndependently of t he ba se m aterial, si nce f or i nstance t hey c an b e b onded (surface glued or embedded) to metal, carbon, composite, etc.

43722_C005.indd 7

5-7

Adaptive s ystems c an i mprove a ircraft re liability, re ducing failures through vibration control or reduction. Vibration reduction me ans d amage, w ear, a nd f ailure p robability re duction, which leads to larger intervals between maintenance operations. This directly leads to a high cost reduction. Aeronautic structures are exible, specially the lifting surfaces. As the airplane is ying, s tructural de formation t akes p lace. Th is leads to a c hange i n t he aerodynamic loads produced by them due to t he shape modication. The aerodynamics forces increment o r de crement c onsequently re sults i n a ne w s tructural deformation. Th is i s a t wo-way i nteraction b etween t he structure a nd t he  uid [1]. The b ehavior c an b e s table, i f a n equilibrium c ondition i s at tained, o r u nstable, w here c atastrophic f ailure c an o ccur. N onetheless, t he si mple v ibration phenomena w ill c ause i ts d amage d uring t he l ife o f t he structures. Vibration control, which provides vibration reduction, utter and d ivergence at tenuation, a nd del ay, c an b e ac hieved u sing actuators to s end p ulses to t he s tructure to opp ose i ts s ensed vibration based on a f requency response methodology. In addition, a multicell cross-section for wing spars can be used. These spars when built from  ber composite laminate materials, with the preferable  ber orientation misaligned with span direction, possess a  exure–torsion coupling. Th is means that when these spars are subjected to  exure, they will present some amount of torsion. For certain ber angles and multicellular cross-section congurations, certain t ypes a nd i ntensities of t hat coupling c an be achieved, such that it can be utilized to counteract aeroelastic de formations a nd ph enomena, a s  utter a nd d ivergence attenuation. Recently, at the Portuguese Air Force Aeronautics Laboratory, Suleman et al. have reported several studies on aeroelastic control and utter alleviation in adaptive ight vehicles [2–4]. These studies led to two active wing concepts: the active spar and active skin c oncepts a nd b oth w ind-tunnel a nd  ight tests were conducted u sing a remote p iloted v ehicle ( RPV) [5]. A sig nicant increase in t he  utter s peed a nd de crease i n t he a mplitude o f vibrations w as ob served. These c onclusions w ere v alidated f or realistic ight conditions [6,7] with the successful results, a analogous proportional control system was applied in this study. 5.2.1.1 Bending–Torsion Coupling Spars Piezoelectric (PZT) actuators and sensors are nowadays reliable for i ndustrial app lications. They h ave t he adv antage o f t heir application b eing e asily ac hieved, i ndependently o f t he ba se material. F or i nstance, t hey c an b e b onded ( surface g lued o r embedded) to me tal, c arbon, c omposite, e tc. The coupling between the actuator and base structure can be easily enhanced in composite structures if one uses a m isaligned layer stacking. If a  at b eam is considered, the misalignment is the ber angle deviation from the length direction. The misalignment can vary from 0° (aligned) to 90° (perpendicular alignment). A structure built with this type of internal layout when loaded vertically will

10/3/2008 8:44:46 PM

5-8

Coupling

Smart Materials

0

90 Degrees

FIGURE 5.4

Actuators and Sensors

The ideal actuator would have a high-energy (mechanical stress/ strain) ac tuation c apability w ith lo w energ y i nput ( voltage/ current), wide frequency range, linear and unidirectional behavior, and c onvenient s hape. I deal s ensors w ould h ave a no ise-free response, h igh p recision, a nd  delity ( low m echanical e nergy stimulation s ensitive), l inear b ehavior, w ide f requency s ensing range, and convenient shape. Both would be easily integrated or bonded into the structures. Due to t his t ype o f s tructure d imensions, lo w-frequency requirements, PZ T ac tuators a nd s ensors w ere s elected. F rom the v arious t ypes o f ac tuators a nd s ensors a vailable i n t he market, Q P40W f rom Q uickPack A CX w as s elected a s t he actuator a nd P ZT m ade of m aterial BM 500 f rom S ensor Technology Ltd. were chosen as the sensor.

5.2.2

Multicell Cross-Section Spar Design

One of t he challenges was to de sign, opt imize, manufacture, and te st a c ontinuous  lament m ulticell c ross-section c omposite s par w ith BTC a nd w ith emb edded PZ T ac tuators (Figure 5.5). The resulting multicell spar will have its upper wall composite bers o riented i n t he s ame d irection a s t he lo wer w all  bers (Figure 5. 5). W hen sub jected to b ending lo ads, t his m ulticell spar will present bending and also torsion deformations (BTC). Orienting the upper and lower wall bers from trailing edge to leading edge, from root to wing tip, when the wing spar bends upwards, it twists in a manner that the angle of attack is decreased (leading edge down). With this kind of orientation, a desirable behavior is at tained. Th is counteracts t he dy namical behavior of utter.

43722_C005.indd 8

Semispars and ber layout.

BTC evolution.

deform vertically and twist. The same effect can be attained if a torsion moment is applied. An inherent consequence is that the vertical d isplacement a nd to rsion a re no lo nger l inear w ith respect to the specic load applied. Early studies i n t his  eld were c arried out by G arnkle and Pastore [ 8]. They a nalyzed t he pa ssive a nd a ugmented ac tive behavior o f st ructures ex hibiting be nding–torsion co upling (BTC) as shown in Figure 5.4. 5.2.1.2

FIGURE 5.5

This i s a pa ssive c haracteristic o f t hese t ypes o f s pars. Using an ac tive s ystem, t his cha racteristic c an be augmented by app lying ac tuators o n t he upp er a nd lo wer m ulticell s par walls. In order to determine the misalignment, a global optimization approach through metamodeling was programmed. For the type of spar presented, considering its constraints and dened objective f unction, t he de sign v ariables s et w as c alculated [ 9]. The optimized solution dened the nal design to be manufactured, as follows: span, 1 m; height, 35 mm; wall thickness, 1 mm; and ber orientation, 10°. 5.2.2.1

Actuators and Sensors Positioning

In order to determine the most advantageous positions to place both actuators and sensors, computational analyses were carried out. With respect to vibration reduction, the location of sensors and actuators was determined by nite element method (FEM) analysis of the vibration modes. This methodology revealed the location of points with larger strain values (near the wing root). The advantage provided by this approach is that a small and light actuator manages to change dramatically the vibration parameters of the wing because it is placed near the points of higher strain. Also, the higher strains permit better sensing denition by t he s ensors due to i ts l arger a mplitudes. This positioning is also adv antageous f or t he f requency-based si mple h armonic motion (SHM). Despite t he fact t hat t he main present work i s not o n SH M, related c onstraints w ere t aken i nto ac count. Wi th re spect to wave propagation methods, it is useful to position the sensors as far away as possible from the actuators. Consequently, the spar was divided into two equal zones, from now on called cells, each one dened by a pair of actuators and one of sensors. The cells have the actuators on its root and sensors on its end on the spanwise direction. The upcoming disadvantage for the vibration reduction is only related to t he sensors amplitude output decrease, which can be easily overcome. 5.2.2.2

Active System

Figure 5. 6 s hows t he ac tive s ystem s cheme. St arting f rom t he PZT sensor, it is assumed that the PZT signal is proportional to

10/3/2008 8:44:46 PM

5-9

Composite Systems Modeling—Adaptive Structures: Modeling and Applications and Hybrid Composites

Piezoelectric actuator

Amplifier

Active wing

Signal conditioning D/A converter

Bending gain

Piezoelectric sensor

control s cheme, t he ac tuators w ill a lways counteract t he spar’s vertical movement.

Signal conditioning

5.2.3

Lowpass filter

Several experiments were conducted during the work progress. So, s par, w ing, a nd R PV  nal a ssembly w ere a ll te sted. The experiments comprehended static and dy namic on passive and active solutions. Results obtained include sensor output analysis, fast F ourier t ransform ( FFT), a nd d amping c alculations ( see Figure 5.8).

A/D converter

Controller

FIGURE 5.6

Proportional control scheme.

5.2.3.1 the spar’s vertical displacement. This signal has to be conditioned because the control, carried out by a MicroAutoBox DSpace control module, has to b e matched to t he I/O requirements. Inside the c ontrol mo dule, a M atlab Si mulink mo del, p reviously uploaded a nd r unning a s a  ash app lication o n t he c ontroller memory, i mplements t he d irect p roportional c ontrol l aw ( see Figure 5.7). A lo w pa ss  lter is included to retain only the frequencies below 50 Hz. The main concern is the low frequency present in the  rst b ending v ibration mo de a lso a voids i nterferences coming from the power current. Afterwards, a gain is applied to the i nput sig nal a nd i t i s  nally sent as an output back to the signal conditioning board. The re ferred ga in h as to b e de termined t hrough e xperiments. O nce a gain t he I /O h as to b e matched to the one required by the amplier, which by its turn, sends t he  nal sig nal to t he PZ T ac tuators. U sing t his si mple RTI Data

DS1401 Flight recorder

Results

Spar

The spar with the actuators and sensors was clamped on its root and an impact was given to t he tip. Figure 5.8 shows the sensor output time evolution. Using the data attained, the rst natural frequency a nd d amping c oefficient w ere de termined f or b oth situations. From the results shown, it is easily seen that the improvement is signicant. Up to 8 0% of d amping i ncrease c an b e achieved and up to 12%, rst natural frequency augment can be attained (see Table 5.2). 5.2.3.2 Wing Design and Fabrication After nishing the spar’s experiments, a wing was built with it to be tunnel tested. This wing was intended to be used in the future as a s pecimen c apable to a llow t he i mplementation o f SH M methods. Thus, t he mo del w as de signed t aking i nto c onsideration t wo SH M me thods, na mely v ibration ba sed a nd w ave

79944

1

.001 0.3999 79

SensorRec Filter on

12:34

(single)

DS1401 flight recorder

Digital clock

Data type conversion3

TimeRec

Input w/filt (single) Data type conversion1

2.18 Coarse gain

ADC

Sensor output1 ADC_type1_M1_con1

0.5

+ − + Clear input

X

+ +

0

Product

1

DAC

Active_gain

Filter off

DAC_type1_M1_C1 Final output

V offset

−1

+ +

−0.0068 Test-offset

Offset_sensor

Pulse generator

Gain1

Test-output + +

0

DAC

0 Test-gain

X

DAC_type1_M1_C2

Active switch + + Pulse generator1 1 Pulse offset

(single) 0 Active off

Multiport switch

Data type conversion2

1

DS1401 flight recorder

Active on

SwitchRec

FIGURE 5.7 SIMULINK control law model.

43722_C005.indd 9

10/3/2008 8:44:47 PM

5-10

Smart Materials

1 0.8

Dimensionalized sensor signal Active Passive

0.6 0.4 0.2 0 −0.2 0

0.05

0.1

−0.4

0.15

0.2 Time (s)

−0.6 −0.8 −1

FIGURE 5.8

Passive and active responses.

TABLE 5.2 Passive and Active Damping Coefficients and First Natural Frequencies Coefficient Passive spar Active spar Increase

Damping Frequency (Hz) 0.054 0.1 80%

First Natural Frequency 28.92 32.46 12%

propagation. One main feature of the wave propagation method application is that an attached wing spar rib or any kind of stiffener limits wave propagation. Thus, i n o rder to h ave s pace to place the actuators and sensors and to a llow the wave to propagate, t he gener al s tructure o f t his sm all w ing mo del f or w ind tunnel testing was divided only into two cells with a continuous wing s par. H owever, to a ssure t he a erodynamic s hape o f t he wing no nstructural r ibs ( not at tached to t he m ain s par) w ere positioned spanwise, 8 cm apart from each other. Beyond these two cells, there was a structural rib at the wing root and another one 4 cm apa rt s panwise, c onnected to e ach ot her by foam (in addition to the wing spar). This w as i ntended to f orm a rei nforced section to guarantee the wing–fuselage connection. Also, the wing skin was not structural and could be made of a t ransparent  lm, w hich a llowed s eeing t he i nternal w ing s tructure, actuators/sensors positioning, systems, etc. Figure 5.9 shows the wind tunnel testing apparatus. For e ach te st, a s ensor sig nal a mplitude o utput v ersus t ime sample was retrieved and an FFT analysis was done. For instance, Figure 5.10 re presents t he F FT re sults ob tained f or t he 2 0 m/s velocity for the passive wing and active wing with a gain of 50% max. A c lear de crease i n t he v ibration a mplitude ne ar f or t he rst natural frequency can be observed. Maximum signal output amplitudes were retrieved and Figure 5.11 shows the results. As shown, the amplitude decreased from passive to ac tive te sts. G enerally, t his a mplitude ten ds to b e smaller when a higher gain is applied. Figure 5.12 shows the percentage of maximum sensor signal output amplitude decrease: a

43722_C005.indd 10

maximum 8.83% decrease is reached for 30 m/s of velocity and a gain of 50% max. As regards the damping analysis, studies were concentrated around t he  rst nat ural f requency. Ha ving t he s ensor sig nal output, FFT a mplitude versus frequency graphs, damping was calculated, re trieving t he f requency v alue for t he p eak a mplitude a nd f requency v alues (around t he previous) obtained for maximum amplitude divided by the square root of 2. With these

FIGURE 5.9

Wind tunnel test setup.

10/3/2008 8:44:47 PM

Composite Systems Modeling—Adaptive Structures: Modeling and Applications and Hybrid Composites

FFT 20 m/s

0.012 Pass Act G5

Amplitude

0.01 0.008 0.006 0.004 0.002 0 0

20

40

60

80

100

Hz

FIGURE 5.10

The 20 m/s passive and active wing FFT analysis.

three values, it was possible to determine a d amping coefficient proportional v alue (not t he d amping c oefficient itself) by sub tracting the highest and lowest frequency values divided by the value of the central frequency. With this algorithm, all the damping coefficients were calculated and plotted on the same graph (Figure 5.13) along with the

5-11

respective qu adratic p olynomial t rendlines. These l ast o nes, allowed us to predict utter speed (that corresponds to the velocity f or t he z ero d amping c oefficient, i .e., w hen t he t rendlines would ideally cross the x/velocity axis). It is noted that the utter s peed i ncreases f rom pa ssive to active tests, and it reaches its highest value for the 50% max gain. Also, d amp i mprovements a re ob served f rom pa ssive to ac tive tests, and they are higher also for the 50% max gain. Overactuation and delays inuence the active gain results for gains higher than 50%, re sulting i n p oorer outcomes. Figure 5.14 s hows t he p ercentage damping improvements: a maximum 93% improvement is reached for 45 m/s of velocity and using 50% maximum gain. The maximum damp i mprovement, now obtained, is higher than t he o ne ob tained f or t he s par w hen i ts me chanical te sts were performed (80% max improvement). This increase can be explained by the BTC, which had no inuence on the spar behavior when its mechanical tests were done, but which is essential on wind tunnel tests. On these nal tests, this coupling affected the aeroelastic behavior of the wing. When the wing bends and consequently rotates along the spanwise axis due to the coupling, the aerodynamics of the wing will be inuenced, i.e., aerodynamic

Max amplitude 0.7

Signal amplitude

0.6 0.5

Pass

0.4

G25

0.3

G50 G100

0.2 0.1 0 20

30

40

45

m/s

FIGURE 5.11

Wing maximum amplitude.

%

Max amplitude decrease

10.00 9.00 8.00 7.00 6.00 5.00 4.00 3.00 2.00 1.00 0.00

G25 G50 G100

20

30

40

45

m/s

FIGURE 5.12

43722_C005.indd 11

Wing maximum amplitude decrease.

10/3/2008 8:44:48 PM

5-12

Smart Materials Damping

0.012 G25 G50 G100 Pass Poly. (G50) Poly. (G100) Poly. (G25) Poly. (Pass)

0.01 Damp

0.008 0.006 0.004 0.002 0 20

30

40

50

60

70

m/s

FIGURE 5.13

Damping for 25%, 50%, and 100% of maximum gain.

%

Damp improvement

100 90 80 70 60 50 40 30 20 10 0

G25 G50 G100

20

30

40

45

m/s

FIGURE 5.14 Damping improvement for 25%, 50%, and 100% of maximum gain.

loads applied and their distribution is altered, affecting the wing structure, cha nging co nsequently t he w ing st ructure deformations. 5.2.3.2.1 0°/10° Wing Performance Comparison In order to truly assess the ber misalignment behavior, a similar 0° ber-oriented spar was manufactured and tested. Figure 5.15 presents t he a mplitude v ariation f or b oth s pars at d ifferent

velocities. It can be inferred that the misalignment on spars (to a certain extent) by itself can lead to an improvement in the wing’s behavior. Afterwards, it can be clearly augmented by implementing active materials. 5.2.3.3

RPV: Flight Tests

After the successful wind-tunnel testing, the wing was prepared to be installed on an RPV for ight testing. It does not compromise

Max amplitude

Max amplitude decrease

35

1

30

0.8

25

10º 0º

0.6

%

Signal amplitude

1.2

20 15

0.4

10

0.2

5 0

0 20

35

40

45

m/s

FIGURE 5.15

43722_C005.indd 12

20

35

40

45

m/s

0° and 10° maximum amplitude and its comparison.

10/3/2008 8:44:49 PM

Composite Systems Modeling—Adaptive Structures: Modeling and Applications and Hybrid Composites

FIGURE 5.16 Flight testing modied wing.

the desired behavior but induces some mass increase. The wing must also be equipped with an aileron and respective servo action mechanism (Figure 5.16). Its root had to be reinforced in order to receive the clamping pin provided by the fuselage. To c arry o ut t he  ight te sts, t he a mplier also had to be changed. The one used on previous tests was adequate and was replaced by a sm all bat tery-powered one. Figures 5.17 a nd 5.18 show a ll t he ac tive e quipment ge ar o n b oard a nd i n-ground. During t he g round e quipment te sting, a n i nterference w as detected between the engine and the sensor system. Due to this condition, it was opted to do gliding ight testing. The procedure consisted in ying the RPV as high as possible, which w as l imited b y t he p ilot’s ob servation c apability. Afterwards the engine was cut off and a glide descent aimed to a speed of 20 m/s was performed. The airspeed was monitored in the P C u sing t he J et-Tronic s oft ware, w hich re ceived t he p ilot telemetry sig nal f rom onboard a nd t he te sting t ime w as b eing correlated with speed, and both values registered. This operation revealed that it was very difficult to conduct, even with two assistants, because of the difficulty in keeping constant conditions. Two c onsecutive  ights w ere c arried o ut. F igure 5.19 s hows the  ight data recorded, including the altitude gain. Figure 5.20

5-13

is focused on the glide part of both ights. The ramp functions represent the ight time. Flying the RPV must be done not very far from the pilot, and the a ircraft p erformed a s eries o f t urns t hat i nuenced the results. In the turns, effective wing lift increased and the vibrations w ere h igher. This m ay e xplain w hy s ometimes w hen t he control was ON (and was supposed to decrease vibrations), some very large amplitude peaks appeared. Due to this, no maximum or minimum comparison was done because t hey do not re present the general wing’s behavior. If one considers that a w indow represents a s tate of active or inactive, ight 1 had six windows and  ight 2 had ve windows. Selecting joined windows believed to represent the most equivalent  ight c onditions, w e ac hieved t he f ollowing F FT a nalysis, shown in Figures 5.21 and 5.22. A c lear i mprovement c an b e ob served. A nother i nteresting fact is that the improvement seems to be higher in the last part of both  ights, the  nal straight approach. This can be justied by the fact that the wind conditions were more severe near the ground, thus implying more vibrations on the wing. Adding to this, since the airplane was aligned with the runway less pilots, inputs were necessary to guide the RPV to the landing.

5.2.4

Concluding Remarks

The proposed solutions have resulted in a successful implementation of a ne w spar concept for a c onventional w ing. The proposed s olution c onsists o f a m ulticell c ross-section s par w ith misaligned  bers, w hich re sulted i n a n i mproved a eroelastic performance a nd supp ression o f v ibrations a nd c onsequent postponement of utter. The rst step in the process was to design the spar itself. The conceptual design of the spar indicated that it should have more than one section, and the ber orientation should be aligned at

FIGURE 5.17 RPV internal equipment placing.

43722_C005.indd 13

10/3/2008 8:44:50 PM

5-14

Smart Materials

FIGURE 5.18

Flight test setup on runway. 1.4

Sensor output (V)

1.2 1 0.8 0.6 0.4 0.2 0

3 Time

4

1

0.9

0.9

0.8

0.8

0.7

0.7

0.6 0.5 0.4 0.3

0.5 0.4 0.3 0.2

0.1

0.1

FIGURE 5.20

6 ⫻105

0.6

0.2

0

5

Flight test sensor pattern.

1

0

43722_C005.indd 14

2

Sensor output (V)

Sensor output (V)

FIGURE 5.19

1

0

1

2

3

4 Time

5

6

7

8 ⫻104

0

0

1

2

3

4 Time

5

6

7 ⫻104

First and second gliding  ights sensor output.

10/3/2008 8:44:51 PM

5-15

Composite Systems Modeling—Adaptive Structures: Modeling and Applications and Hybrid Composites

FFT 0.006 0.005

Passive 0.005

Active

Active

0.004 Amplitude

0.004 Amplitude

FFT

0.006

Passive

0.003 0.002

0.003 0.002 0.001

0.001 0

0

10

10

FIGURE 5.21

20

30 Hz

40

20

30

40

Hz

Second/third and  ft h/sixth windows passive/active FFT analysis.

FFT

FFT

0.006

0.006 Passive

Passive 0.005

0.005

Active

Amplitude

Amplitude

0.003

0.003

0.002

0.002

0.001

0.001 0

0 10

20

30 Hz

40

10

50

20

30 Hz

40

50

Second/third and fourth/ ft h windows passive/active FFT analysis.

an angle with respect to the spanwise direction. On a planview, these bers must start in the trailing edge and end in the leading edge such that the leading edge decreases its displacement when the wing experiences a lift load. For a simple constant rectangular cross-section spar with an overall constant thickness and manufactured with one material, this is the only way of achieving the desired BTC. Next, t he issue was to de cide t he  nal d imensions a nd t he right misalignment of the bers. On one hand, a small misalignment c ould re sult i n a p oor BTC b ut go od v ertical s tiffness and h igh nat ural b ending mo des f requencies. O n t he ot her hand, a h igh m isalignment w ould p resent a b etter c oupling but p oor s tiff ness a nd lo w nat ural b ending f requencies. T o determine t he b est s olution, a n opt imization p roblem w as de ned and solved.

43722_C005.indd 15

Active

0.004

0.004

FIGURE 5.22

50

−0.001

50

The u se of ac tive m aterials w as te sted r ight w ay at t his very early s tage of work. S ensors a nd ac tuators p ositions were c arefully studied. Having in mind future objectives such as structural health mo nitoring, ot her t han v ibration re duction a nd  utter suppression, two pairs of actuators were placed, one at t he root and t he ot her at t he m idspan. I n e ach c ase, o ne ac tuator w as glued on the upper and the other on the lower part of the spar. Two pa irs o f s ensors w ere a lso g lued to t he s par, b ut f or t his work, only one was necessary. Its position is in the midspan and both upper and lower positions were suitable for its task. Experimental ac tive te sts w ere t hen r un a nd c ompared to passive ones. The active methodology consisted of reading t he sensors strain and feed them back to the actuators, via a proportional c ontrol s ystem, to opp ose t he mo vement. Sub stantial damping a nd to rsional nat ural f requency i ncreases w ere

10/3/2008 8:44:53 PM

5-16

attained. This presented a go od mot ivation for continuing t he path followed so far. At t his stage, t he next m ilestone was identied. A w ing was built w ith o ne o f t he m anufactured s pars. Si nce t his s par w as intended to b e u sed f or f uture SH M me thods te sting, a w ing composed of two main cells was designed. One of these methods was ba sed o n w ave p ropagation. St arting f rom t he ro ot, t his wing had a reinforcement box for clamping in wind tunnel and the RPV fuselage. Then it had two cells. Each one was dened by two s tructural r ibs (attached to t he s par). B etween t hese t wo, there were “form” ribs, which were not attached and were placed to g uarantee t he a irfoil s hape a nd t ransmit t he a erodynamic loads to the edges. These were then transmitted to the spar by the attached structural ribs. This way, mechanical waves generated by the actuators could progress through the spar in each cell. Next, w ind t unnel te sts were c arried out. These were r un at several stabilized velocities, starting from 20 m/s to a maximum near the utter speed initiation behavior, approximately around 50 m/s. A similar wing was built at the same time, using a 0° spar to test and compare with the native behavior. The results showed that the native performance is clearly improved when applying a misalignment to the  bers. An approximately 30% reduction in the maximum vertical displacement amplitude was found for a test run at 45 m/s. It also presented a stable vertical displacement behavior a s t he v elocity w as i ncreased. The 0° spar tended to increase this parameter as the velocity rose. When applying active control using the piezo actuators, this aeroelastic performance of the wing was enhanced. A structural damping increase around 90% was attained for 45 m/s. The utter was predicted to be delayed by 8 m/s, which corresponds to a 16% increase. The best results were found when using a 50% gain increase in the proportional feedback control system. At t his s tage, t he s econd m ilestone w as c ompleted a nd t he main objective of the work satised, since the wind tunnel tests closely re produced t he s tructure’s re sponse to t he re al a irow inputs. Nevertheless,  ight testing can push t he limits and test other aspects and act as a determinant to assess the systems viability. So, the  nal milestone was to adapt this wing to the RPV and test it in ight. To do so, the control equipment had to be put onboard. The energ y s ource w as c hanged to bat teries a nd a n RPV, w hich i ndicated a irspeed me asuring e quipment, w as installed. The RPV was manually piloted from the ground. The control system was preprogrammed with the control law and set to save all the  ight data for future postprocessing. During the  ight, t he ac tive s ystem w as a utomatically s et o n a nd off in

43722_C005.indd 16

Smart Materials

intervals of 12 s. Pre ight tests revealed engine interferences of an u ndetermined source. Gliding  ights opt ion was t aken a nd spectral frequency analyses were carried out. The results showed a sig nicant v ibration re duction de spite t he re duced a vailable amplication provided by the internal amplier. This  nal work stage closed the loop of the spar development program and on the passive and active system implementation. As a  nal c onclusion, it c an b e s aid t hat b oth t he pa ssive a nd active multicell cross-section spars proved to b e a v iable solution in a ctive a eroelastic c ontrol s olutions in a daptive  ight vehicles.

References 1. 2.

3.

4.

5. 6.

7. 8. 9.

Bisplinghoff, R .L., Ashley, H., a nd H alfman, R .L., Aeroelasticity, Dover, New York, 1996. Suleman, A., Costa, A.P., and Moniz, P., Experimental utter and buffeting suppression using piezoelectric actuators and sensors, Proceedings of the Smart Structures and Materials— 3674, 1999. Amprikidis, M. and Cooper, J., Adaptive internal structures for active aeroelastic control, Proceedings of the International Forum o n Aeroelasticity a nd S tructural D ynamics—IFASD 2003, 2003. Rocha, J., Moniz, P., C osta, A., and Suleman, A., On ac tive aeroelastic control of an adaptive wing using piezoelectric actuators, Journal of AircraĀ, 42(1), pp 272-278, January— February, 2005. Costa, A., N ovel C oncepts f or P iezoelectric Actuated Adaptive A eroelastic Air craft S tructures, P hD thesis, Instituto Superior Técnico, Portugal, 2002. Rocha, J., Suleman, A., Costa, A., Moniz, P., and Santos, D., Research and development of an active aeroelastic adaptive ight demo nstrator, CEAS/AI AA/NVvL IF ASD 2003, Netherlands, 2003. Rocha, J ., An E xperimental S tudy a nd Flig ht T esting o f Active Aeroelastic Aircraft Wing S tructures, M aSC thesis, University of Victoria, Canada, 2005. Garnkle, M. a nd P astore, C., I ntrinsically sma rt co upled box b eams, Retrie ved February 5, 2007, f rom http://www. pages.drexel.edu/∼garnkm/Spar.html. Silva, C., Rocha, B., and Suleman, A., A metamodeling optimization approach to a wing spar design, 49th AIAA/ASME/ ASCE/AHS/ASC S tructures, S tructural D ynamics a nd Materials, Schaumburg IL, USA, 2008.

10/3/2008 8:44:53 PM

6 Ferromagnetic Shape Memory Alloy Actuators Yuanchang Liang University of Washington

Minoru Taya University of Washington

6.1 I ntroduction ...............................................................................................................................6-1 Review of Shape Memory Alloy • Ferromagnetic Shape Memory Alloy • Driving Mechanisms for FSMA-Based Actuators

6.2 FS MA-Based Actuator...............................................................................................................6-6 References ...........................................................................................................................................6-10

6.1 Introduction 6.1.1 Review of Shape Memory Alloy Shape memory alloys (SMAs), such as Ni–Ti (Nitinol) alloy, have a smaller hysteresis of (austenite Û martensite) phase transformation than that in steel [1]. Ā ere a re n umerous pap ers t hat have e xamined S MAs. Ā is is bec ause S MAs a re sci entically attractive a nd can be used as sensors a nd ac tuators. Ā e phase transformation in SMAs is reversible and usually can be characterized by t he latent heat a nd t he change of lattice pa rameters. To obtain reversible deformation, a heating or cooling device is usually incorporated. SMAs also have been used extensively as key actuator materials with applications in aerospace structures (such a s chevron) a nd me dical i mplant de vices (such a s s tent), where tem perature o r s tress i s a d riving f orce to ob tain l arge strain recovery (shape m emory eff ect [SME]) o r l arge “elastic” strain (superelastic behavior). Ā e stress–temperature phase diagram of SMAs, shown schematically in Figure 6.1, is often u sed to c haracterize t he ph ase transformation b etween s tress a nd tem perature. U sually, i t i s constructed by t he stress–strain curves tested at d ifferent temperatures. Ā e intersections with the temperature axis (stress = 0), TMs, TMf, TAs, a nd TAf a re u sually t aken f rom t he re sults b y differential scanning calorimetry (DSC), which are martensite at the s tart, m artensite at t he  nish, a ustenite at t he s tart, a nd austenite at the nish temperatures, respectively. Ā e se transformation temperatures are functions of stress where the larger the stress, t he h igher t he t ransformation tem perature. F igure 6 .2a and b s chematically s how lo ading a nd u nloading c urves of a n SMA tested at temperatures (T) higher than TAf and below TMs, respectively. As the SMA is loaded at T > TAf from the stress-free stage, O, t he stress–strain curve initially follows a s traight line

(OA) w ith t he s lope o f EA ( Young’s mo dulus o f t he a ustenite phase). Ā is corresponds to the elastic behavior of the SMA with the 100% austenite phase until it reaches the stress (sMs) of the onset s tress-induced m artensitic ( SIM) ph ase t ransformation (austenite → martensite), point A. Upon further loading, it follows a gradual straight line until point B where the curve becomes steep again in a slope of EM (Young’s modulus of the martensite phase). B etween p oints B a nd C , t he S MA i s u nder a n el astic deformation with 100% martensite phase until it is yielded at the yield stress (sy), point C. When the stress is over sy, t he 100% martensite phase of the SMA undergoes plastic deformation by a dislocation gliding mechanism. If under a constant temperature, T > T Af, the loading curve continues up to s ome point below sy followed b y unl oading, th e s tress–strain p ath will f ollow th e elastic u nloading a long t he s traight l ine ( CD) u ntil t he s tress reaches t he o nset o f re verse t ransformation ( martensite → austenite) at point D. Further unloading will render the reverse transformation b y f ollowing a g radual s traight l ine u ntil t he curve h its t he el astic lo ading l ine at p oint E . Ā e unloading beyond point E follows the elastic deformation path (EO) to the origin (O). Ā is loop (OABDEO) in Figure 6.2a is called “superelastic” (SE) behavior of the SMA where the strain at p oint B i s usually in the order of 5% for the polycrystal SMA. On the other hand, i f t he lo ading–unloading process i s c onducted at a c onstant temperature, T < T MS (Figure 6.2b), t he curve follows t he path OAB and leaves a residual strain (eres). Ā is residual strain is re coverable when t he temperature r ises over TAf because t he martensite to austenite phase transformation occurs. Ā e curve will follow the strain axis (BO) and return to the origin (O), i.e., zero strain. Ā is loop (OABO) is the so-called SME. eres is usually on the order of 5% for the polycrystal SMA. Unlike the path OA of SE in Figure 6.2a, the path OA in Figure 6.2b is not linear 6-1

43722_C006.indd 1

10/3/2008 8:45:24 PM

6-2 Stress (s)

Smart Materials

6.1.2 Ferromagnetic Shape Memory Alloy Martensite Austenite

TMf

TMs

TAs

TAf

Temperature (T )

FIGURE 6.1 Stress–temperature diagram of an SMA.

s C

sy

EM B A s Ms D E O

EA e

(a) s

A

B

O (b)

As de scribed b efore, t he ph ase t ransformation o f S MAs c an b e controlled by stress and temperature. Their relationship can be described by a stress–temperature phase diagram as shown in Figure 6.1. Recently, attention has been paid to some SMAs accompanying a change in ferromagnetism at phase transformation and having low hysteresis. Ā is is because the transformation characteristics suc h a s t ransition p oint ( martensite s tart tem perature, TMs) and macroscopically observed strain, caused by the transformation, are possibly controlled by an applied magnetic eld (H). Ā e response of transformation is fast in this case because the characteristic t ime i s governed by t he formation a nd g rowth of martensite, w hich i s i nduced b y a n app lied m agnetic  eld. It i s fast, thus, it is plausible to p roduce an actuator having reversible straining w ith qu ick re sponse to a sig nal i mposed o r de tected. Such an alloy with both ferromagnetic property and SMA behaviors is called ferromagnetic shape memory alloy (FSMA) and it is considered a s a s trong c andidate f or f ast re sponsive ac tuator material. Ā e rel ationship o f ph ase t ransformation, s tress ( s), temperature ( T), a nd m agnetic  eld (H) c an b e p resented a s a three-dimensional ph ase t ransformation d iagram a s s hown schematically in Figure 6.3. Currently, o nly a l imited n umbers o f F SMAs h ave b een found as listed in Table 6.1. Many researchers have extensively studied Ni MnGa, Fe–Ni–Co–Ti, Fe–Pt, a nd Fe–Pd FSMAs to examine SME and superelasticity. Ā ese alloys have also been considered as a possible candidate actuator materials. Among these a lloys, Ni MnGa [ 4,5] a nd F e–Ni–Co–Ti [ 20,21] a lloys have been known to exhibit good SMEs. However, the martensitic transformation start temperature (TMs) of the latter alloy is below − 150°C. I t i s to o lo w to b e p ractically u sed. Fu rther, Kakeshita e t a l. e ven re ported m agnetoelastic m artensitic transformation i n F e–Ni–Co–Ti b y u sing a s trongly app lied magnetic  eld up to 2 .3 × 1 07 A/ m. Ā ey also reported that martensitic transformation is not i nduced by an applied magnetic  eld u ntil 4 .8 × 1 06 A /m i s re ached [ 25]. Ā ese results show that a very strong applied magnetic eld is n eeded t o directly induce martensitic transformation. Ā is is not suitable

e res

e

FIGURE 6.2 Stress–strain curve of an SMA with (a) SE loop at T > TAf and (b) SME loop at T < TMs.

Stress (s )

Martensite Austenite

Temperature (T )

because it is not an elastic deformation. Ā e SMA is in the fully martensite phase at t he temperature below TMs and the loading path OA is involved in the martensite variant rearrangement. However, if the SMA is loaded beyond point A in Figure 6.2b, it is under the elastic deformation with a slope of EM similar to BC in Figure 6.2a.

43722_C006.indd 2

Magnetic field (H )

FIGURE 6.3 Schematics of t hree-dimensional phase transformation diagram of FSMAs.

10/3/2008 8:45:25 PM

6-3

Ferromagnetic Shape Memory Alloy Actuators TABLE 6.1 List of Current FSMAs Alloy

TMs

References

Ni2MnGa Fe–Ni–Co–Ti Fe-23–25Pt Fe-28–31Pd Fe–Pd–Ni NiAlCo

−150°C to 150°C <−150°C <−48°C −197°C to 80°C
[2–19] [20–29] [30–38] [39–53] [54,55] [56,57]

Type Heusler Iron-based alloys

Cobalt-based alloy

for applications as an actuator material where a compact electromagnet is needed. However, t he variant rearrangement mechanism u nder constant magnetic eld maybe still useful for some applications. Ā e change of the surface morphology on a si ngle crystal specimen can b e ob served d uring t he p rocess o f c ooling. A s s hown i n Figure 6.4a of a single crystal FePd whose temperature is just below TMf (= 0°C), the boundary and the contrast of the two variants are not distinct. When this specimen is further cooled at −20°C, its surface morphology exhibits a distinct boundary with a clear contrast of black a nd white colors (Figure 6.4b). Ā is is

−2⬚C

Observation direction

C

) FC

01

(1

P: No contrast j

100 µm

(a)

a CV1 a

−20⬚C

c

CV2

c Q: Less distinct contrast corresponding to photo (a) j

because the lattice parameter change (e.g., c/a ratio, Figure 6.4c [43,44]) of the FePd martensite gradually decreases with decreasing tem perature. Ā e g radual c hange o f t he l attice pa rameter results i n a n i ncrease of t he a ngle j formed by two martensite variants (CV1 and CV2) (Figure 6.4d). As j i ncreases d uring cooling, t he c ontrast o f t he b oundary b etween C V1 a nd C V2 (Figure 6.4a and b) also is enhanced. Ā e change of the volume fraction between variants (variant rearrangement) such as CV1 and CV2 in Figure 6.4d will result in additional straining of the FePd specimen. Ā is can be done by applying a magnetic eld since t he F ePd i s a f erromagnetic m aterial. Suc h a s training, which is due to t he magnetic  eld–induced martensite variants rearrangement, c an b e de tected. W hen t he FePd si ngle c rystal specimen is fully covered by the pair of CVs at −20°C, a magnetic eld i s applied a long t he leng th of t he s pecimen. A s s hown i n Figure 6.5, the strain of the magnetic eld–induced martensite rearrangement is a function of the magnetic ux density, where the ne gative m agnetic  ux den sity me ans t hat t he d irection of the eld is opposite. A strain of 0.4% is obtained and it is reversible a nd repeatable w ith a ny bias stress, a lthough some studies have reported [20] that the application of a bias stress is required simultaneously with the applied magnetic eld in order to induce such a reversible strain in FSMAs as shown in Figure 6.5. A single crystalline FePd specimen is subjected to a compression test at −20°C a nd t he a mount of strain accompanied w ith the stress-induced variant rearrangement is measured. Su rface morphology change of the specimen is also observed during the tests. Figure 6 .6 s hows t he su rface morphology c hange during the  rst lo ading a nd t he s tress–strain c urves o f t he  rst two cycles of loading a nd u nloading te sts. D uring t he  rst loading process ( Figure 6 .6a t hrough e ), t he vol ume f raction o f C V1 increases at the expense of CV2. When the stress reaches 10 MPa, the specimen is fully covered by a single variant of CV1 and the process of the variant rearrangement is completed (Figure 6.6e), while the strain reaches to 5.2%. During the rst unloading process, t he volume f raction of C V1 de creases a nd t he u nloading curve (white circles in Figure 6.6f) is slightly lower than that the loading curve (black circles). Ā e specimen exhibits about 5.0%

a CV1 a

c c

(b)

CV2 0.5

R: Distinct contrast corresponding to photo (b) Austenite P

1 c/a

Q

Strain (%)

Martensite

0.4

(d)

0.3 0.2 0.1

R 0 (c)

Temperature

FIGURE 6.4 (a) and (b) Changing contrast of the twinning boundary after martensitic transformation. (c) Schematic illustration of t he relationship between c/a ratio and temperature in Fe–Pd alloy. (d) Increasing angle j with decreasing temperature.

43722_C006.indd 3

−0.1 −0.4

−0.2

0 Magnetic field (T )

0.2

0.4

FIGURE 6.5 Magnetic eld-induced strain as a function of magnetic ux density measured in Fe–Pd single crystal.

10/3/2008 8:45:25 PM

6-4

Smart Materials

(a) 0.0 MPa

(b) 3.2 MPa

(c) 4.8 MPa

(d) 7.5 MPa

(e) 9.5 MPa

Stress (MPa)

15 1st cycle loading 1st cycle unloading 2nd cycle loading 2nd cycle unloading

10

5

0

0

(f)

FIGURE 6 .6 unloading.

1

2

3 Strain (%)

Driving Mechanisms for FSMA-Based Actuators

FSMAs h ave b oth c haracteristics o f t hermoelastic S MAs a nd ferromagnetic p roperties. I n p rinciple, t he ph ase t ransformation and straining also can be controlled by the magnetic eld. Ā e relationship of phase transformation, stress (s), temperature (T), and magnetic eld (H) can be presented as a three-dimensional ph ase t ransformation d iagram a s s hown s chematically in Figure 6.3. Ā e t hree-dimensional su rface i n t he d iagram

43722_C006.indd 4

5

6

(a)–(e) C hange of vol ume r atio of C Vs for t he  rst lo ading a nd (f) s tress–strain c urves for t he  rst t wo c ycles of lo ading a nd

reversible strain and 0.3% irreversible strain in the rst loading unloading cycle. Ā e stress–strain curve of the second test cycle (square dot s i n Figure 6 .6f) a lmost follows t he one o f t he  rst cycle. However, no reversible strain is measured in the second test a nd t he re peatable s train i s 5. 0%. Ā is r epeatable st rain under compressive stress could be utilized as a useful actuation method i f t he applied magnetic  ux gradient that provides the stress is applied to the specimen so as to induce the strain by the above variant rearrangement mechanism. Similarly, the superelastic behavior, SME, and magnetic properties have been studied in Fe–Pt alloys by some groups [30–38]. As shown in Table 6.1, Fe–Pt alloys also have the TMs much lower than room temperature [30,31]. Kakeshita et al. [38] have also reported that the martensitic transformation can be induced by a large applied magnetic  eld, as large as 2.0 × 107 A/m. Ā e TMs of Fe–Pd alloys can be tailored by compositions of Pd between 80°C and −200°C [39,40,43]. However, currently most of FSMA actuator studies have focused on NiMnGa and Fe–Pd alloys.

6.1.3

4

separates t he m artensite a nd a ustenite ph ases. Ā is diagram can h elp u s to de termine a f avorable d riving me chanism f or FSMAs as actuator materials, especially as controlled solely by an applied magnetic eld. Ā is control method by the magnetic eld on FSMAs will be faster than the one by thermal conduction on SMAs. Several driving mechanisms (by t he magnetic  eld) of actuators based on FSMAs have been proposed and studied. The f ollowing s ections su mmarize t he t hree m ain d riving mechanisms. 6.1.3.1

Magnetic Field–Induced Phase Transformation

Ā e ph ase t ransformations o f F SMAs c an b e c ontrolled o r affected by a m agnetic  eld. Ā is is called t he magnetic  eld– induced martensitic phase transformation [7,60]. Ā er efore, the superelasticity may also be controlled or produced by the magnetic eld a nd a large reversible displacement is obtained. Ā is property is advantageous when applied to an actuator device. Ā e forward transformation (austenite → martensite), induced by a m agnetic  eld, h as b een f ound i n Ni MnGa, w hile t he reverse transformation (martensite → austenite), induced by a magnetic eld, also occurs in FePd [60]. As shown in Figure 6.7, the s lope, d H/dT, of FePd is negative and that of NiMnGa is positive. Ā is phenomenon can be explained by the Clapeyron– Clausius relationship [59] dH h = dT TMs ( M M − M A )

(6.1)

10/3/2008 8:45:26 PM

6-5

Ferromagnetic Shape Memory Alloy Actuators

s

s

Martensite

Martensite Austenite Austenite T

T H

Parallel to H-axis

(a)

H (b)

FIGURE 6.7 Schematics of t he ph ase t ransformation d iagram of ( a) FePd a nd ( b) Ni MnGa. (From L iang, Y., K azo, H ., a nd Taya, M ., Mech. Mater., 38, 564, 2006. With permission.)

where h is the enthalpy H is the magnetic eld MM is the magnetization of martensite phase MA is that of the austenite phase For the case of FePd, (MM –MA) is negative before the saturation of magnetization [60]. Hence, Equation 6.1 implies that dH/dT is negative, re sulting i n p romoting t he m artensite → au stenite phase transformation by applying the magnetic eld. However, for t he c ase o f Ni MnGa, i ts m agnetization i n t he m artensite phase i s l arger t han t hat i n t he a ustenite ph ase [4,58], t hus, dH/dT i n E quation 6 .1 i s p ositive, re sulting i n p romoting t he austenite → m artensite ph ase t ransformation. I t h as b een reported that the magnetic eld can induce phase transformation in FSMAs, but this effect is very small in both NiMnGa and FePd alloys b ecause o f a sm all m agnetization d ifference (MM –MA) between martensite and austenite phases. 6.1.3.2

Variant Rearrangement in Fully Martensite Phase by Magnetic Field

In addition to the above superelasticity, a rearrangement (change) in variants in a f ully martensitic phase due to t he application of stress in SMAs will occur and additional strain can be obtained as shown in Figure 6.2b. Here, stress controls variants with different t ransformation st rain. Ā is v ariant re arrangement i nduced by the application of the magnetic eld is also possible in FSMAs. Many studies have concentrated on this mechanism [2,5,6,8– 19,28,29,53]. I n th is c ase, th e m agnetic  eld co ntrols v ariants with different magnetization. Variants having different magnetization may have different transformation strains. Ā us, the magnetic eld may change overall strain of an FSMA, when it is in a martensitic st ate. Ā is is also advantageous when an FSMA is used as an actuator. If the magnetic eld can really discriminate martensite variants with a different t ransformation st rain, a n FSMA can be used as a high damping material as well. First, the movement of variant or martensite boundaries accompanies energy dissipation. Ā is is manifested by a hysteresis, even though it is small in thermoelastic transformation. Secondly, mechanical

43722_C006.indd 5

vibration interacts with the oscillating magnetic  eld that moves boundaries of magnetic domain or variant in martensite phase. Currently, t he v ariant re arrangement i nduced b y t he m agnetic eld is effective in a single crystal FSMA, i.e., NiMnGa. However, a bias stress is needed in order to induce single variant in single crystal F SMA s o a l arge s train c an b e ob tained. A lthough 5 % reversible strain has been reported by this driving mechanism, the output stress is only as high as 10 MPa. Ā is is not suitable for actuators that require a high output force. Furthermore, NiMnGa is very brittle; therefore, it is only suitable for the applications where NiMnGa is under a compression mode. 6.1.3.3

SIM Phase Transformation by Magnetic Field Gradient

It is well-known that ferromagnetic material feels a force when it is in a nonuniform magnetic eld. Ā is force can be large enough to induce the SIM in a superelastic FSMA. Ā is is called hybrid mechanism [ 59]. Ā e h ybrid me chanism c onsists o f c hain re actions: Magnetic  eld gradient causes magnetic force or moment, which introduces stress in an FSMA → SIM transformation → the FSMA becomes much softer and exhibits a l arge deformation. In reality, the hybrid me chanism c an a lso b e ut ilized, w hen a n ac tuator i s designed. Ā at is, a l arge force can be applied to a n actuator by a compact electromagnet with a high magnetic eld gradient, resulting in a larger displacement with fast response. In this driving mechanism, FePd FSMA, which exhibits 1.6% superelastic strain, is more suitable because it has a l arge magnetization factor, w ith the same order as iron [59]. Ā is can be easily seen by a demonstration as shown in Figure 6.8, where two identical cantilever beams made of FePd and NiMnGa are subjected to a m agnetic eld. As shown in Figure 6.8b, the FePd beam bends signicantly because FePd has a much larger magnetization than NiMnGa. In addition, FePd is very ductile, which is suitable for any actuator application. In summary, Table 6.2 shows the comparison among the three driving mechanisms proposed for FSMAs. Mechanism (a) is not practical because a large magnetic eld is needed to change the transformation tem peratures (i.e., i nducing ph ase t ransformation). For example, for NiMnGa in Table 6.2, TMs can be changed

10/3/2008 8:45:27 PM

6-6

Smart Materials

Top view

FePd

Side view

NiMnGa

NiMnGa

Electromagnet

H field

FePd

CCD camera

5 mm (b)

(a)

FIGURE 6.8 Bending comparison between Fe–Pd and NiMnGa cantilever beams: (a) setup and (b) after applied H-eld.

TABLE 6.2 Summary and Comparison of Magnetic Field (H) Actuation Mechanisms Mechanism

Method

Phase

Reported by

NiMnGa

Magnetic eld-induced phase transformation

TMs is changed by H-eld → forward or reverse transformation

A → M or M → A

∆TMs ≈ 9°C at H = 4 × 10 A/m Inoue (Material Science, UW)

∆TMs ≈ −0.2°C at H = 8 × 105 A/m

Variants rearrangement

H-eld applied to martensite → variant change → deformation of the martensite

M

Ā e shear stress acting on martensite under H = 3.2 × 105 A/m, (calculated) = 2.7 MPa

10 MPa

Hybrid mechanism

Force induced by H-eld gradient applied to austenite → SIM

A→M

NiMnGa Vasil’ev et al. [2] FePd Liang et al. [59] NiMnGa Chernenko et al. [62] Murray et al. [63] James et al. [4] FePd James and Wuttig [50] FePd Liang et al. [59]

Superelasticity (no H-eld) single crystal: 1.6%, 120 MPa, compression Murray et al. [25]

Polycrystal: 1.5%, 450 MPa in tension

by a 4 × 106 A/m magnetic eld, however, the electromagnet will be to o h uge to u se f or p ractical app lications. A c onventional compact s olenoid c annot p roduce suc h a h igh m agnetic  eld. Mechanism (b) is not ac hievable for polycrystal FSMA because only a very small strain is available. For example, only the order of 10−4 strain can be obtained for polycrystal FePd in a eld up to 8 × 105 A/m [61]. Even though a large strain can be produced by a small magnetic eld in single crystals of NiMnGa using mechanism ( b), t he o utput f orce i s s till a s sm all a s s everal M Pa. Ā erefore, for the current available FSMA such as the FePd alloy, mechanism (c) (hybrid mechanism) shows the best performance among these mechanisms.

6.2

FSMA-Based Actuator

Previously, three driving mechanisms were summarized for the magnetic eld for FSMAs to be used as actuator materials. Ā ese mechanisms a re (a) m agnetic  eld–induced ph ase t ransformation, (b) variant rearrangement in fully martensite phase by the

43722_C006.indd 6

Fe–Pd 6

magnetic  eld, and (c) hybrid mechanism (SIM transformation by t he m agnetic  eld g radient). Table 6 .2 s hows t hat t he t hird mechanism (hybrid mechanism) has the best performance and is applicable for a c ompact actuator system. Ā e hybrid mechanism i s ba sed on t he force (or moment) ac ting on a f erromagnetic actuator material due to the magnetic eld gradient. Ā en, this force (or moment) i ntroduces a SI M, w hich w ill c ause t he material (initially in the austenite phase) becoming softer (martensite ph ase) a nd t hen a ble to p roduce l arge de formation a nd force. When the magnetic eld i s remo ved, t he f orce i s a lso removed from t he FSMA material and t he material w ill spring back because of superelasticity. As mentioned before, polycrystal FePd has been identied as a promising FM SA f or c ompact ac tuators ba sed o n t he h ybrid mechanism. It is because FePd can exhibit superelasticity at room temperature and has a large magnetization on the same order as iron and is so ductile, which makes it suitable for many applications. In a te st, a c antilever beam made of polycrystalline FePd was simultaneously actuated with a ferromagnetic Fe beam by a

10/3/2008 8:45:27 PM

6-7

Ferromagnetic Shape Memory Alloy Actuators

Fe FePd Fe

FePd

H field Solenoid magnet

5 mm

Side View (a)

(c)

(b)

FIGURE 6.9 Bending of polycrystal FePd and Fe cantilever beams under applied magnetic eld: (a) the schematic of the experimental setup (the side view from the tip of the beams) (b) before and (c) after applying the eld.

portable electromagnet (H = 8 × 104 A/m). Both beams have the same geometry and Figure 6.9a shows the experimental setup. It is clearly demonstrated in Figure 6.9c that the FePd beam has a much larger deection than the Fe beam. Ā e deection is due to the hybrid mechanism. Since the magnetization of FePd in austenite is similar to that of Fe and the ratio of the Young’s modulus of Fe (210 GPa) to FePd (80 GPa in the austenite phase, 30 GPa in the martensite phase) is about in the 3–7 range, the ratios of the bending deection of FePd to Fe could be also 3–7. However, the demonstration results shows that the tip of the FePd beam exhibits a 7 mm displacement during the application of the magnetic eld, while the Fe beam shows almost no deection. Ā is implies that the FePd beam is not only very ductile but also undergoes austenite → martensite phase transformation (SIM). As the gap between the specimen and the electromagnet becomes smaller during the deformation, t he l arger m agnetic f orce le ads to mo re ph ase change and deformation. Ā is comparison shows t hat t he FePd FSMA as an actuator material is superior to the conventional material. Both t he c antilever beams spring back to t he original position when the electromagnet is turned off.

Because FePd is very ductile, it is also possible to make FePd into v arious s hapes. F igure 6 .10 s hows t he v arious s hapes o f polycrystalline F ePd s pecimens, suc h a s ro d, w ire, a nd h elical spring. Ā erefore, this material can be used without limitations of manufacturing. Ā e helical FePd spring can be produced by winding a nd s hape memo rizing t he F ePd w ire, w hich c an b e extruded from a solid cylinder. A series of pictures were taken to exhibit the FePd spring actuated under a compact electromagnet and shown in Figure 6.11 [64]. Ā e spring actuation is driven by a modied electromagnet [65], which can produce a much higher magnetic eld gradient than a conventional solenoid. By gradually increasing the electrical current, the FePd spring partially shrinks due to magnetic attraction. Ā e partial shrinkage of the spring becomes part of the yoke and attracts the rest of the turns of t he s pring a s t he ele ctrical c urrent i ncreases. F igure 6 .11b shows the complete shrinkage of the FePd spring. Upon removing the electrical power, t he spring completely returns to i ts initial length. Although F ePd app ears a s a p romising ac tuator m aterial based on the hybrid mechanism, it is very expensive because of

Fe-Pd wire f 1.6 mm

Fe-Pd rod f 5 mm

Fe-Pd wire f 1 mm

FIGURE 6.10

43722_C006.indd 7

Fe-Pd helical spring D = 22 mm, L = 55 mm

Different shapes of polycrystalline FePd specimens.

10/3/2008 8:45:28 PM

6-8

Smart Materials

Fe-Pd spring

Electromagnet

FIGURE 6.11

5 cm

(b)

(a)

Actuation of the FePd spring actuator: (a) power off and (b) power on.

system, re spectively. Ā e torque actuator consists of an inner rod, which will rotate counterclockwise upon switching on the electromagnet system, thus, attracting the FSMA plate spring to its inner wall. Ā e rotating motion of the inner rod will provide the torque work for a de ad load that is hanging on the rod by a pulley or belt. Ā e FSMA composite is composed of a ferromagnetic material and a superelastic grade SMA, and it is subjected to a b ending moment, which is not u niform over t he leng th of the F SMA c omposite d ue to i ts v arying c urvature. I n t he  rst prototype torque actuator, the FSMA composite consists of NiTi superelastic wires and several cylindrical soft iron rods as shown in Figure 6.14a. Ā e requirement for designing the FMSA composite based torque ac tuators is to i nduce a l arge stress so t hat the S MA p late o f t he F SMA c omposite c an re ach t he o nset o f SIM transformation, while the stress in the ferromagnetic rods remains below its plastic yield stress. Ā e FSMA composite will be attracted to the inner wall of the actuator due to strong magnetic  ux g radient up on s witching o n t he ele ctromagnetic system. Ā e rst prototype torque actuator has produced 0.736 N

Pd. An alternative way to reduce the cost of the material is to use the FSMA composite, which is composed of t he ferromagnetic material and superelastic SMA, where the ferromagnetic material is provides a l arge force due to t he magnetic eld gradient, resulting in a large deformation on superelastic SMA (i.e., NiTi) due to the SIM transformation. Several cases of FSMA composites h ave b een s tudied a nd t he l aminated F SMA c omposite i s easily made without losing its performance and is the most costeffective [66], for example, the laminated plate (Figure 6.12a) and wire of concentric cylinders (Figure 6.12b). As shown in Figure 6.12a, the outer layer of superelastic NiTi SMA can sustain large stress and the inner core ferromagnetic material is subjected to modest stress, thus, leveraging the extra stress bearing capacity of sup erelastic Ni Ti w hile p rotecting a n ot herwise b rittle s oft ferromagnetic material. Several F SMA c omposite ac tuators h ave b een m ade. O ne of the examples is the torque actuator based on the FSMA composite [67] and its design concept is illustrated in Figure 6.13 where (a) and (b) denote t he cases of switch on and off of the actuator

Superelastic TiNi

Ferromagnetic layer

Superelastic TiNi tube

M

Ferromagnetic core Stress distribution

Stress-induced transformation

s s fm

M l

t t SIM

t fm

s TiNi Stress induced

q

t TiNi

Df s SIM

(a) Plate bending

Transformation

D

(b) Wire torsion (helical spring)

FIGURE 6.12 Two types of FSMA composites composed of soft ferromagnetic core and superelastic NiTi, (a) laminated composite plate and (b) concentric cylinder composite.

43722_C006.indd 8

10/3/2008 8:45:29 PM

6-9

Ferromagnetic Shape Memory Alloy Actuators

FSMA composite

20 mm

20 mm (b)

(a)

FIGURE 6.13

Photos of the torque actuator made of FSMA composite (a) switch on and (b) switch off.

NiTi plate

Ferromagnetic cylindrical bars

Ferromagnetic square bars

NiTi wires

(a)

(b)

FIGURE 6.14 Two types of FSMA composites for the torque actuator (a) made of NiTi superelastic wires and cylindrical soft iron bars and (b) made of NiTi superelastic plate and square soft iron bars.

m w ith a 4 0° maximum a ngle. Fu rther i mprovement has been done by using the FSMA composite plate spring, which is made of a sup erelastic Ni–Ti s heet a nd s quare s oft iron bars (Figure 6.14b). B ecause of t he h igher s tiff ness of t he s tructure a nd t he stronger m agnetic at tracting f orce, i t c an p rovide a 4 .8 N m torque with a 102° rotation angle. A synthetic jet actuator based on the FSMA composite membrane has also been constructed [68]. Ā e composite membrane

was driven by the electromagnetic system and oscillated to create a synthetic jet ow through the exit hole. Ā e FSMA composite membrane was composed of a superelastic NiTi thin sheet and a ferromagnetic soft iron pad. Figure 6.15a and b shows the parts of the membrane actuator, respectively. It has two chambers in the center divided by the composite membrane. Ā e thin laminated yoke is used in the electromagnet unit in order to eliminate the e ddy c urrent f or t he h igh-frequency op eration. W hen t he

Exit hole

Chamber#1 ( a)

Chamber#2

Electromagnet

FSMA composite

(b)

FIGURE 6.15 (a) Photo of the synthetic jet membrane actuator and (b) the schematic of actuator parts. (From Liang, Y., Kuga, Y., and Taya, M., Sens. Actuators A, 125, 512, 2005. With permission.)

43722_C006.indd 9

10/3/2008 8:45:30 PM

6-10

Smart Materials

Peak to peak stroke (mm)

Flow velocity (m/s)

250 200 0.2 mm

150 0.3 mm

100 0.4 mm

50 50 (a)

100

150 200 Frequency (Hz)

250

300

4 3.5 3 2.5 2 1.5

0.3 mm thick NiTi plate 1 50

(b)

100

150 200 Frequency (Hz)

250

FIGURE 6.16 (a) Synthetic jet ow velocity as a function of frequency produced by 0.2, 0.3, and 0.4 mm thick NiTi sheets of the composite membrane and (b) peak to p eak stroke of t he composite membrane. (From L iang, Y., Kuga, Y., and Taya, M., Sens. Actuators A, 125, 512, 2005. With permission.)

FSMA c omposite memb rane o scillates c lose to i ts re sonance frequency, t he s ystem e fficiency i s opt imized a nd t he mem brane exhibits a large stroke to produce a strong jet ow. Since the Ni Ti s heet o f t he c omposite memb rane i s sup erelastic, i t can sustain large stresses without plastic deformation. Ā e synthetic je t  ow velocity f rom t he membrane ac tuator was me asured by the hotwire probe at t he exit hole. Figure 6.16a shows the re sults of je t  ow ve locities a s a f unction of f requency by using different thicknesses of NiTi thin sheets of the composite membrane. Ā e maximum jet velocity, 190 m/s at 2 20 Hz, was obtained by using the 0.3 mm thick NiTi sheet of the composite membrane. I ts d isplacement–frequency re sponse i s s hown i n Figure 6.16b where the peak to p eak stroke of the membrane is 3.72 mm at 1 95 Hz. Ā e energ y density of the FSMA composite membrane was reported as 30 kJ/m3 and its power density was 6000 kW/m3 [68]. Both the energy density and the power density can be further improved if the composite membrane exhibits a m uch l arger s troke. A lso, t he h igher f requency re sponse of t he c omposite memb rane c an f urther i ncrease t he p ower density.

References 1. 2. 3. 4. 5. 6. 7. 8.

43722_C006.indd 10

Z. Nishiyama, Martensitic Transformation, Academic Press, London, 1978. M. Wuttig, L. Li u, K. Tsuchiya, a nd R . D . J ames, J. Ap pl. Phys., 87, 2000, 4707. R. D. James and K. F. Hane, Acta Mater., 48, 2000, 197. P. J. Webster, K. R . A. Z iebeck, S. L. Town, and M. S. Peak, Phil. Mag. B, 49, 1984, 295. R. C. O’Handley, J Appl. Phys., 83, 1998, 3263. V. A. Chernenko, E. Cesari, V. V. Kokorin, and I. N. Vitenko, Scripta Metal. Mater., 33, 1995,1239. S. J. Murray, M. Farinelli, C. Kantner, J. K. Huang, S. M. Allen, and R. C. O’Handley, J. Appl. Phys., 83, 1998, 7297. A. N. Vasil’ev, A. D. Bozhko, V. V. Khovailo, I. E. Dikshitein, V. G. Shavrov, V. D. Buchelnikov, M. Matsumoto, S. Suzuki, T. Takagi, and J. Tani, Phys. Rev. B, 59, 1999, 1113.

9. A. D. Bozhko, A. N. Vasil’ev, V. V. Khovailo, I. E. Dikshitein, V. V. K oledov, S. M. S eletski, A. A. T ulaikova, A. A. Cherechukin, and V. D. Buchelnikov, J. E xp. Ā eo ret. Phys., 88, 1999, 954. 10. R. D. James, R. Tickle, and M. Wuttig, Mater. Sci. Eng., A273, 1999, 320. 11. R. Tickle and R. D. James, J. Magn. Mater., 195, 1999, 627. 12. K. Ullakko, J. K. Huang, V. V. Kokorin, and R. C. O’Handley, Scripta Mat., 36, 1997, 1133. 13. H. Miki, T. Takagi, M. Matsumoto, J. Tani, K. Yamauchi, T. Abe, H. N akamura, a nd V. K hovailo, P roc. S ymp. o n Dynamics of Electro-Magnetic Field, 1999, p. 326. 14. Y. Ezer , A. S ozinov, G. K immel, V. E telaniemi, J . T . Glavatskaya, A. D’Anci, C. Podgursky, V. K. Lindroos, and K. Ullako, Proc. SPIE, 3675, 1999, 244. 15. M. Wuttig, L. Liu, K. Tsuchiya, and R. D. James, J. Appl. Phys., 87, 2000, 4701. 16. Qi-Pan and R. D. James, J. Appl. Phys., 87, 2000, 4702. 17. K. Tsuchiya, D. Ohtoyo, H. Nakamura, M. Umemoto, and P. G. McCormik, Mater. Trans. JIM, 8, 2000, 938. 18. K. Tsuchiya, H. Nakamura, M. Umemoto, and H. Ohtsuka, Trans. MSRJ, 25, 2000, 517. 19. K. T suchiya, D . Oh toyo, M. U memoto, a nd H. Oh tsuka, Trans. MSRJ, 25, 2000, 521. 20. S. J. Murray, M. Marioni, S. M. Allen, R. C. O’Handley, and T. A. Lograsso, Appl. Phys. Lett., 77, 2000, 886. 21. R. C. O’Handley, S. J. Murray, M. Marioni, H. Nembach, and S. M. Allen, J. Appl. Phys., 87, 2000, 4712. 22. S. J. Murray, M. A. Marioni, A. M. Kukla, J. Robinson, R. C. O’Handley, and S. M. Allen, J. Appl. Phys., 87, 2000, 5774. 23. Yu. N. Koval, V. V. Kokorin, and L. G. Khandros, Phys. Metal. Metall., 48, 1979, 162. 24. T. M aki, K. K obayashi, M. Mina to, a nd I. Tamura, Scripta Metall., 18, 1984, 1105. 25. T. Kakeshita, K. Shimizu, T. Maki, I. Tamura, S. Kijima, and M. Date, Scripta Metall., 19, 1985, 973. 26. T. Maki, S. Furutani, and I. Tamura, ISIJ Int., 29, 1989, 438. 27. N. Jost and M. Huhner, Pract. Metal., 26, 1989, 295.

10/3/2008 8:45:33 PM

Ferromagnetic Shape Memory Alloy Actuators

28. E. Hornbogen a nd N. Jost, J.Phys.IV, S uppl. J. Phys.III, C4, 1991, 199. 29. T. K ikuchi a nd S. K ajiwara, Mater. T rans. JIM , 34, 1993, 907. 30. G. Kubla and E. Hornbogen, J.Phys.IV, Suppl. J. Phys. III, C2, 1995, 304. 31. S. J . M urray, R . H ayashi, M. M arioni, S. M. Allen, a nd R. C. O’Handley, Proc. SPIE, 3675, 1999, 204. 32. R. H ayashi, S. J . M urray, M. M arioni, S. M. Allen, a nd R. C. O’Handley, Sensor and Actuator-A, 81, 2000, 219. 33. D. P. Dunne a nd C. M. Wayman, Metall.Trans.A, 4, 1972, 137, 147. 34. M. Umemoto and C. M. Wayman, Acta Met., 26, 1978, 1529. 35. M. Foos, C. Frantz, and M. Gantois, Acta Metall., 29, 1981, 1091. 36. C. M. Wayman, Scripta Metal., 5, 1981, 489. 37. K. Sumiyama, M. Shiga, and Y. Nakamura, Phys. Status Solidi A, 76, 1983, 747. 38. T. Kakeshita, T. Shimizu, S. Funada, and M. Date, Trans. JIM, 25, 1984, 837. 39. R. Oshima, S. Sugimoto, M. Sugiyama, and F. E. Fujita, Trans. JIM, 26, 1985, 523. 40. A. C evecka, H. O tsuka, and H. K. D. H. Bhadeshia, Mater. Sci. Tech. Feb, 1995, 109. 41. M. Johmen, T. Fukuda, T. Takeuchi, T. Kakeshita, R. Oshima, and S. Muto, Proc. SMART2000, Sendai, 26. 42. M. M atsui, T. S himizu, H. Yamada, a nd K. Adachi, J. Ma g. Mag. Mater., 1980, 1201, 15–18. 43. T. Sohmura, R. Oshima, and F. E. Fujita, Scripta Metall., 14, 1980, 855. 44. R. Oshima, Scripta Metall., 15, 1981, 829. 45. M. Matsui and K. Adachi, J. Mag. Mag. Mater., 1983, 115, 31–34. 46. M. S ugiyama, R . O shima, a nd F. E. F ujita, Trans. JIM ., 25, 1984, 585. 47. M. Sugiyama, S. Harada, and R. Oshima, Scripta Metall., 19, 1985, 315. 48. M. Matsui, J. P. Kuang, T. Totani, and K. Adachi, J. Mag. Mag. Mater., 1986, 911, 54–57.

43722_C006.indd 11

6-11

49. M. S ugiyama, R . O shima, a nd F. E. F ujita, Trans. JIM , 27, 1986, 719. 50. S. Muto, R. Oshima, and F. E. Fujita, Acta Metall., 38, 1990, 685. 51. R. Oshima, S. Muto, F. E. Fujita, T. Hamada, and M. Sugiyama, MRS. Intl. Mtg. Adv. Mat., 9, 1989, 475. 52. R. Oshima, M. Sugiyama, and F. E. Fujita, Metall. Trans. A, 19A, 1988, 803. 53. H. Seto, Y. Noda, and Y. Yamada, J. Phys. Soc. Jpn., 57, 1988, 3668. 54. R. Oshima, Bull. Jpn. Inst. Metals, 28, 1989, 493. 55. R. O shima, S. M uto, a nd F. E F ujita, Mater. Trans. JIM , 3, 1992, 197. 56. H. M orito, A. F ujita, K. F ukamichi, R . K ainuma, a nd K. Ishida, Appl. Phys. Lett., 81, 2002, 1657. 57. K. Oika wa, L. Wulff, T . I ijima, F . G ejima, T . O hmori, a nd A. Fujita, Appl. Phys. Lett., 79, 2001, 3290. 58. K. U llakko, J. K. Huang, C. K antner, a nd R . C. O’Handley, R. C., Appl. Phys. Lett., 69, 1996, 1966–1968. 59. Y . Lia ng, H. K ato, a nd M. T aya, Mech. Ma ter., 38, 2006, 564–570. 60. Y. Lia ng, Y. S utou, T. Wada, C. L ee, M. Taya, a nd T. M ori, Scripta Mater., 48, 2003, 1415–1419. 61. T. Wada, Y. Liang, H. Kato, T. Tagawa, M. Taya, and T. Mori, Mater. Sci. Eng. A, 361, 2003, 75–82. 62. H. Hamada, J. H. Lee, K. Mizuuchi, M. Taya, and K. Inoue, Metall. Trans., 29A, 1998, 1127–1135. 63. D. S. Ford and S. R. White, Acta Mater., 44, 1996, 2295–2307. 64. T. Wada and M. Taya, Proc. SPIE, 4699, 2002, 294–302. 65. K. Oguri , Y. O chiai, Y. N ishi, S. Ogino , a nd Y. U chikawa, Proceedings o f 9th S ymposium o n I ntelligent M aterials, March 16, 2000, Tokyo, Japan, pp. 24–25. 66. M. K usaka a nd M. T aya, J. Co mposite M ater, 38, 2004, 1011–1035. 67. V. Chen, M. Taya, J. K. L ee, M. Kusaka, and T. Wada, Proc. SPIE, 5390, 2004, 309–316. 68. Y. Liang, Y. Kuga, and M. Taya, Sens. Actuators A Phys, 125/2, 2005, 512–518.

10/3/2008 8:45:33 PM

43722_C006.indd 12

10/3/2008 8:45:33 PM

7 Aircraft Applications of Smart Structures 7.1 I 7.2 7.3 G 7.4 7.5 7.6

ntroduction ........................................................................................................................... 7-1 Smart Structures for Flight in Nature ................................................................................. 7-3 eneral Remarks on Aircraft Design Aspects ...................................................................7-4 Traditional Active or Adaptive Aircraft Control Concepts .............................................. 7-4 Range of Active Structures and Materials Applications in Aeronautics ......................... 7-4 Aircraft Structures ................................................................................................................. 7-5 De nitions for a Structure • Rigidity of Wing Structures • Structures and Mechanisms • Passive Materials for Aircraft Structures • Typical Load Requirements for Aircraft Structures

7.7 Smart Materials for Active Structures ................................................................................. 7-7 7.8 Ro le of Aeroelasticity ............................................................................................................ 7-7 7.9

Cla

German Aerospace Society

7.1

Overview of Smart Structures Concept for Aircraft Control ........................................ 7-10 ssication of Concepts • Fictitious Control Surface Concepts • Variable Shear Stiff ness Spar Concept • Innovative Control Effector Program • Active Flow Control Actuators • Innovative Aerodynamic Control Surface Concepts • Active Structures and Materials Concepts • Other Innovative Structure Concepts • Adaptive All-Movable Aerodynamic Surfaces

Quality of the Deformations .............................................................................................. 7-14 Achievable Amount of Deformation and Effectiveness from Different Active Aeroelastic Concepts............................................................................................... 7-15 7.12 Need for the Analysis and Analytical Design Optimization of Active Structures Concepts ........................................................................................... 7-16 7.13 Summary and Conclusions ................................................................................................ 7-17 Appendix A: Future Directions........................................................................................................ 7-18 A.7.1 Aerodynamic Drag and Structural Design Issues............................................................ 7-19 A.7.2 New Structural Research Efforts and Achievements ....................................................... 7-19 A.7.3 Example for the Interaction of Structural, Aerodynamic, and Aeroelastic Constraints for Different Wing Tip Design Concepts .................................................... 7-19 References ........................................................................................................................................... 7-20 7.10 7.11

Johannes Schweiger

Reputation of Aeroelasticity • Aeroelastic Effects • Aeroelastic Tailoring and Structural Optimization

Introduction

Probably the most famous photography in aviation (Figure 7.1) depicts not o nly t he  rst successful manned powered  ight on December 17, 1903 by the Wright brothers, but it also shows the rst successful airplane with an active structure. Ā e Wright brothers design must denitely be called smart w ith respect to many a spects a nd de sign f eatures. Ā ey w ere a mong t he  rst

pioneers i n a viation w ho h ad re alized t hat u ncoupled c ontrol about all three vehicle axes was required. Ā ey had done systematic e xperimental a erodynamic re search to ac hieve m aximum possible aerodynamic performance. And they had learned how to design and manufacture light weight structures in their bicycle shop. Ā eir s olution f or ade quate rol l c ontrol o f t he a irplane, however, w as mo re t han o ne c entury a head o f t he s tate-ofthe-art in aviation technology. Since the centennial celebrations 7-1

43722_C007.indd 1

10/3/2008 8:45:48 PM

7-2

Smart Materials

FIGURE 7.1 First ight of Wright Flyer I on December 17, 1903 (available as photo hard copy, postcard).

of this remarkable event, no single airplane exists yet, which makes u se o f sm art s tructure c oncepts to c ontrol t he  ight of the vehicle. Rather than ghting the low torsional stiff ness of their braced biplane w ing de sign, t hey u sed t his characteristic i n a p ositive way. By t he side -way mot ions o f t he p ilot, ly ing o n a s liding cradle, the wing tips were twisted in opposite directions by the wires attached to the cradle, thus producing the desired aerodynamic loads to roll the airplane. Figure 7.2 from Orville Wright’s book [1] demonstrated this principle, which is also a v ery good example of t he i mportance of i ntegrated or m ultidisciplinary

design c oncepts, e specially i n a eronautics. U nfortunately, t his knowledge was lost and forgotten over the years, mainly because of more expert knowledge in single disciplines and more formal and bureaucratic processes for the design of new airplanes. Only in recent years, some prophets in aerospace have tried to spread the news about this old ide a again and develop some new ones. Weisshaar [ 2], f or e xample, quote d t he suc cess o f t he W right Flyer as a good example for the need for integrated design methods i n 1986. Ā e Wright F lyer e xample a lso demonstrates t hat smart aircraft structures do not necessarily rely only on advanced active materials. Even earlier than that, active structure concepts were studied. Alois Wolfmüller (1864–1948), the producer of the world’s  rst series moto rcycle, h ad b ought t he s econd o f t he p roduction glider mo del “N ormal-Segelapparat” ( normal s oaring appa ratus) from Otto Lilienthal in 1894 [3]. Both aviation pioneers were communicating a bout i mprovements o f p erformance a nd maneuverability b y c ontrolling t he a ir lo ads t hrough  exible wing t wist. W olfmüller t ried to i mprove t he p erformance b y introducing a  exible h inge to t he w ings i n order to mo dulate aerodynamic control forces from the exible deformations. Today, a ircraft c ontrol i s ac hieved b y me ans o f c ontrol su rfaces attached to the main aerodynamic surfaces. Ā e se devices, aileron, ele vator, a nd r udder, c reate t he re quired f orces a nd moments to c ontrol t he mot ion o f t he a ircraft a bout a ll t hree axes in space. Depending on t he size and speed of t he aircraft, these surfaces are actuated manually or by hydraulic systems. If the Wright brothers had used separate ailerons to roll the airplane, t he add itional structural weight m ight have been too much for the available power from the engine.

Left

Right

Cradle

Cables attached to cradle-sliding cradle to left of machine pulls trailing edge of right wing downward Cable (not attached to cradle) is moved automatically by downward movement of right wing

The accompanying picture of the Wrights’ powered machine (with motor and propellers removed) shows the method of twisting the rear of the wings. A movement of only an inch or two, to the left or right, of the operators’ hips resting on the little cradle was enough to give greater lift to whichever wing needed it, and to restore sidewise balance.

FIGURE 7.2 Active structural concept of t he Wright Flyer I for rol l control. (Adapted from Wright, O., How We Invented the Airplane, Dover Publications, Mineola, NY, 1988.)

43722_C007.indd 2

10/3/2008 8:45:49 PM

7-3

Aircraft Applications of Smart Structures

Ā e ide a o f ac tive o r sm art s tructures f or t he c ontrol o f a ir vehicles i s a s old a s t he e arliest k nown at tempts to  y with heavier-than-air m achines. E arly at tempts b y m ankind to  y were u sually ba sed o n t he e fforts t o u nderstand a nd c opy t he ight of birds. Besides the difficulty to control an unstable ying vehicle, which requires the high-performance computing power of our days or the complex neural network of animals to control their muscular system, it is even more difficult to sense and actuate t he dy namic mot ions o f c ontinuously de forming,  exible aerodynamic surfaces. In most of these efforts, i t w as t he p ilot h imself w ho w as supposed to ac tuate bird-like aerodynamic surfaces to p roduce the re quired l ift f orce, c reate t he f orward t hrust, s tabilize t he vehicle, and apply the necessary control inputs in time. All these early eff orts w ere do omed to f ailure f rom t he b eginning f or a multitude of reasons: • Very limited knowledge of the laws of aerodynamics and ight mechanics. • Insufficient load c apacity of t he available materials, only primitive manufacturing techniques, and only a rudimentary understanding of structural mechanics. • Not enough sustained power output from the human body to produce t he required t hrust, or a l ack of engines w ith sufficient power density. • Efforts trying to copy structural design principles from nature did not take into account the scaling laws of physics; the resulting designs were too heavy, too fragile, or too exible. • In most cases, also the too complex efforts required to stabilize and control a ying vehicle without natural stability. Only in recent years, after o ne c entury p lus si nce t he  rst successful powered ight, it was possible to design, build, and y vehicles p owered b y t he h uman b ody f or t he s ole p urpose o f winning trophies. For these reasons, the rst successes in aviation were only possible by design concepts with almost “rigid” surfaces and natural stability of the vehicles. Nevertheless, a major contribution to the success of the Wright brothers was their “smart structures” ight control s ystem for t he rol l a xis. Ā ey h ad a c ombustion eng ine with sufficient energy density available “just in time.”

7.2

Smart Structures for Flight in Nature

Although c omplete pl ants do not  y, they have to withstand aerodynamic lo ads b y p roper a eroelastic re actions, a nd s ometimes their existence relies on their aerodynamic performance. Ā e seeds of some plants are optimized for long distance ight in shape a nd m ass d istribution b y m illions o f gener ations i n a genetic optimization process. Ā e interest in micro air vehicles in re cent ye ars h as i ncreased a erodynamic re search e fforts in this area [4]. Although not for free ight, the leaves of trees and their s tems a re b uilt to w ithstand s trong w inds. Ā e joints

43722_C007.indd 3

between le aves a nd b ranches h ave to h ave t he r ight a mount of  exibility i n b ending a nd to rsion to re duce a erodynamic loads a nd at t he s ame t ime a void e xcessive u nsteady lo ads from utter. More i nteresting for a ircraft appl ications of s mart s tructures i s t he  ight o f a nimals. A s a lready men tioned a bove, early aviators tried to learn from birds. However, the required knowledge a bout t he ph ysics o f  ight w as m issing f or t hese early a ttempts. Ā e c omplex i nteractions o f s teady a nd unsteady aerodynamic forces with the active motions and passive deformations of wings and feathers are still not completely understood today. We are only now beginning to u nderstand the f unctions a nd i mportance o f t he i ndividual c omponents for t he e fficiency o f a nimal  ight. P endleton [ 5] g ives s ome good examples from prehistoric ying s aurian t o m odern birds with feathers. But even more astonishing are the achievements in the art of ying for another species. Insects show by far the widest variety and m ost adva nced st ructural co ncepts o f ac tive a nd p assive control. An ordinary y can land and take off on the tip of your nose or on the ceiling of a room. Dragon ies like the one shown in Figure 7.3 use a combination of passive mass balance for utter p revention at t he w ing t ips, a nd a n adv anced s tructural design w ith s tiff chitin “spar” elements, supporting the membrane skins, and exible hinges from Resilin to adjust the shape for all ight conditions. Ā e variety and large number of ying members in the family of insects has been attributed to t heir ability to  y, w hich o ffers the advantages to reach and conquer new territories more easily. In the context of formal optimization methods in aeronautics, genetic algorithms became fashionable in recent years. Ā e question in the context of technical products is: can we really wait as long as in nature to get better products? Or should we better continue to rely on gradient-based methods, which can be seen as t argeted, “articial gene tic m anipulations,” si milar to t he “biological engineering” approaches today?

FIGURE 7.3 “Structural” design concept for dragon ies.

10/3/2008 8:45:49 PM

7-4

7.3

Smart Materials

General Remarks on Aircraft Design Aspects

One reason why we have not seen more progress in smart structures applications in aeronautics may be the lack of understanding of the interactions between the different classical disciplines in aircraft de sign, a nd b etween t hese d isciplines a nd the s pecialists f rom t he sm art m aterials a rea f or e ach ot hers needs and capabilities. In o rder to a ssess t he p ossibilities o f sm art s tructures applications f or a ircraft c ontrol, i t i s adv isable to lo ok at some aspects of aircraft design first. The structural engineers are usually concerned with the strength of their design for c ritical c onditions. A nd t hey may b e su rprised i n s ome cases, h ow rel atively sm all de formations o f t he s tructure can h ave s evere i mpacts o n t he a erodynamic p erformance or on the maneuverability. The aerodynamic and flight control eng ineers want t he structure “as r igid a s possible,” but at t he s ame t ime w ith d ifferent s hapes f or d ifferent f light conditions. These requirements may attract the smart structures specialists to make the design deformable and meet at the s ame t ime t ight re quirements f or t he e xternal s hape at changing f light conditions. This requires a look at the stiffness of the structure from strength needs, as well as at the additional internal loads created by the active deformations. And f inally, a eroelastic a spects l ike f lutter s tability a nd effectiveness of deformed aerodynamic surfaces have to be considered. For t he aerodynamic engineer, an aircraft design is always a compromise b etween diff erent ight c onditions: a sm all  at wing for cruise w ith low d rag at h igh speed, a nd a l arge, cambered wing for take-off and landing at low speed. Ā is can partially be met by extendable control surfaces, attached to the xed surface by complex k inematic systems. On t he ot her hand, t he complexity of these systems will increase the structural weight. It lo oks t herefore at tractive to re place t hese me chanisms a nd integrate their functions into one actively deformable structure. To consider such options, it is necessary to look at basic structural design requirements for aircraft wings. Ā ey have to carry the loads from between 2.5 and 9.0 times the total weight of the airplane. And they must be strong enough in torsion to keep the aerodynamic s hape a nd t ransmit t he lo ads f rom c ontrol su rfaces. To achieve this at an acceptable structural weight, sophisticated lightweight design concepts were developed during t he rst half century of manned aviation. Besides improvements on the materials and manufacturing side, it is mainly the principle of l ightweight de sign b y s hape w ithin t he p rescribed a erodynamic shape. Simplied, this principle leads to the placements of the material to t he external shape of t he a irfoil a nd to c losed cross sections o f m aximum a rea. Ā is w ill p rovide m aximum s trength for a  xed amount of material. At the same time, this structure will also have maximum stiffness. Ā is fact must be kept in mind when active deformable s tructures c oncepts a re c onsidered for airframe components.

43722_C007.indd 4

7.4

Traditional Active or Adaptive Aircraft Control Concepts

One kind of active aircraft c ontrol c oncept w as a lready de veloped and demonstrated in the early 1980s: articial stabilization of t he a irplane’s  ight pat h by fast mot ions of t he c ontrol su rfaces, commanded by a d igital  ight control system. Today this system h elps to re duce t rim d rag a nd i ncrease a gility o f a ll modern ghter aircraft by avoiding negative contributions from a stabilizer surface to the total lift force in order to establish natural static stability. Most a irplanes h ave ad aptive w ings b y me ans o f add itional deployable surfaces like slats at the leading edge or Fowler aps at the trailing edge to provide additional lift for take-off and landing, or p rovide c lean  ow c onditions f or e xtreme m aneuvers. Ā e resulting effects are usually a combination of increased wing surface, i ncreased c amber, a nd ac celerated a ir  ow. Ā ese concepts are w ell de scribed i n mo st a erodynamic te xtbooks l ike Re f. [6] where also more sophisticated or exotic concepts like telescopic wings c an be found. Variable w ing sweep is rather common for combat a ircraft, a mong t hem t he A merican F-111. Ā e mission adaptive wing (MAW) demonstrator version of this aircraft used a complex mechanical system to adjust the wing camber for changing ight conditions [5]. However, this system proved too complex and heavy for applications on production aircraft. Active utter stability enhancement by means of control surface deections to create unsteady aerodynamic damping forces was also demonstrated in the 1970s [7]. Ā ere are two main reasons why we do not see them on today’s airplanes: • Flutter w ithin t he  ight en velope at f ailure c ases o f t he system is considered too critical. • Performance of these systems was not very good because of t he lo ss of s tatic a eroelastic e ffectiveness o f t he u sed control su rfaces w ith i ncreasing a irspeed; i t i s t ypically close to zero for ailerons mounted near the wing tip at the trailing edge. Ā e last point also applies to the effectiveness of maneuver load alleviation systems by means of symmetrical deected outboard aileron surfaces.

7.5

Range of Active Structures and Materials Applications in Aeronautics

Besides aircraft control, including load alleviation, aerodynamic performance improvement by shape, and static aeroelastic applications for m aneuverability a nd s tatic s tability en hancement, which w ill b e t he m ain p oints o f t he f ollowing s ections, t he following m ain a dditional a pplications in a eronautics ar e addressed by theoretical and experimental research: Health monitoring. Because piezoelectric materials can serve as actuators a nd s ensors a s w ell, t hese m aterials c an b e u sed to

10/3/2008 8:45:50 PM

7-5

Aircraft Applications of Smart Structures

monitor changing static or dynamic internal loading conditions, resulting from airplane maneuvers or from failures in the structure. A lthough t hese f unctions h ave b een w idely d iscussed i n the literature for t wo decades now, no app lication to a p roduction airplane is known to the author to date. Vibration c ontrol. Ā is application c an b e sub divided i nto t wo categories. For t he t reatment o f lo cal o r g lobal v ibrations o f s tructural components. Ā is was also the rst ight test demonstration of piezomaterials, applied to a s kin panel affected by e ngine noise of the Rockwell B-1B bomber [8]. For the reduction of vibration levels on equipment, mounted to the aircraft structure internally or externally. Ā is can be done to support the integrity of the equipment, or to improve the performance of equipment acting as sensor systems. Because rather small deformations and forces have to be handled here, the most promising and near-term applications of smart materials on airplanes in this area were already predicted in the survey paper by Crowe and Sater [9]. Active f lutter s uppression. A lot o f re search e ffort h as b een dedicated to t his t ask. T he NASA w ind t unnel test program “Piezoelectric Ae roelastic R esponse Tailoring I nvestigation” (PARTI) [10] dem onstrated t his te chnology. A pplications to real aircraft seem to be rather unlikely because of the abovementioned s afety a spects. Pa nel f lutter a s a s pecial c ase f or the instability of individual skin panels at supersonic speed is also addressed in many publications. The required effort for the installation and control of the active devices seems to b e inappropriate compared to a simple structural reinforcement of the critical panels. Acoustic noise reduction. Several research projects are dealing w ith c abin no ise re duction, e specially f or t urbo-prop airplanes.

7.6 7.6.1

7.6.2

Rigidity of Wing Structures

Considering active deformations of a wing structure, it is useful to look at the differences i n t he p ossible pa ssive de formations under diff erent lo ading c onditions. Ā e lift f orce i s t he l argest component of the aerodynamic forces. It corresponds to multiples of t he tot al weight of t he a irplane; 2 .5 t imes for t ransport aircraft and currently up to n ine times for  ghters. At the same time, the external shape of a wing has the smallest dimension in its height and the largest in spanwise direction. Ā is means that the structure needs the largest cross sections of the skins on the upper and lower surface. To actively deform a w ing in bending would t herefore b e d ifficult b ut not i mpossible. H owever, t he bending deformation of an unswept wing has no impact on the aerodynamic characteristics or loading conditions. Only swept wings a re sensitive in t his respect. Swept forward, t he bending will i ncrease t he lo cal a erodynamic a ngle o f at tack i n s treamwise direction, as indicated in Figure 7.4. Ā is will increase the bending mo ment a nd c auses s tructural d ivergence at a  ight

Aircraft Structures Defi nitions for a Structure

To be able to design active structures systems for airplane control, i t i s ne cessary to u nderstand t he f unctions o f t he structure, the requirements that define its properties, and the reasons why existing aircraft structures are built in a specific way. Ā e main functions of the airframe are • To bear and carry the loads acting on the vehicle. • Provide s pace a nd supp ort f or eng ines, e quipment, pa yload, and fuel. • Satisfy the tight rigidity and smoothness requirements for the external shape from aerodynamics. In other words, the external shape of the vehicle is de ned by aerodynamic p erformance re quirements, a nd t he s tructure has to me et i ts re quirements w ithin t his s hape at m inimum weight. Ā is requires a c areful combination of t he shape a nd load-carrying p urpose o f t he i ndividual s tructural elemen ts

43722_C007.indd 5

with the best available material for this purpose. Early designs used f abric to c ollect t he a erodynamic p ressure lo ads o n a wing s urface an d t ransfer th em t o th e fu selage th rough a wooden framework, supported by thin wires. As soon as it was d iscovered t hat t he a erodynamic d rag o f t hese w ires i s 100 times higher than that of a carefully designed airfoil with the s ame th ickness, d esign e fforts concentrated on cantilevered w ings where t he loads could completely be t ransferred internally. Aluminum skins paved the way to mo dern monocoque airframes, where the main loads are carried by a monolithic shell structure, which mainly needs the internal structure only to m aintain t he s hape. Ā e s hape o f t his s hell p rovides highest re sistance a gainst b ending a nd to rsion lo ads at m inimum weight.

V

Streamwise deformation

V Streamwise deformation Deformations along elastic axis

Deformations along elastic axis Swept back wing

Forward swept wing

FIGURE 7.4 Bending deformation of swept wings and impacts on the aerodynamic angle-of-attack.

10/3/2008 8:45:50 PM

7-6

43722_C007.indd 6

constraints, no app lication o f t his pa ssive a eroelastic c ontrol feature on a realistic wing design is known to the author.

7.6.3

Structures and Mechanisms

To a c ertain e xtent, t he m ain f unctions o f s tructures a nd mechanisms a re opp osite: a s tructure must provide r igidity and a me chanism m ust p rovide l arge de fined mot ions between parts. If active structures are considered, both functions must be integrated into the structure, the performance of t his s ystem should be better, a nd t he tot al weight should be lower compared to a conventional design. This shows the difficulty of the task to develop active structures for aircraft control. Ā e i ntention to m ake t he structure more  exible in order to a llow de formations i s t herefore a c ontradiction. To a llow deformations w ithout p roducing i nternal f orces, h inges a re required. If this function is desired within the structure, the structure has to b ecome more  exible in distinguished small regions. B ut t his a ttempt w ill c reate h igh i nternal l oads i n regions w ith sm all c ross s ections. A nd t he de sired de formations for aircraft control functions will produce a high number of load cycles.

7.6.4 Passive Materials for Aircraft Structures Lightweight aircraft structures are obtained by optimal shape and the best suitable materials for the load levels, and type of loading, Figure 7.5 shows the achievements in aircraft design by new materials. Unfortunately for active structures concepts, today’s h ighperformance c omposite m aterials a re s tiffer t han p revious aluminum structures. It is therefore a misbelief by some active aircraft s tructures re searchers to t alk a bout “ highly  exible” composite s tructures f or t heir de signs. O n t he ot her h and, to day’s skepticism against future applications of smart materials may be as wrong as the statement in one of the earliest textbooks on aircraft s tructures, where the author states that metal will never be used i n a ircraft s tructures b ecause o f i ts to o h igh den sity c ompared to metal [11]. Obviously, the author was only looking at iron at that time.

Aircraft performance

speed, called d ivergence speed, which depends on t he bending stiff ness and geometrical properties of the wing. Ā e same effect causes a reduction of the bending moment on a swept back wing under load, which acts as a pa ssive load alleviation system. But to actively control t hese deformations by i nternal forces wou ld mean to s tretch a nd compress t he skins i n spanwise d irection, which is a rather difficult task. Ā e aerodynamic drag forces, acting in streamwise direction, are smaller than the lift forces by a factor of 10. At the same time, the shape of the airfoil creates a high static moment of resistance in t his d irection. For t hese re asons, t he lo ads i n t his d irection need no s pecial at tention i n t he s tructural de sign. A n ac tive deformation would be both very difficult and meaningless. Torsional loads on a wing can be very high, depending on the chordwise center of pressure locations and on additional forces from deected control surfaces. A center of pressure in front of the ctive elastic axis through the wing cross sections will cause torsional d ivergence at a c ertain  ight s peed, a nd a c enter o f pressure to o f ar b ehind t he el astic a xis w ill c ause t he w ing to twist against the desired angle of attack or control surface deection. Ā e wing torsional deformation is therefore very sensitive for t he loads acting on t he w ing, for t he spanwise lift distribution, which is i mportant for t he aerodynamic d rag a nd for t he effectiveness o f t he c ontrol su rfaces. A s men tioned b efore, a closed torque box w ith maximum cross section is desirable for the structural designer. But the possibility to adjust the torsional exibility would also allow several options for the active control of a erodynamic p erformance, lo ad d istribution, a nd c ontrol effectiveness. Ā is active control by internal forces could theoretically be achieved by active materials in the skins, or by an internal torque device, xed at the wing root and attached to the wing t ip. Suc h a de vice, ba sed on s hape memory a lloy (SMA), was a lready d emonstrated on a w ind-tunnel mo del [ 4]. For practical applications, t he pa rameters t hat de ne t he torsional stiff ness of t he s tructure w ith a c losed c ross s ection s hould b e kept in mind. It is proportional to the square of the complete cross-sectional a rea, a nd l inearly p roportional to t he a verage thickness of the skin. Ā is demonstrates how difficult it would be to modify the stiff ness by changes to the skins or by an internal torque tube with a smaller cross section. Ā e most often mentioned application of a ctive structures for aircraft application is camber control and the integration of control su rface f unctions i nto t he m ain su rface by c amber c ontrol. Ā is would me an a h igh c hordwise b ending de formation of t he wing box. As mentioned above for the spanwise bending, the skins would have to be stretched and compressed considerably, but this time ba sed o n a sm aller re ference leng th a nd w ith a sm aller moment arm. For t his reason, we do not s ee chordwise bending deformations o n c onventional w ings u nder lo ad. A eroelastic tailoring, addressed below, by means of adjusting the carbon ber plies i n t hickness a nd d irection to me et de sired de formation characteristics, was a lso addressing camber control in t he 1970s and 1980s as one specic option. Because of the above mentioned

Smart Materials

Active structures Composite Metal Wood

1900

Year

2000

FIGURE 7.5 A/C performance improvements from materials.

10/3/2008 8:45:50 PM

7-7

Aircraft Applications of Smart Structures

Ā e gure a bove a lso i ndicated t he t ypical p erformance trends f or ne w te chnologies. W hen t hey a re i ntroduced, t hey are i nferior to t he b est a vailable s tate-of-the-art te chnology at t hat t ime. Ā e book “ Ā e Innovator’s Dilemma” by Clayton M. C hristensen [ 12] de scribes t his t rend f or s everal ne w, disruptive te chnologies. F irst, app lications a re t ypical o n low-cost, low-performance products.

7.6.5 Typical Load Requirements for Aircraft Structures A typical ghter aircraft has to be designed to carry a load of nine times its own weight. Applied to a car with an empty weight of 1 t plus 0.5 t payload, this would mean an external load of 14.5 t. Ā e wings f or a tra nsport a ircraft ha ve t o ca rry 2.5 times the t otal weight. Ā is sho ws t hat air plane st ructures ha ve to b e st rong, which means that they also are rather stiff. Ā e upper and lower skins of the torque box have a typical thickness in the order of 10 to 20 mm, compared to body of a car below 1 mm.

7.7 Smart Materials for Active Structures Smart materials for active structure applications are mainly interesting b ecause of their hig h energy density. On the o ther hand, their strain or stroke capacity is rather limited, as compared with other ma terials f or a ircraft s tructures, or w ith ot her a ctuators. And the y a re ra ther hea vy. P robably the most co mplete sur vey paper o n this t opic wi th the ti tle “smart a ircraft st ructures” b y Crowe a nd Sa ter [9] was p resented in 1997 a t the AGARD Symposium on Future Aerospace Technology in t he S ervices of the Alliance. It tries to classify the different concepts and gives an overview on recent and ongoing research activities. Ā ere is also a prediction f or f uture a pplications in r eal system s wi th diff erent purposes and for different classes of airplanes. Ā e piezoelectric materials and SMA are mainly addressed for potential applications. Whereas piezoelectric materials are usually applied as patches with multiple layers, distributed over the surface, or as concentrated stack actuators, SMAs have mainly been investigated in the f orm o f wir es o r t orque t ubes f or the ac tive deformation of aerodynamic surfaces. To date, the achieved results are no t v ery p romising f or a ircraft co ntrol b y ac tive st ructures. For t his pu rpose, w hich me ans l arge st atic d eformations i n a rather s hort t ime, pi ezomaterials re spond f ast e nough, but t he stroke i s ve ry l imited. SM As on t he ot her h and, c ould prov ide larger deformations and higher forces, but are not fast enough for ight control inputs, and their thermal ener gy supply issues a re rather complex within the airframe and aircraft environment. Because of these limitations, the Defense Advanced Research Projects Agency (DARPA) launched an ambitious research program in 1999, called compact hybrid actuators (CHAP), with the aim to multiply the stroke and force output of current actuators by a factor of 10.

43722_C007.indd 7

7.8 7.8.1

Role of Aeroelasticity Reputation of Aeroelasticity

Some y ears a fter the Wright brothers’ s uccess with their a ctive wing, designers began to f ear the  exibility of the structure. Ā e famous MIT Lester B. Gardner Lecture “History of Aeroelasticity” from Raymond L. Bisplinghoff [13] quotes many of the early incidents involving aeroelastic phenomena, and the famous comment from Ā eodore von Kármán: “Some fear  utter because t hey do not understand it, and some fear it because they do.” Also quoted from a re view paper on Aeroelastic Tailoring by T. A. Weisshaar [2]: “As a result, aeroelasticity helped the phrase ‘stiffness penalty’ to enter into the design engineer’s language. Aeroelasticity became, in a manner of speaking, a four-letter-word.…it deserves substantial credit for the widespread belief that the only good structure is a rigid structure.” Ā e role of aeroelasticity in aviation is depicted in Figure 7.6. It shows the impact on aircraft performance over the years, m ainly c aused b y i ncreasing s peed. But t he upp er dot i n 1903 a lso i ndicates t hat a eroelasticity c an a lso ac t i n a p ositive way, if properly used and understood, also today, and on faster airplanes. Smart structures concepts will help to reverse this negative trend of aeroelastic impacts on aircraft performance. Similar to Figure 7.5, the progress in aeronautics can also be connected to the p rogress i n a eroelasticity a nd rel ated e xternal s timuli a nd events, as drafted in Figure 7.7.

7.8.2 Aeroelastic Effects Because of the difficulty to describe aeroelastic effects by proper theoretical models, involving a good description of the structure with its exibility and structural dynamic characteristics as well as steady and unsteady aerodynamic properties, solutions were limited in the early years of aviation to selected cases with only a few degrees of freedom. More general solutions required modern computers w ith re spect to s torage s pace a nd c omputing t ime. Ā e aeroelastic triangle (Figure 7.8), for the  rst time quoted by Collar [14] in 1946, describes the involved types of forces in the different a eroelastic ph enomena. L ooking at t hese f orces a nd

Active aeroelastic concepts

Performance

Aeroelastic impacts

Aeroelastic degradation

Wright flyer I Langley aerodrome

Rigid A/C performance

1903

Year

2000

FIGURE 7.6 Impact of aeroelasticity on aircraft performance.

10/3/2008 8:45:51 PM

7-8

Smart Materials

Active materials

Aircraft performance increase

Finite element methods Computers

Composite materials Aeroelastic tailoring

Active aeroelastic concepts

Analytical methods

Theoretical models Description of phenomena Discover phenomena 1900

2000 Year

FIGURE 7.7 Relationship between aircraft performance, advances in aeroelasticity, and external stimuli.

er ca ati St

s

nic

Dynamic aeroelasticity

ha

ec

oe

tm

las

gh

tic

Fli

ity

Aerodynamic forces

Structural dynamics Elastic forces

Inertia forces

FIGURE 7.8 Aeroelastic triangle.

interactions, it becomes obvious that smart structures for aeronautical appl ications w ill h ave a c lose re lationship w ith a eroelasticity in most cases. 7.8.2.1 Static Aeroelasticity In static aeroelastic effects, no inertia forces are involved by denition. Ā is is true for aileron reversal, an effect, where the rolling moment due to a control surface deection changes the sign at a certain ight speed due to opposite deformation of the  xed surface i n f ront o f t he c ontrol su rface. Ā is effect h as to b e avoided w ithin t he  ight en velope o f t he a ircraft i n o rder to avoid deterioration of t he pilot when he moves t he stick to rol l the airplane. If the Wright brothers had used conventional ailerons on their rst a irplanes, t hey m ight h ave e xperienced a ileron re versal because of the low torsional stiff ness of their wings, even at very low speeds. On the other hand, the Wright brothers’ main competitor, Samuel P. Langley, was very likely less fortunate with his Aerodrome designs because of insufficient aeroelastic stability [13]

43722_C007.indd 8

after scaling the successful smaller unmanned vehicle to l arger dimensions. For ghter airplanes, it is not sufficient to avoid aileron reversal. Even at t he worst  ight condition, a high roll rate must be achieved to provide high agility. Ā is is usually done by reinforcing the wing structure because the basic static design of a ghter wing yields rolling moment effectiveness slightly above or even below zero at t he worst  ight condition. In the case of the U.S. aircraft F-18, the basic design had to be revised after delivery of the  rst batc h o f p roduction a ircraft. An additional weight of 200 lbs p er w ing side w as re ported to t he author for t he Isr ael Lavi lightweight ghter aircraft project to provide sufficient roll power. In addition to the loss of roll power, the adverse deformation of the control surface required larger control surface deections, which resulted in higher hinge moments and required stronger actuators. Ā e difficulty i n predicting t he most effective distribution of a dditional s tiff ness for improved roll effectiveness, especially in c onjunction w ith t he in troduction o f m odern composite materials with highly anisotropic stiffness properties for a irframe de sign, i nspired t he de velopment o f f ormal mathematical structural optimization methods [15]. Aileron reversal has usually the most severe static aeroelastic impact on aerodynamic forces and moments. But all other aerodynamic performance or control characteristics of a n a irplane are affected as well by static aeroelastic deformations and aerodynamic lo ad re distributions to a mo re o r le ss s evere de gree. Weisshaar [16], f or e xample, men tions t he e xcessive t rim d rag due to a eroelastic w ing de formations o n t he del ta w ing o f a supersonic transport aircraft. Roll c ontrol i mprovement b y me ans o f ac tive c oncepts w as and still is the most often studied application of active concepts for s tatic a eroelastic ph enomena o n a ircraft. A lthough ac tive structures or materials are not i nvolved, the Active Aeroelastic Wing project [17] or Active Flexible Wing project, as it was called before, h as u ndergone  ight test trials in 2001 on a modied

10/3/2008 8:45:51 PM

7-9

Aircraft Applications of Smart Structures

Front spar

FIGURE 7.9 Active Aeroelastic Wing model mounted in the NASA transonic dynamics wind t unnel at the N ASA Langley Research Center. http:// www.dfrc.nasa.gov/Newsroom/.X-Press/special-editions/AAW/images/ 121703/windtunnel-500.jpg (July 6, 2008).

F-18. Ā is concept originated in several theoretical studies and wind tunnel demonstrations in the 1980s. A su mmary of these activities is presented in a special edition of the Journal of AircraĀ in 1995 [18]. Figure 7.9 depicts the wind tunnel model installed in the transonic dynamics tunnel at the NASA-Langley Research Center (NASA-LARC). Static a eroelastic e ffectiveness lo sses f or t he l ateral s tability and rudder yawing moment are a well-known design driver for vertical tails. Surprisingly, almost nobody has looked so far for smart structures concepts to obtain better designs. Sensburg [19] suggested a smart passive solution, called the Diverging Tail, by means of aeroelastic tailoring the composite skins and modifying t he  n ro ot at tachment to a si ngle p oint a ft p osition, to achieve higher yawing moments compared to a rigid structure. Aeroelastic d ivergence w as t he mo st s evere i nstability f or early monowing airplanes. If the wing main spar is located too far b ehind t he lo cal a erodynamic c enter o f p ressure ( at 25 % chord), a lack of torsional stiff ness will cause the wing structure to d iverge a nd b reak at a c ertain s peed. A s A nthony F okker describes in his book [20], sufficient strength of the design had already been demonstrated by proof load and  ight tests for his D-VIII ( Figure 7 .10) w hen re gulations c alled f or a rei nforced rear spar with proportional strength capacity to the front spar. Ā is redistribution of stiff ness caused torsional divergence under ight loads. Ā is example also demonstrates the potential effects and impact from smart structures applications to an airplane structure. Ā e introduction of high-strength composite materials, with the p ossibility to c reate b ending–torsion c oupling e ffects from the anisotropic material properties, caused a rena issance of the forward s wept w ing i n t he l ate 1970s [2], w hich w as r uled o ut

43722_C007.indd 9

Rear spar

FIGURE 7.10 Fokker D-VIII monoplane, where aeroelastic divergence caused s everal f atal a ccidents a fter re inforcement of t he re ar s par. (Modied by author, photo from Internet.)

before for h igher s weep a ngles b ecause of t he b ending–torsion divergence, as depicted in Figure 7.4. Static aeroelasticity also includes all effects on aerodynamic load d istributions, t he e ffectiveness of a ctive lo ad a lleviation systems by control su rfaces, a nd  exibility effects on t he aerodynamic p erformance. I n t his c ase, t he v ariable i nertia lo ads from payload or fuel on the structural deformations have to be considered simultaneously. 7.8.2.2

Dynamic Aeroelasticity

Flutter is the best known dynamic aeroelastic stability problem. It belongs to t he category of self-excited oscillation systems. In this case, any small external disturbance from a control surface command or at mospheric t urbulence, w hich e xcites t he eigenmodes of the structure, will at the same time create additional unsteady aerodynamic forces. Depending on the mass and stiffness distribution, and on the phase angles between the involved vibration modes, the aerodynamic forces will dampen the oscillations, or enforce them in the case of utter. Active control for  utter stability en hancement by means of aerodynamic control surfaces was fashionable in the late 1970s [21]. In this case, the effectiveness of the system depended on the static aeroelastic eff ectiveness of t he activated control surfaces. Mainly b ecause o f s afety a spects, b ut a lso b ecause o f l imited effectiveness, no ne o f t hese s ystems h as entered s ervice s o f ar. Active control by active structures devices was a popular research topic i n re cent ye ars [10], b ut i t i s do ubtful i f w e w ill e ver s ee applications because of the same reasons. Panel  utter i s a s pecial c ase, w here o nly i ndividual s kin elements of the structure (panels) are affected. Ā is usually happens at low supersonic speeds, and only structural elements with low static load levels, like fairings, can usually be affected. Active control by smart materials is possible, but there are no considerable impacts on aircraft effectiveness.

10/3/2008 8:45:51 PM

7-10

Smart Materials

Buffeting is a forced vibration where turbulent ow generated by one aerodynamic surface excites this surface itself or another surface lo cated i n t he re gion o f t he t urbulent  ow. A lso h ere, aerodynamic c ontrol su rfaces lo cated o n t he a ffected pa rt c an be used to counteract the vibrations. Compared to utter, the aerodynamic e ffectiveness o f t hese s urfaces is a dditionally reduced because of the turbulent  ow conditions. Active structures systems are more effective in this case. For this reason, and because t he re quired ac tive de formations a re sm all, t he  rst large-scale ac tive s tructures app lication o n a ircraft de alt w ith the buffeting problem of ghter aircraft vertical tails at extreme maneuver conditions. A fter s everal t heoretical [22] a nd sm allscale experimental studies [23], full-scale ground tests were performed in a jo int Australian–Canadian–U.S. research program [24] on a n F-18 a nd i n a G erman program for a si mplied n structure o f t he Eu roghter [ 25]. I n b oth c ases, p iezoelectric material was used.

7.8.3

Aeroelastic Tailoring and Structural Optimization

Weisshaar [26] was one of the rst researchers who tried to give aeroelasticity a better reputation when modern ber reinforced composite m aterials w ith h ighly a nisotropic d irectional s tiffness properties were considered for primary aircraft structures. Ā ey provided the possibility to tailor the materials’ directional stiff ness w ithin t he composite lay-up i n order to me et de sired deformation c haracteristics f or i mproved a eroelastic p erformance. Together with formal mathematical optimization methods f or t he s tructural de sign, t his app roach a llowed f or t he minimization of the impact from aeroelasticity. Any improvement of a technical system is often referred to as an optimization. In structural design, this expression is today mainly used for formal analytical and numerical methods. Some years after the introduction of nite element methods (FEM) for the analysis of aircraft structures, the rst attempts were made to use t hese to ols i n a n a utomated de sign p rocess. A lthough t he structural weight is usually used as the objective function for the optimization, the major advantage of these tools is not the weight saving, but the ful llment of aeroelastic constraints. Other than static strength requirements, which can be met by adjusting the dimensions of the individual nite elements, the sensitivities for the elemen ts w ith re spect to a eroelastic c onstraints c annot b e expressed so easily. Ā e option to tailor the composite material’s properties b y in dividual p ly o rientations an d diff erent layer thickness for t he individual orientations required a nd inspired the development of numerical methods [27].

7.9 Overview of Smart Structures Concept for Aircraft Control 7.9.1 Classification of Concepts Active structures concepts for aircraft control can be subdivided into categories by

43722_C007.indd 10

• Purpose of the active system • Types of devices to activate the structure In t he  rst case, t he i ntended concepts a re a iming at i mprovements of • o C ntrol effectiveness • Aerodynamic drag reduction by adaptive shape • Load alleviation by adaptive deformation • taS bilizer effectiveness for trim and static stability As men tioned e arlier, t he m ajority o f c oncepts a re a iming at improved rol l control power because it u sually has t he h ighest sensitivity from structural deformations. A classication by actuation devices can be given by • Activation of a passive structure by conventional or novel aerodynamic control surfaces • Active structural elements • Actuators or connecting elements with adaptive stiffness between structural components An additional classication can be made by • C oncepts, w here a eroelastic e ffects are intentionally used • Concepts without special aeroelastic considerations As far as aeroelasticity is addressed by the concepts, the intended improvements a re a iming at t he h igh-speed pa rt o f t he  ight envelope, where aeroelastic effects b ecome more i mportant. I n the case of exploiting aeroelastic effects in a positive sense, this also m eans t hat a ctive a eroelastic e ffects c an u sually o nly b e used in a b enecial way at h igher speeds. An exception w ill be shown below.

7.9.2

Fictitious Control Surface Concepts

In o rder to e valuate t he p otential b enets o f s mart s tructure concepts as well as the required energy to activate them, it is useful to s tart w ith a “v irtual c oncept,” a ssuming t hat t he intended structural deformation is created by any device. Khot et al. [28] call this the “ctitious co ntrol s urface” co ncept. Ā ey investigated the aeroelastic loss of roll control power for a conventional t railing e dge c ontrol su rface, a nd t hen t ried to retwist the wing by supplying the same amount of strain energy that was created by the aileron deection. Ā e main purpose of this effort was the analytical evaluation of the required energy that is required to maintain a constant roll rate with increasing dynamic p ressure. Ā e re sult, w hich s howed a n i ncrease o f energy versus dynamic pressure with the same gradient as the reduction of effectiveness, however, may be misleading. Ā e achievable rol ling mo ment f rom a de formation de pends o n the position, where t he deformation is initiated by a n internal force or by a control surface deection. In the same way as on a rigid wing, where a trailing edge surface is much more effective than a le ading e dge su rface, t here a re mo re o r le ss e ffective regions o n a f lexible w ing, w here a de f lection o f a c ontrol

10/3/2008 8:45:52 PM

Aircraft Applications of Smart Structures

7-11

surface o r a de formation o f t he s tructure b y i nternal f orces results in diff erent rol ling mo ments a nd re quires d ifferent efforts to create the deection or deformation.

7.9.3 Variable Shear Stiffness Spar Concept In a si milar w ay to t he  ctitious c ontrol su rface, a s tudy b y Griffi n a nd H opkins [ 29] u sed a “ ctitious” va riable st iffness spar concept to modulate the rolling moment effectiveness of a generic F-16 wing model. Ā ey assumed a small outboard trailing edge control surface on a n a nalytical F-16 w ing model for roll control, which would operate in a conventional mode at low dynamic pressures and the negative “postreversal” effectiveness could be enhanced by turning the spar web shear stiffness off at high dynamic pressures. Ā is concept was explained in a si mplied way by “link elements,” attached to the upper and lower spar c aps b y b olts a nd remo vable p ins. Ā e ba sic p rinciple o f this concept was also experimentally veried on an aeroelastic wind t unnel mo del f or a n u nswept, re ctangular w ing w ith removable spars [30]. Unfortunately, no references were found, for more technical smart structures solutions if they were ever investigated for this concept.

7.9.4

Innovative Control Effector Program

In the Innovative Control Eff ector (ICE) program from NASA, Langley [ 31,32], t he p ositions a nd re quired a mount o f sm all, “ctitious c ontrol su rfaces” w ere de termined b y me ans o f a genetic opt imization p rocess f or a n adv anced “ blended w ingbody” conguration. Ā ese “control effectors” were elements of the su rface g rid i n t he a nalytical a erodynamic mo del t hat created t he “v irtual” shape change. Ref. [31] g ives a n e xcellent overview on all research activities within the NASA’s Morphing Program.

7.9.5

Active Flow Control Actuators

Also as a part of the NASA Morphing Program, “synthetic jet actuators were developed a nd tested” [31]. Ā is device is based on a piezoelectrically driven diaphragm, which sucks and blows air through a small orice. It was originally developed for cavity noise control. To use it for aircraft co ntrol, where much higher forces are required, the power output needs to be multiplied.

7.9.6 Innovative Aerodynamic Control Surface Concepts Although t here a re no ac tive s tructural c omponents i nvolved, these concepts can also be considered as “smart structures.” In this case, the active deformation of the structure is actuated by aerodynamic control surfaces. Ā e January/February 1995 edition of the Journal of AircraĀ [18] was a special issue, dedicated to the U.S. Active Flexible Wing Program, which started in 1985 and turned later into the Active Aeroelastic Wing Program. Ā is basic idea was the improvement of roll effectiveness for a ghter

43722_C007.indd 11

FIGURE 7.11 F-18 Ac tive Ae roelastic W ing d emonstrator a ircraft. (From Sanders, B., Flick, P., and Sensburg, O. International Forum on Aeroelasticity and Structural Dynamics, Madrid, Spain, 2001.)

aircraft w ing by t he c ombination of t wo le ading e dge a nd t wo trailing e dge c ontrol su rfaces, w hich c ould a lso b e op erated beyond reversal speed. Ā is c oncept w as demo nstrated o n a n aeroelastic wind tunnel model, with tests started in 1986. After theoretical studies on F-16 and F-18 wings, reported by Pendleton [5,17,33], the F-18, depicted in Figure 7.11, was selected as t he c andidate f or  ight te st dem onstrations, w hich a re expected to started in 2001. For this purpose, the wing structure was returned into the original stiffness version, which had shown aileron reversal in early ight tests. Flick and Love [34] performed a study on wing geometry sensitivities with respect to the potential improvements from active aeroelastic c oncepts ba sed o n t he c ombination o f le ading a nd trail edge surfaces. Ā e results, as shown in Figure 7.12 from Ref. [5], i ndicates o nly v ery sm all adv antages f or lo w-aspect r atio wings. Ā e t heoretical s tudies for a gener ic w ing mo del o f t he Euroghter wing by the Flick, however, resulted in large improvements also for this conguration, as can be seen in Figure 7.6. Active aeroelastic concepts research by TsAGI in Russia has already demonstrated impressive improvements in ight test. In addition to a lso using leading edge control surfaces to i mprove roll p erformance, a sm all c ontrol su rface w as mo unted at t he tip l auncher. F igure 7 .13 f rom Re f. [ 35] s hows t he ac hievable improvement compared to t he t railing edge a ileron only. Note the si ze of t he special su rface i n c omparison w ith t he c onventional aileron. For h igh a spect r atio t ransport a ircraft wings, especially in combination w ith a w inglet, si milar de vices could be u sed not only for roll control, but also for adaptive induced drag reduction, or load alleviation, as indicated in Figure 7.14 for a concept, called active w ing t ip c ontrol ( AWTC) b y S chweiger a nd S ensburg [36]. In this case, the winglet root provides sufficient space and

10/3/2008 8:45:52 PM

7-12

Smart Materials

Taper ratio = 0.2 Taper ratio = 0.2

1

Conventional

1.1 1.0 AAW concept

5.0

4.5

AAW concept

Aspe4.0 ct ra

tio

3.5

0.03 0.04 s s 0.05 ne ick 3.0 0.06 h T

3.0

3.5 4.0 Aspect ra tio

4.5

5.0

Normalized TOGW

2

Wing skin weight

1.2 3

0.9 0.06 0.05 0.04 0.03 t/c

Conventional

FIGURE 7.12 AAW technology advantages and wing geometry sensitivities for l ight weight ghter wings. (Results from Flick, P.M. and Love, M.H. RTA Meeting on Design Issues, Ottawa, Canada, October 1999;  gures from Sanders, B., Flick, P., and Sensburg, O. International Forum on Aeroelasticity and Structural Dynamics, Madrid, Spain, 2001.)

wdx

Aileron+ special aileron (δs.a. = δa.)

0.2

for: Increased roll effectiveness Drag reduction Load alleviation Reduced aeroelatic effectiveness

Increased aeroelastic effectiveness - By torsion

0.1 Wing box

Aileron Elastic axis

Computations Flight tests 0

700

900

Ve (km/h)

FIGURE 7.13 AAW technology in Russia:  ight test results and comparison w ith a nalysis for a s pecial w ing t ip a ileron. (R esults a dapted from Kuzmina, S.I., Amiryants, G.A., Ishmuratov, F.Z., Mosunov, V.A., and C hedrik, V.V. International Forum on Aeroelasticity and Structural Dynamics, , Madrid, Spain, 2001.)

structural rigidity to i ntegrate the control device and its actuation s ystem. With re spect to  utter s tability, t he forward p osition of the masses increases the utter stability, which is reduced by the aft position of the winglet. Of course, the static aeroelastic effectiveness of a c ontrol surface is a lso important for dy namic applications like  utter suppression or load alleviation for buffeting of vertical tails. Ā is fact is very often forgotten in favor of optimizing the control laws.

7.9.7

Active Structures and Materials Concepts

Dynamic applications for  utter suppression [10] or buffetting load a lleviation [37,38] by means of piezoelectric material were

43722_C007.indd 12

- By increased spanwise moment arm

FIGURE 7.14 wing.

Advantages of an AWTC device on a transport aircraft

demonstrated o n w ind t unnel m odels o r i n f ull s cale g round tests. Ā e involved mass and complexity, mainly for the electric ampliers, precludes practical applications at t he moment. For dynamic app lications h owever, a s emiactive s olution w ith shunted piezomaterials [39] with very little energy demand is an interesting option. Ā e u se o f p iezoelectric m aterials f or s tatic de formations i s limited by the small strain capacity, as well as by the stiffness of the basic structure. Because of these facts, some researchers realized r ather e arly t hat i t i s not adv isable to i ntegrate t he ac tive material directly into the load-carrying skins. In order to achieve large deections, it is necessary to a mplify the active material’s stroke and to uncouple the to-be-deformed (soft) part of the passive structure from the (rigid) main load-carrying part. Because this will usually cause a s evere “strength penalty” for the main structure o f c onventional a irplanes, p ractical app lications a re

10/3/2008 8:45:53 PM

7-13

Aircraft Applications of Smart Structures

limited to u nusual c ongurations l ike sm all Unmanned a erial vehicles (UAVs) or missiles [40]. As an example, Barrett [41] developed such a device, where the external shell of a missile n is twisted by a PZT bender element. Compared to p iezoelectric m aterials, w hich re sponds v ery fast, SMAs are rather slow but they can produce high forces. Ā is precludes applications which require speeds for t he ad aptation which are equal to or higher than the speed of ight or the speed of aircraft control device motions. Two t ypical app lications o f S MAs w ere i nvestigated i n t he DARPA/AFRL/NASA Sm art Wi ng P rogram [ 5,31], a n S MA torque tube to twist the wing of 16% scale wind tunnel model of a generic  ghter aircraft, and SMA wires to ac tuate the hingeless trailing edge control surfaces. Concerning the torque tube, the r atio b etween t he to rque t ube c ross s ection a nd t he w ing torque box cross section should be kept in mind. For the replacement of conventional control surfaces, the efforts to create the deformation on a re alistic structure still need to b e addressed. And, what is even more important than the limitations with respect to the actuation speed, the aeroelastic aspects should be kept in mind from the beginning, in order to evaluate and optimize the effectiveness of such concepts. A s depicted i n Figure 7.15 from Ref. [5], the effectiveness of t he c onformal t railing edge control surface is better than the conventional control surface at low speed, but becomes worse with dynamic pressure. As mentioned in this reference, such concepts are not developed to replace e xisting s ystems, b ut to dem onstrate c apabilities o f active materials. If this is the case, realistic applications still need to be discovered.

1 0.9

Smart materials applications on small remote piloted vehicles (RPVs) a re c urrently being i nvestigated at t he Smart Materials Laboratory of the Portugese Air Force together with the Instituto Superior Técnico in Lisbon [42].

7.9.8

Other Innovative Structure Concepts

Because of the limited stroke of active materials and the inherent stiff ness of a minimum weight aircraft structure, some researchers try to amplify the stroke by sophisticated kinematic systems and enable larger deformations more easily by “articially” reducing the structural stiff ness. So far, all these concepts show the following disadvantages: • High complexity for the actuation system • Higher energ y dem and c ompared w ith t he ac tuation o f conventional control surfaces • Additional internal loads in the structure from the forced deformation • Additional s tructural w eight f rom t he re duced s trength capacity • Reduced s tatic a eroelastic e ffectiveness b ecause o f add itional exibility in the rear wing area • Reduced a eroelastic s tability ( utter) f rom t he re duced stiffness As one example, such a s ystem was described by Monner et al. [43]. Ā is paper summarizes the active structure research activities b y t he G erman a erospace re search e stablishment D LR applied to an Airbus type transport aircraft wing. An old ide a, t he p neumatic a irplane, a s de picted i n F igure 7.16, may be useful, if applied to small UAVs (for storage), or, on larger airplanes to selected structural elements, like spar webs, in order to adjust the shape by a variable pneumatic stiffness to control the aeroelastic load redistribution.

0.8

7.9.9

0.7 Conformal Conventional

0.6 0.5

C Lβ

0.4 0.3 0.2 0.1 0 −0.1 0

100

200

300

400

500

q

FIGURE 7 .15 Comparison of rol ling mome nt e ffectiveness for conventional a nd co nformal t railing ed ge co ntrol s urface. ( From Sanders, B ., F lick, P., a nd S ensburg, O ., i n: P endleton, E . W. E d., International Fo rum o n Ae roelasticity a nd S tructural D ynamics, Madrid, Spain, 2001.)

43722_C007.indd 13

Adaptive All-Movable Aerodynamic Surfaces

Adaptive rot ational at tachment or a ctuation s tiffness for allmovable aerodynamic surfaces can be seen as a s pecial class of active aeroelastic structures concepts. If properly designed, this concept will provide superior effectiveness compared to a rigid structure a lso at lo w s peeds. O ther ac tive a eroelastic c oncepts show their advantages only with increasing speed, in the same way as negative aeroelastic effects are increasing. As an example, a  xed root vertical tail can be made more effective if the structure is tailored in such a way that the elastic axis is located behind the aerodynamic center of pressure. Ā is wash-in eff ect w ill, f or e xample, i ncrease t he l ateral s tability compared to a conventional design on a swept back vertical tail, as depicted in Figure 7.17. Ā e so-called diverging tail [19] has an improved e ffectiveness, b ut a lso e xperiences h igher b ending moments. Instead of tailoring the structure, which essentially will always create a (minimized) weight increase, the tail can be designed as

10/3/2008 8:45:54 PM

7-14

Smart Materials

FIGURE 7.16 Goodyear Inatoplane (1950s). http://upload.wikimedia.org/wikipedia/en/C/C6/Goodyear-inatoplane.jpg (July 6, 2008). Aerodynamic load Active all-movable design: - increased effectiveness Fixed root, diverging tail - decreased bending moment design (passive tailoring): -increased effectiveness -increased bending moment Rigid load Conventional design: -reduced effectiveness -reduced bending moment Span

FIGURE 7 .17 Aerodynamic lo ad d istribution for d ifferent design approaches on a vertical tail.

a reduced size all-movable surface, and the location of the spigot axis is used to tailor the wash-in effect, w hile t he at tachment stiff ness is adjusted to the desired effectiveness. Ā is would allow us to obtain the required effectiveness also at low speeds with a smaller t ail. A s de scribed i n Re f. [44], t he proper s hape of t he surface i n c onjunction w ith t he s pigot a xis lo cation w ill a lso enhance t he  utter s tability. F igure 7 .18 de picts t he re sultant effectiveness for different spigot axis locations at different Mach numbers (and dynamic pressure) with variable stiffness. Ā e crucial element of the all-movable surface with adaptive attachment stiff ness is the attachment or actuation component. Ā is can, for example, be a mechanical spring with variable stiffness a nd a c onventional h ydraulic ac tuator. O r a s a mo re advanced s ystem, a h ybrid ac tuator w ith sm art m aterial ele ments l ike magnetorheological  uids. Ā e objective of t he current DARPA program “Compact Hybrid Actuators” is aimed at the development of such components with high energy density and 10 times the stroke of current systems. Of course, such systems can also be used for horizontal stabilizers or outboard wing sections. Compared to a horizontal tailplane, where t he f uselage  exibility ca uses a eroelastic e ffectiveness

43722_C007.indd 14

losses, a forward surface can exploit additional benets from the fuselage exibility.

7.10 Quality of the Deformations Ā e amount of required internal energy for the desired deformations s trongly de pends o n t he i nvolved s tatic a eroelastic e ffectiveness. As depicted in Figure 7.19, the aerodynamic loads can either deform the structure in the wrong direction and require additional e fforts to compensate the deformations caused by external loads, or the internal, initial deformation is used in such a way, that the desired deformations are only triggered, and the major a mount o f re quired energ y i s supp lied b y t he a ir at no costs. I n t he  rst c ase, t he re quired de formation gener ated b y internal forces will already create a high level of internal strain in the structure, resulting in reinforcements and extra structural weight. In the second case, the required internal actuation forces and the strain levels will be much smaller. For a favorable solution, the design process must aim at reducing the total load level of the structure in the “design case,” thus reducing the required total weight for s tructure a nd ac tuation s ystem c ompared to a conventional design, a s i ndicated i n Figure 7.20, a nd achieve a better performance. With r espect t o spec ic t ype o f de formation, t he re quired effort a nd ac hievable re sults s trongly de pend o n t he t ypical properties of a wing structure, which in most cases can be described a s a b eam. First of a ll, l ift forces will create bending deformations i n t he d irection o f t he l ift f orce. B ecause d rag forces a re much smaller (1/10) a nd because of t he shape of t he airfoil shape, i n-plane bending deformations c an be neglected. Depending on t he chordwise location of t he resulting lift force relative to the beam (torque box) shear center location, it is possible to twist the wing. Ā is can, for example, be used to reduce the b ending de formation. B ecause o f t he h igh s tiff ness of a modern wing in a chordwise section, and because of the resulting aerodynamic pressure distribution, that is usually acting in one direction, a chordwise bending deformation (camber) is very difficult to achieve by internal or external forces. Ā is is also true for a re duced t hickness re ar s ection o f a w ing, f or e xample to replace a deected control surface.

10/3/2008 8:45:54 PM

7-15

Aircraft Applications of Smart Structures

AAAMVT Actuator position and stiffness variation for aeroelastic effectiveness HL at 47% chord

HL at 35% chord

1010

1010

Stiffness K

Ma = 0.2 Ma = 0.6 Ma = 0.9 Ma = 1.2 Ma = 1.5

K

109

108

109

Ma = 0.2 Ma = 0.6 Ma = 0.9 Ma = 1.2 Ma = 1.5

108 0

1 Effectiveness C Y

2

0

1 Effectiveness C Y

2

FIGURE 7.18 Achievable aeroelastic effectiveness with variable attachment stiff ness for different locations of the spigot axis.

Forced deformation and aeroelastic response (Case 1) Force

Required additional energy for initial deformation

Forced deformation and aeroelastic response (Case 2) Force

Aeroelastic deformation Applied force without external load

Aeroelastic deformation

Energy loss Energy for actuation

Deformation without external load

Deformation x

Applied force without external load

Energy for actuation

Additional energy supplied from air

Deformation x Deformation without external load

FIGURE 7.19 Forced structural deformation and aeroelastic response for different design approaches.

7.11

Achievable Amount of Deformation and Effectiveness from Different Active Aeroelastic Concepts

Classical active aeroelastic concepts rely o n the adaptive use of aerodynamic c ontrol su rfaces a nd t heir a eroelastic e ffectiveness

43722_C007.indd 15

at various ight conditions. On conventional designs, aeroelastic effect b ecomes mo re p ronounced w ith i ncreasing a irspeed, a s demonstrated in Figure 7.21 for the potential losses and gains. Conventional active aeroelastic concepts exploit the increasing effectiveness in the upper half of this gure, as well as t he recovering e ffectiveness o f a c onventional ail eron b eyond reversal s peed. A c ombined op eration o f le ading a nd t railing

10/3/2008 8:45:55 PM

7-16

Smart Materials

Concepts performance

Active aeroelastic wing: Blending of leading and trailing edge effectiveness

Achievable deformation (for desired performance) Favourable solution 3

Active static deformation 1 (Good) Aeroelastic deformation 1 Active static deformation 2 (Passive) baseline design

Combined effectiveness

Performance index

Aeroelastic deformation 2

Leading edge

Trailing edge

Required weight (structure + actuation system)

Dynamic pressure

(Bad)

Dynamic pressure

Rigid aircraft

Aileron reversal

FIGURE 7.21 Typical range of aeroelastic effectiveness.

edge surface results in achievable roll rate as indicated in Figure 7.22. In order to better exploit aeroelastic effects in a benecial way, it is required to increase the aeroelastic sensitivity of the design in a wider range of the  ight envelope. Ā is can, for example, be achieved by an a ll-movable aerodynamic surface with adaptive rotational attachment stiffness. Ā is will provide high effectiveness a lso at lo w s peeds, w hile e xcessive lo ads f rom d iverging components o r  utter i nstabilities at h igh s peeds c an b e avoided. Ā e usable aeroelastic effectiveness for conventional concepts is r ather l imited b etween t ake-off and cruise speed. Aileron reversal u sually o ccurs b etween c ruise s peed a nd l imit s peed, and to o h igh e ffectiveness of leading edge surfaces must be avoided at l imit speed. On the other hand, adaptive all-movable concepts c an provide h igh e ffectiveness at a ll s peeds a nd avoid excessive loads at the high end of the speed envelope, as indicated in Figures 7.23 and 7.24. Ā is means that a stabilizer surface can

43722_C007.indd 16

1.0 Rigid aircraft

Dynamic pressure

Effectiveness

1.0

Effectiveness

Performance index

Aeroelastic effectiveness

Achievable roll performance by combined leading and

Active aeroelastic concepts Range of aeroelastic Range of effectiveness effectiveness on for advanced active conventional designs aeroelastic concept

Divergence

FIGURE 7.20 Performance of active structures concepts with respect to weight and performance.

FIGURE 7.22 trailing edge.

Rigid aircraft

Dynamic pressure

FIGURE 7.23 Aeroelastic e ffectiveness for c onventional a nd a daptive-all-movable active aeroelastic concepts.

be built smaller than it would be required from a “rigid” aerodynamic low speed performance requirement.

7.12 Need for the Analysis and Analytical Design Optimization of Active Structures Concepts Of course, active materials and structural components, together with t he stimulating forces, require a c orrect description in t he theoretical structural or multidisciplinary a nalysis a nd opt imization (MDAO) models and methods. Once this is provided, the actively de forming s tructure ne eds a nother app roach f or t he static aeroelastic analysis. For an aircraft with conventional control su rfaces, t he deections of selected control su rfaces c an be predescribed for the aeroelastic analysis. For the actively deformed structure, t he i nitial de formations w ithout e xternal lo ads  rst need to be determined, for example, by a static analysis.

10/3/2008 8:45:55 PM

7-17

Aircraft Applications of Smart Structures

Usable range of aeroelastic effectiveness by flight envelope For conventional active aeroelastic concepts VD

For advanced active aeroelastic concept

Effectiveness

Vmin 1.0

Rigid aircraft

Effectiveness

VC

VC

Dynamic pressure

VD

Vmin Rigid aircraft

Dynamic pressure

FIGURE 7.24 Usable r ange of a eroelastic e ffectiveness for c onventional and advanced active aeroelastic concepts.

As described above, the achievable deformations in conjunction w ith t he d istribution o f e xternal a erodynamic lo ads a re essential f or t he e ffectiveness of a ctive s tructures c oncepts for aircraft control. Ā is requires efficient tools and methods for the simultaneous, m ultidisciplinary a nalytical de sign. To  nd the best design, this task involves the optimization of • x E ternal shape • Arrangement of the passive structure (topology) • Sizing of the passive structure • Placement and sizing of the active elements • Control concept for the active components Ā e a ims of t his approach a re not o nly t he opt imum result for the objective f unction (minimum weight, aerodynamic performance) and ful llment of all constraints like strength), but also optimization of additional objectives, like minimum energy. As depicted in Figure 7.25 for the optimization results of a pa ssive structure w ith diff erent c onstraints f or t he re quired rol ling

Rolling moment effectiveness

Aileron hinge moment [kNm]

50.0 1.0 Baseline static design 10.0 Rigid Structural weight

FIGURE 7.25 Optimization results for the rolling moment and hinge moment of a trailing edge aileron a low aspect ration ghter wing.

43722_C007.indd 17

moment effectiveness, the required energy to actuate the control surface can be considerably reduced, even if the required (low) roll rate is already met. MDO do es not me an to c ombine si ngle d iscipline a nalysis tools b y f ormal c omputing p rocesses. I t me ans  rst a go od understanding of what is going on. Ā is is already essential for a conventional design. Only with this understanding, the creative design of an active concept can start. It is also very important to choose the proper analysis methods for the individual disciplines. Usually, not t he highest level of accuracy is suitable for the simulation of important effects for other d isciplines. Ā is also refers to renement of t he a nalysis models, where local details are usually not interesting for interactions. It is more important to keep the models as versatile as possible for changes of the design concepts and to allow the simulation as many variants as possible. Ā is also means an efficient process for the generation of models, including the knowledge of the user for this process. Fully automated model generators can create terrible results if the user cannot interpret or understand the modeling process. In t he w orld o f a erodynamics, t he de sign o f t he re quired twist a nd c amber d istribution f or a de sired l ift at m inimum drag is also an optimization task. Assuming that minimum drag is achieved by an elliptical lift distribution along the wing span, this task can be solved by a closed formal solution and potential  ow t heory. M ore s ophisticated n umerical me thods are required for the 2D-airfoil design or for Euler and NavierStokes C omputational  uid dy namics ( CFD) me thods, w hich are now maturing for practical use in aircraft design. For the conceptual aircraft design, formal optimization methods have been used for many years. Here, quantities like direct operating costs (DOC) can be expressed by rather simple equations a nd t he s tructural weight c an b e der ived f rom empirical data. F ormal me thods l ike opt imum c ontrol t heory a re a lso available for the design of the ight control system. So one might t hink t hat t hese individual opt imization tasks could easily be coupled for one global aircraft optimization process. Ā e reasons why t his task is not s o si mple is t he d ifferent nature of the design variables in individual disciplines, and their cross sensitivities w ith ot her disciplines. Ā e expression multidisciplinary o ptimization ( MDO) s ummarizes a ll a ctivities in this area, which have been intensied in recent years. It must be admitted that today most existing tools and methods in this area are s till si ngle d iscipline opt imization t asks w ith m ultidisciplinary constraints. In order to design and analyze active aeroelastic aircraft concepts, especially when they are based on active materials or other active s tructural memb ers, ne w qu antities a re re quired to describe t heir i nteraction w ith t he s tructure, t he  ight control system, and the resulting aeroelastic effects.

7.13

Summary and Conclusions

In the same way as it was wrong in the past to dem and that an aircraft design be as rigid as possible, it is wrong now to demand

10/3/2008 8:45:56 PM

7-18

a design that is as exible as possible. It is sometimes said that smart structures concepts could completely replace conventional control su rfaces. But t his lo oks v ery u nrealistic, at le ast at t he moment. Ā e major d ifficulties for a suc cessful application a re that the limited deformation capacity of active materials as well as t heir s train a llowables i s u sually b elow t hose o f t he pa ssive structure. However, this can be resolved by a proper design of t he i nterface b etween pa ssive a nd ac tive s tructure. But t he essential difficulties are the stiffness and strain limitations of the passive structure itself. It cannot be expected that the material of the passive structure just needs to b e replaced by more  exible materials w ithout a n e xcessive w eight p enalty. I t i s a lso not correct to believe that an active aeroelastic concept will become more effective if the exibility of the structure is increased. Ā e aeroelastic e ffectiveness de pends o n p roper a eroelastic de sign, which ne eds a c ertain r igidity o f t he s tructure to p roduce t he desired loads. A very exible structure would also not be desirable from the standpoint of aerodynamic shape, stability of the ight control system, and transmission of static loads. Because large control surface deections a re required at lo w speeds, where aeroelastic effects on a xed surface are small, it is more realistic to u se conventional control surfaces for this part of t he  ight envelope a nd make use of active aeroelastic deformation only at higher speeds. Ā is would still save weight on the control surfaces a nd t heir actuation system due to t he reduced loads and actuation power requirements. To produce u sable deformations of t he structure a lso at lo w speeds, all-movable aerodynamic surfaces with a variable attachment stiffness a re a n i nteresting opt ion. Ā is concept rel ies on developmental e fforts f or ac tive de vices w ith a w ide r ange o f adjustable stiffness. Ā e reasons why we have not seen more progress to date for the suc cessful demo nstration o f sm art s tructures c oncepts i n aeronautics may be because • Specialists i n a ircraft design do not know enough about the ac hievements i n t he a rea o f sm art m aterials a nd structures. • Smart materials and actuation system specialists who try to nd and demonstrate applications in aeronautics do not know or care enough about real world conditions for airplane structures. What we need is more awareness on both sides as well as stronger efforts to learn from each other and work together. Although there are strong doubts about useful applications of smart structures for aircraft control, it should always be remembered how often leading experts had been wrong in the past with their predictions i n m any c ases e ven on t heir own i nventions. Norman R. Augustine quotes some of them in his famous book Augustine’s Laws [45]: • “Ā e [ying] machines will eventually be fast; they will be used in sport but they should not be thought of as commercial c arriers.”—Octave Cha nute, A viation Pi oneer, 1910.

43722_C007.indd 18

Smart Materials

• “Ā e energy produced by the breaking down of the atom is a very poor kind of thing. Anyone who expects a source of power f rom t he t ransformation of t hese atoms is talking moonshine.”—Ernest Rutherford, Physicist, ca. 1910. • “Fooling around with alternative currents is just a waste of time. Nobody will use it, ever. It’s too dangerous… it could kill a man as quick as a bolt of lightning. Direct current is safe.”—Ā omas Edison, Inventor, ca. 1880. Also quoted by Augustine [45], the eminent scientist Niels Bohr remarked: “P rediction i s v ery d ifficult, e specially a bout t he future.” At t he mo ment, i t lo oks mo re re alistic a nd f easible t hat new hybrid, c oncentrated ac tive de vices, p ositioned b etween a passive but properly aeroelastically tailored main aerodynamic surface and the corresponding control surfaces is the way of the future.

Appendix A: Future Directions Whereas an aerodynamic drag was always a design driver in the development of a irplanes, it h as now become even more i mportant. M inimum w eight de signs w ere a lways re quired a nd t he amount of f uel t hat needs to b e carried has d ictated t he eng ine development, but now the price of fuel has also become an important additional design driver. Ā e fuel bill share in the operational costs of an airplane has risen from below 10% to above 25% and soon-to-come CO2 surcharges may double this value. Ā ere are strong development efforts to improve the fuel efficiency of airplanes. Besides very successful efforts in the materials and structural design and manufacturing areas, more efficient engines, a nd  y-by-wire ight c ontrol s ystems, A erodynamic designers were t rying to i mprove a irplane efficiency by adding so-called w inglets to t he w ingtips. St udies a nd R 7D programs with winglets on wing tips in large-scale NASA test programs have en ded up i n u nsuccessful app lications. Re asons w ere t he addition lo ads t hat t hese de vices add to t he w ing s tructure, which r equires m ore s trength ca pacity, as w ell as a eroelastic stability i ssues, b ecause t he add itional m ass a nd a erodynamic forces of these surfaces at unfavorable locations, which requires more s tiff ness o r add itional ba lance m asses i n t he o rder o f several hundred pounds. An in teresting d evelopment h appened in t he mi d-1990s. A small company, Aviation Partners, started to develop winglets in order to i mprove t he r ange of business je ts. Ā is a lso i ncluded the app lication to t he B oeing Bu siness J et, a der ivative o f t he highly successful 737-family. When r ather lo w f uel p rices s tarted s oaring a gain i n 1 999, they had a p roduct ready for applications in commercial transport 737s. Ā is is the rst known case, where a winglet as retrot equipment added essential aerodynamic performance improvement i n t he o rder o f 3 %–4% d rag re duction to a c ommercial transport airplane, which is usually t he magnitude t hat can be expected from a completely new aircraft design. More research is needed o r i s a lready o ngoing f or f uture sm art m aterials a nd

10/3/2008 8:45:56 PM

7-19

Aircraft Applications of Smart Structures

structures applications, because t he add ition of w ingtips to a n existing wing or increasing the span of a new design has severe implications on the static and dynamic structural loads as well as on the aeroelastic (utter) stability of the vehicle. Whereas aircraft a lways h ad to b e l ight b ut a lso c ost-efficient in recent decades, much higher development and manufacturing costs for more sophisticated design are justied today.

A.7.1

Aerodynamic Drag and Structural Design Issues

Minimizing the aerodynamic drag is a classical design problem in aviation. Ā e essential drag components are • Term that is independent of the lift forces, which is essentially dened by friction forces and thus from the overall shape and surface area of the vehicle as well as the surface roughness. • i L ft-dependent, so-called (lift-)induced drag term, which is proportional to the squares mass of the vehicle, divided by the squared span, and a correction term for the distribution of the aerodynamic forces along the span. Ā is last term says that an elliptic shape of the lift forces w ill provide m inimum d rag. Ā is shape on t he ot her hand determines the required structural weight to carry the loads. A mass increase f or h igher lo ads w ill t herefore h ave a s evere i mpact on the drag. But t his assumption for the minimum drag does not ye t t ake i nto ac count t hat a h igher s pan w ith a d ifferent load distribution might provide a superior solution, a fact that was a lready recognized by Ludw ig Prandtl i n t he 1920 [46]. If the span is not  xed, an aerodynamic load distribution with higher lo ads i n t he i nboard s ection a nd re duced lo ading toward the tip will give a better overall design. Aircraft designers with only aerodynamics in mind often do not consider this fact. Prandtl’s assumption and similar, modied versions of it like the Refs. [47,48] are however based on simplied assumptions f or t he w ing s tructure’s lo ad c apacity a nd i ts m ass (a simple b eam mo del). Today a nd i n t he f uture, adv anced analytical de sign a nd opt imization me thods a nd to ols w ill allow much better approaches for this kind of optimization, although t he l inks b etween t he s tructural a nd a erodynamic worlds c ould a nd s hould b e m uch b etter a nd s tronger t han they are today. And, even more important, all these assumptions to date consider the same load distribution in cruise  ight and at extreme, “once-in-a-lifetime” ultimate design loads. Although aeroelastic load redistributions are considered in the determination of aerodynamic conditions and in the assessment of structural design loads, they are not ye t present and are not e xploited in an integrated design process. A g reat deal of effort has been spent developing t he aerodynamic design of w inglets a nd ot her w ingtip devices, a iming at the reduction of the energy of the wing tip vortices that originate at t he t ips d ue to t he p ressure d ifference b etween upp er a nd

43722_C007.indd 19

lower su rface. B ecause a w ing w ith i nnite span produces no lift-induced drag, multiplanes with many slender wings or ring wings h ave a lways b een p resent i n a erodynamic re search. Winglets are considered as a design option, where the horizontal span is xed and the vertical extension helps to reduce the energy of the tip vortices. In this approach, it is also often assumed that the c onstant h orizontal s pan add s le ss add itional lo ads to t he structure than a horizontal extension.

A.7.2

New Structural Research Efforts and Achievements

Ā e DARPA M orphing A ircraft P rogram w as i nitially m ainly aiming a t c reating ( military)  ight ve hicles for m ultipurpose missions b y r adical s hape c hanges [ 49]. Today, a s s ome w ind tunnel demonstrators have already shown and successfully been tested, some of the newly developed active and passive materials and structures concept for this initial purpose will now certainly become very valuable to help reduce the induced drag and structural lo ads o n ne w de signs f or c ommercial t ransport a ircraft wings. Another si gnicant European research program Active Aeroelastic A ircraft St ructure ( 3AS) de veloped i n 2 005 w as mainly aimed at civil applications [50–53]. Some new active and passive structural design concepts, mainly based on the exploitation of benecial aeroelastic effects, were developed there and successfully demonstrated on wing tunnel models. Whereas aeroelastic tailoring by means of sweeping the orientations o f h igh el astic mo dulus c arbon  bers a nd c omposing their l ay-ups w as e ssentially me ant f or app lications o n t hin, plate-like wings structures for ghter airplanes to improve their already high maneuverability at extreme ight conditions, some designers hope t hat such features can a lso be applied to r ather thick, slender transport aircraft wings to achieve favorable structural a nd a erodynamic lo ading c onditions. A lthough c arbon ber composites, which have been successfully used for a lmost all c ommercial t ransport a irplane s tructures, i ncluding t he fuselage, it should be mentioned here that the design space and design conditions are different from a ghter. Ā e wing geometric properties a s w ell a s t he lo wer dy namic p ressure w ill s how smaller a eroelastic t ailoring e ffects in this case. Today it looks like aerodynamic drag, f uel price, and CO2 issues have already created a vast new area for future smart materials and structures applications.

A.7.3

Example for the Interaction of Structural, Aerodynamic, and Aeroelastic Constraints for Different Wing Tip Design Concepts

An e xample w as c hosen to de pict t he p otential b enets o f a proper, m ultidisciplinary de sign app roach. D ifferent w ing t ip

10/3/2008 8:45:56 PM

7-20

Smart Materials

Wing bending moment

Wing sectional load distribution

20.00

3.00

18.00 2.50 Bending moment

16.00

Section load

2.00 1.50 1.00

14.00 12.00 10.00 8.00 6.00 4.00

0.50

2.00

0.00 0.00

1.00

2.00

3.00 4.00 Span station

Basic wing Vertical winglet

5.00

0.00 0.00

6.00

design concepts and design approaches were compared to an overall optimum design with respect to minimum drag and fuel consumption. Figure A.7.1 depicts the load distribution along the wing span for a ba seline design, t he wing with a v ertical winglet added, a horizontal span extension, and a tip with a built-in load alleviation

Wing shear force distribution 8.00 7.00 6.00 5.00 4.00

2.00

Basic wing Vertical winglet

Horizontal extension Tip with load alleviation

FIGURE A .7.1 Wing s pan lo ad d istributions for t he fou r d ifferent design concepts.

1.00

3.00 Span

4.00

5.00

6.00

Horizontal tip extension Tip with load alleviation

FIGURE A .7.3 Bending mome nt d istributions a long w ing s pan for the four different design concepts.

concept. Ā ey s hall a ll h ave t he s ame lo ad d istribution o n t he inboard s ections. Ā e re sulting s hear f orce d istributions a re plotted i n F igure A .7.2, w hich s how h igher le vels f or a ll t hree modications. Ā e advantage of t he de sign w ith load a lleviation on t he t ip (reduction o f t he t ip lo ad b y 5 0%) b ecomes obv ious i n F igure A.7.3, where the bending moment distributions are plotted. Ā e reduction of t he small t ip load re sults i n a c onsiderable reduction o f t he w ing ro ot b ending mo ment c ompared to t he ot her two tip modications. Rather than trying to modify the deformation characteristics of the complete wing, for example by means of aeroelastic tailoring, t he c omposite c over s kin t hickness d istributions a nd  ber orientations, it i s s ufficient i n t his c ase to g ive a “ clever t wist” under lo ad to t he sm all t ip s ection, f or e xample b y me ans o f a smart (passive or active design) tip attachment concept, which can be considered as a local, novel aeroelastic tailoring approach.

3.00

References

2.00

0.00 0.00

1.00

2.00

Basic wing Vertical winglet

3.00 Span

4.00

5.00

6.00

Horizontal extension Tip with load alleviation

FIGURE A.7.2 Shear force distributions along wing span for the four different design concepts.

43722_C007.indd 20

Wright, O.: How We I nvented t he Airplane. D over Publications, Mineola, NY, 1988. 2. Weisshaar, T . A.: Aeroelastic T ailoring—Creative U se o f Unusual Materials. AIAA-87-0976-CP. 3. Schwipps, W.: Schwerer als LuĀ—die Frühzeit der Flugtechnik in Deutschland. Bernhard & Graefe Verlag, Koblenz, Germany, 1984. 4. Woods, M. I., H enderson, J . F., a nd L ock, G. D .: Ener gy requirements f or the  ight o f micr o a ir v ehicles. The Aeronautical Journal, March 2001. 1.

1.00

10/3/2008 8:45:57 PM

Aircraft Applications of Smart Structures

5. Sanders, B., Flick, P., and S ensburg, O.: C ontrollable aeroe lastic lifting systems. In: Pendleton, E. W. Ed., International Forum o n Aeroelasticity a nd S tructural D ynamics 2001, Madrid, Spain. 6. Bertin, J. J. a nd S mith, M. L.: Aerodynamics for E ngineers. Prentice-Hall, Englewood Cliffs, NJ, 1979. 7. Sensburg, O., Hönlinger, H., Noll, T. E., a nd Huttsell, L. J.: Active  utter su ppression o n a n F -4F a ircraft. Journal o f AircraĀ, 19(5), 1981. 8. Larson, C. R ., Falanges, E., a nd D obbs, S. K.: P iezoceramic active vibration suppression control system development for the B-1B aircraft. SPIE’s 5th Annual International Symposium Smart Structures and Materials. San Diego, CA, 1998. 9. Crowe, C. R . a nd Sa ter, J . M.: S mart a ircraft structures. AGARD S ymposium o n F uture Aerospace Technology in the Services of the Alliance. CP-600, Vol. I, Paris, 1997. 10. McGowan, A. R., Heeg, J., and Lake, R. C.: Results of windtunnel t esting f rom the p iezoelectric aer oelastic r esponse tailoring in vestigation. P roceedings o f the 37th AIAA/ ASME/ASCE/AHS/ASC S tructures, S tructural D ynamics and Materials Conference, Salt Lake City, UT, 1996. 11. Skopik, O. L.: Wie Berechnet und Konstruiert Man Selbst Ein Flugzeug. Vienna, 1917. 12. Christensen, C. M.: The I nnovator’s D ilemma. H arvard Business School Press, Boston, MA, 1997. 13. Flomenhoft, H. I.: The R evolution in Str uctural D ynamics. Dynaow Press, Palm Beach Gardens, FL, 1997. 14. Collar, A. R.: Ā e expanding domain of aeroelasticity. Journal of the Royal Aeronautical Society, 613–636, August 1946. 15. Schweiger, J., Simpson, J., Weiss, F., Coetzee, E., and B oller, Ch.: Needs for the analysis and integrated design optimization o f a ctive a nd pa ssive s tructure f or a ctive a eroelastic wings. S PIE’s 6th Annual I nternational S ymposium o n Smart Structures and Materials. Newport Beach, CA, 1999. 16. Weisshaar, T. A.: Air v ehicle desig n—new ho rizons? R TO Meeting P roceedings 36 (A C/323(AVT)TP/17) f rom the RTO AVT S pecialists’ M eeting o n S tructural Aspects o f Flexible Aircraft Control. Ottawa, Canada, October 1999. 17. Pendleton, E. W., Bessette, D., Field, P. B., Miller, G. D., and Griffin, K. E.: Active aeroelastic wing ight research program: Technical p rogram a nd mo del a nalytical de velopment. Journal of AircraĀ 37(4), July–August 2000. 18. Journal o f AircraĀ, 32(1), S pecial S ection: Active Flexib le Wing, 1995. 19. Sensburg, O., S chneider, G ., T ischler, V., a nd Venkayya, V.: A uniq ue desig n f or a div erging  exible v ertical t ail. Specialists’ M eeting o n S tructural Aspects o f Flexible Aircraft Control. RTA Meeting on D esign Issues. Ottawa, Canada; October 1999. 20. Fokker, A. H. G. and Gould, B.: Flying Dutchman. The Life of Anthony Fokker. Henry Holt and Company, New York, year unknown. 21. Noll, T ., H önlinger, H., S ensburg, O ., a nd S chmidt, K.: Active utter suppression design and test, a Joint U.S.–F.R.G. Program. ICAS- CP 80-5.5. Munich, Germany, 1980.

43722_C007.indd 21

7-21

22. Lazarus, K. B ., Saa rmaa, E., a nd Agnes, G. S.: An ac tive smart material system for buffet lo ad a lleviation. SPE Vol. 22447/179, Bellingham, WA, 1995. 23. Hauch, R . M., J acobs, J . H., R avindra, K., a nd Dima, C.: Reduction o f v ertical t ail b uffet r esponse usin g ac tive control. AIAA-CP-95-1080, Washington, 1995. 24. Moses, R . W.: C ontributions to ac tive buffeting alleviation programs by the NASA Langley Research Center. CP, 40th AIAA-SDM conference, St. Louis, MO, 1999. 25. Simpson, J. and Schweiger, J.: Industrial approach to piezoelectric damping of large ghter aircraft components. SPIE’s 5th Annual International Symposium Smart Structures and Materials. San Diego, CA, 1998. 26. Weisshaar, T. A.: Di vergence o f f orward sw ept co mposite wings. Journal of AircraĀ, 17(6), June 1980. 27. Schweiger, J., Krammer, J., and Hörnlein, H.: Development and a pplication o f t he in tegrated st ructural desig n to ol LAGRANGE. 6th AIAA/NASA/ISSMO S ymposium o n Multidisciplinary Analysis a nd Op timization. CP -4169, Seattle, WA, 1996. 28. Khot, N., Eastep, F., and Kolonay, R.: A method for enhancement of the rolling maneuver of a exible wing. AIAA-CP96-1361. 29. Griffin, K. and Hopkins, M.: Aeroelastic tailoring for smart structures. 36th AIAA Structures, Structural Dynamics, and Materials Conference. New Orleans, 1995. 30. Giese, C. L., Reich, G. W., Hopkins, M. A., and Griffin, K. E.: An investigation of the aeroelastic tailoring for smart structures concept. AIAA-96-1575-CP, 1996. 31. McGowan, A. R ., H orta, L. G., H arrison, J . S., a nd R aney, D. L.: Research activities within NASA’s Morphing Program. RTO Meeting Proceedings 36 (A C/323(AVT)TP/17) f rom the RTO AVT Specialists’ Meeting on Structural Aspects of Flexible Aircraft Control. Ottawa, Canada, October 1999. 32. Padula, S. L., Rogers, J. L., and Raney, D. L.: Multidisciplinary techniques a nd no vel a ircraft co ntrol syst ems. 8th AIAA/ NASA/ISSMO S ymposium o n M ultidisciplinary Analysis and Op timization. AIAA-CP-2000-4848, L ong B each, CA, 2000. 33. Pendleton, E., G riffin, K. E., K ehoe, M. W., a nd P erry, B .: A Flig ht Res earch P rogram f or Active Aeroelastic Wing Technology. Conference Proceedings AIAA-96-1574-CP. 34. Flick, P. M. and Love, M. H.: Ā e impact of active aeroelastic wing te chnology on conceptual aircraft design. Specialists’ meeting o n st ructural asp ects o f  exible aircraft control. RTA Meeting o n D esign I ssues. Ot tawa, C anada, O ctober 1999. 35. Kuzmina, S. I., Amiryants, G. A., Ishmuratov, F. Z., Mosunov, V. A., and Chedrik, V. V.: Some applications of active aeroelasticity concepts to aircraft design. International Forum on Aeroelasticity a nd S tructural D ynamics 2001, M adrid, Spain. 36. Schweiger, J. a nd S ensburg, O.: A cri tical r eview o f eff orts and achievements to improve aircraft performance or efficiency b y ac tive str uctures co ncepts. CP , I nternational

10/3/2008 8:45:57 PM

7-22

Forum o n Aeroelasticity a nd S tructural D ynamics 2001, Madrid, Spain. 37. Moses, R . W.: C ontributions to ac tive buffeting alleviation programs by the NASA Langley Research Center. CP, 40th AIAA-SDM Conference, St. Louis, MO, 1999. 38. Simpson, J. and Schweiger, J.: Industrial approach to piezoelectric damping of large ghter aircraft components. SPIE’s 6th Annual International Symposium Smart Structures and Materials. San Diego, CA, 1998. 39. McGowan, A. R.: A feasibility study of using shunted piezoelectrics to reduce aeroelastic response. SPIE’s 5th Annual International S ymposium S mart S tructures a nd M aterials. Newport Beach, CA, 1999. 40. W eisshaar, T . A.: Aeroelasticity’s r ole in inno vative U AV design and optimization—the way things ought to b e. 41st Annual Israel C onference on Aerospace Sciences, Tel Aviv and Haifa, February 2001. 41. Barrett, R.: Active aeroelastic tailoring of an adaptive exspar s tabilator. Smart M aterials a nd S tructures N o. 5, I OP Publishing Ltd., Bristol, UK, 1996. 42. Costa, P., Moniz, P. A., and Suleman, A.: Design and testing of a n adaptive RPV aer oelastic demo nstrator. 42nd AIAA SDM Conference, Seattle, WA, 2001, AIAA-2001-1361. 43. Monner, H. P., B reitbach, E., B ein, Ā ., a nd H anselka, H.: Design asp ects o f the ada ptive win g—the elastic tra iling edge and the lo cal spoiler bump. The Aeronautical Journal, 104(1032), 2000 44. Schweiger, J., Weiss, F., a nd Kullrich, T.: Active aer oelastic design o f a v ertical t ail f or a  ghter aircraft. 8th AIAA/ USAF/NASA/ISSMO S ymposium o n M ultidisciplinary Analysis a nd Op timization, L ong B each, CA, S eptember 2000.

43722_C007.indd 22

Smart Materials

45. Augustine, N. R.: Augustine’s Laws. AIAA, Reston, VA, 1997. 46. Prandtl, L.: I nduced dra g o f m ultiplanes, NACA TN 182, 1924. 47. Klein, G. a nd Viswanathan, S.: ZeitschriĀ für Angewandte Mathematik und Physik, 24(6), November, 1973. 48. Löbert, G.: S panwise lift d istribution of for ward-and aft-swept wings in comparison to the optimum distribution form, AIAA-81-4187, Journal of AircraĀ, 18(6), June 1981 49. McGowan, A. M.: Morphing technologies for future aircraft. Online transcript of speech at the Air & Space Conference and Technology E xposition. Washington, D.C., S eptember 26, 2006.Ci ted a t the Air F orce Association URL: h ttp:// www.afa.org/media/scripts/conf2006_McGowan.asp [ci ted 31 March 2007]. 50. Johannes, S., Afzal, S., S vetlana, K., a nd Vasilij, C.: MD O concepts for a European research project on active aeroelastic str uctures. 9th AIAA/NASA/ISSMO S ymposium o n Multidisciplinary Analysis a nd Op timization. CP, Atlanta, GA, 2002. 51. Johannes, S. and Svetlana, K.: Successful Russian–European cooperation in the ac tive aeroelastic aircraft research. ILA Conference, Berlin, Germany, 2004. 52. Moulin, B., Feldgun, V., Karpel, M., Anguita, L., Rosich, F., and C liment, H.: Alleviation of dynamic gust lo ads under special co ntrol surfaces. CP , I nternational F orum o n Aeroelasticity a nd S tructural D ynamics 2005, M unich, Germany. 53. Amiryants, G. A., M ullov, Yu. M., S halaev, S. V., a nd Zichenkov, M. Ch.: D esign, manufacture and wind t unnel testing of the multi-functional European research aeroelastic model (EuRAM). CP, European Conference for Aerospace Sciences, Moscow, 2004.

10/3/2008 8:45:57 PM

8 Smart Battery Materials 8.1 I 8.2 8.3 8.4

Arumugam Manthiram The University of Texas at Austin

8.1

ntroduction ...............................................................................................................................8-1 Electrochemical Concepts Involved in a Battery ...................................................................8-1 Types of Batteries .......................................................................................................................8-2 Lithium Ion Batteries.................................................................................................................8-2 Layered Oxide Cathodes • Spinel Oxide Cathodes • Olivine Oxide Cathodes • Ca rbon Anodes

8.5 C onclusions ................................................................................................................................8-7 Acknowledgment .................................................................................................................................8-7 References .............................................................................................................................................8-7

X

Introduction

Batteries a re t he m ajor p ower s ources f or p ortable ele ctronic devices a s w ell a s f or a utomobile s tarting a nd ig nition. Ā e worldwide bat tery m arket i s c urrently w orth o ver $ 40 b illion. Batteries a re a lso i ntensively p ursued f or a w idespread h ybrid electric vehicle (HEV) and plug-in hybrid electric vehicle (PHEV) applications to add ress t he g rowing en vironmental c oncerns and increasing global energ y demands. Moreover, bat teries are critical a s energ y s torage de vices f or e ffectively ut ilizing s olar and wind energies, which are intermittent in nature. Ā e performance factors of batteries are linked to the materials used as well as t he eng ineering i nvolved i n f abricating t hem. Ā is chapter, after providing a b rief background to t he electrochemical concepts i nvolved i n bat teries, focuses on t he m aterials a spects of high energy density (lithium ion) batteries.

8.2 Electrochemical Concepts Involved in a Battery A battery is an electrochemical device, which converts stored chemical energy directly into electrical energy [1–5]. It consists of an anode, a c athode, and an e lectrolyte (Figure 8 .1). During the battery operation, while the anode (A) undergoes an oxidation reaction as shown in Equation 8.1, the cathode (X) undergoes a re duction reaction as shown in Equation 8.2 by ac cepting t he ele ctrons rele ased at t he a node t hrough t he external circuit: A

→ An+ + ne− (8

.1)

+ ne− → X n− (8

.2)

During t his p rocess, t he ele ctrolyte ac ts a s a me dium f or t he transfer of charge between the anode and cathode in the form of ions inside the cell, resulting in the formation of AX as the reaction product as shown in Equation 8.3: A + X → AX

(8.3)

Ā e free energy change involved in this chemical reaction is tapped out as useful electrical energy. Ā e battery performance is characterized by several electrochemical performance factors, which a re de termined b y b oth t he i ntrinsic p roperties o f t he anode, c athode, a nd ele ctrolyte m aterials u sed a s w ell a s t he engineering i nvolved i n de signing a nd f abricating t he bat tery. Some o f t he p erformance pa rameters a re b riey discussed below. Ā e capacity Q of an electrode (anode or cathode) material is given as the product of the current I owing through the external circuit and the time t, which in turn is related to the number of electrons n involved in the reaction as shown by Q = It = nFNm (8

.4)

where F an d Nm a re, re spectively, t he F araday c onstant (96,487 C/mol) and number of moles of the anode A or cathode X. Specic capacity is expressed as Ah/kg, which is the capacity delivered b y 1 kg o f t he ele ctrode m aterial. F or e xample, o ne gram-equivalent (atomic or molecular weight in grams divided by t he n umber o f ele ctrons n in volved in t he r eaction) o f a n 8-1

43722_C008.indd 1

10/3/2008 8:46:23 PM

8-2

Smart Materials

e−

e−

Cathode X

Anode A

−

+

Separator

Electrolyte

Electrolyte

FIGURE 8 .1 Operating pr inciples of a b attery, w hich i s a n e lectrochemical cell.

electrode material will deliver a capacity of 96,487 C or 26.8 A h. Ā e cell voltage is the difference between the electrode potentials of the anode and cathode materials. Specic energy is a product of the specic capacity and cell voltage, which is expressed in W h/kg. Specic p ower i s t he p roduct o f t he c ell vol tage a nd t he current  owing, w hich i s e xpressed i n W /kg. Ā e capacity, energy, a nd p ower c an a lso b e s pecied p er u nit vol ume. F or example, energ y den sity a nd p ower den sity a re e xpressed, respectively, in W h/L and W/L.

8.3

Types of Batteries

Batteries can be classied into primary and secondary batteries. P rimary bat teries a re no nrechargeable a s t he c hemical

reactions i nvolved i n t he ele ctrode m aterials a re i rreversible, while s econdary bat teries a re re chargeable a s t he c hemical reactions i nvolved a re re versible. S ome e xamples o f p rimary and secondary battery systems are given in Tables 8.1 and 8.2 [2]. Ā e tables give the cell reactions, cell voltage, and cell capacity f or e ach s ystem. W hile mo st o f t he p rimary a nd secondary bat tery s ystems i n Tables 8 .1 a nd 8 .2 a re ba sed on aqueous ele ctrolytes, t he p rimary bat tery s ystems em ploying lithium a node i n T able 8 .1 a nd t he s econdary l ithium io n battery system i n Table 8.2 a re based on nonaqueous electrolytes. Ā e cell voltages of the battery systems based on aqueous electrolytes are limited to ≤2.1 V due to a narrower thermodynamic, e lectrochemical s tability w indow of w ater. I n ot her words, it is due to a sm aller energ y separation E g between the highest occupied molecular orbital (HOMO) and lowest unoccupied mole cular o rbital ( LUMO) o f w ater a nd a c onsequent susceptibility of water to unwanted reduction/oxidation (redox) reactions at h igher c ell vol tages. I n c ontrast, a w ider ele ctrochemical stability window offered by the nonaqueous electrolytes in t he c ases of b oth primary l ithium bat teries a nd s econdary lithium ion batteries allows in principle the realization of cell voltages of as high as 5 V.

8.4

Lithium Ion Batteries

Since the use of nonaqueous electrolytes allows the realization of high c ell vol tages o f a round 4 V i n p ractical c ells, l ithium io n cells o ffer much higher volumetric (W h/L) and gravimetric (W h/kg) energ y densities compared to ot her rechargeable systems a s s een i n F igure 8 .2 [ 6]. A s a re sult, t hey h ave b ecome attractive power sources for portable electronic devices such as laptop computers and cell phones, revolutionizing the electronics industry. Although the concept of rechargeable lithium bat-

TABLE 8.1 Major Primary Battery Systems and Ā e ir Characteristics Battery

Anode

Cathode

Cell Reaction

Cell Voltage (V) Capacity (mA h/g)a

Leclanche

Zn

MnO2

Zn + 2MnO2 → ZnO . Mn2O3

1.6

224

Magnesium

Mg

MnO2

Mg + 2MnO2 + H2O → Mn2O3 + Mg(OH)2

2.8

271

Alkaline MnO2

Zn

MnO2

Zn + 2MnO2 → ZnO + Mn2O3

1.5

224

Mercury

Zn

HgO

Zn + HgO → ZnO + Hg

1.34

190

Zinc–air

Zn

O2

Zn + 0.5O2 → ZnO

1.65

658

Li–SO2

Li

SO2

2Li + 2SO2 → Li2S2O4

3.1

379

Li–MnO2

Li

MnO2

Li + MnO2 → LiMnO2

3.1

286

a

Based on active electrode materials (cathode and anode) only.

TABLE 8.2 Major Secondary Battery Systems and Ā e ir Characteristics Battery

Cathode

Cell Reaction

Cell Voltage (V)

Capacity (mA h/g)a

Lead–acid

Pb

PbO2

Pb + PbO2 + 2H2SO4 → 2PbSO4 + 2H2O

2.1

120

Nickel–cadmium

Cd

NiOOH

Cd + 2NiOOH + 2H2O → 2Ni(OH)2 + Cd(OH)2

1.35

181

Nickel–hydrogen

H2

NiOOH

H2 + 2NiOOH → 2Ni(OH)2

1.5

289

Nickel-metal hydride

MH

NiOOH

MH + NiOOH → M + Ni(OH)2

1.35

206

Lithium ion

Li

Li0.5CoO2

0.5Li + Li0.5CoO2 → LiCoO2

3.7

137

a

43722_C008.indd 2

Anode

Based on active electrode materials (cathode and anode) only.

10/3/2008 8:46:23 PM

8-3

Smart Battery Materials

Smaller 200

•

Lighter

Gravimetric energy density (W h/kg)

250

150

Lithium-ion

100

•

50 Ni/Cd

Ni/MH

Lead-acid

0 0

100

200

300

400

•

Volumetric energy density (W h/L)

FIGURE 8 .2 Comparison of t he e nergy d ensities ( gravimetric a nd volumetric) of various rechargeable battery systems.

• •

teries was initially demonstrated with a su lde cathode such as TiS2 and a metallic lithium anode about three decades ago [7–9], the dendritic short-circuiting and the safety concerns associated with the metallic lithium anode remained as an impediment for several years to commercialize the rechargeable lithium battery technology. It is because of the design and development of smart materials as cathodes and anodes [7–17], the present-day lithium ion battery technology has become a commercial reality, following the rst announcement by Sony in 1991. Figure 8.3 shows the operating principles involved in a lithium ion cell. It involves the reversible extraction and insertion of lithium ions from or into two lithium insertion hosts during the cha rge–discharge p rocess. F or ex ample, i n co mmercial lithium ion cells, the lithium ions are extracted from the layered L iCoO2 c athode a nd i nserted i nto t he l ayered g raphite during c harge, w hile t he ele ctrons  ow f rom the c athode to the a node t hrough t he e xternal c ircuit. D uring d ischarge, exactly t he opp osite h appens. Si nce t he p rocess i nvolves a shuttling or rocking of lithium ions between the cathode and anode during the charge–discharge pr ocess, t he c ell i s a lso some t imes r eferred t o a s r ocking chair cell. Although the concept looks simple, the anode and cathode materials should satisfy several criteria outlined below in order for them to b e successful: • i D fference in the lithium chemical potential between the cathode and anode should be large to maximize the cell voltage. Ā e c ell vol tage i s de termined b y t he energ ies involved in both the electron transfer and the Li+ transfer. W hile t he en ergy i nvolved i n el ectron t ransfer i s related to t he work functions of the cathode and anode, the energy involved in Li+ transfer is determined by the crystal s tructure a nd t he c oordination ge ometry o f t he site i nto/from w hich L i+ ions are inserted or extracted

43722_C008.indd 3

[18]. I f we c onsider only the e lectron t ransfer, then the Mn+ ion in the insertion compound Li xMy Oz should have a high oxidation state in order for it to be employed as a cathode a nd a lo w o xidation s tate i n o rder f or i t to b e employed as an anode. Insertion compound LixMyOz should allow an insertion or extraction of a large amount of lithium x to maximize the cell c apacity. Ā is de pends o n t he n umber o f a vailable lithium ion sites and the accessibility of multiple valences for M in the insertion host. Insertion c ompound L i xMyOz should support both high electronic a nd l ithium io n c onductivities to off er high power density. Lithium insertion and extraction process should be reversible w ith no o r m inimal s tructural c hanges i n t he h ost structure over t he entire r ange x o f l ithium i nsertion o r extraction in order to provide good cycle life. Insertion compound Li xMyOz should be chemically stable without undergoing any reaction with the electrolyte over the entire range x of lithium insertion or extraction. Insertion compound Li xMyOz should be low cost and environmentally benign to be employed in practical cells and commercially feasible.

Considering s ome o f t hese c riteria, t ransition me tal o xides crystallizing in l ayered L iMO2 , s pinel L iM 2O 4 , a nd ol ivine LiMPO4 structures (M = transition metal) have become appealing as cathodes, while carbon-based materials such as graphite have become attractive anodes for lithium ion batteries. Ā is is because w hile t hese o xides off er high voltages of 3.5–4.3 V versus metallic lithium, graphite offers a low voltage of <0.5 V versus metallic lithium, resulting in a c ell voltage of 3–4 V on coupling t hese o xide c athodes a nd g raphite a node. Ā e high voltage o ffered b y t hese o xide c athodes i s d ue to a l arge

Discharge

Charge

e−

e−

Anode

Cathode Discharge Li+

Charge Li+

Li+

Li+

Electrolyte Lix C6

Li1−x CoO2

FIGURE 8 .3 Operating pr inciples of a l ithium ion c ell, i llustrating the charge–discharge process.

10/3/2008 8:46:24 PM

8-4

Smart Materials

Madelung energ y i n o xides a nd t he c onsequent a bility to stabilize high oxidation states such as Co3+/4+. However, t hese oxide cathodes differ signicantly in some of their performance factors such as energ y and power densities as well as cost and environmental impact. Also, although the lithium ion battery technology has revolutionized the portable electronics market, they are yet to enter the HEV and PHEV arena. Moreover, the rapid developments in the miniaturization and features of portable ele ctronics de vices dem and a n i ncrease i n t he energ y density of t he lithium ion bat teries while lowering t he cost of electrode ma terials. Ā e s ections b elow p resent t he re cent developments a nd c urrent s tatus o f t he sm art c athode a nd anode materials mentioned above.

8.4.1

Layered Oxide Cathodes

Oxides with a general formula LiMO2 (M = Co and Ni) adopt a layered s tructure i n w hich t he L i+ an d M 3+ ions occupy the alternate (111) planes of a rock salt structure as shown in Figure 8.4 to g ive a l ayer s equence o f – O–Li–O–M–O– a long t he c-axis. With an oxygen stacking sequence of ABCABC along the c-axis, this structure is designated as O3 layer structure as the L i+ io ns o ccupy t he o ctahedral si tes a nd t here a re t hree

FIGURE 8.4 Crystal structure of layered LiMO2 (M = Co or Ni).

43722_C008.indd 4

MO2 s heets p er u nit c ell. W hile t he L i+ ion s pre sent b etween the strongly bonded MO2 layers provide facile, re versible L i+ion diff usion, t he e dge-shared M O6 oc tahedral a rrangement with a direct M–M interaction provides good electronic conductivity. As a result, for example, layered LiCoO2 offers acceptable p ower c apability ( charge–discharge r ate) f or p ortable applications such as laptop computers and cell phones. On the other h and, a l arge w ork f unction a ssociated w ith t he h ighly oxidized C o3+/4+ couple provides a h igh cell voltage of a round 4 V and the discharge voltage does not change signicantly with the degree of lithium extraction/insertion x in Li1−x CoO2 (Figure 8.5). Ā e high energy density with acceptable power capability and cycle life has made LiCoO2 an attractive cathode, and commercial l ithium ion c ells presently u se pre dominantly l ayered LiCoO2 as the cathode. However, o nly 5 0% o f t he t heoretical c apacity o f L iCoO2, which corresponds to a re versible extraction or insertion of 0.5 lithium p er C o a nd a p ractical c apacity of 140 mA h /g, c an b e utilized in practical cells as seen in Figure 8.5. Although this limitation wa s o riginally a ttributed t o st ructural d istortions around x = 0. 5 i n Li1−xCoO2, re sulting f rom a n ordering of L i+ ions [19], recent characterization of chemically delithiated samples suggest t hat chemical instability arising f rom a sig nicant overlap of Co3+/4+:3d energy band with the top of the O2−:2p band may p lay a c ritical role i n c ontrolling t he re versible c apacity values [20]. Ā is conclusion is consistent with the introduction of holes (removal of electrons) into the O 2−:2p band rather than into t he Co 3+/4+:3d ba nd d uring ele ctrochemical c harge a s revealed b y x -ray a bsorption s pectroscopic [ 21] a nd ele ctron energy loss spectroscopic [22] studies. Additionally, Co is relatively expensive and toxic. Moreover, the chemical instability of LiCoO2 le ads to s afety c oncerns u nder t he c onditions o f o vercharge, which becomes a serious issue particularly in the case of large bat teries suc h a s t hose ne cessary f or H EV a nd P HEV applications. Also, the power capability of LiCoO2 is not attractive for vehicle applications. Ā e above drawbacks associated w ith LiCoO2 have created interest i n t he de velopment o f a lternative c athode ma terials. In this regard, the nickel-rich LiNi0.85Co 0.15O2 that has the layered s tructure s hown i n F igure 8 .4 b ecame app ealing a f ew years a go [23] a s i t off ers a h igher c apacity of 180 mA h /g a s seen in Figure 8.5 and Ni is less expensive than Co. However, more i n-depth i nvestigation h as s hown t hat L iNi0.85Co 0.15O2 suffers f rom i mpedance g rowth d uring c ycling at ele vated temperatures, which may partly be related to the migration of Ni 2+/3+ io ns f rom t he t ransition me tal p lane to t he l ithium plane at elevated temperatures [24]. Recently, layered compositions such as LiNi1/3Mn1/3Co1/3O2 [25–27] and LiNi1/2Mn1/2O2 [28,29] have become attractive as they offer capacities close to 200 mA h /g at a lo wer c ost c ompared to L iCoO2 . H owever, their p ower c apability i s lo wer t han t hat o f L iCoO2 du e t o some c ation d isordering b etween t he t ransition me tal a nd lithium planes, and the power capability may not be adequate for vehicle applications.

10/3/2008 8:46:25 PM

8-5

Smart Battery Materials

Cell voltage (V)

4.5 4

LiCoO2

3.5

LiMn2O4 LiNi0.85Co0.15O2

3

LiFePO4

2.5 0

50

100

150

200

Capacity (mA h/g)

FIGURE 8.5 Comparison of the discharge curves of layered LiCoO2, layered LiNi0.85Co0.15O2, spinel LiMn2O4, and olivine LiFePO4.

More recently, complex layered oxide solutions between Li[M] O2 (M = Ni, Mn, and Co) and Li[Li1/3Mn2/3]O2 that have the same O3 layered structure shown in Figure 8.4 have become appealing as they exhibit high capacities of up to 280 mA h/g [29–32], which is two times higher than that of LiCoO2. However, these oxides encounter a n i rreversible lo ss o f o xygen ga s f rom t he l attice during  rst c harge a nd a h uge i rreversible c apacity lo ss o f 40–100 mA h /g i n t he  rst c ycle. A lso, t hey ne ed a h igh c utoff charge vol tage o f 4 .8 V a nd s low c harge–discharge r ate ( low power c apability). Ā e i rreversible c apacity lo ss c ould b e s uppressed to s ome extent by a su rface modication of t he oxides with inert oxides like Al2O3 [32]. Ā e amount of oxygen loss from the lattice and the reversible capacity values have been found to increase with increasing lithium content in the transition metal layer i n suc h L i[M1−h Li h ]O2 so lid so lutions [ 33,34]. A lso, t he irreversible o xygen lo ss f rom t he l attice h as b een f ound to b e sensitive to t he substitution of ot her elements l ike A l3+ for L i+ and F − for O 2− [35]. Optimization of such complex oxide solid solutions as well as development of robust electrolytes that are stable up to 4 .8 V c ould m ake t hese h igh-capacity c athodes attractive f or p ortable ele ctronic de vices a lthough t heir lo w power capability due to cation disorder may not be adequate for vehicle applications.

8.4.2 Spinel Oxide Cathodes Although layered oxides exhibit high capacity, some of their performance parameters such as power capability are often limited by c ation d isorder b etween t he t ransition me tal a nd l ithium planes as well as structural and chemical instabilities occurring during cycling. Ā e compositions with low cation disorder tend to t ransform f rom t he i nitial O3 s tructure to, f or e xample, O1 structure at deep charge [20,27,36]. Ā e limited power capability as well as the safety and cost concerns particularly in compositions with high Co contents makes them less appealing for HEV and PHEV applications. In this regard, LiMn2O4 crystallizing in the spinel structure has appealing for vehicle applications as Mn is i nexpensive a nd en vironmentally b enign. A lso, t he t hreedimensional spinel structure (Figure 8.6) w ith LiO6 tetrahedra and e dge-shared M nO6 octahedra provides good structural

43722_C008.indd 5

FIGURE 8 .6 Crystal s tructure of s pinel L iMn2O4, i llustrating t he three-dimensional framework with LiO4 tetrahedra and MnO6 octahedra.

stability d uring t he c ycling w hile t he ly ing o f t he M n3+/4+:3d band well above t he O 2−:2p band offers good chemical stability unlike, for example, layered LiCoO2. However, LiMn2O4 that has been investigated extensively over the years has been plagued by severe capacity fade particularly at elevated temperatures. Several mechanisms such as Mn dissolution f rom t he l attice [ 37], s tructural ( Jahn–Teller) d istortion during over discharge [38], formation of two cubic phases during the charge–discharge process [39], and loss of crystallinity have been p roposed to ac count f or t he c apacity f ade [ 40]. S everal approaches such as cationic and anionic substitutions as well as surface modication have been pursued to improve the capacity retention, but they could not completely overcome the problem. More re cently, f rom a s ystematic i nvestigation o f a n umber o f singly a nd do ubly sub stituted s pinel c ompositions, i nitial M n valence and the lattice parameter difference ∆a between the two cubic phases formed during the charge–discharge process have been f ound to p lay a c ritical role i n c ontrolling t he c apacity retention [41–43].

10/3/2008 8:46:27 PM

8-6

Smart Materials

However, cationic substitutions often result in an increase in the oxidation state of Mn and a consequent decrease in capacity values to unattractive levels of <100 mA h/g. Ā is p roblem h as b een overcome by a partial substitution of F− for O2− in stabilized cation-substituted spinel oxide compositions. For example, stabilized spinel oxyuoride compositions such a s LiMn1.8Li0.1Ni0.1O3.8F0.2 exhibit e xcellent c apacity re tention at ele vated tem peratures with a capacity of >100 mA h/g, high power capability, superior storage properties, and low irreversible capacity loss in the  rst cycle [44]. Ā e excellent electrochemical performance is found to be due to a sm aller lattice mismatch (low ∆a) between t he t wo cubic phases formed during the charge–discharge process, suppressed m anganese d issolution, a nd a n i nitial M n v alence o f >3.6+. From a systematic investigation of several spinel compositions, t he c apacity f ade i s found to b ear a c lear rel ationship to initial Mn valence, ∆a, and Mn dissolution [45]. Ā e Mn dissolution c ould a lso b e suppressed f urther by m ixing t he s tabilized spinel oxyuoride compositions with a layered oxide like LiCoO2 and charging to high enough voltages (4.7 V) in the rst cycle [46]. Ā is is because the overcharged layered Li1−xCoO2 traps the trace amounts of protons present in the electrolyte and thereby suppresses the disproportionation of Mn3+ and the Mn dissolution from the spinel lattice. Ā is composite strategy involving a spinel + layered oxide mixture could become a viable strategy to adopt the spinel cathodes for automotive applications. In add ition to t he 4 V s pinel c athodes de scribed a bove, 5 V spinel cathodes based on LiMn1.5Ni0.5O4 [47] are appealing particularly for high-power applications. As in the case of 4 V spinel cathodes, t he ele ctrochemical p erformance o f 5 V s pinel a lso bears a c lear rel ationship to t he l attice m ismatch a mong t he three cubic phases formed during the charge–discharge process [48]. Stabilized spinel compositions such as LiMn1.42Ni0.42Co0.16O4 that i nvolve a m uch smaller lattice m ismatch a mong t he c ubic phases exhibit better cyclability. However, development of robust electrolyte compositions t hat off er lo ng-term s tability at 5 V i s critical to realize the full potential of these 5 V cathodes.

8.4.3

Olivine Oxide Cathodes

Another candidate that exhibits excellent structural and chemical stabilities along with superior safety characteristics is LiFePO4 crystallizing in the olivine structure as shown in Figure 8.7 [17]. Ā e excellent safety is due to t he chemically more stable Fe2+/3+ redox c ouple a nd t he c ovalently b onded P O4 g roups. D espite the Fe2+/3+ couple, LiFePO4 exhibits a h igh discharge voltage of 3.6 V as seen in Figure 8.5 due to the inductive effect caused by the c ountercation P 5+ a nd t he lo wering o f t he F e2+/3+ re dox energy [ 49,50]. H owever, t he m ajor d rawback w ith ol ivine LiFePO4 i s t he lo w ele ctronic a nd l ithium io n c onductivities and t he c onsequent de crease i n t he p ower c apability. Ā e low electronic conductivity is due to the highly localized Fe2+/3+ redox couple, the presence of FeO6 octahedra, and PO4 tetrahedra, a nd t he coexistence of L iFePO4 and FePO4 as line phases during t he c harge–discharge p rocess w ithout a ny de tectable mixed valent Fe2+/3+.

43722_C008.indd 6

FIGURE 8.7 Crystal structure of olivine LiFePO4, consisting of FeO6 octahedra and PO4 tetrahedra.

However, t his p roblem h as b een o vercome sig nicantly by making the LiFePO4 powder at the nanoscale to reduce the lithium diffusion length as well as by an intimate mixing or coating with conductive carbon to improve the electrical conductivity [51–53]. A lthough c onventional L iFePO4 p owder w ith a la rger particle size undergoes a two-phase reaction involving LiFePO4 and FePO4 without any solid solubility range during the charge– discharge process, reducing the particle size to t he nanoscale is believed to sh ift t he m iscibility gap to lo wer temperatures a nd provide s ome s olid s olubility r ange a nd m ixed v alency. Wi th nanoscale powder and conductive carbon along with associated engineering, sig nicant progress has been made recently by A123 company to employ LiFePO4 for high-power applications. LiFePO4 is currently employed in lithium ion bat teries used in power tools, a nd it is i ntensively being pursued for automotive applications.

8.4.4

Carbon Anodes

With a lightweight and low electrochemical potential lying close to that of metallic lithium, graphite has become an attractive anode, a nd c ommercial l ithium io n c ells c urrently u se mo stly the carbon anodes [54,55]. It offers a t heoretical c apacity o f 372 mA h/g, which corresponds to a n insertion of one lithium per six carbon atoms (x = 1 in Li xC6). One of the drawbacks with the carbon anodes is the occurrence of signicant a mount o f irreversible capacity loss during the  rst discharge–charge cycle due to unwanted, irreversible side reactions with the electrolyte. Although natural graphite cannot be charged with electrolytes consisting of propylene carbonate (PC) as it leads to an evolution of gas at around 1 V, this problem could be overcome with electrolytes consisting of ot her solvents such as ethylene carbonate

10/3/2008 8:46:28 PM

8-7

Smart Battery Materials

(EC) a nd d iethyl c arbonate (DEC). Additionally, h ard c arbons (glassy carbon) obtained by a thermal decomposition of phenolic a nd e poxy re sins a nd p roducts f rom p etroleum p itch s how higher capacities than graphite as they could accommodate extra lithium at the edges and surfaces of the graphene sheets and into nanometer si ze c avities. A lso, h ard c arbons c an b e u sed w ith PC-based electrolytes unlike graphite. However, the hard carbons show a sloping discharge pro le between 0 and 1 V unlike graphite, which shows a nearly flat discharge profile between 0 and 0.3 V.

8.5

Conclusions

Lithium io n bat teries h ave re volutionized t he p ortable ele ctronics market because of the development of smart cathode, anode, a nd ele ctrolyte c ompositions. A lthough o nly 5 0% o f the t heoretical c apacity o f t he c urrently em ployed l ayered LiCoO2 ca thode ca n b e us ed in p ractical c ells, s ignicant progress i s b eing m ade i n re cent ye ars a nd c omplex l ayered oxide s olid s olutions del ivering t wo t imes h igher c apacity than LiCoO2 have been identied. Ā ese high-energy density cathodes a re at tractive f or p ortable app lications, b ut f urther work i s ne eded to i mprove t he c harge–discharge r ates a nd electrolyte stability to higher charge voltages as well as to address t he i rreversible rele ase o f o xygen ga s d uring  rst charge. D espite t he suc cess i n p ortable de vices, l ithium io n batteries are yet to enter the HEV and PHEV arena mainly due to the high cost, safety concerns, and limited power capability of the layered LiCoO2 cathode. In this regard, olivine LiFePO 4 and stabilized spinel LiMn 2O4 (cation- and anion-substituted) cathodes are appealing as they offer high power capability and better safety while both Fe and Mn are inexpensive and environmentally b enign. A ccordingly, t hese c athodes a re b eing pursued for HEV and PHEV applications. Although mo st o f t he l ithium io n c ells a re c urrently m ade with carbon anode, several new materials systems such as nanostructured tin alloys and nanocomposites based on Si are being intensively pursued. Some major issues with some of these materials are large volume changes occurring during the charge–discharge processes and high irreversible capacity loss during the rst cycle. Another attractive anode system is the spinel Li4Ti5O12, but the drawback is its higher discharge voltage (∼1.5 V vs. lithium) and the consequent decrease in cell voltage as well as lower capacity (∼175 mA h/g) compared to the carbon anode. Overall, c ontinued f ocus o n t he de velopment o f ne w, sm art materials and processes has the potential to increase the energy density o f t he l ithium io n bat teries f or p ortables a s w ell a s to make them viable for HEV and PHEV applications. Ā e interest in HEV and PHEV is becoming more prominent and automakers around the globe are beginning to be intensively involved in this effort. Successful development of lithium ion technology for automobiles w ill h ave a p rofound i mpact i n add ressing one of the critical challenges facing humankind in the twenty-rst century: i ncreasing energ y ne eds, en vironmental c oncerns, a nd global warming.

43722_C008.indd 7

Acknowledgment Ā is w ork w as supp orted b y t he W elch F oundation g rant F-1254.

References 1. C. A. Vincent a nd B . S crosati, Modern B atteries: An Introduction to Elec trochemical P ower S ources, E dward Arnold, London, 1984, p. 16. 2. D. Linden, e d., Handbook of B atteries, 2nd e dn., McGrawHill, New York, 1995. 3. J. O. Besenhard, ed., Handbook of Battery Materials, WileyVCH, Weinheim, 1999. 4. P. J. Gellings and H. J. M. Bouwmeester, Ā e CRC Handbook of Solid State Electrochemistry, CRC Press, Boca Raton, FL, 1997. 5. D. R . Lide , e d., CRC H andbook o f Ch emistry a nd Ph ysics, 81st edn., CRC Press, Boca Raton, FL, 2000, p. 821. 6. Lithium B attery Ener gy S torage (LIBES) Pub lication, Technological Research Association, Tokyo, 1994. 7. M. S. Whittingham a nd A. J . J acobson, Intercalation Chemistry, Academic Press, New York, 1982. 8. J. P. Ga bano, Lithium B atteries, Academic Pr ess, L ondon, 1983. 9. H. V. Venkatasetty, Lithium Battery Technology, John Wiley, New York, 1984. 10. G. Pistoia, ed., Lithium Ba tteries: N ew M aterials, D evelopments and Perspectives, Vol. 5, Elsevier, Amsterdam, 1994. 11. C. J ulien a nd G. A. N azri, Solid S tate Ba tteries: M aterials Design and Optimization, Kluwer, Boston, MA, 1994. 12. G.-A. Nazri and G. Pistoia, eds., Science and Technology of Lithium Batteries, Kluwer Academic, Boston, MA, 2003. 13. M. Wakihara and O. Yamamoto, ed., Lithium Ion Batteries: Fundamentals a nd P erformance, W iley-VCH, W einheim, 1998. 14. K. Mizushima, P . C. J ones, P . J . Wiseman, a nd J . B . Goodenough, Mater. Res. Bull. 15, 783, 1980. 15. J. B. Goodenough, K. Mizushima, and T. Takeda, Jpn. J. Appl. Phys. 19, 305, 1983. 16. M. M. Ā ackeray, W. I. F . D avid, P . G. B ruce, a nd J . B . Goodenough, Mater. Res. Bull. 18, 461, 1983. 17. A. K. Padhi, K. S. Nanjundaswamy, and J. B. G oodenough, J. Electrochem. Soc. 144, 1188, 1997. 18. M. K. Aydinol and G. Ceder, J. Electrochem. Soc. 144, 3832, 1997. 19. J. N. Reimers and J. R. Dahn, J. Electrochem. Soc. 139, 2091, 1992. 20. J. Cho i, E. Alvarez, T. A. Arunkumar, a nd A. M anthiram, Electrochem. Solid State Lett. 9, A241, 2006. 21. L. A. Montoro, M. Abbate, and J. M. Ros olen, Electrochem. Solid State Lett. 3, 410, 2000. 22. A. H ightower, J . G raetz, C. C. Ahn, P. Rez, a nd B . F ultz, 198th Meeting of the Electrochemical Society, Phoenix, AZ, October 22–27, 2000, Abstract No. 177.

10/3/2008 8:46:29 PM

8-8

23. W. Li and J. Curie, J. Electrochem. Soc. 144, 2773, 1997. 24. A. M. Kannan and A. Manthiram, J. Elec trochem. S oc. 150, A349, 2003. 25. T. Ohzuku and Y. Makimura, Chem. Lett. 642, 2001. 26. K. M. Shaju, G. V. S. Rao, and B. V. R. Chowdari, Electrochim. Acta 48, 145, 2002. 27. J. Choi and A. Manthiram, J. Electrochem. Soc. 152, A1714, 2005. 28. T. Ohzuku and Y. Makimura, Chem. Lett. 744, 2001. 29. Z. Lu, D. D. Macneil, and J. R. Dahn, Electrochem. Solid State Lett. 4, A191, 2001. 30. S. H. Kang, Y. K. Sun, and K. Amine, Electrochem. Solid-State Lett. 6, A183, 2003. 31. Y. J. Park, Y. S. Hong, X. Wu, K. S. Ryu, and S. H. Chang, J. Power Sources 129, 288, 2004. 32. Y. Wu a nd A. M anthiram, Electrochem. S olid S tate Le tt. 9, A221, 2006. 33. A. R. Armstrong, M. Holzapfel, P. Novak, C. S. Johnson, S. Kang, M. M. Ā ackeray, a nd P. G. B ruce, J. Am. C hem. S oc. 128, 8694, 2006. 34. T. A. Arunkumar, Y. Wu, and A. Manthiram, Chem. Mater. 19, 3067, 2007. 35. Y. Wu a nd A. M anthiram, Electrochem. S olid S tate Le tt. 10, A151, 2007. 36. S. Venkatraman, J . Cho i, a nd A. M anthiram, Electrochem. Commun. 6, 832, 2004. 37. R. J. Gummow, A. de Kock, and M. M. Ā ac keray, Solid State Ionics 69, 59, 1994. 38. M. M. Ā ackeray, Y. Shao-Horn, A. J. Kahaian, K. D. Kepler, E. S kinner, J . T. Vaugney, a nd S. A. H ackney, Electrochem. Solid State Lett. 1, 7, 1998. 39. Y. Shin and A. Manthiram, Electrochem. Solid-State Lett. 5, A55, 2002.

43722_C008.indd 8

Smart Materials

40. H. Huang, C. A. Vincent, and P. G. Bruce, J. Electrochem. Soc. 146, 3649, 1999. 41. Y. Shin and A. Manthiram, Chem. Mater. 15, 2954, 2003. 42. Y. S hin a nd A. M anthiram, J. Elec trochem. S oc. 151, A204, 2004. 43. A. Manthiram, J. Choi, and W. Choi, Solid State Ionics 177, 2629, 2006. 44. W. Choi and A. Manthiram, Electrochem. Solid-State Lett. 9, A245, 2006. 45. W. Choi and A. Manthiram, J. Electrochem. Soc. 154, A792, 2007. 46. A. Manthiram and W. Choi, Electrochem. Solid State Lett. 10, A228, 2007. 47. F. G. B. Ooms, E. M. Kelder, J. Schoonman, M. Wagemaker, and F. M. Mulder, Solid State Ionics 152, 143, 2002. 48. T. A. Arunkumar and A. Manthiram, Electrochem. Solid State Lett. 8, A403-A405, 2005. 49. A. Manthiram a nd J. B. G oodenough, J. Power S ources 26, 403, 1989. 50. A. Manthiram and J. B. Goodenough, J. Solid State Chem. 71, 349, 1987. 51. A. Yamada, S. C. Chung, and K. Hinokuma, J. Electrochem. Soc. 148, A224, 2001. 52. H. Huang, S. C. Yin, and L. F. Nazar, Electrochem. Solid State Lett. 4, A170, 2001. 53. S. Y. Chung, J. T. Bloking, and Y. M. Chiang, Nature Materials 1, 123, 2002. 54. N. I manishi, Y. Takeda, a nd O. Yamamoto, in Lithium I on Batteries: Fundamentals and Performance, M. Wakihara and O. Yamamoto, ed., Wiley-VCH, Weinheim, 1998, p. 98. 55. M. Winter a nd J . O . B esenhard, in Lithium I on Ba tteries: Fundamentals a nd P erformance, M. Wakihara a nd O. Yamamoto, ed., Wiley-VCH, Weinheim, 1998, p. 127.

10/3/2008 8:46:29 PM

9 Piezoelectric and Electrostrictive Ceramics Transducers and Actuators 9.1

Smart Ferroelectric Ceramics for Transducer Applications ................................................ 9-1 Introduction • Piezoelectric and Electrostrictive Effects in Ceramic Materials • Measurements of Piezoelectric and Electrostrictive Effects • C ommon Piezoelectric and Electrostrictive Materials • Piezoelectric Composites • Applications of Piezoelectric and Electrostrictive Ceramics • Current Research and Future Trends

References .......................................................................................................................................... 9-11 9.2 Smart Ceramics: Transducers, Sensors, and Actuators ...................................................... 9-12 Introduction • P iezoelectricity • P iezoelectric Materials • Appl ications of Piezoelectricity

References .......................................................................................................................................... 9-27 9.3 Noncontact Ultrasonic Testing and Analysis of Materials ................................................ 9-27 Introduction • NC U Transducers • NC U System and Signal Processing • NC U Techniques and Applications • Perusal of NCU

References .......................................................................................................................................... 9-40

9.1

Smart Ferroelectric Ceramics for Transducer Applications

A.L. Kholkin, D.A. Kiselev, L.A. Kholkine, and A. Safari 9.1.1 Introduction In the fast developing world, the use of smart materials becomes more and more important to implement sophisticated functions of t he de signed de vice s tarting f rom t he m aterials le vel. I n a common denition [1], the smart materials differ from the ordinary materials in that they can perform two or several functions sometimes w ith a u seful c orrelation o r f eedback me chanism between t hem. I n t he c ase o f p iezoelectric o r ele ctrostrictive materials, it means that the same material can be used for both sensor and actuator functions. A piezoelectric or an electrostrictive sensor converts mechanical variable (displacement or force) into a measurable electrical quantity by means of a direct piezoelectric or e lectrostrictive e ffect. A lternatively, t he ac tuator co nverts the electrical signal into useful displacement or force. Typically, the term transducer is used to describe actuator (transmitting ) and s ensor (receiving) f unctions. Si nce piezoelectrics

and electrostrictors inherently possess both direct (sensor) and converse ( actuator) eff ects, t hey c an b e c onsidered a s sm art materials in the sense indicated above. The degree of smartness can vary in piezoelectric and electrostrictive materials: it is very often t hat t he merely sm art material (only sensor a nd ac tuator functions) can be eng ineered into a “v ery smart” (i.e., tunable) device or even into an “intelligent structure” where the sensor and ac tuator f unctions a re m utually c onnected v ia i ntegrated processing electronics, sometimes with learning function. The g rowth i n t he t ransducer m arket h as b een r apid a nd i s predicted to c ontinue o n i ts c urrent pac e i n t he b eginning o f twenty-rst century. Piezoelectric and electrostrictive sensors and actuators t ake up a sig nicant p ortion of t he t ransducer m arket with a g rowing t rend, e specially d ue to a utomobile p roduction, active vibration damping, navigation systems, and medical imaging. In this chapter, the principles of piezoelectric and electrostrictive sensors and actuators will be considered along with the properties of the most useful materials and examples of successful devices.

9.1.2

Piezoelectric and Electrostrictive Effects in Ceramic Materials

Piezoelectricity, discovered in Rochelle salt by Jacques and Pierre Curie, is the term used to describe the ability of certain 9-1

43722_C009.indd 1

10/3/2008 9:23:25 PM

9-2

Smart Materials

materials to develop an electric charge that is directly proportional to an applied me chanical s tress (Fi gure 9 .1a). T he piezoelectric materials also show the converse effect, i.e., they deform (strain) proportionally to an applied electric eld (Figure 9 .1b). To e xhibit p iezoelectricity, the crystal should belong to one of the 20 noncentrosymmetric crystallographic classes. An important subgroup of piezoelectric crystals is ferroelectrics, which possess the averaged dipole moment per unit cell (spontaneous polarization) that can be reversed by an application of t he e xternal e lectric eld. A bove t he c ertain temperature (Curie point), most ferroelectrics lose their ferroelectric and piezoelectric properties and become paraelectrics, i.e., materials having centrosymmetric crystallographic structure w ith no s pontaneous p olarization. E lectrostriction i s a second order effect, which refers to the ability of all materials to deform under the application of an electric eld. The phenomenological master equation describing the deformations of an insulating c rystal subjected to b oth el astic s tress a nd ele ctric eld is given by x ij = sijkl X kl + d mij E m + M mnij E m E n ,

in Equation 9.1, i.e., the electrostrictive deformations are much smaller than piezoelectric strains. In this case, under zero stress, Equation 9.1 simply transforms to x ij ≈ d mij E m .

(9.2)

Equation 9 .2 de scribes t he c onverse p iezoelectric e ffect where the electric eld changes dimensions of the sample (Figure 9.1b). In centrosymmetric materials, the piezoelectric effect is absent and the elastic strain is only due to the electrostriction. In single domain f erroelectrics h aving c entrosymmetric p araelectric phase, t he piezoelectric a nd ele ctrostriction coefficients can be described in terms of their polarization and dielectric constant. For e xample, lon gitudinal c oefficients ( both ele ctric  eld and deformation are along tetragonal axis, symmetry 4 mm) can be described as follows:

(9.1)

d 33 = 2Q 11e 0e 33 P3 ,

(9.3a)

M 11 = Q 11 (e 0e 33 )2 ,

(9.3b)

where e33 an d P3 a re t he d ielectric c onstant a nd p olarization along the polar direction, e0 = 8.854 × 10−12 F/m is the permittivity of v acuum, a nd Q11 i s t he p olarization ele ctrostriction c oefficient, which couples longitudinal strain and polarization due to general electrostriction equation:

In this equation, the Einstein summation rule is used for repeating indices. Typically, the electrostriction term (µ EmEn) is more than an order of magnitude smaller than the piezoelectric term

The ma thematical de nition o f t he dir ect p iezoelectric e ffect where applied elastic stress causes charge on the major surfaces of the piezoelectric crystal is given by

P

S3 = Q 11P32 .

P X − ∆X

X

Y + ∆Y

Y

where xij are the components of elastic strain sijkl is the elastic compliance tensor Xkl are the stress components dmij are the piezoelectric tensor components Mmnij is the electrostrictive tensor Em and En are the components of an external electric eld

(9.4)

+ Voltage −

(a)

Y

+ P

Charge −

P

X (b) Force

FIGURE 9.1 Schematic representations of the direct and converse piezoelectric effect: (a) an electric eld applied to the material changes its shape and (b) a mechanical force on the material yields an electric eld across it.

43722_C009.indd 2

10/3/2008 9:23:25 PM

9-3

Piezoelectric and Electrostrictive Ceramics Transducers and Actuators

Pm = d mi X i ,

(9.5)

where Pm is the component of electrical polarization. In the case of electrostriction (centrosymmetric crystals), no charge appears on t he su rface o f t he c rystal up on s tressing, a nd t he c onverse electrostriction effect is simply a change of the inverse dielectric constant under mechanical stress: ∆(l /(ee 0 )) = 2Q 11 X 3 .

(9.6)

It should be noted t hat t he reduced matrix notation [2] for t he piezoelectric a nd electrostriction coefficients, a nd stress tensor is used in Equations 9.3 through 9.6. The p iezoelectric a nd ele ctrostrictive e ffects w ere d escribed for the case of single domain crystals in which the spontaneous polarization i s c onstant e verywhere. A te chnologically i mportant class of materials is piezoelectric and electrostrictive ceramics, w hich c onsist o f r andomly o riented g rains, s eparated b y grain boundaries. Ceramics are much less expensive in processing than single crystal and typically offer comparable piezoelectric and electrostrictive properties. Apparently, in nonferroelectric ceramics, piezoelectric effect of individual grains is canceled out by averaging over the entire sample and the whole structure will have a m acroscopic c enter o f s ymmetry a nd ne gligible p iezoelectric properties. Only electrostriction can be observed in such Poled

Unpoled

Ep

(a)

(a)

.7)

where Ei are t he components of electric  eld arising due to t he external stress Xj. The charge coefficients dij a nd voltage coefficients are related by the following equation: g ij = dij /(e 0e 33 ).

(9.8)

An i mportant p roperty o f p iezoelectric a nd ele ctrostrictive transducers is t heir e lectromechanical co upling coeffi cient, k, dened as

(b)

V

43722_C009.indd 3

E i = g ij X j , (9

k 2 = resulting mechanical energy/input electrical energy, or (9.9) 2 k = resulting electrical energy/input mechanical energy.

FIGURE 9.2 Schematic of t he poling process in piezoelectric ceramics: ( a) i n t he a bsence of a n e lectric  eld, t he dom ains h ave r andom orientation of p olarization w ith z ero pie zoelectric a ctivity; ( b) t he polarization within the domains and grains are aligned in the direction of the electric eld.

FIGURE 9.3 electric eld.

ceramics. Si ntered f erroelectric m aterials ( single c rystals o r ceramics) consist of regions with different orientations of spontaneous p olarization, t he s o-called f erroelectric do mains. Domains appear when the material is cooled down through the Curie point to m inimize t he electrostatic a nd elastic energ y of the s ystem. D omain b oundaries o r do main w alls a re mo vable under applied electric eld, so the ferroelectric can be poled, i.e., domains become oriented in the crystallographic direction closest to the direction of applied electric eld (Figure 9.2). Typically, poling is performed under high electric eld at elevated temperature to facilitate domain alignment. As a result, initially centrosymmetric c eramic s ample lo ses t he i nversion c enter a nd becomes piezoelectric (symmetry ∞∞m). There are t hree independent piezoelectric coefficients: d33, d31, a nd d15, which relate longitudinal, t ransverse, a nd s hear de formations, re spectively, to the applied electric eld (see Figures 9.1 and 9.3). Another s et o f m aterials c onstants t hat i s f requently u sed to characterize the piezoelectric properties of ceramics are the piezoelectric voltage coefficients, gij dened in a matrix notation as

P

The co upling coeffi cient r epresents t he e fficiency of t he piezoelectric in converting electrical energy into mechanical energy and v ice versa. Si nce t he energ y c onversion i s ne ver c omplete, the coupling coefficient is always less than unity.

V

P

(b)

Schematic of t he longitudinal (a), t ransverse (a), a nd shear deformations (b) of t he piezoelectric ceramic material u nder applied

10/3/2008 9:23:30 PM

9-4

Smart Materials

9.1.3

Measurements of Piezoelectric and Electrostrictive Effects

Different means have been proposed to characterize the piezoelectric and electrostrictive properties of ceramic materials. The resonance technique involves the measurement of the characteristic resonance frequencies when a su itably shaped specimen is driven by a sinusoidally varying electric eld. To a rst approximation, t he b ehavior o f a p oled c eramic s ample c lose to i ts fundamental re sonance f requency c an b e re presented b y a n equivalent circuit as shown in Figure 9.4a. The schematic behavior of the reactance of the sample as a function of frequency is shown in F igure 9.4b. By me asuring t he c haracteristic re sonance f requencies of the sample, the material constants including piezoelectric coefficients can be calculated. The equations used for the calculations of the electromechanical properties are described in the IEEE Standard on Piezoelectricity [3]. The simplest example of piezoelectric measurements by resonance technique relates to a c eramic ro d ( typically 6 mm i n d iameter a nd 1 5 mm lo ng) poled along its length. It can be shown that the coupling coefcient for t his conguration (k33) i s e xpressed a s a f unction of the s eries a nd pa rallel re sonance f requencies, fs an d fp, respectively: k33 = (p /2)( f s / f p )tan[(p /2)( f p − f s )/2]. (9

.10)

The lon gitudinal pie zoelectric c oefficient d33 is then calculated using k33, el astic c ompliance s33, a nd lo w-frequency d ielectric constant e33: d 33 = k33 (e 33 s33 )1/2 . (9

.11)

In a similar manner, other coupling coefficients and piezoelectric moduli c an b e der ived u sing d ifferent v ibration mo des o f t he ceramic sample. The disadvantage of the resonance technique is that me asurements a re l imited to t he s pecic frequencies determined by the electromechanical resonance. The resonance measurements are not easy in electrostrictive samples, which require application of a strong dc bias eld to induce piezoelectric effect in relaxor ferroelectrics (see Section 9.1.4).

Subresonance techniques are frequently used to evaluate piezoelectric p roperties o f c eramic m aterials at f requencies much lower t han t he f undamental re sonance f requency o f t he sample. They include both the measurements of the piezoelectric charge under the action of mechanical force (direct piezoelectric effect) a nd t he me asurement of ele ctric  eld-induced displacement (converse piezoelectric effect). It can be shown that piezoelectric coefficients obtained by direct and converse piezoelectric effects a re t hermodynamically e quivalent. E lectrostrictive properties of ceramics are most easily determined by measurements o f d isplacements a s a f unction o f t he ele ctric  eld or polarization. Thus M and Q electrostriction coefficients can be evaluated ac cording to E quations 9 .1 a nd 9 .4, re spectively. Alternatively, Equations 9.3b and 9.6 can be also used for electrostriction measurements. A d irect te chnique i s w idely u sed to e valuate t he s ensor capabilities of piezoelectric and electrostrictive materials at sufciently low frequencies. The mechanical deformations can be applied in different directions to obtain different components of the p iezoelectric a nd ele ctrostriction ten sors. I n t he si mplest case, the metal electrodes are placed onto the major surfaces of the p iezoelectric t ransducer no rmal to i ts p oling d irection (Figure 9 .1b). Thus the charge is produced on the electrode plates upon mechanical loading, which is proportional to the longitudinal piezoelectric coefficient d33 and the force F exerted on the ceramic sample: Q = d33F. The charge can be measured by means of charge amplier using an etalon capacitor in the feedback loop. Details of the direct piezoelectric measurements can be found in a number of textbooks (see, e.g., Ref. [4]). Measurements of electric eld-induced d isplacements can be performed b y a n umber o f te chniques i ncluding s train ga uges, linear v ariable d ifferential t ransformers ( LVDTs), c apacitance method,  ber optic sensor, and laser interferometry. Metal wire strain gauges are the most popular sensors used to measure strain with the resolution of about 10−6. The strain gauge is glued to the ceramic sample and the resistance of gauge changes according to its deformation. The resistance variation is measured by a precise potentiometer up to the frequency of several MHz. Several gauges need to be used in order to get the complete set of the piezoelectric and electrostrictive coefficients of the sample.

L R

L Reactance

Cm

fr

fa

Frequency

C (a)

Ce

(b)

FIGURE 9.4 (a) E quivalent c ircuit of t he pie zoelectric s ample ne ar it s f undamental e lectromechanical re sonance (top br anch re presents t he mechanical part and bottom branch represents the electrical part of the circuit) and (b) electrical reactance of the sample as a function of frequency.

43722_C009.indd 4

10/3/2008 9:23:33 PM

9-5

Piezoelectric and Electrostrictive Ceramics Transducers and Actuators Piezoelectric Lamp Photo detector Secondary coils Optical fibers Vout Probe Vin Gap

Target surface

Primary coil

FIGURE 9.5 Principle of LVDT used for the measurement of the electric eld-induced deformations in a piezolectric sample.

The design of LVDT is illustrated in Figure 9.5. The moving surface of the sample is attached to the magnetic core inserted in the center of primary and secondary electromagnetic coils. The change of the core position causes the change of the mutual inductance of the coils. The primary coil is supplied by a n ac current, and the signal in the secondary coils is proportional to the displacement of the core. The response speed is limited by the f requency o f ac sig nal a nd t he me chanical re sonance o f the coil, a nd t ypically does not e xceed 100 Hz. The resolution is s ufficiently h igh (10–100 nm) a nd de pends o n t he n umber of turns. Capacitive te chnique for s train me asurements i s ba sed on a change of the capacitance of the parallel-plate capacitor with the air gap between two opposite plates. One of the plates is rigidly connected to the moving surface of the sample and another plate is xed by the holder. The capacitance change due to vibration of the s ample c an b e me asured precisely by a z ero-point p otentiometer and a lo ck-in amplier. Therefore, a h igh resolution (in the Angstrom range) can be achieved by this technique. The measurement frequency must be much lower than the frequency of an ac input signal and typically does not exceed 100 Hz. All the previously described techniques require a mechanical contact between the sample and the measurement unit. Thi s, of course, l imits t he re solution a nd m aximum f requency, a nd prevents the accurate measurements of the phase angle between the d riving vol tage a nd d isplacement ( so-called p iezoelectric loss). Signicant force is exerted on the moving surface, which in the case of ceramic thin lms, may damage the sample. Ther efore, the no ncontact me asurements a re p referred f or t he ac curate determination of electric  eld-induced displacements of piezoelectric an d e lectrostrictive m aterials. P hotonic  ber-optic sensor offers noncontact measurements of the displacement of a at reecting surface. The principle of operation is illustrated in

43722_C009.indd 5

FIGURE 9 .6 Schematic of t he  ber-optic photon ic s ensor u sed for nondestructive evaluation of electric eld-induced strains.

Figure 9.6 [5]. The sensor head consists of a bundle of transmitting and receiving optical bers located in the immediate vicinity of the vibrating surface. The intensity of the reected light depends on the gap width between the moving object and the probe. This intensity v ersus w idth de pendence a llows f or t he ac curate determination o f t he d isplacement i n b oth dc a nd ac mo des. Using a lo ck-in a mplier f or t he me asurements o f t he o utput signal, which is proportional to the light intensity, the resolution of the order of 1 Å can be achieved [5]. The frequency response is determined b y t he f requency ba nd o f t he ph otodiode a nd amplier (typically of t he ord er of s everal hundreds of k Hz). A modication of this techniques based on the cantilever gently touching t he sample has been recently i ntroduced to a void t he inuence of the surface roughness and to decrease the electrode size [6]. This a llowed to me asure piezoelectric and electrostrictive displacements in thick lms. Another t echnique a llowing nonc ontact a ccurate me asurement of the electric eld-induced d isplacements i s t he opt ical interferometry. I nterferometric me thods o f me asuring sm all displacements include homodyne [7], heterodyne [8], and FabriPerot [ 9] te chniques. The c ommon te chnique i s a h omodyne interferometer with an active stabilization of the working point to prevent drift due to thermal expansion. When two laser beams of the same wavelength λ i nterfere, t he l ight i ntensity c hanges periodically ( l/2 p eriod) w ith t he c hange o f t he opt ical pat h length difference between two beams. When one of the beams is reected by the moving object, the intensity of the output light changes depending on the electric eld-induced displacement of the sample. In the early measurements, simple Michelson interferometer w as u sed [10] w hich, i n p rinciple, ena bles v ery h igh resolution (∼10−5 Å). However, the measurements are limited to a narrow f requency range si nce t he sample is at tached to a r igid substrate a nd d isplacement o f o nly t he f ront su rface o f t he

10/3/2008 9:23:34 PM

9-6

Smart Materials

sample is monitored [10]. The error due to bending effect of the sample c an b e v ery h igh e specially i n t he c ase o f f erroelectric thin  lms. In response to that, a double beam (Mach–Zender) interferometer w as su ggested i n w hich t he d ifference o f t he displacements of both major surfaces of the sample is taken into account [11]. The modied version of the double-beam interferometer specially adapted for the thin lm measurement demonstrated t he resolution a s h igh a s 10−4 Å i n t he f requency range (10–105 Hz) and a long-term stability (<1%) [12].

9.1.4 9.1.4.1

Common Piezoelectric and Electrostrictive Materials Single Crystals

A number of single crystals (ferroelectric and nonferroelectric) have dem onstrated p iezoelectricity. Tho ugh nonferroelectric piezoelectrics exhibit much lower piezoelectric coefficients than the ferroelectric materials, t he former are still extensively used in some applications requiring either high-temperature stability or low loss. The most i mportant material i s qu artz (crystalline SiO2), which possess small but very stable piezoelectric properties (e.g., d11 = 2.3 pC/N, x-cut [2]). LiNbO3 and LiTaO3 are ferroelectrics with high Curie temperatures (1210°C and 660°C, respectively) t hat a re mo stly u sed i n su rface ac oustic w ave ( SAW) devices. Recent investigations [13] have shown that rhombohedral single cr ystals i n t he P b(Zn1/3Nb1/3)-PbTiO3 s ystem e xhibit exceptionally l arge piezoelectric c oefficients (d33 = 25 00 pm/V) and coupling coefficients (k = 0.94). Ultrahigh strain of 1.7% was observed in these materials under high electric eld. The se single crystals a re no w b eing i ntensively i nvestigated a nd s how a signicant promise for next generation of smart materials. 9.1.4.2

Piezoelectric and Electrostrictive Ceramics

As i ndicated e arlier, t he r andomness o f t he g rains i n a s-prepared polycrystalline ferroelectric ceramics yields a no npiezoelectric c entrosymmetric m aterial. P iezoelectric b ehavior i s induced b y “poling” t he c eramic ( Figure 9.2). A lthough a ll o f the domains in a c eramic can never be fully a ligned a long the poling a xis due to s ymmetry l imitations, t he ceramic ends up with a ne t p olarization a long t he p oling a xis a nd su fficiently high piezoelectric properties. The largest class of piezoelectric ceramics is made up of mixed oxides co ntaining co rner-sharing oc tahedra o f O 2− io ns. The most technologically i mportant materials i n t his class a re perovskites with a gener al formula ABO3 where A = Na, K, Rb, Ca, Sr, B a, Pb , e tc. a nd B = Ti, Sn, Z r, N b, T a, o r W. P iezoelectric ceramics having this structure include barium titanate (BaTiO3), lead t itanate (PbTiO3), lead zirconate titanate (PbZrxTi1−xO3 or PZT), lead lanthanum zirconate titanate {Pb1−xLa x(ZryTi1−y)1−x/4O3, or P LZT}, a nd le ad m agnesium n iobate { PbMg1/3Nb2/3O3 or PMN}. Several of these ceramics are discussed below. The piezoelectric effect in BaTiO3 was discovered in the 1940s [14] and it became the rst piezoelectric ceramic developed. The Curie p oint o f B aTiO3 i s a bout 1 20°C–130°C. A bove 130°C, a

43722_C009.indd 6

(a)

(b) Ti4+

Ba2+

O2−

FIGURE 9.7 Crystal structure of B aTiO3: (a) above t he Cu rie point, the c ell i s c ubic; ( b) b elow t he Cu rie p oint, t he c ell i s tetragonal w ith Ba 2+ and Ti4+ ions displaced relative to O2− ions.

nonpiezoelectric cubic phase is stable, where the center of positive charge (Ba2+ and Ti4+) coincides with the center of the negative charge (O2−) (Figure 9.7a). When cooled below the Curie point, a tetragonal structure (shown in Figure 9.7b) develops where the center of positive charge is displaced relative to the O2− ion s, leading to t he f ormation o f ele ctric d ipoles. The piezoelectric coefficients of BaTiO3 a re rel atively h igh: d15 = 2 70 a nd d33 =1 90 pC/N [14]. The coupling coefficient for BaTiO3 is approximately 0.5. Due to its high dielectric constant, BaTiO3 is widely used as a capacitor material. Lead titanate, PbTiO3 rst reported to be ferroelectric in 1950 [14], h as a si milar s tructure to B aTiO b ut w ith a sig nicantly higher Curie point (TC = 490°C). Pure lead titanate is difficult to fabricate in a b ulk form. W hen cooled t hrough t he Curie temperature, t he g rains go t hrough a c ubic to te tragonal ph ase change, le ading to l arge s train a nd c eramic f racturing. This spontaneous strain has been decreased by additions of dopants such as Ca, Sr, Ba, Sn, and W. Calcium-doped PbTiO3 [15] has a dielectric constant of about 200, and longitudinal piezoelectric coefficient (d33) o f 65 pC/N. A s a re sult o f su fficiently high piezoelectric coefficient a nd low d ielectric constant, t he voltage piezoelectric coefficients are exceptionally high, which led to the applications of lead titanate as hydrophones and sonobuoys [16]. PZT is a b inary solid solution of PbZrO3 (a n a ntiferroelectric or thorhombic s tructure) a nd PbTiO3 (a f erroelectric t etragonal perovskite structure) [14]. It has a perovskite structure, with the Zr4+ and Ti4+ ions occupying t he B si te of t he A BO3 structure at r andom ( Figure 9.8). PZ T s hows a tem peratureindependent mo rphotropic ph ase b oundary ( MPB) b etween tetragonal and rhombohedral phases at the Zr:Ti ratio of 52:48 (Figure 9.9). Recently, a narrow region of the monoclinic phase was observed in poled PZT ceramics [17]. The existence of this phase m ay e xplain e asy rot ation of t he p olarization b etween tetragonal a nd rhombohedral phases. As a re sult of t he large number of polarization orientations, this material shows efficient

10/3/2008 9:23:35 PM

9-7

Piezoelectric and Electrostrictive Ceramics Transducers and Actuators Cubic paraelectric phase

Tetragonal ferroelectric phase

500 450

PC

Temperature (⬚C)

400 350

AT

300 250

F R(HT) 8 domain states <111>

200 150

FT 6 domain states <100> FM

100 Ps = 0

A0

FR(LT)

PbZrO3 10

20

50

Ps =/ 0 Pb

O

Ps

Ti

30

40 50 60 70 PbTiO3 (Mole %)

80

90 PbTiO3

FIGURE 9.8 Schematic of the crystal structure of Pb(Zr,Ti)O3 ceramics: above the Curie point, the cell is cubic (left); below the Curie point, the cell is tetragonal/rhombohedral.

FIGURE 9.9 Phase diagram of PZT piezoelectric ceramics as a function of mole % of PbTiO3.

poling and large enhancement in the piezoelectric properties at this composition due to domain wall motion. Piezoelectric PZT at t he MPB is usually doped by a v ariety of ions to form what are known as “hard” and “soft” PZTs. Hard PZT (abbreviated typically as PZT 4) is doped with acceptor ions, such as K+ or Na+ at the A site, or Fe3+, Al 3+ or Mn3+ at the B site. This doping lowers the piezoelectric properties and makes the PZT more difficult to pole or depole. Soft PZ T (PZT 5H) i s doped with donor ions, such as La 3+ at t he A si te, or N b5+ or Sb5+ at the B site. It has very high piezoelectric properties and is easy to pole or depole. The comparison of piezoelectric properties of several major piezoelectric ceramics is given in Table 9.1. Another group of perovskite ceramics is represented by PMN (PbMg1/3Nb2/3O3), which belongs to a family of so-called relaxor ferroelectrics or fe rroelectrics w ith a d iff use ph ase t ransition. Unlike nor mal fe rroelectrics, w hich h ave we ll-dened Curie points to s eparate ferroelectric and paraelectric phases, relaxor ferroelectrics e xhibit a b road t ransition p eak i n t he w eak-eld dielectric constant [18]. This kind of transition is often referred to a s d iffuse ph ase t ransition. A d istinct f eature o f rel axor ferroelectrics is a strong frequency dispersion of the dielectric

constant a nd a s hift of its maximum with frequency. It is suggested that relaxor properties are due to local inhomogeneity of B-ions (e.g., Mg2+ and Nb5+) in the perovskite lattice. Relaxors do not possess a p iezoelectric effect w ithout a dc b ias  eld, which breaks nonpie zoelectric c ubic ph ase i nto t he rhomb ohedral ferroelelectric ph ase. The adv antage o f u sing rel axors a s t he actuators i s d ue to t heir l arge i nduced p olarization a nd ele ctrostrictive strain of the order of 10−3 combined with negligible hysteresis. A comparison of electric eld-induced strain in typical piezoelectric ( PZT) a nd ele ctrostrictive c eramics ( PMN) i s shown in Figure 9.10. 9.1.4.3

Processing of Piezoelectric Ceramics

The ele ctromechanical properties of piezoelectric c eramics a re largely in uenced by their processing conditions. Each step of the process must be carefully controlled to yield the best product. Figure 9 .11 i s a  owchart o f a t ypical o xide m anufacturing process for piezoelectric ceramics. The high-purity raw materials are ac curately w eighed ac cording to t heir de sired r atio, a nd mechanically or chemically mixed. During the calcination step, the s olid ph ases re act to y ield t he p iezoelectric ph ase. A fter

TABLE 9.1 Comparison of Piezoelectric and Related Properties of Several Important Piezoelectric Materials Parameter

Quartz

BaTiO3

PbTiO3:Sm

PZT 4

PZT 5H

LF4T

d33 (pC/N) g33 (10−3 V m/N) kt kp

2.3 57.8 0.09

e33X/e0 Qm

5

190 12.6 0.38 0.33 1700

65 42 0.50 0.03 175

289 26.1 0.51 0.58 1300

593 19.7 0.50 0.65 3400

410 20.2 0.45 0.6 2300

900

500

65

—

120

355

328

193

253

Tc (°C)

43722_C009.indd 7

>105

10/3/2008 9:23:35 PM

9-8

Smart Materials

S3 (10 −3 )

2

PZT

PMN-PT 1

10

20

E3 (kV cm −1 )

FIGURE 9.10 Comparison of the electric eld-induced strain in typical piezoelectric (PZT) and relaxor (0.9PMN-0.1PT) ceramics.

calcining, the solid mixture is ground into ne particles by milling. Shaping i s done by a v ariety of c eramic processing te chniques including p owder c ompaction, t ape c asting, s lip c asting, o r extrusion. During the shaping operation, organic materials are typically added to t he ceramic powder to i mprove its  ow and binding c haracteristics. These o rganics a re remo ved i n a lo wtemperature (500°C–600°C) b inder b urnout s tep. A fter binder burnout, the ceramic structure is sintered to a n optimum density at a n ele vated temperature. For t he lead-containing piezoelectric an d e lectrostrictive c eramics ( PbTiO3, PZ T, P MN), sintering i s p erformed i n s ealed c rucibles w ith a n opt imized PbO atmosphere. This is because lead loss occurs in these ceramics above 800°C.

9.1.5

Piezoelectric Composites

Single-phase piezoelectric and electrostrictive materials are not idealIy suited for hydrostatic and ultrasonic applications where the c eramic elemen t r adiates a nd re ceives t he ac oustic w aves. Thou gh the d33 and d31 piezoelectric coefficients are exceptionally high in PZT ceramics, their hydrostatic voltage response is relatively low due to the high dielectric constant and low hydrostatic charge coefficient dh =d33 + 2d31. Since d31 ≈ −0.4d33 in PZT ceramics [14], the hydrostatic sensor capabilities of single-phase ceramics

Raw materials

Characterization

FIGURE 9.11

43722_C009.indd 8

are rather low. In addition, the high density of ceramics results in h igh ac oustic i mpedance m ismatch b etween t he t ransducer and t he me dium i n w hich t he ac oustic w aves a re propagating. On the other hand, piezoelectric polymers have low density (low impedance) and low dielectric constant and piezoelectric coefficients. In the past three decades, researchers have been focusing on methods to combine the best characteristics of ceramics and polymers to ove rcome t he a forementioned d eciencies. Combination of a p iezoelectric c eramic w ith a p olymer a llows one to t ailor p iezoelectric p roperties o f t he c omposites. The mechanical a nd ele ctrical p roperties o f a c omposite s trongly depend upon the characteristics of each phase and the manner in which they are connected. In a d iphasic composite, there are 10 different ways in which the materials can be oriented in threedimensional (3D) space [19]. The possible connectivity patterns are 0–0, 1–0, 2–0, 3–0, 1–1, 2–1, 3–1, 2–2, 3–2, and 3–3. As a matter of convention, the rst and second numbers in the connectivity denote the continuity of the piezoelectric and polymer phases, re spectively. Figure 9.12 shows some of t he composites made in the past 30 years [20]. The important connectivity patterns are 0–3, 1–3, 3–3, and 2–2. The 0-3 composites are made with a homogeneous d istribution o f p iezoelectric c eramic pa rticles within polymer matrix. The primary advantage of these composites is their ability to be formed into shapes while remaining piezoelectrically active. The d isadvantage is t hat t he 0 –3 composites cannot b e s ufficiently p oled b ecause t he c eramic ph ase i s not self-connected in the poling direction. On the other hand, 3–0 composites, which are simply the ceramic matrix with low concentration of polymer inclusions or voids, can be effectively poled and exhibit hydrostatic properties superior to that of single phase PZT [20]. In composites with 3–3 connectivity, the piezoceramic and polymer phases are continuous in three dimensions in t he form of t wo interlocking skeletons having intimate contact with one another. The rst composite with 3–3 connectivity was formed by t he replamine process using coral skeleton [21]. Another e ffective method of making 3–3 composites is called BURPS ( acronym f or B URnedout P lastic Sph eres) [ 22] w ith properties si milar to t he re plamine c omposites. A m ixture o f PZT powder a nd burnable plastic spheres a re used to f abricate the PZ T–polymer c omposites. O ther te chniques, suc h a s rel ic processing [ 23] a nd d istorted re ticulated c eramics [ 24], h ave been de veloped to f abricate 3 –3 c omposites. Re cently, f used deposition mo deling ( FDM) a nd f used de position o f c eramics (FDC) have been used to make ladder and 3D honeycomb composites [25]. I n t he F DM te chnique, a 3 D p lastic mold i s

Mixing

Calcining

Milling

Shaping

Poling

Electroding

Sintering

Binder burnout

Flowchart for the processing of piezoelectric ceramics.

10/3/2008 9:23:35 PM

9-9

Piezoelectric and Electrostrictive Ceramics Transducers and Actuators

Particles in a polymer (0-3)

Ceramic–Air–Polymer composite (1-1-3)

Perforated composite (3-1)

FIGURE 9.12

PVDF composite model (0-3)

Sheet composite (2-2)

Perforated composite (3-2)

Ceramic rods in a polymer (1-3)

Moonie (3-0)

BURPS composite (3-3)

Transverse reinforcement (1-2-3)

Honeycomb composite (3-1)

Ladder composite (3-3)

Schematic of various piezoelectric composites of different connectivities.

Applications of Piezoelectric and Electrostrictive Ceramics

Piezoelectric and electrostrictive effects provide a direct coupling between me chanical a nd ele ctrical qu antities. Ther efore, they have b een e xtensively u sed i n a v ariety o f ele ctromechanical

43722_C009.indd 9

Ceramic–Air composite (3-0)

Replamine composite (3-3)

prepared and  lled with a PZT slurry. The FDC process directly deposits a m ixture o f PZ T a nd p olymer i n t he f orm o f a 3 D ladder structure. Either structure is heat-treated to burn the organic, sintered, and embedded in epoxy polymer. The most extensively studied and used in transducer applications are the composites with 1–3 connectivities. They consist of individual PZT rods, or  bers, aligned in the direction parallel to poling and embedded in a polymer matrix. A decoupling of d and d33 and d31 responses of the PZT elements signicantly enhances dh. The rod diameter, spacing between them, composite thickness, volume fraction of PZT, and polymer compliance inuences the composite performance. The most common methods of forming 1–3 c omposites a re d ice a nd  ll te chnique [ 26] a nd i njection molding [27]. In the former method, the composite is fabricated by dicing deep grooves in perpendicular directions into a s olid sintered blo ck o f p oled PZ T. The g rooves a re bac klled with polymer and the base is removed via either grinding or cutting. In the latter method, a thermoplastic mixture of ceramic powder and organic binder is i njected i nto a c ooled mold . The process can be used to form composites with a variety of rod sizes, shapes, and spacings. This technique has recently been used for the mass production of SmartPanelsÔ by Materials Systems, Inc. [28].

9.1.6

Diced composite (1-3)

devices f or b oth s ensor a nd ac tuator app lications. The direct piezoelectric effect is apparently used for the generation of charge (voltage) i n app lications suc h ga s ig niters, ac oustic p ressure sensors, v ibration s ensors, a ccelerometers, h ydrophones, e tc. [29]. U sing a c onverse e ffect, me chanical d isplacements a re produced u nder t he app lied ele ctric  eld. The b est k nown examples of actuators are the piezoelectric motors, piezoelectrically driven relays, ink-jet heads for printers, noise cancellation systems, V CR h ead t rackers, p recise p ositioners, def ormable mirrors for correction of optical images, etc. [30]. Acoustic and ultrasonic vibrations can be generated by an ac eld at resonance conditions a nd de tected b y a p iezoelectric re ceiver. Very often acoustic sender a nd receiver a re combined i n t he same device, the piezoelectric transducer. Ultrasonic transducers are used for medical imaging, nondestructive testing,  sh nders, etc. [31]. Another important application of piezoelectrics at high frequencies include frequency control, bulk and SAW resonators, lters, and delay lines. The ultrasonic transducers operate in a s o-called pulse-echo mode. The transducer sends an acoustic wave, which is reected from the interfaces and is received by the same transducer. These echoes vary in intensity according to t he type of interface (e.g., tissue and bone). Therefore, an ultrasonic image is created, which represents the mechanical properties of the human bodies. Thus anatomic structures of d ifferent organs c an be recognized i n a real t ime. C learly, w hen a lo w i ntensity o f u ltrasonic w aves i s used a nd a t ransducer i s s ensitive eno ugh, u ltrasonic me dical imaging is one of the safest diagnostic techniques. The transducer is c omposed o f m atching a nd bac king l ayers a nd p iezoelectric

10/3/2008 9:23:36 PM

9-10

Smart Materials

material itself. The backing and matching layers are added to the transducer to reduce the acoustic impedance mismatch between the imaged object and a transducer and to damp acoustic backwave. Composite materials instead of single-phase materials are frequently u sed i n order to i ncrease p erformance of t he t ransducers [20]. Actuator appl ications re quire si gnicant d isplacement a nd generative force to b e produced under moderate electric elds. For precise positioning, a re producibility of t he displacements is also required. In this case, electrostrictive materials (PMN or its solid solution with PT) are preferred because of their small hysteresis a nd sm all rem anence. F igure 9 .13 s hows s everal possible de signs o f t he p iezoelectric a nd ele ctrostrictive actuators. In simple structures, as shown in Figure 9.13a through d, t he ac tuator d isplacements a re s olely d ue to d33, d31, o r d15 effects of the ceramic rod, plate, or tube. Since strain is limited to 10−3, the typical displacement of a 1 cm long actuator is of the order of 10 µm. The multilayer actuator (Figure 9.13b) utilizes the parallel connection of the ceramic plates cemented together. In this case, the displacement of many individual sheets of a piezoelectric ceramics is summed up. The advantage of the multilayer ac tuators i s t he re duced op erating vol tage, f ast s peed, and large generative force. The u seful design is a p iezoelectric tube ( Figure 9 .13c), w hich i s p oled a nd d riven b y t he vol tage applied in radial direction (through t he wall w idth). The axial response is due to d31 coefficient of the material and is proportional t o t he l ength/width ratio. The r adial re sponse c an b e tuned to almost zero by manipulating the geometry of the tube

Structures with strain amplification

Simple structures

V

d

V

(a)

[32]. Th is configuration is beneficial in suppressing unwanted lateral displacements. A nother i mportant de sign i s a s hear actuator (Figure 9 .13d) d irectly t ransforming volt age appl ied normal to t he polarization vector into the pure rotation due to d15- coefficient [33]. As has been previously indicated, all the simple structures are based on pure d33, d31, or d15 displacements, so the displacements are limited to tens of microns. The amplication of the strain at the expense of generative force can be achieved by using monomorph and bimorph structures (Figure 9.13e and f). The se types of actuators produce large displacements (up to several millimeters) but low generative force and slow response. Another way of strain amplication is represented by extensional transducers. One o f t he de signs c alled M OONIE i s s hown i n F igure 9 .13g [34]. This type of actuator utilizes the bending effect o f t he moon-shaped metallic cap attached to both sides of a multilayer actuator. The d31 motion of the actuator is amplied by bending of the metallic cap. Another examples of extensional actuators are RAINBOWs and CYMBALs (not shown in the gure) [35,36]. Flextensional ac tuators ha ve i ntermediate cha racteristics between multilayers and bimorphs and are now intensively used in various actuator applications. An example of the smart structure using extensional actuator (MOONIE) is shown on Figure 9.14. The actuator portion of the device consists of the standard MOONIE a nd a sm all p iezoelectric c eramic emb edded i n t he upper cap that serves as a s ensor. The sensor detects vibrations normal to the actuator surface and, via a feedback loop, sends a signal of appropriate amplitude and phase to the actuator so that it e ffectively c ancels t he e xternal v ibration. P otential app lications of the smart structure shown in Figure 9.14 include active optical systems, rotor suspension systems, and other noise cancellation devices. Recent trends to miniaturization have resulted in the extensive use o f p iezoelectric a nd ele ctrostrictive m aterials i n m icroelectromechanical s ystems ( MEMS). Si nce m iniaturization o f b ulk ceramics i s l imited, t hese m aterials a re u sed i n a t hin o r t hick

(e)

Sensor

V

d

Brass (b)

(f)

V

Feedback loop

PZT V d

V

Brass

d

Vibration (c)

(d)

(g)

FIGURE 9 .13 Typical a ctuator d esigns: s imple s tructures [ (a)–(d)] and structures with strain amplication [(e)–(g)].

43722_C009.indd 10

FIGURE 9.14 Example of the smart system using a PZT sensor incorporated i n MOONIE actuator. (Adapted f rom Dogan, A ., Uchino, K ., and Newnham, R .E., IEEE Trans. Ultrasonics Ferroelectrics Frequency Control, 44, 597, 1997.)

10/3/2008 9:23:36 PM

Piezoelectric and Electrostrictive Ceramics Transducers and Actuators

Feedback controller Feedback signal

Frequency synthesizer + Ref.

Lock-in amplifier

A Differential current amplifier

Sample Tube scanner (x–y scanning)

FIGURE 9.15 Schematic of t he A FM c antilever sensor a nd actuator based on t he PZT thin  lm. (Adapted from Fujii, T. and Watanabe, S., Appl. Phys. Lett., 68, 487, 1996.)

lm f orm. Thin lm ac tuators ba sed o n p iezoelectric e ffect in PZT materials have been demonstrated by several groups. They include micromotors [37], acoustic imaging devices [38], components for atomic force microscope (AFM) [39], and micropumps [40]. The design of the AFM utilizing PZT  lm for both sensing and actuating functions is shown in Figure 9.15. The excitation ac voltage signal superimposed on the actuation dc voltage is applied to PZT lm deposited on the Si cantilever. The vibrational amplitude, which is sensitive to t he atomic force between t he tip a nd investigated s urface, i s d etected b y m easuring t he diff erence between t he cantilever current a nd reference current. The feedback s ystem maintains t his c urrent constant while scanning i n x–y plane. This system does not require the optical registration of the v ibration t hat m akes PZ T-based A FM c ompact a nd a llows the m ultiprobe s ystems to b e ac hieved. Si nce PZ T  lm i s very sensitive to v ibrations, t he v ertical re solution o f suc h A FM approaches that of conventional systems. This electromechanically driven AFM gives an excellent example of using piezoelectric ceramic t hin  lm a s sm art m aterial. The e xamples of ot her promising applications can be found in a recent review [41].

9.1.7

Current Research and Future Trends

Piezoelectric and electrostrictive ceramics currently mostly rely on le ad-oxide ba sed A BO3 p erovskites d ue to t heir e xcellent properties as sensors and actuators. However, with the increased public awareness on health problems associated with lead and environmental protection p olicies, re search w ill b e focused on nding lead-free compounds having similar piezoelectric properties to PZT. The newly developed lead-free piezoelectric ceramics with comparable piezoelectric performances to those of PZT materials [ 42] c onsist o f a ne w c omposition ba sed o n ( K,Na) NbO3 alkaline niobate system with textured microstructure. As in the case of PZT, the highest properties are observed near the

43722_C009.indd 11

9-11

MPB between orthorhombic and tetragonal crystal phases of the alkaline n iobate s ystem t hat ge nerated h igh pie zoelectric c onstant (d33 ≈ 400 pC/N, see Table 9.1). Relaxor single-crystal materials with giant piezoelectric effect already found a wide range of applications i n c omposite t ransducers f or me dical i maging to MEMS. The current research has focused on the preparation of these m aterials i n t he c eramic f orm a nd u nderstanding t he nature o f t heir u nusually h igh p iezoelectric p erformance [43]. The current trend to miniaturization will continue giving rise to complex s ensors a nd ac tuators i ntegrated d irectly on a Si c hip [41]. U sing mo dern pat terning te chniques, t he s cale o f p iezoelectric and electrostrictive ceramics can be reduced to submicron d imensions, g iving r ise to na noscale m aterials ( see, e .g., Ref. [44] and references therein) that can be used in nanoelectromechanical systems (NEMS). The research on piezoelectric and electrostrictive materials operating under severe external conditions (temperature, pressure, harsh chemical environments) will be obviously continued for applications in space and deep ocean exploration, noise cancellation in airplanes and helicopters, etc.

References 1. R. B . N ewnham a nd G. R . R uschau, J. In tell. M ater. S yst. Struct. 4, 289, 1993. 2. J. F . N ye, Physical P roperties of Cr ystals, Ox ford: Ox ford University Press, 1985. 3. IEEE Standards on Piezoelectricity, IEEE Std. 176, 1978. 4. V. F. Janas, A. Safari, A. Bandyopadhyay, and A. Kholkin, in Measurements, Instrumentation and Sensor Handbook, J. G. Webster ed., Boca Raton, FL: CRC Press. 5. MTI-2000 F otonic™ S ensor, I nstruction M anual, MTI Instruments Division, Latham NY. 6. N. Vyshatko, P. Brioso, P. Vilarinho, and A. L. Kholkin, Rev. Sci. Instr. 76, 085101, 2005. 7. S. Sizgoric and A. A. Gundjian, Proc. IEEE 57, 1312, 1969. 8. R. M. De La Rue, R. F. Humphryes, I. M. Mason, and E. A. Ash, Proc. IEEE 119, 117, 1972. 9. V. E. Bottom, Rev. Sci. Instrum. 35, 374, 1964. 10. Q. M. Zha ng, W. Y. P an, a nd L. E. Cr oss, J. Ap pl. Phys. 65, 2807, 1988. 11. W. Y. Pan and L. E. Cross, Rev. Sci. Instrum. 60, 2701, 1989. 12. A. L. Kholkin, Ch. Wuethrich, D. V. Taylor, and N. Setter, Rev. Sci. Instrum. 67, 1935, 1996. 13. S. -E. Park and T. R. Shrout, J. Appl. Phys. 82, 1804, 1997. 14. B. Jaffe, W. R. Cook, Jr., and H. Jaffe, Piezoelectric Ceramics, Marietta, OH: R.A.N. 1971. 15. Piezoelectric P roducts, S ensor T echnology Limi ted, Collingwood, Ontario, Canada 1991. 16. K. M. Ri ttenmyer a nd R . Y. T ing, Ferroelectrics 110, 171, 1990. 17. R. Guo, L. E. Cross, S. -E. Park, B. Noheda, D. E. C ox, and G. Shirane, Phys. Rev. Lett. 84, 5423, 2000. 18. L. E. Cr oss, in Ferroelectric C eramics: T utorials, Re views, Theory, Processing and Applications, N. S etter and E. C olla eds., Basel: Birkhouser, 1993.

10/3/2008 9:23:36 PM

9-12

19. R. E. Newnham, D. P. Skinner, and L. E. Cross, Mat. Res. Soc. Bull. 13, 525, 1978. 20. T. R. Gururaja, A. Safari, R. E. Newnham, and L. B. Cross, in Electronic Ceramics, L. M. Levinson ed., New York: Marcel Dekker, p. 92. 21. D. P. Skinner, R. E. Newnham, and L. E. Cross, Mat. Res. Soc. Bull. 13, 599, 1978. 22. T. R. Shrout, W. A. Schulze, and J. V. Biggers, Mat. Res. Soc. Bull. 14, 1553, 1979. 23. S. M. Ting, V. F. Janas, and A. Safari, J. Am. Ceram. Soc. 79, 1689, 1996. 24. M. J . Cr eedon a nd W. A. S chuize, Ferroelectrics 153, 333, 1994. 25. A. Bandyopadhayay, R. K. Panda, V. F. Janas, M. K. Agarwala, S. C. D anforth, a nd A. Sa fari, J. Am. Ceram. S oc. 80, 1366, 1997. 26. H. P. Savakus, K. A. Klicker, and R. E. Newnham, Mat. Res. Soc. Bull. 16, 677, 1981. 27. R. L. Gentilman, D. F. Fiore, H. T. Pham, K. W. French, and L. W. Bowen, Ceram. Trans. 43, 239, 1994. 28. D. Fiore, R. Toni, and R. Gentilman, Proc. of the 8th US–Japan Seminar on Dielectric and Piezoelectric Ceramics, Plymouth, MA, 1997, p. 135. 29. A. J . M oulson a nd J . M. H erbert, Electroceramics: Materials, Properties, Applications, L ondon: Chapman & Hall, 1990. 30. K. Uchino, Piezoelectric Actuators a nd Ul trasonic M otors, Boston: Kluwer, 1997. 31. J. M. H erbert, Ferroelectric Transducers a nd S ensors, N ew York: Gordon & Breach, 1982. 32. Q. M. Zhang, H. Wang, and L. E. Cross, Proc. SPIE, Smart Structures and Materials 1916, 244, 1993. 33. A. E. Glazounov, Q. M. Zhang, and C. Kim, Appl. Phys. Lett. 72, 2526, 1998. 34. A. Dogan, Q. C. Xu, K. Onitsuka, S. Yoshikawa, K. Uchino, and R. E. Newnham, Ferroelectrics 156, 1, 1994. 35. G. Haertling, Bull. Am. Ceramic Soc. 73, 93, 1994. 36. A. D ogan, K. U chino, a nd R . E. N ewnham, IEEE T rans. Ultrasonics Ferroelectrics Frequency Control 44, 597, 1997. 37. P. M uralt, M. K ohli, T . M aeder, A. K holkin, K. B rooks, N. Setter, and R. Luthier, Sensors Actuators A48, 157, 1995. 38. R. E. Newnham, J. F. Fernandez, K.A. Markowski, J. T. Fielding, A. Dogan, and J. Wallis, Mat. Res. Soc. Proc. 360, 33, 1995. 39. T. Fujii and S. Watanabe, Appl. Phys. Lett. 68, 487, 1996. 40. J. Bernstein, K. Houston, L. Niles, K. Li, H. Chen, L. E. Cross, and K. U dayakumar, Proc. o f t he 8t h. S ymp. o n I ntegrated Ferroelectrics, Tempe, AR, 1996, p. 68. 41. N. S etter, D . D amjanovic, L. En g, G. F ox, S. G evorgian, S. Hong, A. Kingon, H. Kohlsedt, N. Y. Park, G. B. Stephenson, I. S tolichnov, A. T agantsev, D . T aylor, T . Yamada, a nd S. Streiffer. J. Appl. Phys. 100, 151606, 2006. 42. Y . Sa ito, H. T akao, T . T ani, T . N onoyama, K. T akatori, T . Homma, T. Nagaya, and M. Nakamura, Nature 432, 84, 2004. 43. Z. Kutnjak, J. Petzelt, and R. Blinc, Nature 441, 956, 2006. 44. A. Gruverman and A. L. Kholkin, Rep. Prog. Phys. 69, 2443, 2006.

43722_C009.indd 12

Smart Materials

9.2

Smart Ceramics: Transducers, Sensors, and Actuators

Kenji Uchino and Yukio Ito 9.2.1 Introduction 9.2.1.1

Smart Material

Let us start with the smartness of the material. Table 9.2 lists the v arious e ffects t hat rel ate t he i nput (electric f ield, m agnetic field, stress, h eat, a nd l ight) to t he o utput (charge/current, magnetization, s train, tem perature, a nd l ight). Conducting a nd el astic m aterials t hat gener ate c urrent a nd strain o utputs, re spectively, f or t he i nput, vol tage, o r s tress (well-known phenomena) are sometimes called trivial materials. C onversely, py roelectric a nd p iezoelectric m aterials that generate an electric field from the input of heat or stress (unexpected phenomena) are called “smart” materials. These off-diagonal c ouplings have a c orresponding c onverse effect such as electrocaloric and converse-piezoelectric effects, and both sensing a nd actuating f unctions can be realized in t he same materials. I ntelligent m aterials s hould p ossess a drive/control or processing f unction, which is adaptive to the change in environmental conditions in addition to the actuation and sensing functions. Note that the ferroelectric materials exhibit most of t hese effects, except magnetic-related phenomena. Thus, the ferroelectrics are said to be very smart materials. The ac tuator i n a na rrow me aning s tands f or m aterials o r devices t hat gener ate me chanical strain (or stress) output. A s indicated by the solid columnar border in Table 9.2, solid state actuators use converse piezoelectric, magnetostriction, elasticity, thermal e xpansion, or photo striction phe nomena. A s hapememory a lloy i s a k ind o f t hermally e xpanding m aterial. O n the other hand, a sensor requires charge and current output in most cases. Thus, c onducting a nd s emiconducting, m agnetoelectric, piezoelectric, pyroelectric, and photovoltaic materials are u sed f or de tecting ele ctric  elds, ma gnetic  elds, stress, heat, and light, respectively (see the dashed columnar border in Table 9.2). In this sense, piezoelectric materials are most popularly used in sm art s tructures a nd s ystems, b ecause t he s ame m aterial i s applicable for both sensors and actuators, in principle. We treat the piezoelectric t ransducers, s ensors, a nd ac tuators m ainly i n this chapter. Even though transducers, in general, are devices that convert input energy to a different energy type of output, the piezoelectric “transducer” is often used to denote a de vice t hat possesses both sensing and actuating functions, exemplied by underwater sonar. 9.2.1.2

Piezoelectric Effect

Certain materials produce electric charges on their surfaces as a consequence of applying mechanical stress. W hen t he induced charge is proportional to the mechanical stress, it is called direct piezoelectric effect and was discovered by J. and P. Curie in 1880.

10/3/2008 9:23:37 PM

9-13

Piezoelectric and Electrostrictive Ceramics Transducers and Actuators TABLE 9.2 Various Effects in Ferroelectric and Ferromagnetic Materials Material device

Input

Output Input

Charge Current

Output

Magnetization

Strain

Temperature

Light

Electro magnetic effect

Converse piezoeffect

Electric caloric effect

Electric-optic effect

Magnetostriction

Electric field

Permittivity Conductivity

Magnetic field

Magnetic electric effect

Permeability

Stress

Piezoelectric effect

Piezomagnetic effect

Pyroelectric effect

—

Heat

Photovoltaic effect

—

Light

Thermal expansion

Specific heat

Photostriction

—

43722_C009.indd 13

Photoelastic effect —

Refractive index

Sensor Actuator

Smart material

Materials that show this phenomenon also conversely have a geometric strain generated that is proportional to an applied electric eld. This is the converse piezoelectric effect. The root of the w ord “ piezo” i s t he Gre ek w ord f or “ pressure;” h ence t he original meaning of the word piezoelectricity implied “pressure electricity” [1,2]. Piezoelectric m aterials c ouple ele ctrical a nd me chanical parameters. T he m aterial u sed e arliest f or i ts p iezoelectric properties w as si ngle-crystal qu artz. Q uartz c rystal re sonators for frequency control appear today at the heart of clocks and a re a lso used i n T Vs a nd computers. Ferroelectric polycrystalline ceramics, such as barium titanate and lead zirconate t itanate (P ZT), e xhibit pie zoelectricity w hen e lectrically poled. Because t hese ceramics possess sig nificant a nd stable piezoelectric effects, that is, high electromechanical coupling, they can produce large strains or forces and hence are extensively u sed a s t ransducers. P iezoelectric p olymers, n otably polyvinylidene dif luoride ( PVDF) a nd i ts c opolymers w ith trifluoroethylene a nd piezoelectric composites t hat combine a p iezoelectric c eramic a nd a pa ssive p olymer h ave b een developed a nd o ffer h igh p otential. Re cently, t hin f ilms o f piezoelectric m aterials a re b eing re searched d ue to t heir potential u se i n m icrodevices (sensors a nd m icroelectromechanical syst ems [ MEMS]). Pi ezoelectricity is bei ng ex tensively used in fabricating various devices such as transducers, sensors, actuators, surface acoustic wave (SAW) devices, and frequency controls. We describe the fundamentals of piezoelectric effect first, then p resent a b rief h istory o f p iezoelectricity, f ollowed b y present day piezoelectric materials that are used, and f inally various p otential app lications o f p iezoelectric m aterials a re presented.

—

Elastic constant

Diagonal coupling Off-diagonal coupling =

Magnetic caloric Magnetic optic effect effect

9.2.2 Piezoelectricity 9.2.2.1 Relationship between Crystal Symmetry and Properties All crystals can be classied into 32 point groups according to their crystallographic symmetry. These point groups are divided into t wo c lasses; o ne h as a c enter o f s ymmetry a nd t he ot her lacks it. There are 21 noncentrosymmetric point groups. Crystals that belong to 20 of these point groups exhibit piezoelectricity. Although cubic class 432 lacks a center of symmetry, it does not permit piezoelectricity. Of these 20 point groups, 10 polar crystal classes c ontain a u nique a xis, a long w hich a n ele ctric d ipole moment is oriented in the unstrained condition. The pyroelectric effect appears in any material that possesses a p olar s ymmetry a xis. The ma terial i n t his c ategory de velops electric c harge o n t he su rface o wing to t he c hange i n d ipole moment as temperature changes. The pyroelectric crystals whose spontaneous polarization are reorientable by applying an electric eld of sufficient magnitude (not exceeding the breakdown limit of the crystal) are called ferroelectrics [3,4]. Table 9.3 shows the crystallographic classication of the point groups. 9.2.2.2

Piezoelectric Coefficients

Materials a re de formed by s tresses a nd t he re sulting de formations are represented by strains (∆L/L). When the stress X (force per unit area) causes a proportional strain x,

x = sX , (9

.12)

where all quantities are tensors; x and X are second rank and s is fourth rank. Piezoelectricity creates additional strains by an applied electric eld E. The piezoelectric equation is given by

10/3/2008 9:23:37 PM

9-14

Smart Materials TABLE 9.3 Crystallographic Classication According to Crystal Centrosymmetry and Polarity Crystal System

Polarity

Symmetry

Cubic

Centro (11)

m3m m3

Nonpolar (22)

Hexagonal

432

6/mmm 6/m 4/mmm 4/m

Polar (Pyroelectric) (10)

Noncentro (21)

6

3m

3

Ortho rhombic

Mono clinic

mmm

2/m

Triclinic

4

32

4

3m

222

42m

6m2

2 6mm

6

4mm

3

mm2

1 m

Note:

Inside the bold line are piezoelectrics.

x ij = sijkl X kl + dijk E k , (9

.13)

where E is the electric eld d is the piezoelectric constant, which is a third rank tensor

0 d 31 ⎤ ⎡ x1 ⎤ ⎡ 0 ⎢x ⎥ ⎢ 0 0 d 31 ⎥ ⎥ ⎡E ⎤ ⎢ 2⎥ ⎢ 0 d 33 ⎥ ⎢ 1 ⎥ ⎢x3 ⎥ ⎢ 0 ⎥ ⎢ E 2 ⎥ . (9 ⎢x ⎥ = ⎢ ⎢ 4 ⎥ ⎢ 0 d15 0 ⎥ ⎢ E ⎥ ⎢ x 5 ⎥ ⎢d15 0 0 ⎥⎣ 3⎦ ⎥ ⎢ ⎥ ⎢ 0 0 ⎦⎥ ⎣⎢ x 6 ⎦⎥ ⎣⎢ 0

.14)

d , ee 0 (9

= d 2 /e 0es.

.15)

where e0 is the permittivity of free space. A measure of the effectiveness of the electromechanical energy conversion is the electromechanical coupling factor k, which measures the fraction of the electrical energy converted to mechanical energy when an electric eld is applied or vice versa when a material is stressed [5]: k 2 = (Stored mechanical energy/input electrical energy) (9.16) or = (Stored electrical energy/input mechanical energy). (9 .17) Let u s c alculate k2 in Equation 9.16, when an electric eld E is applied to a p iezoelectric material. Because t he i nput electrical

(9

.18)

Note t hat k is always less than 1. Typical values of k are 0.10 for quartz, 0.4 f or B aTiO3 c eramic, 0. 5–0.7 f or PZ T c eramic, a nd 0.1–0.3 for PVDF polymer. Another important material parameter is the mechanical quality factor Qm, which determines the frequency characteristics. Qm is given by an inverse value of mechanical loss: Qm =

Another frequently used piezoelectric constant is g, which gives the electric eld produced when a stress is applied (E = gT). The g constant is related to the d constant through the relative permittivity e: g=

energy is (1/2) · e0 · e E2 per unit volume and the stored mechanical energy per unit volume under zero external stress is given by (1/2) x2 /s = (1/2) (d E)2/s, k2 can be calculated as k 2 = [(1/2) (d E )2 /s]/[(1/2) e 0e E 2 ]

This e quation c an b e a lso e xpressed i n a m atrix f orm suc h a s given for a poled ceramic:

43722_C009.indd 14

Rhombo hedral

422

622 23

43m

Tetragonal

1 . (9 tan d

.19)

A low Qm is preferred for sensor applications in general to cover a w ide f requency range. C onversely, a h igh Qm is most desired for ultrasonic actuations (e.g., ultrasonic motors [USM]) to suppress heat generation through the loss. 9.2.2.3

History of Piezoelectricity

As stated a lready, Pierre a nd Jacques Cu rie d iscovered piezoelectricity in quartz in 1880. The d iscovery of ferroelectricity accelerated t he c reation o f u seful p iezoelectric m aterials. Rochelle salt was the rst ferroelectric discovered in 1921. Until 1940, o nly t wo t ypes o f f erroelectrics w ere k nown, Ro chelle salt a nd p otassium d ihydrogen ph osphate a nd i ts i somorph. From 1 940 to 1 943, u nusual d ielectric p roperties suc h a s a n abnormally high dielectric constant were found in barium titanate ( BaTiO3) i ndependently b y W ainer a nd S almon, Ogawa, a nd Wul a nd G olman. A fter t he d iscovery, co mpositional modications for BaTiO3 led to improvement in the temperature s tability a nd h igh-voltage o utput. P iezoelectric transducers based on BaTiO3 ceramics became well-established in a number of devices.

10/3/2008 9:23:37 PM

9-15

Piezoelectric and Electrostrictive Ceramics Transducers and Actuators

In 1 950s, J affe a nd c oworkers e stablished t he PZ T s ystem induces strong piezoelectric effects. The maximum piezoelectric response was found for PZT compositions near the morphotropic phase b oundary b etween t he rh ombohedral a nd te tragonal phases. Since then, the PZT system containing various additives has become the dominant piezoelectric ceramic for a v ariety of applications. The de velopment o f PZ T-based ter nary s olid solution systems was a major success of the piezoelectric industry for these applications. In 1969, Kawai et al. discovered that certain polymers, notably PVDF, are piezoelectric when stretched during fabrication. Such piezoelectric p olymers a re a lso u seful f or s ome t ransducer applications. I n 1 978, N ewnham e t a l. i mproved c omposite piezoelectric m aterials b y c ombining a p iezoelectric c eramic with a passive p olymer w hose p roperties c an b e t ailored to t he requirements of various piezoelectric devices. Another c lass o f c eramic m aterial h as re cently b ecome important: r elaxor-type e lectrostrictors, s uch a s l ead m agnesium n iobate ( PMN), t ypically dop ed w ith 1 0% le ad t itanate (PT), which have potential applications in the piezoelectric actuator  eld. A recent breakthrough in the growth of high-quality, large, s ingle-crystal r elaxor p iezoelectric co mpositions ha s created interest in these materials for applications ranging from high-strain actuators to high-frequency transducers for medical ultrasound de vices d ue to t heir sup erior ele ctromechanical characteristics. More recently, thin lms of piezoelectric materials suc h a s z inc o xide ( ZnO) a nd PZ T h ave b een e xtensively investigated and developed for use in MEMS devices.

9.2.3 Piezoelectric Materials This s ection su mmarizes t he c urrent s tatus o f p iezoelectric materials: single crystal materials, piezoceramics, piezopolymers, piezocomposites, a nd p iezolms. Table 9.4 s hows t he m aterial parameters of s ome re presentative pie zoelectric m aterials described here [6,7]. 9.2.3.1

Single Crystals

Piezoelectric c eramics a re w idely u sed at p resent f or a l arge number of applications. However, single-crystal materials retain their utility; they are essential for applications such as frequency stabilized o scillators a nd su rface ac oustic de vices. The most popular single-crystal piezoelectric materials are quartz, lithium

niobate ( LiNbO3), a nd l ithium t antalate ( LiTaO3). The single crystals are anisotropic in general and have different properties depending on the cut of the materials and the direction of bulk or surface wave propagation. Quartz i s a w ell-known p iezoelectric m aterial. α-Quartz belongs to t he t riclinic crystal system w ith point g roup 32 a nd has a ph ase transition at 5 37°C to t he β-type t hat is not pie zoelectric. Quartz has a c ut with a z ero temperature coefficient of the resonance frequency change. For instance, quartz oscillators using t he t hickness s hear mo de o f t he A T-cut a re e xtensively used as clock sources in computers and as frequency stabilized oscillators i n T Vs a nd v ideo c assette re corder ( VCRs). O n t he other hand, a n ST-cut qu artz substrate t hat has X-propagation has a z ero temperature c oefficient for SAWs a nd s o i s u sed for SAW de vices t hat h ave h ighly s tabilized f requencies. A nother distinguishing characteristic of quartz is its extremely high mechanical quality factor Qm > 105. Lithium n iobate a nd l ithium t antalate b elong to a n i somorphous crystal system and are composed of oxygen octahedron. The Curie temperatures of LiNbO3 and LiTaO3 are 1210°C a nd 660°C, re spectively. The c rystal s ymmetry o f t he f erroelectric phase of these single crystals is 3 m and the polarization direction is along the c-axis. These materials have high electromechanical coupling coefficients for SAWs. In addition, large single crystals can easily be obtained from their melts by using the conventional Czochralski technique. Thus, both materials are very important in SAW device applications. 9.2.3.2

Perovskite Ceramics

Most of the piezoelectric ceramics have the perovskite structure ABO3, s hown i n F igure 9.16. This ide al s tructure c onsists o f a simple c ubic u nit cell t hat has a l arge c ation A at t he corner, a smaller cation B in the body center, and oxygens O in the centers of the faces. The structure is a network of corner-linked oxygen octahedra surrounding B cations. The piezoelectric properties of perovskite-structured materials can be easily tailored for applications b y i ncorporating v arious c ations i n t he p erovskite structure. 9.2.3.2.1 Barium Titanate Barium titanate (BaTiO3) is one of the most thoroughly studied and most widely used piezoelectric materials. Figure 9.17 shows the tem perature de pendence o f d ielectric co nstants i n B aTiO3

TABLE 9.4 Material Parameters of Representative Piezoelectric Materials Parameter

43722_C009.indd 15

Quartz

BaTiO3

PZT 4

PZT 5H

(Pb, Sm)TiO3

PVDF-TrFE

d33 (pC/N) g33 (10−3 V m/N) kt kp

2.3 57.8 0.09 5

289 26.1 0.51 0.58 1300

593 19.7 0.50 0.65 3400

65 42 0.50 0.03 175

33 380 0.30

e3T/e0 Qm TC (°C)

190 12.6 0.38 0.33 1700

>105 120

500 328

65 193

900 355

6 3–10

10/3/2008 9:23:41 PM

9-16

Smart Materials

9.2.3.2.2 A

O B

FIGURE 9 .16 Perovskite s tructure A BO3. ( Modied from Encyclo pedia of Smart Materials.)

that demonstrate the phase transitions in BaTiO3 single crystals. Three anomalies can be observed. The discontinuity at the Curie point (130°C) is due to a transition from a ferroelectric to a paraelectric phase. The other two discontinuities are accompanied by transitions f rom one ferroelectric ph ase to a nother. A bove t he Curie point, the crystal structure is cubic and has no spontaneous dipole moments. At t he Curie point, t he crystal becomes polar and t he s tructure c hanges f rom a c ubic to a te tragonal ph ase. The dipole moment and the spontaneous polarization are parallel to t he te tragonal a xis. J ust b elow t he Cu rie tem perature, t he vector o f t he s pontaneous p olarization p oints i n t he [ 001] direction (tetragonal ph ase) a nd b elow 5° C, i t re orients i n t he [011] (orthrhombic phase), and below −90°C in the [111] (rhombohedral ph ase). The d ielectric a nd p iezoelectric p roperties o f ferroelectric ceramic BaTiO3 can be affected by its own stoichiometry, microstructure, and by dopants entering into the A or B site so lid so lution. M odied BaTiO3 ceramics that contain dopants such as Pb or Ca ions have been used as commercial piezoelectric materials.

Rhombohedral

Piezoelectric P b(Ti, Zr )O3 s olid s olutions ( PZT) c eramics a re widely u sed b ecause o f t heir sup erior p iezoelectric p roperties. The phase diagram of the PZT system (Pb(ZrxTi1−x)O3) is shown in Figure 9.18. The crystalline symmetry of this solid solution is determined by the Zr content. PT also has a tetragonal ferroelectric phase of the perovskite structure. As the Zr content x increases, the t etragonal d istortion d ecreases, a nd w hen x > 0. 52, t he structure c hanges f rom t he te tragonal 4 mm ph ase to a nother ferroelectric phase of rhombohedral 3 m symmetry. Figure 9.19 shows the dependence of several d constants on the composition near the morphotropic phase boundary between the tetragonal and rhom bohedral ph ases. The d c onstants h ave t heir h ighest values ne ar t he mo rphotropic ph ase b oundary. This enhancement in the piezoelectric effect is attributed to the increased ease of re orientation of t he polarization i n a n electric  eld. Doping the PZT material with donors or acceptor changes the properties dramatically. D onor dop ing w ith io ns suc h a s N b5+ or T a 5+ provides soft PZTs like PZT-5, because of the facility of domain motion due to the charge compensation of the Pb vacancy, which is gener ated d uring si ntering. O n t he ot her h and, ac ceptor doping w ith Fe 3+ or S c3+ le ads to h ard PZ Ts suc h a s PZ T-8 because oxygen vacancies pin the domain wall motion. 9.2.3.2.3 Lead Titanate PbTiO3 has a tetragonal structure at room temperature and has large tetragonality c/a = 1.063. The Curie temperature is 490°C. Densely si ntered P bTiO3 ceramics cannot be obtained easily because t hey b reak up i nto p owders w hen c ooled t hrough t he Curie temperature. This is pa rtly due to t he large spontaneous strain that occurs at the transition. PT ceramics modied by small amounts of additives exhibit high piezoelectric anisotropy. Either (Pb, S m)TiO3 [8] or (Pb, Ca)TiO3 [9] h as e xtremely lo w planar c oupling, t hat i s, a l arge kt/kp ratio. Here, kt and kp are thickness-extensional a nd p lanar ele ctromechanical c oupling

Orthorhombic

Tetragonal

Ps

Ps

Ps

Dielectric constant

Lead Zirconate–Lead Titanate

10,000

Cubic

ea

5,000

ec 0 −150

−100

−50

0

50

100

150

Temperature (⬚C)

FIGURE 9.17 Dielectric constants of BaTiO3 as a function of temperature. (Modied from Encyclopedia of Smart Materials.)

43722_C009.indd 16

10/3/2008 9:23:41 PM

9-17

Piezoelectric and Electrostrictive Ceramics Transducers and Actuators 500 Cubic a

400

a

Temperature (⬚C)

a 300 Tetragonal

Morphotropic phase boundery

200 c

Ps

0

0 10 PbTiO3

a Ps

a

a

100

Rhombohedral

a

20

30

40

50

60

70

80

90

100 PbZrO3

PbZrO3 (Mole %)

Phase diagram of the PZT system. (Modied from Ceramics, transducers, in Encyclopedia of Smart Materials.)

factors, respectively. (Pb, Nd)(Ti, Mn, In)O3 ceramics that have a zero temperature coefficient of SAW delay have been developed as superior substrate materials for SAW devices [10]. 9.2.3.2.4 Relaxor Ferroelectrics Relaxor fe rroelectrics d iffer f rom nor mal fe rroelectrics; t hey have broad phase transitions from the paraelectric to t he ferroelectric state, strong frequency dependence of the dielectric constant (i.e., dielectric relaxation) and weak remanent polarization at temperatures close to t he dielectric maximum. Leadbased rel axor m aterials h ave c omplex d isordered p erovskite structures of the general formula Pb(B1, B 2)O3 (B1 = Mg 2+, Zn2+, Sc3+, B 2 = N b5+, Ta 5+, W 6+). The B si te c ations a re d istributed

randomly in the crystal. The characteristic of a relaxor is a broad and fr equency di spersive di electric m aximum. I n a ddition, relaxor-type m aterials suc h a s le ad m agnesium n iobate Pb(Mg1/3Nb2/3)O3–lead titanate PbTiO3 solid solution [PMN-PT] exhibit electrostrictive phenomena that are suitable for actuator applications. Figure 9.20 shows a n electric-eld-induced strain curve t hat w as ob served f or 0. 9PMN-0.1PT a nd re ported b y Cross et al. in 1980 [11]. Note that a strain of 0.1% can be induced by an electric eld a s sm all a s 1 k V/mm a nd t hat h ysteresis i s negligibly small for this electrostriction. Because ele ctrostriction i s t he s econdary ele ctromechanical coupling observed in cubic structures, in principle, the charge is

1.00 800

dij (⫻ 10−12 C/N)

600 d15

Strain (⫻10−3)

FIGURE 9.18

0.50 400 d33 200 −d31 0 48

50

52 54 56 PbZrO3 (Mole %)

58

60

FIGURE 9.19 Piezoelectric d st rain coeffi c ients versus composition for t he PZ T s ystem. ( Modied f rom Ce ramics, t ransducers, i n Encyclopedia of Smart Materials.)

43722_C009.indd 17

−15

−10

−5 0 5 10 Electric field (kV/cm)

15

FIGURE 9.20 Field-induced electrostrictive strain in 0.9PMN-0.1PT. (Modied f rom Ce ramics, t ransducers, i n Encyclopedia of Sm art Materials.)

10/3/2008 9:23:41 PM

9-18

Smart Materials

not i nduced u nder applied s tress. The converse electrostrictive effect, which can be used for sensor applications, means that the permittivity (rst der ivative of p olarization w ith re spect to a n electric eld) is changed by stress. In relaxor ferroelectrics, the piezoelectric effect can be induced under a bias eld, that is, the electromechanical coupling factor kt varies as the applied DC bias eld changes. As the DC bias eld increases, the coupling increases and saturates. The se materials can be used for ultrasonic transducers that are tunable by a bias eld [12]. The re cent d evelopment of si ngle-crystal pie zoelectrics started in 1981, when Kuwata, Uchino, and Nomura rst reported a n e normously l arge e lectromechanical c oupling f actor k33 = 92%–95% a nd a p iezoelectric constant d33 = 1 500 pC/N i n solid-solution si ngle c rystals b etween rel axor a nd no rmal ferroelectrics , Pb(Zn1/3Nb2/3)O3-PbTiO3 [1 3]. A fter a bout 1 0 years, Y. Yamashita et al. (Toshiba) and T. R. Shrout et al. (Penn State) independently reconrmed that these values are true, and much more improved data were obtained in these several years, aimed at medical acoustic applications [14]. Important data have been ac cumulated f or Pb( Mg1/3Nb2/3)O3 [P MN], P b(Zn1/3Nb2/3) O3 [PZN], and binary systems of these materials combined with PbTiO3 ( PMN-PT a nd PZ N-PT) f or ac tuator app lications. Strains as large as 1.7% can be induced practically for a morphotropic phase boundary composition of the PZN-PT solid-solution single crystals. Figure 9.21 shows the eld i nduced strain curve for [ 001] o riented 0. 92PZN-0.08PT [14]. I t i s not able t hat t he highest v alues a re ob served f or a rh ombohedral c omposition only when the single crystal is poled along the perovskite [001] axis, not along the [111] spontaneous polarization axis. 9.2.3.3 Polymers PVDF or PVF2 is a piezoelectric when stretched during fabrication. Thin sheets of the cast polymer are drawn and stretched in the plane of the sheet in at least one direction and frequently also

in t he p erpendicular d irection to c onvert t he m aterial i nto i ts microscopically polar phase. Cr ystallization from a mel t forms the no npolar α-phase, w hich c an b e c onverted i nto a nother polar β-phase by uniaxial or biaxial drawing; these dipoles are then reoriented by electric poling. Large sheets can be manufactured a nd t hermally f ormed i nto c omplex s hapes. C opolymerization of vinilydene diuoride with triuoroethylene (TrFE) results i n a r andom copolymer (PVDF-TrFE) t hat has a s table, polar β phase. This polymer does not need to be stretched; it can be poled directly as formed. The thickness-mode coupling coefcient of 0.30 has been reported. Such piezoelectric polymers are used for directional microphones and ultrasonic hydrophones. 9.2.3.4 Composites Piezocomposites co mprised o f p iezoelectric ce ramics a nd polymers are promising materials because of excellent tailored properties. The geometry of two-phase composites can be classied ac cording to t he c onnectivity of e ach ph ase (0, 1, 2 , or 3 dimensionality) into 10 structures; 0-0, 0-1, 0-2, 0-3, 1-1, 1-2, 1-3, 2-2, 2-3, and 3-3 [15]. A 1-3 piezocomposite, or PZT–rod/ polymer–matrix composite is a most promising candidate. The advantages of this composite are high coupling factors, low acoustic i mpedance (square ro ot of t he product of its den sity and elastic stiff ness), a go od match to w ater or human t issue, mechanical exibility, a broad bandwidth in combination with a low mechanical quality factor, and the possibility of making undiced a rrays by s tructuring only t he ele ctrodes. The thickness-mode ele ctromechanical c oupling o f t he c omposite c an exceed the kt (0.40–0.50) of the constituent ceramic and almost approaches the value of the rod-mode electromechanical coupling, k33 (0.70–0.80), of that ceramic [16]. The acoustic match to tissue or water (1.5 Mrayls) of typical piezoceramics (20–30 Mrayls) i s s ignicantly i mproved b y f orming a c omposite structure, that is, by replacing a heavy, stiff ceramic by a light, soft p olymer. P iezoelectric c omposite m aterials a re e specially useful for underwater sonar and medical diagnostic ultrasonic transducers.

2.0 d33 ~ 480 pC/N

9.2.3.5

Tetragonal

Strain (%)

1.5

1.0

0.5

0.0

0

20

60 80 40 Electric field (kV/cm)

100

120

FIGURE 9.21 Field-induced strain curve for [001] oriented 0.92PZN0.08PT. (Modied from Encyclopedia of Smart Materials.)

43722_C009.indd 18

Thin Films

Both zinc oxide (ZnO) and aluminum nitride (AlN) are simple binary c ompounds t hat h ave W urtzite-type s tructures, w hich can be sputter-deposited in a c-axis-oriented thin lm on a variety of substrates. ZnO has reasonable piezoelectric coupling and its thin lms are widely used in bulk acoustic and SAW devices. The fabrication of h ighly c-axis oriented ZnO lms has been extensively studied and developed. The performance of ZnO devices is, h owever, l imited d ue to t heir sm all p iezoelectric c oupling (20%–30%). PZT thin lms are expected to exhibit higher piezoelectric p roperties. A t p resent, t he g rowth o f PZ T t hin  lm is being carried out for use in microtransducers and microactuators. A s eries o f t heoretical c alculations o n p erovskite-type ferroelectric crystals suggests that large d and k values of magnitudes similar to t hose of PZN-PT can also be expected in PZT. Crystal orientation de pendence of piezoelectric properties w as

10/3/2008 9:23:42 PM

9-19

Piezoelectric and Electrostrictive Ceramics Transducers and Actuators

d33 Ps

E1

Ps

Ps

E1

E E2

Strain

d 33eff

E

d15

Ps

E2

FIGURE 9.22 Materials.)

Principle of the enhancement in electromechanical couplings in a perovskite piezoelectric. (Modied from Encyclopedia of Smart

phenomenologically c alculated for c ompositions a round t he morphotropic phase boundary of PZT [17]. The maximum longitudinal p iezoelectric c onstant d33 ( four to  ve t imes t he enhancement) a nd t he ele ctromechanical c oupling f actor k33 (more than 90%) in the rhombohedral composition were found at angles of 57° and 51°, respectively, canted from the spontaneous p olarization d irection [111], w hich c orrespond ro ughly to the perovskite [100] axis. Figure 9.22 shows the principle of the enhancement in electromechanical couplings. Because the shear coupling d15 is the highest in perovskite piezoelectric crystals, the applied eld should be canted from the spontaneous polarization direction to obtain the maximum s train. Ep itaxially g rown, [ 001] o riented t hin/thick lms using a rhomboherial PZT composition reportedly enhance the effective piezoelectric constant by—four to ve times.

9.2.4

Applications of Piezoelectricity

Piezoelectric materials can provide coupling between electrical and mechanical energy and thus have been extensively used in a variety o f ele ctromechanical de vices. The dir ect p iezoelectric effect is most obviously used to generate charge or high voltage in applications such as the spark ignition of gas in space heaters, cooking stoves, and cigarette lighters. Using the converse effect, small mechanical displacements and vibrations can be produced in actuators by applying an electric eld. Acoustic and ultrasonic vibrations can be generated by an alternating  eld tuned at t he mechanical re sonance f requency of a p iezoelectric de vice, a nd can b e de tected by a mplifying t he  eld gener ated by v ibration incident o n t he m aterial, w hich i s u sually u sed f or u ltrasonic transducers. A nother i mportant app lication o f p iezoelectricity is f requency c ontrol. The application of piezoelectric m aterials ranges o ver m any t echnology f ields, i ncluding u ltrasonic

43722_C009.indd 19

transducers, ac tuators a nd USMs; ele ctronic c omponents such as resonators, wave  lters, delay lines, SAW devices, and transformers a nd h igh-voltage app lications; ga s ig nitors, u ltrasonic cleaning, a nd m achining. P iezoelectric-based s ensors, f or instance, ac celerometers, a utomobile k nock s ensors, v ibration sensors, s train g ages, a nd  ow meters have been developed because pressure and vibration can be directly sensed as electric signals t hrough t he p iezoelectric e ffect. E xamples o f t hese applications are given in the following sections. 9.2.4.1

Pressure Sensor/Accelerometer/Gyroscope

The ga s ig niter i s one of t he ba sic applications of piezoelectric ceramics. Very high voltage generated in a piezoelectric ceramic under applied mechanical stress can cause sparking a nd ig nite the gas. Piezoelectric ceramics can be employed as stress sensors and acceleration sensors, because of their “direct piezoelectric effect.” Kistler (Switzerland) i s m anufacturing a 3 -D s tress s ensor. By combining a n app ropriate n umber o f qu artz c rystal p lates (extensional a nd shear t ypes), t he multilayer de vice c an de tect three-dimensional stresses [18]. Figure 9.23 shows a cylindrical gyroscope commercialized by NEC-Tokin (Japan) [19]. The cylinder has six divided electrodes; one pa ir i s u sed to e xcite t he f undamental b ending v ibration mode and the other two pairs are used to detect the acceleration. When rotation acceleration is applied around the axis of this gyro, the voltage generated on the electrodes is modulated by the Coriolis force. By subtracting the signals between the two sensor electrode pairs, a voltage directly proportional to the acceleration can be obtained. This type of gyroscope has been widely installed in handheld video cameras to monitor the inevitable hand vibration during operation and to compensate for it electronically on a display by using the sensed signal.

10/3/2008 9:23:42 PM

9-20

Smart Materials Piezoelectric element Matching layer Backing

Vibrator

Lead

Holder Ultrasonic beam

Support

FIGURE 9.23 Cylindrical gyroscope commercialized by NEC-Tokin (Japan). (Modied from Encyclopedia of Smart Materials.)

9.2.4.2 Ultrasonic Transducer One of the most important applications of piezoelectric materials is based on ultrasonic echo  eld [20,21]. Ultrasonic transducers convert electrical energy into a mechanical form when generating an acoustic pulse and convert mechanical energy into an electrical sig nal w hen de tecting i ts e cho. N owadays, p iezoelectric transducers a re b eing u sed i n me dical u ltrasound f or c linical applications t hat r ange f rom d iagnosis to t herapy a nd su rgery. They are also used for underwater detection, such as sonars and sh nders, and nondestructive testing. The u ltrasonic tr ansducers o ften op erate i n a p ulse-echo mode. The transducer converts electrical input into an acoustic wave o utput. The t ransmitted w aves p ropagate i nto t he b ody, and echoes are generated that travel back to be received by the same transducer. These echoes vary in intensity according to the type of tissue or body structure, and thereby create images. An ultrasonic i mage re presents t he me chanical p roperties o f t he tissue, such as density and elasticity. We can recognize anatomical structures in an ultrasonic image because the organ boundaries and uid-to-tissue interfaces are easily discerned. The ultrasonic imaging can also be done in real time. This means that we can follow rapidly moving structures such as heart without motional distortion. In addition, ultrasound is one of the safest diagnostic imaging techniques. It does not use ionizing radiation like x-rays and t hus i s ro utinely u sed f or f etal a nd ob stetrical i maging. Useful a reas for u ltrasonic i maging i nclude cardiac structures, the v ascular s ystem, t he f etus, a nd a bdominal o rgans suc h a s liver and kidney. In brief, it is possible to s ee inside the human body by using a beam of ultrasound without breaking the skin. There a re v arious t ypes o f t ransducers u sed i n u ltrasonic imaging. Mechanical s ector t ransducers c onsist of si ngle, rel atively large resonators that provide images by mechanical scanning such as wobbling. Multiple element array transducers permit the imaging s ystems to ac cess d iscrete el ements i ndividually a nd

43722_C009.indd 20

Input pulse

FIGURE 9.24 Geometry of t he fundamental transducer for a coustic imaging. (Modied from Encyclopedia of Smart Materials.)

enable electronic focusing in the scanning plane at various adjustable p enetration de pths b y u sing ph ase del ays. The two basic types of array transducers are linear and phased (or sector). Linear array transducers are used for radiological and obstetrical e xaminations, a nd ph ased a rray t ransducers a re u seful f or cardiological applications where positioning between the ribs is necessary. Figure 9.24 shows t he geometry of t he basic u ltrasonic t ransducer. The transducer is composed mainly of matching, piezoelectric material, and backing layers [22]. One or more matching layers are used to increase sound transmissions into tissues. The backing is attached to the transducer rear to damp the acoustic return wave and to reduce the pulse duration. Piezoelectric materials are used to g enerate a nd de tect u ltrasound. I n g eneral, b roadband transducers should be used for medical u ltrasonic i maging. The broad ba ndwidth re sponse c orresponds to a s hort p ulse leng th that results in better axial resolution. Three factors are important in de signing b road ba ndwidth t ransducers. The rst is acoustic impedance m atching, t hat i s, e ffectively co upling t he aco ustic energy to the body. The second is high electromechanical coupling coefficient o f t he t ransducer. The t hird i s ele ctrical i mpedance matching, t hat is, effectively coupling electrical energy from the driving electronics to the transducer across the frequency range of interest. The op erator of pu lse-echo t ransducers i s b ased on t he thickness mo de re sonance o f t he p iezoelectric t hin p late. The thickness mode coupling coefficient, kt, is related to the efficiency of converting electric energy into acoustic and vice versa. Further, a low planar mode coupling coefficient, kp, is benecial for limiting e nergies fr om b eing e xpended in a n onproductive l ateral mode. A l arge d ielectric c onstant i s ne cessary to ena ble a go od electrical impedance match to the system, especially in tiny piezoelectric sizes. Table 9 .5 c ompares t he p roperties o f u ltrasonic t ransducer materials [7,23] Ferroelectric ceramics, such as PZT and modied PT, a re a lmost u niversally u sed a s u ltrasonic t ransducers. The success of ceramics is due to their very high electromechanical

10/3/2008 9:23:42 PM

Piezoelectric and Electrostrictive Ceramics Transducers and Actuators TABLE 9.5 Comparison of the Properties of Ultrasonic Transducer Materials PZT Ceramic PVDF Polymer kt Z (Mrayls) e33T/e0 tan d (%) Qm r (g/cm3)

PZT–Polymer Composite

ZnO Film 0.20–0.30 35 10

0.45–0.55 20–30 200–5000

0.20–0.30 1.5–4 10

0.60–0.75 4–20 50–2500

<1

1.5–5

<1

<1

10–1000 5.5–8

5–10 1–2

2–50 2–5

10 3–6

coupling coeffi cients. I n pa rticular, s oft PZ T c eramics suc h a s PZT-5A and 5H type compositions are most widely used because of t heir e xceedingly h igh c oupling p roperties a nd b ecause t hey can be relatively easily tailored, for instance, in the wide dielectric constant range. On the other hand, modied PTs such as samarium -doped m aterials h ave h igh p iezoelectric a nisotropy: t he p lanar coupling factor kp is much less than the thickness coupling factor kt. Because the absence of lateral coupling leads to reduced interference from spurious lateral resonances in longitudinal oscillators, this is very useful in high-frequency array transducer applications. O ne d isadvantage to PZ T a nd ot her le ad-based ceramics is t heir large acoustic i mpedance (approximately 30 kg m−2 s−1 (Mrayls) compared to body tissue (1.5 Mrayls). Single or multiple matching layers of intermediate impedances need to be used in PZT to improve acoustic matching. On t he ot her h and, p iezoelectric p olymers, suc h a s PV DFtriuoroethylene, h ave m uch lo wer ac oustic i mpedance ( 4–5 Mrayls) t han c eramics a nd t hus m atch s oft t issues b etter. However, piezopolymers are less sensitive than the ceramics and they h ave rel atively lo w d ielectric c onstants t hat re quire l arge drive vol tage a nd g iving p oor no ise p erformance d ue to m ismatching of electrical impedance. Piezoelectric ceramic/polymer composites are alternatives to ceramics a nd p olymers. P iezocomposites t hat h ave 2- 2 o r 1 -3 connectivity are commonly used in ultrasonic medical applications. They c ombine t he low ac oustic i mpedance adv antage o f polymers with the high sensitivity and low electrical impedance advantages of ceramics. The design frequency of a transducer depends on the penetration depth required by the application. Resolution is improved as frequency increases. Although a high-frequency transducer can produce a h igh-resolution i mage, h igher f requency ac oustic energy is more readily attenuated by the body. A lower frequency transducer is used as a compromise when imaging deeper structures. Most of me dical u ltrasound i maging s ystems op erate i n the f requency range f rom 2 to 1 0 MHz a nd c an resolve objects approximately 0.2–1 mm in size. At 3.5 MHz, imaging to a depth of 10–20 cm is possible, and at 50 MHz, increased losses limit the depth to le ss t han 1 cm. H igher-frequency t ransducers (10–50 MHz) are used for endoscopic imaging and for catheterbased in travascular im aging. U ltrasonic mi croscopy i s b eing done at frequencies higher than 100 MHz. The op erating f requency of the transducer is directly related to the thickness and

43722_C009.indd 21

9-21

velocity o f s ound i n p iezoelectric m aterials em ployed. A s f requency increases, resonator thickness decreases. For a 3.5 MHz transducer, PZ T c eramic t hickness m ust b e ro ughly 0.4 mm. Conventional ceramic t ransducers, such as PZT, a re limited to frequencies below 80 MHz because of the difficulty of fabricating thinner devices [24]. Piezoelectric thin- lm transducers such as ZnO have to be used for microscopic applications (at frequencies higher than 100 MHz, corresponding to a t hickness of less than 20 µm) [25]. 9.2.4.3

Resonator and Filter

When a p iezoelectric b ody v ibrates at i ts re sonant f requency, i t absorbs considerably more energy than at other frequencies, resulting i n a f all o f t he i mpedance. This p henomenon enab les u sing piezoelectric materials as wave lters. A  lter is required to pass a certain selected frequency band or to stop a g iven band. The band width o f a  lter fabricated f rom a p iezoelectric ma terial i s de termined by the square of the coupling coefficient k. Quartz crystals that h ave ve ry lo w k v alue o f a bout 0.1 c an pa ss v ery na rrow frequency ba nds o f app roximately 1% o f t he c enter re sonance frequency. On the other hand, PZT ceramics whose planar coupling coefficient of about 0.5 can easily pass a ba nd of 10% of the center resonance frequency. The sharpness of the passband depends on the mechanical quality factor Qm of the materials. Quartz also has a very h igh Qm of a bout 106, w hich re sults i n a s harp c utoff of t he passband and well-dened frequency of the oscillator. A simple resonator is a thin disk electroded on its plane faces and v ibrating r adially f or app lications i n  lters w hose c enter frequency ranges from 200 kHz to 1 MHz and whose bandwidth is s everal p ercent o f t he c enter f requency. The di sk di ameter must b e a bout 5. 6 mm for a f requency of 4 55 kHz. However, i f the required f requency is h igher t han 10 MHz, ot her modes of vibration such as the thickness extensional mode are exploited, because of its smaller size disk. Trapped-energy type lters made from PZT ceramics have been widely used in the intermediate frequency (IF) range, for example, 10.7 MHz for FM radio receivers and t ransmitters. By em ploying t he t rapped-energy ph enomenon, the overtone frequencies are suppressed. The plate is partly covered w ith ele ctrodes o f a s pecic a rea a nd t hickness. The fundamental f requency o f t he t hickness mo de b eneath t he electrode is less than that of the unelectroded portion because of the extra inertia of the electrode mass. The longer wave characteristic of t he electrode region cannot propagate in t he unelectroded re gion. The h igher-frequency o vertones c an p ropagate into t he unelectroded region. This is called the trapped-energy principle. Figure 9.25 shows a s chematic drawing of a t rappedenergy  lter. In t his structure, t he top ele ctrode is split so t hat coupling b etween t he t wo pa rts i s e fficient o nly a t r esonance. More stable  lters suitable for telecommunication systems have been made from single crystals such as quartz or LiTaO3. 9.2.4.4

Piezoelectric Transformer

The t ransfer o f v ibration energ y f rom o ne s et o f ele ctrodes to another on a piezoelectric ceramic body can be used to transform voltage. The device is called a piezoelectric transformer. Recently,

10/3/2008 9:23:43 PM

9-22

Smart Materials

Electrode

SAW

Input

Output

Ceramic plate Interdigital electrode Top

Bottom Piezoelectric substrate

FIGURE 9.25 Trapped-energy  lter. (Modied from Encyclopedia of Smart Materials.)

office automation equipment that has a liquid crystal display such as notebook-type personal computers and car navigation systems has been successfully commercialized. This equipment that uses a liquid crystal display requires a very thin transformer without electromagnetic no ise to s tart t he g low o f a  uorescent backlamp. This application has recently accelerated the development of t he p iezoelectric t ransformers. F igure 9 .26 s hows t he ba sic structure, where two differently poled parts coexist in one piezoelectric plate. The plate has electrodes on half of its major faces and on an edge. The plate is then poled in its thickness direction at one end and parallel to the long axis over most of its length. A low-voltage AC supply is applied to the large-area electrodes at a frequency t hat e xcites a leng th e xtensional mo de re sonance. Then, a high-voltage output can be taken from the small electrode and from one of the larger electrodes. Following the proposal by Rosen men tioned b efore, p iezoelectric t ransformers o f s everal different s tructures h ave b een re ported [26]. A m ultilayer t ype transformers a re proposed to i ncrease t he voltage s tep-up r atio [27]. The i nput pa rt has a m ultilayer structure a nd has i nternal electrodes, and the output electrodes are formed at t he side su rface of the half of the rectangular plate. This transformer uses the piezoelectric longitudinal mode for the input and output parts. 9.2.4.5

Saw Device

SAW, also called a Rayleigh wave, is composed of a coupling between longitudinal and shear waves in which the SAW energy is conned near the surface. An associated electrostatic wave exists for a SAW on a piezoelectric substrate that allows electroacoustic coupling via a transducer. The advantages of SAW technology are that a wave can be electroacoustically accessed and trapped at the substrate su rface a nd i ts v elocity i s app roximately 1 04 tim es slower than an electromagnetic wave. The SAW wavelength is of the s ame ord er of m agnitude a s line d imensions t hat c an be

FIGURE 9 .27 Typical SAW bid irectional  lter t hat c onsists of t wo interdigital tr ansducers. ( Modied from Encyclopedia of Sm art Materials.)

photolithographically p roduced, a nd t he leng ths f or b oth s hort and long delays are achievable on reasonable size substrates [28,29]. There i s a v ery b road r ange o f c ommercial s ystem app lications, i ncluding f ront-end a nd I F  lters, c ommunity a ntenna television (CATV), and VCR components, synthesizers, analyzers, and navigators. In SAW transducers, nger electrodes provide the ability to sample or tap the wave, and the electrode gap gives the relative delay. A SAW  lter is composed of a minimum of two transducers. A s chematic of a si mple SAW bidirectional  lter is shown in Figure 9.27. A bidirectional transducer radiates energy e qually f rom e ach side o f t he t ransducer. E nergy not received is absorbed to eliminate spurious reection. Various materials are currently being used for SAW devices. The most p opular si ngle-crystal SAW m aterials a re l ithium n iobate and l ithium t antalate. The m aterials h ave d ifferent properties depending on their cuts and the direction of propagation. The fundamental pa rameters a re t he SAW v elocity, t he tem perature c oefficients of delay (TCD), t he electromechanical coupling factor, a nd the propagation loss. SAWs can be generated and detected by spatially p eriodic, i nterdigital ele ctrodes o n t he p lane su rface o f a piezoelectric plate. A periodic electric eld is produced when an RF source is connected to the electrode, thus permitting piezoelectric coupling to a traveling surface wave. If an RF source of a frequency f is applied to an electrode whose periodicity is p, energy conversion from an electrical to mechanical form will be maximum when f = f0 =

Vs , (9 p

.20)

where Vs is the SAW velocity f0 is the center frequency of the device

Low voltage input

High voltage output

FIGURE 9.26

43722_C009.indd 22

Rosen-type piezoelectric transformer. (Modied from Encyclopedia of Smart Materials.)

10/3/2008 9:23:43 PM

9-23

Piezoelectric and Electrostrictive Ceramics Transducers and Actuators

SAW velocity is an important parameter that determines the center frequency. Another important parameter for many applications is the temperature sensitivity. For example, the temperature stability of the center frequency of SAW bandpass lters is a direct function of tem perature co efficient f or t he v elocity a nd del ay t ime o f t he material used. The rst-order TCD time is given by ⎛ 1 ⎞ ⎛ d t ⎞ = ⎛ 1 ⎞ ⎛ d L ⎞ − ⎛ 1 ⎞ ⎛ dV s ⎞ , ⎝ t ⎠ ⎜⎝ dT ⎟⎠ ⎝ L ⎠ ⎜⎝ dT ⎟⎠ ⎜⎝ Vs ⎟⎠ ⎜⎝ dT ⎟⎠ (9

.21)

where t = L/Vs is the delay time L is the SAW propagation length The surface wave coupling factor, ks2, is dened in terms of the change in SAW velocity that occurs when the wave passes across a surface coated by a thin massless conductor, so that the piezoelectric  eld a ssociated wi th th e w ave i s e ffectively shortedcircuited. The coupling factor, ks2, is expressed by ks2 = 2

(V f − V m ) , (9 Vf

(128°-rotated-Y-cut a nd X-propagation)—LiNbO3 an d X-112°Y (X-cut a nd 112°-rotated-Y-propagation)—LiTaO3 have bee n ex tensively employed as SAW substrates for VIF  lters. ZnO t hin  lms, c-axis oriented and deposited on a f used quartz, glass, or sapphire substrate, have also been commercialized for SAW devices. 9.2.4.6

Actuators

Currently another important application of piezoelectric materials exists in the actuator eld [30]. Using t he c onverse p iezoelectric effect, a small displacement can be produced by applying an electric eld to a piezoelectric material. Vibrations can be generated by applying an alternating electric eld. There is a demand in advanced precision eng ineering for a v ariety of t ypes of ac tuators t hat can adjust position precisely (micropositioning devices), suppress noise vibrations (dampers), and drive objects dynamically (USMs). These devices are used in areas, including optics, astronomy,  uid control, and precision machinery. Piezoelectric strain and electrostriction induced by an electric eld are used for actuator applications. Figure 9 .28 s hows t he de sign c lassication o f c eramic a ctuators. Simple devices composed of a disk or a multilayer type use the strain

.22) z

where Vf is the free surface wave velocity Vm is the velocity on the metallized surface

Multilayer

In SAW applications, the value of ks2 relates to the maximum bandwidth obtainable and the amount of signal loss between input and output that determines the fractional bandwidth versus minimum insertion loss for a given material and a  lter. Propagation loss, one of the major factors that determines the insertion loss of a device, is caused by wave scattering by crystalline defects and surface irregularities. Materials that have high electromechanical coupling factors combined with small TCD time are likely to b e required. The free surface velocity, V0, of t he m aterial i s a f unction of c ut a ngle a nd propagative d irection. The TC D i s a n i ndication o f t he f requency shift expected from a transducer due to a temperature change and is also a f unction of the cut angle and the propagation direction. The substrate is chosen on the basis of the device’s design specications for operating temperature, fractional bandwidth, and insertion loss. Table 9.6 shows some important material parameters of representative SAW materials. Piezoelectric single crystals such as 128°Y–X

v

z

Bimorph

v

z

Moonie

z v

FIGURE 9 .28 Structures o f ce ramic a ctuators. ( Modied from Encyclopedia of Smart Materials.)

TABLE 9.6 Material Parameters for Representative SAW Materials Material Single crystal

Ceramic Thin lm

43722_C009.indd 23

Quartz LiNbO3 LiTaO3 Li2B4O7 PZT-In(Li3/5W2/5)O3 (Pb,Nd)(Ti, Mn, In)O3 ZnO/glass ZnO/Sapphire

Cut–Propagation Direction ST—X 128°Y—X X112°—Y (110) —<001>

k2 (%)

TCD (ppm/C)

V0 (m/s)

er

0.16 5.5 0.75 0.8 1.0 2.6 0.64 1.0

0 −74 −18 0 10 <1 −15 −30

3158 3960 3290 3467 2270 2554 3150 5000

4.5 35 42 9.5 690 225 8.5 8.5

10/3/2008 9:23:44 PM

9-24

Smart Materials

induced in a ceramic by the applied electric eld d irectly. Complex devices do not use the induced strain directly but use the amplied displacement t hrough a s pecial magnication mechanism such a s unimorph, bimorph, and moonie. The most popularly used multilayer and bimorph types have the following characteristics: The multilayer t ype do es not h ave a l arge d isplacement ( 10 µm), b ut h as advantages i n gener ation force (1 kN), re sponse s peed (10 µs), l ifetime ( 1011 c ycles), a nd t he ele ctromechanical c oupling f actor k33 (0.70). The bimorph type has a large displacement (300 µm), but has disadvantages in generation force (1 N), response speed (1 ms), lifetime (108 cycles), and the electromechanical coupling factor keff (0.10). For instance, in a 0.65 PMN–0.35 PT multilayer actuator with 99 layers of 100 µm thick sheets (2 × 3 × 10 mm3), a 8.7 µm displacement is generated by a 100 V voltage, accompanied by a s light hysteresis. The t ransmitted re sponse o f t he i nduced d isplacement a fter the application of a re ctangular voltage is as quick as 10 µs. The multilayer has a eld-induced strain of 0.1% along the length [30]. Unimorph and bimorph devices are dened by the number of piezoelectric c eramic p lates: o nly o ne c eramic p late i s b onded onto an elastic shim, or two ceramic plates are bonded together. The bimorph c auses b ending de formation b ecause e ach piezoelectric plate bonded together produces extension or contraction in an electric eld. In general, there are two types of piezoelectric bimorphs: t he a ntiparallel p olarization t ype a nd t he pa rallel polarization type, as shown in Figure 9.29. Two poled piezoelectric

V (a)

V

(b)

FIGURE 9.29 Two t ypes of pie zoelectric bi morphs: (a) a ntiparallel polarization t ype a nd ( b) p arallel p olarization t ype. ( Modied from Encyclopedia of Smart Materials.)

plates t/2 thick and L long are bonded so that their polarization directions opposite or parallel to each other. In the cantilever bimorph c onguration w here o ne en d i s c lamped, t he t ip d isplacement d z under an applied voltage V is d z = (3/2)d 31 (L2/t 2 ) V (antiparallel type) (9

.23)

d z = 3d 31 ( L2/t 2 ) V (parallel type). (9

.24)

The resonance frequency fr for both types is given by f r = 0.16 /t L2 ( rs11E )−1/2 , (9

.25)

where r is the density s11E is the elastic compliance A metallic s heet ( called a s him) i s o ccasionally s andwiched between the two piezoelectric plates to increase the reliability; the structure can be maintained even if the ceramics fracture. Using the bimorph structure, a large magnication of the displacement is e asily ob tainable. H owever, t he d isadvantages i nclude a lo w response speed (1 kHz) and low generative force [30]. A c omposite ac tuator s tructure c alled “ moonie” h as b een developed to amplify the small displacements induced in piezoelectric c eramics [31]. The moonie consists of a thin single or multilayer element and two metal plates that have narrow moonshaped cavities bonded together. This device has characteristics intermediate between the conventional multilayer and bimorph actuators; i t h as a n o rder o f m agnitude l arger d isplacement (100 µm) t han t he multilayer, and much larger generative force (100 N) and quicker response (100 µs) than the bimorph. Some e xamples of appl ications of pie zoelectric a nd e lectrostrictive ac tuators a re de scribed h ere. The piezoelectric impact dot-matrix printer is the rst mass-produced device that uses multilayer ceramic actuators (Figure 9.30) [32]. The advantage of a piezoelectric printer head compared to conventional magnetic type are low energy consumption, low heat generation, and fast printing speed. Longitudinal multilayer actuators do not have a

Head element Platen Paper Ink ribbon Guide Piezoelectric actuator

Wire Wire (a)

FIGURE 9.30

43722_C009.indd 24

Wire guide

Stroke amplifier (b)

Impact dot matrix printer head commercialized by NEC (Japan). (Modied from Encyclopedia of Smart Materials.)

10/3/2008 9:23:45 PM

9-25

Piezoelectric and Electrostrictive Ceramics Transducers and Actuators

large displacement and thus a suitable displacement magnication mechanism is necessary. The displacement induced in a m ultilayer ac tuator p ushes up t he f orce p oint a nd i ts d isplacement magnication i s c arried o ut t hrough h inge le vers to gener ate a large wire stroke. When the displacement in the piezoactuator is 8 µm, the wire stroke of 240 µm can be obtained; the magnication rate is 30 times. Bimorph structures are commonly used for VCR head tracking actuators because of their large displacements. An auto tracking scan s ystem u ses t he p iezoelectric ac tuators s o t hat t he h ead follows t he recording t rack even d riven i n both still a nd qu ick modes [33]. As can be anticipated, the bimorph drive is inevitably accompanied by a rot ational motion. A s pecial mechanism has to be employed to obtain perfectly linear translational motion. Piezoelectric p umps f or ga s o r l iquid t hat u se t he a lternating bending motion of the bimorph have been developed for intravenous d rip i njection i n hospitals a nd for medication d ispensing for chemotherapy, chronic pain, and diabetes [34]. Piezoelectric fans f or c ooling ele ctronic c ircuits a re m ade f rom a pa ir o f bimorphs that are driven out of phase so as to blow effectively [35]. Fu rthermore, pie zobimorph t ype c amera s hutters h ave been widely commercialized by Minolta [36]. Lenses a nd m irrors i n opt ical c ontrol s ystems re quire micropositioning and even the shapes of mirrors are adjusted to correct image distortions. For instance, a space-qualied active mirror called articulating fold mirror utilizes six PMN electrostrictive multilayer actuators to position and tilt a mirror tip precisely to correct the focusing aberration of the Hubble space telescope [37]. Piezoelectric actuators are also useful for vibration suppression systems o f a n a utomobile. A n ele ctronic c ontrolled s hock absorber w as de veloped b y T oyota [ 38]. P iezoelectric s ensors that detect road roughness are composed of ve layers of 0.5 mm thick PZ T d isks. The actuator is made of 88 layers of 0.5 mm thick disks. Under 500 V, it generates about 50 µm displacement, which is magnied by 40 times by a piston and plunger pin combination. This s troke p ushes t he c hange v alve o f t he d amping force do wn a nd t hen op ens t he b ypass o il ro ute, le ading to decrease i n  ow r esistance. Th is e lectronically co ntrolled shock absorber h as b oth c ontrollability a nd p rovides c omfort simultaneously. The U .S. A rmy i s i nterested i n de veloping a roto r c ontrol system for helicopters, because a slight change in the blade angle drastically en hances c ontrollability. F igure 9 .31 s hows a b earingless rotor exbeam that has piezoelectric strips attached [39]. Various types of PZT-sandwiched beam structures have been investigated f or suc h a  exbeam app lication a nd f or ac tive vibration control. In order to increase the diesel engine efficiency, high-pressure fuel a nd qu ick i njection c ontrol a re re quired. I n o ne eng ine cycle (typically 60 Hz), t he multiple i njections should be realized in a v ery sharp shape for t he diesel engine. For t his purpose, piezoelectric actuators, specically multilayer types, are adopted. The highest reliability of these devices at a n elevated temperature ( 150°C) f or a lo ng p eriod ( 10 ye ars) h as b een

43722_C009.indd 25

Lag pin Piezoelectric crystals

Torque tube Hub

Flexbeam Pitch link Blade

FIGURE 9.31 Bearingless rotor  exbeam with attached piezoelectric strips. (Modied from Encyclopedia of Smart Materials.)

achieved. The so-called common-rail type injection valves have been widely commercialized by Siemens, Bosch, and Denso Corp. (Figure 9.32) [40]. 9.2.4.7

USMs

An USM is an example of piezoelectric actuator that uses resonant vibration. Linear motion in USMs is obtained by frictional force from the elliptical vibration. The motor consists of a highfrequency power supply, a v ibrator, and a s lider. The vibrator is composed of a p iezoelectric d riving c omponent a nd a n el astic vibratory part, and the slider is composed of an elastic moving part a nd a f riction c oat. The c haracteristics o f U SMs a re lo w speed a nd h igh to rque c ompared to t he h igh s peed a nd lo w torque of conventional electromagnetic motors [41,42]. USMs a re c lassied i nto t wo t ypes: t he s tanding-wave t ype and the propagating-wave type. The displacement of a standing wave is expressed by

Piezoelectric actuator

Control valve

Displacement amplification unit Injector body Nozzle

FIGURE 9.32 Common rail type diesel injection valve with a pie zoelectric multilayer actuator. (Courtesy of Denso Corp.) (Modied from Encyclopedia of Smart Materials.)

10/3/2008 9:23:47 PM

9-26

Smart Materials Oscillator Rotor

Vibratory piece

FIGURE 9 .33 Vibratory c oupler t ype USM . ( Modied from Encyclopedia of Smart Materials.)

x s (x , t ) = A cos(kx ) cos(wt ), (9

.26)

for a propagative wave displacement is given by x p (x , t ) = A cos(kx − wt ) = A cos(kx ) cos(w t ) + A cos(kx − p /2) cos(w x − p /2).

(9.27)

Slider Moving direction

Propagation direction

Elliptical particle motion Elastic body

FIGURE 9.34 Principle of the propagating wave type USM. (Modied from Encyclopedia of Smart Materials.)

A p ropagating w ave c an b e gener ated b y sup erimposing t wo standing waves whose phases differ from each other by 90° in both time and space. The s tanding-wave t ype i s s ometimes called a v ibratory coupler or a “woodpecker” type; a v ibratory piece is connected to a piezoelectric driver, and the tip portion generates  at-elliptic move ment (Fi gure 9 .33). The vibratory piece is attached to a rotor or a slider at a slight cant angle. The standing-wave t ype h as h igh e fficiency a nd up to 9 8% o f t he theoretical value. However, a problem of this type is lack of control i n b oth c lockwise a nd c ounterclockwise d irections. The principle of the propagation type is shown in Figure 9.34. In the propagating-wave type, also called the “sur ng-type,” a surface particle of the elastic b ody d raws a n e lliptical l ocus due to coupling of the longitudinal and transverse waves. This type generally requires two vibrational sources to generate one propagating w ave; t his le ads to lo w e fficiency (not more than 50%), but is controllable in both t he rotational directions. An ultrasonic rotary motor is successfully used in autofocusing camera t o pr oduce pre cise r otational d isplacements. The advantages of this motor over the conventional electromagnetic motor are silent drive (inaudible), thin motor design, and energy savings. Tiny conventional electromagnetic motors, smaller t han 1 cm, that have sufficient energy efficiency are rather difficult to produce. Therefore, the USM is gaining widespread attention. USMs whose efficiency is independent of size are superior in the minimotor area. The Penn St ate University a nd S amsung E lectromechanics developed a z oom me chanism f or a c ellular ph one c amera w ith two microrotary motors [43–45]. A micro motor called “metal tube type” consisting of a metal hollow cylinder and two PZT rectangular plates is used as basic microactuators (see Figure 9.35a). When we drive one of the PZT plates, Plate X, a bending vibration is excited basically along x-axis. However, because of an asymmetrical mass (Plate Y), another hybridized bending mode is excited with some phase lag along y-axis, leading to an elliptical locus in a clockwise direction, like a “Hula-Hoop” motion. The rotor of this motor is a c ylindrical rod w ith a pa ir of stainless ferrule pressed with a s pring. The a ssembly is shown i n Figure 9.35b. The metal

Y

Elastic hollow cylinder y⬘

x⬘

Plate X X

Plate Y (a)

(b)

FIGURE 9.35 “Metal tube” motor that uses a metal tube and two rectangular PZT plates. (a) Schematic structure and (b) photo of a world smallest motor (1.5 mmφ). (Modied from Encyclopedia of Smart Materials.)

43722_C009.indd 26

10/3/2008 9:23:47 PM

Piezoelectric and Electrostrictive Ceramics Transducers and Actuators

cylinder motor 2.4 mm in diameter and 12 mm in length was driven at 62.1 kHz in both rotation directions. A no-load speed of 1800 rpm a nd a n output torque up to 1 .8 mN m a re obtained for rotation in both directions under an applied voltage of 80 V. Rather high maximum efficiency of about 28% for t his small motor is a noteworthy feature. The smallest camera in the world adopted two micro-USMs with 2.4 mm diameter and 14 mm length to control zooming and focusing lenses independently in conjunction w ith screw mechanisms.

References B. Jaffe, W. Cook, and H. Jaffe, Piezoelectric Ceramics, London: Academic Press, 1971. 2. W. G. Cady, Piezoelectricity, New York: McGraw-Hill, revised edition by Dover, 1964. 3. F. J ona a nd G. S hirane, Ferroelectric C rystals, L ondon: Pergamon Press, 1962. 4. M. E. Lines a nd A. M. Glass, Principles and Applications of Ferroelectric Materials, Oxford: Clarendon Press, 1977. 5. IEEE S tandard o n P iezoelectricity, N ew York: IEEE, I nc., 1978. 6. Landold a nd B oernstein, N umerical da ta a nd f unctional relationships. Science and Technology: Crystal and Solid State Physics, Vol.11, Berlin: Springer-Verlag, 1979. 7. W. A. Smith, Proc. S PIE—The I nternational S ociety f or Optical Engineering, 1733, 1992. 8. H. Takeuchi, S. Jyomura, E. Yamamoto, and Y. Ito, J. Acoust. Soc. Am., 74, 1114, 1982. 9. Y. Yamashita, K. Yokoyama, H. Honda, and T. Takahashi, Jpn. J. Appl. Phys., 20 (Suppl. 20-4), 183, 1981. 10. Y. Ito, H. Takeuchi, S. Jyomura, K. Nagatsuma, and S. Ashida, Appl. Phys. Lett., 35, 595, 1979. 11. L. E. Cr oss, S. J . J ang, R . E. N ewnham, S. N omura, a nd K. Uchino, Ferroelectrics, 23, 187, 1980. 12. H. Takeuchi, H. Masuzawa, C. Nakaya, and Y. Ito, Proc. IEEE 1990 Ultrasonics Symposium, 697, 1990. 13. J. Kuwata, K. Uchino, and S. Nomura, Jpn. J. Appl. Phys., 21, 1298, 1982. 14. T. R. Shrout, Z. P. Chang, N. Kim, and S. Markgraf, Ferroelec. Lett., 12, 63, 1990. 15. R. E. Newnham, D. P. Skinner, and L. E. Cross, Mater. R es. Bull., 13, 525, 1978. 16. W. A. Smith, Proc. 1989 IEEE U ltrasonic Symposium, 755, 1989. 17. X. H. Du , J. Zhen g, U. B elegundu, a nd K. Uchino, J. Ap pl. Phys. Lett., 72, 2421, 1998. 18. Kistler, Stress Sensor, Production Catalog, Switzerland. 19. Tokin, Gyroscope, Production Catalog, Japan. 20. B. A. Auld, Acoustic F ields a nd Waves i n S olids, 2nd e dn., Melbourne: Robert E. Krieger, 1990. 21. G. S. K ino, Acoustic Waves: D evice I maging a nd Analog Signal Processing, Englewood Cliffs, NJ: Prentice-Hall, 1987. 22. C. S. Desilets, J. D. Fraser, and G. S. Kino, IEEE Trans. Sonics Ultrason., SU-25, 115, 1978.

1.

43722_C009.indd 27

9-27

23. T. R. Gururaja, Am. Ceram. Soc. Bull., 73, 50, 1994. 24. F. S. F oster, L. K. R yan, a nd D . H. Turnbull, IEEE T rans. Ultrason. Ferroelec. Freq. Cont., 38, 446, 1991. 25. Y. Ito, K. Kushida, K. Sugawara, and H. Takeuchi, IEEE Trans. Ultrason. Ferroelec. Freq. Cont., 42, 316, 1995. 26. C. A. Ros en, P roc. Ele ctronic C omponent S ymp., p . 205, 1957. 27. S. K awashima, O . Ohnishi , H. H akamata, S. T agami, A. Fukuoka, T. Inoue, and S. Hirose, Proc. IEEE Int’l Ustrasonic Symp. ‘94, France, Nov., 1994. 28. C. Campbell, Surface Acoustic Wave Devices and Thei r Signal Processing A pplications, Sa n Diego , CA: Academic Pr ess, 1989. 29. H. Matthews, Surface W ave Filters, N ew York: Wiley Interscience, 1977. 30. K. U chino a nd J . R . G iniewicz, Micromechatronics, N ew York: Marcel Dekker, 2003. 31. Y . S ugawara, K. Oni tsuka, S. Yoshikawa, Q . C. X u, R . E. Newnham, and K. Uchino, J. Am. Ceram. Soc., 75, 996, 1992. 32. K. Yano, T. Inoue, S. Takahashi, and I. Fukui, Proc. Jpn. Electr. Commun. Soc., pp. 1–159, Spring, 1984. 33. A. Ohgoshi and S. Nishigaki, Ceramic Data Book ‘81, Tokyo: Institute o f I ndustrial M anufacturing T echology, p . 35, 1981. 34. T. Narasaki, Japan Patent Disclosure, S. 57-137671, 1978. 35. M. Yorinaga, D. Makino, K. Kawaguchi, and M. Naito, Jpn. J. Appl. Phys., 24, Suppl. 24-3, 203, 1985. 36. Minolta Camera, Product Catalog “Mac Dual I, II”. 37. J. L. Fanson and M. A. E aley, Active and Adaptive Optical Components a nd S ystems II, S PIE 1920, Albuquerque, 1993. 38. Y. Yokoya, Electron. Ceramics, 22(111), 55, 1991. 39. F. K. Straub, Smart Mater. Struct., 5, 1, 1996. 40. A. Fujii, Proc. JTTAS Meeting on Dec. 2, Tokyo, 2005. 41. S. U eha a nd Y. T omikawa, Ultrasonic M otors, O xford: Clarendon Press, 1993. 42. K. Uchino, Piezoelectric Actuators a nd Ul trasonic M otors, Boston, MA: Kluwer Academic, 1996. 43. B. Koc, S. Cagatay, and K. Uchino, IEEE Ultrasonic Ferroelec. Freq. Control Trans., 49(4), 495–500, 2002. 44. S. Cagatay, B. Koc, and K. Uchino, IEEE Trans. UFFC, 50(7), 782–786, 2003. 45. K. U chino, P roc. N ew Actuator 2004, B remen, G ermany. June 14–16, p. 127, 2004.

9.3

Noncontact Ultrasonic Testing and Analysis of Materials

Mahesh C. Bhardwaj 9.3.1 Introduction Noncontact (air–gas coupled) u ses of sound waves a re perhaps buried i n t he a ntiquities o f o ur c ivilization. I n mo dern t imes, presumably one of the  rst applications is related to A ntarctic’s

10/3/2008 9:23:49 PM

9-28

Smart Materials

wood, lumber, e tc. p olymer Z-matched t ransducers have to be excited w ith a bnormally h igh p owers, t hus r isking t ransducer failure besides causing other problems. In 1997, Bhardwaj after decades of attempts nally succeeded in producing a transducer that is characterized by phenomenally high t ransduction i n a ir [6] f rom 5 0 kHz to mo re t han 5 MHz frequencies (Fi gure 9 .37). This t ransducer is Z-matched w ith compressed  ber m aterial t he ac oustic i mpedance o f w hich i s extremely close (arguably the closest) to that of air, which is primarily re sponsible f or e xtraordinary t ransduction i n a ir a nd other gases. Simply known as NCU transducers, their efficiency relative to polymer Z-matched transducers is order of magnitude more. Further, in ambient air, the sensitivity of the NCU transducers is 14 to 30 dB lesser (depending on the frequency) than their e quivalent c ontact a nd i mmersion c ounterparts, w hen tested with appropriate solids or with water. Considering the magnitude of NCU applications made possible by these transducers, we believe a new eld of signicance in materials and biomedical diagnostics has now become a reality. Details of NCU transducers, acoustic measurements, and applications have been provided by Bhardwaj in the Encyclopedia of Smart Materials [7]. In this chapter, our focus is on providing the reader with introduction to NCU transducers, systems, and applications for a w ide range of materials by utilizing the wellknown ultrasonic techniques.

ice t hickness me asurement b y s ending h igh-intensity s ound waves from an airplane in the 1920s. But close to reality, noncontact ultrasound (NCU) is highly desirable for testing and analyzing e arly s tage m aterials f ormation ( green c eramics, p owder metals, el astomers, c omposite p repregs, e tc.), a nd f or c ontact and liquid-sensitive materials (porous materials, food, pharmaceutical and hygroscopic materials, burn and wounded victims, or w here c ontact w ith t he te st me dia i s si mply a n uisance). Materials such as these cannot be reliably tested by conventional ultrasound, which utilizes liquid coupling of the transducer to the te st m aterial. L iquid c oupling i s ne cessary i n order to e fficiently t ransmit u ltrasound i n t he m aterial a s e xplained b y Brunk [1] in this book. However, t here a re re alities t hat de fy N CU, w hich a re: extremely h igh absorption of u ltrasound by a ir or ot her ga ses, phenomenally high acoustic impedance (Z) mismatch between air a nd t he test media, a nd i nefficient ultrasound transduction from a p iezoelectric material into the raried media like air or other ga ses (Figure 9.36). The rst t wo a re nat ural phenomena about which nothing can be done. Thus we are left to do “something” to t he piezoelectric material. To this effect, a h andful of researchers have been busy developing t ransducer designs t hat are based upon acoustic impedance, Z matching to air by utilizing a v ariety o f p olymers a s t he  nal m atching l ayer o n t he piezoelectric material, Fox et al. [2], Bhardwaj [3], Yano et al. [4], Haller and Khouri-Yakub [5]. These transducers exhibit reasonable effi ciency, but only for limited NCU applications, and that too at rel atively low f requencies suc h a s <5 00 kHz. Fu rther, i n order to i nvestigate attenuative materials such as concrete, very thick multilayered composites a nd s andwich structures, rocks,

9.3.2

NCU Transducers

It should be pointed out t hat early NCU transducers t hat are based upon conventional solid and polymer matrix piezoelectric

2 3

Gross Z mismatch between air/ gases to solids

1 Exorbitant absorption by air– gases —>100 dB @ 1 MHZ for Air

Ultrasonic transducer

Ambient air 2.s

Z1, 410 kg/m

>99.99% Ultrasound is reflected from most material surfaces R, Ref. Co-eff.

Z2 - Z1

2

Inefficient transmission from piezoeletric to air/gases

PZT Air

Z2 + Z1 Air-Mat. R = 0.99992

T = −97 dB

<0.01% Water-Mat. R = 0.74

Round trip loss to and from material ~190 dB

FIGURE 9.36 Realities t hat d efy NC U. U ltrasound a bsorption by a ir–gases a nd m assive a coustic i mpedance m ismatch b etween a ir a nd t est media are natural phenomena about which nothing can be done. Therefore, piezoelectric material must be treated in a manner that it exudes enormous acoustic pressure in air.

43722_C009.indd 28

10/3/2008 9:23:49 PM

9-29

Piezoelectric and Electrostrictive Ceramics Transducers and Actuators 160 140 Acoustic pressure (Pa/V)

1 MHz 50 mm 120 100

500 kHz 50 mm

80 100 kHz 50 mm

60

200 kHz 50 mm

40 20 0 0

25

50

75 100 125 150 175 200 225 250 Transducer to reflecting target distance in air (mm)

275

300

FIGURE 9.37 Acoustic pressure (Pa/V) of several NCU transducers as a function of reecting target distance from the transducer in ambient air.

materials allowed relatively easy production from 1 to 5 MHz frequencies. Dissatised with the performance and signal-tonoise ratio (SNR) of NCU transducers between 50 and 500 kHz based on conventional piezoelectric materials, forced us to think e ven mo re u nconventionally. I n 2 002, t his re sulted i n the development of a ne w piezoelectric composite. K nown as gas matrix piezoelectric (GMP) composite [8], it is characterized b y e xtremely u nusual f eatures. F or e xample, t he t hickness coupling constant of this material is nearly equal to longitudinal coupling constant. GMP is also characterized by zero ac oustic c ross-talk, v ery l arge d isplacement i n t he c oupling direction, negative Poisson’s ratio, very low mechanical Q, low dielectric constant, low density, and many other useful properties. Further, the manufacturability of GMP allows the

production of extremely large transducers, hitherto considered arduous or i mpossible. GM P i s pa rticularly su itable b etween <50 a nd >500 kHz. Th is ma terial ha s n ot o nly en hanced t he performance of NCU transducers, but has also elevated those contact a nd i mmersion app lications t hat re quire lo w u ltrasound frequencies. NCU t ransducers h ave b een suc cessfully p roduced f rom <50 kHz to >5 MHz and in the dimensional range from <1 mm to >250 mm i n p lanar, p oint a nd c ylindrical f ocused c ongurations. For high-throughput applications besides the large transmitters, multielement arrays have also be produced and deployed in t he i ndustry. Table 9.7 provides t he s alient features of NCU transducers a nd F igure 9 .38 s hows a n umber o f t hese de vices varying in frequency and dimensions.

TABLE 9.7 Salient Characteristics and Features of NCU Transducers Characteristics/Features Frequency range Dimensional range Shape Field geometry Sensitivity

Bandwidth Housing Coaxial connector Mechanical construction Environment limitations

43722_C009.indd 29

Description <50 KHz to >5 MHz <1 mm to >250 mm (extremely large dimensions also possible) Circular, square, or rectangular Planar, point, and cylindrical focused Extremely high in ambient air—typically 14–30 dB below conventional contact and immersion transducers Typically between 30% and 50% at bandwidth center frequency Two-part aluminum protected with plastic cover Standard BNC Robust, factor suitable Temperature: −20 to 60°C. RH: Up to 80%, higher with special provisions

10/3/2008 9:23:50 PM

9-30

Smart Materials

FIGURE 9.38 Wide range of NCU transducers including large transmitters and multielement arrays varying in frequency from 50 kHz to 5 MHz.

9.3.3

NCU System and Signal Processing

NCU transducers can be used with conventional pulsers and receivers ( designed f or c ontact a nd i mmersion u ltrasonic testing), but only for limited applications. However, for wide range of materials and testing objectives, it is necessary that suitable t ransducer e xcitation a nd r eceiver a mplification b e used. One such s ystem is U ltran’s 2nd Wave M 510, which is composed o f co mputer co ntrolled b urst p ulser, s ingle a nd multichannel receiver amplifier, and high-speed signal processing (Figure 9.39). This system is sufficient to generate NCU i mages o f m aterials i n v arious f ormats a nd me asurements, w hich c an a lso b e p ostprocessed to c onvert ac oustic

data i nto c haracteristics a nd f eatures rele vant to t he te st material.

9.3.4

NCU Techniques and Applications

Techniques such as through transmission, T-R pitch-catch, and single transducer pulse-echo that are well known in conventional ultrasound can also be applied to NCU. In this section, we provide brief introduction to these techniques and exhibit NCU analysis of a w ide range of materials. Details of materials a nalyzed are given in the gure captions. 9.3.4.1 Direct or through Transmission NCU This technique, which requires access to both sides of the test material (Figure 9.40) is the easiest to use in NCU mode. By utilizing suitable transducers here, we show imaging and analysis of a number of materials: • Uncured ( prepregs), c ured C arbon F iber Rei nforced Plastic (CFRP), C -C, a nd s andwich c omposites ( Figures 9.41 through 9.45). • Rubber, plastics, and foams (Figures 9.46 through 9.48). • Green, si ntered, a nd  red c eramics a nd re fractories (Figures 9.49 through 9.53) • Concrete (Figure 9.54) • Lumber and wood (Figures 9.55 and 9.56) • Food (Figure 9.57) 9.3.4.2

FIGURE 9.39 NCU Ultran’s 2nd Wave M510 system shown here with an X–Y raster scanning stage for imaging applications.

43722_C009.indd 30

T-R Pitch-Catch Same Side Reflection NCU

By suitably angulating transmitting and receiving transducers on the same side of the test material where access is only from one side, it is also possible to launch bulk waves in materials (Figure 9.58). By doing so, either the far side reflection o r re flections f rom w ithin t he m aterial c an b e i nvestigated to de cipher t he i nternal f eatures o r co ndition o f t he

10/3/2008 9:23:50 PM

9-31

Piezoelectric and Electrostrictive Ceramics Transducers and Actuators 150

100

50

mV

Transmitting transducer Ambient air

0

Test material −50

−100

Receiving transducer

−150 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 130 135 140 µS

FIGURE 9.40 NCU direct transmission technique showing an example of ultrasound transmission signal through a material. In this case from 8 mm CFRP composite with 1 MHz transducers.

20

X (mm) 40 60

80

0

100

0

0

20

20

40

40

60

60

Y (mm)

Y (mm)

0

X (mm) 40 60

80

100

0 dB, 100%

80

80

100

8 dB, 40% 3 dB, 100 69% 3 dB, 74%

120

120

0 dB

20

0 dB

FIGURE 9.41 Uncured (left) and cured (right) CFRP stacks transmission NCU images. Darker regions in the uncured material image are indicative of poor adhesion. During the curing process, several such areas are bonded, while some are not such as the ones indicated by the lighter areas in the cured material. These observations were also conrmed by optical imaging of sample cross-sections. Transducers: 500 kHz, 12.5 mm active diameter. (Reproduced from HEXCEL Composites, U.K.)

43722_C009.indd 31

10/3/2008 9:23:51 PM

9-32

Smart Materials Sem (nm) 40 45 50 55 60 65 70 75 80 85 90 95 100 105110115 120125130135140145150155160165170175180185190195 5 10

Index (nm)

15 20 25 30 35 40 45 50

−40 −30 −20 −10 0

−3

−12

0 to <1 % porosity

−21

1.5 to >4 % porosity

−30 −39

−48

FIGURE 9.42 Varying porosity (10 mm) c ured CFRP composites NCU i maging a nd a nalysis, left 0% to 0.1% porosity and right 1.5% to >4% porosity. Porosity variation across both samples is also shown by cross-sectional pro le, bottom, where y-axis scale is in dB. Transducers: 500 kHz, 12.5 mm active diameter.

0

18

36

54

72

90

108

126

144

162

180

198

216

234

245

0 17 34 51 68 85 102 119 139 0

50

100

FIGURE 9.43 Graphitized wove n c arbon c omposite NC U t ransmission i mage e xhibiting g ross p orosity a nd app arent d efects. Transducers: Focused 500 kHz 12.5 × 12.5 mm active area.

43722_C009.indd 32

10/3/2008 9:23:52 PM

9-33

Piezoelectric and Electrostrictive Ceramics Transducers and Actuators

0

50

100

X (mm) 150

200

250

0

300

100

150

200

250

50

100

150

200

250

50

50

100

100

Y (mm)

50

0

0

150

150

200 200 250 250 300 300 350 0%

100%

100

FIGURE 9.44 Nomex hone ycomb (50 mm t hick) C FRP s kins s andwich NCU transmission imaging showing articially created defects at skin hone ycomb i nterfaces. T ransducers: 1 40 kHz, 2 5 mm a ctive diameter.

50

FIGURE 9 .45 2D c arbon–carbon d isc br ake NC U t ransmission imaging showing regions of ve ry high and very low ultrasound transmission, presumably indicative of varying material density. Transducers: Focused 200 kHz, 25 mm active diameter. 56

63

70

77

84

91

100

0

6

12

18

24

30

36

42 mm 50.0

0 6 12 18 24 30 36 42 mm

FIGURE 9.46 diameter.

43722_C009.indd 33

Tire r ubber m ixes NC U i maging. L eft: Homo geneous m ix. R ight: He terogeneous m ix. Transducers: 5 00 kHz, 1 2.5 mm a ctive

10/3/2008 9:23:57 PM

9-34

Smart Materials

0

50

100

150

200

250

0

0

50

100

150

200

250

300

350

400

0

50

100

150

200

250

300

350

400

0

50

50

100 100 150 200

150

250 200 300 250 80 60

FIGURE 9 .47 Sheet mold ed pl astic ( SMP) NC U i maging s howing unwanted  ber or ientation. T ransducers: Fo cused 5 00 kHz, 1 2.5 × 12.5 mm.

40 20

FIGURE 9.48 A 100 mm polyurethane foam NCU imaging, presumably s howing p orosity a nd d ensity v ariations. T ransducers: 2 00 kHz, 25 mm active diameter.

2150 2100

r (kg/m3)

2050 2000 Gres rojo 1950

1850

Revestimiento blanco

1800

Gres porcelánico

1750 6.8

FIGURE 9.49 diameter.

43722_C009.indd 34

Gres blanco

1900

6.9

7.0

7.1

7.2 7.3 ln(vL(m/s))

7.4

7.5

7.6

Green ceramic oor tiles NCU velocity–density relationships for a v ariety of compositions. Transducers: 500 kHz, 19 mm active

10/3/2008 9:24:01 PM

9-35

Piezoelectric and Electrostrictive Ceramics Transducers and Actuators

100 0

150

200

X (mm) 250

300

350

(Figure 9.60). By monitoring the amplitude and time of ight of surface re ections, m aterials c an b e c haracterized f or su rface roughness, texture, etc. Figures 9.61 and 9.62 provide examples to this effect. While highly desired, pulse-echo NCU technique to obtain the far side materials ref lection under ambient conditions is at best ex tremely a rduous, i f n ot i mpossible a t t he t ime o f this w riting. H owever, i f t he te st i s co nducted u nder h igh air–gas pressure, then it is relatively easy to observe far side ref lections i n p ulse-echo N CU mo de. F igure 9.63 s hows a n example o f multiple t hickness re f lections f rom 10 mm s teel at 5 ba r a ir p ressure b y u sing a s pecial b roadband 3 MHz transducer.

400

50

Y (mm)

100

150

200

250

9.3.5

300

0%

100% %

FIGURE 9.50 Green c eramic  oor t ile h igh-resolution NC U i mage. Transducers: Focused 500 kHz, 19 mm active diameter.

material. An example of this is shown in Figure 9.59. Here we see the artificially embedded defects in a 4 mm thick panel of CFRP composite. 9.3.4.3 Single Transducer Pulse-Echo NCU Under ambient conditions, it is very easy to obtain high-quality signals corresponding to the reection from the material surface

0

8

16

24

32

40

48

Perusal of NCU

In this chapter, we have provided an introduction to an extraordinarily h igh a ir–gas t ransduction p iezoelectric transducer. Examples of materials testing by NCU with wellknown ultrasonic techniques have also been given. In our laboratory, work continues to en hance NCU by considering the ge ometrical ac oustic i n a ir a nd u ltrasound i nteraction with materials. This work exhibits the unique nature of NCU with respect to very high resolution and detectability of discontinuities in m aterials, b esides o ther a pplications o f significance. Underscoring t he sig nificance a nd no velty o f N CU, i n recent years, several ultrasound and materials research institutes and organizations have undertaken the task of pursuing this field. It suffices to say that their work also concerns testing a nd a nalysis o f i ndustrial m aterials, f ood, b iomedical,

56

64

72

80

88

96

104 mm 110

0 7 14 20.0 4

6

8

10

12 dB

12 dB

10 8 6 4 10

20

30

40

50

60

70

80

90

100 mm 110

FIGURE 9.51 Green alumina varying in density NCU image analysis. Left: Pressed at 9000 psi, right: pressed at 15,000 psi. Note as the density increases (higher pressing pressure), in NCU mode ultrasound transmission decreases. Transducers: 500 kHz, 12.5 mm active diameter.

43722_C009.indd 35

10/3/2008 9:24:06 PM

9-36

Smart Materials

0

9

18

27

36

45

54

63

72

81

90

99

108

117 mm 125

0 8 15.0 −20

−10

0

dB

dB

−5 −10

−15 20

40

60

80

100

120 mm

FIGURE 9 .52 Sintered a lumina NC U t ransmission i maging a nd a nalysis of s amples v arying i n d ensity. L eft: 65 % ( 2.58 g/cc), m iddle: 7 0% (2.8 g/cc), right: 77% (3.08 g/cc). Transducers: 500 kHz, 12.5 mm active diameter.

FIGURE 9.53

Fired silica refractory NCU imaging. Left: Defect-free. Right: Cracked. Transducers: 500 kHz, 12.5 mm active diameter.

0

13

26

39

52

65

78

91

104

117

130

143 mm 150

0

13

26

39 mm 50.0 −40

−30

60

80

dB

dB −30

−40

20

40

100

120

140 mm

FIGURE 9.54 Concrete NCU (200 mm) imaging showing internal defects. Transducers: Mixed 140 kHz.

43722_C009.indd 36

10/3/2008 9:24:08 PM

9-37

Piezoelectric and Electrostrictive Ceramics Transducers and Actuators

FIGURE 9.55 Oriented strand board (OSB) NCU imaging as a f unction of moisture content. Left to right: Dry, 30 min soak, 2 h soak, and 11 h soak. T ransducers: 2 00 kHz. N ote a s t he moi sture c ontent i ncreases t he NC U t ransmission d ecreases, w hich i s more e vident i n t he co lored images.

0

25

50

75

100

125

150

175 mm 200

0 24 48 72 96 120 144 mm

170 −30

FIGURE 9.56

−20

−10

0 dB 10

Porous wood core NCU high-resolution imaging. Transducers: Mixed 1 MHz.

FIGURE 9.57 Cheddar cheese NCU imaging. Left: Reduced fat. Right: Extra sharp. It appears as the fat content in cheese increases so does the ultrasound attenuation. Transducers: 1 MHz, 12.5 mm active area diameter.

43722_C009.indd 37

10/3/2008 9:24:11 PM

9-38

Smart Materials 250 200 150

Transmitting transducer

100

Receiving transducer mV

50

Ambient air

0 −50

−100 −150

Test material

−200 −250

10

15

20

25

30

35

40 µS

45

50

55

60

65

70

75

FIGURE 9.58 NCU T-R pitch-catch reection technique showing far side (bottom surface) reection from a m aterial. In this case the bottom surface of 12.5 mm thick aluminum with 1 MHz transducers.

0

32

64

96

128

160

192

224

256

288

320

352 mm 380

0

32

64

96

128

160

192

224

256

288

320

352 mm 380

0

0

1 33

1 33

1

2 66

2 66

2

99

99

3 132

3 132

165 4

165 4

198

198

5 228 mm

5 228 mm 20

30

40

50

60

70

3

4

5 % 20

80 %

40

60

80

FIGURE 9.59 Same side NCU image of a CFRP panel with embedded defects (left). As a comparison, the right hand image shows similar image, but acquired by direct transmission mode. Transducers: 500 kHz, 12.5 mm active diameter.

450 300 Transmitting/ receiver mV

150

Ambient air

Material surface

0 −150 −300 −450

30

35

µS

40

45

FIGURE 9.60 NCU single transducer, pulse-echo technique showing surface reection signal from a material. In this case the surface of dense alumina with 3 MHz transducer.

43722_C009.indd 38

10/3/2008 9:24:13 PM

Piezoelectric and Electrostrictive Ceramics Transducers and Actuators

9-39

FIGURE 9.61 Very high-resolution NCU surface images of familiar objects generated by 3 MHz sharply focused transducer.

and other applications. A partial list of these organizations is as following: • • • •

Iowa State University, United States University of Bordeaux, France University of Irvine at California, United States University of Ancona, Italy

• • • • • • •

Institute of Technical Ceramics, Spain Yamanashi University, Japan California State University at Carson, United States Katholieke Hogeschool (KATHO), Belgium Katholieke University at Leuven, Belgium Stanford University at Stanford, United States University of Windsor, Canada

FIGURE 9.62 Surface i maging of ot her materials: L eft: S emiconductor polishing pad showing surface rou ghness. R ight: Fi ne celled polymer foam showing its texture, the details of which are more evident in color image.

43722_C009.indd 39

10/3/2008 9:24:15 PM

9-40

Smart Materials Tek Run

Trig’d

T

T

Transmitter/ receiver

1

High air/gas pressure Test material Ch1

100mV

M 4.00µs

A

Ext

397mV

FIGURE 9.63 NCU single transducer, pulse-echo at high gas pressure. In this case, far side (bottom surface) reections from a 10 mm steel plate at 5 bar air pressure by using a broadband 3 MHz NCU transducer. (Reproduced from Gas Technology Institute, U.S.A.)

• Fraunhofer Institute at Saarbrucken, Germany • Penn State University at University Park, United States • Queens University at Kingston, Canada • UWarwick University, nited Kindom • UNational Physical Laboratory, nited Kindom • Los Alamos National Laboratory, United States • Gas Technology Institute, United States • Southwest Research Institute, United States • Nagoya Institute of Technology, Japan • Johns Hopkins University, United States It i s o ur h ope t hat t he m aterials a nd b iomedical i ndustries, research institutes, and government agencies take a serious look at the status of modern ultrasound, particularly NCU transducers and a ssociated te sting k now-how a nd i ts ne cessity i n s olving important p rocess a nd qu ality-related m aterial te sting ne eds. The advantages of NCU are too numerous and far beyond conventional ultrasound to be ignored. We believe sooner it is applied for specic tasks along with increment in practical educational foundation, b etter it would b e for our c omplex s ociotechnical world.

3. Bhardwaj, M.C., M odern ul trasonic tra nsducers, C ommercial Catalog, Ultran Laboratories, Inc., 1986. 4. Yano, T., Tone, M., a nd F ukumoto, A., R ange  nding and surface c haracterization usin g hig h f requency a ir tra nsducer, IEEE Trans. UFFC, 34(2), 222–236, 1987. 5. Haller, M.I. and Khuri-Yakub, B.T., 1–3 composites for ultrasonic a ir tra nsducer, IEEE U ltrasonics S ymposium, p p. 937–939, 1992. 6. Bhardwaj, M.C., U ltrasonic tra nsducer f or hig h tra nsduction in gas es and method for non-contact transmission in solids, U.S. Patent 6,311,573, November 6, 2001. Japan Patent 3225050, August 24, 2001. E uropean C ooperation T reaty, Pending. WIPO, PCT/US98/12537. 7. Bhardwaj, M.C., N on-destructive e valuation: in troduction of non-contact ultrasound, Encyclopedia of Smart Materials, M. Schwartz ed., John Wiley & Sons, New York, pp. 690–714, 2002. 8. Bhardwaj, M.C., Gas ma trix p iezoelectric co mposite, U.S. and international patents pending.

References 1. Brunk, J.A., Ultrasonic nondestructive testing and materials characterization, this volume. 2. Fox, J .D., K huri-Yakub, B .T., a nd K ino, G.S., H igh f requency wave measurements in a ir, 1983 IEEE U ltrasonics Symposium, pp. 581–592, 1983.

43722_C009.indd 40

10/3/2008 9:24:16 PM

10 Chitosan-Based Gels and Hydrogels 10.1 C hitosan-Based Gels .............................................................................................................10-1 Introduction • Supramolecular Interactions and Gel Formation • Applications

References .........................................................................................................................................10-10 10.2 Chitosan-Based Hydrogels in Biomedical and Pharmaceutical Sciences ..................10-13 Introduction • C hitosan and Chitosan Derivatives • H ydrogels • C hitosan Hydrogel Applications • C onclusion

References .........................................................................................................................................10-17

10.1 Chitosan-Based Gels Kang De Yao, Fang Lian Yao, Jun Jie Li, and Yu Ji Yin 10.1.1 Introduction Chitosan i s a nat ural biopolymer obtained by de acetylation of chitin, which is produced from marine shellsh, such as crabs, shrimp, fungal cell walls, and other biological sources. Chemically it is a linear cationic poly(β-(1–4)-2-amino-2-deoxy-d-glucan) derived f rom c hitin, a p oly(β-(1–4)-2-amino-2-acetamido2-deoxy-d-glucan) b y de acetylation. C hitosan i s de scribed i n terms o f t he de gree o f de acetylation ( DDA) a nd a verage molecular weight. And next to cellulose, it is the second most plentiful biomass and is already known as a b iocompatible and biodegradable material. Many researchers have examined tissue response to v arious c hitosan-based i mplants. Re sults i ndicate that these materials evoke minimal foreign body reactions [1]. Gels c onsist o f 3 D p olymer ne tworks s wollen i n s wellers, whereas h ydrogels s well i n w ater. Sm art ( intelligent) gel s ( or hydrogels) can swell or contract in response to stimulus changes, e.g., tem perature, p H, io nic s trength, c hemical, a nd ele ctrical elds. Chitosan h as b een u sed f or de veloping sm art gel s v ia ne tworks o r c omplex f ormations. D ue to t heir v arious p roperties that de pend o n en vironmental v ariables suc h a s p H, io nic strength, and electric eld, these materials have potential applications such as biomaterials, separation membranes, a nd  eldresponsive materials.

10.1.2

Supramolecular Interactions and Gel Formation

We know that chitosan-based gels are chemically or physically cross-linked networks. To make them smart, special supramolecular interactions are often introduced into the network to provide reversible re sponsiveness to en vironmental s timuli. The refore, different interpolymer c omplexes a re f ormed wi thin th e c hitosan-based networks. These c omplexes c an f orm o r d issociate i n re sponse to changes in their environment. They can be divided into the following major categories based on the dominant type of interaction. 1. Hydrogen bonding complexes 2. Polyelectrolyte complexes (PEC) 3. Grafted and block network 4. Self-assembly 5. Coordination complexes The presence of amino groups in chitosan (pKa = 6.5) results in pH-sensitive, temperature-responsive, and electric eld-sensitive behavior. Research efforts have been aimed at tailoring the properties o f c hitosan t hrough c ross-linking ( include c ovalence cross-link a nd p hysical c ross-link), f unctionalization, c opolymerization, and blending to t smart requirements. 10.1.2.1 Hydrogen Bond Complexes The preparation of a hydrogel containing a covalently cross-linked chitosan m inimally r equires c hitosan an d a c ross-linker i n an appropriate s olvent, u sually w ater. The a mino g roups o f g lucosamine residue within chitosan chains can serve as cross-linking 10-1

43722_C010.indd 1

10/3/2008 9:31:29 PM

10-2

Smart Materials

sites, for example, reacting with glutaraldehyde (GA) [2,3], glyoxa [4], proanthocyanidin [5], 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide ( EDC), o r gen ipin [ 6] to f orm i mine c ross-linking between linear chitosan chains that lead to gel formation. Other components can be added, such as additional polymers to form a hybrid polymer network (HPN) or a semi- or full-IPN (interpenetrating p olymer ne twork). A uxiliary mole cules c an also be used to i nitiate reactions during t he preparation of t he network. An IPN is dened as a combination of two cross-linked polymers; at le ast one of them synthesizes or cross-links in the presence of the other. If one of the polymers is linear (without being cross-linked), a s emi-IPN results. The IPN technique can be used to improve the properties of chitosan-based gels. Lee a nd C hen [ 7] e xamined GA -cross-linked c hitosan– isopropylacrylamide(chitosan/PNIPAAm) s emi-IPN an d I PN hydrogels. They found that the swelling ratios of IPN gels are lower than those of semi-IPN gels and the reason is that the addition of cross-linked chitosan in the PNIPAAm hydrogel makes the network structure of the gel denser and more hydrophobic. The swelling ratios of PNIPAAm/chitosan gels slightly decrease with the addition of chitosan, but the swelling ratios are not sig nicantly affected by the amount of chitosan added to the PNIPAAm gel. Because o f t he e xistence o f m any – NH2 g roups on c hitosan chains, h ydrogels c omposed o f c hitosan a re a lso u sed a s p Hsensitive materials [8]. A novel pH-sensitive hydrogel composed of N, O-carboxymethyl c hitosan a nd a lginate c ross-linked b y genipin (NOCC/alginate) and the swelling ratio of this hydrogel at pH 1.2 was about 2.5 w ithin 30 min subsequent to s welling, while i t w as app roximately 6 .5 at p H 7.4 (as s hown i n F igure 10.1). This nding may contribute to the fact that, at low pH, the swelling ratio was limited due to f ormation of hydrogen bonds between NOCC (–COOH and –OH) and alginate (–COOH and –OH). However, at higher pH 7.4, the carboxylic groups on the genipin-cross-linked NOCC/alginate hydrogel became progressively 16 (n = 5)

NOCC:Alginate = 1:0, 0.75mM Genipin, pH 1.2 NOCC:Alginate = 1:1, 0.75mM Genipin, pH 1.2 NOCC:Alginate = 1:0, 0.75mM Genipin, pH 7.4 NOCC:Alginate = 1:1, 0.75mM Genipin, pH 7.4

14

ionized (–COO–). I n t his c ase, t he hydrogel s welled mo re signicantly due to a large swelling force created by the electrostatic repulsion between the ionized acid groups. The swelling ratio of the h ydrogel wi th al ginate w as s ignicantly l ower th an th at without alginate due to the formation of hydrogen bonds between NOCC (–NH2 and –OH) and alginate (–OH) [9]. Yao et al. [10] examined GA cross-linked chitosan–poly(propylene glycol) (PPG), semi-IPN as a pH-sensitive hydrogel. Chitosan–PPG s emi-IPN s hows p H-dependent s welling; t he highest swelling degree (SD) is at pH 4.0 and the lowest at pH 7.0 (Fi gure 10.2). The s tructure f eatures o f c hitosan–PPG a re reversible (e.g., –N H3 to – NH2) w hen t he gel s a re t ransferred from a pH 1.0 to a pH 7.8 buffer [11], because it is a hydrophobic and w ater-insoluble p olymer a nd t he i ncorporated P PG i s expected to decrease the equilibrium SD in a low pH medium. In addition, linear PPG in the network can enhance the  exibility of a semi-IPN. For the same reason, at a higher pH it may not fully ionize all the –NH 2 , the pH-sensitive, biodegradable, semi-IPN c hitosan–polyvinyl a lcohol ( PVA) h ydrogel c rosslinked w ith GA a nd s hows lo w-bound w ater c oncentrations (CBW, mg/mg dry hydrogel). On the other hand, if the pH goes too low, the high concentration of H + will actually impair the swelling, s o t he m aximum CBW, mg /mg d ry h ydrogel w ould appear at pH 3 medium (Figure 10.3) [12,13]. The SD of hydrogels based on [N-(2-carboxybenzyl) chitosan, CBCS] (CBCSG) was signicantly increased with the raise of the pH in the range of pH 5.0–9.0 and the lowest swelling occurred almost always at pH 5.0 among CBCSGs, as shown in Figure 10.4 [14]. This maybe depends mainly on the osmotic pressure difference between the inside of the gel and the surroundings caused by the redistribution of mobile ions [15], and along with decreasing pH. The a mount o f – COO i s g radually re duced i nside t he hydrogels, w hich le ads to a de crease i n o smotic p ressure a nd makes the SD of the hydrogels smaller. When the pH was reduced to 1 , t he – COO w as c ompletely i nhibited, a nd t he do minant ionic g roups were – NH3+, but t he number i s sm all b ecause t he majority of the –NH2 groups were linked, therefore the CBCSGs swelled sluggishly as pH diminished in the range of pH 1.0–5.0.

12 Equilibrium degree of swelling

Swelling ratio

60 10 8 6 4 2 0 0

50

100

150 200 Time (min)

250

300

FIGURE 1 0.1 Swelling c haracteristics of t he NO CC a nd NO CC/ alginate hydrogels cross-linked w ith 0.75 mM genipin at p H 1.2 or 7.4 at 37°C.

43722_C010.indd 2

50 40 30 20 10 0

1

2

3

4

5

6

7 pH

8

9

10

11

12

13

FIGURE 10.2 Equilibrium d egree of s welling v s. pH for t he s emi-IPN, synthesized from different amounts of cross-linking agent. Molar ratios of –CHO:–NH2 are 2.21° and 4.83°, respectively.

10/3/2008 9:31:30 PM

10-3 10

4.0 Equilibrium degree of swelling

Bound water concentration (C BW) g/g dry gel

Chitosan-Based Gels and Hydrogels

3.0

2.0

1.0

4

2

1

2

0.0 2

3

4

5

6

7

8

9

10 11 12

pH

FIGURE 1 0.3 Bound w ater c oncentration i n t he c hitosan–PVA hydrogel a t d ifferent p H v alues ( chitosan:PVA mol ar r atio 1 :10, G A concentration 33.3 M).

14 12 10 Swelling ratio

6

0

1

8 pH 9.0 pH 5.0 pH 1.0

6

pH 7.4 pH 3.0

4 2 0

0

FIGURE 1 0.4 solutions.

4

8

12 16 Time (h)

20

24

Swelling b ehavior o f C BCSG in diff erent buffer

Collagen is k nown to b e the most promising of biomaterials and has been found diverse applications, which can form a complex w ith c hitosan [16]. The collagen–chitosan complex c an be used to mimic the components of the native extracellular matrix (ECM). It has also been reported that the hybrids of collagen– chitosan manufactured by cross-linking [17]. Gelatin is a denatured form of collagen, composed of glycine, praline, hydroxyproline, arginine, and other amino acids. The amphiphilic protein (isoelectric point = 4.96) can provide amino groups for cocross-linking with c hitosan v ia GA t hus p reparing a c hitosan–gelatin ( CG) HPN. The pH-sensitive swelling behavior of HPN gel is displayed in Figure 10.5. The data show that the degree of swelling declines

43722_C010.indd 3

8

3

4

5

6

7 8 pH

9

10 11 12 13

FIGURE 10.5 Swelling behavior of CG hybrid network specimen with a –CHO:–NH2 molar ratio of 10 in solutions of different pH with ionic strength I= 0.1 at 37°C.

sharply at pH 7.0 and this can be explained by the fact that the hydrogen bonds within the chirosan–gelatin HPN dissociate in an acidic medium [18]. The properties of many chitosan-based hydrogen bond complexes depends on not only pH but also temperature. Carmen et al. investigated the swelling behavior of different crosslink chitosan–poly (N-isopropylacrylamide) interpenetrated networks (CS/PNIPA). The SD of all the CS/PNIPA IPN decrease with the enhancement of temperature, and the SD at low pH 3 i s higher than that at high pH 8 for any temperature (as shown in Figure 10.6) [19]. Temperature/pH-sensitive poly(2-ethyl-2-oxazoline)/ chitosan(PEtOz/CS) I PN hydrogels exhibited considerable shrinkage over the temperature range of 35°–45°, and thermoresponsive behavior has decreased with increasing chitosan content [20]. The ele ctrical  eld i s a nother i mportant f actor, w hich i nuences t he properties of chitosan-based hydrogels. The response of electrosensitive hydrogels is generally exhibited in the form of either swelling and shrinking or bending behaviors. Figure 10.7 shows the degree of bending of the chitosan–poly(hydroxyethyl methacrylate) s emi-IPN h ydrogel a s a f unction o f t he v arious applied voltages in an aqueous 1.0% by weight NaCl solution. The degree and the increase in the speed of bending as well as an increase in the voltage applied across the hydrogel indicates that the b ending i s i nduced b y t he ele ctric c urrent, h owever, t he bending did not o ccur in pure water [21]. The chitosan–polyallylamine IPN hydrogel exhibited the same characteristic [22]. 10.1.2.2 PEC PEC are generally obtained either by the reaction of polycations and polyanions or by polymerizing monomers that have suitable functional groups onto polymeric templates of known structures. PECs have numerous applications such as membranes, medical prosthetic m aterials, en vironmental s ensors, a nd c hemical detectors. C hitosan, a c ationic p olysaccharide, h as b een c omplexed with anionic polymers. The properties of PEC are mainly

10/3/2008 9:31:30 PM

10-4

Smart Materials

100

Degree of swelling (%)

pH 3 (Glycine buffer)

80

60

40

20

0 100

Degree of swelling (%)

pH 8 (Phosphate buffer)

80

60

40

20

0 20

25

30 35 Temperature (⬚C)

40

45

FIGURE 10.6 Dependence of the degree of swelling (%) of PNIPA/CS IPNs on t he pH a nd temperature of t he aqueous medium; I PN-0 (O), IPN-1 (
), IPN-3 (), IPN-5 (▼), and IPN-7 (▼). PNIPA/chitosan-I IPNs exhibited a similar pattern. (IPN-X, where X denotes the concentration (Vol.%) of GA.

90

Degree of bending (%)

80 70 60 50 40 30 20 0

5

10 15 Voltage (V)

20

25

FIGURE 10.7 Degree of bending of chitosan/PHEMA semi-IPN in response to various voltages in a 1.0 wt % NaCl solution at T = 35°C.

43722_C010.indd 4

determined by the degree of interaction between the individual polymers. This l atter c ondition de pends e ssentially o n t heir global charge densities and determines their relative composition in t he P EC. The lo wer t he c harge den sity o f t he p olymer, t he higher is the polymer proportion in the PEC, since more polymeric chains are required to react with the other polymers. A PEC can be f ormed v ia t wo io nizable p olymers w ith opp osite c harges individually. This means that a PEC-forming reaction can only occur at pH values in the vicinity of the pKa interval of the two polyelectrolytes. CS i s a p H-sensitive p olymer ( pKa i s c a. 6 .5) w hose a mino groups can be protonated depending on the pH of the environment. When CS is mixed with ampholytic gelatin (Gel, its pHiso is approximately 4.7), which in the case of pH of medium above the isoelectric point of G el, a nd where a ne t charge is negative one, t here o ccurs ele ctrostatic i nteraction b etween t he a mmonium ions of CS salts and carboxylate groups of Gel. Yao et a l. [23] found that there was strong interaction between Gel and CS in t he aque ous me dium t hat w as enough to f orm PEC i n situ, and the CS/Gel PEC only yielded at pH value larger than 4.7 and above pH 6.2 CS could precipitate from the solution. Pectin is an acidic polysaccharide that has a repeating unit of α-(1,4)-l -rhamnose u nits. A P EC h as b een f ormatted f rom anionic pectin and cationic chitosan. The PEC swells obviously at pH < 3 and pH > 8, and does not swell in the range of 3 < pH < 8. Moreover, its degree of swelling in an acidic medium is by far higher than that in an alkaline medium. The swelling of the PEC correlates with its composition and is also affected by the degree of deacetylation (DDA) and the methoxy level of pectin [24]. Alginate comprises a linear chain of (1,4)-linked β-d-mannuronate and α-l -glucuronate residue arranged blockwise. A gel-like chitosan/ alginate PEC was at one time formed with coexisting polyions of opposite charges (–NH3+ and –COO−). Polysaccharides (chitosan and alginate) h ave b ulky py ranose r ings a nd a h ighly s tereoregular conguration in their rigid linear backbone chains. At a g iven pH, the composition of the PEC shifted to a lower alginate content as the degree of N-acetylating of chitosan increased and the higher the pH, the lower the alginate content of the PEC for a given chitosan sample [25]. W. Argüelles-Monal et al. [26] found that the degree of polymerization of chitosan and the chemical composition of alginic acid have shown to have an inuence on the PEC formation. Yoshitsune et al. [27,28] found that the chitosan/polyalkyleneoxide–maleic acid copolymer (CS/PAOMA) PEC lms swelled at a low pH, shrunk at pH between 4.8 and 6.5 and their swelling enhanced after pH 6.5. This swelling behavior can be attributed to the electrostatic interaction between the protonation of amino groups in CS and carboxyl groups in the PAOMA. The pKa of CS is 6.5 and the pKa of PAOMA is 4.8. In different pH medium corresponding g roups ( e.g., N H3+ a nd C OO−) p lay d ifferent roles between intermolecules. 10.1.2.3 Grafted and Block Network Many p olymers, suc h a s ph osphatidylethanolamine, p oly(ethylene glycol) (PEG), poly(vinyl acetate), poly(vinyl alcohol), poly(3-hydroxyalkanoate), linoleic acid, poly(l -lactic acid), and

10/3/2008 9:31:30 PM

10-5

Chitosan-Based Gels and Hydrogels

OH O

OH O

O HO

HO NH X O

OH O

O

O HO

NH2

NHAc z

X = Linker

x O

O

y

OMe

n

Structure of chitosan-g-PEG.

so on, have been grafted onto the chitosan in order to obtain some type of special performance. PEG i s a w ater-soluble p olymer t hat h as lo w to xicity a nd immunogenicity. Hydrophilic PEG can form a diff use layer containing water molecules, which has protein resistance when PEG is adsorbed on a surface. Many researchers have turned their attention to graft copolymers of chitosan and PEG, as shown in Figure 1 0.8, b ecause o f t heir h igh b iocompatibility [ 29–31]. Hydrogen bonds exist between the hydrogen of the amino groups of chitosan and the oxygen of the polyether in alkaline regions, whereas the hydrogen bonds dissociate at acidic pH. Additionally, the SD is very sharp at a h igh pH 6 [32], and the polymers were soluble in water over a w ide pH range w ithin a s hort t ime a nd the viscosities of the solution were very low. The degradation rate of the copolymer declined with the enhancing degree of substitution of PEG to chitosan, which was probably due to the relative stability of PEG to lysozyme [33]. Poly(α-hydroxy ac ids) g enerates ac idic deg radation p roducts at t he i mplanted si te, w hich e vokes u ndesirable t issue reaction [ 34]. The ac id b y-product m ay le ad to lo cal d isturbances due to poor vascularization in the surrounding tissue. Chitosan may be combined with acid-producing biodegradable polymers, s o t hat lo cal to xicity d ue to t he ac id b y-products can b e a lleviated. Y ao e t a l. ob tained t he c ytocompatible poly(chitosan-g-l -lactic ac id) t hrough g rafting oligo(l -lactic acid) o nto t he a mino g roups i n c hitosan w ithout a c atalyst. These graft copolymers had a higher strength than chitosan and their swelling behavior was inuenced by pH (as shown in Figure 10.9). When a pH < 2 and with an increase in the buffer pH, the concentration of the charged ionic groups in the  lms also i ncreased. The s welling o f t he s amples w ere s hown to increase due to t he enhancement of the osmotic pressure and charge repulsion. While at a h igher pH, the degree of ionization was reduced due to the deprotonation of the amino units of c hitosan a nd t he s welling of t he  lms de creased. I n add ition, t he h ydrophobic side -chain a ggregation a nd h ydrogen bonds in the copolymers were much stronger, which also lead to lower swellability of the samples [35,36]. 10.1.2.4 Self-Assembly The te chnique o f s elf-assembly l ayer b y l ayer ( LbL) PEC multilayer lms opens up many new opportunities for us to achieve the

43722_C010.indd 5

24 20 Swelling degree

FIGURE 10.8

CL-5 CL-3 CL-1

16 12 8 4 0 0

2

4

6

8

10

12

14

pH

FIGURE 1 0.9 Effect of t he L A/CS fe ed r atio a nd p H v alue on t he equilibrium water uptake of the CL  lms (CL-1, CL-3, and CL-5 denote that LA/CS = 0.5, 2.0, and 4.0(w/w), respectively).

ideal mo del su rface w hose p roperties a re c ontrollable [37]. The self-assembly polyelectrolyte systems are sensitive to the fabrication orders in terms of the orders of static electric interactions. It has long been known that one of the challenges of the materials construction is t he co ntrollable p roperties o f t he ma terials, including shape, size, morphology, thickness, roughness, chemical and t ransport p roperties, de gree o f i nterpenetration [ 38], a nd surface evidence associated with the mechanisms of the absorption and formation of polyelectrolytes on charged surfaces [39]. Poly(acrylic acid) (PAA) is a synthetic anionic electrolyte that has been synthesized for use in transmucosal drug delivery and self-curing memb ranes w ith C S. C S-PAA w ith a na nobrous structure have been formed by self-assembling. The nanobers are of diff erent leng ths ranging f rom 1 µm to s everal m icrons, but the widths of the nanobers are on a na noscale from 50 to 150 nm. In the CS/PAA system, ionic interactions occur between the two oppositely charged macromolecules forming a polycomplex t hat p recipitates a s a n i nsoluble a ggregate b ecause t he charged groups, which are responsible for solubility (carboxylic groups of PAA in the form of COO− and amino groups of CS in the form of NH3+), are involved in the complex [40]. The formation of nanoparticles by self-assembly is an energysaving p rocess. S elf-assembling na noparticles c an b e f ormed spontaneously u nder m ild c onditions. F igure 10.10 s hows t he

10/3/2008 9:31:31 PM

10-6

Smart Materials −+ −+ −+ −+ −+ −+ −+ −+ −+ −+

+ + + + +

+ + + + + Chitosan hydrolysate −

+ + + + − + + − + + − − + − − +

− − − −

Side-by-side linkage Mixing

− − − − − − − − −

Carboxymethyl cellulose (CMC) hydrolysate

End-to-end linkage

+ + − + + −+ + − + − + + + − + − + − + + + − − + − − − − + − − + + − − + − − +

Chitosan–CMC nanoparticles

Random linkage Self-assembly by polyionic complexation

FIGURE 10.10 Schematic illustration of the formation of chitosan–CMC nanoparticles.

formation p rocess f or c hitosan/carboxymethyl c ellulose ( CS– CMC) n anoparticles. The s ize of d istribution w as re latively narrow and the mean particle size decreased from 226 to 165 nm with the decreasing molecular weight of chitosan hydrolysate from 9.5 to 6.8 kDa. Moreover, with a decreasing pH, the average size was enhanced however; the size was not affected by low temperature [41]. 10.1.2.5

Coordination Complexes

Chitosan has one amino group in addition to hydroxyl groups in the re peating u nit. S o c hitosan c an c oordinate s trongly w ith transition-metal io ns [ 42] w here t he p recious me tal a nd r are metal ions form coordination complexes [42]. They may be used as a bsorbents f or me tal io ns a nd s eparation memb ranes. The order f or t he s tability c ontents o f t hese c hitosan–transitionmetal c omplexes i s M n
10.1.3 Applications 10.1.3.1

Controlled Release Matrixes

The aim of the controlled release of drugs is either to mo dulate tissue drug levels or spatially or temporarily place a drug in some region of the body to maintain efficacy a nd de press side re actions. Therefore, a matrix used for controlled release should offer modulating ability for drug delivery in addition to maintaining drug stability with the matrix. Chitosan has structure characteristics si milar to t hose g lycosaminoglycans ( GAGs); i ts sm art

43722_C010.indd 6

material properties have been explored for use in controlled release matrixes. The use of chitosan in the development of oral sustained release stemmed from the intragastric “oating tables” of c hitosan. C hitosan h as gel -forming p roperties i n a lo w p H range in addition to its antacid, antiulcer characteristics. Thu s, it has promising potential for use in oral sustained delivery systems. Gastrointestinal, respiratory, ophthalmologic, cervical, and vaginal mucins are hydrophilic saline gels that are thickened by nature in anionic glycoproteins. In searching for drug delivery systems for suc h sub strates, su itable c ationic p olymers a re ne eded. Chitosan is known to be bioadhesive to mucosal surfaces, such as in the eye, nose, and vagina. The DDA of chitosan and molecular weight are the main factors that affect the role of chitosan in therapeutic and intelligent drug delivery systems [45]. The DDA also controls the degree of crystallinity and hydrophobicity in chitosan due to variations in hydrophobic interactions. These hydrophobic interactions u ltimately control t he loading and release characteristics of chitosan matrices. Gupta and Jabrail [46] revealed that microspheres prepared using chitosan with 62% (w/w) of DDA were suitable for sustained release of centchroman in comparison to m icrospheres prepared using chitosan with low (48% w/w) and high (75% w/w) DDA. The cross-link of chitosan microspheres has an inuence on the rele ase b ehavior f or d rug, c hitosan m icrospheres, w hich when c ross-linked b y gen ipen m ay del ay a nd e xtend d rug delivery [47]. The pH-sensitive CG HPN microsphere has been prepared v ia t he i nverse su spension c ross-linking w ith GA . The c imetidine-loaded m icrospheres c an rele ase t he d rug alternately p ulsed a nd c hange t he p H o f t he me dium. The HPN microspheres appear to b e a promising carrier for drug delivery [48]. Chitosan–starch  bers a re e xcellent b iomaterials f or d rug release. Their release behavior can be controlled by changing the

10/3/2008 9:31:31 PM

10-7

Chitosan-Based Gels and Hydrogels

pH 1.0 pH 3.6 pH 5.0 pH 7.4

Cumulative release (%)

100

80

60

40

20

0

0

5

10

15

20 25 Time (h)

30

35

40

FIGURE 10.11 pH inuence of the release medium on the drug controlled release process from chitosan/starch bers (30/70, w/w).

composition r atio, t he d rug-loaded a mount, p H, a nd io nic strength o f t he rele ase s ystem. F or e xample, t he rele ase w as accelerated w ith t he decrease of t he pH (Figure 10.11), because the ele ctrostatic i nteraction b etween a nions a nd c hitosan w as greatly inuenced by the pH of the solutions; the decrease of pH also w eakened t he s alt b onds a nd t herefore, f acilitated  ber swelling. A s a re sult, t he d rug rele ase w as ac celerated. O n t he other hand, the pH also has a slight effect on the solubility of the drug. A higher pH leads to a better solubility of the drug, which results in a h igher drug release rate. However, compared to t he strong in uence of pH on t he  ber matrix, pH effects on drug could be neglected [49,50]. Macroporous ch itosan sc affolds re inforced by β-tricalcium phosphate ( β-TCP) a nd c alcium ph osphate i nvert g lasses a lso were used as a drug carrier for controlled drug release. This scaffold effectively reduced the initial burst release of antibiotic gentamicin sulfate (GS) in PBS. In comparison with pure chitosan scaffolds, t he c omposite s caffolds s howed a lo wer i nitial b urst release followed by a slow sustained release of GS up to 3 w eeks after i mmersing t hem i n P BS. A p ossible me chanism f or c ontrolled GS release from the composite scaffolds might be the formation o f s trong c ross-linking b etween t he a mine g roups o f chitosan a nd ph osphate io ns i n P BS s olution. The composite scaffolds would form more c ross-linked b onds t han t he origin chitosan scaffold due to the presence of a higher concentration of phosphates in PBS solution [51]. Protein a nd en zymes re present a g rowing a nd p romising eld of t herapeutics a nd a re c urrently ad ministrated by i njection. As a result of development of biotechnology, many peptides a nd p roteins h ave b een e xplored a s a ne w gener ation o f drugs. Maintaining their bioactivity in the drug delivery is one of the key problems for these drugs. Oral peptide drug delivery is t he e asiest me thod of ad ministration; however, proteins a re

43722_C010.indd 7

quickly denatured and degraded in the stomach. Moreover, the intestinal t ransport o f p eptide d rugs i s gener ally v ery p oor. Chitosan can enhance the intestinal transport of peptide drugs by increasing the paracellular permeability of the intestinal epithelium. For example, chitosan nanocapsules w ill increase t he intestinal absorption of salmon calcitonin [52,53], and the extra PEG a ffect with the release behavior of the resulting nanocapsules for chitosan–PEG nanocapsules. Additionally, the increase in the pegylation degree of chitosan can lead to a signicant decrease in the percentage of salmon calcitonin released. This could be related to b oth t he g reater a mount o f p eptide associated w ith th e c hitosan–PEG n anocapsules an d t o th e different composition and organization of the coating [30]. The p rogress i n b iotechnology, c oupled w ith a n i ncreased understanding o f t he mole cular me chanism o f a v ariety o f diseases at the gene level, has effected dramatic changes in therapeutic modalities. Recombinant DNA itself has been used like a “drug” i n gene t herapy, w here gene s a re app lied to p roduce therapeutic proteins i n t he pat ient. Ol igonucleotides, relatively small s ynthetic D NA de signed to h ybridize s pecic mRNA sequences, have been used to block gene expression. To achieve these goals, the gene drugs must be administered via an appropriate ro ute a nd b e del ivered i nto t he i ntercellular si te o f t he target cells where gene expression occurs. Gene d rugs h ave sub stantial p roblems suc h a s p olyanionic, nucleic, a nd lo w p ermeability. Therefore, a su itable c arrier system is the key to successful in vivo gene therapy. Considering that viral vectors have a number of potential limitations involving safety, a c ationic l ipid polymer developed as N DA carriers can improve in vivo lipid transfection efficiency. Chitosan has been considered to be a good candidate for gene delivery systems. It has also been found to interact with negatively charged DNA in a similar fashion [54]. Chitosan h as b een s hown to c ondense D NA e ffectively and protect DNA from nuclease degradation [55,56]. However, poor water s olubility a nd low t ransfection e fficiency of chitosan are major drawbacks for its use as a gene del ivery carrier [57]. The water solubility of chitosan can be improved by various approaches suc h a s qu aternization o f t he a mino g roup, N-carboxy methylation, and PEGylation. An N-dodecylated chitosan (CS-12) was synthesized and assembled with DNA to form a DNA–CS-12 complex. Its AFM image (Figure 10.12) revealed that the DNA was aggregated and the complex appeared globular in structure. The diameter varied from 90 to 230 nm, and the average height was approximately 18 nm. A si ngle globule consisted of 40–115 DNA molecules [58,59]. Yang et al. [31] synthesized the folate-PEG-grafted chitosan (FA–PEG–CS) for targeted plasmid D NA del ivery to t umor c ells, a nd t he re sults s howed that t he w ater s olubility o f F A–PEG–CS a nd P EG–CS w ere excellent and increased as a function of the PEG grafti ng degree. Moreover, there was strong binding between the modied chitosan and the DNA [60]. The PEG–CS DNA complexes injected via the bile duct showed an improved gene expression in a rat liver compared with unmodied chitosan [61].

10/3/2008 9:31:31 PM

10-8

Smart Materials

FIGURE 10.12 AFM images of DNA–CS-12 complex. The chitosan– DNA c harge r atio w as 2/1 a nd t he me asurement w as c arried out i n a tapping mode.

10.1.3.2

Separation Membranes

Chemically m odied chitosan membranes have been used in various  elds. F or e xample, me tal-ion s eparation, ga s s eparation, reverse osmosis, prevaporation separation of alcohol–water mixtures, ultra ltration of biological macromolecular products, and the affinity precipitation of protein isolates. Prevaporation i s a v ery u seful memb rane s eparation te chnique for separating organic liquid mixtures, such as azetropic mixtures and mixtures of materials that have boiling points. In prevaporation, the characteristics of permeation and separation are signicantly governed by the solubility and diff usion of the permeate. The separation mechanism of prevaporation is ba sed on t he solution-diff usion th eory, i n th at th e a dsorption–diffusion– desorption process of t he component i n t he feed comes ac ross the membrane from one side to the other. The refore, prevaporation properties can be improved by enhancing the adsorption of one component in the feed to the membrane or accelerating the diff usion of one component in the feed through the membrane. Chitosan has been used to from prevaporation membranes for separating water–alcohol mixtures and has shown good performance i n deh ydrating a lcohol s olutions. C S/PVA blen d mem branes have been formed by aqueous solution-casting, followed by c ross-linking w ith f ormaldehyde o r GA ; t hese memb ranes showed i mproved st rength a nd  exibility u nder b oth w et a nd dry conditions as compared to those in membranes of the individual c omponent. H owever, c ross-linking o f t he memb ranes with GA re duced t he h ydrophilicity o f t he blen d memb ranes [62,63]. CS/PVA blend membranes exhibited higher  ux than a CS membrane. With an increasing amount of PVA in the blend, polymer c hains b ecame more  exible, t hereby i ncreasing t he free-volume s pace o f t he blen d m atrix a nd i ncreasing t he  ux [64]. Wi th a n i ncreasing memb rane t hickness, s electivity h as improved, but the ux has decreased [65].

43722_C010.indd 8

Smart p olymers a re t he ba sis f or a ne w p rotein i solation technique–affin ity precipitation. The p recipitation app lies a ligand coupled to a w ater-soluble polymer known as an affinity macroligand, which forms a complex with the target. The phase sepa ration o f t he c omplex t riggered b y sm all c hanges i n environment, for example, pH, temperature, ionic strength, or addition o f re agent, m akes t he p olymer bac kbone i nsoluble; afterward, the target protein can be recovered via elution or dissolution. Chitosan i tself h as b een suc cessfully u sed a s a m acroligand for affinity precipitation of isolated wheat germ agglutinin and formed its extract and glycosides from cellulose preparation by changes in the pH of media [66]. Long-term s tability o f op eration i s c ritical f or memb ranes when used in commercial practice. Chitosan–silica hybrid material (CSHM) ba sed memb ranes, w hich w ere ob tained t hrough blending c hitosan an d γ-(glycidyloxypropyl)trimethoxysilane (GPTMS) i n ac etic ac id aque ous s olution, h ad a sig nicant improvement i n t heir lo ng-term s tability. C SHM-5 ( GPTMS with 5% (wt %) ) had an initial ux of 1730 g/(m2 h) and showed the b est lo ng-term s tability o n p revaporative deh ydration o f 70 wt % . i sopropanol–water (I PA–H2O) s olution ( as s hown i n Table 10.1). The dehydration performance of CSHM-5 was maintained after 140 days where CSHM-10 failed in 4 days in the test owing to its brittleness [67]. Yao et al. prepared CG network  lms cross-linked by GA and synthesized nanohydroxyapatite (nHA) on it in situ. The results showed t hat t he e xit c hemical i nteractions b etween n HA a nd CG tem plates, t he C OO−, C =O, o r – NH2 g roups m ay b e t he especially ac tive si tes f or t he c oordination o f Ca 2+ to form ion complexes, which are the active sites of nucleation and the formation nuclei of nHA controlled by these special groups. In order to mo dulate t he si ze o f n HA c rystalline, o ne c an c hange t he charged den sity o f a C G ne twork b y c hanging t he c ontent C S and Gel in the CG network  lm [68]. Rusua et al. also prepared the si ze-controlled h ydroxyapatite na noparticles i n a c hitosan matrix [69]. Chitosan is also a cationic polyelectrolyte, and chitosan membranes have been used for active chloride ion conductivity in aqueous solutions. In the same way, if these membranes were set in a closed circuit, this kind of active ionic conductivity should be present. In its natural state (dry state), a chitosan membrane TABLE 10.1 Long-Term Stability Test on Various Membranes for Pervaporation Dehydration of 70 wt % Isopropanol–Water (IPA–H2O) Solution, where X (wt % Based on Chitosan) Indicates the Amount of GPTMS Added into the Preparation Solution Membrane Pure CS CSHM-5 CSNM-2.5 CSHM-7.5 CSHM-10

Thick ness (µm) 11.8 ± 1.4 12.8 ± 1.0 12.0 ± 1.2 13.8 ± 0.5 14.2 ± 0.4

Initial Flux (g/(m2 h) )

Membrane Life (days)

4470 2000 1730 1560 1240

1 47 141 70 4

10/3/2008 9:31:32 PM

10-9

Chitosan-Based Gels and Hydrogels

has very low electrical conductivity. However, when it is swollen in w ater, i ts a mino g roups c an b e p rotonated a nd t hus m ay contribute to io nic c onduction i n t he memb rane. A c hitosan membrane made from unmodied chitosan with a high molecular weight and an appropriate DDA has an ionic conductivity of around 10 −4 S cm−1 after hydration for 1 h at 25°C. However, this i s s till rel atively lo w f or p ractical app lications [ 70]. Compared w ith t he u ncross-linked c hitosan memb rane, t he cross-linked chitosan membranes with lower DDA values and higher molecular weights have a relatively high conductivity [71] a nd t he io nic c onductivity o f h ydrated mo died chitosan membranes has been greatly improved over that of unmodied chitosan membrane. Si nce ionic c onductance w ill o ccur only after memb ranes a re h ydrated, i t c an b e c oncluded t hat t he hydrated properties of membranes will critically affect the ion permeability t hrough t he memb ranes, a nd t he memb ranes with a rel atively h igh s welling i ndex m ay a llow io ns to go through more easily in the swollen state of the membranes [72]. Lithium ( Li+) a nd p roton (H +)-conducting c hitosan-based electrolytes have been prepared using the solution cast technique a nd h ave s hown e xcellent c onductivity [ 73]. The chitosan-based polymer electrolyte has the potential for use in practical electrochemical devices. 10.1.3.3

Immobilization Supports

Immobilization i s a n e ffective me asure f or u sing en zymes o r microbial cells as recoverable, stable, and specic industrial biocatalysts. Immobilization must be carried out under mild conditions. Immobilized enzymes can be easily separated from soluble reaction products and unreacted substrate, thus simplifying the work-up a nd p reventing p rotein c ontamination o f t he  nal product [74]. An ideal support for enzyme immobilization should be chosen to achieve essential properties such as chemical stability, hydrophilicity, r igidity, me chanical s tability, l arge su rface a rea, a nd resistance to microbial attack. Several methods have been used to modify the supports to improve stability and mechanical strength, and to modify diff erent f unctional groups t hat may be superior for en zyme i mmobilization. O ne o f t he w ays to i mprove t hese properties is the grafting of a monomer onto a matrix. Chitosan has been used as a supp ort for enzyme immobilization, since it offers considerable advantages such as form versatility (powder, gel b eads,  ocks, bers, c apsules, a nd memb ranes), b iodegradability, low toxicity, high affinity toward proteins and good biocompatibility [75]. C hitosan b eads w ere u sed f or i mmobilizing pepsin and the pepsin immobilized onto chitosan beads exhibits an i mproved re sistance a gainst t hermal denat uration, t hermal stability, as well as storage stability of immobilized pepsin which was greater than that of free pepsin [76]. Hydrophobically modifying chitosan can increase the enzyme activity, which is a more optimal matrix for enzyme immobilization [77]. Pancreatic lipase, fungal laccases, cellulose, trypsin, glutamater dehydrogenase, p enicillin a cylase, a nd g lucosidase h ave b een immobilized via chitosan gels. The result is that immobilization improves the stability of enzymes.

43722_C010.indd 9

10.1.3.4

ECM for Tissue Engineering

Various s ynthetic a nd nat urally der ived h ydrogels h ave re cently been used as articial ECMs for cell immobilization, cell transplantation, an d ti ssue e ngineering. N ative E CMs ar e c omplex chemically a nd physically c ross-linked ne tworks of prot eins a nd GAGs. Articial ECMs replace many functions of the native ECM, such as organizing cells into a 3-D architecture, providing mechanical integrity to new tissue, and providing a hydrated space for the diffusion of nutrients and metabolites to and from cells. Chitosan is similar to GAGs. Therefore, it is promising for application as a biomaterial in addition to its use as a controlled delivery matrix. Chitosan is a basic polysaccharide, so it is possible to evaluate the percentage of amino functions, which remain charged at the pH of cell cultivation (7.2–7.4). These cationic charged materials have a de nite inuence o n c ell at tachment b y t heir p ossible interaction w ith ne gative c harges lo cated at t he c ell su rface. Nomizu a nd coworkers developed biomaterials using i ntegrinbinding peptides covalently bound to c hitosan membranes and demonstrate t hat p eptide–chitosan m embranes c an r egulate specic integrin-mediated cell responses and are useful constructs as ECM mimetics [78]. In order to improve the cytocompatibility, Yao et al. prepared the CG–hyaluronic acid scaffold. The scaffold showed a h igher water uptake and retention ability and  broblast and keratinocytes were cocultured in the scaffold. The cells grew and proliferated well and t hey exhibited a s trong v iability. Keratinocytes were cocultured with broblasts i n C G–hyaluronic ac id s caffolds a nd t his construct resulted i n a n a rticial bilayer skin in vitro. The articial s kin o btained w as  exible a nd h ad go od mechanical p roperties, w hich su ggested t hat C G–hyaluronic acid scaffolds are suitable for preparing a bilayer skin substitute [79,80]. Mesenchymal s tem c ells ( MSC), a h eterogeneous s tem c ell population within bone marrow, are reliant in the extracellular environment and not only can they survive but also can differentiate with a similar structure to GAGs and collagen. The PEC derived from gel atin a nd c hitosan h as a lso b een u sed a s supp orting scaffolds, m imicking t he nat ural EC M, f or t issue eng ineering bone and cartilage with promising results. Zhao et al. examined the effects of the incorporation of hydroxyapatite in the chitosan and gel atin p olymer ne twork o n h uman M SC m icroenvironment formation and 3D tissue development in vitro. The human MSCs maintained the higher propagation and multilineage differentiation potentials with an enhanced osteogenic differentiation upon induction in 3-D scaffolds [81]. 10.1.3.5 Field-Responsive Materials Chitosan-based gels consist of a positive charged network and a uid (e.g., water) that  lls t he i nterstitial space of t he network. The gel s e xhibit a v ariety of u nique  eld-responsive behaviors, such as electromechanical (EMC) phenomena. EMC b ehavior de als w ith t he c ontraction o f p olymers i n a n electric  eld. The effect o f a n e lectric  eld o n p olyelectrolyte hydrogels relates to the proton action of its alkaline amino groups

10/3/2008 9:31:32 PM

10-10

Smart Materials 45 40

Degree of bending (%)

35 30 25 20 15 10 5 0 Off

On 0

4

8

12

16

20 24 Time (s)

28

32

36

40

44

FIGURE 10.13 Bending kinetics of the chitosan/PHEMA semi-IPN, in a 1.0 wt % NaCl solution, voltage (10 V), at T = 35°C.

and the redistribution of mobile counter ions. The EMC behavior of c hitosan a nd p oly(hydroxyethyl me thacrylate) i n a 1 .0 wt % NaCl solution in a 10 V electric eld is shown in Figure 10.13 [21]. The gure a lso s hows t hat t he re versible b ending b ehavior o f chitosan–PAN semi-IPN hydrogel depends on the application of the electric eld [82].

8. 9.

References 1. J.-K.F. S ub a nd H.W.K. M atthew, Application o f c hitosanbased polysaccharide biomaterials in cartilage tissue engineering: A review, Biomaterials 21, 2589–2598, 2000. 2. G.A.F. Rob ert a nd K.E. T aylor, The f ormation o f g els b y reaction of chitosan with glutaraldehyde. Makromol. Chem. 190, 995–960, 1989. 3. M. Ruiza, A.M. Sastreb, and E. Guibal, Palladium sorption on glutaraldehyde-crosslinked chitosan, React. Funct. Polym. 45, 155–173, 2000. 4. M.N. K halid, L. H o, J .L. Agnely, J .L. G rossiord, a nd G. C ouarraze, Swelling properties and mechanical characterization of a s emiinterpenetrating chitosan/polyethylene oxide network; C omparison with a c hitosan reference gel, STP Pharm. Sci. 9, 359–364, 1999. 5. S. Kim, M.E. Nimni, Z. Yang, and B. Han, Chitosan/gelatin– based  lms cr osslinked b y p roanthocyanidin. J. B iomed. Mater. Res. Part B: Appl. Biomater. 75B, 442–450, 2005. 6. Y.-C. K uo a nd C.-Y . Lin, Eff ect o f g enipin-crosslinked chitin–chitosan scaffolds with hydroxyapatite modications on the cultivation of bovine knee chondrocytes, Biotechnol. Bio Eng. 95(1), 132–144, 2006. 7. W.F. Lee and Y.J. Chen, Studies on preparation and swelling p roperties o f t he nis opropylacrylamide/chitosan

43722_C010.indd 10

10.

11. 12. 13. 14.

15.

16.

semi-IPN a nd IPN h ydrogels, J. Appl. P olym. S ci. 82, 2487–2496, 2001. X. Qu, A. Wirsen, and A.C. Albertsson, Novel pH-sensitive chitosan h ydrogels: S welling b ehavior a nd st ates o f wa ter, Polymer 41, 4589–4598, 2000. S.C. Chen, Y.C. Wu, F.L. Mi, Y.H. Lin, L.C. Yu, and H.W. Sung, A novel pH-sensitive hydrogel composed of N,O-carboxymethyl chitosan and alginate cross-linked by genipin for protein drug delivery, J. Control. Release 96, 285–300, 2004. K.D. Yao, T. P eng, M.F.A. G oosen, J.M. Min, a nd Y.Y. H e, pH-sensitivity o f h ydrogels ba sed o n co mplex f orming chitosan–polyether in terpenetrating p olymer netw ork, J. Appl. Polym. Sci. 48(2), 354–353, 1993. T. P eng, K.D . Yao, C. Yuan, a nd M.F .A. G oosen, S tructural changes of pH-sensitive chitosan/polyether hydrogels in different pH solution, J. Polym. Sci., Part A. 32(3), 591–593, 1994. T. Wang a nd S. G unasekaran, S tate o f wa ter in c hitosan– PVA hydrogel, J. Appl. Polym. Sci. 101, 3227–3232, 2006. T. Wang, M. Turhan, and S. Gunasekaran, Selected properties o f pH-s ensitive, b iodegradable c hitosan–poly(vinyl alcohol) hydrogel, Polym. Int. 53, 911–918, 2004. Y. Lin, Q . Chen, a nd H. L uo, P reparation a nd c haracterization of N-(2-carboxybenzyl)chitosan as a p otential pHsensitive h ydrogel f or dr ug deli very, Carbohyd. Re s. 342, 87–95, 2007. E.K. Sarkyt and B.S. Vladimir, Swelling, shrinking, deformation, and oscillation of p olyampholyte gels based on vinyl 2-aminoethyl ether a nd s odium acr ylate, Langmuir 15, 4230–4235, 1999. A. Sionkowska, M. Wisniewski, J. Skopinska, C.J. Kennedy, and T .J. Wess, M olecular in teractions in co llagen a nd chitosan blends, Biomaterials 25, 795–801, 2004.

10/3/2008 9:31:32 PM

10-11

Chitosan-Based Gels and Hydrogels

17. Z. Chen, X. Mo, and F. Qing, Electrospinning of collagen– chitosan complex, Mater. Lett. 61, 3490–3494, 2007. 18. K.D. Yao, Y.J. Yin, M.X. Xu, and Y.F. Wang, Investigation of pH-sensitive dr ug deli very syst em o f c hitosan/gelatin hybrid polymer network, Polym. Int. 38(1), 77–82, 1995. 19. G. I varez-Lorenzo, A. C oncheiro, A.S. Dub ovik, N.V . Grinberg, T .V. Bu rova, and V.Ya. Gr inberg, T emperaturesensitive chitosan-poly(N-isopropylacrylamide) in terpenetrated networks with enhanced loading capacity and controlled release properties, J. Control. Release 102, 629–641, 2005. 20. S.J. Kim, K.J. Lee, I.Y. Kim, D.I. Shin, and S. Kim, Temperature and pH-r esponse sw elling b ehavior o f p oly(2-ethyl-2oxazoline)/chitosan in terpenetrating p olymer netw ork hydrogels, J. Appl. Polym. Sci. 99, 1100–1103, 2006. 21. S.J. K im, S.R . S hin, S.M. L ee, I.Y . K im, a nd S.I. K im, Electromechanical properties of hydrogels based on chitosan and poly(hydroxyethyl methacrylate) in NaCl solution, Smart. Mater. Struct. 13, 1036–1039, 2004. 22. S.J. Kim, S.J. Park, M.-S. Shin, and S.I. Kim, Characteristics of electrical responsive chitosan/polyallylamine interpenetrating p olymer netw ork h ydrogel, J. A ppl..Polym. Sci. 86, 2290–2295, 2002. 23. Y. Yin, Z. Li, Y. Sun, and K. Yao, A preliminary study on chitosan/gelatin p olyelectrolyte co mplex f ormation, J. Ma ter. Sci. 40, 4649–4652, 2005. 24. K.D. Yao, H.L. Tu, F. Chang, J.W. Zhang, and J. Liu, pH sensitivity o f the sw elling o f a c hitosan–pectin p olyelectrolyte complex, Angew. Makromol. Chem. 245, 63–72, 1997. 25. K.Y. Lee, W.H. Park, and W.S. Ha, Polyelectrolyte complexes of sodium alginate with chitosan or its derivatives for microcapsules, J. Appl. Polym. Sci. 63(4), 425–432, 1997. 26. L. B echerán-Marón, C. P eniche, a nd W. Argüelles-Monal, Study of the interpolyelectrolyte reaction between chitosan and alginate: Inuence of alginate composition and chitosan molecular weight, Int. J. Biol. Macromol. 34, 127–133, 2004. 27. T. Yoshizawa, Y. Shin-ya, K.-J. Hong, and T. Kajiuchi, pH- and temperature-sensitive p ermeation thr ough p olyelectrolyte complex  lms co mposed o f chi tosan a nd p olyalkyleneoxide–maleic acid copolymer, J. Membrane Sci. 241, 347–354, 2004. 28. T. Yoshizawa, Y. S hin-ya, K.-J. H ong, a nd T. K ajiuchi, pHand temperature-sensitive release behaviors from polyelectrolyte complex  lms co mposed o f c hitosan a nd PAOMA copolymer, Eur. J. Pharm. Biopharm. 59, 307–313, 2005. 29. E. F ernandez-Megia, R . N ovoa-Carballal, E. Quiñoá, a nd R. Riguera, Conjugation of bioactive ligands to PEG-grafted chitosan a t the dist al end o f PEG, Biomacromolecules 8, 833–842, 2007. 30. C. P rego, D . T orres, E. F ernandez-Megia, R . N ovoaCarballal, E. Quiñoá, a nd M.J . Alonso, Chi tosan–PEG nanocapsules as ne w ca rriers f or o ral p eptide deliv ery effect o f c hitosan peg ylation d egree, J. Co ntrol. Re lease, 111, 299–308, 2006. 31. P. Chan, M. Kurisawa, J.E. Chung, and Y.-Y. Yang, Synthesis and characterization of chitosan-g-poly(ethylene glycol)-folate as

43722_C010.indd 11

32.

33. 34. 35. 36. 37. 38. 39.

40.

41.

42.

43. 44.

45. 46.

47.

a no n-viral ca rrier f or t umor-targeted g ene deli very, Biomaterials 28, 540–549, 2007. M.S. Beena, T. Chandy, and C.P. Sharma, Heparin immobilized chitosan–poly(ethylene glycol) interpenetrating network: Antithrombogenicity. Art. Cells Blood Subst. Immobil. Biotech. 23(2), 175–192, 1995. Y. Hu, H. Jiang, C. Xu, Y. Wang, and K. Zhu, Preparation and characterization of poly(ethylene glycol)-g-chitosan with water and organosolubility, Carbohyd. Polym. 61, 472–479, 2005. C.A. Landes and S. Kriener, Resorbable plate osteosynthesis of s agittal spli t oste otomies wi th ma jor b one mo vement, Plast. Reconstr. Surg. 111, 1828–1840, 2003. F. Yao, W. Chen, H. Wang, H. Li u, K. Yao, P . S un, a nd H. Lin, A study on cytocompatible poly(chitosan-g-l-lac tic acid), Polymer 44, 6435–6441, 2003. Y. F anglian, L. Cha ng, C. Wei, B . Yun, T . Zhi yuan, a nd Y. Kangde. Synthesis and characterization of chitosan grafted oligo(l-lac tic acid), Macromol. Biosci., 3, 653–656, 2003. A.R. Esker, C. Mengel, G. Wegner, Ultrathin lms of a polyelectrolyte with layered architecture, Science 280, 892–895, 1998. P.T. Hammond, Recent explorations in ele ctrostatic multilayer thin  lm assembly, Curr. O pin. C olloid In terf. S ci. 4, 430–442, 2000. Q. Feng, G. Zeng, P. Yanga, C.Wang, and J. Caia, Self-assembly and char acterization of p olyelectrolyte c omplex  lms of hyaluronic acid/chitosan, Colloid. Surface A 257–258, 85–88, 2005. C.-Y. Chen, J.-W. Wang, and M.-H. Hon, Polyion complex nanobrous structure formed by self-assembly of chitosan and poly(acrylic acid), Macromol. Mater. Eng. 291, 123–127, 2006. S. Ichikawa, S. Iwamoto, and J. Watanabe, Formation of biocompatible na noparticles b y s elf-assembly o f enzyma tic hydrolysates of chitosan and carboxymethyl cellulose, Biosci. Biotechnol. Biochem., 69(9), 1637–1642, 2005. G. Huai-min and C. X ian-su, Study of cobalt(II)–chitosan coordination polymer and its catalytic activity and selectivity for vinyl monomer polymerization, Polym. Adv. Technol. 15, 89–92, 2004. N. K obota, P ermeability p roperties o f c hitosan–transition metal complex membrane, J. Appl. Polym. Sci. 64, 819–822, 1997. R. T erreux, M. D omard, C. Viton, a nd A. D omard, Interactions study between the copper II ion and constitutive elemen ts o f c hitosan str ucture b y D FT calc ulation, Biomacromolecules 7, 31–37, 2006. D. Thacharodi a nd K.P . R uo, P ropranolol h ydrochloride release behavior of cross-linked chitosan membranes, J. Chem. Technol. Biotechnol. 58, 177–181, 1993. K.C. Gupta and F.H. Jabrail, Eff ects of degree of deacetylation and cross-linking on physical characteristics, swelling and r elease b ehavior o f c hitosan micr ospheres, Carbohyd. Polym. 66, 43–54, 2006. Y. Yuan, B.M. Chesnutt, G. Utturkar, W.O. Haggard, Y. Yang, J.L. Ong, and J.D. Bumgardner, The effect of cross-linking of

10/3/2008 9:31:32 PM

10-12

48.

49. 50. 51. 52. 53. 54.

55. 56. 57. 58. 59. 60.

61.

62. 63. 64.

43722_C010.indd 12

chitosan m icrospheres w ith ge nipin on prote in re lease, Carbohyd. Polym. 68, 561–567, 2007. K.D. Yao, M.X. Xu, Y.J. Yin, J.Y. Zhao, a nd X.L. Chen, pHsensitive c hitosan/gelatin h ybrid p olymer netw ork micr ospheres for delivery of cimetidine, Polym. Int. 39, 333–337, 1996. Q. Wang, N . Z hang, X . H u, J . Yang, and Y. D u, C hitosan/ starch bers and their properties for drug controlled release. Eur. J. Pharm. Biopharm. 66, 39–404, 2007. Q. Wang, Y.M. Du, and L.H. Fan, Properties of chitosan/poly (vinyl alco hol)  lms f or dr ug co ntrolled r elease, J. Ap pl. Polym. Sci. 96(3), 808–813, 2005. Y. Zha ng a nd M. Zha ng, C alcium p hosphate/chitosan composite scaffolds controlled in vitro antibiotic drug release, J. Biomed. Mater. Res. 62, 378–386, 2002. C. Prego, M. García, D. Torres, and M.J. Alonso, Transmucosal macr omolecular dr ug deli very, J. Co ntrol. Re lease, 101, 151–162, 2005. C. Prego, M. Fabre, D. Torres, and M.J. Alonso, Effi cacy and mechanism of a ction of ch itosan n anocapsules for or al peptide delivery, Pharm. Res. 23(3), 549–556, 2006. F.C. M acLaughlin, R .J. M umper, J . Wang, J .M. Tagliaferri, I. G ill, M. H inchcliffe, et al., C hitosan a nd dep olymerized chitosan oligomers as co ndensing carriers for in vi vo plasmid delivery, J. Control. Release 56, 259–272, 1998. S. Mao, X. Shuai, F. Unger, M. Simon, D. Bi, and T. Kissel, The depolymerization of ch itosan: E ffects o n ph ysicochemical and biological properties, Int. J. Pharm. 281, 45–54, 2004. W.G. Li u a nd K.D . Yao, Chi tosan a nd i ts deri vatives—a promising nonviral vector for gene transfection, J. Control. Release 83, 1–11, 2002. M. Prabaharan and J.F. Mano, Chitosan-based particles as controlled drug delivery systems, Drug Deliv. 12(1), 41–57, 2005. W.G. Li u, K.D . Yao, a nd Q .G. Li u, F ormation o f a D NA/ N-dodecylated c hitosan co mplex a nd s alt-induced g ene delivery, J. Appl. Polym. Sci. 82, 3391–3395, 2001. W.G. Li u a nd K.D . Yao, Chi tosan a nd i ts deri vatives—a promising non-viral vector for gene transfection, J. Control. Release 83, 1–11, 2002. P. Chan, M. Kurisawa, J.E. Chung, and Y.-Y. Yang, Synthesis and char acterization of chi tosan-g-poly(ethylene g lycol)folate as a non-viral carrier for tumor-targeted gene delivery, Biomaterials, 28, 540–549, 2007. X. Jiang, H. Dai, K.W. Leong, H.Q. Hao, and Y.Y. Yang, PEGg-chitosan complexes for liver targets gene delivery, Abstract for I nternational C onference o n M aterials F or Advanced Technologies, P12, Singapore, 2003. Y.J. Ming, S.W. Su, L.T. Lang, and Y.M. Chen, Evaluation of chitosan/PVA blended hydrogel membranes, J. Membr. Sci. 236, 39–51, 2004. S.B. B ahrami, S.S. Kordestani, H. Mirzadeh, and P. Mansoon, Poly(vinyl alcohol)–chitosan blends: Preparation, mechanical and physical properties, Iranian Polym. J. 12, 139–146, 2003. K.S.V. K rishna R ao, M.C.S. S ubha, M. Sa iram, N.N. Mallikarjuna, a nd T.M. Aminabhavi, B lend mem branes o f

Smart Materials

65.

66. 67.

68. 69.

70. 71. 72.

73. 74. 75. 76. 77.

78.

79.

80.

chitosan and poly(vinyl alcohol) in pervaporation dehydration of isopropanol and tetrahydrofuran, J. Appl. Polym. Sci. 103, 1918–1926, 2007. D. Anjali D evi, B . Smi tha, S. Sr idhar, T .M. Aminabhavi, Dehydration o f 1,4-dio xane thr ough b lend mem branes of p oly(vinyl alco hol) a nd c hitosan b y p ervaporation, J. Membrane Sci. 280, 138–147, 2006. I.Y. Galaev, M.N. Gupta, and B. Mattiasson, Use smart polymers for bioseparations, Chemtech, 26(12), 19–25, 1996. Y.-L. Li u, Y.-H. S u, K.-R . L ee, a nd J .-Y. L ai, Cr osslinked organic–inorganic hybrid chitosan membranes for pervaporation dehydration of isopropanol–water mixtures with a long-term stability, J. Membrane. Sci. 251, 233–238, 2005. J. Li , Y.P. Chen, Y. Yin, F. Yao, a nd K. Yao, M odulation o f nano-hydroxyapatite size via formation on chitosan-gelatin network lm in situ, Biomaterials, 28(5), 781–790, 2007. V.M. R usua, C.-H. Ng, M. Wilke, B . T iersch, P. F ratzl, a nd M.G. Pete, Size-controlled hydroxyapatite nanoparticles as self-organized o rganic–inorganic co mposite ma terials, Biomaterials, 26, 5414–5421, 2005. Y. Wan, K.A.M. Creber, B. Peppley, and V.T. Bui, Ionic conductivity of chitosan membranes, Polymer 44, 1057–1065, 2003. Y. Wan, K.A.M. Cr eber, B . P eppley, a nd V. Tam B ui, I onic conductivity and related properties of crosslinked chitosan membranes, J. Appl. Polym. Sci. 89, 306–317, 2003. Y. Wan, K.A.M. Cr eber, B . P eppley, a nd V. Tam B ui, I onic conductivity and te nsile prop erties of h ydroxyethyl and hydroxypropyl c hitosan mem branes, J. P olym. S ci. P art B: Polym. Phys. 42, 1379–1397, 2004. T. Winie, S.R. Majid, A.S.A. Khiar, and A.K. Arof, Ionic conductivity of chitosan membranes and application for electrochemical devices, Polym. Adv. Technol. 17, 523–527, 2006. J. L alonde a nd A. M argolin, I mmobilization o f enzymes. I n: Drauz, K. a nd Waldmann, H., e ditors. Enzyme Ca talysis i n Organic Synthesis. Weinheim: Wiley–VCH, pp. 163–184, 2002. N. Ha and S.P. Meyers, Preparation and characterization of chitin and chitosan, a review, J. Aquat. Food. Prod. Tech. 4, 27–52, 1995. G.D. Altun and S.A. Cetinus, Immobilization of pepsin on chitosan beads, Food Chem. 100, 964–971, 2007. T. Klotzbach, M. Watt, Y. Ansari, and S.D. Minteer, Effects of hydrophobic modication of chitosan and Naon on transport p roperties, io n-exchange ca pacities, a nd enzyme immobilization, J. Membrane. Sci. 282, 276–283, 2006. M. M ochizuki, N. Yamagata, D . P hilp, K. H ozumi, T. Watanabe, Y. K ikkawa, Y. K adoya, H.K. K leinman, a nd M. N omizu, I ntegrin-dependent cell b ehavior o n ECM peptide-conjugated chitosan membranes, Biopolymers (Pept. Sci.), 88(2), 122–130, 2007. J.S. Mao, H.F. Liu, Y.J. Yin, and K.D. Yao, The prop erties of chitosan–gelatin mem branes a nd s caffolds modied with hyaluronic acid b y diff erent methods, Biomaterials, 24, 1621–1629, 2003. H. Liu, J. Mao, K. Yao, G. Yang, L. Cui, and Y. Cao, A study on a chitosan–gelatin–hyaluronic acid scaffold as articial skin

10/3/2008 9:31:32 PM

10-13

Chitosan-Based Gels and Hydrogels

in vitro and its tissue engineering applications, J. Biomater. Sci. Polymer Edn. 15(1), 25–40, 2004. 81. F. Zhao, W.L. Graysona, T. Maa, B. Bunnellb, W.W. Lu, Effects of hydroxyapatite in 3-D chitosan–gelatin polymer network on human mesenchymal stem cell co nstruct development, Biomaterials, 27, 1859–1867, 2006. 82. S.J. Kim, S.R. Shin, J.H. Lee, S.H. Lee, and S.I. Kim, Electrical response characterization of chitosan/polyacrylonitrile hydrogel in NaCl solutions, J. Appl. Polym. Sci. 90, 91–96, 2003.

10.2

Chitosan-Based Hydrogels in Biomedical and Pharmaceutical Sciences

Claire Jarry and Matthew S. Shive

demonstrated unique properties, which it brings to c hitosanbased h ydrogels, i ncluding b iocompatibility a nd i f de sired, biodegradability, a s w ell a s m eeting a cceptable l evels o f effi cacy a nd s afety. I ndeed, c hitosan-based hydrogels h ave b een investigated for specic applications in areas as varied as drug delivery systems for controlled release of drugs [4,5] and proteins [6], to scaffolds for tissue engineering [7]. In addition to hydrogels, chitosan in the form of  lms, beads, microspheres, and s olutions h as b een well-studied [8], a nd owing to c hitosan’s b iological a nd ph ysicochemical p roperties, h ave le d to the d evelopment of n umerous i nnovations not on ly i n t he pharmaceutical [8–10] and biomedical [11] industries, but also cosmetics [12], paper improvement [13], waste water treatment [14], f ood a nd a griculture [ 15], a nd te xtile [ 16] i ndustries. Numerous p roperties a nd appl ications of c hitosan a re p resented in Table 10.2.

10.2.1 Introduction In t he pa st few decades, hydrogels ba sed on nat ural biopolymers such as chitosan, collagen [1], alginate [2], and hyaluronic acid [3] h ave b een w idely s tudied b ecause o f t heir b iological properties and potential uses in biomedical and pharmaceutical f ields. T he p olysaccharide c hitosan, i n pa rticular, h as

10.2.2

Chitosan and Chitosan Derivatives

Chitosan is a natural copolymer of d-glucosamine and N-acetyld-glucosamine u nits b elonging to t he f amily o f β(1–4)-linked polysaccharides. C hitosan i s der ived f rom c hitin, w hich i s

TABLE 10.2 Applications of Chitosan and Chitosan Derivatives Property Absorption enhancement

Pharmaceutical

Angiogenic

Pharmaceutical and biomedical Food and agriculture Cosmetics Pharmaceutical and biomedical Biomedical Pharmaceutical Waste water treatment Chemistry Textile Paper Biomedical

Antimicrobial Cationic Cell substrate Chelation

Coating

43722_C010.indd 13

Industry

Cytocompatibility

Cosmetics Biomedical

Hemostatic Gelling Mucoadhesive

Biomedical Pharmaceutical Biomedical

Osteoconductive Wound healing

Biomedical Biomedical

Applications Increase permeability of tight junctions for hydrophilic macromolecules for oral drug and nasal vaccine deliveries Cartilage regeneration and tissue engineering Seed and food preservation Body lotion, toothpaste Gene delivery through anionic complexation, Drug delivery from microspheres Bone regeneration Fat and cholesterol binding Removal of metal ions Separation of inorganic anions Dye sorption Paper mechanical properties improvement Improvement of osseointegration of orthopedic devices Hair conditioning Intervertebral disk regeneration bone and cartilage regeneration Hemostatic dressings Local drug delivery Postsurgical adhesion prevention and tissue adhesive Cartilage regeneration Bone regeneration Wound occlusion Wound dressings

References [6,86–88]

[4,76] [15,83,89–91] [12] [46,47,92,93] [66,94] [95,96] [14] [97] [98] [13] [99] [100] [41,73,79,81] [101–103] [5,59] [31,32] [75–77] [61,68,104] [85,105,106]

10/3/2008 9:31:33 PM

10-14

Smart Materials OH

OH

OH

O

O O

HO

O O

HO

OH

O

HO

OH

OH

Cellulose OH

OH

OH

O

O O

HO

O

HO

NH C

O O

HO

NH O

C

CH3

NH O

C

CH3

O

CH3

Chitin OH

OH

OH

O

O O

HO

O O

HO

NH2

O

HO

NH2

NH2

Chitosan

FIGURE 10.14 Structures of cellulose, chitin, and chitosan.

plentiful in nature, and is found in exoskeletons of crustaceans, as well as some insects, yeast, and fungi. Very similar to cellulose in structure, chitin has N-acetyl groups on its C2 units as opposed to t he h ydroxyl g roups o f c ellulose ( Figure 1 0.14). The partial substitution of chitin’s N-acetyl groups for NH2 groups (deacetylation) under alkaline conditions yields chitosan and the degree of substitution is called DDA. Chitosan with a DDA between 70% a nd 95 % h as s hown ut ility i n t he ph armaceutical a nd biomedical  elds. A t o r b elow i ts p Ka of ∼6.3 [17], c hitosan i s positively charged in dilute acids since the free NH2 groups are protonated, t hus o vercoming t he a ssociative f orces b etween chains and permitting solubilization. On the other hand, if the solvent p H i s a bove t he p Ka, t he f raction o f p rotonated N H2 groups on the chitosan is insufficient and the polymer remains insoluble. DDA is one parameter that plays an important role in the b iodegradability o f c hitosan, a mong ot her t hings. I ndeed, the degradation of chitosan in vivo may occur primarily through the endogenous enzyme lysozyme, and is inversely proportional to its DDA, i .e., a lo wer DDA permits degradation to a g reater extent than a higher DDA, and in cases of very high DDA (>90%), enzymatic degradation slows such that these chitosans are considered nonbiodegradable [18–21]. Further chemical modification of the chitosan chain other than deacetylation has been studied as a way of bringing new functionalities t o t his p olymer. D erivatization o f c hitosan through c hemical sub stitution o f t he p rimary N H 2 gr oups can result in desired functional groups randomly distributed along t he bac kbone. T he mo dified c hitosan, o r c hitosan

43722_C010.indd 14

derivative, a nd t he sub stituted f unctional g roups c an b ring improved ph ysicochemical prop erties to t he p olymer. For example, t he mucoadhesivity of chitosan for a d rug delivery application was improved using thiol substitution. The thiolated chitosan enhanced the pH-dependent formation of both inter- and intramolecular disulfide bonds with mucus glycoproteins, and prolonged the controlled release of therapeutic agents b y p roviding a s trong c ohesion a nd s tability to t he delivery system [22].

10.2.3 Hydrogels By de nition, a gel is a colloidal dispersion (dispersed solid in a liquid solvent) that spontaneously forms a homogenous and  ne network (coagulation). Both t he liquid and solid are in a continuous phase bringing attributes of both to the gel. By convention, w hen t he s olvent i s w ater, t he ter m h ydrogel i s used. A xerogel describes the removal of the liquid phase leaving only the gel network structure. In the pharmaceutical eld, hydrogels a re c ommonly app lied a s d rug del ivery s ystems because they can be designed to respond to physiological environments. For instance, dramatic behavioral changes in swelling p roperties a nd me chanical s trength c an b e i nduced a s consequence of exposure to different stimuli such as pH, temperature, lo cal p rotein c oncentrations, a nd io nic s trengths [23]. Hydrogels can be classied into two groups, chemical or physical, de pending o n t he nat ure o f t he b onds b etween t he network chains.

10/3/2008 9:31:33 PM

10-15

Chitosan-Based Gels and Hydrogels

Chemical hydrogels a re s o c alled b ecause o f c ovalent b onding between h ydrophilic m acromolecules c rosslinking a t hreedimensional network. The stabilizing cross-links give these irreversible hydrogels the ability to swell in water while maintaining that network. Chitosan chemical hydrogels most commonly achieve t his c ross-linking u sing g lutaraldehyde [ 24–26], b ut sclearaldehyde [27], glyoxal [28,29], genipin [30], and ultraviolet irradiation [31,32] have also been employed. Use of such chemical cross-linkers, however, has raised safety concerns regarding human u se d ue to t he p otential to xicity o f f ree u nreacted molecules and has motivated further development of alternative types of chitosan hydrogels. 10.2.3.2 Physical Hydrogels Physical hydrogels are formed through the physical gelation of chitosan a s a c onsequence o f mo difying t he p olymers h ydrophilic–hydrophobic ba lance, w hich p ermits t he f ormation o f both hydrophobic interactions as well as hydrogen bonding. These hydrogels are often reversible because, as opposed to the chemical h ydrogels, t hey ar e m aintained v ia m uch w eaker intermolecular f orces, suc h a s ele ctrostatic o r h ydrophobic interactions as well as hydrogen bonding. Several parameters or conditions c an b e re gulated to c ontrol t he p roperties o f t hese gels, i ncluding p H, me dia, tem perature, c hitosan D DA, a nd charge densi ty. I ntelligent u se o f ch itosan der ivatives, i n pa rticular, c an f acilitate t he m anipulation o f t hese pa rameters, thus en hancing spec ic c hitosan in teractions a nd f urther improving the hydrogel characteristics, in general. This strategy was u sed w ith a p olyethylene g lycol ( PEG) der ivative t hat increased t he solubility of chitosan in neut ral pH [33], g rafted N-isobutyryl g roups, w hich i mparted t hermosensitivity to a hydrogel [34], alkyl chains that favored the formation of hydrophobic domains thus improving gelation [35,36], a polyampholyte h ydrogel, i .e., a h ydrogel t hat c ontains b oth p ositive a nd negative charges and thus, being pH-dependant, that was achieved via carboxymethyl groups [37], and an ethylenediaminetetraacetic acid (EDTA) derivative that improved the antimicrobial nature of topical gels [38]. Physical chitosan hydrogels can also be achieved through the complexation of polyelectrolytes of opposite charges, called ionotropic gelation [39]. Chitosan, by its cationic nature, will ionically i nteract w ith polyanion polymers, such a s a lginate [40,41], pectin [42,43], xanthan [44], collagen [45], and DNA [46,47]. A special group of physical hydrogels have been studied called injectable t hermosensitive hydrogels, which have been adapted for in vivo use, where they solidify in situ upon injection, and are attractive in many elds due to less invasive delivery. Although hydrogel systems, which gel with a decrease in temperature, may be more familiar (e.g., poly(vinyl alcohol), gelatin, or agar), those that a re l iquid at ro om tem perature o r b elow a nd re quire a n increase in temperature to solidify may be more amenable to pharmaceutical and biomedical applications. Poloxamer, a copolymer o f p oly(ethylene o xide) a nd p oly(propylene o xide), a nd

43722_C010.indd 15

poly(N-isopropyl ac rylamide) a re t he mo st s tudied s ystems o f this kind [48]. Few thermosensitive systems, however, have contained chitosan due to i ts physiologically incompatible low pKa (6.3), and the systematic formation of unusable hydrated gel-like precipitates at pH values above it. Attempts to develop chitosanbased systems that are both injectable and liquid at physiological pH, as well as thermosensitive have proven to be challenging. One such attempt grafted PEG onto chitosan to increase its solubility to a pH near 7. By controlling both the degree of substitution of PEG on the chitosan chain as well as the concentrations of b oth p olymers, a n i njectable t hermosensitive [30] a nd t hermoreversible [49] hydrogel was obtained, which gel led w ith a n increase in temperature above 25°C. While not fully understood, the mechanism of gelation seems to i nclude hydrophobic interactions b etween c hitosan p olymer c hains, h ydrogen b onding between hydroxyl and amino groups of chitosan, and intermolecular bonding between PEG chains. Another i nteresting s ystem ut ilizing c hitosan a nd p olyol salts was successfully de veloped [50,51], y ielding a n i njectable and thermosensitive hydrogel at a pH near 7 called chitosan-GP. Glycerophosphate (GP), a p olyol-phosphate s alt, w as adde d to an aqueous chitosan solution where it acted as a buffering agent and permitted chitosan solutions w ith pH values of 7 w ithout precipitation o r gel ation. I ncreasing tem peratures i ncreased gelation i nvolving a c ombination o f ele ctrostatic re pulsions between l ike-charged c hitosan c hains, ele ctrostatic at traction between t he opp ositely c harged c hitosan a nd G P, a nd h ydrophobic i nterchain at tractions. The tem perature o f g elation i n this chitosan–GP system can be controlled, as shown in Figure 10.15, p rimarily t hrough t he re gulation o f t he s olution p H. I f desired, gelation at human body temperature can be achieved, such as a solution with a 2% chitosan concentration and a pH of 7.2, w hich gel s at 3 7°C, w hile s light ac idication to pH 6.85 increases the gelation temperature to nearly 50°C.

80

Temperature (⬚C)

10.2.3.1 Chemical Hydrogels

70 Gel 60 50 40 Sol 30 6.4

6.6

6.8

7.0

7.2

7.4

pH

FIGURE 10.15 pH-Dependent gelation temperature of chitosan–GP solutions. (R eprinted f rom C henite, A ., Bu schmann, M ., W ang, D ., Chaput, C ., a nd K andani, N., Carbohydr. P olym., 4 6, 3 9, 2 001. W ith permission.)

10/3/2008 9:31:33 PM

10-16

Smart Materials

10.2.4 Chitosan Hydrogel Applications Chitosan-based hydrogels have been widely studied in the development of controlled drug delivery systems, tissue engineering, and other biomedical applications. 10.2.4.1

Drug Delivery Systems

Drug delivery systems are designed to act as depots, safely releasing physically entrapped bioactive molecules at a desired rate. In most cases, drug delivery systems based on chitosan or chitosan derivative hydrogels a re i nitially i n a d ry state, such a s  lms, tablets, o r mi crospheres. The d ry h ydrogel s wells f ollowing contact w ith b iological  uids, t hus p ermitting d issolution o f the drug and subsequently its diff usion and release through the polymeric matrix of t he hydrogel. Fu rthermore, t he ut ility of the chitosan matrix in protecting sensitive drugs, proteins, and peptides from harsh environmental conditions at t he delivery site is particularly interesting in applications targeting the gastrointestinal t ract, w here t hese p roteins a nd p eptides h ave short h alf-lives a nd a re p oorly a bsorbed w hen ad ministrated orally, due to their hydrophilic nature and large molecular size. The su sceptibility of suc h mole cules to de gradation by proteolytic enzymes a nd harsh conditions can be decreased by use of chitosan, improving their absorption from the gastrointestinal t ract w hen u sed i n t he f orm o f gel b eads [52,53], m icrospheres [ 54,55], or nonc ovalently c ross-linked c hitosan hydrogels [56,57]. Chitosan microspheres encapsulated insulin and p revented i ts p roteolytic en zymatic de gradation, a nd improved i ts b ioavailability b y c ontrolling t he rele ase r ate o f the d rug f rom t he m icrospheres [ 54]. p H-sensitive m icrospheres of c hitosan/sodium t ripolyphosphate/dextran s ulfate resisted hydrolysis in strong acid and biodegradation in enzymatic surroundings and achieved successful intestinal delivery of the hydrophobic drug, ibuprofen [58]. Typically, the systemic administration of repeated doses of a therapeutic a gent le ads to l arge p lasma d rug c oncentration variations. Sustained release of macromolecules was achieved over a p eriod of s everal hours to a f ew d ays, u sing t he c hitosan-GP in jectable f ormulation [ 5]. The s ame s ystem l oaded with pac litaxel f or lo cal del ivery demo nstrated t he a bility to inhibit tumor growth in a subcutaneous tumor mouse model at the same level as four intravenous injections of Taxol ® (registered t rademark of Br istol-Myers S quibb C o., New York), but with less toxicity [59]. Because the release is usually monitored by the diffusion of the drug through the matrix, a burst release could s till o ccur i n t he  rst h ours a fter i mplantation. T o manage t his b urst e ffect a nd to m aintain t herapeutic p lasma levels o f t he d rug, add itional c ross-links c an b e i ntroduced into a physical hydrogel to further reduce the mobility of polymer chains and thus slow drug release. Improved sustained release can be achieved using the addition of another interactive p olymer, suc h a s c hondroitin su lfate, i nto t he c hitosan hydrogel beads to prolong the release of drug [52], or by increasing gel hydrophobicity and thereby limiting both drug dissolution a nd c hain re laxation i n nonc onvalently c ross-linked

43722_C010.indd 16

hydrogels [56]. O ther techniques c an be employed to p rolong protein rele ase, suc h a s c ross-linking P EG-grafted chitosan with genipin, transforming a physical hydrogel into an insoluble chemical hydrogel [30]. Also, the rate of bioerosion and biodegradation o f t hese h ydrogels c an b e m anipulated b y u sing chitosans w ith lower DDAs, which degrade more r apidly a nd release drug faster than those containing more permanent chitosans with higher DDAs. In the latter case, the delivery rate will be more dependent on drug diff usion f rom t he hydrogel matrix than degradation. 10.2.4.2 Tissue Engineering Tissue eng ineering i s t he b iomedical f ield app roaching t he repair of damaged tissues through the creation of new tissue or the modification or repair of existing tissues. Hydrogels, in general, are an attractive material for tissue engineering because t hey a re s tructurally si milar to t he EC M o f m any tissues, c an o ften b e p rocessed u nder rel atively m ild c onditions, a nd m ay b e app lied i n v ivo i n a m inimally i nvasive manner. Because of regenerative and healing properties associated wi th c hitosan an d c hitosan d erivatives [11], c hitosan hydrogels have been used as scaffold for applications in tissue engineering of bone, cartilage and intervertebral disk, as well as in wound repair. 10.2.4.2.1

Bone Repair

Chitosan e xhibits o steoconductivity [ 60–62] i .e., i t supp orts the at tachment of new bone c ells (osteoblasts) a nd osteoprogenitor c ells, a nd f acilitate t he f ormation o f ne w b one (osteogenesis) [63]. C onsequently, i t h as b een e valuated a s a potential vehicle for bioceramics particles as an alternative to the gold standard of autograft ing (same donor/receptor) a nd its associated difficulties of procurement morbidity, increased operative time, and limited availability especially when treating large osseous defects. Due to a chemical composition similar to t hat o f nat ural b one a nd e xcellent o steoconductivity [64,65], calcium phosphate bioceramics have been mixed with biodegradable p olymers, w hich ac t a s c arriers to f acilitate implantation and to favor bone regeneration as both the polymer and the calcium phosphate degrade over time. And while biphasic calcium phosphate (BCP), tricalcium phosphate, and hydroxyapatite have been mixed with chitosan to produce sponges [66], pastes or composites [67–69], and cements [70], use o f c hitosan h ydrogels h as b een qu ite l imited [ 41]. T he chitosan–GP h ydrogel w as c ombined w ith B CP pa rticles, resulting i n a f ully i njectable s olution c apable o f f orming a solid bo ne g raft subst itute a t b ody tem perature [71]. BCP i s known to s timulate a nd p romote bone re generation [ 65,72], and c ombined w ith t he o steoconductivity o f c hitosan, t his injectable s ystem c ould b e app lied i n b oth o rthopedic a nd dental app lications. St udies u sing en vironmental s can ele ctronic microscopy (ESEM) showed that BCP particles are uniformly embedded into the hydrogel matrix (Figure 10.16) where it is hypothesized to act synergistically with chitosan–GP to stimulate the formation of new bone.

10/3/2008 9:31:33 PM

10-17

Chitosan-Based Gels and Hydrogels

or prosthetic device to restore the intervertebral space, and fusion of t he v ertebral b one. These t reatments p rovide s ymptomatic relief, but may compromise spine biomechanics and also exacerbate the degenerative process in adjacent spinal segments. Another approach would be to u se a del ivery system to i mplant, i.e., cells which can be transformed into a de sired intervertebral cell type. Ideally, the delivery system would be injectable, the delivered cells will e ventually r egenerate t he t issue i n t he i ntervertebral spac e, and the disk should obtain sufficient mechanical strength to resist long-term and cyclical biomechanical compressions. Hydroxybutyl chitosan [79], chitosan–genipin [80], and chitosan–GP [81] injectable t hermosensitive h ydrogels h ave b een e valuated f or suc h a delivery system. These gels have shown in early stage research to have the ability to maintain inter vertebral disk and mesenchymal stem c ell v iability, a lthough a ttaining s ufficient biomechanical properties in the disk is still problematic. 10.2.4.2.4

FIGURE 10.16 Granules of BCP embedded in chitosan-GP as viewed under ESEM. Addition of BCP to chitosan-GP results in a f ully injectable s olution c apable of for ming a s olid b one g raft subs titute at b ody temperature.

10.2.4.2.2

Cartilage Repair

Cartilage d amage re presents a m ajor c linical c hallenge f or orthopedic su rgeons b ecause c artilage do es not h ave i ntrinsic properties, p rimarily d ue to t he l ack o f v ascularity a nd a c ell population incapable of mounting a repair response. The se cells, called c hondrocytes, a re emb edded i n t he h ighly h ydrated cartilage tissue composed primarily of type II collagen and GAGs. Having a similar macromolecular structure to cartilage, hydrogels of different types have been investigated to mimic the cartilage EC M to w hich c ells c an at tach a nd p roliferate. This tissue eng ineering app roach re quires t he i solation o f a rticular chondrocytes or t heir precursor cells, followed by the cell proliferation i n v itro a nd t he seeding of c ultured cells into a hydrogel matrix that will eventually be implanted into the damaged cartilage area [73,74]. Unfortunately, processes involving in vitro cell culture for clinical use are prohibitive due to their complexity and extreme cost. Another form of chitosan-GP has been investigated in the treatment of articular cartilage damage of the k nee. The chitosan–GP matrix, when mixed with autologous whole blood and su rgically app lied, demo nstrated i n a nimals t he a bility to regenerate reproducible high-quality cartilage [75–77] through a mechanism i nvolving v ascularization o f t he u nderlying b one and an increase in stem cell migration. Called BST-CarGel ® (registered t rademark o f Bio s yntech C anada I nc., L aval, Canada), this material has advanced to the clinical testing stage as a medical device [78]. 10.2.4.2.3

Intervertebral Disk Replacement

Current t reatment o f i ntervertebral d isk de generation i nvolves surgical excision of the damaged tissue, insertion of a metal cage

43722_C010.indd 17

Wound Healing

Wound healing encompasses such clinical problems as chronic ulcers and burns. An ideal wound dressing should, in addition of being biocompatible, protect the wound from bacterial infection and provide a mo ist healing environment. A h ydrogel dressing made from chitosan, with its wound healing [82] and antimicrobial p roperties [ 83], c ould t herefore b e a p erfect m arriage o f these p roperties. I ndeed, ph otocross-linkable c hitosan h ydrogels applied on full thickness skin incisions on the back of mice signicantly induced wound contraction and accelerate wound closure [84]. And the negative effects of oxygen free radicals on wound healing were eliminated by the antioxidant taurine incorporated into a c hitosan hydrogel [85]. Finally, the chitosan–GP hydrogel has been applied to d iabetic ulcers on the foot. Called BST-DermOn™ (trademark o f Bio s yntech C anada I nc., L aval, Canada), this material has advanced to the clinical testing stage as a medical device [78].

10.2.5 Conclusion Chitosan-based hydrogels draw their versatility from the biological and physicochemical properties associated with chitosan and chitosan derivatives. As evidenced by their varied uses from controlled drug delivery to tissue engineering, the physical forms of chitosan hydrogels appear to hold the most promise due mainly to their improved safety. In particular, the development of injectable and thermosensitive chitosan hydrogels has opened the door for innovative treatments using less invasive delivery methods.

References 1. Purna, S.K. a nd B abu, M., C ollagen bas ed dr essings—a review. Burns, 26, 54, 2000. 2. Tonnesen, H.H. a nd K arlsen, J., Alginate in dr ug deli very systems. Drug. Dev. Ind. Pharm., 28, 621, 2002. 3. Price, R.D., et al., The role of hyaluronic acid in wound healing: assessment o f clinica l e vidence. Am. J . C lin. D ermatol., 6, 393, 2005.

10/3/2008 9:31:33 PM

10-18

4. Ishihara, M., et al., C hitosan hydrogel as a dr ug deli very carrier to control angiogenesis. J. Artif. Organs, 9, 8, 2006. 5. Ruel-Gariepy, E., et al., Characterization of thermosensitive chitosan gels for the sustained delivery of drugs. Int. J. Pharm., 203, 89, 2000. 6. Bernkop-Schnurch, A., Chi tosan a nd i ts deri vatives: potential ex cipients f or p eroral p eptide deli very syst ems. Int. J. Pharm., 194, 1, 2000. 7. Gutowska, A., Jeong, B., and Jasionowski, M., Injectable gels for tissue engineering. Anat. Rec., 263, 342, 2001. 8. Felt, O., Buri, P., and Gurny, R., Chitosan: a unique polysaccharide for drug delivery. Drug Dev. Ind. Pharm., 24, 979, 1998. 9. Singla, A.K. and Chawla, M., Chitosan: Some pharmaceutical and biological aspects–an update. J. Pharm. Pharmacol., 53, 1047, 2001. 10. Paul, W. and Sharma, C.P., Chitosan, a dr ug carrier for the 21st century: A review. S.T.P. Pharma. Sci., 10, 5, 2000. 11. Mattioli-Belmonte, M., M uzzarelli, B ., a nd M uzzarelli, R.A.A., Chi tin a nd c hitosan in w ound healin g a nd o ther biomedical applications. Carbohydr. Eur., 19, 30, 1997. 12. Muzzarelli, R., et al., Characteristic properties of N-carboxybutyl chitosan. Carbohydr. Polym., 11, 307, 1989. 13. Mucha, M. and Miskiewicz, D., Reactive chitosan blends for paper improvement, inChitosan in Pharmacy and Chemistry. Muzzarelli, R.A.A. and Muzzarelli, C., Eds., Atec, Italy, 2002, p. 251. 14. Lalov, I., e t al ., Treatment o f w aste w ater f rom d istilleries wtih chitosan. Wat. Res., 34, 1503, 2000. 15. El Gaouth, A., Arul, J., and Asselin, A., Potential use of chitosan in p ostharvest preservation of fruits and vegetables, in Advances in Chitin and Chitosan. Brine, C.J., Sandford, P.A., and Zikakis, J.P., Eds., Elsevier Applied Science, New York, 1992, p. 440. 16. Lim, S.H. and Husdon, M., Review of chitosan and its derivatives as antimicrobial agents and their uses as textile chemicals. J. Macromol. Sci., Part C Polym. Rev., C43, 223, 2003. 17. Muzzarelli, R .A.A., M uzzarelli, C., a nd T erbojevich, M., Chitin chemistry, upgrading a renewable resource. Carbohydr. Eur., 19, 10, 1997. 18. Gori, S., G élication et dég radation d’un biopolymère p ar voie enzyma tique, in M asters o f Applied S ciences in Biomedical Engineering, Montréal, École Polytechnique de Montréal, 2002. 19. Vårum, K.M., et al ., In vitro degradation rates of partially N-acetylated c hitosans in h uman s erum. Carbohydr. Re s., 299, 99, 1997. 20. Nordtveit, R.J., Vårum, K.M., and Smidsrød, O., Degradation of fully water-soluble, partially N-acetylated chitosans with lysozyme. Carbohyd. Polym., 23, 253, 1994. 21. Sashiwa, H., et al ., L ysozyme sus ceptibility o f pa rtially deacetylated chitin. Int. J. Biol. Macromol., 12, 295, 1990. 22. Bernkop-Schnurch, A., Hornof, M., and Guggi, D., Thiola ted chitosans. Eur. J. Pharm. Biopharm., 57, 9, 2004. 23. Gupta, P., Vermani, K., and Garg, S., Hydrogels: from controlled release to pH-responsive drug delivery. Drug Discov. Today, 7, 569, 2002.

43722_C010.indd 18

Smart Materials

24. Argüelles-Monal, W., et al., Rheological study of the chitosan/ glutaraldehyde chemical gel system. Polym. Gels Networks, 6, 429, 1998. 25. Goycoolea, F.M., et al ., Eff ect o f chemical cr osslinking o n the sw elling a nd shrinkin g p roperties of t hermal an d pH-responsive c hitosan h ydrogels. Macromol. B iosci., 3, 612, 2003. 26. Roberts, G.A.F. and Taylor, K.E., Chitosan gels, 3. The formation of gels by reaction of chitosan with glutaraldehyde. Makromol. Chem., 190, 951, 1989. 27. Guo, B., Elgsaeter, A., and Stokke, B.T., Gelation kinetics of sclealdehyde–chitosan co-gels. Polym. Gels Networks, 6, 113, 1998. 28. Park, H., P ark, K., a nd K im, D ., P reparation a nd sw elling behavior of chitosan-based superporous hydrogels for gastric retention application. J. Biomed. Mater. Res. A, 76, 144, 2006. 29. Park, H. and Kim, D., Swelling and mechanical properties of glycol c hitosan/poly(vinyl alco hol) IPN-typ e su perporous hydrogels. J. Biomed. Mater. Res. A, 78A, 662, 2006. 30. Bhattarai, N., et al ., PEG-gra fted c hitosan a s a n i njectable thermosensitive h ydrogel f or sust ained p rotein r elease. J. Control. Release, 103, 609, 2005. 31. Yeo, Y., et al ., P eritoneal a pplication o f c hitosan a nd UV-cross-linkable c hitosan. J. B iomed. Mater. R es. A, 78A, 668, 2006. 32. Ono, K., et al ., Photocrosslinkable chitosan as a b iological adhesive. J. Biomed. Mater. Res., 49, 289, 2000. 33. Ouchi, T., N ishizawa, H., a nd Oh ya, Y., Aggregation p henomenon o f PEG-gra fted chi tosan in aq ueous s olution. Polymer, 39, 5171, 1998. 34. Felix, L., et al., Kinetics of gelation and thermal sensitivity of N-isobutyryl chitosan hydrogels. Biomacromolecules, 6, 2408, 2005. 35. Rinaudo, M., et al ., Specic interactions in mo died chitosan systems. Biomacromolecules, 6, 2396, 2005. 36. Holme, K.R . a nd H all, L.D ., Chi tosan deri vatives b earing C10-alkyl glycoside branches: A temperature-induced gelling polysaccharide. Macromolecules, 24, 3828, 1991. 37. Chen, L., Tian, Z., and Du, Y., Synthesis and pH sensitivity of ca rboxymethyl c hitosan-based p olyampholyte h ydrogels f or p rotein ca rrier ma trices. Biomaterials, 25, 3725, 2004. 38. Valenta, C., Christ en, B ., a nd B ernkop-Schnurch, A., Chitosan–EDTA conjugate: a novel polymer for topical gels. J. Pharm. Pharmacol., 50, 445, 1998. 39. Berger, J ., et al., S tructure a nd in teractions in chi tosan hydrogels formed by complexation or aggregation for biomedical ap plications. Eur. J . Ph arm. B iopharm., 57, 35, 2004. 40. Lin, Y.H., et al ., P hysically cr osslinked algina te/N,O-carboxymethyl chitosan hydrogels with calcium for oral delivery of protein drugs. Biomaterials, 26, 2105, 2005. 41. Park, D.J., et al., Injectable bone using chitosan–alginate gel/ mesenchymal st em cells/BMP -2 co mposites. J. C raniomaxillofac. Surg., 33, 50, 2005.

10/3/2008 9:31:34 PM

Chitosan-Based Gels and Hydrogels

42. Marudova, M., M acdougall, A.J., a nd Rin g, S.G., P ectin– chitosan in teractions a nd g el f ormation. Carbohydr. Re s., 339, 1933, 2004. 43. Nordby, M.H., et al ., Thermoreversible gelation of aqueous mixtures o f p ectin a nd chi tosan. R heology. Biomacromolecules, 4, 337, 2003. 44. Chellat, F., et al ., I n vi tro a nd in vi vo b iocompatibility o f chitosan–xanthan polyionic complex. J. Biomed. Mater. Res., 51, 107, 2000. 45. Tan, W., Krishnaraj, R., and Desai, T.A., Evaluation of nanostructured composite collagen–chitosan matrices for tissue engineering. Tissue Eng., 7, 203, 2001. 46. Guang Li u, W. a nd D e Yao, K., Chi tosan a nd i ts deri vatives—a p romising no n-viral v ector f or g ene tra nsfection. J. Control. Release, 83, 1, 2002. 47. Borchard, G., Chitosans for gene delivery. Adv. D rug D eliv. Rev., 52, 145, 2001. 48. Ruel-Gariepy, E. a nd L eroux, J .C., I n si tu-forming h ydrogels—review of temperature-sensitive systems. Eur. J. Pharm. Biopharm., 58, 409, 2004. 49. Bhattarai, N., Matsen, F.A., and Zhang, M., PEG-grafted chitosan as an injectable thermoreversible hydrogel. Macromol. Biosci., 5, 107, 2005. 50. Chenite, A., et al ., Rheological characterisation of thermogelling c hitosan/glycerol-phosphate s olutions. Carbohydr. Polym., 46, 39, 2001. 51. Chenite, A., et al., Novel injectable neutral solutions of chitosan form biodegradable gels in situ. Biomaterials, 21, 2155, 2000. 52. Kofuji, K., et al., The controlled release of a drug from biodegradable chitosan gel beads. Chem. Pharm. Bull. (Tokyo), 48, 579, 2000. 53. Shu, X.Z. and Zhu, K.J., Controlled drug release properties of io nically cr oss-linked c hitosan b eads: The inuence of anion structure. Int. J. Pharm., 233, 217, 2002. 54. Wang, L.Y., et al ., Preparation and characterization of uniform-sized ch itosan m icrospheres c ontaining i nsulin b y membrane emulsication and a two-step solidication process. Colloids Surf. B Biointerfaces, 50, 126, 2006. 55. Giunchedi, P., et al ., Formulation and in vi vo evaluation of chlorhexidine b uccal t ablets p repared usin g dr ug-loaded chitosan m icrospheres. Eur. J . Ph arm. B iopharm., 53, 233, 2002. 56. Martin, L., et al., The release of model macromolecules may be co ntrolled b y the h ydrophobicity o f palmi toyl g lycol chitosan hydrogels. J. Control. Release, 80, 87, 2002. 57. Knapczyk, J., Chitosan hydrogel as a base for semisolid drug forms. Int. J. Pharm., 93, 233, 1993. 58. Lin, W.C., Yu, D.G., and Yang, M.C., pH-sensitive polyelectrolyte co mplex g el micr ospheres co mposed o f c hitosan/ sodium tripolyphosphate/dextran sulfate: Swelling kinetics and drug delivery properties. Colloids Surf. B Biointerfaces, 44, 143, 2005. 59. Ruel-Gariepy, E., et al ., A thermos ensitive c hitosan-based hydrogel for the lo cal delivery of paclitaxel. Eur. J. P harm. Biopharm., 57, 53, 2004.

43722_C010.indd 19

10-19

60. Muzzarelli, R.A., et al., Stimulatory effect on bone formation exerted b y a mo died chitosan. Biomaterials, 15, 1075, 1994. 61. Muzzarelli, R.A., et al., Osteoconductive properties of methylpyrrolidinone chitosan in an animal model. Biomaterials, 14, 925, 1993. 62. Muzzarelli, R.A., et al ., Osteoconduction exerted by methylpyrrolidinone chitosan used in dental surgery. Biomaterials, 14, 39, 1993. 63. Klokkevold, P.R., et al., Osteogenesis enhanced by chitosan (poly-N-acetyl glucosaminoglycan) in vitro. J. Periodontol., 67, 1170, 1996. 64. Legeros, R.Z., et al., Biphasic calcium phosphate bioceramics: p reparation, p roperties a nd a pplications. J. M ater. S ci. Mater. Med., 14, 201, 2003. 65. Daculsi, G., Biphasic calcium phosphate concept applied to articial bone, implant coating and injectable bone substitute. Biomaterials, 19, 1473, 1998. 66. Seol, Y.J., et al., C hitosan sp onges as t issue en gineering scaffolds for b one for mation. Biotechnol. Le tt., 26, 1037, 2004. 67. Carey, L.E., et al., Premixed rapid-setting calcium phosphate composites for bone repair. Biomaterials, 26, 5002, 2005. 68. Liu, H., et al., Novel injectable calcium phosphate/chitosan composites for bone substitute materials. Acta Biomater., 2, 557, 2006. 69. Mukherjee, D.P., et al ., An animal e valuation of a past e of chitosan glutamate and hydroxyapatite as a syn thetic bone graft material. J. Biomed. Mater. Res. B Appl. Biomater., 67, 603, 2003. 70. Leroux, L., et al ., Eff ects o f va rious ad juvants (lac tic acid , glycerol, a nd c hitosan) o n th e i njectability o f a cal cium phosphate cement. Bone, 25, 31S, 1999. 71. Jarry, C., Shive, M., and Chenite, A., An injectable biomaterial fo r bo ne r epair, material s cience for um, 539-543, 535, 2007. 72. Nery, E.B., et al., Tissue response to biphasic calcium phosphate ceramic with different ratios of HA/beta TCP in periodontal osseous defects. J. Periodontol., 63, 729, 1992. 73. Suh, J.K. and Matthew, H.W., Application of chitosan-based polysaccharide biomaterials in cartilage tissue engineering: A review. Biomaterials, 21, 2589, 2000. 74. Hoemann, C.D., et al., Tissue engineering of cartilage using an i njectable a nd a dhesive c hitosan-based ce ll-delivery vehicle. Osteoarthritis Cartilage, 13, 318, 2005. 75. Hoemann, C., et al ., Chi tosan-glycerol p hosphate/blood implants elici t h yaline ca rtilage r epair in tegrated wi th porous sub chondral b one in micr odrilled ra bbit def ects. Osteoarthritis Cartilage, 15, 78, 2007. 76. Chevrier, A., et al ., Chi tosan-glycerol p hosphate/blood implants increase cell recruitment, transient vascularization and sub chondral b one r emodeling in dri lled ca rtilage defects. Osteoarthritis Cartilage, 15, 316, 2007. 77. Hoemann, C.D., et al ., Chi tosan-glycerol p hosphate/blood implants improve hyaline cartilage repair in o vine microfracture defects. J. Bone Joint Surg. Am., 87, 2671, 2005.

10/3/2008 9:31:34 PM

10-20

78. www.clinicaltrials.gov 79. Dang, J .M., et al ., T emperature-responsive h ydroxybutyl chitosan f or the c ulture o f mes enchymal st em cells a nd intervertebral disk cells. Biomaterials, 27, 406, 2006. 80. Mwale, F., et al., Biological evaluation of chitosan salts crosslinked to genipin as a cell scaffold for disk tissue engineering. Tissue Eng., 11, 130, 2005. 81. Roughley, P., et al., The potential of chitosan-based gels containing intervertebral disc cells for nucleus pulposus supplementation. Biomaterials, 27, 388, 2006. 82. Muzzarelli, R .A.A., Human enzyma tic ac tivities r elated t o the thera peutic administra tion o f c hitin deri vatives. Cell. Mol. Life Sci., 53, 131, 1997. 83. Rabea, E.I., et al., Chitosan as antimicrobial agent: Applications and mode of action. Biomacromolecules, 4, 1457, 2003. 84. Ishihara, M., et al., Photocrosslinkable chitosan as a dressing for w ound o cclusion a nd accelera tor in healin g p rocess. Biomaterials, 23, 833, 2002. 85. Degim, Z., et al., An investigation on skin wound healing in mice with a taurine–chitosan gel formulation. Amino Acids, 22, 187, 2002. 86. Thanou, M., Verhoef, J.C., and Junginger, H.E., Chitosan and its derivatives as intestinal absorption enhancers. Adv. Drug Deliv. Rev., 50 Suppl 1, S91, 2001. 87. Thanou, M., Verhoef, J .C., a nd J unginger, H.E., Oral dr ug absorption en hancement b y chi tosan a nd i ts deri vatives. Adv. Drug Deliv. Rev., 52, 117, 2001. 88. Illum, L., et al., Chitosan as a novel nasal delivery system for vaccines. Adv. Drug Deliv. Rev., 51, 81, 2001. 89. Hirano, S., Applications of chitin and chitosan in the e cological a nd environmental  elds, in Applications o f Ch itin and Chitosan. Goosen, M., Eds., Lancaster, PA, 1997, p. 31. 90. Struszczyk, H. a nd P ospieszny, H., N ew a pplications o f chitin and its derivatives in plant protection, in Applications of Ch itin a nd Ch itosan. G oosen, M., E ds., L ancaster, P A, 1997, p. 171. 91. Freepons, D ., E nhancing f ood p roduction w ith c hitosan seed-coating t echnology, i n Applications o f Ch itin a nd Chitosan. Goosen, M., Eds., Lancaster, PA, 1997, p. 129. 92. Lavertu, M., et al., High efficiency gene transfer using chitosan/DNA na noparticles wi th sp ecic co mbinations o f molecular weight and degree of deacetylation. Biomaterials, 27, 4815, 2006.

43722_C010.indd 20

Smart Materials

93. Sinha, V.R., et al., Chitosan microspheres as a potential carrier for drugs. Int. J. Pharm., 274, 1, 2004. 94. Lahiji, A., et al ., Chi tosan su pports the exp ression o f extracellular ma trix p roteins in h uman ost eoblasts a nd chondrocytes. J. Biomed. Mater. Res., 51, 586, 2000. 95. Bokura, H. a nd K obayashi, S., Chi tosan de creases t otal cholesterol in w omen: A randomized, double-blind, placebo-controlled trial. Eur. J. Clin. Nutr., 57, 721, 2003. 96. Muzzarelli, R., Clinical and biochemical evaluation of chitosan for hypercholesterolemia and overweight control. EXS, 87, 293, 1999. 97. Takayanagi, T . a nd M otomizu, S., Chi tosan as ca tionic polyelectrolyte for the modication of electroosmotic ow and its utilization for the separation of inorganic anions by capillary zone electrophoresis. Anal. Sci., 22, 1241, 2006. 98. Jocic, D., et al., Dye sorption by wool treated with low temperature plasma and chitosan, in Chitosan in Pharmacy and Chemistry. Muzzarelli, R.A.A. and Muzzarelli, C., Eds., Atec, Italy, 2002, p. 317. 99. Bumgardner, J.D., et al ., Chitosan: potential use as a b ioactive coa ting f or o rthopaedic a nd cra niofacial/dental implants. J. Biomater. Sci. Polym. Ed., 14, 423, 2003. 100. Lang, G., Wendel, H., and Konrad, E., Quaternary hydroxyethyl-substituted chi tosan deri vatives, cosmet ic co mpositions bas ed ther eon a nd p rocesses f or the p roduction thereof, Patent, US4772690, 1988. 101. Wedmore, I., et al ., A special report on the c hitosan-based hemostatic dressing: Experience in c urrent combat operations. J. Trauma., 60, 655, 2006. 102. Lee, K.Y., Ha, W.S., and Park, W.H., Blood compatibility and biodegradability o f pa rtially N-ac ylated c hitosan deri vatives. Biomaterials, 16, 1211, 1995. 103. Rao, S.B. and Sharma, C.P., Use of chitosan as a biomaterial: studies o n i ts s afety a nd hemost atic p otential. J. B iomed. Mater. Res., 34, 21, 1997. 104. Kawakami, T., et al., Experimental study on osteoconductive properties of a c hitosan-bonded hydroxyapatite s elf-hardening paste. Biomaterials, 13, 759, 1992. 105. Ueno, H., M ori, T., a nd F ujinaga, T., Topical f ormulations and wou nd he aling a pplications of ch itosan. Adv. D rug Deliv. Rev., 52, 105, 2001. 106. Muzzarelli, R., et al., Biochemistry, histology and clinical uses of chitins and chitosans in wound healing. EXS, 87, 251, 1999.

10/3/2008 9:31:34 PM

11 Films, Coatings, Adhesives, Polymers, and Thermoelectric Materials 11.1 Sm art Adhesives .................................................................................................................... 11-1 References ...........................................................................................................................................11-4 11.2 Oxides as Potential Ther moelectric Materials ..................................................................11-4 Introduction • P erovskite LaCoO3: A n- or p-Type Oxide • Orthochromites Pr1−x Ca x CrO3: Role of the Spin and Orbital Degeneracies • Large Ther mopower in Metallic Oxides: The Mist Layer Oxides • SrRuO3: A Metallic Perovskite with a Thermoelectric Power Driven by the Spin Degeneracy Term • Conclusion

Acknowledgments ........................................................................................................................... 11-11 References ......................................................................................................................................... 11-11 11.3 Electrically Conductive Adhesives ................................................................................... 11-12 Introduction • I sotropically Conductive Adhesives • A nisotropic Conductive Adhesives and Films • Nonconductive Adhesives and Films • Summary

References ......................................................................................................................................... 11-23 11.4 Electrochromic Sol–Gel Coatings .................................................................................... 11-25 De nition of Electrochromism • Principle of the Galvanic Cell • Sol–Gel Processing • C onclusions

Acknowledgments ........................................................................................................................... 11-29 References ......................................................................................................................................... 11-29

11.1

Smart Adhesives

James A. Harvey and Susan Williams Smart m aterials a re m aterials t hat re spond to t heir environments in a timely manner. Smart materials can receive, transmit, or process a stimulus and respond by producing a useful effect t hat m ay i nclude a sig nal t hat t he m aterials a re ac ting upon it. Some of the stimuli that may act upon these materials are strain, stress, temperature, chemicals (including pH stimuli), electric eld, magnetic eld, hydrostatic pressure, different t ypes o f r adiation, a nd ot her f orms o f s timuli. A nother important criterion for a smart material is its ability to receive stimuli a nd re sponding to t he s timuli to p roduce a u seful effect that is and it must be reversible. Another feature that is an important factor in determining if a material is smart pertains to its asymmetrical nature. From the purist point of view, materials are smart if at some point within their performance history, they act reversible to a stimulus. Materials that formally had t he l abel o f b eing sm art i nclude p iezoelectric m aterials,

electrostrictive m aterials, e lectrorheological m aterials, m agnetorheological ma terials, t hermoresponsive ma terials, p Hsensitive ma terials, U V-sensitive ma terials, sma rt p olymers, smart gel s ( hydrogels), sm art c atalysts, a nd s hape memo ry alloys [1]. Wi th t he p rogress e xperienced b y sm art m aterials and t he suc cessful i ntroduction o f na notechnology, a ne w class of smart materials has been developed. Th is class consists of the smart adhesives. Smart adhesives do not h ave the full range of capabilities or actions a s t he t ypical sm art m aterials. H owever, sm art ad hesives do have the abilities to do many smart-like actions. These include, but not limited to the following: mixing, curing monitoring the presence of hazardous gases and changes in temperatures v ia c olor c hanges, a nd t he rele ase o f ad hesion up on demand. The development of a truly smart adhesive hinges around the ability of the adhesive to behave as if it could have a brain, nerves, and muscles. This type of behavior has not yet been developed. This indeed will be accomplished through the efforts of adhesive, smart adhesives, and smart materials practitioners. 11-1

43722_C011.indd 1

10/3/2008 9:34:57 PM

11-2

Smart Materials TABLE 11.1

Examples of Smart Adhesive Properties and Typical Applications

Smart Adhesive Properties Aesthetics—color surface nish

Disassembly Materials mimicry Thermal conductivity Electrical conductivity Electrical resistance/insulation Energy and vibration adsorption

Heat resistance, re resistance, re extinguishing Auto-surface coupling Self-indicating—presence of cured adhesive Security Ultralow density Vibration adsorption Regulatory compliance

Applications Color matching High contrast for visible security Cross-pigmentation for two-part security Opaque or translucent Disbond on demand for materials recycling Ability to be machined, drilled, or cut Magnetic Thermal management in electronics and electric motors Heat sink attachment Organic solder replacement Electrical terminal potting Thr ead locking Gas seal in threading ttings Mirror bonding in transportation Joint disassembly Mass transport and aerospace interior trim Hot joints in electrical devices or high-temperature environments Minimizing surface treatment on aluminum and composites UV uorescence for automated quality Color change on cure for hazardous gas ttings Leak sealing and repair Weight saving in transportation—particularly for aerospace duties Noise abatement, vibration damping Intrinsically safe in use: medical devices, drug-delivery patches and food contact (direct or indirect).

So urce: From Fakley, M., Chem. Ind., 21, 691, 2001. With permission.

One of the initial mentions of the term smart adhesives in the open literature was a review article with that title. The foci of the review were how to work smartly and the factors to consider when choosing an adhesive. The article did contain a table (Table 11.1) that is co nsidered to b e t he f orerunner o f ma ny co ncepts a nd applications of smart adhesives [2]. Another re view d iscussed t he u ses of smart ad hesives i n t he information technological  elds. Such u sages described t he u se of su rface mount ad hesives to  x a nd hold soldering a nd ot her operations. Si ngle-component he at-cured e poxies we re u sed i n this app lication. O ther te chnology-related sm art ad hesives l ike conformal coatings provide t hin d ielectric coatings t hat ex tend circuit board lifetimes by protecting components and traces from corrosion, shorts, and mechanical damage. Solid systems such as UV curing acrylics and silicones were used in this function [3]. An interesting and comprehensive review of medical adhesives for use in products that are applied directly to the body or inside the body by surgical procedures has been reported. The article focused on pre ssure-sensitive a dhesives a nd t issue a dhesives. These two classes of adhesives were further divided into synthetic and biological adhesives. The report cited the progress gained in the understanding of wound management and the development and for mulation of p olymers for pre ssure-sensitive a dhesives that facilitated the wound healing process. Adhesives with highmoisture v apor p ermeation a llow wound d ressings to c ontrol exudates m ore eff ectively. T rauma ma nagement ha s se en

43722_C011.indd 2

advances m ade i n s mart pre ssure-sensitive a dhesives t hat changed adhesive properties with external stimuli [4]. Researchers reported an example of smart adhesion at a polymer/metal (oxide) interface that responded reversibly to changes in tem perature. The tem perature de pendence i n t his s ystem arose from the rubber elasticity of the polymer, 1,4-polybutadiene, and m irrored t he i nterfacial b ehavior o f t he s ame p olymer against w ater. These s ystems off er un ique o pportunities f or designing responsive materials whose properties can be actively controlled [5]. Interesting uses of the waste from the chromium compound treatment of the collagen polymers resulting from the treatment of leather from the tanning process were reported in the literature. Covalent i ntramolecular a nd i ntermolecular c ross-links w ere formed between modied lysine side chains within the collagen brils. High molecular weight versions expanded the potential commercial applications to the veterinary, medical, and cosmetic elds. I t s hould b e p ointed o ut t hat t he b iopolymers i solated from chrome-tanned collagenic wastes were applied directly to the following industries: paper, wood, ink, textile and leather, as binders, biodegradable biopolymers, and protein-based self-feeding biodegradable  ower p ots,  llers, re tanning a gents, a nd c asein substitute in photographic adhesives. A second possibility would be to p roduce low formaldehyde content ecoadhesives a nd collagen-based smart fabric formed by chemical reaction with these biopolymers [6].

10/3/2008 9:34:58 PM

Films, Coatings, Adhesives, Polymers, and Thermoelectric Materials

In t he re alm of sm art m aterials, sm art ad hesives, a nd ot her smart types, there exists a unique class of materials, which utilize a smart material in conjunction with a nonsmart material. These loosely belong to the classication of smart structures. Ther e are smart s tructures t hat mo nitor t he b ehavior o f ad hesives or enhance the manufacture of adhesives. The rst e xample o f t hese t ypes o f sm artness i s t he u se o f piezoelectrics to monitor the behavior of adhesives. The action of piezoelectrics is discussed in Chapter 9. Metalized poly(vinylidene  uoride)  lms w ere e tched to p rovide m ultipoint s ensors, which were imbedded within adhesive joints to measure the peel stresses. Peel stresses were thought to be the most critical stresses responsible for failure of a v ariety of ad hesive joints. The technique successfully demonstrated the peel stress trends expected in si ngle a nd do uble l ap jo ints a nd b utt jo ints. A dhesive c ure monitoring a nd vo id/porosity de tection w ere a lso p erformed with ultrasound using the lm [7–9]. Other l iterature c itations de scribing mo dications or improvements to t he u se of piezoelectrics to mo nitor ad hesive joints have been recently published. These along with the previously men tioned e fforts i ndicate t he p otential o f h aving w ires embedded w ith a b ond to mo nitor t he behavior, which in turn can lead to the prediction of the lifetime (or out of service time) of a s tructure [ 10–13]. This c ited l iterature p ertains to t he through-life, n ondestructive m onitoring o f a dhesively b onded structures. It presented a discussion of the concept of microelectromechanical s ystems ( MEMS) sm art s ensors. The MEMS smart s ensors w ere p ermanently i nstalled i n suc h l arge s tructures as ships, aircrafts, or land vehicles [14]. Another app roach o f i ncorporating sm art m aterials w ith adhesives i nvolves t he add ition of smart pa rticles i n t he ad hesive f ormulation to i ncrease t he re activity o f t he ad hesive. Nanoparticles with crystal structure having properties of ferromagnetic, fe rrimagnetic, s uperparamagnetic, or pie zoelectric materials were heated by electric, magnetic, or electromagnetic alternating elds aided in the hardening of adhesives [15,16]. As e xpected, t he re verse has b een ac complished. By h eating nanoparticles with crystal structure having properties of fe rromagnetic, fe rrimagnetic, s uperparamagnetic, or pie zoelectric materials b y ele ctric, m agnetic, o r ele ctromagnetic a lternating elds, thermoplastic adhesives can be loosened [17]. Another concept that allows one to work smarter with adhesives consists of depositing adhesives or adhesion control agents in a m anufacturing p rocess u sing i nk je t p rinting h eads. The authors a re advo cates o f t his te chnique o f app lying m inute amounts to conned or limited spaces. We would like to caution users of this application of chemistries that the consistence of the material is the same after deposition and is the same as formulated or initially manufactured [18]. Packaging of m icroelectronic de vices h as b ecome more a nd more important. Applications for i nterconnection technologies range from highly specialized processors to be assembled on ex for mobile phones, highly dened packages as for hearing aids to low-cost transponders as for smart cards and smart labels. Each of t hese i nterconnection t ypes h as d ifferent r equirements i n

43722_C011.indd 3

11-3

technologies and costs. The use of isotropic conductive adhesive (ICA) f or b umping a nd a ssembly f or sm art c ards re sults i n a number of advantages such as a si mple versatile process, lower temperatures, and environment friendliness. Using an anisotropic c onductive ad hesive ( ACA) o r a no nconductive ad hesive (NCA) for smart label production allows a high throughput from reel to reel [19]. One of t he applications of s pecial ad hesives h as b een i n t he fabrication of smart cards. UV-curable ACAs were used for these mobile electronic products. The adhesives were subjected to different mechanical or environmental attacks like warps, stretching, changing humidity and temperature, which could affect the performance and reliability of the electronic products [20]. A si mple-design Cu c oil pl ated on p olyethylene t erephthalate was used as the microstrip antenna of the smart card. The entire fabrication process of the UV-curable ACA was divided into three parts: high-power UV curing, chip-on-ex (COF) bonding, a nd p ostcuring. By v arying t he U V c uring a nd postcuring pa rameters, a n umber o f c ontactless sm art c ards were made [21]. A c lass of sm art m aterials t hat h as pa ralled t he g rowth of and is closely related to t he smart adhesives are the smart coatings. The advances of the smart coatings can be illustrated in the top ics c overed at a sm art c oating c onference h eld i n February 2005. Examples o f sm art c oatings i ncluded t he f ormation o f a n enigmatic p olymer w ith t he u nique p roperty o f b eing h ydrophilic (water loving) when dry and hydrophobic (water hating) when wet. The researchers produced an antimicrobial coating by adding hydantoin into uorine-containing polymer chains. The researchers c laimed t hat w ater-induced h ydrophobic su rfaces would re sult i n ne w me dicine de vices, s witching de vices, a nd drag-reducing coatings [22]. A new uoropolymer additive for coating has been developed by Ausimont. This new coating additive assists in the removal of graffiti from outside surfaces without aff ecting the coating and weathering properties of the surface [23]. Millenium Chemicals commercialized a polysiloxane silicone coating, Ecopaint. This unique coating in conjunction with either titanium dioxide or calcium carbonate or a c ombination of the two i norganics ren ders re spirator-causing p roblems, n itrogen oxides (NOx gases), harmless. Buildings with this type of system will be more environmentally friendly [24]. In conclusion, one of the authors remembers an article by a group of French adhesive scientists in which they commented on t he s tatus of a dhesion t heories. They i ndicated that there was s till a g reat de al o f i nformation ne eded to e xplain w hy things s tick t ogether. The a uthors m ust t ake t hese rem arks into account and apply similar methodologies to the development of smart adhesives. The growth of smart adhesives is in its early development s tages. M ost e xamples o f t he sm art ad hesives found in the open literature do not truly  ll all the criteria o f b eing sm art t hat i s h aving a “ brain,” “ nerves,” a nd “muscles.” With time and research and development efforts, this will change.

10/3/2008 9:34:58 PM

11-4

References 1. Harvey, J .A. S mart ma terials, in Mechanical Eng ineers’ Handbook, 3r d e d., K utz, M. (e d.), J ohn Wiley & S ons, New York, 2005. 2. Fakley, M. Chemistry & Industry, 21, 691–695, 2001. 3. Birkett, D. Chemistry in Britain, 38(11), 29–31, 2002. 4. Webster, I. and West, P.J. in Polymeric Biomaterials, 2nd ed., Dumitriu, S. (ed.), Marcel Dekker, New York, pp. 703–737, 2002. 5. Khongtongad, S. and Ferguson, G.S. Journal of the American Chemical Society, 124(25), 7254–7255, 2002. 6. Cot, J. Journal of the American Leather Chemists Association, 99(8), 322–350, 2004. 7. Dillard, D.A., Anderson, G.L., and Davis, D.D. Jr., Journal of Adhesion, 29(1–4), 245–255, 1989. 8. Anderson, G.L. a nd Dilla rd, D .A. NTIS. Rep ort (AR O23759.3-EG-F; Order No. AD-A233 880 1991). 9. Anderson, G.L., Mommaerts, J., Tang, S.L., Duke, J.C., and Dillard, D .A. P roceedings o f a C onference o n Re cent Advances in Adaptive S ensing M aterials a nd Their Applications, Rogers, C.A and Rogers, R.C. (eds.), pp. 266– 273, 1992. 10. Shankar, K., Tahtali, M., Chester, R., and Torr, G., Proceedings of S PIE-Ā e I nternational S ociety f or O ptical E ngineering, 4317 (Experimental Mechanics), 363–368, 2001. 11. Kwon, J.W., Chin, W.S., a nd L ee, D.G. Journal of Adhesion Science and Technology, 17(6), 777–796, 2003. 12. Hodges, C.A., M ossi, K.M., a nd S cott, L.A. Proceedings of S PIE-Ā e I nternational S ociety f or O ptical E ngineering, 5053 (Active Materials: Behavior and Mechanics), 467–474, 2003. 13. Hwang, H.Y. and Lee, D.G. Journal of Adhesion Science and Technology, 19(12), 1053–1080, 2005. 14. Wilson, A., Christina, O .-J., a nd M uscat, R . Materials Australia, 34(3), 15–17, 2002. 15. Kirsten, C.N., H enke, G., Cla udia, M.-J ., U nger, L., a nd Meier, F. PCT I nt. Appl. WO 2002018499 A1 20020307, 2002. 16. Kirsten, C.N. PCT Int. Appl. WO 2002012409 A1 20020214, 2002. 17. Kirsten, C.N., Chris ophliemk, P ., N onninger, R ., S chirra, H., and Schmidt, H. Ger. Offen., DE 19924138 A1 20001130, 2000. 18. Saksa, T.A. and Lee, S. U.S. Pat. Appl. Publ., US 2003070740 A1 20030417, 2003. 19. Seidowski, T ., Kr iebel, F ., a nd Ga lties, J . Pr oceedings— International Conference on Adhesive Joining and Coating Technology in E lectronics M anufacturing, p resented a t Adhesive in Electronics 2000, 4th, Espoo, Finland, pp. 52–54, June 18–21, 2000. 20. Lee, K., N g, K.T., Tan, C.W., Cha n, Y.C., a nd Chen g, L.M. Proceedings o f I nternational C onference o n t he Business of El ectronic P roduct Re liability a nd Lia bility, S hanghai, China, pp. 134–139 Apr. 27–30, 2004.

43722_C011.indd 4

Smart Materials

21. Tan, C.W., Siu, Y.M., Lee, K.K., Chan, Y.C., and Cheng, L.M. Proceedings o f I nternational C onference o n t he Business of El ectronic P roduct Re liability a nd Lia bility, S hanghai, China, pp. 140–144, Apr. 27–30, 2004. 22. NASA Tech Brief Briefs, Insider, May 5, 2005. 23. Locaspi, A. a nd M archetti, R . P aint & C oatings I ndustry, March, 2000. 24. www.SpecialChem.com, I ntelligent Adhesives, S ept. 14, 2005.

11.2

Oxides as Potential Thermoelectric Materials

S. Hébert and A. Maignan 11.2.1 Introduction The search for new sources of energy is an active topic. Among them, t he t hermoelectric gener ation i s a p owerful te chnique, which i mplies no me chanical mo vements a nd no  uid, a nd i s thus a reliable technique to recover waste heat and generate electricity. Due to t he small conversion efficiencies, thermoelectric generation is for the moment restricted to specic applications such a s s patial app lications (with r adioisotopic t hermoelectric generator). However, recovering i ndustrial waste heat could be an efficient way to generate power [1], and thermoelectric materials op erating at h igh tem peratures (T up to 1400°C for steel industry [ 1]) a re ne eded. A t suc h tem peratures, t he c lassical thermoelectric materials a s Bi2Te3 su ffer f rom t hermal decomposition or detrimental oxidations. Recently, t his has created a rush f or ne w t hermoelectric m aterials suc h a s s kutterudites, clathrates, a mong ot hers. The p rinciple o f a t hermoelectric device i s s hown i n F igure 11.1. I t i s ba sed o n pa irs o f n - a nd p-type le gs ele ctrically a nd t hermally a ssembled i n s erial a nd parallel mo des, re spectively. The i ntrinsic p erformances o f a thermoelectric material, n- or p-type, are dened by its gure of merit, Z = S2/rk, or its p ower f actor PF = S2/r, w here S is t he Seebeck coefficient, r the electrical resistivity, and k the thermal conductivity. The dimensionless ratio ZT should be at least equal to 1 for applications. The discovery of a large thermoelectric power in the metallic oxide Na xCoO2 has shown that oxides are promising materials for thermoelectric applications [2]. Due to t heir good thermal stability in high temperatures and in air, they can be used for recovery of waste heat and thermoelectric generation at high temperatures. Following this report, several studies have been dedicated to the search for new good thermoelectric oxides. We will present here the results obtained in our group. Among them, the mist oxides look promising. This family of oxides was rst discovered in our laboratory [3], and their layered structure consists i n t he st acking o f C oO2 layers of the CdI2 type as in Na x CoO2 , s eparated b y N aCl-like l ayers. T heir t ransport properties a re very c lose to t he ones of Na xCoO2, w ith a l arge thermopower associated to a me tallic behavior above 100 K [4]. A ZT value close to 1.2 at 973 K has been estimated by Funahashi

10/3/2008 9:34:58 PM

11-5

Films, Coatings, Adhesives, Polymers, and Thermoelectric Materials

p

n

Th : Heater

the spin and orbital degeneracy terms, as proposed by Doumerc [7] and Koshibae et al. [8], can enhance the Seebeck coefficient in these oxides.

∆T = T h −T c

11.2.2

Perovskite LaCoO3: A n- or p-Type Oxide

The Heikes formula has been proposed to calculate the Seebeck coefficient of localized carriers [6]. Th is formula is valid i n t he limit of innite temperature and can be written as:

T c: Heat sink

S=

Thermogeneration

FIGURE 1 1.1 Schematic d escription of a t hermogeneration d evice showing one pair made of n and p legs.

et a l. [5]. The o ptimization o f t he t hermoelectric p roperties o f this family of oxides has been investigated and this work has then been extended to the search for new thermoelectric oxides. In the following, the chapter will be divided in two parts. First, the results concerning semiconducting oxides and the possibility to create n- and p-type thermoelements will be presented. Then, the results obtained in the case of metallic oxides will be described. Most of the results will be analyzed using the Heikes formula [6], which states that, at high temperature, the Seebeck coefficient is a direct measure of the entropy per carrier, and thus mainly depends on the carrier concentration. We will show how

−kB ⎛ 1 − x ⎞ (1 ln ⎜ ⎝ x ⎟⎠ e

1.1)

where x is the carrier density. For small doping, it is thus theoretically possible to create n- or p-type thermoelements, with large Seebeck coefficient. Following this formula, small doping has been used in LaCoO3 to introduce mixed cation valency of Co2+/Co3+ (n-type) or Co3+/ Co4+ ( p-type). F igure 1 1.2 p resents t he S eebeck c oefficient of LaCoO3, L a0.98Sr0.022+Co3+/4+ O3, a nd L a0.99Ce0.014+Co2+/3+O3. As e xpected, i n t he s toichiometric c ompound, t he S eebeck coefficient i s v ery l arge, a nd p ositive (+600 µV/K at 3 00 K), suggesting a sm all c oncentration o f h oles i n t he pa rent c ompound (x ∼ 9 × 10−4 Co4+/Co following the Heikes formula). S is reduced b y t he i ntroduction o f Sr 2+, a nd t he sig n o f S cha nges when C e4+ i s i ntroduced. U sing t he H eikes f ormula, a go od agreement is obtained between t he nominal and actual doping concentration in the case of p-type materials, with x ∼ 2.22 × 10−2 Co4+/Co to be compared to the nominal 2 × 10−2 Co4+/Co calculated from the chemical formula. The agreement is not so good in the n-type case w ith x ∼ 3.6 × 10−2 Co 2+/Co, to be compared to 10−2,

1100 1000 900

LaCoO3

800 700

S (µV/K)

600 La0.98Sr0.02CoO3

500 400 300 200 100 0

La0.99Ce0.01CoO3

−100 −200 −300 −400 50

FIGURE 11.2

43722_C011.indd 5

100

150

200 T (K)

250

300

S(T) of LaCoO3 and p (La0.98Sr0.022+Co3+/4+O3) and n (La0.99Ce 0.014+Co2+/3+O3) type doped LaCoO3.

10/3/2008 9:34:58 PM

11-6

Smart Materials

107 107

106 105

x = 0.05

105

La0.99Ce0.01CoO3

103 r (Ω cm)

r (Ω cm)

106

LaCoO3

104

102 101 100

10−2 0

103

La0.98Sr0.02CoO3 101 50

100

150

200 250 T (K)

300

350

400

FIGURE 1 1.3 r(T) of L aCoO3, L a0.98Sr0.022+Co3+/4+O3, a nd L a0.99 Ce 0.014+Co2+/3+O3.

but this difference could be linked to a problem of homogeneity due to t he very sm all dopa nt c oncentration. A lso, ot her ter ms linked to t he spin a nd orbital degeneracies should be added to the Heikes formula, as shown by Taskin et al. [9]. The most important result is that n- or p-type materials can be created by doping this oxide with holes or electrons [10]. Concerning t he t hermoelectric per formances, a st rong a symmetry i s ob served f or t he re sistivities, w ith l arge r in La0.99Ce0.014+Co2+/3+ O3, which will lead systematically to a smaller power factor for n-type LaCoO3 (Figure 11.3). As shown in Ref. [11], c arrier t ransport i n doped L aCoO3 is easier i n t he c ase of holes than in the case of electrons. This phenomenon, called the “spin blockade,” is due to the spin states of cobalt and to the different  lling o f t he o rbitals: a h ole c an h op i n t he t 2g or bitals, while hopping is more difficult in the case of electron doping, due to the formation of a (Co3++ electron), which is not energetically favorable. The p erformances o f t he n -type L aCoO3 a re t hus a lways smaller. N evertheless, t he p ossibility to c reate n - a nd p -type thermoelements st arting f rom a st oichiometric co mpound is very appealing. This result can be extended to other compounds, as shown in the 1D Ca3Co2O6 [12].

11.2.3

104

102

10−1

Orthochromites Pr1−xCa xCrO3: Role of the Spin and Orbital Degeneracies

Following t he  rst pap ers re porting o n t he H eikes f ormula, many calculations have been performed to gener alize it, taking into ac count t he p ossible s pin [7] a nd o rbital de generacies [8]. Among them, a detailed analysis has been performed by Marsh and Pa rris to c alculate t he t hermopower o f s ystems w ith t 2g orbitals (L aCrO3 [13]), or with eg orbitals (i.e., LaMnO3 [1 4]). Taking into account t he polaronic nature of t ransport in t hese

43722_C011.indd 6

Pr1−x Cax CrO3 x = 0.0

x = 0.3

100 10−1

0

FIGURE 11.4

100

200 T (K)

300

400

r(T) of the Pr1−x Ca x CrO3 series.

materials, and the spatial and spin degeneracies associated to the orbitals, they calculate the Seebeck coefficients of these materials. Following these results, we have measured the thermopower of the Pr1−xCa xCrO3 system for x ≤ 0.3 [15]. Figure 11.4 presents the resistivity curves. These materials are semiconductors, a nd f or x = 0, r ∼ 10 0 Ω c m a t 40 0 K. A s x increases, the carrier concentration increases and the semiconducting-like resistivity decreases. The t hermopower d ata a re p resented i n F igure 1 1.5. The Seebeck c oefficient of x = 0 i s p ositive, a nd v ery l arge, w ith S∼900 µV/K at 3 00 K. The introduction o f carri ers in duces a strong decrease of S as x increases, but S re mains p ositive a s expected f or h ole d oping ( introduction o f C r4+ in t he C r3+ matrix through the substitution of Ca 2+). S is almost constant from 2 00 to 3 00 K, a nd t he h igh-temperature me asurements show that S is independent of temperature at least up to 650 K (Figure 11.6). The constant value of S is consistent with polaronic transport. The evolution of S measured at 300 K, as a function of x, is presented in Figure 11.7. The s olid l ine re presents t he t heoretical evolution o f S f rom t he H eikes f ormula ( Equation 1 1.1). The dotted l ine i s t he e volution o f S ve rsus x calculated from the Marsh and Parris formula [13], which takes into account the spin and orbital degeneracies: S=

−kB ⎛ 1 − x ⎞ k ln ⎜ + ∆S d , with ∆S d = B ln(Γ orbΓ spin ) (1 ⎝ x ⎟⎠ e e

1.2)

where Γorb and Γspin are the orbital and spin degeneracies. For Cr3+/ Cr4+, in the case of weak magnetic coupling, ∆Sd = +69.9 µV/K [13]. Figure 11.7 shows that the experimental results  t perfectly well

10/3/2008 9:35:02 PM

11-7

Films, Coatings, Adhesives, Polymers, and Thermoelectric Materials Pr1−x Cax CrO3

1200

1000 900

x = 0.0 x = 0.05 x = 0.3

1000

700 600 S (µV/K)

800 S (µV/K)

Pr1−xCaxCrO3 Calculated value taking orbital degeneracy Calculated value from Heikes formula

800

600

500 400 300 200

400

100 0

200

−100

0 100

0.0 150

250

200

300

0.1

350

T (K)

0.2 0.3 0.4 x (Ca concentration)

0.5

FIGURE 11.7 S as a function of the Ca 2+ concentration. The solid line corresponds to the Heikes formula (Equation 11.1), and the dotted line to the Marsh and Parris model, which takes into account the spin and orbital degeneracies (Equation 11.2).

FIGURE 11.5 S(T) of the Pr1−x Ca x CrO3 series.

with the theory. The spin a nd orbital degeneracies can strongly enhance the thermopower, and maximizing this term is crucial to optimize the Seebeck coefficient.

11.2.4 Large Thermopower in Metallic Oxides: The Misfit Layer Oxides Na xCoO2 i s at t he o rigin o f t he s earch f or ne w t hermoelectric oxides. Na xCoO2 is a layered material with CoO2 layers of tilted edge-shared octahedra s eparated by r andomly  lled Na layers. Two mo dels h ave b een p roposed to e xplain t he l arge S va lue. The rst one takes into account the spin and orbital degeneracies associated to a mixture of low spin Co3+ and low spin Co4+, which is responsible for a large entropy and therefore a large S [8]. This

model should be applied only for localized systems and cannot explain t he me tallicity o f N a xCoO2. The s econd o ne h as b een proposed by Singh who calculated the band structure of Na xCoO2 [16]. Due to the rhombohedral symmetry of the CoO2 layers, the t2g orbitals a re s plit i n t wo suble vels. There would t herefore b e two t ypes of c arriers, l ight ones i n t he e g′ ba nd re sponsible for metallicity, and heavy ones in the a1g band responsible for a large thermopower. The mist layered oxides have been discovered in our laboratory a decade ago [3]. Their structure has the same CoO2 layers of tilted edge-shared octahedra, separated in this case by NaCl-like layers (Figure 11.8). There can be three or four separating layers. Recently, ne w m ists with two separating layers have been reported [ 17,18]. The s tructure c an b e de scribed w ith t wo monoclinic sublattices, with common a, c, and b parameters but

400 350

b2

S (µV/K)

300

CoO2

Pr0.95Ca0.05CrO3 250

Pr0.7Ca0.3CrO3

CaO

200 CoO CaO

300

FIGURE 11.6 and 0.3).

43722_C011.indd 7

S1

b1

150 100

S2

350

400

450

500 T (K)

550

600

650

700

S(T) measured up to 7 00 K of P r1−x Ca x CrO3 (x = 0 .05

CoO2

FIGURE 1 1.8 Schematic d escription of t he m ist s tructure of [Ca 2CoO3][CoO2]1.62, with the CoO2 layers and the NaCl-like layers.

10/3/2008 9:35:04 PM

11-8

Smart Materials

incommensurate b parameters, with the ratio b1/b2 ∼ 1.6–2 (b1 for NaCl-like layer and b2 for CoO2 layer). The mist o xides po ssess co mmon f eatures w ith N a xCoO2: they a re me tallic, w ith a l arge t hermopower at ro om temperature [4]. The aim of our work was to dop e t he CoO2 layers and understand the evolution of S (and r) with doping. 11.2.4.1

Influence of Doping

Substitutions t ake p lace mo st o f t he t imes i n t he N aCl-like layers, except for t he case of R h, as detailed in t he following discussion. The doping of the CoO2 layers is therefore induced by the charge balance between the two sublattices: the NaCllike l ayers ar e p ositively c harged ( a , w hich i s u nknown), while t he C oO2 layers is negatively charged, the two sublattices b eing ele ctrostatically b ounded. T aking i nto ac count this c harge ba lance, t he f ollowing e quations c an b e w ritten for the Co valency: v Co = x + 3 = −

α +4 b1/b2

where x is the Co4+ concentration. x depends both on a and on the b1/b2 ratio, where b1 and b2 are the b parameters of the NaCl-like layers and CoO2 layers, respectively. Another important parameter is the oxygen stoichiometry, which directly affects a. The investigation of the different families of mists has shown a general trend. First, the thermopower does not c hange much, from +90 µV/K for [Tl0.81Co0.2Sr1.99O3]RS[CoO2]1.79 [19] to 170 µV/K for [Ca 2Co0.6Ti0.4O3][CoO2]1.62 [20]. S econd, t wo d ifferent families o f b ehaviors a re ob served. S ome m ists s how a m etallic behavior down to low T ( 2 K), a nd a sm all p ositive m agne-

toresistance (M R), a s f or [ Tl0.81Co0.2Sr1.99O3]RS[CoO2]1.79 an d [Bi2Ba1.8Co0.2O4][CoO2]2 [21]. On t he ot her hand, most of t hem show a strong increase of resistivity at low T, with a strong negative MR. The thermopower is always smaller in the rst family. Due to t he l arge n umber o f u nknown pa rameters ( oxygen stoichiometry in the two sublattices, cobalt valency in the NaCllike layer, among others), it is very difficult to precisely know the Co v alency i n t he C oO2 l ayer a nd c heck t he v alidity o f t he Koshibae’s formula [8]. We h ave ne vertheless found t hat, for a given family of mist, S a lways i ncreases w hen b1/b2 de creases (Figure 11.9). The s ame t rend i s ob served i n t he [ Pb0.7Sr2−xCa xCo0.3O3] [CoO2]b1/b2 s ystem ( Figure 11.10) w here S i ncreases f rom S = 120 µV/K for x = 0 (b1/b2 = 1.79) to S = 165 µV/K for x = 2 (b1/b2 = 1.62) [ 22]. A lso, t his e volution i s ob served f rom 1 10 µV/K f or [Sr2CoO3][CoO2]1.80 (b1/b2 = 1 .80) to 1 20 µV/K f or [ Ca 2CoO3] [CoO2]1.62 (b1/b2 = 1.62) [23]. Following t he vCo equation, vCo = x + 3 = −( a/(b1/b2)) + 4 , x theoretically decreases when b1/b2 decreases. From the Koshibae’s formula, S i ncreases w hen x de creases. The re sults p resented here c an t herefore qu alitatively b e u nderstood i n t his f ramework: S should increase when b1/b2 decreases. For a constant b1/b 2 , it is also possible to modify a, and thus vCo. By doping [Ca 2CoO3][CoO2]1.62 with Ti4+ [20], or Pb3+, b1/b 2 is u nchanged ( 1.62), b ut S i s e qual to 1 65 µVK w ith T i4+ or 160 µV/K for Pb3+, instead of 120 µV/K. If the oxygen content is unchanged, a is increased by the substitution, and vCo decreases, consistently with an increase of S th rough th e K oshibae’s formula. Recent experiments of annealing and titration by Karppinen et al. have shown that doping indeed modies the S value a nd that the smaller content of Co4+ corresponds to the larger S [24].

160

Bi/Ca/Co/O

140

BiPb/Ca/Co/O

120

Bi/Sr/Co/O Bi/Ba/Co/O

80

140

60

S 300k (µV/K)

S (µV/K)

100

40 20

BiCa

130

BiPbCa

120

BiSr

110 100 90

0

BiBa 1.7

1.8

−20 0

50

100

150 T (K)

1.9

2.0

250

300

b 1/b 2

200

FIGURE 11.9 S(T) of the Bi-based family of mist. Inset: S at 300 K as a function of b1/b2.

43722_C011.indd 8

10/3/2008 9:35:05 PM

11-9

Films, Coatings, Adhesives, Polymers, and Thermoelectric Materials

106

200 x = 1.5

105 160

104

x =2 S (µV K−1)

r (Ω cm)

103 102 x = 1.5

101 100

x = 0.5

x =0

0

50

100

(a)

150 200 T (K)

120 x = 0.5 80

x = 0.0

40

10−1 10−2

x = 2.0

250

300

0

350

0

50

100

(b)

150 200 T (K)

250

300

350

FIGURE 11.10 (a) r(T) of the Pb/Sr/Ca/Co/O mists; (b) S(T) of the Pb/Sr/Ca/Co/O mists.

The q uantitative an alysis o f S as a function of doping can unfortunately not be made here due to the number of unknown quantities. O ther te chniques a re no w ne eded to e valuate t he Co4+ concentration and make a more quantitative analysis. 11.2.4.2 Importance of Low Spin State

11.2.4.3

20

100

15

80

10

Electrical Resistivity and Thermal Conductivity

Two different kinds of behaviors are observed in the misfits. At lo w tem perature, o nly t wo m isfits me asured s o f ar a re metallic down to 2 K [19], or with a small increase of resistivity at lo w tem perature [21], a ssociated w ith a sm all p ositive MR. A ll t he ot hers p resent a t ransition to a mo re re sistive behavior b elow 1 00 K, a ssociated to a s trong ne gative M R (−90% at 2 .5 K i n 7 T f or BiC aCo [28]). I n mo st of t hem, a n increase of Seebeck coefficient is associated to an increase of

S (µV K−1)

r (mΩ cm)

Following t he gener alized H eikes f ormula, t he l arge en tropy associated to the low spin states of Co3+ and Co4+ i n t he C oO2 layers w ould b e at t he o rigin o f t he l arge t hermopower. Interestingly, rhodium can partially [25] or totally be substituted to cobalt in the CoO2 l ayer [ 26]. R hodium i s i soelectronic to cobalt, a nd a lways lo w s pin, w ith R h3+ (t 2g6) a nd R h4+ (t 2g5) electronic conguration. We have thus investigated the magnetotransport prop erties of [Bi1.95Ba1.95Rh0.1O4][RhO2]1.8, k nowing that le Rh species are in the low spin states, to compare them to the cobalt mists [27]. The m aterial i s me tallic a nd shows a l arge S of +90 µV/K at 300 K (Figures 11.11 and 11.12). These properties are very close to

the ones measured in BiBaCoO [21], and this similarity is indirect proof of t he low spin states proposed i n t he c ase of cobalt mists. Also, a small positive MR of +5% is observed at low temperature in a magnetic eld of 7 T. This is, to o ur knowledge, a unique case of MR in the rhodium oxides, and this emphasizes the p eculiarity o f t he t ransport p roperties ob served i n t hese CdI2-like layers.

60

40

5 20 0

0

100

200 T (K)

300

FIGURE 11.11 r(T) of the Bi/Ba/Rh/O samples.

43722_C011.indd 9

400

0

0

50

100

150 T (K)

200

250

300

FIGURE 11.12 S(T) of Bi/Ba/Rh/O.

10/3/2008 9:35:06 PM

11-10

Smart Materials 3.0

10−2

2.5

1.5

Sr0.8La0.2RuO3 r (Ω cm)

k (W m−1 K−1)

SrRu0.92O3 2.0

1.0 0.5 0.0

0

50

100

150 T (K)

200

250

SrRuO3 10−3

300 SrRu0.98O3

FIGURE 11.13 k(T) of [Ca 2CoO3][CoO2]1.62. 0

the re sistivity a nd a t ransition to a mo re localized behavior, in the whole temperature range, as shown in Figure 11.10a for the Pb SrCaCoO s eries. N evertheless, i n a ll t hese m aterials, the resistivity remains in the same range at room temperature with r ∼ 10–100 mΩ cm. The thermal conductivity measured in polycrystals is always small w ith k ∼ 2 –3 W/m/K at 3 00 K (Figure 11.13). Following the Wiedemann-Franz law, it can be shown that the electronic part of the thermal conductivity is very small, with k el ∼ 0.03 W/m/K at 3 00 K. These small v alues a re c lose to t he ones re ported i n disordered materials [29] and show that these materials could be go od e xample o f ph onon g lasses a nd ele ctron c rystals (PGEC) [30]. Combining all these values, maximum power factors close to 2 × 1 0−4 W/m/K 2 can be obtained at 300 K. By optimizing the texturation process in these materials, ZT values close to 1 have been reported [5].

11.2.5

SrRuO3: A Metallic Perovskite with a Thermoelectric Power Driven by the Spin Degeneracy Term

The coexistence of a l arge thermopower in the metallic sodium cobaltites or m ists ra ises m any q uestions. In p articular, t he importance of the spin and orbital degeneracies is very interesting a nd maximizing t hese ter ms should be crucial to ge t large Seebeck c oefficients. F ollowing t his ide a, w e h ave de cided to investigate ot her metallic oxides. A mong t hem, t he r uthenium oxides h ave b een i ntensively i nvestigated due to t he ferromagnetism coexisting with good metallicity in SrRuO3 [31], a nd to the superconductivity observed in Sr2RuO4 [32]. Also electronic correlations are expected to be less important in these 4d materials than in the 3d [33]. Only f ew re ports c an b e f ound i n t he l iterature a bout t he thermoelectric p roperties o f S rRuO3 a nd rel ated c ompounds. The thermopower of SrRuO3 is close to +21 µV/K at 3 00 K [34]. The thermoelectric properties of SrRuO3 and related compounds

43722_C011.indd 10

50

100

150

200 T (K)

250

300

350

400

FIGURE 1 1.14 r(T) of Sr RuO3 a nd re lated c ompounds, me asured under 0T for the solid line and under 7T for the dashed line.

have b een i nvestigated, a s a f unction of doping. Fou r d ifferent samples have been investigated to modify the formal ruthenium oxidation state f rom 3.8 (for Sr 0.8La0.2RuO3), 4 ( for SrRuO3 and SrRu0.97O2.94), to 4.35 (for SrRu0.92O3). Figure 11.14 shows the resistivity of the four different samples investigated here. SrRuO3 is metallic with a clear change of slope around Tc ∼ 160 K. A similar behavior is observed for SrRu0.97O2.94, with an even smaller resistivity. This co unterintuitive r esult might be only related to e xtrinsic effects such as a re duction of the g rain b oundaries s cattering. W hen t he f ormal o xidation state is different from 4, the resistivity increases. r is still metallic in S r0.8La0.2RuO3, b ut l arger, a nd r s hows a mo re lo calized behavior in the case of SrRu0.92O3, with dr/dT < 0 at low T. The t hermopower i s p resented i n F igure 1 1.15. A t lo w temperatures, all the curves are very close, with S ∼ T up to 100– 150 K. For T > 100–150 K, S saturates in the case of SrRuO3 and SrRu0.97O2.94 to S300 K ∼ 33 µV/K. F or S r0.8La0.2RuO3, S is o nly slightly a ffected with S300 K ∼ 33 µV/K, still i ncreasing at 3 00 K, and the accident observed at Tc = 160 K in SrRuO3 [35] has disappeared. O n t he ot her h and, i n SrRu0.92O3, S300 K ∼ 21 µV/K a nd dS/dT < 0. For the four different samples, S is modied mostly at T > 100 K, but the differences are not very large. The s pin de generacy a ssociated to t he m ixed v alency Ru 3+/ Ru4+ or Ru 4+/Ru5+ c an b e c alculated f ollowing D oumerc [ 7]: S spin = (−kB/|e|)ln[(2Sn + 1)/(2Sn+1 + 1)], and with S = 1/2 for Ru 3+, S = 1 f or Ru 4+, a nd S = 3 /2 for Ru 5+, we obtain t he t wo l imits: S = + 3 5 µV/K for Ru 3+/Ru4+ and S = + 25 µV/K for Ru 4+/Ru5+. Surprisingly, t hese t wo v alues c orrespond to t he e xperimental limits f ound i n o ur e xperiment. The sp in de generacy ter m directly gives a good estimate of the Seebeck coefficient. Again, this result is not in agreement with the metallic behavior of most of the samples. All metallic oxides do not behave like SrRuO3, as

10/3/2008 9:35:07 PM

11-11

Films, Coatings, Adhesives, Polymers, and Thermoelectric Materials 40 SrRuO3

35

SrRu0.97O2.94

30

S (µV/K)

25 20

SrRu0.92O3

Sr0.8La0.2RuO3

15 10 5 0

0

50

100

150 T (K)

200

250

300

FIGURE 11.15 S(T) of SrRuO3 and related compounds.

for e xample L a1−xSrxTiO3, w hich c an b e de scribed b y a si ngle band m odel, w ith S linear in T [36]. N evertheless, m ists and ruthenates po ssess s imilar f eatures ( S a lmost co nstant f rom 100 K to h igh T, with the S value depending on the spin degeneracy), which are also observed in other oxides. For example, in BaMoO3, S is constant from 200 to 1000 K, close to −30 µV/K. The s pin de generacy ter m g ives −25 µV/K f or M o3+/Mo4+ (S = 3/2/S = 1) or −35 µV/K for Mo4+/Mo5+ (S = 1 /S = 1 /2) [37]. This shows that this spin degeneracy term is important to maximize the Seebeck coefficient.

11.2.6 Conclusion The present results obtained for oxides show the richness of their thermoelectric p roperties. I t m ust b e o utlined t hat f or s everal examples among those given here, these oxides synthesized in air are c hemically s table up to v ery l arge tem peratures ( ∼1000°C). Interestingly, for such h igh temperatures, t he oxides cha racterized b y s emiconducting-like ele ctrical re sistivities e xhibit lo wresistivity values. In this respect, it is remarkable that either metallic o r s emiconducting-like o xides c an h ave l arge S eebeck coefficient v alues, w hich i s c ounterintuitive w hen c ompared to classical thermoelectric materials. The origin of the large thermoelectric power in oxides reects the strong contribution of the spin or orbital degeneracies to the Heikes formula. Again, this experimental conclusion is not predicted in the case of a metal but the example of the metallic ruthenates with constant Seebeck values over a wide range of temperatures demonstrates that theoretical models have still to be developed in the future to explain these properties. With regard to possible applications of thermoelectric oxides, the third important physical parameter is their thermal conductivity. The electronic part is very small due to small charge carrier mean free path, and is mainly phononic and small (∼ 1 to 1 0 W/K/m).

43722_C011.indd 11

At present, though several devices for waste-heat conversion into electricity based on all oxides materials have been made, the best results f or tem perature g radients o f a bout 5 00°C rem ain lo wer than 1 W/cm 2 . Th is tel ls u s t hat ne w o xides w ith en hanced thermoelectric properties are still needed to increase the efficiency of such devices. In particular, oxides containing transition metal cations with large spin and orbital entropies appear to be the most promising candidates.

Acknowledgments The a uthors w ould l ike to t hank t heir c ollaborators i n t he Crismat laboratory: Denis Pelloquin, Christine Martin, Maryvonne Hervieu, B ernard R aveau, R ichard Re toux, D elphine F lahaut, Yannick K lein, C édric Yaicle, Sud ipta Pa l, Vi ncent Ha rdy, a nd Christophe Goupil. Also fruitful collaborations with Jiri Hejtmanek (Institute of Physics, Prague) and Bogdan Dabrowski (Argonne National Laboratory) are acknowledged.

References 1. 2. 3. 4. 5. 6. 7.

T. Kajikawa, Ā ermoelectric H andbook, M acro to N ano, M. Rowe (ed.), CRC Press, Boca Raton, FL, 2006. I. Terasaki, Y. Sas ago, a nd K. Uchinokura, Phys. Re v. B 56, R12685, 1997. P. B oullay, B . D omengès, M . H ervieu, D . G roult, a nd B. Raveau, Chem. Mater. 8, 1482, 1996. A. C. M asset, C. Mic hel, A. M aignan, C. M artin, a nd B. Raveau, Phys. Rev. B 62, 166, 2000. R. Funahashi, I. Matsubara, H. Ikuta, T. Takeuchi, U. Mizutani, and S. Sodeoka, Jpn. J. Appl. Phys. 39, L1127, 2000. P. M. Chaikin and G. Beni, Phys. Rev. B 13, 647, 1976. J. P. Doumerc, J. Solid Stat. Chem. 110, 419, 1994.

10/3/2008 9:35:08 PM

11-12

8. W. Koshibae, K. Tsutsui, a nd S. Maekawa, Phys. Re v. B 62, 6869, 2000. 9. A. A. Taskin, A. N. L avrov, a nd Y. Ando, Phys. Re v. B 73, 121101, 2006. 10. A. Maignan, D. Flahaut, and S. Hébert, Eur. Phys. J. B 39, 145, 2004. 11. A. M aignan, V. C aignaert, B . R aveau, D . K homskii, a nd G. Sawatzky, Phys. Rev. Lett. 93, 026401, 2004. 12. S. H ébert, D . Fla haut, C. M artin, S. L emonnier, J . N oudem, C. Goupil, and A. Maignan, Progress in S olid State Chemistry, 35, 457, 2007. 13. D. B. Marsh and P. E. Parris, Phys. Rev. B 54, 7720, 1996. 14. D. B. Marsh and P. E. Parris, Phys. Rev. B 54, 16602, 1996. 15. S. Pal, S. Hébert, C. Yaicle, C. Martin, and A. Maignan, Eur. Phys. J. B 53, 5, 2006. 16. D. J. Singh, Phys. Rev. B 61, 13397, 2000. 17. H. Yamauchi, K. Sakai, T. Nagai, Y. Matsui, and M. Karppinen, Chem. Mater. 18, 155, 2006. 18. M. Isobe, M. Onoda, M. Shizuya, M. Tanaka, E. TakayamaMuromachi, Journal of the American Chemical Society, 129, 14585, 2007. 19. S. Hébert, S. Lambert, D. Pelloquin, and A. Maignan, Phys. Rev. B 64, 172101, 2001. 20. A. M aignan, S. H ébert, L. P i, D . P elloquin, C. M artin, C. Michel, M. Hervieu, and B. Raveau, Cryst. Eng. 5, 299, 2002. 21. M. Hervieu, A. Maignan, C. Michel, V. Hardy, N. Créon, and B. Raveau, Phys. Rev. B 67, 045112, 2003. 22. A. M aignan, S. H ébert, D . P elloquin, C. Mic hel, a nd J. Hejtmanek, J. Appl. Phys. 92, 1964, 2002. 23. D. P elloquin, S. H ébert, A. M aignan, a nd B . R aveau, Solid Stat. Sci. 6, 167, 2004. 24. M. Karppinen, H. Fjellvag, T. Konno, Y. Morita, T. Motohashi, and H. Yamauchi, Chem. Mater. 16, 2790, 2004. 25. D. Pelloquin, S. Hébert, A. Maignan, and B. Raveau, J. Solid Stat. Chem. 178, 769, 2005. 26. S. Okada and I. Terasaki, Jpn. J. Appl. Phys. 44, 1834, 2005. 27. Y. Klein, S. Hébert, D. Pelloquin, V. Hardy, and A. Maignan, Phys. Rev. B 73, 165121, 2006. 28. A. Maignan, S. Hébert, M. Hervieu, C. Michel, D. Pelloquin, and D. Khomskii, J. Phys. Cond. Matt. 15, 7055, 2003. 29. D. G. Cahill, S. K.Watson, and R. O. Pohl, Phys. Rev. B 46, 6131, 1992. 30. G. A. S lack, in CRC H andbook o f Ā er moelectrics, D . M. Rowe (ed.), CRC Press, Boca Raton, FL, 1995, p. 407. 31. P. B . Allen, H. B erger, O . Cha uvet, L. F orro, T . J arlborg, A. Junod, B. Revaz, and G. Santi, Phys. Rev. B 53, 4393, 1996. 32. Y. Maeno, H. Hashimoto, K. Yoshida, S. Nashizaki, T. Fujita, J. G. Bednorz, and F. Lichtenberg, Nature 372, 532, 1994. 33. I. I. Mazin, D. and J. Singh, Phys. Rev. B 56, 2556, 1997. 34. K. Yamaura, D. P. Young, and E. Takayama-Muromachi, Phys. Rev. B 69, 024410, 2004. 35. Y. Klein, S. Hébert, A. Maignan, S. Kolesnik, T. Maxwell, and B. Dabrowski, Phys. Rev. B 73, 052412, 2006. 36. T. Ok uda, K. Nakanishi, S. Mizas aka, a nd Y. Tokura, Phys. Rev. B 63, 113104, 2001. 37. K. Kurosaki, T. Oyama, H. Muta, M. Uno, and S. Yamanaka, J. Alloys Comp. 372, 65, 2004.

43722_C011.indd 12

Smart Materials

11.3

Electrically Conductive Adhesives

Yi Li, Myung Jin Yim, Kyoung-sik Moon, and C.P. Wong 11.3.1 Introduction Electrically c onductive a dhesives (E CAs) a re c omposites of polymeric m atrices a nd ele ctrically c onductive f illers. T he polymeric resin, such as an epoxy, a silicone, or a polyimide, provides ph ysical a nd me chanical prop erties s uch a s a dhesion, me chanical s trength, i mpact s trength, a nd t he me tal filler (such a s, si lver, gold , n ickel, or copper) conducts electricity. M etal-filled t hermoset p olymers w ere f irst pat ented as EC As i n t he 1 950s [ 1–3]. Re cently, EC A m aterials h ave been identified as one of the major alternatives for lead-containing solders for microelectronics packaging applications. ECAs offer numerous advantages over conventional solder technology, suc h a s en vironmental f riendliness, m ild p rocessing c onditions ( enabling t he u se o f h eat-sensitive a nd low-cost components and substrates), fewer processing steps (reducing processing cost), low stress on the substrates, and fine pitch interconnect capability (enabling the miniaturization of electronic devices) [4–7]. Therefore, conductive adhesives have been used in f lat panel displays such as liquid crystal d isplay ( LCD), a nd sm art c ard app lications a s a n interconnect m aterial a nd i n f lip-chip a ssembly, c hip-scale package (CSP), a nd ba ll g rid a rray ( BGA) app lications a s a replacement f or s older. H owever, no c urrently c ommercialized EC As c an re place t in-lead me tal s olders i n a ll applications due to some challenging issues such as lower electrical conductivity, conductivity fatigue (decreased conductivity at elevated temperature and humidity aging or normal use condition) i n rel iability te sting, l imited c urrent-carrying c apability, a nd p oor i mpact s trength. Table 11.2 g ives a gener al comparison b etween t in-lead s older a nd gener ic c ommercialized ECAs [8]. Depending on the conductive filler loading level, ECAs are divided into ICAs, ACAs, and NCAs. For ICAs, the electrical conductivity in all x-, y-, and z-directions is provided due to high f iller c ontent e xceeding t he p ercolation t hreshold. For ACAs o r N CAs, t he ele ctrical c onductivity i s p rovided

TABLE 11.2 Conductive Adhesives Compared with Solder Characteristic Volume resistivity Typical junction R Thermal conductivity Shear strength Finest pitch Minimum processing temperature Environmental impact Thermal fatigue

Sn/Pb Solder

ECA

0.000015 Ω cm 10–15 mW 30 W/m °K 2200 psi

0.00035 Ω cm <25 mW 3.5 W/m °K 2000 psi

300 mm 215°C Negative Yes

<150–200 mm 150°C–170°C Very minor Minimal

10/3/2008 9:35:08 PM

11-13

Films, Coatings, Adhesives, Polymers, and Thermoelectric Materials

only i n z-direction b etween t he ele ctrodes o f t he a ssembly. Figure 11.16 shows the schematics of the interconnect structures and typical cross-sectional images of flip-chip joints by ICA, ACA, and NCA materials illustrating the bonding mechanism for all three adhesives.

11.3.2

Isotropically Conductive Adhesives

ICAs, also called as “polymer solder,” are composites of polymer re sin a nd c onductive  llers. The ad hesive m atrix i s used to f orm a me chanical b ond f or t he i nterconnects. B oth thermosetting a nd t hermoplastic m aterials a re u sed a s t he polymer matrix. Epoxy, cyanate ester, silicone, polyurethane, etc. are widely used thermosets, and phenolic epoxy, maleimide acrylic preimidized polyimide, etc. are the commonly used thermoplastics. An attractive advantage of thermoplastic ICAs is that they are reworkable, i.e., can easily be repaired. A major drawback of thermoplastic ICAs, however, is the degradation of a dhesion at h igh t emperature. A nother d rawback of p olyimide-based I CAs i s t hat t hey gener ally c ontain s olvents. During heating, voids are formed when the solvent evaporates. Most of commercial ICAs are based on thermosetting resins. Thermoset epoxies are by far the most common binders due to the sup erior ba lanced p roperties, suc h a s e xcellent ad hesive strength, go od c hemical a nd c orrosion re sistances, a nd lo w cost, w hile t hermoplastics a re u sually adde d to a llow s oftening and rework under moderate heat. The conductive llers provide the composite with electrical conductivity through contact between the conductive particles. The possible conductive  llers include silver (Ag), gold (Au), nickel (Ni), copper (Cu), and

carbon i n v arious f orms ( graphites, c arbon na notubes, e tc.), sizes, and shapes. Among different metal particles, silver akes are t he mo st c ommonly u sed c onductive  llers f or c urrent commercial I CAs b ecause o f t he h igh c onductivity, si mple process, a nd t he m aximum c ontact w ith  akes. I n addition, silver is unique a mong a ll t he cost-effective metals by nature of its conductive oxide. Oxides of most common metals are good ele ctrical i nsulators, a nd c opper p owder, f or e xample, becomes a p oor c onductor a fter a ging. N ickel- a nd co pperbased conductive adhesives generally do not h ave good resistance s tability, b ecause b oth n ickel a nd c opper a re e asily oxidized. ICAs have been used for die attach adhesives [9,10]. Recently, ICAs have also been considered as an alternative to tin/lead s olders i n su rface mo unt te chnology ( SMT) [11,12],  ip-chip [13], a nd ot her app lications a nd a l arge a mount o f effort has been conducted to improve the properties of ICAs in the past few years. 11.3.2.1

Improvement of Electrical Conductivity of ICAs

Polymer-based ECAs typically have lower electrical conductivity than S n/Pb s olders. T o enha nce the ele ctrical co nductivity o f ICAs, various methods have been used and signicant improvement of conductivity of ICA has been achieved. 11.3.2.2

Increase of Polymer Matrix Shrinkage

In general, ICA pastes exhibit insulative property before cure, but the conductivity increases dramatically after c ure. ICAs achie ve electrical conductivity during the polymer curing process caused by the shrinkage of polymer binder. Accordingly, ICAs with high

Silver flakes

Contact pads

IC electrode Polymer matrix Silver flakes

Substrate pad

Substrate (a)

Component or IC

Substrate (c)

Polymer matrix

(b)

IC Bump

Conductive spheres Polymer matrix

Substrate pad Polymer matrix

Conductive ball

(d)

IC Bump Polymer matrix Substrate electrode

(e)

(f)

FIGURE 11.16 Schematic illustrations and cross-sectional views of (a,b) ICA, (c,d) ACA, and (e,f) NCA ip-chip bondings.

43722_C011.indd 13

10/3/2008 9:35:08 PM

11-14

Smart Materials

cure shrinka ge g enerally exhib it hig her co nductivity [14]. With increasing cr oss-linking den sity of IC As, t he s hrinkage of t he polymer ma trix incr eases, a nd subs equently, the r esistivity o f ICAs decreases. For epoxy-based ICAs, a small amount of a multifunctional epoxy resin can be added into an ICA formulation to increase cross-linking density, shrinkage, and thus increase electrical conductivity.

Component

Metallic network

11.3.2.3 In Situ Replacement of Lubricants on Ag flakes

Substrate

An ICA is generally composed of a polymer binder and Ag akes. There is a thin layer of organic lubricant on the Ag ake surface. This lubricant layer plays an important role for the performance of ICAs, including t he dispersion of t he Ag  akes in t he ad hesives and the rheology of the adhesive formulations [15,16]. This organic lubricant layer, typically a fatty acid such as stearic acid, forms a silver salt complex between the Ag surface and the lubricant. However, this lubricant layer affects conductivity of an ICA because it is electrically insulating. To improve conductivity, the organic lubricant l ayer should b e pa rtially or f ully removed or replaced during the curing of ICA. Short chain dicarboxylic acid is one of t he su itable lubricant removers b ecause of t he s trong affinity and more acidic of carboxylic functional group (–COOH) to silver. With the addition of only small amount of short chain dicarboxylic acid, the conductivity of ICA can be improved signicantly due to t he easier electronic tunneling/transportation between A g  akes a nd sub sequently t he i ntimate  ake-ake contact [17]. 11.3.2.4

Incorporation of Reducing Agent in Conductive Adhesives

Silver  akes a re b y f ar t he mo st u sed  llers f or c onductive adhesives due to t he unique properties of the high conductivity o f si lver o xide c ompared to ot her me tal o xides, mo st o f which are insulative. However, the conductivity of silver oxide is s till i nferior to me tal i tself. Therefore, i ncorporation of reducing agents would further improve the electrical conductivity of ICAs. A ldehydes a re one of t he re ducing a gents t hat can b e i ntroduced. Obv iously i mproved c onductivity w as reported i n ICAs d ue to t he re action b etween a ldehydes a nd silver oxides that exist on the surface of metal llers during the curing process [18]: R-CHO + Ag 2 O → R − COOH + 2Ag (1

1.3)

The oxidation product of aldehydes, carboxylic acids, which are stronger ac ids a nd h ave s horter mole cular leng th t han s tearic acid, can also partially replace or remove the stearic acid on Ag akes and contribute to the improved electrical conductivity. 11.3.2.5

Low-Temperature Transient Liquid Phase Fillers

Another app roach f or i mproving ele ctrical c onductivity i s to i ncorporate t ransient l iquid-phase me tallic  llers i n I CA formulations. The  ller u sed i s a m ixture o f a h igh-melting-

43722_C011.indd 14

Polymer binder

FIGURE 11.17 Diagram of t ransient l iquid phase si ntering c onductive adhesives.

point metal powder (such as Cu) and a low-melting-point alloy powder (such a s Sn -Pb o r Sn -In). The lo w-melting-alloy  ller melts when its melting point is achieved during the cure of the polymer m atrix. The l iquid ph ase d issolves a nd wets t he h igh melting point particles. The liquid exists only for a short period of t ime a nd t hen f orms a n a lloy a nd s olidies. The electrical conduction i s e stablished t hrough a p lurality o f me tallurgical connections in situ formed from t hese t wo powders in a p olymer bi nder (Figure 11.17). The polymer binder w ith acid f unctional ingredient uxes both the metal powders and the metals to b e jo ined a nd f acilitates t he t ransient l iquid b onding o f t he powders to f orm a s table me tallurgical ne twork f or ele ctrical conduction, a nd a lso f orms a n i nterpenetrating p olymer ne twork p roviding ad hesion. H igh ele ctrical c onductivity c an b e achieved using this method [19,20]. One critical limitation of this technology is that the numbers of combinations of low and high mel ting  llers a re l imited. O nly c ertain c ombinations o f two metallic llers, which are mutually soluble, exist to form this type of metallurgical interconnection. 11.3.2.6

Low-Temperature Sintering of Nanosilver Fillers

Recently, na nosized c onductive pa rticles w ere p roposed to b e used as conductive  llers i n I CAs f or h igh-performance a nd ne-pitch interconnects. Although the nanosilver  llers in ICAs can re duce t he p ercolation t hreshold, t here h as b een c oncern that incorporation of nanosized llers may introduce more contact s pots d ue to h igh su rface a rea a nd c onsequently i nduce higher r esistivity co mpared t o m icron-sized  llers. A re cent study s howed t hat na nosilver pa rticles c ould e xhibit si ntering behavior at curing temperature of ICAs [21]. Typically, applicavtion of nano llers i ncreases t he contact resistance a nd reduces the electrical performance of the ICAs. The number of contacts between the small particles is larger than that between the large particles as shown in Figure 11.18a and b. The overall resistance of the ICA formulation is the sum of the resistance of llers, the resistance between  llers, and the resistance between  ller and pads ( Equation 1 1.4). I n o rder to de crease t he o verall c ontact resistance, the reduction of the number of contact points between the pa rticles m ay b e e ffective. I f na noparticles a re si ntered

10/3/2008 9:35:09 PM

11-15

Films, Coatings, Adhesives, Polymers, and Thermoelectric Materials

FIGURE 11.18 Schematic illustration of particles between the metal pads. (From Jiang, H.J., Moon, K., Lu, J., and Wong, C.P., J. Electron. Mater., 34, 1432, 2005.)

together, then the contact between  llers will be fewer. This will lead to smaller contact resistance (Figure 11.18c). By using effective su rfactants f or t he d ispersion a nd e ffective c apping t hose nanosized si lver  llers i n EC As, obv ious si ntering b ehavior o f the nano llers can be achieved. The sintering of nanosilver  llers improved the interfacial properties of conductive llers and polymer m atrices a nd re duced t he c ontact re sistance b etween llers. Therefore, an improved electrical conductivity of nanosilver- lled ICAs can be achieved. R total = R btw fillers + R filler to bond pad + R fillers (1

1.4)

11.3.2.2 Reliability Enhancement of ICA Interconnects One of the critical reliability issues of ICAs is that contact resistance between an ICA and nonnoble metal nished components increases dramatically during an elevated temperature and humidity a ging. The N ational C enter o f M anufacturing a nd Science (NCMS) dened the stability criterion for solder replacement conductive adhesives as a c ontact resistance shift (increase) of less than 20% after a ging at 85° C/85%RH f or 5 00 h [ 22]. However, most currently commercialized ICAs cannot pass the reliability test on nonnoble metal surfaces such as Sn/Pb, Sn, Cu, Ni, etc. However, some of the improved ICAs have been developed recently to satisfy the reliability requirements. 11.3.2.2.1

11.3.2.2.2 Low-Moisture Absorption Moisture i n p olymer c omposites h as b een k nown to h ave a n adverse e ffect on both mechanical and electrical properties of epoxy la minates. Eff ects of moisture absorption on conductive adhesive j oints in clude d egrading b ulk m echanical s trength,

Electrical conduction

Ag flake Ag Polymer binder Nonnoble metal (Before corrosion)

Mechanism Underlying Unstable Contact Resistance

Simple oxidation and galvanic corrosion of the nonnoble metal surfaces a re t he t wo p ossible me chanisms for u nstable c ontact resistance of ICAs. Studies have shown that galvanic corrosion at the interface between an ICA and nonnoble metal is the dominant me chanism f or t he u nstable c ontact re sistance [ 23,24] (Figure 1 1.19). U nder t he a ging c onditions ( for e xample, 85°C/85%RH), the nonnoble metal acts as an anode, and is oxidized by losing electrons, and then turns into metal ion (M − ne− = M n+). The noble metal acts as a cathode and its reaction generally is 2H2O + O 2 + 4 e− = 4 OH−. Then Mn+ combines with OH− to f orm a me tal h ydroxide t hen f urther o xidizes to metal oxide. After corrosion, a layer of metal hydroxide or metal oxide is formed at the interface. Because this layer is electrically

43722_C011.indd 15

insulating, the contact resistance increases dramatically. A ga lvanic c orrosion p rocess h as s everal c haracteristics: ( 1) o ccurs only under wet conditions, (2) an electrolyte must be present, and (3) oxygen generally accelerates the process, and (4) dissimilar metals should be present. Based on the mechanism of unstable contact resistance of ICAs, several methods can be applied to stabilize the contact resistance and improve the reliability.

Electrical conduction

Ag flake

Polymer binder

Ag

Condensed water solution Nonnoble metal

Metal hydroxide or oxide

(After corrosion)

FIGURE 1 1.19 corrosion.

Metal h ydroxide or o xide for mation a fter galvanic

10/3/2008 9:35:09 PM

11-16

Smart Materials

decreasing interfacial ad hesion strength, and causing delamination, p romoting t he g rowth o f vo ids p resent i n t he jo ints, giving rise to swelling stress in the joint interfaces and inducing t he f ormation o f me tal o xide l ayers re sulted f rom c orrosion. The water condensed f rom t he absorbed moisture at t he interface between an ECA and metal surface forms the electrolyte solution required for galvanic corrosion. Ther efore, one way to p revent ga lvanic corrosion at t he interface between an ICA and the nonnoble metal surface and achieve high reliability is to select ICAs with lower moisture absorption. Typically, ECAs formulated with high purity, low moisture absorption, and less p olar re sins h ave lo wer mo isture a bsorption a nd c ontact resistance [25]. 11.3.2.2.3

Oxygen Scavengers

Since oxygen accelerates galvanic corrosion, oxygen scavengers can be used in ICAs to slow down the corrosion rate [23]. When ambient oxygen molecules diff use through the polymer binder, they re act w ith t he o xygen s cavenger a nd a re c onsumed. The main mechanism for oxygen scavengers to inhibit the corrosion is t he c athodic me chanism, w hich i s ba sed o n t he lowering o f oxygen concentrations. In the system consisting of metal, water, oxygen, and oxygen scavenger, the cathodic reaction of O2 competes with its chemical reduction. If the overall corrosion process is cathodically controlled, the reduction in corrosion rate depends on t he rel ative m agnitudes o f t he r ates o f t he ele ctrochemical cathodic process Rel a nd t he c hemical p rocess Rch. The overall rate of oxygen reduction Rox is given by the sum of the rates of the two processes indicated in the following equation: R corr = R ox − R ch (1

1.5)

Since the rate of corrosion, Rcorr, is equal to t he rate of the electrochemical cathode process Rel, the corrosion rate is given by R corr = R ox − R ch (1

1.6)

It can be seen from Equation 11.6 that the smaller the corrosion rate, t he g reater t he r ate o f t he c hemical re duction o f o xygen. Therefore, the reactivity of an oxygen scavenger with oxygen is an i mportant c onsideration o f i ts p roperties. S ome c ommonly used oxygen scavengers include sultes (such as sodium sulfates (Na 2SO4), hyd razine ( H2N-NH2), c arbohydrazide ( H2N-NHCO-NH-NH2), diethylhydroxylamine ((C2H5)2N-OH), and hydroquinone (HO-C6H4-OH). 11.3.2.2.4 Corrosion Inhibitors Another method of preventing galvanic corrosion and stabilizing contact resistance is t he use of corrosion i nhibitors i n ICA formulations [25,26]. In general, organic corrosion inhibitors are chemicals that adsorb on metal surfaces and act as a passivation barrier layer between the metal and the environment by forming an inert lm over the metal surfaces [27–31]. Thus, the metal nishes can be protected. Some chelating compounds are especially effective in preventing metal corrosion. Appropriate selection of

43722_C011.indd 16

corrosion inhibitors can be very effective in protecting the metal nishes from corrosion. However, the effectiveness of the corrosion inhibitors is highly dependent on the types of contact surfaces, which is governed by the affi nity o f t he pa ssivating materials toward t he coated metals. E ffective corrosion i nhibitors h ave b een d iscovered f or Sn /Pb, Cu , A l, a nd Sn su rfaces, respectively [23,26,31]. 11.3.2.2.5 Sacrificial Anode To improve the contact resistance stability, applying a sacricial anode is a nother efficient method. For ECAs during aging, t he larger the difference in electrochemical potential, the faster the corrosion de velops. T able 1 1.3 s hows t he ele ctrode p otential values of s ome c ommon me tals. G enerally, me tals w ith a lo w potential ten d to c orrode f aster a nd s how i ncreased c ontact resistance t han t hose w ith a h igh p otential v alue. Ther efore, when applying sacr icial m aterials w ith lower ele ctrochemical potential than those of electrode-metal pads into ICAs, the sacricial materials are preferably corroded rst and, thus, can protect the metal nishes. This corrosion control is very important in reliability issues of the conductive adhesive joints. The addition o f lo w c orrosion p otential i ndividual me tals, me tal m ixtures, or metal alloys reduces the electrode potential of ICAs, or in other words, narrows down the potential gap between the ICA and the metal nishes. Thus, these sacricial anode materials act as an an ode i n th is c onguration a nd t hey a re c orroded  rst instead of the metal nishes, resulting in protecting the surfaces at the cathode [32–34]. 11.3.2.2.6

Oxide-Penetrating Particles

Another app roach o f i mproving c ontact re sistance s tability during aging is to incorporate some electrically conductive particles, w hich h ave s harp e dges, i nto t he ICA f ormulations. The pa rticle i s c alled o xide-penetrating  ller. F orce m ust b e provided to drive the oxide-penetrating particles through oxide layer and hold them against the adherent materials. Th is can be accomplished b y em ploying p olymer b inders t hat s how h igh shrinkage when cured [35]. (Figure 11.20) Th is concept is used in p olymer-solder, w hich h as go od c ontact re sistance s tability

TABLE 11.3 Electrode Potentials of Selected Metal–Ion Pairs Electrode Reaction Au − 3e = Au Pt − 2e− = Pt2+ Ag − e− = Ag− 2H2O + O2 + 4e− = 4OH− Cu − 1e− = Cu+ Cu − 2e− = Cu2+ 2H+ + 2e− = H2 Pb − 2e− = Pb2+ Sn − 2e− = Sn2+ Ni − 2e− = Ni2+ −

3+

Standard Potential (eV) 1.50 1.20 0.80 0.40 0.34 0.52 0.00 −0.13 −0.14 −0.25

10/3/2008 9:35:10 PM

11-17

Films, Coatings, Adhesives, Polymers, and Thermoelectric Materials

Component

Oxide layer

Ag particle Polymer binder

Oxide-penetrating particles

Substrate

FIGURE 11.20 Joint connected with an ICA containing oxide-penetrating particles and silver powders.

with standard surface-mounted devices (SMDs) on both soldercoated and bare circuit boards. 11.3.2.3

Enhancement of Adhesion Strength

Another i ssue o f c onductive ad hesive i s t he lo wer ad hesion strength. High adhesion strength is a c ritical parameter in  ne pitch i nterconnection t hat i s f ragile to s hocks en countered during assembly, handling, and lifetime usage. There are two types of adhesion mechanisms, physical bonding and chemical bonding, which contribute to the overall adhesion s trength o f p olymer o n a su rface [36]. C hemical b onding involves the formation of covalent or ionic bonds to link between the p olymer a nd t he sub strate. P hysical b onding i nvolves mechanical interlocking or physical adsorption such as van der Waals f orce b etween t he p olymer a nd t he su rface o f t he sub strate. I n c ases w here t he mole cules of t he p olymer a re h ighly compatible with the molecules of the substrate, they interact to form an interdiff usion l ayer. I n me chanical i nterlocking, p olymer a nd substrate interact on a mo re macroscopic level, where the p olymer  ows i nto t he c revices a nd t he p ores o f sub strate surface to establish adhesion. Therefore, a polymer is expected to have higher adhesion on a rougher surface because there is more surface area and “anchors” to allow for interlocking between the polymer and the substrate. 11.3.2.3.1

Plasma Cleaning and Vacuum Process

Plasma cleaning of surfaces has been considered as one of the effective approaches to enhance the adhesion strength of conductive adhesives [37]. During the plasma etching process, the plasma charged pa rticles, such as radicals react w ith t he contaminants a nd t hose lo ng c hains o rganic mole cules c an b e broken do wn i nto sm all ga seous pa rticles ( mostly ga seous water a nd c arbon-oxide conjunctions). These pa rticles c an be evacuated from the system by the vacuum pump. After the surface is cleaned of oxide, a layer of atoms available in the chamber will deposit on the clean surface, protecting the metal from new oxidation. Furthermore, the plasma can also etch the surface t o en hance t he m echanical i nterlocking m echanism f or adhesion improvement. With plasma treatment, the adhesion strength of conductive adhesives could be improved. However,

43722_C011.indd 17

the i mprovement a lso requires a h igh vacuum process due to the possible corrosion on the highly cleaned surface caused by the moisture in ECAs. A fter vacuum process, the surface was changed b y remo val o f a ir a nd w ater a nd t he c ontraction o f adhesives. 11.3.2.3.2 Application of Coupling Agents in ECA Another app roach to i mprove ad hesion i s b y u sing c oupling agents [38]. C oupling a gents a re o rganofunctional c ompounds based o n si licon, t itanium, o r z irconium. F or e xample, R -X(O-R′)3, where, X = Si, Ti or Zr, R = organic chain that interacts with the polymer, and R′ = organic chain that interacts with the substrate. A c oupling a gent c onsists o f t wo pa rts a nd ac ts a s intermediary to “couple” the inorganic substrate and polymer. Silane-coupling agents have been commonly used to improve the a dhesion p erformance. I n c onductive a dhesives, appl ication o f s pecic si lane-coupling a gents w ith app ropriate c oncentration c an i ncrease t he a dhesion s trength on d ifferent metal surfaces. Although silane-coupling agents are mostly used, some other coupling agents with various functional groups, such as thiol, carboxylate-coupling agents, are also used for specic surfaces [39]. 11.3.2.3.3

Optimization of Elastic Modulus

In order to en hance the adhesion, another approach is to lower the elastic modulus of adhesive resins. By using low elastic modulus resins, t he t hermal stress at t he ad hesion i nterface c an be reduced and, thus, the improved the adhesion strength [40,41]. In g eneral, th e a dhesion s trength i ncreases wi th l owering th e elastic mo dulus. H owever, to o lo w mo dulus v alue de teriorates the c ohesive f orce a nd t hus, de creases t he ad hesion s trength. Therefore, the elastic modulus needs to be optimized to improve the adhesion properties. In addition to the methods listed above, some other factors such as curing conditions, surface roughness and s tructures o f I C pac kaging m ay a lso a ffect t he ad hesion strength of conductive adhesives. 11.3.2.4

Improvement of Impact Performance

Impact performance is a critical property of solder replacement ICAs. Effort has been continuing in developing ICAs that have

10/3/2008 9:35:11 PM

11-18

Smart Materials

higher i mpact s trength a nd c an pa ss t he d rop te st, a s tandard test used to evaluate the impact strength of ICAs. Nanosized metal particles were used in ICAs to improve the electrical c onduction a nd me chanical s trength. U sing na nosized particles, agglomerates are formed due to surface tension effect, a nd h igh-power u ltrasonication is n eeded t o d isperse these agglomerates. Another approach is to simply decrease the  ller lo ading to i mprove t he i mpact s trength [ 42]. H owever, such a p rocess reduces t he electrical properties of t he conductive a dhesives. R ecent d evelopment w as re ported w here c onductive ad hesives were developed using resins of low modulus so t hat t his c lass o f c onductive ad hesives c ould a bsorb t he impact energy developed during the drop [43]. Conformal coating w ith si licone o r p olyurethanes o f t he S MDs w as u sed to improve mechanical strength. It was demonstrated that conformal c oating c ould i mprove t he i mpact s trength of c onductive adhesives jo ints [44]. Fu rthermore, el astomer-modied epoxy resins w ith en hanced lo ss mo dulus ( high t an q va lues) w ere also u sed to en hance t he d amping p roperties a nd t he i mpact performance of ICAs [45]. 11.3.2.5

Improvement of Reliability in Thermomechanical Cycling Test

The poor t hermal c ycling (TC) performance of t he ECA joints has been another reliability issue for board level interconnects. It has been found rather difficult to c ontrol the TC f ailures of the ECA joints. Generally, the failure of the electrical interconnection during the TC test could be caused by many factors such as coefficient o f t he t hermal e xpansion (CTE) m ismatch b etween the IC component chip, t he interconnection materials, a nd t he substrates, elastic moduli difference of these components, adhesion strength of t he i nterconnect materials on t he IC chip a nd the substrate, the mechanical properties of the IC chips, the glass transition or the softening point of the ECA materials, moisture uptake in the interface and the bulk ECAs, the surface or interface property change and so forth. Especially, the thermal stress in t he EC A jo ints gener ated b y a h uge tem perature d ifference during the TC and the interfacial delamination due to the adhesion degradation could be t he critical reasons. In t his aspect, a feasible solution to the TC failure problem is to introduce exible molecules into the epoxy resin matrix. By releasing the thermal stress with the exible molecules, the thermomechanical stresses could b e d ramatically re duced a nd t he EC A/component jo int interfaces c ould ke ep i ntact t hrough t he TC te st [46]. I n add ition, application of low CTE adhesives by using high loading of  llers has been proved to i nduce lower shear strain induced by CTE m ismatch b etween c hip a nd b oard u nder tem perature cycling [47–49]. 11.3.2.6

Electrochemical Migration Control

Although silver is the most widely used conductive ller in ICAs, silver migration has long been a rel iability concern in the electronic industry. Metal migration is an electrochemical process, whereby metal (e.g., silver), in contact with an insulating material, in a humid environment and under an applied electric eld,

43722_C011.indd 18

leaves its i nitial lo cation i n ionic form a nd deposits at a nother location [ 50]. I t i s c onsidered t hat a t hreshold vol tage e xists above which the migration starts. Such migration may lead to a reduction in electrical spacing or cause a s hort circuit between interconnections. The m igration p rocess b egins w hen a t hin continuous lm of water forms on an insulating material between oppositely charged electrodes. When a potential is applied across the electrodes, a c hemical reaction takes place at t he positively biased ele ctrode w here p ositive me tal io ns a re f ormed. These ions, t hrough io nic c onduction, m igrate to ward t he ne gatively charged cathode and over time, they accumulate to form metallic dendrites. A s t he dendrite g rowth i ncreases, a re duction of electrical spacing occurs. Eventually, the dendrite silver growth reaches the anode and creates a me tal bridge between the electrodes, resulting in an electrical short circuit [51]. Although other metals may also migrate under specic environment, silver is more susceptible to migration, mainly due to the high solubility of silver ion, low activation energy for silver migration, high tendency to form dendrite shape and low possibility to form stable passivation oxide layer [52–54]. The rate of silver m igration i s i ncreased b y (1) a n i ncrease i n t he app lied potential; (2) a n i ncrease i n t he t ime of t he applied p otentials; (3) an increase in the level of relative humidity; (4) an increase in the presence of ionic and hydroscopic contaminants on the surface of the substrate; and (5) a decrease in the distance between electrodes of the opposite polarity. In order to reduce silver migration and improve the reliability, several me thods h ave b een re ported. T he me thods i nclude: (1) alloying t he si lver w ith a n a nodically s table me tal suc h a s palladium [51] or platinum [55]; (2) using hydrophobic coating over t he PW B to s hield i ts su rface f rom h umidity a nd io nic contamination [56], si nce water a nd contaminates c an ac t a s a transport medium and increase the rate of migration; (3) plating of silver with metals such as tin, nickel, or gold to protect the silver llers and reduce migration; (4) coating the substrate with polymer [57]; (5) applying benzotriazole (BTA) and its derivatives in the environment [58]; (6) employing siloxane epoxy polymers as diff usion ba rriers d ue to t he e xcellent ad hesion o f si loxane epoxy polymers to conductive metals [59]; and (7) chelating silver llers in ECAs with molecular monolayers [60].

11.3.3

Anisotropic Conductive Adhesives and Films

ACAs or anisotropic conductive lms ( ACFs) p rovide u nidirectional electrical conductivity in the vertical or Z-axis. This directional c onductivity i s a chieved b y u sing a r elatively l ow v olume loading of conductive llers (5 to 20 vol.%). The low-volume loading is insufficient for interparticle contact and prevents conductivity in the X–Y plane of the adhesives. The Z-axis adhesive, in lm or paste form, is interposed between the surfaces to be connected. Heat and pressure are simultaneously applied to t his stack-up until the particles bridge the two conductor surfaces. The ACF b onding me thod i s a t hermocompression b onding process as shown in Figure 11.21. In case of tape-carrier packages

10/3/2008 9:35:11 PM

11-19

Films, Coatings, Adhesives, Polymers, and Thermoelectric Materials

Circuit ACF

Electrode Substrate

an i ntimate c ontact, a nd t hus to a lo w c ontact re sistance. I n addition, b ecause t he b onding o f a ll b umps o f a c hip i s p erformed simultaneously, uniform conductivity of all joints in the chip re quires t he b onding si tuation o f e very jo int to b e c ompletely the same. 11.3.3.1.1

Heat and pressure

Circuit Substrate

FIGURE 11.21 Thermocompression bonding using ACF.

(TCP) bonding, the ACF material is attached on a glass substrate and p reattached. F inal b onding i s e stablished b y t he t hermal cure o f t he A CF re sin, t ypically at 1 50°C–220°C, 3 –20 s, a nd 100–200 MPa, and the conductive particle deformation between the electrodes of TCP and the glass substrate by applied bonding pressure. Interconnection te chnologies u sing ACFs a re major pac kaging methods for at panel display modules and have high resolution, a re l ight w eight, h ave t hin p ro le, a nd lo w c onsumption power [61], and already successfully implemented in the forms of out le ad b onding ( OLB),  ex to P CB b onding ( PCB), rel iable direct ch ip a ttach such a s ch ip-on-glass ( COG), ch ip-on-lm (COF) f or  at pa nel d isplay mo dules [ 62–65], i ncluding L CD, plasma d isplay pa nel ( PDP), a nd o rganic l ight em itting d iode display (OLED). As for the small and ne pitched bump of driver ICs to be packaged, ne pitch capability of ACF interconnection is much more desired for COG, COF, and even OLB assemblies. There h ave b een a dvances i n d evelopment works for i mproved material system and design rule for ACF materials to meet ne pitch capability and better adhesion characteristics of ACF interconnection for at panel displays. In add ition to t he L CD i ndustry, ACA/ACF i s no w  nding applications i n f lex c ircuits a nd SM T for C SP, appl icationspecic integrated circuit (ASIC), and  ip-chip attachment for cell ph ones, r adios, p ersonal d igital a ssistants ( PDAs), s ensor chip in digital cameras, and memory chip in laptop computers. In s pite of t he w ide applications of ACA/ACF, t here a re s ome key i ssues t hat h inder t heir i mplementations a s h igh p ower devices. T he ACA/ACF joints gener ally h ave lower ele ctrical conductivity, poor current carrying capability, and electrical failure during TC. 11.3.3.1 Improvement of Electrical Properties of ACA/ACF As the conductivity of ACA joints is directly determined by the mechanical contact between the terminals of chips and the electrodes on chip carriers, the bonding force plays a critical role in the electric performance. High bonding pressure is favorable to

43722_C011.indd 19

Organic Self-Assembly Monolayer Enhancement

In order to en hance the electrical performance of ACA materials, organic monolayers have been introduced into the interface between metal  llers and the metal-nished bond pad of ACAs [66,67]. These organic molecules adhere to the metal surface and form ph ysiochemical b onds, w hich a llow e lectrons to  ow, as such, it reduces electrical resistance and enables a h igh current ow. The unique electrical properties are due to t heir tuning of metal w ork f unctions b y t hose o rganic mo nolayers. The metal surfaces can be chemically modied by the organic monolayers and the reduced work functions can be achieved by using suitable o rganic mo nolayer c oatings. A n i mportant c onsideration when examining the advantages of organic monolayers pertains to t he a ffinity a nd t hermal s tability o f o rganic c ompounds to specic me tal su rfaces. Table 11.4 g ives t he e xamples of mole cules p referred f or m aximum i nteractions w ith s pecic metal nishes. 11.3.3.1.2 Low-Temperature Sintering of Nano Ag Filled ACA One of the concerns for ACA/ACF is the higher joint resistance since i nterconnection u sing A CA/ACF rel ies o n me chanical contact, unlike the metal bonding of soldering. An approach to minimize t he joint resistance of ACA/ACF is to m ake t he conductive  llers f use each ot her a nd form metallic joints such a s metal s older jo ints. H owever, to f use me tal  llers i n p olymers does not app ear feasible, since a t ypical organic printed circuit board (Tg ∼ 125°C), on which the metal lled polymer is applied, cannot withstand such a high temperature; the melting temperature (Tm) of Ag, for example, is around 960°C. Research showed that Tm a nd si ntering tem peratures o f m aterials c ould b e d ramatically reduced by decreasing the size of the materials [68–70]. It h as b een re ported t hat t he su rface p remelting a nd si ntering processes are a primary mechanism of the Tm depression of the ne n anoparticles (<50 nm). For n anosized p articles, si ntering

TABLE 11.4 Potential Organic Monolayer Interfacial Modiers for Different Metal Finishes Formula

Compound

R-S-H R-COOH R-CºN R-N=C=O N

Metal Finish

Thiols Carboxylates Cyanides

Au, Ag, Cu, Pt, Zn Fe/FexOy, Ti/TiO2, Ni, Al, Ag Au, Ag, Pt

Isocyanates Imidazole and derivatives

Pt, Pd, Rh, Ru Cu

Organosilicone derivatives

SiO2, Al2O3, quartz, glass, mica, ZnSe, GeO2, Au

NH

R-SiOH

10/3/2008 9:35:12 PM

11-20

Smart Materials Polymer binder

Conductive filler Electronic

(a)

Circuit substrate

(b)

FIGURE 11.22 Schematic illustrations of ACA/ACF with (a) microsized conductive  llers and (b) nanosized conductive  llers.

behavior could occur at m uch lower temperatures, as such, the use of the ne me tal particles in ACAs would be promising for fabricating h igh ele ctrical p erformance A CA jo ints t hrough eliminating the interface between metal  llers. The application of nanosized particles can also increase the number of conductive  llers o n e ach b ond pad a nd re sult i n mo re c ontact a rea between  llers and bond pads as illustrated in Figure 11.22. For the sintering reaction in a certain material system, temperature and duration are the most important parameters, in particular, the sintering temperature. 11.3.3.2 Improvement of Thermal Properties of ACA/ACF The thermal performance of an adhesively assembled chip is of vast interest as power dissipation in the high-performance, highfrequency devices. The current high-performance microprocessor operates in excess of 5 GHz, as a consequence, enormous heat has b een gener ated w hich ne eds to b e d issipated for rel iability concern. Power dissipations have been simulated by Sihlbom et al. for both ICA and ACA ip-chip joints [71]. They concluded that the ACA ip-chip joint is more effective in transferring heat to t he sub strate f rom t he p owered c hip t han t he I CA jo int, because th e a dhesive th ickness i s s o m uch th inner th an th e ICA joint. For t he ACA interconnect joints to del iver high current, not only a low electrical resistance, but also a high thermal conductivity of the interconnect materials is required. The higher thermal c onductivity c an h elp d issipate h eat mo re e fficiently from adhesive joints generated at high current. Therefore, a h igher thermal conductivity could contribute to an improved currentcarrying c apability. Wi th t he i ntroduction o f su itable o rganic monolayer t reatment, a n i mproved i nterface property b etween metal  llers a nd p olymer re sins c ould b e ac hieved, w hich enhanced the thermal conductivity of ACA joints. Studies show that organic monolayer enhanced ACA joints with best electrical properties a lso h ad a h igher t hermal c onductivity. I t i s a lso reported that with the addition of high thermal conductivity llers (e.g., SiC, AlN) into the ACA formulation, the higher thermal conductivity could be achieved [72,73], which also rendered a high current-carrying capability. Furthermore, carbon nanotubes, with h igh th ermal c onductivity ( >3000 W/k f rom th eoretical calculations) can also be aligned within the composite matrix to enhance the thermal conductivity.

43722_C011.indd 20

11.3.3.3

ACAs for High-Frequency Interconnections

Recently, high-frequency modeling and the characterization for ACA ip-chip interconnects have been performed to understand the h igh-frequency c haracteristics o f  ip-chip interconnects using ACFs, and thereby design higher performance ACF materials and bump system [74–76]. The effect of low dielectric  ller additions on high-frequency behavior o f A CF w as i nvestigated. The e xtracted i mpedance model pa rameters o f a 1 00 mm100 mm b onding pad w ere p resented for conventional ACF with the conductive ball only, and ACF w ith t he conductive ba ll a nd SiO 2  ller [77]. I n ACF  ipchip in terconnects a t hi gh fr equency, in terconnection ca pacitance formed is relatively high due to the high dielectric constant of the ACF resin and the large area, the small gap of the parallel metal p late s tructure, c ompared w ith t he s older ba ll  ip-chip structure. Therefore, t he re sonance f requency o f t he ACF  ipchip interconnect is lower than the solder ball ip-chip interconnects. I n add ition, t he b ump me tallurgy c an a lso a ffect the high-frequency behaviors of ACF interconnect. High-frequency p erformance o f s everal  ip-chip interconnects using ACFs at RF and high-frequency range were demonstrated and t he ACF  ip-chip assembly was proved as a si mple and cost effective method for high frequency devices. 11.3.3.4 Nanowire ACF for Ultrafi ne Pitch Flip-Chip Interconnection In o rder to s atisfy t he re duced I /O p itch a nd a void ele ctrical shorts, a possible solution is to use high-aspect-ratio metal post. Nanowires e xhibit h igh p ossibilities d ue to t he sm all si ze a nd extremely h igh a spect r atio. I n t he l iterature, na nowires h ave been shown to be applied to eld effective transistor (FET) sensor for ga s detection, magnetic hard-disk, na noelectrodes for electrochemical sensor, thermal-electric device for thermal dissipation, tem perature c ontrol, a mong ot hers [ 78–80]. T o p repare nanowires, it is important to dene nanostructures on photoresist. High-cost l ithographic methods such a s e-beam, X-ray, or scanning p robe l ithography h ave b een u sed b ut t he leng th o f nanowires c annot b e a chieved i n m icrometers. A nother le ss expensive alternative is electrodeposition of metal into nanoporous template such as a nodic a luminum oxide (AAO) [81] or a block-copolymer self-assembly template [82]. The disadvantages of a blo ck-copolymer tem plate i nclude t hin t hickness ( that

10/3/2008 9:35:12 PM

11-21

Films, Coatings, Adhesives, Polymers, and Thermoelectric Materials

means s hort na nowires), no nuniform d istribution, a nd p oor parallelism of nanopores. However, AAO has benets of higher thickness (>10 mm), u niform p ore si ze a nd den sity, l arger si ze and very parallel pores.

after TC test. However, pressure has to be large enough to ensure electrical conductivity and strong bonding.

11.3.3.5

Since thermal stress or inuence of humidity-induced cracks is the main degradation mechanism of ACF joints, the thermomechanical properties of ACF materials play a signicant role in the joint reliability during t he TC te st [85,86]. Generally, materials with low CTE should show improved reliability in temperature cycling [87]. This is because matched CTEs between the matching electrode pairs of the IC and the substrate (two contact partners) is necessary to minimize thermomechanical stresses on conductive joints within the polymeric layer (Figure 11.23). To enhance t he reliability of ACA  ip-chip assemblies, nonconductive  llers, such as silica  llers w ere adde d i n t he ACA materials with different  ller loadings [88]. It has been reported that the modulus of ACA materials increased and CTE below the Tg decreased as the content of silica increased. Consequently, the reliability of ACA  ip-chip a ssemblies u sing ACA w ith h igher content of nonconductive llers was signicantly improved than that of ip-chip assembly using conventional ACAs.

Reliability of ACA/ACF

Reliability is always one of the most important considerations to determine the value of materials. Failure behaviors of the interconnection a re c haracterized b y p erforming ac celerated a ging tests suc h a s h igh a nd lo w tem perature s torages, tem perature cycling, and humidity storage. Some main degradation mechanisms h ave b een f ound, i ncluding o xide g rowth o n le ss noble metallization causing increased contact resistance, stresses due to th e th ermal l oad o r i nuence of humidity inducing cracks within or delamination of the adhesive layer. 11.3.3.5.1 Effects of Bonding Conditions Bonding conditions of ACF, such as bonding temperature, temperature ra mping ra te, a nd p ressure a ffect t he c uring o f A CF materials a nd t he b onding qu ality o f A CF jo ints. I t h as b een reported t hat h igher b onding t emperatures re sulted i n h igher cross-linking den sity a nd s tronger ad hesion s trength o f A CF [83,84]. The increased adhesion strength with increasing bonding temperature is considered to be due to the interdiffusion or reaction between ACF polymeric material and the  ex. In addition, it h as b een found t hat ACF joints w ith h igher de grees of curing showed smaller increase in contact resistance after aging. At lower bonding temperatures a nd subsequently lower degree of c ure, t he c onductive pa rticles c an m igrate to t he p olymer layer. Therefore, the conductive particles could be squeezed out of the joints due to t he applied pressure through the soft adhesive. A s a re sult, i n s ome jo ints, h igher a nd u nstable c ontact resistance and even some open joints were found. At higher bonding temperature and higher degree of cure, the conductive particles remain throughout the ACF joints. The bonding pressure also affects the reliability of ACF joints signicantly. The lower the used pressure, the lower stress accumulated at t he interface and consequently the higher reliability

11.3.3.5.2 Effect of Mechanical Properties on the Reliability of ACF Joints

11.3.4

Nonconductive Adhesives and Films

ECA jo ints c an b e f ormed w ithout a ny c onductive  llers. The electrical co nnection o f N CAs is a chieved b y sea ling t he t wo contact pa rtners u nder pressure a nd heat. Thus, t he sm all gap contact is created, approaching the two surfaces to the distance of t he su rface a sperities, a s i llustrated i n F igure 1 1.16e. The formation of contact spots depends on the surface roughness of the c ontact pa rtners. App roaching t he t wo su rfaces ena bles a small number of contact spots to form, which allows the electric current to  ow. When the parts are pressed together during the sealing process, the number and area of the single contact spots are increased according to the macroscopic elasticity or exibility of the parts and the microhardness and plasticity of the surfaces, respectively.

(a)

(b)

FIGURE 11.23 (a) Deformation mechanism of the ACA joint due to thermally induced stress and (b) schematic diagram showing that thermal mismatch strain reduces, as solder bump height increases.

43722_C011.indd 21

10/3/2008 9:35:12 PM

11-22

11.3.4.1

Smart Materials

Electrical Properties of NCA

Conductive joints with NCA/NCF provide a number of advantages c ompared to ot her ad hesive b onding te chniques. N CA/ NCF joints avoid short-circuiting and are not l imited, in terms of particle size or percolation phenomena, to a reduction of connector pit ches. The si ze re duction o f ele ctronic de vices c an b e realized by t he shrinkage of t he package a nd t he chip. Fu rther advantages include cost-effectiveness, ease of processing regarding t he possibility of nonstructured ad hesive application, good compatibility w ith a w ide r ange of contact materials, a nd lowtemperature cure. In fact, the pitch size of the NCF joints can be limited only by the pitch pattern of the bond pad, rather than the adhesive materials. In spite of the advantages of NCF joints, there are still some challenging properties of NCF. The electrical conduction of NCF between the two surfaces of the IC bump and the substrate bond pad is achieved by bonding the two contact parts with pressure and heat. Thus, t he direct mechanical contact is formed by t he very  ne up-and-hill structure of t he bottom a nd top pad su rfaces with the roughness during bonding and a permanent joint is then formed by NCF resin curing or solidication. Since t he electrical conductivity of NCF is achieved through physical and mechanical contact and no metallurgical joints are formed, it has limited electrical conductance and current-carrying capability. L ow c ontact re sistance a nd h igh c urrent-carrying capability of NCF joints are demanding properties for lead-free solder a lternatives a nd h igh-current densi ty ap plication. T o ensure low contact resistance and high current density, interface between electrodes plays important role. The interfaces between electrodes for NCF joints s hould b e de fect f ree a nd o ccupy t he stable contact area even under high electrical current and harsh environments. This interface control in NCF joints contributes to performance a nd reliability of t heir joints in electronic packaging. In most NCF joints, electrical conductance between contacts is de pendent up on t he c onstriction re sistance a nd t unneling

Au-stud bump on the aluminum pad

resistance due to the presence of ultrathin insulating lm between contacts. The control of the tunneling resistance is important in reducing t he c ontact re sistance f or N CF jo ints [ 89–91]. S elfassembled p-conjugated mole cular w ires we re re cently d iscovered to functionalize the NCF joints and tailor their physical and chemical properties. Conjugated molecules tend to have a smaller band gaps b etween t he h ighest o ccupied mole cular o rbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) and possess delocalized conjugated p-electrons that can contribute to c onduction. In semiconductors, self-assembled molecular wires have s een i mportance i n t uning t he me tal work f unction (f) and electrical conduction of metal–molecule contact [92–96]. 11.3.4.2

NCAs with Improved Reliability for Flip-Chip Assembly

For the full implementation of ip-chip using NCAs, it is necessary to provide good reliability data to prove the availability of NCAs ip-chip te chnology. The m ost co mmonly o bserved  ip-chip failure h as o ccurred d uring t he TC te st, w hich i s d ue to t he thermal e xpansion m ismatch b etween c hips a nd sub strates. Therefore, t he p roblem o f C TE m ismatch b etween c hips a nd substrates b ecomes s erious w ith t he N CAs  ip-chip assembly because of high CTE of NCAs materials without any ller. Unlike ACAs with some conductive  llers, even small delamination or deformation of the NCA joint can easily lose the electrical joint. For this reason, novel NCAs that have low CTE for underll-like function has been developed. Figure 11.24 shows the schematic illustration of ip-chip CSP using NCF as rst level interconnection and its cross-section view of Au stud bump joint [97]. The add ition o f no nconducting si lica f iller o f t he N CA composite materials has control on the curing behavior, thermomechanical p roperties, a nd rel iability f or t he N CA  ip-chip assembly on an organic substrate. The content of nonconducting ller was optimized for the desirable thermomechanical properties of NCA composite materials such as proper curing prole, high

Underfilled NCF (nonconductive film)

FIGURE 11.24 Schematics of  ip-chip CSP using NCF and cross-section of NCF interconnection.

43722_C011.indd 22

10/3/2008 9:35:12 PM

Films, Coatings, Adhesives, Polymers, and Thermoelectric Materials

Tg, low CTE and modulus, and strong adhesion. These effects of nonconducting  ller add ition o n t he N CA m aterial p roperties were veried by reliability tests. The reliability of NCA ip-chip assembly by using a modied NCA with nonconducting llers is signicantly higher than that of a commercial ACF. Ther efore, the incorporation of nonconductive  llers in the NCA composite material signicantly improves the reliability of ip-chip CSP using NCAs materials. NCA materials continue to increase their applications with low cost, ner pitch interconnection, high reliability, and processability. But the technical concerns for process and p erformance of NCAs suc h a s h igh b onding pressure a nd electrical instability at high temperature should be resolved [98]. Recently, nanosized metallic llers were incorporated into NCA formulation f or i mproved ele ctrical a nd t hermal c onductivity and reduced bonding pressure [99].

11.3.5 Summary Polymer–metal-based E CAs h ave b een e volved to me et t he high el ectrical/mechanical/thermal per formance,  ne p itch capability, low-temperature process, and strong adhesion and reliability re quirements f or ele ctronic pac kaging mo dule and assemblies. As one of the most promising lead-free alternatives in electronic and optoelectronic device packaging interconnect ma terials, EC As ha ve sh own r emarkable adva ntages a nd attracted ma ny research interests. Signicant improvements in the ele ctrical, me chanical, a nd t hermal p roperties o f d ifferent types o f c onductive ad hesives h ave b een ac hieved. EC As w ith high performance and acceptable reliability have been developed for applications in die attach, ip-chip and surface-mount interconnect app lications. These le ad-free m aterials a nd p rocessing methods h ave g reat p otential to re place t he le ad-containing solders in the electronic industry.

References 1. H. Wolfson and G. Elliot, Electrically Conducting Cements Containing E poxy Resin s a nd S ilver, U .S. P atent, 2,774, 747, 1956. 2. K.R. Matz, Electrically Conductive Cement and Brush Shunt Containing the Same, U.S. Patent, 2,849,631, 1958. 3. D.P. Beck, Printed Electrical Resistors, U.S. Patent, 2,866,057, 1958. 4. Y. Li , K. M oon, a nd C.P . Wong, Science, 308, 2005, 1419–1420. 5. J.S. Hwang, (Ed.), Environment-Friendly Electronics: Lead-Free Technology, Electrochemical Publications Ltd., Port Erin, UK, 2001, Chapter 1, pp. 4–10. 6. Y. Li and C.P. Wong, Mater. Sci. Eng., R 51, 2006, 1–35. 7. J. L au, C.P. Wong, N.C. L ee, a nd S.W.R. L ee, In Electronics Manufacturing: with Lead-Free, Halogen-Free, and ConductiveAdhesive Materials, McGraw Hill, New York, 2002. 8. K. Gilleo, in: J.S. Hwang (Ed.), Environment-Friendly Electronics: Le ad-free T echnology, E lectrochemical Publica tions Ltd., Port Erin, UK, 2001, Chapter 24.

43722_C011.indd 23

11-23

9. R.L. Dietz et al., Soldering Surface Mount Tech., 9, 1997, 55. 10. J. Miragliotta, R.C. Benson, T.E. Phillips, and J.A. Emerson, in Proc. Mater. Res. Soc. Symp., 515, 1998, 245. 11. D. C avasin, K. B rice-Heams, a nd A. Arab, in P roc. 53r d Electronic C omponents a nd Technology C onf., 2003, p p. 1404–1407. 12. R. Kisiel, J. Electron. Packag., 124, 2002, 367. 13. H. de Vries, J. van Delft, and K. Slob, IEEE Trans. Compon. Packag. Technol., 28, 2005, 499. 14. D. L u, Q .K. T ong, a nd C.P. Wong, IEEE T rans. Co mpon. Packag. Manuf. Technol. Part C, 22, 1999, 223–227. 15. E.M. J ost a nd K. M cNeilly, P roceedings o f I nternational Society f or H ybrids a nd Micr oelectronics S ociety, 1987, pp. 548–553. 16. D. L u, Q .K. T ong, a nd C.P. Wong, IEEE T rans. Co mpon. Packag. Manuf. Technol. Part A, 22–3, 1999, 365–371. 17. Y. Li, K. Moon, and C.P. Wong, IEEE Trans. Compon. Packag. Technol., 29–1, 2006, 173–178. 18. Y. Li , K. M oon, A. Whitman, a nd C.P. Wong, IEEE Trans. Compon. Packag. Technol., 29–4, 2006, 758–763. 19. C. Gallagher, G. Matijasevic, and J.F. Maguire, Proceedings of 47th IEEE Ele ctronic C omponents a nd T echnology Conference, May 18–21, 1997, pp. 554–560. 20. J.W. Ro man a nd T .W. E agar, P roceedings o f the I nternational Society for Hybrids and Microelectronics Society (San Francisco, CA), Oct. 1992, p. 52. 21. H.J. Jiang, K. Moon, J. Lu, and C.P. Wong, J. Electron. Mater., 34, 2005, 1432–1439. 22. M. Z wolinski, J . H ickman, H. R ubon, a nd Y. Z aks, P roceedings o f the 2nd IEEE I nternational C onference o n Adhesive J oining & C oating T echnology in E lectronics Manufacturing, (Stockholm, Sweden), June 1996, p. 333. 23. D. L u, Q .K. T ong, a nd C.P. Wong, IEEE T rans. Co mpon. Packag. Manuf. Technol. Part C, 22(3), 1999, 228–232. 24. K. Persson, A. Nylund, J. Liu, and I. Olefjord, Proceedings of the 7t h E uropean C onference o n Applications o f S urface and Interface Analysis, Gothenburg, June 1997. 25. D. Lu and C.P. Wong, J. Appl. Polym. Sci., 74, 1999, 399–406. 26. Y. Li , K. M oon, a nd C.P. Wong, J. Adhesion S ci. Technol., 19–16, 2005, 1427–1444. 27. C. Cheng, G. Fredrickson, Y. Xiao, Q.K. Tong, and D. Lu, U.S. Patent No. 6,344,157, 2002. 28. G. T rabanelli a nd V. C arassiti, in: G. F ontana a nd R .W. Staehle (Eds.), Advanced Corrosion Science and Technology, Plenum Press, New York, 1970, Vol. 1, pp. 147–229. 29. G. Trabanelli, in: F. Mansfeld (Ed.), Corrosion Mechanisms, Marcel Dekker, New York, 1987, pp. 119–164. 30. O.L. Rig gs, J r., in: C.C. N athan (E d.), Ā eor etical Aspects of Corrosion Inhibitors and Inhibition, The National Association o f C orrosion En gineers (NACE), Houston, T X, 1973, pp. 2–27. 31. L.J. Matienzo, F.D. Egitto, and P.E. Logan, J. Mater. S ci., 38, 2003, 4831–4842. 32. H. Li, K. Moon, and C.P. Wong, J. Electron. Mater., 33, 2004, 106–113.

10/3/2008 9:35:13 PM

11-24

33. H. Takezawa, T. Mitani, T. Kitae, H. Sogo, S. Kobayashi, and Y. Bessho, Proceedings of 8th IEEE International Symposium on Advanced P ackaging M aterials: P rocesses, P roperties and Interfaces, Atlanta, GA, March 3–6, 2002, pp. 39–143. 34. K. Moon, S. Liong, H. Li, and C.P. Wong, J. Electron. Mater., 33–2, 2004, 1381–1388. 35. D. Dura nd, D. Vieau, A.L. Chu, a nd T.S. Weiu, U.S. Patent 5,180,523, November 1989. 36. A.N. Gent and G.R. Hamed, in: J.I. Kroschwitz, H.F. Mark, N.M. B ikales, C.G. O verberger, a nd G. M enges, (E ds.), Encyclopedia of Polymer Science and Technology, Wiley, New York, Vol. 1, 1985. 37. J.E. M orris a nd S. P robsthain, P roceedings. 4th IEEE International Conference on Adhesive Joining and Coating Technology in Electronics Manufacturing, June 8–21, 2000, pp. 41–45. 38. S. Liong, C.P. Wong, and W.F. Burgoyne, IEEE Trans. Compon. Packag. Technol., 28–2, 2005, 327–336. 39. K. Moon, C. Rocket, and C.P. Wong, J. Adhesion Sci. Technol., 18–2, 2004, 153–167. 40. A. N agai, K. Takemura, K. I saka, O. Watanabe, K. K ojima, K. Matsuda, and I. Watanabe, 2nd IEMT/IMC Symposium, 15–17, April 1998, pp. 353–357. 41. I. Watanabe, T. Fujinawa, M. Arifuku, M. Fujii, and Y. Gotoh, Proceedings o f 9th IEEE I nternational S ymposium o n Advanced P ackaging M aterials: P rocesses, P roperties a nd Interfaces, Atlanta, GA, March 24–26, 2004, pp. 11–16. 42. S. M acathy, P roceedings o f S urface M ount I nternational, San Jose, CA, August 27–31, 1995, pp. 562–567. 43. S.A. Vona and Q.K. Tong, Proceedings of 4th International Symposium an d E xhibition on Advanced P ackaging Materials, Pr ocesses, Pr operties a nd I nterfaces, B raselton, GA, March 14–16, 1998, pp. 261–267. 44. J. Liu and B. Weman, Proceedings of the 2nd I nternational Symposium o n Ele ctronics P ackaging T echnology, December 9–12, 1996, pp. 313–319. 45. D. L u a nd C.P . Wong, P roceedings—IEEE I nternational Symposium o n Advanced P ackaging M aterials: P rocesses, Properties and Interfaces, Braselton, GA, March 6–8, 2000, pp. 24–31. 46. H. Li, K. Moon, Y. Li, L. Fan, J. Xu, and C.P. Wong, Proceedings of 54th IEEE Ele ctronic C omponents a nd T echnology Conference, Las Vegas, Nevada, June 1–4, 2004, pp. 165–169. 47. J .H. Zha ng, a nd Y.C. Cha n, J. Elec tron. M ater., 32, 2003, 228–234. 48. W.-S. Kwon, M.-J. Yim, K.-W. Paik, S.-J. Ham, and S.-B. Lee, J. Electron. Packag., 127, 2005, 86–90. 49. M.J. Rizvi, Y.C. Chan, C. Bailey, H. Lu, and A. Sharif, Soldering Surface Mount Technol., 17, 2005, 40–48. 50. G. D avies a nd J . Sa ndstrom, Cir cuits M anufacturing, October 1976, pp. 56–62. 51. G. Harsanyi and G. Ripka, Electrocomp. Sci. Tech., 11, 1985, 281–290. 52. G.D. G iacomo, in: J . M cHardy a nd F . L udwig (E ds.), Electrochemistry of Semiconductors and Electronics: Processes

43722_C011.indd 24

Smart Materials

53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77.

and D evices, N oyes Pub lications, P ark Ridg e, NJ , 1992, Chapter 6, pp. 255–295. R. Manepalli, F. Stepniak, S.A. Bidstrup-Allen, and P.A. Kohl, IEEE Trans. Adv. Packag., 22, 1999, 4–8. G.D. Giacomo, Reliability of Electronic Packages and Semiconductor Devices, McGraw-Hill, New York, 1997, Chapter 9. R. Wassink, Hybrid Circ. no. 13, May 1987, pp. 9–13. A. D er M arderosian, R atheon C o. E quipment Di vision, Equipment Development Laboratories, pp. 134–141. H. Schonhorn and L.H. Sharpe, Prevention of Surface Mass Migration b y a P olymeric S urface C oating. U .S. P atent 4,377,619, 1983. V. Brusic, G.S. Frankel, J. Roldan, and R. Saraf, J. Electrochem. Soc., 142, 1995, 2591–2594. P.-I. Wang, T .-M. L u, S.P. M urarka, a nd R . Ghoshal , U .S. Patent No. 2,005,023,6711, 2005. Y. Li and C.P. Wong, Appl. Phys. Lett., 89, 2006, 112112. I. Watanabe et al ., In Proc. Asia Display/IDW, pp. 553–556, 2001. H. Nishida, K. Sakamoto, and H. Ogawa, IBM J. Res. Dev., 42, 1998, 517. D.J. Williams et al., Soldering Surface Mount Tech., 1993, 4. J. Li u, A. Tolvgard, J . M almodin, a nd Z. L ai, IEEE T rans. Compon. Packag., Manuf. Technol., 22, 1999, 186. P. Clot, J.F. Zeberli, J.M. Chenuz, F. Ferrando, and D. Styblo, in P roc. Ele ctronics M anufac. Technol. S ymp., 24th IEEE/ CPMT, 36, 1999. Y. Li, K. Moon, and C.P. Wong, J. Electron. Mater., 34–3, 2005, 266–271. Y. Li , K. M oon, a nd C.P. Wong, J. Elec tron. M ater., 34–12, 2005, 1573–1578. K. Moon, H. Dong, R. Maric, S. Pothukuchi, A. Hunt, Y. Li, and C.P. Wong, J. Electron. Mater., 34, 2005, 132–139. Y. Matsuba, Erekutoronikusu J isso G akkaishi, 6(2), 2003, 130–135. Y. Li, K. Moon, and C.P. Wong, J. Appl. Polym. Sci., 99, 2006, 1665. A. Sihlbom, R. Sihlbom, and J. Liu, Proceedings of IEEE 2nd Electronic P ackaging T echnology C onference, D ecember 8–10, 1998, Singapore, pp. 251–257. M.J. Yim, H.-J. Kim, and K.-W. Paik, J. Electron. Mater., 34, 2005, 1165–1171. L. Ekstra nd, H. K rstiansen, a nd J. Li u, P roceedings o f the IEEE 28th I nt. Spring Seminar on Electronics Technology, Wiener Neustadt, Austria, May 19–20, 2005, pp. 35–39. M.J. Yim, W.H. Ryu, Y.D. Jeon, J.H. Lee, S.Y. Ahn, J.H. Kim, and K.W . P aik, IEEE T rans. Co mpon. P ackag., M anuf. Technol., 22, 1999, 575. R. S ihlbom, et al ., IEEE T rans. Co mpon. P ackag. M anuf. Technol., 21, 1998, 469. G. Dou, Y.C. Chan, J.E. Morris, and D.C. Whalley, Soldering Surf. Mount Technol., 18, 2006, 3. M.J. Yim, I.H. Jeong, H.K. Choi, J.S. Hwang, J.Y. Ahn, W.S. Kwon, a nd K .W. P aik, IEEE Trans. Comp. Packag. Technol., 28, 2005, 789.

10/3/2008 9:35:13 PM

11-25

Films, Coatings, Adhesives, Polymers, and Thermoelectric Materials

78. C.M. Lieber, Science, 293, 2001, 1289–1292. 79. G.A. Prinz, Science, 282, 1998, 1660. 80. C.R. M artin a nd V.P. M enon, Anal. C hem., 67, 1995, 1920–1928. 81. J.M. Xu, Appl. Phys. Lett., 79, 2001, 1039–1041. 82. T.P. Russell, Science, 290, 2000, 2126–2129. 83. Y.W. Chiu, Y.C. Chan, and S.M. Lui, Microelectron. Reliab., 42, 2002, 1945–1951. 84. L. Cao, Z. Lai, and J. Liu, Proceedings of the IEEE 6th CPMT Conference o n H igh D ensity Micr osystem D esign a nd Packaging a nd C omponent F ailure Analysis (HD P’04), Shanghai, China, June 30–July 3, 2004, pp. 254–258. 85. L. C ao, Z. L ai, a nd J . Li u, J. Elec tron. P ackag., 127, 2005, 43–46. 86. R.A. Islam, Y.C. Chan, and B. Ralph, J. Mater. Res., 19, 2004, 1662–1668. 87. A. S chubert, P roceedings o f IEEE 4th I nternational Symposium o n Advanced P ackaging M aterials, B raselton, GA, March 15–18, 1998, pp. 153–160. 88. M.-J. Yim a nd K.-W. P aik, IEEE T rans. Co mpon. P ackag. Technol., 24–1, 2001, 24–32. 89. R. Holm, In Electrical Contacts, Springer, New York, 1967. 90. M. Chin, K.A. Iyer, and S.J. Hu, IEEE Trans. Compon. Packag. Technol., 27(2), 2004, 317–326. 91. L. K ogut a nd K. K omvopoulos, J. Ap pl. Phys., 95(2), 2004, 576–585. 92. B. de Boer, A. Hadipour, M.M. Mandoc, T. van Woudenbergh, and P.W.M. Blom, Adv. Mater., 17, 2005, 621–625. 93. T. D adosh, Y. G ordin, R . K rahne, I . K hivrich, D . M ahalu, V. Frydman, J. Sperling, A. Yacoby, and I. Bar-Joseph, Nature, 436, 2005, 677–680. 94. Q. Sun, A. Selloni, and G. Scoles, J. Phys. Chem. B, 110, 2006, 3493–3498. 95. G. Heimel, L. Romaner, J.-L. Brédas, and E. Zojer, Phys. Rev. Lett., 96, 2006, 196806. 96. Y. Li, M.J. Yim, and C.P. Wong, J. Electron. Mater., 36–5, 2007, 549–554. 97. M.J. Yim, J.S. Hwang, W.S. Kwon, K.W. Jang, and K.W. Paik, IEEE Trans. Electr. Packg. Manuf., 26, 2003, 150. 98. H.C. Cheng, C.L. Ho, K.N. Chiang, and S.M. Chang, IEEE Trans. Compon. Packag. Tech., 27, 2004, 398. 99. Y. Li , M.J . Yim, K. M oon, a nd C.P. Wong, P roceedings o f 57th IEEE Electronic Components and Technology Conference, Reno, Nevada, May 29–June 1, 2007, pp. 1350–1355.

11.4

Electrochromic Sol–Gel Coatings

L.C. Klein 11.4.1

Defi nition of Electrochromism

Certain i norganic c ompounds, e specially o xides o f p olyvalent metals, exhibit coloration that depends on the oxidation state of their cations. This property leads to electrochromism, which is a

43722_C011.indd 25

reversible a nd v isible c hange i n t ransmittance o r re ectance. The oxidation–reduction reactions are electrochemically induced, using low voltages, on the order of ±1 V dc. An electrochromic (EC) device is a multilayer construction for which one of the layers shows EC properties (Figure 11.25) [1–5]. An EC device operates on the principle of a galvanic cell. The best known EC material is tungsten trioxide (WO3), which forms deep blue alkali and hydrogen tungsten bronzes (M xWO3) on re duction. The re action i s e xpressed b y t he f ollowing equation: 6+

+

6

+

+

WO3 + xM + xe′ = M x WO3 = M x Wx W1− x O3 (1 (Clear) (Blue)

1.7)

where M i s hydrogen or a lkali, e′ is a n electron, a nd 0 < x < 1 . Typically, a low-voltage EC device with an EC cathode is colored in the charged state and bleached in the discharged state. A m ajor app lication o f EC de vices i s s o-called “ smart w indows.” These ele ctronically c ontrolled w indows, w hich c an lighten or darken depending on charge insertion/extraction, are designed to adjust to the amount of sunlight, the time of the day or t he s eason o f t he ye ar. C ontrol o f t ransmittance t hrough windows is a factor in the energy usage in a building, as well as the creation of a pleasant environment in interior spaces. Using them on buildings reduces the heating lost through architectural windows, if they are colored on bright, summer days and bleached on cloudy, winter days. EC windows that fulll these expectations have been demonstrated using thin lm technologies other than sol–gel processing [6,7]. Substitution of a sol–gel layer for any one of the layers—the EC electrode, the counter electrode, or the electrolyte—is a breakthrough in making affordable EC windows [8–10].

11.4.2

Principle of the Galvanic Cell

Recall for a moment, the basic galvanic cell. The cell consists of an anode, a c athode, and an e lectrolyte. I n an E C d evice, the layers are referred to as electrode, counterelectrode, and electrolyte. The performance of an EC device is shown schematically in Figure 11.26. Assuming that the EC device is operating with a lithium ion (M = Li+), on charging the Li is oxidized at the cathode (electrode) Li → Li + + e′ (1

1.8)

and reduced at the anode (counterelectrode) Li + + e′ → Li (1

1.9)

In t he case of a re versible cell, both t he anode and cathode are intercalation compounds with a layered or framework structure. In addition to tungsten oxides, other common electrode materials

10/3/2008 9:35:13 PM

11-26

Smart Materials

Counter electrode

Solid electrolyte

Active electrochromic layer - electrode

Interior glass with ITO

Exterior glass with ITO

Typical window section

FIGURE 11.25 Schematic of an EC device, substrate (ITO-coated commercial  oat glass), electrode (tungsten oxide), solid electrolyte (lithium containing inorganic gel), counterelectrode (vanadium oxide), and substrate. Lithium reduced

Lithium reduced

Lithium oxidized Li+

Li+ e−

e−

Li+ Li+ e−

Li+ Li+

Li+ eLi+

e−

Li+ Li+

(−)

e−

(+) (+) +3 V

e−

Bleached state sunlight transmitted

(−) e−

−1.5 V Colored state sunlight

FIGURE 11.26 Schematic operation of t he EC device, comparing t he colored state (charged) a nd bleached state (discharged). The open circle indicates a lithium, where the oxidation/reduction reaction takes place, according to Equations 11.8 and 11.9.

are vanadium oxides, niobium oxides, and titania-doped cerium oxides [11–13]. The requirements for the electrodes are: • • • • •

43722_C011.indd 26

High coloration efficiency Fast switching between colored and bleached states High diffusivity for lithium High capacity for lithium Compatibility with electrolyte

A critical feature in a galvanic cell is the electrolyte, which has to transport ions back a nd forth between t he cathode a nd a node. The electrolyte has to have • Wide electrochemical potential window • Stability d uring c harge/overcharge a nd d ischarge/overdischarge

10/3/2008 9:35:14 PM

11-27

Films, Coatings, Adhesives, Polymers, and Thermoelectric Materials

• • • •

Deposits uniform thin  lms on large areas Allows the coating of any size or shape Allows control of lm microstructure Delivers m ulticomponent  lms w ithout c hanging t he deposition equipment • Provides exibility in manufacturing

Their low cost Their aqueous nature Their ability to be recycled The presence of W–O–W bonds in the sols Their amorphous nature and nondirectional morphology

A drawback of tungstic acid sols is that they gel in 30 min. A s olution of a queous s odium t ungstate, Na 2WO4 .2H2O, i s made by mixing tungstic acid powder, sodium hydroxide, and water. The s olution i s p oured t hrough a p roton e xchange re sin, which is conditioned before use by washing with 2 N HCl. Excess acid is removed from the column by washing with distilled water. The sol is collected in a beaker containing a dened amount of solvent. The sol, which is clear yel low, contains polymeric species. The mechanism of polymerization is approximately:

O

O

O

O

O—W—OH + HO—W—O – –> O—W— O—W—O + H2O

(1

—

43722_C011.indd 27

• • • • •

—

There are three basic approaches to the preparation of transition metal oxides [19–21]. One of t he routes is to p repare a s olution using alkoxides, metal organic compounds with alkoxy groups. The a vailable a lkoxides a re o nly s lightly s oluble i n c ommon

While there are several ways to ob tain WO3 thin lms [11], the least expensive precursor is tungstic acid. The reasons for choosing tungstic acid sols are:

—

• Exhibit signicant spectral switching over the visible and infrared regions of the solar spectrum. • Have a cceptable du rability, b oth c hemical a nd ph ysical, over a long time of exposure (∼30 year lifetime of architectural windows [18]).

Tungsten Oxide Gels for the Active Electrochromic Layer

—

Another important consequence of the sol–gel process is the fact that the lm is less than 1 µm thick. While the  lm is rigid, it is thin enough to re sist cracking, so t hat d imensional changes in the electrodes can be accommodated even when the electrolyte is in place. Sol–gel t hin  lms u sed i n EC de vices h ave e xceptional requirements:

11.4.3.1

—

The denition of a gel is a material having solid-like properties, meaning a gel h as  xed s hape. O n t he o ne h and, t he o rganic “gels,” which are frequently used in batteries, are rigid, but only in the sense that they “set” when shear stress is absent. Inorganic gels, on the other hand, that result from the sol–gel process are irreversible gels (see Ref. [17] for general background). The se gels are based on a n oxide network. Not only a re t hese r igid structures, they are inert in contact with organics and solvents. A nat ural adv antage of t he s ol–gel process i s t he f act t hat a  lm of oxide can be placed on a substrate directly from solution. Other me thods o f de positing  lms o f o xides a re s puttering o r chemical vapor deposition (CVD). These methods may succeed in some cases, but the sol–gel process is simpler than either one. Dipping a nd s pinning a re c ommon p rocesses f or p reparing sol–gel thin  lms. Among the available techniques, dip-coating is more widely used. Its advantages include the following:

—

Sol–Gel Processing

—

11.4.3

organic solvents and tend to be very reactive with water. The second route is to u se salts such as oxychlorides as the network former oxide precursor. The third route is to use colloidal sols. A sol, by denition, is a colloidal suspension of small particles (1–500 nm) in a liquid. Oxide sols are obtained after hydrolysis– condensation re actions o f t he p recursors. H ydrolysis o ccurs when me tal c ations ( M) a re s olvated b y w ater mole cules. Condensation occurs as soon as hydroxyl groups are present in the c oordination s phere of M . These re actions le ad to M -O-M bonds. When the number of links is high enough, the sol becomes a gel . The gel p oint occurs when t he solid phase becomes continuous. After t his p oint, t he gel c ontinues to a ge, p rimarily causing changes in structure. Aging is followed by drying, when evaporation of the liquid can cause further shrinkage of the gel structure.

—

Most commercial electrolytes are polymers and “gel” electrolytes [14]. Some problems with commercial electrolytes are that aqueous ele ctrolytes m ay gener ate H 2 a nd no naqueous ele ctrolytes have conductivities that are too low. Solid ele ctrolytes a re at tractive b ecause t here a re e ffectively no mov ing p arts, me aning no mobi le proton s. W hile a s olid electrolyte i s f unctioning a s a s eparator a nd memb rane, i t i s mechanically r igid, h olding t he c omponents o f t he EC de vice x ed. The p roblem i s t hat mo st s olid ele ctrolytes do not h ave high eno ugh io nic c onductivity at a mbient tem perature [ 15]. One way around this problem is thin lm technology [16], which is, in part, the motivation for using sol–gel processing.

O

O

O

O

1.10)

As-prepared W O3  lms, w hich a re p oorly c rystallized, h ave a higher coloring efficiency than crystallized WO3. Crystallization of t he W O3 th in  lms is ex pected t o occ ur a fter a h eating at 400°C, so lms should not c rystallize when heat treated for 1 h only at 90°C. Comparisons between uncoated glass and coated glass showed that t he t ransmittance i n t he v isible r ange i s lo wer b y a bout 10%–15% for coated samples. Nevertheless, the transmittance of the coated glass is acceptable for most applications. The WO3 sol is de posited o n co mmercial  oat g lass, w hich i s c oated w ith indium tin oxide (ITO).

10/3/2008 9:35:15 PM

11-28

Smart Materials

11.4.3.2

Vanadium Oxide Gels for the Counter Electrode

Again, there are several ways to obtain V2O5 thin lms [22–24]. These include the nitric acid dissolution of vanadate salts, proton exchange, h ydrolysis o f v anadium a lkoxides, a nd h ydration o f amorphous v anadia. The le ast e xpensive p recursor i s v anadia powder that is melted at 1000°C (Tmelt = 6 90°C) a nd quenched into water to produce amorphous vanadia. Th is glassy material is soluble in water and yields a golden brown, colloidal sol that is suitable for dip coating. The behavior of the vanadia sols parallels, in many respects, the tungstic acid sols, so that thin, transparent coatings can be produced. 11.4.3.3

Solid Electrolyte for Lithium Ion Conduction

Sol–gel processing has been used for more than 15 years to prepare s olutions c ontaining o xide c omponents de sirable f or f ast ion conductors (FIC) [15]. Fast ion-conducting gels and glasses are desirable because they • reA isotropic. • Have few restrictions on composition. • Have more channels for fast ion conduction than crystalline compounds. Ionic transport in solids is an activated process. In most FIC, the activation e nergy i s in dicative o f t he e nergy o f m otion. The desired features in a FIC are • • • • •

Large number of mobile ions Large number of sites for the ions Relatively low activation energy Wide compositional range Ease of fabrication in complex shapes and thin lms

Classically, high ionic conductivity requires high alkali concentration. T he s ystem i nvestigated o riginally f or t he s olid electrolyte w as l ithium si licates. I n t he L i 2O-SiO2 ge ls, t he second co mponent wa s i ntroduced t ypically a s l ithium nitrate. T he n itrate s alt d issolves i n w ater a nd t hen c rystallizes in the dried gel as nitrate nanocrystals [25]. Alternatively, lithium h ydroxide w as u sed to c reate v ery h igh p H c onditions, increasing the chance of a reaction between LiOH and the si lica ne twork to g ive Si-O-Li b onds [26]. T he a lkoxidebased gel s, o r s o-called p olymerized gel s, s howed v alues o f ionic conductivity that are stable and in the range needed for operation of a l ithium battery at a round 250°C. The lithium nitrate gels had a characteristic behavior described by a composite mechanism [27]. Thermodynamic calculations were performed in order to  nd out w hich o xides, suc h a s a lumina, z irconia, o r t itania, c ould improve the stability of the gel. Using the standard heats of formation, it is easy to see that no oxide addition makes lithium silicates stable compared to lithium metal. At best, the addition improves the k inetic s tability at lo w temperatures. A lumina i s one c andidate for increasing the kinetic stability of the lithium silicate gels.

43722_C011.indd 28

Ionic conductivity measurements were carried out by t he ac complex i mpedance me thod u sing a S olartron 1250 f requency response a nalyzer a nd 1 186 ele ctrochemical i nterface, w hich were programmed by a Hewlett-Packard 9816 desktop computer for data collection and analysis. Experiments were performed on pressed pellets w ith both sides coated w ith a l ayer of platinum paste to provide contacts with platinum electrodes. Ionic c onductivity me asurements w ere p erformed o n gel s heated f rom 100°C to a round 3 60°C at a h eating r ate of 10°C/ min i n t he f requency r ange of 10 Hz to 65 kHz. Measurements were re peated a nd re produced to el iminate a ny c oncern t hat there was a contribution to t he measurements from protons. In bulk materials, the conductivities are too low at room temperature for use in an EC device. In t he cas e of measurements on t hin  lms, t he su rface o f a microelectrode ( Microsensor S ystems SAW 3 02), c onsisting o f an interdigitated array of gold ele ctrodes deposited on a qu artz crystal sub strate, w as c oated w ith t he s olution [28]. A s ample containing 10 mol% A l2O3 w as s elected b ecause me asurements on t he b ulk s howed go od io nic c onductivity f or t his c omposition. The values of ionic conductivity measured on thin lms matched those for bulk ionic conductivity. While stable and reproducible conductivities have been demonstrated in t hin  lms, t he v alues of c onductivity a re s till to o low. Further advances in EC devices are dependent on nding a solution to t he ele ctrolyte p roblem. S ome re cent s tudies o n organically modied electrolytes are leading in the right direction [29,30].

11.4.4

Conclusions

In further development of all sol–gel EC devices, physical properties need to be considered, in addition to the ionic conductivity. Of course, the electrolyte has to have • High electronic resistivity • High ionic conductivity • o N defects In addition, all of the layers require • Film t hickness le ss t han a bout 1 000 nm ( generally le ss than 500 nm) • High transparency over the visible range • Good adhesion between adjacent layers • Thermodynamic stability with adjacent layers over a wide electrochemical window Overall, a n EC de vice s hould op erate u nder t he app roximate conditions of +2.5 V (100 s)/−1.5 V (100 s) for more t han 10,000 cycles at 15 mC/cm2. For practical EC w indows, t here h as b een sig nicant recent progress using a variety of technologies [31,32]. Sol–gel processing, of course, is only one route under investigation. Ther e continues to be interest in sol–gel processing because precursors for all of the layers are readily available. The resulting gels are inorganic a nd r igid. I n pa rticular, t he s ol–gel ele ctrolyte op erates

10/3/2008 9:35:15 PM

Films, Coatings, Adhesives, Polymers, and Thermoelectric Materials

while p laying t he role o f a me chanical d ivider b etween t he electrodes. F inally, s ol–gel p rocessing i s a si mple p rocess f or large area coatings.

Acknowledgments This work in the past was supported by the Assistant Secretary for C onservation a nd Rene wable E nergy, O ffice of Transportation Technologies, E lectric a nd H ybrid P ropulsion D ivision of the US Department of Energy under Contract No. DE-AC03-76SF00098, Sub contract N o. 4 593610 w ith t he Lawrence Berkeley Laboratory. My thanks go to G. Amatucci, L. Laby, G. Lous, A. Wojcik, M. Greenblatt, A. Salkind, and J. Van Dine for many useful discussions.

References 1. C. G. Granqvist, Journal of the European Ceramic Society, 25, 2005, 2907. 2. F. G. K. B aucke, Materials S cience a nd E ngineering, B10, 1991, 285. 3. C. B. Greenberg, Ā in Solid Films, 251, 1994, 81. 4. C. B . G reenberg, Journal of Elec trochemical S ociety, 140, 1993, 11. 5. I. D . Watkins, G. E. T ulloch, T . M aine, L. S piccia, D . R . McFarlane, a nd J . L. Woolfrey, in Sol–Gel Pr ocessing o f Advanced Materials, L. C. Klein, E. J. A. Pope, S. Sakka, and J. L.Woolfrey (eds.), Ceramic Transactions,Vol. 81:American Ceramic Society, Westerville, OH, 1998, p. 223. 6. J. -G. Zhang, C. E. Tracy, D. K. Benson, and S. K. Deb, Journal of Materials Research 8, 1993, 2649. 7. J. P. Cronin, D. J. Tarico, J. C. L. Tonazzi, A. Agrawal, and S. R. Kennedy, Solar Energy Materials and Solar Cells, 29, 1993, 371. 8. J. E. Van Dine, V. D. Parkhe, L. C. Klein, and F. A. Trumbore, U.S. Patent 5,404,244, April 4, 1995. 9. J. E. Van Dine, V. D. Parkhe, L. C. Klein, and F. A. Trumbore, U.S. Patent 5,659,417, August 19, 1997. 10. J. E. Van Dine, V. D. Parkhe, L. C. Klein, and F. A. Trumbore, U.S. Patent 5,699,192, December 16, 1997. 11. J. Livage, in Solid State Ionics III, G. -A. Nazri, J. M. Tarascon, and M. Armand (eds.), MRS Symposium Proceedings, Vol. 293, Pittsburgh, PA, 1993, p. 261. 12. A. Pawlicka, C. Avellaneda, M. S chmitt, S. Heusing, a nd M. A. Aegerter, in Sol–Gel Processing of Advanced Materials,

43722_C011.indd 29

13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32.

11-29

L. C. K lein, E. J . A. P ope, S. Sak ka, J . L. Woolfrey (e ds.), Ceramic Transactions, Vol. 81: American C eramic S ociety, Westerville, OH, 1998, p. 229. B. Munro, P. C onrad, S. Kra mer, H. S chmidt, a nd P. Z app, Solar Energy Materials and Solar Cells, 54, 1998, 131. C. Marcel and J. -M. Tarascon, Solid State Ionics, 143, 2001, 89. J. P. Boilot and Ph. Colomban, in Sol–Gel Technology for Āin Films, Fib ers, P reforms, E lectronics an d Sp ecialty S hapes, L. C. Klein (ed.), Noyes Publication, NJ, 1988, p. 303. B. Wang, M. Greenblatt, J. Yan, and Y. Wu, Journal of Sol–Gel Science and Technology, Park Ridge, NJ, 2, 1994, 323. C. J. Brinker and G.W. Scherer, Sol–Gel Science: Ā e Physics and Ch emistry o f S ol–Gel Pr ocessing, A cademic Press, Boston, MA, 1990. C. E. Tracy, J.-G. Zhang, D. K. Benson, A. W. Czanderna, and S. K. Deb, Electrochimica Acta 44, 1999, 3195. A. Chemseddine, M. Henry, and J. Livage, Revue de Chimie Minerale 21, 1984, 487. B. Munro, S. Kramer, P. Zapp, and H. Krug, J. Sol-Gel Science and Technology, 13, 1998, 673. B. Pecquenard, H. Lecacheux, S. Castro-Garcia, and J. Livage, J. Sol–Gel Science and Technology 13, 1998, 923. J. Livage and J. Lemerle, Ann. Rev. Mat. Sci., 12, 1982, 103. S. Mege, M. Verelst. P. Lecante, E. Perez, F. Amsart, and J. M. Savariault, Journal of Noncrystalline Solids, 238, 1998, 37. R. D. Cussler, M. G. Kulkarni, and E. L. Cussler, Journal of Membrane Science, 127, 1997, 153. S. -P . Szu , L. C. K lein, a nd M. G reenblatt, Journal o f Noncrystalline Solids, 121, 1990, 119. S. -P . Szu , M. G reenblatt, a nd L. C. K lein, Journal o f Noncrystalline Solids, 124, 1990, 91. S. -P. Szu, M. Greenblatt, and L. C. Klein, Solid State Ionics, 46, 1991, 291. L. L aby, L. C. K lein, J. Yan, a nd M. G reenblatt, Solid S tate Ionics, 81, 1995, 217. K. D ahmouche, M. Atik, N. C. Mello, T. J. B onagamba, H. Panepucci, M. A. Aegerter, a nd P. J udeinstein, Journal o f Sol–Gel Science and Technology, 8, 1997, 711. P. J udeinstein, J . T itman, M. S tamm, a nd H. S chmidt, Chemical Materials, 6, 1994, 127. C. G. Granqvist, Electrochimica Acta, 44, 1999, 3005. C. G. G ranqvist, A. Azens, A. H jelm, L. K ullman, G. A. Niklasson, D. Ronnow, M. Stromme Mattsson, M. Veszelei, and G. Vaivars, Solar Energy, 63, 1998, 199.

10/3/2008 9:35:16 PM

43722_C011.indd 30

10/3/2008 9:35:16 PM

12 Cure and Health Monitoring Tatsuro Kosaka Osaka City University

12.1 I ntroduction ...........................................................................................................................12-1 12.2 C ure Monitoring ....................................................................................................................12-1 12.3 H ealth Monitoring .................................................................................................................12-3 Bibliography .......................................................................................................................................12-5

12.1 Introduction Because safety is the most important keyword, companies have made enormous efforts to a ssure that the quality of their products when shipped and when in operation perform frequent and accurate i nspections. The i nspection me thods h ave b een improved to b ecome more a ccurate, qu icker, a nd more l abor saving therefore reducing cost and time. However, even the most efficient inspection method requires downtime of the machines and cannot insure the safety when an emergent accident occurs in service. In order to a void t hese problems, successive inspection methods have been desired. Cure and he alth monitoring is a ne w technology to mo nitor materials and structures from birth to death successively. During the ma nufacturing p rocess o f p olymer p roducts, the c ure monitoring is applied to observe processes of the c ure reaction, temperature, pressure, residual stress, homogeneity, and integrity over the materials. After the shipment, built-in sensors measure mechanical and thermal responses such as strain, stress, and temperature in real time and report the information to remote operating cen ters. The ma in diff erence b etween the t raditional inspection method and the cure and health monitoring method is the us e of built-in s ensors and remote monitoring technologies. Recent developments of sensor technologies has enabled us to in stall minia ture s ensors in to ma terials a nd str uctures f or observing the internal conditions in real time, and with the evolution of the I nternet has made i t practical to collect data from remote sensing systems at high speed. This paper introduces the cure and health monitoring method from the viewpoint of builtin sensors.

12.2 Cure Monitoring The c uring p rocess i s o ne o f t he mo st i mportant s teps i n t he manufacturing of thermosetting polymers because it signicantly controls the properties of the materials. Monitoring of the cure

process is essential in order to  nd the best conditions for temperature and pressure. During the curing process, the material changes f rom l iquid to s olid s tate b y a c hemical re action, a nd therefore, any property such as density, elasticity, viscosity, glass temperature, c rystallinity, a nd d ielectricity d ramatically changes. Many cure monitoring methods have been proposed but the most popular methods are thermal analyses by differential scanning calorimeters (DSC), viscoelastic analyses by thermomechanical analyzers (TMA), and chemical analyses of the molecular structures by in frared spectrometers. The degree of cure (DOC), which is dened by a r atio between reaction heats and total exothermic heats, is the most important index to evaluate during the curing process, and thus the rst purpose of cure monitoring is to obtain DOC. Although t hese traditional methods are useful and sophisticated, they are available only in the laboratory due to the requirement of small samples. On the other hand, the cure monitoring of practical products has become important with the increasing demand of large and complex-shaped polymer matrix composites (PMC) because of the local concentrations of the incomplete cure, t he re sidual s tress a nd voids, w hich s ometimes o ccur. I n order to mo nitor t he i nternal s tate o f t he p ractical p roducts, miniature and embeddable sensors are necessary. Table 12.1 lists the sensors available for cure monitoring with measurable values and usability for the cure monitoring. Optical ber sensors have advantages due to t heir small size, exibility, lo cality, a nd h igh-temperature re sistance f or c ure monitoring. In addition, many kinds of physical values such as temperature, p ressure, s train, re fractive i ndex, a nd mole cular structure can be monitored. The optical  bers used in the spectrometers a nd re ectometers re quire e xposed su rfaces of t heir core in order to transmit light through materials. Three types of sensor structures are the transmission type, reection type, and evanescent type and have been proposed for use as illustrated in Figure 12.1. The evanescent type sensor utilizes evanescent light leaked from the core of the stripped optical ber. Fourier transform 12-1

43722_C012.indd 1

10/3/2008 8:48:42 PM

12-2

Smart Materials

TABLE 12.1 Embeddable Sensors for Cure Monitoring Measurable Values

Use in Operation

Optical ber sensors Spectrometers Reectometers Strain/temperature sensors Dielectric sensors

Molecular structures Refractive index Strain, temperature Permittivity, loss factor

Poor Poor Good Poor

Piezoelectric sensors Impedance measurement Ultrasonic measurement

Impedance Sound velocity

Poor Good

infrared spectroscopy (FTIR) and UV uorimetry are available as ber optic spectrometers. The FTIR measures the absorption spectrum of the infrared light whose bands identify the specic molecular structures. Since the specic molecular groups appear or disappear by the advance of the cure reaction, the DOC can be ob tained f rom c hanges i n t he m agnitude o f t he s pecic absorption ba nd. O n t he ot her h and, t he U V  uorimetry i s a technique ba sed o n t he ph otoluminescent ph enomenon w here the incident light irradiates t he  uorescent material. The emission light from the  uorescent curing agent can be used for the cure monitoring of polymers. It has been reported that the peak of t he e xcitatory s pectra o f e poxy re sin s hifts d uring c ure. Although the spectrometers are very powerful in monitoring the curing process, the typical measuring speed is slow. The ber optic reectometers measure Fresnel’s reection rate at t he b oundary b etween t he opt ical  ber a nd p olymer. The reection rate varies monotonically when the refractive index of the p olymer i ncreases d uring t he c uring p rocess. I t s hould b e noted t hat t emperature mon itoring ne ar t he  ber t ip i s ne cessary b ecause t he tem perature e ffect o n t he re fractive i ndex cannot b e ne gligible. The m ain adv antage o f t his me thod i s a

simple, low-cost, a nd fast measurement system. The strain a nd temperature sensors are mainly used for health monitoring and are men tioned l ater. A lthough t he s train i s not c onvenient i n evaluating t he D OC, t he me asurements o f t he re sidual s train during the curing process yield benets especially if the sensors are i nstalled b efore t he p rocess i s i nitiated. A ny t ype o f  ber optic strain sensors can be utilized for the purpose. A d ielectric sensor consists of t wo pa rallel electrodes whose pattern i s p rinted o n a t hin no nconductive ba se a s s hown i n Figure 12.2. Si nce t he electrodes  lled w ith t he polymer forms an equivalent RC circuit, the sensor is able to measure the complex dielectric constant of the polymer. For the cure monitoring, the reciprocal of a product of the alternating current frequency and the loss factor, which is called ion viscosity, is more convenient t han t he p ermittivity b ecause i t v aries o n a loga rithmic scale d uring t he c uring p rocess. I n t he l iterature, i t h as b een shown that the log ion viscosity of epoxy resin behaves similar to the DOC during the cure process. Piezoelectric m aterials s how t he re versible p iezoelectric effect, i n w hich e xternal app lied vol tage c hanges t he m aterial shapes. Therefore, piezoelectric materials can be used as actuators as well as sensors. For cure monitoring, piezoelectric sensors have been employed to monitor the impedance of polymers or to me asure the ultrasonic velocity. For the impedance measurement, t he p iezoelectric w afer emb edded i n t he v iscous polymer can be modeled by a mass-spring-damper system and represented as an electromechanical circuit. Therefore the electrical re sponse f or t he re sonance f requency i s s ensitive to changes i n s hear f or t he mo dulus a nd v iscosity o f t he c uring polymer a round t he sensor. On t he ot her hand, t he piezoelectric s ensors/actuators c an b e u sed to t ransmit a nd re ceive t he ultrasonic sig nals p ropagating t hrough t he p olymer a nd t hen the sound velocity, which is proportional to t he square root of the s tiff ness, c an b e ob tained. I t h as b een re ported t hat t he behavior of the sound velocity of the polymer can be monitored during the curing process. Optical fiber

Optical fiber Resin

Resin

Base (a)

Transmission type

Stripped fiber (High index)

Base (b)

Evanescent type

Optical fiber

Resin

(c)

FIGURE 12.1

43722_C012.indd 2

Reflection type

Illustrations of the sensing parts of ber optic spectrometers and reectometers.

10/3/2008 8:48:43 PM

12-3

Cure and Health Monitoring Electrodes

Resin

Base film

FIGURE 12.2 Interdigital dielectric sensors where the electrodes were printed on the base  lm.

12.3 Health Monitoring Like a h uman b ody, s tructures de teriorate i n lo ng-term u se b y mechanical and thermal fatigue, overload, impact load, and chemical c orrosion, e tc. The de terioration o f s tructures, w hich often causes catastrophic accidents, is the most signicant problem, and therefore, periodic and accurate inspections are essential in order to ensure safe operation. Although the most common inspection technique i s a v isual i nspection, d amages i n m aterials t hat a re internal are not detectable. Thus, nondestructive evaluation (NDE) techniques h ave b een de veloped. A lthough t he N DE te chniques such as an ultrasonic scan, X-radiography, and thermography are widely used, they can be used only during the downtime of the machines due to the large and heavy apparatuses. With health monitoring the use of built-in sensors is an attractive approach to monitor structures successfully. Especially in aeronautical and civil engineering, successive health monitoring has been expected to dramatically improve safety because shocking disasters such as the crash of aircraft and collapse of buildings have occurred recently. In health monitoring, temperature distribution, strain responses, vibration responses, and damage initiation as well as development are the targets to be monitored since t hey le ad to de gradation a nd de struction o f s tructures. Table 12.2 lists the major sensors and sensing methods available

for h ealth m onitoring wi th th eir m easurable v alues an d th eir area of coverage. Optical  ber s ensors a re t he mo st p romising f or i nternal health monitoring in the long term due to their small size, light weight, h igh me chanical p erformance, a nd h igh d urability. There a re m any t ypes o f opt ical  ber s ensors t hat h ave b een developed, w hich a re more ac curate, lower i n c ost, a nd h ave a higher functionality, as listed in Table 12.3. In the table, optical ber s ensors h ave b een c ategorized b y t heir opt ical s ystems, which strongly affect costs, sensing speed, and sensing area. Optical lo ss s ensors a re ba sed o n t he lo ss o ccurring i n t he sensing parts. The earliest idea of health monitoring by the sensors using the  ber breaks to denote t he damage progression of materials can be seen in literature. Today, many kinds of optical loss sensors have been proposed as shown in Figure 12.3. The microbend s ensor u ses opt ical lo ss c hanges s hown b y d amage development or the strain of a material. The tapered sensors have their sensing parts constructed to induce a leak of light and then the opt ical lo ss app ears a s a f unction o f s train. The vibration sensor c an me asure t ransverse v ibration w hen app lied to t he sensor f rom t he t ransmission lo ss c aused b y t he opt ical a xial slippage between the input and output bers. The advantage of the optical loss sensor is the simplest and cheapest optical system consisting of a laser-emitting diode (LED), a photodetector (PD), and a communication optical ber. In addition to the cost advantage, t he s ensor s ystem nat urally h as a f unction of h igh-speed measurement due to their high optical frequency. The M ichelson i nterferometric, p olarimetric, a nd l aser Doppler velocitymeter (LDV) sensors a re f undamental sensing methods w hich u se opt ical  bers. F or c ure mo nitoring, t he interference between two polarized lights is often m easured by the p olarimetric s ensing de vice. The opt ical s ystem i s v ery simple, but it i s d ifficult to m aintain stability for t hese sensors when the two different optical bers for measuring and reference are em ployed. A s f or L DV s ensors, s ome ide as to a mplify Doppler’s effect in the sensing part have been proposed recently in order to improve their sensitivity and stability. Although t he e xtrinsic F abry-Perot i nterferometer ( EFPI) sensor is one of t he optical interferometers, t he difference from the ot hers i s t hat t he i nput a nd o utput l ights p ropagate i n t he same ber. Therefore, the sensor shows very good stability when

TABLE 12.3 Optical Fiber Sensors for Health Monitoring TABLE 12.2 Sensors and Sensing Techniques for Health Monitoring Optical ber sensors

Piezoelectric sensors Electric resistance measurement

43722_C012.indd 3

Measurable Values

Multiplexing

Optical System

Switching

Power meter

Michelson Polarimeter LDV EFPI FBG, LPG

Damage, strain, vibration Strain, vibration Strain, vibration Velocity Strain Strain, temperature

Interferometer Interferometer Interferometer Interferometer Spectrometer

Distributed

Strain, temperature

Switching Switching Switching Switching Frequency domain Distributed

Optical loss

Measurable Values

Measurable Area

Temperature, strain, vibration, velocity, damages Vibration, impact, acoustic emission, damages Strain, damages

Around ber

Multi points Whole

OTDR

10/3/2008 8:48:43 PM

12-4

Smart Materials

Deformation

Optical fiber

Breaks ensor Deformation

(b)

Tapered section

FIGURE 12.3

Host material

Air gap

Vibration

Optical fiber

Tapered strain sensor

Microbend

Microbend sensor Capillary

Optical fiber (c)

Crack

Deformation

Host material

Optical fiber (a)

Crack

(d)

Vibration sensor

Optical loss sensors for health monitoring.

it is embedded in materials. The sensor is constructed f rom a n input/output  ber, a reector, a g lass capillary, and the two atend opt ical  bers, which compose a F abry-Perot i nterferometer as illustrated in Figure 12.4. The typical gauge length and strain resolution are 1 mm and 10−6, respectively. The EFPI sensor has been m ainly app lied to mo nitor s train w hile s pecialized t ypes have been used to measure the pressure of existing temperature. The ber Br agg g rating ( FBG) s ensor i s t he mo st p opular sensor for health monitoring due to the high sensitivity, durability, and i ts m ultiplexing c apability. The F BG s ensor h as a s ensing part whose typical length is 10 mm in which the refractive index of the core varies periodically to form a Bragg grating. When a broadband l ight is i ncident i nto t he sensor pa rt, t he na rrowband diffracted light reects and the center wavelength changes proportionally to s train a nd tem perature. The t ypical re solutions are 10−6 strains and 0.1°C. Although the FBG sensor is not

suited f or me asuring no nuniform s train d istribution suc h a s strain concentration around a crack tip, however, this weak point becomes a n adv antage f or t he p urpose o f mo nitoring d amage development by placing the sensor near neighboring weak parts of structures. The long-period grating (LPG) sensor is similar to the F BG s ensor b ut t he p eriod o f t he g rating i s m uch lo nger. Since the strain and temperature sensitivities of the several loss peaks of the transmitted light are different from each other, the LPG sensor can measure strain and temperature simultaneously by the single sensor. The main advantage of the FBG sensors for health monitoring is their ease and quick multiplexed measurements in the frequency domain. The distributed ber optic sensors can measure strain and temperature d istribution a long t he op tical  bers b y m easuring t he reected light at a ny part. An optical time domain reectometer (OTDR) can be employed to scan the intensity, spectrum, and

Optical fiber

Capillary tube Cavity length

Light from source

Transmitted light

Adhesion

Reflected light Adhesion

Gauge length Half mirror

FIGURE 12.4 Structure of the EFPI sensor.

43722_C012.indd 4

10/3/2008 8:48:43 PM

12-5

Cure and Health Monitoring

Piezoelectric sensors Vibration

(a)

Structural vibration

(b)

Impact damage detection

Damage

Acts as an actuator

Acts as a sensor (c)

Damage diagnostics

FIGURE 12.5 Health monitoring by piezoelectric sensors/actuators.

position o f t he r eected l ight suc h a s R aman s cattering a nd Brillouin scattering. The distributed sensors are practical for the point of wide area measurement of large structures but the typical spatial resolution is almost 1 m and the scanning speed is slow. Lead z irconate t itanate ( PZT) c eramic w afers a nd p olyvinylidene uoride (PVDF) lms are popular piezoelectric sensors for h ealth mo nitoring. PZ T i s mo re p owerful t han PV DF a nd operates very efficiently as an actuator as well as a sensor while PVDF can be formed into any desired shape due to its high exibility. Typical h ealth mo nitoring me thods u sing p iezoelectric sensors are structural vibration monitoring, impact monitoring, and d amage d iagnostics a s i llustrated i n Figure 12.5. I n s tructural v ibration mo nitoring, t he dy namic s train re sponses a re measured by the sensors and then modal shapes and frequencies are analyzed, which change when the structures deteriorate. The impact monitoring i s a lso i mportant for t he s tructure b ecause the i mpact d amages often c ause de struction of s tructures. The sensors measure the impact signals and the impact energy, location, a nd gener ated d amages c an b e e stimated. The passive damage d iagnostics u se a coustic e mission (AE) si gnals, w hich occur t hrough d amage i nitiation a nd p rogress. The active method has been achieved by allowing the piezoelectric sensors/ actuators to operate and thus transmit and receive the diagnostics sig nals. A lthough t he ac tive d amage d iagnostics re quire a power s ource, t hey a re t he mo st p owerful b ecause t hey ac t at any time without structural vibration, impact or damage development. The ele ctric re sistance me asurement o f c arbon rei nforced composites and carbon/steel reinforced concretes is attractive to monitor s train a nd d amages. The m ain adv antage f or t his method is the unnecessity of embedding of sensors because the materials work as sensors. The resistance changes by deforming the shape of the materials and the damage initiation. The mechanism of the resistivity variation is very complicated and particularly for the reinforcing structures of the materials because the

43722_C012.indd 5

Matrix crack

Fiber breakage

0
plies

90
plies

Delamination

FIGURE 12.6

Current path

Current paths in CFRP laminates with damages.

complex current paths are constructed mainly by the conductive reinforcements as illustrated in Figure12.6. Th is method has an interesting f unction i n w hich t he re sistance c an re cord t he degree of damages.

Bibliography Abry, J.C., Bochard, S., Chateauminois, A., Salvia, M., and Giraud, G. 1999. In situ detection of damage in CFRP la minates by electrical r esistance me asurement. Composite S cience a nd Technology, 59:925–935. Arregui, F.J., Matias, I.R., and Amo, M.L. 2000. Optical ber strain gauge bas ed o n a t apered sin gle-mode  ber. Sensors a nd Actuators, 79:90–96. Doyle, C. and Fernando, G. 1998. Detecting impact damage in a composite m aterial wi th a n o ptical  ber vib ration s ensor system. Smart Materials and Structures, 7:543–549. Glossop, N.D .W., Dub ois, S., T saw, W., L eBlanc, M., L ymer, J ., Measures, R .M., a nd T ennyson, R .C. 1990. Op tical  ber damage detection for an aircraft composite edge. Composites, 21:71–80.

10/3/2008 8:48:44 PM

12-6

Islam, A.S. and Craig, K.C. 1994. Damage detection in composite structures usin g p iezoelectric ma terials. Smart M aterials and Structures, 3:318–328. Kageyama, K., Murayama, H., Ohsawa, I., Kanai, M., Nagata, K., Machijima, Y., and Matsumura, F. 2005. Acoustic emission monitoring of a reinforced concrete st ructure by applying new  ber-optic sensors. Smart M aterials a nd S tructures, 14:52–59. Kalamkarov, A.L., Fitzgerald, S.B., and MacDonald, D.O. 1999. The use of Fabry Perot  ber optic s ensors to mo nitor residual strains d uring p ultrusion o f FRP co mposites. Composites: Part B, 30:167–175. Kim, J.S. and Lee, D.G. 1996. Measurement of the degree of cure of carbon ber epoxy composite materials. Journal of Composite Materials, 30:1436–1457. Leung, C.K.Y., Elvin, N., Olson, N., Morse, T.F., and He, Y.F. 2000. A novel distributed optical crack sensor for concrete structures. Engineering Fracture Mechanics, 65:133–148. Liu, Y.M., Ganesh, C., Steele, J.P.H., and Jones, J.E. 1997. Fiber optic sensor development for real-time in-situ epoxy cure monitoring. Journal of Composite Materials, 31:87–102. Murukeshan, V.M., Chan, P.Y., Ong, L.S., and Seah, L.K. 2000. Cure monitoring of smart composites using  ber Bragg grating based em bedded s ensors. Sensors and A ctuators: A , 79:153–161. Okabe, Y., Yashiro, S., Kosaka, T., and Takeda, N. 2000. Detection of transverse cracks in CFRP co mposites using embedded ber Bragg grating sensors. Smart Materials and Structures, 9:832–838.

43722_C012.indd 6

Smart Materials

Paik, H.J. and Sung, N.H. 1994. Fiberoptic intrinsic  uorescence for in-situ cure monitoring of amine cured epoxy and composites. Polymer Engineering and Science, 34:1025–1032. Powerll, G.R., Crosby, P.A., Waters, D.N., France, C.M., Spooncer, R.C., and Fernando, G.F. 1998. In-situ cure monitoring using optical bre sensors—a comparative study. Smart Materials and Structures, 7:557–568. Rath, M., Döring, J., Stark, W., and Hinrichsen, G. 2000. Process monitoring o f mo ulding co mpounds b y u ltrasonic me asurements in a compression mould. NDT & E International, 33:123–130. Smith, J ., B rown, A.W., D eMerchant, M.D ., a nd B ao, X. 1999. Simultaneous strain and temperature measurement using a Brillouin sca ttering ba sed d istributed se nsor. Proc. S PIE, 3670:366–373. Todoroki, A., 1998. Health monitoring of internal dela mination cracks for graphite/epoxy composites by ele ctric p otential method. P roc. 4th ESS M a nd 2nd MIMR C onference, pp. 429–434. Tracy, M., Roh, Y.S., and Chang, F.K., 1996. Impact damage diagnostics for composite structures using built-in sensors and actuators. Proc. of Third ICIM/ECSSM ‘96, pp. 118–123. Vries, M., Bhatia, V., D’Alberto, T., Arya, V., and Claus, R.O. 1998. Photoinduced g ranting-based opt ical  ber s ensors f or structural a nalysis a nd co ntrol. Engineering Str uctures, 20:205–210. Wang, X., Ehlers, C., Kissinger, C., Neitzel, M., Ye, L., and Mai, Y.W. 1998. S ensitivity o f p iezoelectric wa fers t o the c uring o f thermoset resins and thermoset composites. Smart Materials and Structures, 7:113–120.

10/3/2008 8:48:44 PM

13 Drug Delivery Systems 13.1 Smart Drug Delivery Systems ..............................................................................................13-1 Introduction • S timuli-Modulated Release Systems • Na noparticles in Drug Delivery • C onclusions

Acknowledgment ...............................................................................................................................13-7 References ...........................................................................................................................................13-7 13.2 Drug Delivery Systems: Smart Polymers ..........................................................................13-10 Introduction • Development of Controlled Drug Delivery Systems • Pulsatile Systems • Self-Regulated Systems (Environmentally Responsive Systems) • Concluding Remarks

References ......................................................................................................................................... 13-14

13.1

Smart Drug Delivery Systems

Il Keun Kwon, Sung Won Kim, Somali Chaterji, Kumar Vedantham, and Kinam Park 13.1.1

Introduction

Polymers h ave b een in dispensable in p reparing v arious c ontrolled release formulations. Of the many routes of drug administration, oral delivery is most attractive due to high patient convenience and compliance, but it is oĀen very challenging. For example, oral delivery of drugs that are poorly soluble, or are not stable in the gastrointestinal tract, or have high molecular weights has been difficult. Delivery of high molecular weight drugs, such as peptide and protein drugs and oligonucleotides, by parenteral administration has also been difficult due to their short circulation half-lives as well as lack of suitable delivery systems for long-term delivery. A number of approaches have been explored to increase the b ioavailability o f h ydrophilic/hydrophobic d rugs, p eptides, and proteins by conjugating drugs with natural or synthetic macromolecules or lo ading i n p olymeric d rug d elivery s ystems. Polymers with functional properties have been used to overcome hurdles in delivery of diverse drugs. Stimuli–responsive po lymers a re po lymers wh ich r espond with physical or chemical transformations in response to changes

in en vironmental c onditions, suc h a s tem perature, p H, mechanical stress, electric/magnetic response, light, and ionic strength. Figure 13.1 s chematically presents t he c oncept of a stimuli–responsive drug delivery system using glucose-dependent self-regulated in sulin r elease as a n e xample. The s ensor t hat recognizes the stimulus, glucose in this case, should have high specicity a nd s ensitivity. The i nformation p rocess s hould handle t he sig nal a s f ast a s p ossible, a nd s o t he k inetics o f information p rocessing is h ighly i mportant. The actuator should h ave t he a bility to re verse t he ac tion, i .e., t ransition between o n a nd off st ates, w ith s ufficient m agnitude to b e useful. St imuli–responsive p olymers c an b e i n t he f orm o f a drug re servoir, w hich c an rele ase i ts c ontents i n re sponse to environmental signals. Th is release can be via changes in their shapes, surface properties, solubility parameters, or by molecular self-assembly, and sol–gel transition [1–7]. Nanotechnology, in combination with diverse targeting strategies, c an b ring adv antages to d rug del ivery, o vercoming t he limitations o f c onventional t herapeutics [ 8]. The advantages include delivery of much higher therapeutic payloads per target biorecognition e vent; t he a bility to c arry m ultiple, p otentially different targeting agents for enhanced selectivity; the ability to integrate me ans to b ypass biological ba rriers; a nd t he c olocalized delivery of multiple agents, resulting in targeted combination therapy. This c hapter s ummarizes v arious s timuli-modulated drug delivery systems as well as the applications of nanoparticles in drug delivery.

13-1

43722_C013.indd 1

10/3/2008 8:49:03 PM

13-2

Smart Materials

Changes in environmental factors

Changes in glucose level in blood ( CH2 Glucose sensor

Sensor

Information processor

Determine the amount of insulin to be released

Actuator

Feedback: Stop insulin release Insulin release

FIGURE 13.1 Schematic i llustration of t he on– off d rug re lease i n a stimuli-modulated d rug del ivery s ystem. Gl ucose-dependent m odulated insulin release is shown on the right as an example.

13.1.2 Stimuli-Modulated Release Systems 13.1.2.1 Thermoresponsive Polymers The sig nals a nd ph ysiological c onditions t hat c an t rigger en vironment-sensitive p olymers a re tem perature, p H, me chanical stimulus, biomolecule concentration, electric eld strength, and salt concentration. Of these, temperature is the most commonly studied t riggering sig nal f or mo dulated d rug del ivery [ 7,9,10]. Temperature i s a u seful s timulus f or c ontrolled d rug del ivery, because human body temperature deviates from 37°C in the presence of pyrogens [8,9]. The common characteristic of temperaturesensitive polymers is the presence of a h ydrophobic group, such as methyl, ethyl, and propyl groups. The properties of thermoresponsive polymers are governed by lower critical solution temperature (LCST), which is dened as the temperature at which the polymer solution undergoes a ph ase transition from a s oluble to an insoluble state [11]. This phase transition is mainly determined by t he hydrophilic–hydrophobic ba lance of t he p olymer [12]. I n general, the solubility of most of the polymers increases with increase in temperature. In the case of polymers that exhibit LCST, increase i n tem perature de creases t he w ater s olubility d ue to hydrophobic associations of polymer molecules and reduction in hydrogen bonding between polymer and water molecules [10]. The s tudy o f t hermoresponsive gel s s tarted i n 1 978 w hen Tanaka reported the thermodynamics behind the collapse of a polyacrylamide g el [ 13]. M ost co mmonly st udied s ynthetic thermoresponsive p olymers a re p oly(N-isopropylacrylamide) (PNIPAAm), p oly(N,N′-diethylacrylamide) ( PDEAAm), a nd poly(2-carboxyisopropylacrylamide) (PCIPAAm) (Figure 13.2). PNIPAAm is one of the most useful thermoresponsive polymers in drug delivery, because its LCST in water is around 32°C, close to the body temperature [14]. The LCST can be altered by adjusting the ratio of the hydrophilic and the hydrophobic segments of the polymer [15–17].

43722_C013.indd 2

( CH2

H3C

CH (n

C

O

C

N

H

N

CH (a)

Feedback

CH (n

CH3

(b)

( CH2

O

H2C

CH2

H3C

CH3

O

H

C

O

C (n C

O

N

H

CH (c)

H3C

CH3

FIGURE 1 3.2 Chemical structures of poly(N-isopropylacrylamide) (a), p oly(N,N′-diethylacrylamide) ( b), a nd p oly(2-carboxyisopropylacrylamide) (c).

An in teresting  nding re garding tem perature-modulated release w as t hat PN IPAAm h ydrogels w ith mo re h ydrophobic comonomers, such as n-butylmethacrylate, formed a dense skin during t he deswelling process when t he temperature increased above its LCST [18–20]. This dense skin formation (surface regulation) is able to block the release of loaded drugs from the matrix or diff usion t hrough t he membrane, resulting i n t he reversible off-control f or so lute r elease. P (NIPAAm-co-methacrylic a cid) was app lied f or t he p ulsatile rele ase o f a t hrombolytic a gent, streptokinase, b y s imultaneous p H an d t emperature c hanges, which c ould o ccur i n blo od pH a nd temperature at t he site of blood clotting [21]. The pulsatile release pattern of loaded streptokinase was obtained by uctuating tem perature a nd p H, simultaneously. The opp osite o n–off d rug rele ase pat tern ( i.e., release, o r “on-control,” at h igh tem perature) w as ob tained b y using the interpenetrating polymer network (IPN) structures of poly(acrylamide-co-butylmethacrylate) a nd p oly(acrylic ac id) (PAAc) [22] or poly(N,N-dimethylacrylamide [DMA]) and PAAc [23]. The polymer networks are formed by hydrogen bonds at a lower temperature. At a higher temperature, the hydrogen bonds are dissociated, resulting in release of the loaded drug. The sensitivity of sol–gel transition can be accelerated by structural molecular design, e.g., making comb-type PNIPAAm [24,25] and by changing the synthesis condition, such as a weak alkaline condition, for P(NIPAAm-co-AAc) [26]. Dissociation of the carboxyl g roups (–COOH) of A AC to c arboxylate a nions (–COO−) results in electrostatic repulsion between the carboxylate anions, leading to e xpanded c onformation o f t he p olymer c hains. The expanded conformation induces fast temperature sensitivity and improves the oscillating swelling–deswelling property. A l arge n umber o f p oly(ethylene o xide) ( PEO) a nd poly(propylene o xide) ( PPO) b lock co polymers possess a n inverse tem perature-sensitive m icellization a nd gel ation p roperty. M any a re c ommercially a vailable u nder t he na mes o f Pluronics (or Poloxamers) and Tetronics [27]. Gelation of PEOPPO-PEO can be induced by the three-dimensional packing of micelles formed by the hydrophilic–hydrophobic balance at over the critical temperature [28]. The PPO hydrophobic block can be replaced with other hydrophobic groups, such as poly(1,2-butylene

10/3/2008 8:49:04 PM

13-3

Drug Delivery Systems

oxide) (PBO) [29,30] a nd polyesters, such a s poly(l -lactic acid) (PLLA), a nd ( dl- lactic ac id-co-glycolic ac id) ( PLGA) [ 31–33]. The PEO-PLLA-PEO triblock copolymer, which combined thermogelation a nd b iodegradability w ith lo nger gel d uration, showed the sol state at 45°C and transformed to a gel at a lo wer temperature (e.g., 37°C) [32]. This t ransition helps loading a nd delivering of a d rug at a su stained rate in t he body. A s eries of low mole cular w eight P LGA-PEG-PLGA c opolymers a lso showed si milar t hermoreversible s ol–gel t ransitions to t hat o f PEO-PLLA-PEO block copolymer [34]. The lower sol–gel transition is related to the increase in aggregation numbers and size of micelles consisting of a polyester core and a PEO shell at higher temperatures, and the upper gel–sol transition induced from the shrinking of micelles or shape changes caused by dehydration of the PEO block [32]. The sol–gel transition temperature is a function of both concentration and composition of the block copolymers. The ABA-type triblock copolymers have potential as an injectable drug delivery system with their sharp phase transition property. Some nat ural p olymers i ncluding p roteins [ 35], p olysaccharides [36,37], and their complexations [38–40] possess thermoresponsive properties and have been used to form hydrogels by chemical or ph ysical c ross-linking. For e xample, ge latin c ombined with chitosan [38] or polyol salt [39] has been used for injectable drug delivery. Other naturally available thermoresponsive polymers are cellulose derivatives. Methylcellulose (MC) and hydroxypropylmethylcellulose ( HPMC) ar e t ypical e xamples. The transition temperatures of MC and HPMC are known to be

40–50°C and 75–90°C, respectively [41]. Kawasaki et al. studied xyloglucan, a naturally occurring thermoresponsive polysaccharide, which is derived from tamarind seeds as a v ehicle for oral delivery. Degradation of xyloglucan by β-galactosidase results in the formation of a product that undergoes sol–gel transition [42]. Recently, recombinant articial elastin-like polypeptides (ELPs), which a re b iopolymers w ith re peating p entapeptides ( Val-ProGly–X–Gly, where the “guest residue” X can be any of the natural amino ac ids e xcept p roline), w ere re ported to u ndergo a t hermally reversible phase transition [43,44]. With current advances in biotechnology, coupling of stimuli-sensitive peptide motifs to a synthetic polymer or other polypeptide can create novel stimuli-sensitive hybrid systems [45]. 13.1.2.2 pH-Sensitive Polymers The pH variation in the GIT is pronounced. The stomach has an acidic environment, pH of 1–2 in a fasting condition, and pH 4 during d igestion. The p H i n t he d uodenum i s 5. 5 d ue to t he mixing of acidic chyme with bicarbonate from pancreatic juices. The extracellular and intracellular pH values in most cancers are more ac idic (about p H 6 .7 a nd p H < 7 .0, re spectively) t han i n normal t issues or c ells [46–48]. Suc h d ifferences i n pH v alues, even t hough sm all, c an b e e xploited f or c ancer t argeted d rug delivery using pH-responsive polymeric drug delivery system. The pH-sensitive polymers include ionizable groups such as carboxylic, su lfonic ac id, o r ba sic a mino g roups t hat c an accept or release protons in response to the environmental pH [10]. Figure 13.3 shows chemical structures of representative CH3

CH3

H ( CH 2

(a)

C (n C

O

O

H

( CH 2

(b)

CH3 ( CH 2

H2C (e)

C

O

O

H

( CH 2

(c)

CH3

C (n C

C (n

( CH 2

O

C (n C O

CH2

CH2

CH2

CH2

N

N

(f)

CH2

CH2

CH2

C (n C

O

O

H

( CH 2

(d)

CH3 ( CH 2

C (n

H3C

CH3

O

O

H

( CH 2

C (n

N CH3 ( CH 2

CH2

C

N (g)

H2C

C (n

CH3

O

O

CH2

CH3

(h)

C (n

N

FIGURE 13.3 Chemical structures of representative pH-responsive polymers: (a) poly(acrylic acid), (b) poly(methacrylic acid), (c) poly(3-ethyl acrylic a cid), ( d) p oly(2-propyl a crylic a cid), ( e) p oly(N,N′-dimethylaminoethyl me thacrylate), ( f) p oly(N,N-diethylaminoethyl m ethacrylate), (g) poly(4 or 2-vinypyridine), and (h) poly(vinyl imidazole).

43722_C013.indd 3

10/3/2008 8:49:04 PM

13-4

pH-responsive polyacids or polybases. Depending on pH, such polyelectrolytes u ndergo a t ransition b etween p rotonation and de protonation d ue to t heir pK a or p Kb. I n gener al, w eak acids a re w ater-soluble b y de protonation w hile w eak ba ses become le ss s oluble. Thus, i n a n aq ueous en vironment, t he ionization le ads to s welling o f p olymer h ydrogels a s w ell a s disintegration of physically assembled polymer particles [48]. In addition to pH, however, pH sensitivity also is inuenced by other parameters, such as ionic strength or hydrogen bonding. For e xample, PAAc, w ith it s p K a a round 4 –4.5 [49], h as extended chain structure at p H above t he pK a va lue, because of the polymer chain hydration and electrostatic repulsion of charged p olymer c hains. H owever, at a lo w p H, P AAc c an make a complex with any proton-accepting polymers, such as PEG or polyvinylpyrrolidone (PVP) [49,50]. Complex formation through hydrogen bonding between these polymers leads to squeezing out of water molecules from the complex. On the other h and, h igh io nic s trength i n p olymer s olution u sually masks t he charge of a p olymer enough to re duce t he electrostatic repulsion between polymer chains. Thus, the pH-sensitive hydrogel showed higher swelling capacity at low ionic strength than at high ionic strength [51]. The pH-sensitive polymers, which can bypass drug resistance and deliver sufficient amounts of the anticancer drug, were evaluated both in vitro and in vivo experiments [52–54]. A new pHsensitive p olymer ba sed o n su lfonamide der ivatives h as b een used for targeting to a receptor that was overexpressed by acidic extracellular matrix (ECM) of cancers [55–57]. Other pH-sensitive polymers, block copolymers of poly(l -histidine), which is a polybase, a nd P EG, w ere a lso s ynthesized to m ake p olymeric micelles t hat c an re spond to t he lo cal pH changes i n t he b ody [58,59]. The local pH change around cancer cells has a lso been used to del iver D NA f rom na noparticle c omplexes f ormed b y electrostatic interactions between DNA, cationic polymer (PEI), and amphiphilic copolymer. The complex formed at physiological p H d issociated to rele ase D NA at p H 6 .6 [60]. C hitosan, a polysaccharide, is also known to be complexed by citrate in the pH r ange o f 4 .3–7.6 [ 61]. Thus, c hitosan, b eing a p H-sensitive natural polymer, can be used to del iver d rugs in a c ontrolled manner by exploiting the pH-sensitive formation and disintegration of chitosan-containing polyelectrolyte complexes. Exploring the wide diversity in the behavior of polymer gels, Luo et al. s tumbled up on a n i nteresting property i n a c lass of gels-silicone ba sed p olysilamines [ 62]. The c hemistry o f t his novel class of hydrogels comprises of a lternating d iamine a nd organosilyl units. Surprisingly, these gels hardened on swelling based o n a re versible ro d–globule t ransition. I n add ition, i n aqueous me dia, p olysilamine h ydrogels e xhibit re producible swelling/syneresis behavior in response to the protonation degree of t he ne twork. The gel ne twork c hains e xpand a nd rigidify their conformation on protonation and anion binding, which induce the gel swelling consistently. Since a drawback in the use of most conventional hydrogels as biomedical devices is the decrease in mechanical properties on swelling, t he unique

43722_C013.indd 4

Smart Materials

property o f t his c lass o f si licone-based h ydrogels m akes t hem attractive for use in controlled drug delivery, especially as pump systems in pulsatile drug delivery. 13.1.2.3 Biomacromolecule-Sensitive Polymers 13.1.2.3.1 Glucose In the treatment of diabetes caused by the malfunction of pancreas, g lucose-responsive p olymers, w hich u ndergo s welling changes i n re sponse to g lucose, c an c onstruct s elf-regulated insulin-delivery s ystems, in w hich a n a ppropriate a mount o f insulin c an b e rele ased i n re sponse to t he blo od g lucose concentration. Glucose-sensitive po lymers ha ve bee n d ivided i nto bo ronic acid derivatives, g lucose oxidase (GOx) conjugates, a nd concanavalin A ( ConA) c onjugates. The b oronic ac id mo iety c an accept o ne h ydroxyl io n to b e ne gatively c harged w hen o ver a certain pH (pKa of boronic acid). The ionized form of b oronic acid can bind to a trans-diol such as glucose w ith high a ffinity [63,64]. pH-sensitive polymers in coupling with the GOx can change the hydrophilicity depending on the existence of glucose, since it can degrade glucoses to yield proton ions, which lead the local environments to b e ac idic. PAAc w as u sed for forming a gate to re gulate s olute p ermeability i n a memb rane. I f, at h igh glucose c oncentration, G Ox gener ates a l arge a mount o f g luconic ac id, lo cal pH de creases a nd protonated A Ac c auses low permeability [ 65,66]. L ikewise, su lfonamide-based h ydrogels also de creased its s welling c apacity at h igh g lucose c oncentration [67]. Since the ConA can hold four glucose units per molecule, C onA blen ded w ith g lucose-bearing p olymers f orms cross-linked g lucose-sensitive h ydrogels. A s t he c oncentration of glucose increases, the polymer network becomes loose enough to s peed up t he lo aded d rug ( e.g., i nsulin) rele ase [ 68–70]. Glucose-responsive h ydrogels ha ve be en ma nipulated w ith ConA a nd g lycopolymers t hat c ontains le ctin-interacting s accharide moieties [71]. You et al. designed PEO-b-poly(2-glucosyloxyethyl ac rylate) de signed f or a g lucose-responsive m icellar structure, which might release the loaded insulin only when the glucose c oncentration i n blo od w as i ncreased [72]. It i s note d, however, that ConA may leak out into the environment during swelling in an aqueous glucose solution, since it is not covalently bound in the hydrogel networks [73]. 13.1.2.3.2 Glutathione Glutathione i s a n i mportant b iomolecule i n m aintaining t he homeostasis o f cel lular r edox st ate. K akizawa et al . re ported glutathione-sensitive m icelles t hat w ere re ady to b e c leaved inside cell rather than circulating blood [74]. Usually, the micelles selectively deliver therapeutic oligodeoxynucleotides (ODNs) or small interfering RNAs (siRNAs) into cells without degradation during blo od c irculation, w hich h ave t herapeutic e fficacy only inside a c ell. A c onjugate b etween P EG a nd O DN i nhibiting luciferase ex pression v ia d isulde b ond suc cessfully p revented luciferase gene expression in HuH-7 cells, because of the reaction

10/3/2008 8:49:05 PM

13-5

Drug Delivery Systems

between d isulde b onds a nd i ntracellular g lutathiones [ 75]. PEG-b-thiolated poly(l -lysine) formed micelle with a cross-linked core due to disulde bridges. In this case, intracellular glutathiones made the micelle so unstable that the loaded ODN or siRNA were l iberated f rom t he io nic c omplex [ 74,76]. E l-Sayed et al . recently p roposed a p H- a nd g lutathione-sensitive p olymer to deliver therapeutic ODNs [77]. This dual stimuli-sensitive polymer provided tunable range of pH to maximize the therapeutic efficiency by varying the composition of pH-sensitive and hydrophobic moiety (butyl acrylate) or glutathione-responsive moiety (pyridyl disulde acrylate). 13.1.2.4 Other Stimuli-Responsive Polymers 13.1.2.4.1 Mechanical Stress Lee et al. prepared a de gradable and injectable hydrogel as an ECM, which was designed to respond to mechanical stimuli in the b ody. The ba sic s tructure w as ba sed o n p olyguluronate (PG) p repared f rom t he a lginate. A Āer o xidization, a ldehyde groups on the PG were treated with dihydrazide to form a cross-linked hydrogel, which released a protein drug in a pulsatile manner [78]. Pressure-sensitive polymers (PSPs) or adhesives ( PSAs) h ave b een w idely s tudied f or t he t ransdermal delivery of drugs [79]. PSAs should have lower surface energ y than (or, at le ast e qual to) t hat o f s kin w here t hey a re to b e attached. The ad hesiveness c an b e i ncreased w ith i ncrease o f pressure, which makes the elastic polymer temporally viscous. Osada a nd M atsuda s tudied a s hape-memory p olymer. A poly(N,N′-methylenebisacrylamide) h ydrogel w as t hermally deformed, and then recovered its original shape by cooling to room temperature [80]. The shape memory properties of thermoplastic polyurethane allow design of a f ully polymeric selfexpandable drug eluting stent [81]. 13.1.2.4.2 Electrical Signal In a pioneering work by Okahata et al. [82], a t hin, but porous nylon-2,12 c apsule w as p repared f rom e thylenediamine a nd 1,10-bis(chlorocarbonyl)decane. The pore was lled with bilayerforming a mphiphiles suc h a s d ialkyldimethylammonium bromide (cationic), s odium d idodecylphosphate (anionic), a nd 1,3-dihexadecyl-rac-glycero-2-phosphocholine ( zwitter ion ic), which re sulted i n re versible gat ing o f i ntracapsular p ermeants (e.g., glucose, ions) by the external electric eld ( 0–80 V). A model protein drug, zinc insulin, was loaded into a c omplex of polyethyloxazoline (PEOx) and poly(methacrylic acid) (PMAA) [83]. Because t he complex was based on t he hydrogen bonding between two polymers, the complex could form a s table matrix that was easily erodible at pH over 5.4 by deionization of acid moieties o f P MAA. E lectrical c urrent m ight h ydrolyze w ater molecules, le ading to m any O H− io ns. Thus, t he m atrix c ould turn on the insulin–release by application of the electrical current (5 mA). A nother p ulsatile d rug (insulin) del ivery w ith a n electroresponsive hydrogel was studied [84]. The gel led to plasma glucose pulsatility when stimulated with a c onstant current of

43722_C013.indd 5

1.0 mA (0.36 mA/cm2) in vivo. The mechanism of insulin release from t he c ross-linked p olymer, w hich h as a h igh pKa (10.4), i s associated w ith e lectrokinetic  ow of solvated insulin with water; that is, transportation process of counterions (electrophoresis) a nd water molecules (electroosmosis) i n t he cross-linked polyelectrolyte gel network. 13.1.2.4.3

Ultrasound

With t he de velopment o f b iodegradable p olymers, u ltrasound has been used to externally regulate the release pro le of a drug loaded in a c arrier [85,86]. For example, Pluronic micelles containing doxorubicin showed a fast drug release rate as a function of u ltrasound f requency as well as power density, which effectively enhanced intracellular drug uptake in a breast cancer cell line w ith multidrug re sistance (MDR) [87]. U ltrasound c an b e utilized not only to facilitate the degradation rate of biodegradable polymers, but also to modulate the mass transport property of nondegradable polymers. Lavon et al. investigated the effect of the u ltrasound on model d rug release f rom a no nerodible polymer hydrogel, p oly(hydroxyethyl me thacrylate-co-N,N′dimethylaminoethyl m ethacrylate) ( PHEMA-N,N′-DMAEMA) [88]. The ultrasound could enhance the drug release rate from a matrix where the mass transport is limited. 13.1.2.4.4

Magnetic Field

Polymeric coating onto the magnetic or composite particles has been u sed f or t he m agnetic re sonance i maging ( MRI) a gents [89,90]. Hsieh et al. prepared a p olymer matrix (ethylene–vinyl acetate copolymer) containing magnetic beads and a model protein drug, bovine serum albumin (BSA) [91]. When exposed to an oscillating magnetic  eld, t he polymer matrix released BSA in a sustained manner, with the small magnetic beads embedded in the matrix might enlarge either pores or channels where the BSA moved through. 13.1.2.4.5 Redox State At t heir e arly s tage o f de velopment, re dox-sensitive p olymers have been used to modulate the permeability of vesicle membranes [92]. For example, a recent study using ABA-type triblock copolymer o f P EG-b-poly(propylene s ulde) (PPS)-b-PEG reported disintegration of the vesicle membranes responding to the oxidation signal [93]. The hydrophobic PPS block is oxidized to be more hydrophilic, which nally leads to vesicle disruption. 13.1.2.4.6

Light

In a s tudy by Sh imoboji e t a l., biological ac tivity of a p rotein conjugated to a photoresponsive polymer was regulated by light irradiation [94]. Two kinds of photosensitive polymers prepared in the study were N,N-DMA-co-4-phenylazophenyl acrylate and DMA-co-N-4-phenylazophenyl acr ylamide. The former became soluble i n w ater b y i rradiating u ltraviolet (U V), b ut b ecame insoluble by irradiating visible light, while the latter possessed an opposite solubility proles. By conjugation of each polymer to a genetically engineered streptavidin, on–off switching of biotin

10/3/2008 8:49:05 PM

13-6

Smart Materials

affi nity induced by UV or visible light was conrmed. Tao et al. engineered a hollow sphere with controllable optical property as well as membrane permeability in response to the light [95]. 13.1.2.4.7

Enzymes

Recent s tudies o n t he s timuli-sensitive p olymers h ave qu ickly moved toward the enzyme-sensitive polymers. Because a disease usually ac companies at le ast o ne ph ysiologically o r b iochemically impaired function, detection and utilization of such a small change c an b e g reatly u seful for t argeted d rug del ivery. It w as already re ported t hat do xorubicin c onjugation to P EG [96] o r N-(2-hydroxypropyl)methacrylamide c opolymer [ 97] v ia p eptide linkers could be cleaved by lysosomal enzymes, leading to a great potential for cancer treatment. Recently, enzyme-sensitive hydrogels a lso have been i ntensively de veloped for t issue eng ineering [98]. 13.1.2.4.8

Ionic Strength

Most of ionic strength-sensitive polymers have also demonstrated pH-sensitivity, because these polymers are basically polyelectrolytes. Ion–polymer i nteraction re sults i n re ducing ele ctrostatic repulsion b etween io nic p olymer bac kbones, w hich m akes i t difficult for solutes to pass through. One recent study reported a specific ion selectivity of a p olymer m icrocapsule to c ontrol the solute rele ase [ 99]. P oly(NIPAAm-co-benzo-18-crown[6]acrylamide) sh owed Ba 2+-selective c losure o f t he memb rane pores, while Na+ did not have much inuence on the membrane permeability.

13.1.3 Nanoparticles in Drug Delivery Recent advances in nanotechnology utilizing various polymers and hydrogels enhanced the ability to de sign better drug delivery systems. Among them, injectable nanoparticles are of great interest for application i n cancer t reatments. The size of na noparticles c an b e sm all eno ugh to h ave i ntracapillary o r t ranscapillary passage. When the nanoparticles have long circulating stealth property achieved by PEGylation, nonspecic opsonization a nd i nteraction w ith c ells a nd rele ase o f t he lo aded d rug during circulation can be minimized [8]. Uptake of nanoparticles by cells or tissues was much higher as compared to largersize microparticles [100]. In a size-dependent cytotoxicity study, nanoparticles less than 150 nm in diameter induced signicant apoptosis i n t he en dothelial c ells. I n 2 005, t he U .S. F ood a nd Drug A dministration ( FDA) a pproved i ntravenously ad ministered pac litaxel-loaded 1 30 nm a lbumin na nogels (Abraxane™) for cancer therapy [101]. MDR i s a m ajor p roblem i n c hemotherapy, a nd i t o ccurs when tumor cells acquire the ability to pump toxic anticancer agents out t hrough t heir c ell membrane b efore t hey h ave t he ability to k ill t he c ell. T o o vercome t his p roblem, “ smart drug delivery s ystems,” such a s p H-dependent d rug del ivery, drug t argeting, a nd m olecular im aging o f dru g-containing nanoparticles, have been investigated. The targeting nanoparticle can be obtained by conjugation of site-recognition moieties,

43722_C013.indd 6

such as synthetic ligands [102], growth factors [103], antibodies [104], a nd f olic ac id ( folate) [ 105,106], to t he su rface o f t he nanoparticles. Lee et al. have developed targeted polymeric micelles using a pH-sensitive po lymer, a co polymer o f po ly(l -histidine) a nd PEG, linked to folate of which receptor is overexpressed by many t ypes o f t umors. The p olymers c an del iver su fficient amounts of t he a nticancer d rug doxorubicin to k ill ot herwise resistant t umor c ells [59]. Broz et al . h ave c onstructed a ne w class of nanoparticles using an easily constructed biocompatible polymer, b iotin-functionalized (poly(2-methyloxazoline)-bpoly(dimethylsiloxane)-b-poly(2-methyloxazoline), w hich i s capable o f del ivering t herapeutic a gents to a w ide v ariety o f cells involved in cancer and other diseases [107]. These nanoparticles, w hich we re c alled n anocontainers, autom atically incorporated payload molecules as they self-assembled in water. The uorescent prob e-labeled n anocontainers, r anging f rom 100 to 25 0 nm, w ere t argeted a gainst the scavenger receptor A1 from macrophages, an important cell in human disease. Soppimath et al. constructed pH-triggered thermally responsive n anoparticles f rom p oly(NIPAAm-co-N,N′-DMA-co-10undecenoic acid) [108]. At pH 7.4, the nanoparticles released an initial b urst o f do xorubicin ac counting f or 18% o f t he lo aded drug. In contrast, at pH 6.6 and lower the nanoparticles released approximately 7 0% o f t he l oaded drug over the course of 48 h. Xu et al . g enerated den dritic core-shell nanocarriers based on hyperbranched polyethyleneimine cores and different shells containing aliphatic chains and PEG chains, respectively [109]. Fast pH-sensitive cleavage of the imine bond occurred at pH 5–7 while relatively high stability of t he i mine b ond w as observed at a pH of 8. This pH shift corresponds to the pH shiĀ observed in malignant tissues from that of normal tissues and hence is a valuable trigger for releasing encapsulated drugs at t he t arget site. Photodynamic t herapy (PD T), which uses a light-sensitive chemical k nown a s a photosensitizer (such as porphyrins, which is a component of hemoglobin) to produce cell-killing “reactive oxygen species,” has become an important option for the treatment of esophageal cancer and nonsmall cell lung cancer [110]. While there is no doubt that nanotechnology for drug delivery has b rought a ne w d imension to d rug del ivery te chnologies, further de velopment a nd i mprovements a re ne cessary f or t he nanoparticulate formulations to be useful in clinical applications.

13.1.4

Conclusions

The ideal drug delivery systems should keep a drug at a de sired therapeutic level at the target site while avoiding frequent admi nistration. Wi th t he r apid adv ances i n p olymer f abrication keeping pace with the strides in nanotechnology, an ever-increasing number of smart polymers with unique characteristics are being engineered. With the rapid development of such stimuli-responsive polymers, a w hole l ine o f no vel d rug del ivery p latforms w ith smart f unctionalities a nd en hanced pat ient c ompliance a re bound to emerge.

10/3/2008 8:49:05 PM

Drug Delivery Systems

Acknowledgment This study was supported in part by grants from NIH t hrough 511-1336-1761 and from Purdue Cancer Center.

References 1. Bignotti, F., Penco, M., Sartore, L., Peroni, I., Mendichi, R ., Casolaro, M., and D’Amore, A., Synthesis, characterisation and s olution b ehaviour o f thermo- a nd pH-r esponsive polymers b earing L -leucine r esidues in t he side cha ins, Polymer, 41, 8247, 2000. 2. Gan, L.H., Ga n, Y.Y., a nd D een, G.R ., Poly(N-acryloyl-N′propylpiperazine): A ne w stim uli-responsive p olymer, Macromolecules, 33, 7893, 2000. 3. Jeong, B. and Gutowska, A., L essons f rom nature: Stimuliresponsive p olymers and t heir b iomedical a pplications, Trends Biotech., 20, 305, 2002. 4. Kaneko, Y. a nd Oka no, T., Temperature-responsive hydrogels as intelligent materials, in Biorelated Polymers and Gels, Okano, T., Ed., Academic Press, London, 1998, Chapter 2. 5. Peng, T. and Cheng, Y.L., PNIPAAm and PMAA co-graĀed porous PE mem branes: Li ving radical co-gra Āing mechanism a nd m ulti-stimuli r esponsive p ermeability, Polymer, 42, 2091, 2000. 6. Pinkrah, V.T., S nowden, M.J ., Mi tchell, J .C., S eidel, J ., Chowdhry, B.Z., and Fern, G.R., Physicochemical properties of poly(N-isopropylacrylamide-co-4-vinylpyridine) cationic polyelectrolyte c olloidal mic rogels, Langmuir, 19, 585, 2003. 7. Kikuchi, A. a nd Oka no, T., H ydrogels: stim uli-sensitive hydrogels, in Polymeric D rug D elivery S ystems, K won, G.S., Ed., Taylor & F rancis Group, Boca Raton, FL, 2005, Chapter 7. 8. Kwon, I.K., J eong, S.H., K ang, E., a nd P ark, K., N anoparticulate dr ug deli very f or ca ncer therapy, in Polymeric Drug De livery S ystems, N alwa, H.S., Webster, T ., E ds., American Scientic Publishers, New York, 2006, Chapter 3. 9. Bae, Y.H., Smart polymers in drug delivery, Pharm. News, 9, 417, 2002. 10. Qiu, Y. a nd P ark, K., En vironment-sensitive h ydrogels f or drug delivery, Adv. Drug. Deliv. Rev., 53, 321, 2001. 11. Kwon, Y.M. and Kim, S.W., Thermos ensitive biodegradable hydrogels for the delivery of therapeutic agents, in Polymeric Drug De livery S ystems, Kwon, G.S., E d., Taylor & F rancis Group, Boca Raton, FL, 2005, Chapter 6. 12. Song, S.-C., Lee, S.B., Jin, J.-I., and Sohn, Y.S., A new class of biodegradable thermos ensitive p olymers. I. S ynthesis a nd characterization of poly(organophosphazenes) with methoxy-poly(ethylene g lycol) a nd a mino acid ester s as side groups, Macromolecules, 32, 2188, 1999. 13. Tanaka, T., Dynamics of critical concentration uctuations in gels, Phys. Rev., 17, 763, 1978. 14. Schild, H.G., P oly(N-isopropylacrylamide): E xperiment, theory and application, Prog. Polym. Sci., 17, 163, 1992.

43722_C013.indd 7

13-7

15. Feil, H., Bae, Y.H., Feijen, J., and Kim, S.W., Mutual inuence of pH a nd t emperature o n the sw elling o f io nizable a nd thermosensitive hydrogels, Macromolecules, 25, 5528, 1992. 16. Hirotsu, S., C oexistence o f p hases a nd the na ture o f  rstorder phase transition in poly(N-isopropylacrylamide) gels, Adv. Polym. Sci., 10, 1, 1993. 17. Irie, M., S timuli-responsive p oly(N-isopropylacrylamide). Photo- and chemical-induced phase transitions, Adv. Polym. Sci., 110, 49, 1993. 18. Bae, Y.H., Okano, T., and Kim, S.W., “On–off ” thermocontrol of solute transport. I. Temperature dependence of swelling of N-isopropylacrylamide netw orks mo died with hydrophobic components in water, Pharm. Res., 8, 531, 1991. 19. Bae, Y.H., Okano, T., and Kim, S.W., “On–off ” thermocontrol of solute transport. II. Solute release from thermosensitive hydrogels, Pharm. Res., 8, 624, 1991. 20. Yoshida, R., Sakai, K., Okano, T., and Sakurai, Y., Modulating the phase transition temperature and thermos ensitivity in N-isopropylacrylamide co polymer gels, J. B iomat. S ci. Polym. Ed., 585, 1994. 21. Brazel, C.S. a nd P eppas, N.A., Puls atile lo cal deli very o f thrombolytic a nd a ntithrombotic a gents usin g p oly(Nisopropylacrylamide-co-methacrylic acid) h ydrogels, J. Control. Rel., 39, 57, 1996. 22. Katono, H., Maruyama, A., Sanui, K., Ogata, N., Okano, T., and Sakurai, Y., Thermoresponsive swelling and drug release switching of interpenetrating polymer networks composed of poly(acrylamide-co-butyl methacrylate) and poly(acrylic acid), J. Control. Rel., 16, 215, 1991. 23. Aoki, T., Kawashima, M., Katono, H., Sa nui, K., Ogata, N., Okano, T., a nd Sak urai, Y., Temperature-responsive in terpenetrating polymer networks constructed with poly(acrylic acid) and p oly(N,N-dimethylacrylamide), Macromolecules, 27, 947, 1994. 24. Annaka, M., Tanaka, C., Nakahira, T., Sugiyama, M., Aoyagi, T., a nd Oka no, T ., Fl uorescence st udy o n the sw elling behavior o f co mb-type gra Āed poly(N-isopropylacrylamide) hydrogels, Macromolecules, 35, 8173, 2002. 25. Yoshida, R ., Uchida, K., K aneko, Y., Saka i, K., K ikuchi, A., Sakurai, Y., a nd Oka no, T ., C omb-type gra Āed hydrogels with ra pid de-sw elling r esponse t o t emperature c hanges, Nature, 374, 240, 1995. 26. Zhang, X.-Z., Yang, Y.-Y., Wang, F .-J., a nd Ch ung, T .-S., The rmosensitive poly(N-isopropylacrylamide-co-acrylic acid) h ydrogels wi th expa nded netw ork str uctures a nd improved os cillating sw elling–deswelling p roperties, Langmuir, 18, 2013, 2002. 27. Bromberg, L.E. and Ron, E.S., Temperature-responsive gels and thermogelling p olymer matrixes for protein and p eptide delivery, Adv. Drug. Deliv. Rev., 31, 197, 1998. 28. Zhang, K. a nd K han, A., P hase b ehavior o f p oly(ethylene oxide)–poly(propylene o xide)–poly(ethylene o xide) triblock copolymers in water, Macromolecules, 28, 3807, 1995. 29. Calvo, P., Rem unan-Lopez, C., Vila-Jato, J .L., a nd Alonso, M.J., Chitosan and chitosan/ethylene oxide-propylene oxide

10/3/2008 8:49:05 PM

13-8

30.

31. 32. 33.

34.

35.

36. 37. 38.

39.

40.

41. 42.

43.

43722_C013.indd 8

Smart Materials

block copolymer nanoparticles as novel carriers for proteins and vaccines, Pharm. Res., 14, 1431, 1997. Shinkai, M.,Yanase, M., Honda, H.,Wakabayashi, T., Yoshida, J., and Kobayashi, T., Intracellular hyperthermia for cancer using ma gnetite ca tionic li posomes: In v itro study, Jpn. J . Cancer Res., 87, 1179, 1996. Jeong, B., Bae, Y.H., and Kim, S.W., The rmoreversible gelation of PEG-PLGA-PEG triblock copolymer aqueous solutions, Macromolecules, 32, 7064, 1999. Jeong, B., Bae, Y.H., Lee, D.S., and Kim, S.W., Biodegradable block co polymers a s i njectable d rug-delivery sy stems, Nature, 388, 860, 1997. Jeong, B ., L ee, D .S., S hon, J .-I., B ae, Y.H., a nd K im, S.W., Thermoreversible g elation o f p oly(ethylene o xide) b iodegradable p olyester blo ck co polymers, J. Polym. S ci. Polym. Chem., 37, 751, 1999. Lee, D.S., Shim, M.S., Kim, S.W., Lee, H., Park, I., and Chang, T., Novel thermo-reversible gelation of biodegradable PLGAblock-PEO-block-PLGA t riblock co polymers in aq ueous solution, Macromol. Rapid. Commun., 22, 587, 2001. Kuijpers, A.J., En gbers, G.H.M., F eijen, J., D e S medt, S.C., Meyvis, T.K.L., Demeester, J., Krijgsveld, J., Zaat, S.A.J., and Dankert, J ., Cha racterization o f the netw ork str ucture o f carbodiimide cr osslinked gela tin gels, Macromolecules, 32, 3325, 1999. Dentini, M., Desideri, P., Crescenzi, V., Yuguchi, Y., Urakawa, H., a nd K ajiwara, K., S olution a nd g elling p roperties o f gellan benzyl esters, Macromolecules, 32, 7109, 1999. Ramzi, M., Ro chas, C., a nd G uenet, J .-M., S tructure– properties r elation f or aga rose thermo reversible g els in binary solvents, Macromolecules, 31, 6106, 1998. Chen, T., Embree Heather, D., Wu, L.-Q., and Payne Gregory, F., In vitro protein–polysaccharide conjugation: Tyrosinasecatalyzed conjugation of gelatin and chitosan, Biopolymers, 64, 292, 2002. Chenite, A., Chaput, C., Wang, D., Combes, C., Buschmann, M.D., Hoemann, C.D., Leroux, J.C., Atkinson, B.L., Binette, F., and Selmani, A., Novel injectable neutral solutions of chitosan form biodegradable gels in situ, Biomaterials, 21, 2155, 2000. Nordby M arianne, H., K joniksen, A.-L., N ystrom, B ., a nd Roots, J., Thermoreversible gelation of aqueous mixtures of pectin a nd c hitosan. R heology, Biomacromolecules, 4, 337, 2003. Ruel-Gariepy, E. and L eroux, J.-C., In situ-forming hydrogels-review of temperature-sensitive systems, Eur. J. Pharm. Biopharm., 58, 409, 2004. Kawasaki, N., Ohkura, R., Miyazaki, S., Uno, Y., Sugimoto, S., and Attwood, D., Thermally reversible xyloglucan gels as vehicles f or o ral dr ug deliv ery, Int. J . Pha rm., 181, 227, 1999. Furgeson D arin, Y., Dr eher M atthew, R ., a nd Chilk oti, A., Structural optimization of a “smart” doxorubicin–polypeptide co njugate f or thermal ly t argeted deli very to s olid tumors, J. Control. Rel., 110, 362, 2006.

44. Meyer, D .E. a nd Chilk oti, A., Puri cation of re combinant proteins by fusion with thermally-responsive polypeptides, Nat. Biotechnol., 17, 1112, 1999. 45. Wang, C., S tewart, R .J., a nd Kopecek, J., Hybrid hydrogels assembled from synthetic polymers and coiled-coil protein domains, Nature, 397, 417, 1999. 46. Simon, S., Roy, D., and Schindler, M., Intracellular pH a nd the co ntrol o f m ultidrug r esistance, Proc. N atl. Acad. S ci. USA, 91, 1128, 1994. 47. Tannock, I.F . a nd Ro tin, D ., Acid pH in t umors a nd i ts potential for therapeutic exploitation, Cancer Res., 49, 4373, 1989. 48. Na, K. and Bae, Y.H., pH-sensitive polymers for drug delivery, in Polymeric Dr ug D elivery S ystems, Kwon, G.S., E d., Taylor & Francis Group, Boca Raton, FL, 2005, Chapter 3. 49. Pradip, Maltesh, C., S omasundaran, P., Kulkarni, R .A., and Gundiah, S., Polymer–polymer complexation in dilute aqueous s olutions: Poly(acrylic acid)–p oly(ethylene oxide) and poly(acrylic acid)–p oly(vinylpyrrolidone), Langmuir, 7, 2108, 1991. 50. Fujiwara, M., G rubbs, R .H., a nd B aldeschwieler, J .D., Characterization of pH-dependent poly(acrylic acid) complexation wi th phospholi pid v esicles, J. C olloid Interf. S ci., 85, 210, 1997. 51. Gemeinhart, R.A., Chen, J., Park, H., and Park, K., pH-sensitivity of f ast re sponsive s uperporous hydrogels, J. Biomter. Sci. Polym. ed., 11, 1371, 2000. 52. Taillefer, J., Brasseur, N.,Van Lier, J.E., Lenaerts, V., Le Garrec, D., and Leroux, J.C., In-vitro and in-vivo evaluation of pHresponsive p olymeric micel les in a pho todynamic ca ncer therapy model, J. Pharm. Pharmacol., 53, 155, 2001. 53. Gaur, U., Sahoo, S.K., De, T.K., Ghosh, P.C., Maitra, A., and Ghosh, P .K., B iodistribution o f  uoresceinated dextran using no vel na noparticles e vading r eticuloendothelial system, Int. J. Pharm., 202, 1, 2000. 54. Potineni, A., L ynn, D .M., L anger, R ., a nd Amiji, M.M., Poly(ethylene o xide)-modied poly(b-amino est er) na noparticles as a pH-s ensitive biodegradable system for paclitaxel delivery, J. Control. Release, 86, 223, 2003. 55. Chu, L.-Y., Yamaguchi, T., and Nakao, S., A molecular-recognition micr ocapsule f or en vironmental stim uli-responsive controlled release, Adv. Mater, 14, 386, 2002. 56. Kang, S.I. and Bae, Y.H., pH-induced volume-phase transition o f h ydrogels co ntaining su lfonamide side g roup b y reversible cr ystal f ormation, Macromolecules, 34, 8173, 2001. 57. Kang, S.I. and Bae, Y.H., pH-Induced solubility transition of sulfonamide-based polymers, J. Control. Rel., 80, 145, 2002. 58. Kim, G.M., B ae, Y.H., a nd J o, W.H., pH-ind uced micelle formation o f p oly(histidine-co-phenylalanine)-blockpoly(ethylene glycol) in aq ueous media, Macromol. Biosci., 5, 1118, 2005. 59. Lee, E.S., Na, K., and Bae, Y.H., Polymeric micelle for tumor pH a nd f olate-mediated t argeting, J. Co ntrol. Re l., 91, 103, 2003.

10/3/2008 8:49:05 PM

13-9

Drug Delivery Systems

60. Sethuraman, V.A., N a, K., a nd B ae, Y.H., pH-r esponsive sulfonamide/PEI syst em f or t umor sp ecic g ene deli very: An in vitro study, Biomacromolecules, 7, 64, 2006. 61. Shu, X.Z., Zhu, K.J., and Song, W., Novel pH-sensitive citrate cross-linked chitosan lm for drug controlled release, Int. J. Pharm., 212, 19, 2001. 62. Luo, L., Kato, M., Tsuruta, T., Kataoka, K., and Nagasaki, Y., Stimuli-sensitive p olymer g els tha t st iffen u pon sw elling, Macromolecules, 33, 4992, 2000. 63. Hisamitsu, I., K ataoka, K., Oka no, T ., a nd Sak urai, Y., Glucose-responsive g el f rom p henylborate p olymer a nd poly(vinyl alco hol): P rompt r esponse a t p hysiological pH through the interaction of borate with amino group in the gel, Pharm. Res., 14, 289, 1997. 64. Matsumoto, A., Yoshida, R ., a nd K ataoka, K., Gl ucoseresponsive polymer gel bearing phenylborate derivative as a glucose-sensing moiety operating at the p hysiological pH, Biomacromolecules, 5, 1038, 2004. 65. Chu, L.Y ., Li , Y., Zh u, J .H., Wang, H.D ., a nd Lia ng, Y.J., Control of pore size and permeability of a glucose-responsive gating membrane for insulin delivery, J. C ontrol. R el., 97, 43, 2004. 66. Chu, L.Y., Liang, Y.J., Chen, W.M., Ju, X.J., and Wang, H.D., Preparation o f g lucose-sensitive micr ocapsules wi th a porous mem brane a nd f unctional ga tes, Colloids S urf. B Biointerf., 37, 9, 2004. 67. Kang, S.I. a nd B ae, Y.H., A su lfonamide b ased g lucoseresponsive h ydrogel wi th co valently immob ilized g lucose oxidase and catalase, J. Control. Rel., 86, 115, 2003. 68. Obaidat, A.A. a nd P ark, K., Cha racterization o f p rotein release thr ough g lucose-sensitive h ydrogel mem branes, Biomaterials, 18, 801, 1997. 69. Sato, K., Kodama, D., and Anzai, J., Sugar-sensitive thin lms composed of concanavalin A and sugar-bearing polymers, Anal. Sci., 21, 1375, 2005. 70. Tanna, S., Saho ta, T.S., Sa wicka, K., a nd Taylor, M.J ., The effect o f deg ree o f acr ylic der ivatisation o n dext ran a nd concanavalin A g lucose-responsive ma terials f or c losedloop insulin delivery, Biomaterials, 27, 4498, 2006. 71. Miyata, T., J ikihara, A., a nd N akamae, K., P reparation o f poly(2-glucosyloxyethyl methacr ylate)–concanavalin A complex h ydrogel a nd i ts g lucose-sensitivity, Macromol. Chem. Phys., 197, 1135, 1996. 72. You, L.-C., L u, F .-Z., Li , Z.-C., Zha ng, W., a nd Li , F .-M., Glucose-sensitive a ggregates f ormed b y p oly(ethylene oxide)-block-poly(2-glucosyloxyethyl acrylate) with concanavalin A in dilute aqueous medium, Macromolecules, 36, 1, 2003. 73. Miyata, T., J ikihara, A., N akamae, K., a nd H offman, A.S., Preparation o f r eversibly g lucose-responsive h ydrogels b y covalent immob ilization o f le ctin in p olymer netw orks having p endant g lucose, J. B iomater. S ci. P olym. Ed ., 15, 1085, 2004. 74. Kakizawa, Y., H arada, A., a nd K ataoka, K., Gl utathionesensitive stabilization of block copolymer micelles composed

43722_C013.indd 9

75.

76.

77.

78. 79. 80. 81.

82.

83. 84.

85.

86. 87.

88.

of antisense DNA and thiolated poly(ethylene glycol)-blockpoly(l-l ysine): A p otential ca rrier f or syst emic deli very o f antisense DNA, Biomacromolecules, 2, 491, 2001. Oishi, M., H ayama, T., Akiyama, Y., Takae, S., H arada, A., Yamasaki, Y., N agatsugi, F ., S asaki, S ., N agasaki, Y., a nd Kataoka, K., Supramolecular assemblies for the cytoplasmic delivery of antisense oligodeoxynucleotide: polyion complex (PIC) micelles bas ed o n p oly(ethylene g lycol)-SS-oligodeoxynucleotide conjugate, Biomacromolecules, 6, 2449, 2005. Kakizawa, Y., Furukawa, S., and Kataoka, K., Block copolymer-coated calcium phosphate nanoparticles sensing intracellular environment for oligodeoxynucleotide and siRNA delivery, J. Control. Rel., 97, 345, 2004. El-Sayed, M.E., H offman, A.S., a nd S tayton, P.S., R ational design o f co mposition a nd ac tivity co rrelations f or pHresponsive and glutathione-reactive polymer therapeutics, J. Control. Rel., 104, 417, 2005. Lee, K.Y., Alsberg, E., a nd M ooney, D .J., D egradable a nd injectable po ly(aldehyde g uluronate) h ydrogels f or bo ne tissue engineering, J. Biomed. Mater. Res., 56, 228, 2001. Venkatraman, S. and Gale, R., Skin adhesives and skin adhesion. 1. Transdermal drug delivery systems, Biomaterials, 19, 1119, 1998. Osada, Y. a nd M atsuda, A., S hape memo ry in h ydrogels, Nature, 376, 219, 1995. Wache, H.M., Tartakowska, D.J., Hentrich, A., and Wagner, M.H., Development of a polymer stent with shape memory effect as a drug delivery system, J. Mater. Sci. Mater. Med., 14, 109, 2003. Okahata, Y., Hachiya, S., Ariga, K., and Seki, T., The electrical breakdown a nd p ermeability co ntrol o f a b ilayer-corked capsule membrane in an external electric eld, J. Am. Chem. Soc., 108, 2863, 1986. Kwon, I.C., B ae, Y.H., a nd K im, S.W., Ele ctrically er odible polymer gel for controlled release of drugs, Nature, 354, 291, 1991. Kagatani, S., Shinoda, T., Konno, Y., Fukui, M., Ohmura, T., and Osada, Y., Electroresponsive pulsatile depot delivery of insulin f rom p oly(dimethylaminopropylacrylamide) gel in rats, J. Pharm. Sci., 86, 1273, 1997. Agrawal, C.M., K ennedy, M.E., a nd Micallef, D .M., The effects o f ul trasound irradia tion o n a b iodegradable 50%–50% copolymer of polylactic and polyglycolic acids, J. Biomed. Mater. Res., 28, 851, 1994. Kost, J ., L eong, K., a nd L anger, R ., U ltrasound-enhanced polymer degrada tion a nd r elease o f inco rporated substances, Proc. Natl. Acad. Sci. USA, 86, 7663, 1989. Marin, A., Sun, H., Husseini, G.A., Pitt, W.G., Christensen, D.A., and Rapoport, N.Y., Drug delivery in pluronic micelles: Effect of hig h-frequency u ltrasound on dr ug release f rom micelles a nd in tracellular u ptake, J. Co ntrol. Re l., 84, 39, 2002. Lavon, I. and Kost, J., Mass transport enhancement by ultrasound in non-degradable polymeric controlled release systems, J. Control. Rel., 54, 1, 1998.

10/3/2008 8:49:05 PM

13-10

89. Gupta, A.K. and Curtis, A.S., Surface mo died superparamagnetic nanoparticles for drug delivery: Interaction studies wi th h uman  broblasts in c ulture, J. Ma ter. S ci. Ma ter. Med., 15, 493, 2004. 90. Diwan, M., Elamanchili, P., Lane, H., Gainer, A., and Samuel, J., Biodegradable nanoparticle mediated antigen delivery to human cord blood derived dendritic cells f or induction of primary T cell responses, J. Drug Target., 11, 495, 2003. 91. Hsieh, D.S., Langer, R., and Folkman, J., Magnetic modulation of re lease of m acromolecules f rom p olymers, Proc. Natl. Acad. Sci. USA, 78, 1863, 1981. 92. Okahata, Y., Ariga, K., and Seki, T., Redox-sensitive permeation from a capsule membrane graĀed with viologen-containing polymers, J. Chem. Soc. Chem. Commun., 73, 1988. 93. Napoli, A., Valentini, M., Tirelli, N., Muller, M., and Hubbell, J.A., Oxida tion-responsive p olymeric v esicles, Nat. Ma ter., 3, 183, 2004. 94. Shimoboji, T., Din g, Z.L., S tayton, P.S., a nd H offman, A.S., Photoswitching of ligand association with a photoresponsive polymer–protein conjugate, Bioconjug. Chem., 13, 915, 2002. 95. Tao, X., Li, J., and Mohwald H., Self-assembly, optical behavior, and permeability of a novel capsule based on an azo dye and polyelectrolytes, Chemistry, 10, 3397, 2004. 96. Veronese, F .M., S chiavon, O ., P asut, G., M endichi, R ., Andersson, L., Tsirk, A., Ford, J., Wu, G., Kneller, S., Davies, J., and Duncan, R., PEG-doxorubicin conjugates: Inuence of p olymer structure on dr ug release, in v itro c ytotoxicity, biodistribution, a nd a ntitumor ac tivity, Bioconjug. Chem ., 16, 775, 2005. 97. Kovar, M., Kovar, L., Subr, V., Etrych, T., Ulbrich, K., Mrkvan, T., Loucka, J., and Rihova, B., HPMA copolymers containing doxorubicin b ound b y a p roteolytically o r h ydrolytically cleavable bond: comparison of biological properties in vitro, J. Control. Rel., 99, 301, 2004. 98. L utolf, M.P. a nd H ubbell, J .A., S ynthetic b iomaterials as instructive ext racellular micr oenvironments f or mo rphogenesis in tissue engineering, Nat. Biotechnol., 23, 47, 2005. 99. Chu, L.-Y., Yamaguchi, T., and Nakao, S., A molecular-recognition micr ocapsule f or en vironmental stim uli-responsive controlled release, Adv. Mater., 14, 386, 2002. 100. Desai, M.P., L abhasetwar, V., Amidon, G.L., a nd L evy, R .J., Gastrointestinal u ptake o f b iodegradable micr oparticles: effect of particle size, Pharm. Res., 13, 1838, 1996. 101. Gupta, R.B., Fundamentals of Drug Nanoparticles, Taylor & Francis Group, New York, 2006. 102. Nasongkla, N., S huai, X., Ai, H., Weinberg, B .D., P ink, J ., Boothman, D.A., and Gao, J., Bioorganic chemistry: cRGDfunctionalized p olymer micelles f or t argeted do xorubicin delivery, Angew Chem. Int. Ed., 43, 6323, 2004. 103. Kurihara, A., Deguchi, Y., and Pardridge, W.M., Epidermal growth fac tor radio pharmaceuticals: I n c helation, co njugation t o a b lood–brain ba rrier deli very v ector via a biotin–polyethylene linker, pharacokinetics, and in vivo imaging o f exp erimental b rain t umors, Bioconjugate Chem., 10, 502, 1999.

43722_C013.indd 10

Smart Materials

104. Kao, G.Y., Chang, L.-J., and Allen, T.M., Use of targeted cationic liposomes in en hanced DNA delivery to cancer cells, Cancer Gene Ā er ., 3, 250, 1996. 105. P ing, P., Wang, W., Chen, X., a nd J ing, X., P oly(epsiloncaprolactone) p olyurethane a nd i ts sha pe-memory p roperty, Biomacromolecules, 6, 587, 2005. 106. Salmaso, S., Semenzato, A., Caliceti, P., Hoebeke, J., Sonvico, F., D ubernet, C ., a nd C ouvreur, P., S pecic a ntitumor t argetable b-cyclodextrin-poly(ethylene glycol)-folic acid drug delivery bioconjugate, Bioconjug. Chem., 15, 997, 2004. 107. Broz, P., Benito, S.M., Saw, C., Burger, P., Heider, H., Psterer, M., Marsch, S., Meier, W., and Hunziker, P., Cell targeting by a g eneric r eceptor-targeted p olymer na nocontainer p latform, J. Control. Rel., 102, 475, 2005. 108. Soppimath, K.S., Tan, D.C.-W., and Yang, Y.-Y., pH-Triggered thermally r esponsive p olymer co re-shell na noparticles f or drug delivery, Adv. Mat., 17, 318, 2005. 109. Xu, S., Kraemer, M., and Haag, R., pH-responsive dendritic core-shell a rchitectures as a mphiphilic na nocarriers f or polar drugs, J. Drug Target, 14, 367, 2006. 110. Sharman, W.M., va n Lier , J .E., a nd Allen, C.M., T argeted photodynamic therapy via receptor mediated delivery systems, Adv. Drug. Deliv. Rev., 56, 53, 2004.

13.2

Drug Delivery Systems: Smart Polymers

Joseph Kost 13.2.1

Introduction

The rapid advancement of biomedical research has led to m any creative app lications f or b iocompatible p olymers. A s mo dern medicine discerns more mechanisms, both of physiology and of pathophysiology, the approach to healing is to mimic, or if possible, to recreate the physiology of healthy functioning. Thu s, the area of smart polymers for responsive drug delivery has evolved. The developments fall under two categories: externally regulated or p ulsatile s ystems (also k nown a s “ open-loop” s ystems) a nd self-regulated s ystems ( also k nown a s “ closed-loop”). The following chapter outlines the fundamentals of this research area.

13.2.2

Development of Controlled Drug Delivery Systems

13.2.2.1

Control of Drug Concentration Levels Over Time

While newer and more powerful drugs continue to be developed, increasing at tention is being g iven to t he methods of ad ministering these active substances. In conventional drug delivery, the drug c oncentration i n t he blo od r ises w hen t he d rug i s t aken, then peaks, and declines. Maintaining drug in the desired therapeutic r ange w ith just a si ngle dose, or t argeting t he d rug to a

10/3/2008 8:49:05 PM

13-11

Drug Delivery Systems

specic a rea ( lowering t he s ystemic d rug le vel), a re go als t hat have been successfully attained with commercially available controlled release devices [1,2]. However, there are many clinical situations where the approach of a constant drug delivery rate is insufficient, such as the delivery of insulin for patients with diabetes mel litus, a ntiarrhythmics for pat ients w ith heart rhythm disorders, ga stric ac id i nhibitors f or u lcer c ontrol, n itrates f or patients with angina pectoris, as well as selective β-blockade, birth control, general hormone replacement, immunization, and cancer chemotherapy. Furthermore, studies in the eld of chronopharmacology i ndicate t hat t he o nset o f c ertain d iseases exhibit strong circadian temporal dependence. Thu s, treatment of these diseases could be optimized through the use of responsive (smart) delivery systems [3,4], which are, in essence, manmade imitations of healthy functioning. 13.2.2.2 Classification of “Smart” Polymers Smart controlled drug release devices can be classied as open or c losed lo op s ystems. O pen-loop c ontrol s ystems a re t hose where information about the controlled variable is not automatically u sed to ad just t he s ystem i nputs to c ompensate f or t he change in the process variables. In the controlled drug delivery eld, op en-loop s ystems a re k nown a s p ulsatile o r e xternally regulated. The externally controlled devices apply external triggers for pulsatile del ivery suc h a s u ltrasonic, t hermal, ele ctric, light, and chemical or biochemical agents. Closed-loop control systems, on the other hand, are dened as systems where the controlled variable is detected, and as a result, the system output is adjusted accordingly. In the controlled drug delivery eld, closed-loop systems are known as self-regulated, in which t he rele ase r ate i s c ontrolled b y f eedback i nformation, without a ny e xternal i ntervention. The s elf-regulated s ystems utilize several approaches for the rate control mechanisms such as: p H-sensitive p olymers, en zyme–substrate re actions, p Hsensitive drug solubility, competitive binding, antibody interactions, and metal concentration-dependent hydrolysis [5–7]. Many app roaches f or m imicking t he ph ysiological h ealthy state a re u ndergoing re search. The f ocus o f t his c hapter i s o n smart polymers; therefore, other important areas, such as using pumps for controlled drug delivery, liposomes, microencapsulation of living cells or gene therapy, are not covered. An additional area of signicant research that is not covered in this chapter is site-directed or targeted drug delivery, where the release is constant (as in chemotherapy for cancer treatment). Here, the fundamental principles and recent advances of responsive drug delivery for both pulsatile and self-regulated systems are reviewed.

rates were enhanced when samples were exposed to ultrasound. The system’s response to the ultrasonic triggering was rapid and reversible. The rele asing a gents, p-nitroaniline, p-aminohippurate, B SA, a nd i nsulin, w ere te sted f or i ntegrity f ollowing exposure to u ltrasonic energ y, a nd w ere f ound to b e i ntact. Enhanced drug release rate was also observed when nonerodible polymeric systems (where drug release rate is diffusion dependent) were exposed to ultrasound [8]. When d iabetic rats, receiving nonerodible hydrophilic polymer implants containing insulin, were exposed to ultrasound, a sharp d rop i n blood g lucose levels was observed a Āer t he i rradiation. This i ndication o f a r apid rele ase o f i nsulin i n t he implanted site has suggested the feasibility of ultrasound mediated drug delivery [8,9]. Recently ultrasound was also applied, in the clinic, for the enhancement of drugs transport through the skin [10]. 13.2.3.2

Electrically Stimulated Systems

Electrically controlled systems provide drug release by the application of electric eld o n a b iodegradable p olymer [11], r atelimiting membrane/hydrogel [12], and/or directly on the solute [13]. The ele ctrophoretic m igration o f a c harged m acrosolute within a hydrated membrane results from the combined response to the electrical forces on the solute and its associated counterions in the adjacent electrolyte solution [13]. Approaches involving microelectromechanical systems (MEMS) [6] or microchips are b eing de veloped. F or e xample, c arbon na notubes [ 14] o r polymeric b iodegradable m icrochips [15], c ontaining na nosize drug re servoirs, m ight b e a ble to del iver ph armaceuticals f or long time periods in a controlled manner by electrically opening the c aps o f t hese re servoirs. E lectrically c ontrolled memb rane permeability has been also of interest in the eld of electrically controlled or enhanced t ransdermal d rug delivery (e.g., iontophoresis, electroporation) [16]. 13.2.3.3

Photostimulated Systems

Photoinduced phase transition of gels [17] and polymers degradation [18] have been also proposed for temporal drug delivery. Photoresponsive gels reversibly change their physical or chemical prop erties up on photor adiation, t hus a ffecting t he rele ase rate o f d rugs i ncorporated i n t he p olymer. A ph otoresponsive polymer c onsists o f a ph otoreceptor, u sually a ph otochromic chromophore, a nd a f unctional pa rt. The opt ical sig nal i s c aptured by t he photochromic mole cules a nd t hen t he i somerization of t he chromophores i n t he photoreceptor converts it to a chemical signal. 13.2.3.4 Calcium-Stimulated Systems

13.2.3

Pulsatile Systems

13.2.3.1 Ultrasonically Stimulated Systems Release rates of substances can be repeatedly modulated at w ill from a p osition e xternal to t he del ivery s ystem b y u ltrasonic irradiation [ 8]. W hen b ioerodible p olymers w ere u sed a s t he drug carrier matrices, both the polymer erosion and drug release

43722_C013.indd 11

An e xample f or t hese i s a c alcium-responsive b iodegradable drug del ivery s ystem [ 19], w here a no nactive en zyme ( alpha amylase) is incorporated into a matrix composed of its substrate (starch). The s ystem re sponds to a t rigger mole cule ( calcium) that causes the nonactive enzyme to become active. The matrix then starts to degrade, resulting in release of the drug incorporated in the matrix.

10/3/2008 8:49:06 PM

13-12

13.2.4

Smart Materials

Self-Regulated Systems (Environmentally Responsive Systems)

Polymers that alter their characteristics in response to c hanges in their environment have been of great recent interest. Several research g roups ha ve b een de veloping d rug del ivery s ystems, based on these responsive polymers, that more closely resemble the no rmal ph ysiological p rocess [ 3,4]. I n t hese de vices, d rug delivery i s re gulated b y me ans o f a n i nteraction w ith t he su rrounding en vironment ( feedback i nformation) w ithout a ny external i ntervention. The m ost co mmonly st udied po lymers, having environmental sensitivity, are either pH- or temperaturesensitive. There are also inammation-sensitive systems and systems utilizing enzymes and specic binding interactions, which will be discussed as well. 13.2.4.1 Temperature-Sensitive Systems Temperature-sensitive polymers are based on polymer–water interactions, e specially s pecific h ydrophobic–hydrophilic balancing effects, and the conguration of side groups. When polymer networks swell in a s olvent, there is usually a ne gligible or small positive enthalpy of mixing or dilution. Although a positive enthalpy change opposes the process, the large gain in the entropy drives it. In aqueous polymer solutions, the opposite is oĀen observed. Th is unusual behavior is associated with a phenomenon of polymer phase separation as the temperature is r aised to a c ritical v alue, k nown a s t he L CST. P olymers characterized b y L CST u sually s hrink, a s t he tem perature i s increased through the LCST. Lowering the temperature below LCST results in the swelling of the polymer. Bioactive agents such as drugs, enzymes, antibodies, and genes may be immobilized on or w ithin t he temperature sensitive polymers. Responsive drug/genes release patterns regulated by environmental temperature changes have been recently demonstrated by several groups [4,20,21].

exhibit p H-dependent s welling b ehavior d ue to t he re versible formation–dissociation o f i nterpolymer co mplexes. Dr ug delivery systems based on this copolymer showed pH-dependent drug delivery rates. Heller and Trescony [23] were the  rst to p ropose the use of pH-sensitive bioerodible polymers. In their approach, described in t he s ection o n s ystems ut ilizing en zymes, a n en zyme– substrate reaction produces a pH change that is used to modulate the erosion of pH-sensitive polymer containing a dispersed therapeutic agent. 13.2.4.3

Inflammation-Responsive Systems

Inammation-responsive drug delivery systems are based on biodegradable hydrogels t hat a re s pecically degraded by hydroxyl radicals or enzymes produces at inammatory sites [24]. The radicals o r s pecic en zymes c ause de gradation o f t he p olymer a nd thus r elease of t he bioactive a gent, ma king t his del ivery s ystem potentially useful for treatment of osteoarthritis. 13.2.4.4 Systems Utilizing Antibody Interactions Pitt [25] proposed utilizing morphine (hapten)–antibody interactions to supp ress enzymatic degradation and permeability of polymeric reservoirs or matrix drug delivery systems. The delivery device consists of t he drug na ltrexone contained in a p olymeric reservoir or dispersed in a polymeric matrix conguration. The device is coated by covalently graĀing morphine to the surface. Exposure of the graĀed surface to antibodies to morphine results in coating of the surface by the antibodies, a process that can be reversed by exposure to exogenous morphine. The presence of the antibodies on the surface, or in the pores of the delivery device, will block or impede the permeability of naltrexone in a reservoir conguration or enzyme-catalyzed surface degradation and concomitant release of the drug from a matrix device. The re versible binding of a ntigen to a ntibody s erve a lso a s t he basis for hydrogel swelling that could lead to rele ase of a bioactive agent [26,27].

13.2.4.2 pH-Sensitive Systems

13.2.4.5

Systems Utilizing Enzymes

The pH r ange of f luids i n v arious s egments of t he GI T m ay provide environmental stimuli for responsive drug release. Studies by several research groups [3,20,21] have been performed on polymers containing weakly acidic or basic groups in the polymeric backbone. The charge density of the polymers depends on pH and ionic composition of t he outer solution (the solution i nto which the polymer is exposed). Altering the pH of the solution will cause swelling or deswelling of the polymer. Polyacidic polymers will be unswollen at lo w pH, si nce t he acidic g roups w ill be protonated and hence u nionized. With i ncreasing pH, p olyacidic p olymers will swell. The opp osite h olds f or p olybasic polymers, since the ionization of the basic groups will increase with decreasing pH. Thus, d rug rele ase f rom de vices m ade f rom t hese p olymers w ill display pH-dependent release rates. Lowman and Peppas [22] proposed pH-responsive drug delivery ba sed o n a c omplexation–decomplexation me chanism. Copolymer networks of PMAA graĀed with poly(ethylene glycol)

13.2.4.5.1

Urea-Responsive Delivery

43722_C013.indd 12

Heller and Trescony [23] were the rst to attempt using immobilized enzymes to alter local pH and thus cause changes in polymer erosion rates. The proposed system is based on the conversion of urea to NH4HCO3 and NH4OH by the action of urease. As this reaction c auses a p H i ncrease, a p olymer t hat i s sub jected to increased erosion at high pH is required. 13.2.4.5.2

Morphine Triggered Naltrexone Delivery System

Heller [28] also proposed a biodegradable drug delivery system that would be passive until the appearance of morphine external to t he de vice. A ctivation o f rele ase i s ba sed o n t he re versible inactivation of enzymes achieved by the covalent attachment of hapten, close to the active site of the enzyme–hapten conjugate, with the hapten antibody. Because the antibodies are large molecules, access of the substrate to the enzyme’s active site is sterically inhibited, thus effectively rendering the enzyme inactive.

10/3/2008 8:49:06 PM

13-13

Drug Delivery Systems

Triggering of drug (naltrexone) release is initiated by the appearance of morphine (hapten) in t he t issue, causing d issociation of the enzyme–heptan–antibody complex, rendering the enzyme active and therefore release of naltrexone. Many of the glucose-responsive s ystems d iscussed i n t he ne xt s ection a lso utilize enzymes. 13.2.4.6

Glucose-Responsive Insulin Delivery

The present modes of treatment, including insulin pumps, not fully mimic t he physiology of insulin secretion. Ther efore the development of a sma rt insulin-delivery system could signicantly help patients with diabetes in controlling their blood glucose level, thus avoiding the various severe diabetes complications including eye disease, gangrene of the extremities, cardiovascular d isease, a nd rena l f ailure. The d evelopment of glucose-sensitive insulin-delivery systems has made use of several approaches, including polymer–complex system, immobilized GOx in pH-sensitive polymers, and competitive binding. 13.2.4.6.1 Polymer Complex System Kitano et al . [ 29] p roposed a g lucose s ensitive i nsulin rele ase system based on a sol–gel transition. A phenylboronic acid (PBA) moiety wa s i ncorporated i nto po ly(N-vinyl-2-pyrrolidone) (poly(NVP-co-PBA)). I nsulin w as i ncorporated i nto a p olymer gel formed by a complex of poly(vinyl alcohol) with poly(NVPco-PBA). P BA c an for m re versible c ovalent c omplexes w ith molecules having d iol u nits, such as g lucose or PVA. With t he addition o f g lucose, PV A i n t he PV A–boronate c omplex i s replaced by glucose. This leads to a transformation of the system from gel to s ol state, facilitating the release of insulin from the polymeric complex. 13.2.4.6.2

Competitive Binding

The basic principle of competitive binding and its application to c ontrolled d rug del ivery w as  rst p resented b y Bro wnlee and C erami [30]. They su ggested t he p reparation o f b iologically active glycosylated insulins, which are complementary to the m ajor c ombining si te o f c arbohydrate b inding p roteins such a s C oncavalin A ( Con A ). C on A i s i mmobilized o n sepharose b eads. The g lycosylated i nsulin i s d isplaced f rom the Con A by glucose in response to, and proportional to, the amount of glucose p resent, w hich c ompetes f or t he s ame binding si tes. K im et a l. [31] f ound t hat t he rele ase r ate o f glycosylated i nsulin a lso de pends o n t he b inding a ffi nity of the insulin derivative to t he Con A a nd can be inuenced by the c hoice o f s accharide g roup i n t he g lycosylated i nsulin. By encapsulating the glycosylated insulin-bound Con A w ith a su itable p olymer t hat i s p ermeable to g lucose a nd i nsulin, the g lucose i nf lux a nd i nsulin e ff lux w ould b e c ontrolled by the encapsulation membrane. Similarly, hydrogels or membranes formed by mixing Con A with copolymers containing allyl glucose will undergo a reversible sol–gel phase transition in the presence of glucose due to the competitive binding between the free glucose and Con A (Con A acts as a cross-linker). The free glucose would replace the Con A

43722_C013.indd 13

and t herefore d isrupt t he c ross-links a nd m ake t he p olymer more permeable to insulin [32]. 13.2.4.6.3 Immobilized GOx in pH Sensitive Polymers Responsive d rug d elivery s ystems b ased on p H-sensitive p olymers have been developed along three different approaches: pHdependent swelling, degradation, and solubility. 13.2.4.6.3.1 pH-Dependent Solubility Gl ucose-dependent insulin release was proposed by Langer and coworkers [33,34] based on the fact that insulin solubility is pH-dependent. Insulin was incorporated i nto e thylene v inyl ac etate ( EVAc) copolymer matrices in solid form. GOx was immobilized to sepharose beads, which w ere i ncorporated a long w ith i nsulin i nto t he E VAc matrices. W hen g lucose entered t he matrix, t he produced g luconic ac id c aused a r ise i n i nsulin s olubility a nd c onsequently enhanced release. 13.2.4.6.3.2 pH-Dependent Degradation Heller [28] s uggested a system in which insulin is immobilized in a pH-sensitive bioerodible polymer, which is surrounded by a hydrogel containing immobilized GOx. When glucose diffuses into the hydrogel and is o xidized to g luconic ac id, t he re sultant lo wered p H t riggers enhanced polymer degradation and release of insulin from the polymer in proportion to the concentration of glucose. 13.2.4.6.3.3 pH-Dependent Swelling Systems ba sed o n p Hsensitive polymers consist of immobilized GOx in a pH-responsive h ydrogel, en closing a s aturated i nsulin s olution o r incorporated with insulin were  rst proposed by Horbett et al. [35] and Ishira et al. [36]. As glucose diff uses into the hydrogel, GOx catalyzes its conversion to gluconic acid, thereby lowering the p H i n t he m icroenvironment o f t he h ydrogel a nd c ausing swelling. Si nce i nsulin s hould p ermeate t he s welled h ydrogel more rapidly, faster delivery of insulin in the presence of glucose is anticipated. As the glucose concentration decreases in response to the released insulin, the hydrogel should contract and decrease the r ate o f i nsulin del ivery. U p to no w, s everal p H-responsive hydrogels h ave b een de vised [37]. However, none of these systems could f ully m imic t he physiology of i nsulin secretion as yet. F or g lucose-responsive s elf-regulated i nsulin rele ase s ystems, b iocompatibility, b ioavailability, lo ng-term s tability a nd system response kinetics are essential to ensure safety and efficacy [38–43].

13.2.5

Concluding Remarks

The approaches discussed represent attempts conducted over the past t hree de cades to ac hieve re sponsive rele ase. I t s hould b e pointed out that these drug delivery systems are still in the developmental stage and much research will have to be conducted for such s ystems to b ecome p ractical c linical a lternatives. Cr itical considerations a re t he biocompatibility a nd toxicology of t hese multicomponent polymer-based systems, the response times of these systems to stimuli, the ability to provide practical levels of the desired drug, and addressing necessary formulation issues in

10/3/2008 8:49:06 PM

13-14

dosage or design (e.g., shelf life, sterilization, reproducibility). A key i ssue i n t he p ractical ut ilization o f t he pulsatile, e xternally triggered s ystems ( i.e., u ltrasound, ele ctrically re gulated, a nd photoresponsive) will be the design of small portable trigger units that t he pat ient c an e asily u se. I deally, suc h s ystems c ould b e worn by the patient, such as a wristwatch-like system, and it could be either preprogrammed to go on and off at specic times or the patient could turn it on when needed. A critical issue in the development of responsive, self-regulated systems, such as those containing en zymes o r a ntibodies, i s t he s tability a nd/or p otential leakage a nd p ossible i mmunogenicity of t hese bioactive a gents. The successful development of responsive polymer delivery systems w ill be a sig nicant challenge. Nevertheless, t he considerable ph armacological b enet t hese s ystems c ould p otentially provide, pa rticularly g iven o ngoing re search i n b iotechnology, pharmacology and medicine, which may provide new insights on the de sirability a nd re quirements f or p ulsatile rele ase, s hould make this an important and fruitful area for future research.

References 1. Langer, R., Where a pill won’t reach, Sci. Am., 288, 50, 2003. 2. Langer, R ., Dr ug deli very a nd t argeting, Nature, 392, 5, 1998. 3. Kost, J. and Langer, R., Responsive polymeric delivery systems, Adv. Drug Deliver. Rev., 46, 125, 2001. 4. Kopecek, J., Smart and genetically engineered biomaterials and drug delivery systems, Eur. J. Pharm. Sci., 20, 1, 2003. 5. Goldbart, R., Traitel, T., Lapidot, S., and Kost, J., Enzymatically controlled responsive drug delivery systems, Polym. Advan. Technol., 13, 1006, 2002. 6. Galaev, I.Y. and Mattiasson, B., “Smart” polymers and what they co uld do in B iotechnology a nd M edicine, Trends Biotechnol., 17, 335, 1999. 7. Kaetsu, I. et al., Intelligent type controlled release systems by radiation techniques, Radiat. Phys. Chem., 55, 193, 1999. 8. Kost, J ., L eong, K., a nd L anger, R ., U ltrasound-enhanced polymer degrada tion a nd r elease o f inco rporated substances, Proc. Natl. Acad. Sci., 86, 7663, 1989. 9. Miyazaki, S., Yokouchi, C., and Takada, M., External control of drug release: controlled release of insulin from a hydrophilic p olymer im plants b y u ltrasound irradia tion in diabetic rats, J. Pharm. Pharmacol., 40, 716, 1988. 10. Mitragotri, S. and Kost, J., Low-Frequency Sonophoresis: A Review, Adv. Drug Deliver. Rev., 56, 589, 2004. 11. Kwon, I.C., B ae, Y.H., a nd K im, S.W., Ele ctrically er odible polymer gel for controlled release of drugs, Nature, 354, 291, 1991. 12. Madou, M.J. and He, K.Q., Exploitation of a novel articial muscle for controlled drug delivery, Polym. Mater. Sci. Eng., 83, 405, 2000. 13. Grodzinsky, A.J. and Grimshaw, P. E., Electrically and chemically controlled hydrogels for drug delivery, in Pulsed and self r egulated dr ug d elivery, K ost, J., E d., CR C P ress, B oca Raton, 1999, 47.

43722_C013.indd 14

Smart Materials

14. Bianco, A., K ostarelos, K., a nd P rato, M., Applications o f carbon nanotubes in drug delivery, Curr. Opin. Chem. Biol., 9, 674, 2005. 15. Grayson, A. et al., Multiple-pulse drug delivery from a resorbable polymeric microchip device, Nat. mater., 2, 767, 2003. 16. Prausnitz, M.R., Mitragotri, S., and Langer, R., Current status and future potential of transdermal drug delivery, Nat. Rev. Drug Discov., 3, 115, 2004. 17. Mamada, A. et a l., Ph otoinduced ph ase t ransition of ge ls, Macromolecules, 23, 1517, 1990. 18. Yui, N., Okano, T., and Sakurai, Y., Photo-responsive degradation of heterogeneous hydrogels comprising crosslinked hyaluronic acid a nd lipid microspheres for temporal dr ug delivery, J. Control. Release, 26, 141, 1993. 19. Goldbart, R . a nd K ost, J ., C alcium r esponsive b ioerodible drug delivery systems, Pharm. Res., 16, 1483, 1999. 20. Alexander, C., Temperature- and pH-responsive smart polymers for gene delivery, Expert Opin. Drug Deliver., 3, 573, 2006. 21. de Las Heras Alarcam, C., Pennadam, S., and Alexander, C., Stimuli r esponsive p olymers f or b iomedical a pplications, Chem. Soc. Rev., 3, 276, 2005. 22. Lowman, A.M. a nd P eppas, N.A., Puls atile dr ug deli very based o n a co mplexation-decomplexation me chanism, in Intelligent M aterials f or Co ntrolled Re lease, Dinh, S., DeNuzzio, J., and Comfort, A., Eds., ACS Symposium Series 728, American Chemical Society, 30, 1999. 23. Heller, J. and Trescony, P.V., Controlled drug release by polymer dissolution II. An enzyme mediated delivery system, J. Pharm. Sci., 68, 919, 1979. 24. Yui, N., Okano, T., and Sakurai, Y., Inammation responsive degradation o f cr osslinked h yaluronic acid gel, J. Co ntrol. Release, 22, 105, 1992. 25. Pitt, C.G., S elf-regulated a nd trig gered dr ug deli very systems, Pharm. Int., 88, 91, 1986. 26. Miyata, T., Asami, N., and Uragami, T., A reversibly antigenresponsive hydrogel, Nature, 24, 766, 1999. 27. Nakayama, G.R., Roskos, K., Fritzinger, B., and Heller, J., A study of reversibly inactivated lipases for use in morphinetriggered naltrexone delivery system, J. Biomed. Mater. Res., 29, 1389, 1995. 28. Heller, J., Feedback-controlled dr ug delivery, in Controlled drug de livery: c hallenges a nd str ategies, P ark, K., E d., American Chemical Society, Washington, DC, 1997,127. 29. Kitano, S., Kitano, S., Koyama, Y., Kataoka, K., Okanu, T., and Sakurai, Y., A novel drug delivery system utilizing a glucose responsive p olymer co mplex b etween p oly(vinyl alco hol) and p oly(N-vinyl-2-pyrrolidone) wi th p henylboronic acid moiety, J. Control. Release, 19, 161, 1992. 30. Brownlee, M. and Cerami, A., A glucose-controlled insulindelivery system: s emisynthetic in sulin b ound to le ctin, Science, 26, 1190, 1979. 31. Kim, S.W., Pai, C.M., Makino, L.A., Seminoff, L.A., Holmberg, D.L., Gle eson, J .M., Wilson, D .E., a nd M ack, E.J ., S elf regulated g lycosylated in sulin deli very, J. Co ntrol. Re lease, 11, 193, 1990.

10/3/2008 8:49:06 PM

Drug Delivery Systems

32. Kim, J.J. and Park, K., Modulated insulin delivery from glucose sensitive hydrogel dosage forms, J. Control. Release, 77, 29, 2001. 33. Brown, L., Edelman, E., Fischel-Ghodsian, F., and Langer, R., Characterization of glucose-mediated insulin release from implantable polymers, J. Pharm. Sci., 85, 1341, 1996. 34. Fischel-Ghodsian, F., Brown, L., Mahiowitz, E., Branderburg, D., and Langer, R., Enzymatically controlled drug delivery, Proc. Natl. Acad. Sci., 85, 2403, 1988. 35. Horbett, T., Kost, J., and Ratner, B., Swelling behavior of glucose s ensitive mem branes, Am. C hem. S oc., D iv. P olym. Chem., 24, 34, 1983. 36. Ishihara, K., Kobayashi, M., a nd S honohara, I., C ontrol o f insulin p ermeation thr ough a p olymer mem brane wi th responsive f unction for g lucose, Macromol. Ch em., Ra pid Commun., 4, 327, 1983. 37. Kost, J ., I ntelligent Dr ug D elivery S ystems in Ā e Encyclopedia o f C ontrolled D rug D elivery, M athiowitz, E. Ed., John Wiley & Sons, NY, 2000, 445.

43722_C013.indd 15

13-15

38. Podual, K., Doyle, F.J., and Peppas, N.A., Glucose-sensitivity of g lucose o xidase-containing ca tionic co polymer h ydrogels having poly(ethylene glycol) graĀs, J. Control. Release, 67, 9, 2000. 39. Chu, L.Y. et al., Control of pore size and permeability of a glucose-responsive gating membrane for insulin delivery, J. Control. Release, 97, 43, 2004. 40. Misra, G. and Siegel, R ., L ong term autonomous rhythmic hormone r elease acr oss h ydrogel mem brane, J. Co ntrol. Release, 81, 1, 2002. 41. Traitel, T., C ohen, Y., and Kost, J., Characterization of glucose-sensitive in sulin r elease system s in sim ulated in v ivo conditions, Biomaterials, 21, 1679, 2000. 42. Matsumoto, A., Yoshida, R ., a nd K ataoka, K., Gl ucoseresponsive polymer gel bearing phenylborate derivative as a glucose-sensing moiety operating at the p hysiological pH, Biomacromolecules, 3, 1038, 2004. 43. Peppas, N.A., Is there a future in glucose-sensitive, responsive insulin delivery systems? STP, Pharma. Sci., 14, 247, 2004.

10/3/2008 8:49:06 PM

43722_C013.indd 16

10/3/2008 8:49:07 PM

14 Fiber Optic Systems: Optical Fiber Sensor Technology and Windows 14.1

Introduction and Application of Fiber Optic Sensors ......................................................14-1 Introduction • Fi ber Optic Sensors • O ptical Fiber Modulation Mechanisms • E merging Optical Fiber Concepts • Selected Applications of Fiber Optic Sensors • Conclusions

References ......................................................................................................................................... 14-14 14.2 S mart Windows .................................................................................................................... 14-17 Outline • I ntroduction • A rchitectural Glazing Applications for Smart Windows • Su rvey of Smart Windows • Electrochromic Smart Windows • Future Directions

References .........................................................................................................................................14-30

14.1 Introduction and Application of Fiber Optic Sensors Nezih Mrad and Henry C.H. Li 14.1.1 Introduction Recent advances in sensors, microelectronics, and adaptive signal p rocessing te chnologies h ave sig nicantly s haped t he fundamental approaches to de aling with traditional problems. Imagine a w orld w here me chanical s ystems ac hieve opt imum functionality dem onstrated b y b iological s ystems such a s t he human b rain a nd b ody. Ā ese s ystems t hat a re k nown a s smart, adaptive, a nd i ntelligent c ould m imic human muscular, sensory, a nd ner vous s ystems b y em ploying emb edded sensors, actuators, advanced signal processing and control capabilities. Micro and nanosensors act as the system’s nerve endings sending sig nals to t he p rocessor ( brain), w hich i n t urn s ends signals to the actuator that take on the role of responsive or adaptive m uscles, le ading to sm art a nd ad aptive re sponse. Furthermore, d riven b y i ncreased s afety a nd rel iability at reduced cost and enhanced performance, key technologies such as e lectrorheological  uids, s hape memo ry a lloys, p iezoelectric materials, m agnetostrictive an d e lectrostrictive m aterials, an d micro a nd na noelectromechanical devices (MEMS/NEMS) are identied to play a major role in the development of such adaptive systems. One main experienced limitation is the integration

of sensory nodes and sensor networks within a complex system. Among t he few s ensors su ited for i ncorporation i nto materials and structures and for the realization of such adaptive and smart systems, optical bers are considered to lead the pack in sensor technology. I f suc cessfully i mplemented, t his te chnology i s posed to revolutionize the current approaches to systems’ health monitoring a nd m anagement, p erformance a ssessment, a nd implementation of adaptive systems and structures. Compared to more traditional measurement techniques, ber optic s ensors off er unique capabilities such as monitoring the manufacturing p rocess o f c omposite a nd me tallic pa rts, performing nondestructive testing once fabrication is complete, enabling s tructural a nd c omponent h ealth mo nitoring f or prognostics h ealth m anagement, a nd s tructural c ontrol f or component life extension. Because of their very low weight, small size, h igh ba ndwidth, a nd i mmunity to ele ctromagnetic a nd radio f requency i nterferences,  ber opt ic s ensors h ave sig nicant performance advantages over traditional sensors. In fact, in recent ye ars, opt ical  bers a re stead ily beco ming m ore costeffective due to recent advances in telecommunications, optoelectronic, a nd m icro a nd na no i ndustries. I n c ontrast to classical s ensors t hat a re l argely ba sed o n me asurement o f electrical parameters such as resistance or capacitance, ber optic sensors make use of a variety of novel phenomena inherent in the structure of the ber itself. Some of these phenomena are briey presented in this document along with an introduction to the technology and its application in several industrial sectors.

14-1

43722_C014.indd 1

10/16/2008 7:16:01 PM

14-2

Smart Materials

US$ (million)

1600 Multiplexed (FBG) and distributed (Brillouin, Raman) sensors

1200

800 Point sensors

400

0

2002

2004

2005 Year

2006

2007

2008

Estimation of worldwide ber optic sensor market.

According to t he Bu siness C ommunications C ompany, I nc. (BCC) [1], which is one of the market research companies, ber optic sensors are classied as by-product of optical ber research and de velopment for t he communication i ndustry. Ā is industry, regardless of its sluggish recent market, remains the largest application m arket f or t his s ensor te chnology, i n add ition to medicine, defense, and aerospace applications. Ā e world market for  ber opt ic s ensors w as w orth $ 175 m illion i n 1 998 a nd expected to reach $600 million by 2011. However other estimates expect t he m arket to b e w orth $1600 m illion b y 2 008 ( Figure 14.1) [ 2]. F igure 1 4.1 f urther i llustrates t he p otential f or distributed versus d iscrete sensor t ype. According to B CC, t he total worldwide revenues for  ber optic sensors are projected to increase f rom a c urrent e stimate o f $2 88.1 m illion to $ 304.3 million in 2006 with an increase at a modest average annual growth rate (AAGR) of 4.1% to $371.8 million by 2011. BCC further c oncludes t hat w hile  ber opt ic s ensors te chnology h as tremendous p otential, t his p otential ha s o nly b een r ealized within the telecommunications market. Figure 14.2 [1] illustrates an example of the market potential for only extrinsic and intrinsic sensor t ype. E xtrinsic s ensors a re f orecast to i ncrease at a n AAGR o f 4 .1% u ntil 2 011, to $2 74.4 m illion w hereas i ntrinsic sensors will rise from $79.9 million in 2006 to $9 7.4 million in 2011 with an AAGR of 4%. Intrinsic sensors market revenues lag that of e xtrinsic s ensors; however, t his c ategory of s ensors i s where t he g reatest opp ortunity f or ne w m arket en trants i s expected. Cu rrently, t he l argest app lication s egment f or suc h intrinsic sensors is military technologies that tend to be very specialized and project focused, seldom leading to m ainstream market co nsumption r esulting i n m arket s uccess. G rowth i n intrinsic sensors is expected to be rather slow in some application markets due to ongoing challenges with the technology.

been underway since the early 1970s. In recent years, accelerated progress was experienced due to t he signicant development of new, low-cost materials, and devices and the emergence of micro and na notechnologies f or t he tele communications i ndustry. The shape a nd f orm o f opt ical  bers a re si milar to t hose reinforcing bers used in  ber-reinforced composite materials. However, the diameter of optical fibers is much larger, usually in the order of 40–250 µm, compared to glass and carbon bers used in composites that are typically 10 µm or sm aller. Optical  bers consist o f a l ight w aveguide i nner si lica-based core surrounded by an annular-doped silica cladding that is protected by a polymer coating (Figure 14.3). Ā is optical ber can also be made using other materials such as plastic [3]. Ā e ber core refractive index is relatively large compared to that of the cladding i ndex. Ā e c hange i n re fractive i ndices, b etween t he core and the cladding, provides the required mechanics for light propagation within the ber core. Depending on the wavelength of t he l ight i nput, w aveguide ge ometry a nd d istribution o f i ts

300 250

US$ (million)

FIGURE 14.1

2003

200 150 100 50 0 2004

2005

2006

2011

Year

14.1.2 Fiber Optic Sensors Fiber optic sensor development capitalized on the successful discovery of communication optical  bers in the 1950s and has

43722_C014.indd 2

Extrinsic

Intrinsic

FIGURE 14.2 Worldwide ber optic sensor revenues by category.

10/16/2008 7:16:02 PM

14-3

Fiber Optic Systems: Optical Fiber Sensor Technology and Windows Buffer (protective jacket) Sensor parameter

Coating Core

Single mode fiber (µm)

Multimode fiber (µm)

Material

2−10

50−150

Silica-based

100−250

Doped silica

Core Cladding

80−120

Coating

250

250

Polymer

Buffer

900

900

Plastic

Cladding

FIGURE 14.3

Schematic of the basic structure of an optical ber.

refractive i ndices, s everal mo des c an p ropagate t hrough t he ber, re sulting i n t he s o-called, si ngle a nd m ultimode opt ical bers. Both ber types are used in the construction of ber optic sensors. However, single-mode bers are more sensitive to strain variation and are thus the preferred choice. Table 14.1 [4] provides a summary of the typical properties of various optical ber types (e.g., single and multimode). From a p ractical p oint o f v iew, opt ical  ber s ensors c an b e divided i nto t wo b road c lasses, na mely s hort a nd lo ng ga uge length optical  ber sensors, also known as discrete and distributed. Short gauge leng th s ensors a re c omparable to t hose c onventional sensors such a s r esistive f oil s train g auges a nd thermocouples. They typically provide a me asure of physical parameters over distances of several millimeters (≤20 mm). Long gauge leng th s ensors h ave ga uge leng ths r anging f rom s everal centimeters to hundreds of meters and kilometers. Ā ey typically measure physical para meters over distances of few centimeters to few meters (40 cm to 5 m). Ā e latter type is generally known as distributed, spatially distributed, distributed-effect, integrating, a nd/or a veraging s ensors. Figure 14.4 illustrates the c lassication a nd c onguration o f t hese se nsors. Ā e two

TABLE 14.1

Typical Properties of Various Optical Fiber Types

Typea Multimode step index Glass–clad–glass (bundle) Plastic–clad–silica Glass–clad–glass Single mode (low loss) Glass–clad–glass a

43722_C014.indd 3

classes of s ensors c an a lso b e d ivided i nto t wo t ypes, i ntrinsic and extrinsic. Ā is refers to the sensing region of the ber sensor being either outside or inside the ber, respectively. Both intrinsic and extrinsic sensors congurations that also employ single mode and multimode  bers are illustrated in Figure 14.5 [5]. In the extrinsic type, the optical ber acts only as a light-transmitting a nd c ommunication me dium. Re ceived l ight i ntensity i s affected by the variation in the physical parameters in the transducer (sensitive) element. Ā e transducer element does not necessarily consist of optical bers and the optical effect occurs in a different medium. Such sensors usually employ multimode optical bers. In the intrinsic type, the optical ber itself acts as the sensing element, which includes both transducer and communication media. When the conditions of the sensed medium change, t he l ight p ropagation p roperties o f t he opt ical  ber also change, providing a measurement of the change in condition. Sensors of t his t ype a re i mmune to d irty environments due to t heir c losed opt ical pat h a nd a re gener ally f ound i n ber optic gyroscopes and  ber optic hydrophones. Ā ese are expected to have the greatest developmental opportunity regardless of their slow market growth.

Loss (dB/km)

Numerical Aperture (NA)

400–600 3–10 2–6

0.4–0.6 0.3–0.4 0.2–0.3

2–6

0.15

Core Size (OD Microns) 50–70 200–600 50–200 5–8

Core to Cladding Ratio

Bandwidth (MHz km)

0.9–0.95 0.7 0.4–0.8

20 20 20

0.04

1000

A variety of material combinations can be used for both core and cladding; however, it is generally application dependent. Some combination examples include: cladding and core of quartz glass; plastic core and cladding; normal glass core and cladding; plastic cladding and quartz glass core (known as plastic-cladded bers [PCFs]).

10/16/2008 7:16:02 PM

14-4

Smart Materials Point or short gauge length sensor

Signal processing unit Communication link

Long gauge length or integrating or distributed-effect sensor

Signal processing unit Communication link

Signal processing unit

Signal processing unit

Distributed sensor

Individually addressed, quasidistributed or multiplexed sensor array

Communication links

Quasidistributed or multiplexed sensor array

Signal processing unit Communication link

FIGURE 14.4 Classication of optical ber sensors.

Traditionally, optical ber sensors are classied by their transducer o r mo dulation me chanisms. Ā e transducer mechanism affects the properties of the transmitted light in the ber, such as intensity, polarization, phase, and modal content (wavelength). Sensors with these transducer mechanisms are known as intensity, polarimetric, phase, or modal content sensors, respectively. Fiber optic strain, temperature, pressure, and chemical sensors have been developed, which employ many different optical modulation m echanisms in cluding in tensity ( distributed mi crobend), phase (interferometric), w avelength (Bragg g ratings), a nd s tate

of polarization (polarimetric). Some of t he technology benets, concerns, and constraints are provided in Table 14.2.

14.1.3

Optical Fiber Modulation Mechanisms

14.1.3.1 Intensity-Modulated Sensors Intensity-modulated sensors rely o n the reduction of transmitted light intensity through an optical ber as a function of the perturbing en vironmental pa rameters [ 6,7]. Ā e l ight i ntensity at tenuation or loss can be associated with transmission, reection,

Sensitive element

Sensitive element Optical fiber

Optical fiber

Disturbance LED

LED

Disturbance

Photodetector

Photodetector Extrinsic

Intrinsic

FIGURE 14.5 Extrinsic and intrinsic optical ber sensors congurations.

43722_C014.indd 4

10/16/2008 7:16:02 PM

14-5

Fiber Optic Systems: Optical Fiber Sensor Technology and Windows TABLE 14.2 Partial List of Applications, Benets, and Concerns of Optical Fiber Sensors Applications

Benets

Industrial processes Industrial control Biotechnology Energy industry (generation, distribution, exploration, and extraction) Off-shore oil rigs Undersea pipelines Navel vessels Military/aerospace industry Spacecraft and space structures Robotic systems Structural monitoring Aerospace guidance and control Passive/active damping Damage localization in civil, mechanical, aerospace structures Automotive monitoring industry Intelligent transportation systems Utility infrastructure Environmental monitoring Intelligent building systems Security systems

Load cells Pressure vessel monitoring

Concerns

Compact size and low weight Geometric versatility and exibility EMI/RFI immunity High speed and high bandwidth capacity

Moderate to high cost Lack of standardization Impractical modulation units Calibration difficulties due to drift

Remote sensing capability Tolerance to extreme environments Intrinsic safety Low maintenance High sensitivity to multiple quantities Distributed sensing potential Passive effect on measured environment Increased networking and multiplexing capabilities Potentially inexpensive Greater sensitivity

Selectivity Long-term stability, durability, and reliability Lack of suitable temperature compensation Accuracy/linearity Low demand Complex signal processing Potential damage when handling Isolation from unwanted parameters Availability of optical sources Low general awareness

Electrical passiveness Low maintenance and reduced maintenance costs Wide dynamic range Insensitivity to corrosion Minimum impact on mechanical properties of host structure Ideally suited to be embedded into composite structures

Availability of suitable instrumentation Existence of competing technologies Embedding protocol Bonding robustness Multiplexing capability

Potentially high temperature capability Both point and distributed conguration

microbending, absorption, and scattering (Figure 14.6) [4,8]. Āese sensors a re si mple to c onstruct, a re low cost, a nd do not re quire complex signal interpretation techniques. However, like all intensity-modulated sensors, the signal can be affected by the light source’s uctuations affecting the accuracy of the measurement. In addition, they are difficult to calibrate and require a ber optic network to de termine t he lo cation o f t he de sired me asurement ( e.g., damage). A classical application of these single use sensors is large area damage detection in structural components [6].

14.1.3.2

Technological (e.g., temperature compensation, signal handling, multiplexing, optical signal processing) Operational (e.g., in-service applications) Demand (e.g., military continue to be the primary market)

Polarimetric Sensors

Polarimetric sensors rely on coherent interference between two light beams from a common source traveling along two different (orthogonal) p olarization a xes o f a c ommon opt ical  ber. Ā is sensor uses a highly birefringent (HiBi) ber t hat d ivides linearly incident polarized light into the two orthogonal modes traveling along the ber at different phase velocities. Ā is phase velocity diff erence p revents t he t ransfer o f o ptical en ergy between s tates a nd t hus m aintains t he p olarization s tate o f

Microbend

Irregular end finish

Bubble

FIGURE 14.6

43722_C014.indd 5

Impurity (absorption)

Density change (scattering)

Schematic of the causes of light intensity attenuation in an optical ber.

10/16/2008 7:16:03 PM

14-6

Smart Materials Hi-Bi (Polarization maintaining lead-in fiber) Fiber polarizer (or polarizing fiber) 450 splice 3 dB coupler

Source

∗

Mirrored end On-axis splice

Hi-Bi active fiber

Reference Output

FIGURE 14.7 Schematic conguration of a polarimetric sensor conguration.

light t ransmitted t hrough t he  ber. Figure 14.7 illustrates a schematic conguration of a p olarimetric sensor [8]. A t ransverse loading of the ber will produce a change in the birefringence that manifests itself as a change in the polarization state at t he  ber o utput. Various e xperimental a rrangements h ave been de veloped to c onvert t his p olarization s tate c hange to a change in optical output power [9–11]. Ā e major disadvantage of this type of sensor is the three-dimensional nature of strain sensitivity i n t hese  bers, w hich c an b e d ifficult to resolve. Various approaches have recently been investigated to address this limitation such as writing gratings or overlaying two gratings with signicantly different wavelengths into the birefringent ber [12]. 14.1.3.3 Interferometric Sensors Interferometric s ensors rely o n t he ph ase c hange b etween t wo interfering paths of light in response to e xternal stimuli, such as strain, pressure, temperature, and/or chemical reaction. Ā e two paths of light can either come from two different bers, such as the case for Mach–Zehnder and Michelson sensors, or from a si ngle optical  ber, suc h a s t he c ase i n F abry-Perot a nd Br agg g rating sensors [3]. In a Mach–Zehnder interferometric sensor, the sensing ber is either embedded or bonded to a h ost structure while the

Source

∗

3 dB coupler

reference ber remains external to it, as shown in Figure 14.8 [8]. Ā e necessity of such reference ber makes the use of such a system impractical for large-scale applications due to the requirement of preserving the state of the reference arm. In a Michelson interferometric sensor, both the sensing and reference bers are embedded or bonded to the host structure. Both bers have mirrored ends and t he sensing region is formed by t he d ifferential path length between t he t wo pa rallel  bers, a s s hown i n F igure 1 4.9 [ 8]. Although this type of sensor possesses high sensitivity and is simple to construct, it suffers from phase preservation problems at the  ber–host i nterface, g reater i ntrusion to t he h ost s tructure (due to the fact that the sensing and reference bers must be closely spaced), and higher susceptibility to noise [3]. Unlike M ach–Zehnder a nd M ichelson s ensors, F abry–Perot interferometric (FPI) sensor only requires a si ngle optical  ber and c an b e c onstructed i n t wo c ongurations: intrinsic (IFPI) and extrinsic (EFPI). As shown in Figure 14.10 [8], t he sensing region of the IFPI sensor is formed by two fusion-spliced reective mirrors in the optical ber whereas the sensing region of the EFPI sensor is formed using two optical bers fused to a hollow glass tube by either fusion-splicing or adhesive bonding, forming the sensing cavity (air gap). Ā e mirrored ends of two optical bers are perpendicular to the ber axis and separated by an air

Sensing arm

3 dB coupler

Reference arm

FIGURE 14.8

Output

Schematic conguration of a Mach–Zehnder interferometer.

Source

∗

Output

3 dB coupler

Sensing arm

Reference arm

Mirrored fiber ends

FIGURE 14.9 Schematic conguration of a localized Michelson interferometer conguration.

43722_C014.indd 6

10/16/2008 7:16:03 PM

14-7

Fiber Optic Systems: Optical Fiber Sensor Technology and Windows

Sensing region Source

3 dB

∗

Lead-in/out fiber

Reference

Output

Partially mirrored Mirrored end Fabry–Perot cavity or air gap

Schematic configuration of interferometric sensors

Cladding

Internal semireflective

Alignment

Thin film mirror

Single mode fiber

Multimode fiber

R

Air Splice

Gauge length

Epoxy

IFPI

EFPI Gratings (semireflective mirrors)

Gauge length Cladding

Incident light Cladding Hollow core fiber

Transmitted light

Reflected light Gauge length

ILFE

FBG

Cladding Cladding Incident light

Incident light

Reflected light (Low) Grating period

Transmitted light

Reflected light Grating period

LPG

Transmitted light

TFBG

Sensing surface Gratings Core Cladding

FIGURE 14.10 Schematic conguration of IFPI, EFPI, ILFE, FBG, LPG, TFBG, and D-shaped optical ber.

43722_C014.indd 7

10/16/2008 7:16:03 PM

14-8

Smart Materials

gap, forming the EFPI. Any changes in the air gap length of the EFPI or gauge length of the IFPI affect the phase change between the le ad-in/lead-out a nd t he re ection bers, re sulting i n t he appropriate measurement. It is well known that fusion splicing or adhesive bonding of the optical  bers during t he construction of t hese s ensors i nadvertently introduce potential weak points at the physical connections for both IFPI and EFPI. In particular, severe stress concentrations exist in EFPI sensors at the alignment tube fusion points. It is also established that the relatively large size of the EFPI may compromise the structural strength and integrity if embedded into host materials. Ā ese sensors are known to have reduced multiplexing and networking capability with the need for sophisticated interrogation equipment [3]. To reduce the stress concentration effect, an inline ber etalon (ILFE) was proposed [13]. Ā is sensor, illustrated in Figure 14.10 [8], provides the same properties as EFPI with a w ell-dened m icrosize ga uge leng th ( ∼20 µm), r educed temperature sensitivity, and high transverse strain insensitivity. Ā e se adde d adv antages o ver E FPI c ome at t he e xpense o f increased f abrication d ifficulties, l imited multiplexing c apabilities, and reduced technology exposure. Unlike Fabry Perot and ILFE sensors, the transducer element of a  ber Bragg g rating (FBG) sensor is a n i ntegral pa rt of t he optical  ber, providing this type of sensor with increased networking, m ultiplexing, a nd re duced i ntrusiveness c apabilities .Ā e light in the sensing ber of the FBG is also split into two b eams at t he b eginning o f t he g rating. O ne l ight b eam i s reected back at the gratings (the semireective mirrors), while the ot her l ight beam t raverses t he g rating t hrough t he sensing region o nly to en ter t he ne w g ratings a s i llustrated i n F igure 14.11. It is noted that the reection wavelength is a function of the average refractive index and the grating pitch. Straining the ber causes a change in the grating pitch and hence the reected wavelength. Ā is si ngle-mode s ensor c onverts ph ase c hanges into spectral information more suitable for measurement. Ā e gratings are formed through periodic modulation of core refractive Reflected (all gratings)

Grating 1

index u sing u ltraviolet l aser or ph ase m ask te chnology [14,15]. Ā ese gratings only reect a narrow band of wavelengths propagating in the ber. FBGs show excellent linearity between measured s train a nd app lied s tress [ 16,17]. Wi th app ropriate calibration and thermal compensation, accurate strain measurement c an b e ob tained. Ā e i ncreased m ultiplexing c apability using t his single-ber provides d istributed me asurements over small a nd l arge su rface a reas, t hrough w avelength d ivision multiplexing (WDM) [17,18] or optical frequency domain reectometry (OFDR) [16,19], making them more suitable for largescale applications where t he weight penalties a ssociated w ith a large n umber o f si ngle-ended s ensors c an e asily b e m itigated. Drawbacks accompanying t his t ype of sensor i nclude t he h igh cost o f i nterrogation s ystem, s ensitivity to tem perature v ariations, and limited dynamic response (restricted by the interrogation hardware) amongst others. Nevertheless, the overall benets of Bragg grating sensors make them the ideal choice amongst the family of ber optic sensors for use in health monitoring systems for both composite and metallic structures.

14.1.4 Emerging Optical Fiber Concepts In recent years, several ber optic sensor concepts have emerged as a result of the increasing interest in online in-situ structural health monitoring and in response to addressing some of the limitations presented b y c urrent te chnologies. Ā ese c oncepts a re gener ally based on the above presented modulation mechanisms with novel approaches f or t ransducer f abrication a nd i nterrogation. A mong the several concepts that can be found on the open literature and briey l isted i n Table 14.3, o nly t hree a re b riey p resented: lo ng period gratings (LPG), tilted ber Br agg g ratings (T FBG), a nd D-shaped  ber Br agg g ratings (DFBG). Ā ese s ensors a re mainly introduced to either enhance the exciting  ber optic sensor capabilities, address current limitation such as temperature compensation, dynamic and range, high multiplexing and distributed sensing, or to increase sensing system sensitivity to different parameters.

Grating n−1

Reflected

Reflected

Grating n

......

Transmitted

Transmitted

Transmitted

Transmitted light

λo

λo

λ

λ

(a) Single grating fiber

Incident light

λ

Reflected light

λ

Reflected light

x

Incident light

Index of refraction

(all gratings)

Transmitted light

Incident

.... λo

λo

λn

λn

λ

λ

(b) Multiple grating fiber

FIGURE 14.11 Schematic conguration of a FBG sensor.

43722_C014.indd 8

10/16/2008 7:16:03 PM

14-9

Fiber Optic Systems: Optical Fiber Sensor Technology and Windows TABLE 14.3

Some Concepts on Optical Fiber Sensor Technology and Application

Concept Sagnac interferometer Fluorescent-based ber sensors White light interferometry Magnetic eld sensors Semiconductor absorption sensors Velocity sensors Ultrasonic/laser NDE sensors Long period Bragg gratings (LPGs) D-type ber Bragg gratings (DFBG) Tilted ber Bragg gratings (TFBG) Optical time domain reectometry (OTDR) Polarimetric ber optic sensors

Parameter Measurement Rotation

Global positioning systems, electric current sensing In vivo chemical agent measurements Chemical/biological applications Dispersion of optical materials, displacement, Photonics, material engineering surface proles AC/DC magnetic elds Banking (credit card detection) and computer engineering (CD reading) Temperature In harsh conditions (such as radiation, high voltage, and EMI prone areas) In vivo ow Medical/clinical environments Structural elements and faults/defects Concrete testing—civil engineering Strain, external refractive index, temperature, SHM, civil, chemical engineering, photonics bend radius Bending, temp, strain, etc. Structural health monitoring Simultaneous temperature and strain, Aerospace, civil, manufacturing, photonics wavelength interrogation, refractive index Fiber breaks, loss spectra, strain temperature SHM and photonics External perturbations Civil, mechanical, aerospace, etc.

14.1.4.1 Long Period Bragg Grating Sensors LPG s ensors do not re quire t he f abrication c omplexity ob served with regular FBGs [20] and have very low back reection, low insertion lo ss, a nd e xhibit p olarization i ndependence [ 21]. Ā e LPG periodicity is typically above 100 µm, compared to the submicron pitch of regular Bragg gratings. Ā ese sensors work in transmission relying on the coupling between the fundamental core mode and the propagating cladding modes (Figure 14.10) [22,23]. Ā is produces a series of interference fringes within the attenuation bands in t he transmission spectrum centered at speci c wavelengths of the L PG  ber [ 24]. C hanges i n t he c enter w avelengths o f t hese bands depend on the grating period/pitch, the length of the sensor and external factors such as temperature, strain, external refractive index, a nd b end r adius. Ā e l atter e xternal p erturbations i nherently modify the physical characteristics of the ber in terms of its grating p itch a nd i ts re fractive i ndex o f t he c ore a nd c ladding modes [22] and are of signicant interest in the eld of smart structures, c hemical/biological s ensing, opt ical c ommunication, a nd health monitoring applications. Ā eir exceedingly high sensitivity to transverse as well as longitudinal strain and their acute sensitivity to tem perature advocate t heir potential ut ilization i n applications, requiring multiaxis strain and temperature sensing [22,25]. LPGs can also be customized to largely remove the sensitivity to all but one selected parameter [21,26]. Ā ese sensors have the difficulty of allowing localized measurements and embedding into composite structures. I nterrogation of t he output of a n L PG a nd multiplexing capabilities are also deemed to b e complex [27]. Methods have been proposed based on derivative spectroscopy interrogation techniques employed with two pairs of FBGs and LPGs [24]. 14.1.4.2

Tilted Fiber Bragg Grating Sensors

As t he na me su ggests, T FBG, u nlike t raditional g ratings, a re offset vertically (angular orientation) inside the  ber core. Ā ey

43722_C014.indd 9

Applications

are p roduced b y rot ating t he i ncidence o f t he i nterfering l aser light used to write the grating from the usual perpendicular orientation. Ā is results in gratings that are inclined or tilted to the ber a xis. Ā e p eriodicity, a nd h ence t he Br agg w avelength o f the tilted grating, is greater than that in the untilted conguration. Tilting of the grating is a s imple method of increasing the Bragg w avelength w ithout ne eding e xtra w riting e quipment, although it has been shown that the grating reectivity decreases steadily to zero for tilt angles from 0° to 45° [28]. A tilted Bragg grating can be combined with a re gular grating to s eparate the cross-sensitivity o f tem perature a nd st rain e ffects [ 29,30]. As shown in Figure 14.10, the coupling of light from the ber core is directed to toward the cladding, which suggests that the narrow attenuation ba nds i n t he t ransmission spectrum a re sensitive to the external refractive index [31]. 14.1.4.3

DFBG Sensors

Ā ese a re Br agg g ratings, u sually o f lo ng p eriod, w ritten i nto D-shaped optical ber (Figure 14.10). Ā ese sensors show strong ber o rientation de pendence i n t heir s pectral re sponse w hen subjected to bending [32,33]. DFBGs can form the basis for new vector sensors with the ability to measure curvature with direction recognition. Ā ey can also be used as temperature sensors [34]. Ā e prime reason for the usage of the D-shaped  ber sensors in the industry as opposed to standard circular bers is the ability to interact e ffectively w ith t he p ropagating opt ical  eld i n t he area external to the ber. Ā is is because the core of the DFBG is closer to t he o uter re gions b eyond t he  attened c ladding [35]. Ā e opt ical  eld ba sically a llows t he i nteraction b etween t he lightwave and any external perturbations, in other words dening t he sensitivity of t he sensor to en vironmental disturbances [36]. A p roposed c onguration i s a h ybrid F BG-LPG s ensor inscribed in a D -shaped  ber that allows for both temperature

10/16/2008 7:16:03 PM

14-10

Metal-ion sensor element

Cl Sensor element

NO Sensor element

pH Sensor element

Corrosion sensor demodulator box

SO Sensor element

Areas prone to corrosion

HO Sensor element

Smart Materials

Centralized monitoring system

FIGURE 14.12 Illustration of a ghter jet health monitoring system.

and e xternal re fractive i ndex s ensitivity [37]. O ne d rawback when dealing with D-shaped sensors is that the sensitivity to curvature of these FBGs is as pronounced as their sensitivity to other physical parameters such as axial strain and temperature. One proposition is to join two D-shaped FBGs at their horizontal level planes, t hus making it less susceptible to t he effects of temperature and axial strain changes [27].

14.1.5

Selected Applications of Fiber Optic Sensors

Optical  ber s ensors p rovide t he a bility to me asure v arious physical pa rameters, suc h a s tem perature, p ressure, s train, displacement, rot ation, m agnetic/electric  eld, a nd c orrosion (Table 14.3) [3,7,38–41]. To date, it is known that more than 60 different parameters can be measured using ber optic sensors. In addition, these sensors can be used in applications where no electrical analog is suitable. When embedded directly into materials, optical ber sensors can be used to provide in-situ process monitoring [42,43]. In conjunction with nondestructive evaluation schemes, they can also be used to monitor and evaluate not only the integrity of these structures once they have been manufactured, but also to check for aws that may have resulted from processing a nd h andling. Ā ese s ensors c an f urther b e c onnected into a health-monitoring network to determine maintenance requirements a nd operational readiness, t hus increasing worthiness, re ducing f requent m aintenance de pendency c osts, and providing engineers with an opportunity to improve designs and safety of their respective products. Figure 14.12 illustrates a concept o f a p otential h ealth mo nitoring s ystem f or a t ypical ghter jet. Such concept is increasingly emerging in all sectors of the industry, pa rticularly s pace a nd a erospace. I n a erospace,

43722_C014.indd 10

sensors are used w idely to mo nitor v ital components of aircraft, fighter jets, and rocket engines. Fiber optic sensors are used i n t he A rnold E ngineering D evelopment C enter to supervise the air engine and fuel pressure in the combustor of a dual combustor ramjet ( DCR) ( Figure 14.13 [44]). I n t his case, a pre ssure t ransducer ba sed o n re f lected l ight a nd Fabry Perot interferometry principles are used (Figure 14.14 [44]). A l ight-emitting diode (LED) light source incident on the fiber is directed toward a diaphragm, which def lects due to p ressure f luctuation. Ā e l ight i s re ected b ack t oward a photodiode that measures the intensity o f t he r eected light, which is determined to be proportional to the deection experienced b y t he t ransducer d iaphragm [ 44]. It is determined that t hese t ypes o f s ensors a re mo re d iffic ult to manufacture due to their signicant precision requirements [44]. NASA e mployed  ber opt ic Br agg g rating tem perature a nd strain s ensors f or t he s tructural h ealth mo nitoring i n t he X-38 spacecraft (Figure 14.15). Ā ese sensors are either attached to t he structure or strain isolated for temperature compensation application depending on the situation of interest [45]. Upon reentry, the aft o r re ar s ection o f t he v ehicle e xperiences a sig nicant amount of structural load (due to drag) and thermal gradient. Eight s train a nd f our tem perature s ensors e xperienced temperatures a nd s trains r anging f rom − 40°C to 1 90°C a nd −1000 µε t o 3 000 µε, re spectively [ 45]. Ā ese Br agg g rating sensors a re de termined to b e t he b est a lternative to u sing conventional t echniques w ith st rain g auges a nd t hermocouples requiring additional electronic equipment and controls. Previous work on the X33 reusable space vehicle focused on the use of ber optic s ensors f or s tructural h ealth mo nitoring o f t he c omposite fuel tank. Figure 14.16 [46] illustrates a p ostinstallation check of the instrumentation of a  ber optic sensing system on t he composite fuel tank prototype of the X33 reusable space vehicle.

10/16/2008 7:16:04 PM

14-11

Fiber Optic Systems: Optical Fiber Sensor Technology and Windows

Fiber optic ports

Fiber optic cables

(a)

(b)

FIGURE 14.13 Installation of ber optic sensor in a combustor of the DCR.

Sensor LED

Sensor housing Fiber optic cable

Drive electronics Fibers Pressure seal

Photodiode

Diaphragm

FIGURE 14.14 Conguration of pressure transducer sensor used in the compressor of DCR.

Sensors locations

FIGURE 14.15 Illustration of the X-38 reentry vehicle and the sensor targets.

Ā e European Space Agency (ESA) has also vigorously pursued the use  ber optic sensors in various aerospace projects undertaken or in progress [47], such as their use of embedded sensors on t heir ne xt generation l aunchers. Ā ese ro cket l aunchers a re

43722_C014.indd 11

generally exposed to e xtreme mechanical a nd t hermal stresses because of their operational harsh environments. As part of their reusable launch vehicles (RLV), ESA has conducted assessments on the use of palladium-coated D-shaped FBG sensors in their

10/16/2008 7:16:04 PM

14-12

Smart Materials Palladium coating in D-fiber

D-fiber with elliptical core

FIGURE 14.17 Palladium-coated ber Bragg sensor.

Recent advent of composite repair technology for aging metallic military aircraft structures has seen the development of “smart patches” with embedded ber optic strain sensors for the assessment o f patc h i ntegrity, w hich de tect d amage t hrough a measurable loss in load transfer in the region of repair [48–50]. Signicant efforts b y s everal de fense o rganizations f ocused o n the e xploitation o f t hese s ensors f or b onded c omposites patc h repair integrity assessment, monitoring, and repair certication. Arrayed  ber opt ic Br agg g rating s ensors w ere b onded to a nd embedded into a simulated repair to monitor any patch disbond or delamination. As shown in Figure 14.18 [51], the multiplexing exibility o f t hese s ensors p rovided a su itable le ss i ntrusive approach to re pairs monitoring. Ā e s ame  gure further illustrated the effectiveness of these sensors in disbond and delamination m onitoring f or spe cic re pair c onguration. Ā e

Fiber optic sensors (In white)

FIGURE 1 4.16 Fiber opt ic s ensors for pre ssure mon itoring i n t he composite fuel tank of the X-33 reusable space vehicle.

cryogenic fuel tanks to monitor the changes in strain, temperature, and any hydrogen leakage that might occur [45,47]. Figure 14.17 i llustrates suc h c oated s ensor t hat h as s train re sponse ranging from 1000 to 3000 µε for temperatures as low as −253°C. It is observed that the palladium coating decreased the response time of the sensor at low temperatures (≤30°C).

Fiber optic instrumented patch 13 plies graphite/epoxy

Flaw in adhesive

Sensor 1 2

3

4

Flaw in patch

Al inner adherend thickness 1/4"

Embedded optical fiber with Bragg grating sensors Results for fiber optic patch Embedded optical fiber sensor output (microstrain)

1200 1000 800 Sensor 1 Sensor 2 Sensor 3 Sensor 4

600 400 200 0

Sensor 1

10

2

3

4

20 30 40 50 60 70 Damage length from edge of patch (mm)

80

FIGURE 14.18 FBG in bonded repair health monitoring.

43722_C014.indd 12

10/16/2008 7:16:05 PM

14-13

Fiber Optic Systems: Optical Fiber Sensor Technology and Windows

Dams and bridges

Building structures

Communication and power lines

FIGURE 14.19 Fiber optic application to dams, power lines, and smart buildings.

environmental issues impacting sensors performance in actual aircraft operating environment remain unaddressed, leading to reduced p roduct c ommercialization a nd m arket p enetration. More re cently, t he h ealth mo nitoring o f c omposite m arine structures using  ber optic sensors is a lso beginning to at tract noticeable research attention [52–54], in addition to t he signicant research that has expended on the implementation of ber optic strain sensors for delamination detection and load monitoring in aerospace carbon/epoxy composite structures as well as traditional metallic airframes [16,19,55–58]. In industrial engineering applications, infrastructural health monitoring h as si gnicant a nd de trimental i mpact o n p ower supply, general health, and medical services. As witnessed by the 1998 ice storm of eastern Canada, 4%–5% of the population, or approximately 140,000 people, were forced to spend at least one night i n one of t he 4 54 emergency s helters, w hich t he d isaster organization had established due to t he highly uninstrumented

infrastructure. D istributed o r m ultiplexed  ber opt ic s ensors suitable for l arge-scale monitoring c an b e found i n m ines a nd tunnels, where pressure, temperature, and strains are monitored for t he p urpose o f s tructural i ntegrity mo nitoring a nd s afety enhancement. N ew gener ation b uilding s tructures, “ smart buildings,” em ploy d istributed  ber opt ic s ensor ne tworks f or earthquake mo nitoring,  re de tection, a nd no ise mo nitoring (Figure 14.19). Ā ese sensors are usually interfaced with a c ontrol s ystem f or s tructural ad aptive c ontrol f or t he p urpose o f enhancing s tructural s afety a nd re ducing e arthquake-induced damage. D istributed, m ultiplexed, o r h ighly ne tworked  ber optic sensors are seen to adv ance the highway system and contribute to t he i ntelligent t ransportation s ystems (I TS). S ensors are embedded into pavement for the dual purpose of pavement deterioration mo nitoring a nd w eigh-in-motion ( WIM). F igure 14.20 [59] illustrates a schematic of a Bragg grating–based WIN and SHM and control system.

Command and control center

Bridge deck

Solar panel wireless telemetry

Local control unit

Steel / FRP reinforcement and fiber optic sensors

Actuators and sensors

Attached/imbedded sensors

Steel/FRP jacket

Actuators and LVDT

FIGURE 14.20 Fiber optic sensor network for WIM and SHM and control.

43722_C014.indd 13

10/16/2008 7:16:05 PM

14-14

Smart Materials

12 10.5 100 cm 77 cm

9-1 9-2 7.5

0

1.5 3-2 3-1 6

4.5

Locations of 10 sections of sensing fibers 100 cm Sensing fibers

FIGURE 14.21 Location of the sensors around the pipe section.

In t he oil a nd ga s i ndustry, t here is a g rowing concern w ith the maintenance costs associated with repair or remediation of underground p ipelines t hat a re su sceptible to v arious f aults including dents, leakages, buckling, and corrosion. Ā e se faults can signicantly c ompromise t he i ntegrity of t he pipeline a nd cause major economic and health disruption. Ā e use of distributed Brillouin ber optic sensors (DBFS) for environmental and pipeline s train mon itoring h as b een u nderway for s ometime now ( see F igure 1 4.21) [ 60]. Ā ese temperature-compensated strain or pressure sensors have been demonstrated to have high resolution to tem perature a nd s train a nd p rovide b oth measurements in harsh environments.

14.1.6

Conclusions

With t he emergen ce o f a reas, suc h a s d iagnostics, p rognostics, health monitoring and management, and smart structures comes the ne ed f or adv anced s ensory s ystems c apabilities. D espite t he existence of an extensive list of sensors, advanced features, such as reduced m ultiplexing ele ctronics, a nd m ultisensor d ata f usion, calls f or ne w a nd mo re adv anced s ensor c oncepts suc h a s  ber optic sensors. Ā is chapter focused on providing a brief overview of  ber optic sensor technologies and its applications to s elected industries. Several long- and short-gauge ber optic sensors were introduced. Ā ese sensors included Mach–Zehnder and Michelson interferometers, p olarimetric, i nterferometric, lon g a nd s hort period Bragg grating sensors, IFPI and EFPI se nsors, ILFE sensors, among others. Ā e benets, applications, and concerns associated with several of these sensors were also highlighted. Long gauge  ber opt ic s ensors a re mainly employed i n c ivil infrastructure app lications, suc h a s b ridges, p ipelines, p ower lines, d ams, s kyscrapers, m ines, a nd u nderground e xcavation. Although t hese i nterferometric a nd Br illouin s cattering-based sensors e xhibit t he p otential f or u se i n l arge a reas, s tructural applications f or si multaneous s train/pressure a nd tem perature measurements, t hey l ack t he s patial re solution re quired f rom some applications like aerospace. Ā ese sensors also suffer from

43722_C014.indd 14

the e levated c ost a ssociated wi th th eir m anufacturing, o peration, and signal demodulation. Short ga uge leng th s ensors p rovide p otential s olutions to challenges in the aerospace eld and continue to provide technological adv ances f or t he de velopment o f sm art s tructures and t heir app lications. Br agg g rating a nd E FPI o ptical  ber sensors w ere f ound to b e t he mo st p romising a nd re latively mature te chnology for i ntegration i nto sm art s tructures a nd health mon itoring a nd m anagement. Ā ese sensors provide high resolution, improved accuracy, high sensitivity and millimeter-size gauge leng ths suitable for con ned environments. Bragg g rating s ensors c ontinue to b e t he u ndisputed c hampion b ecause o f t heir sup erior m ultiplexing c apabilities; however, t hey s till s uffer f rom temperature a nd st rain crosssensitivity. Regardless o f t he adv ances i n f iber opt ic s ensor te chnology, i llustrated b y t he app lications p resented i n t his do cument a nd ot hers, issues related to adv ances i n opt ical sig nal processing, de velopment o f s pecialty opt ical f iber, s ystem integration, a nd s ystem e conomics ne ed to b e add ressed a s these form t he ner vous system of structures f rom t he stages of i ts i nception to i ts re tirement. Cu rrent e fforts a re u nderway in introducing new designs, such as tilted Bragg grating, dual laid LPG, D-shaped Bragg grating, and integrating different s ensing me chanisms to ren der t he s ensing s ystems more efficient and cost-effective.

References 1. Business C ommunications C ompany, I nc. h ttp://www. bccresearch.com/ias/IAS002C.asp 2. Krohn, D.A., Light wave venture LLC, in Proceedings of SPIE, 2004, 5589, 34. 3. Zhou, G. and Sim, L.M., Damage detection and assessment in  bre-reinforced co mposite st ructures wi th em bedded bre optic sensors—review, Smart Materials and Structures, 11, 925, 2002.

10/16/2008 7:16:07 PM

Fiber Optic Systems: Optical Fiber Sensor Technology and Windows

4. Krohn, D.A., Fiber optic sensors: Fundamentals and applications, Instrument Society of America, 138, 1988. 5. Hauptmann, P., Sensors: Principles and Applications, Prentice Hall, Englewood Cliffs, NJ, 1993. 6. Rippert, L., Wevers, M., a nd Van H uffel, S., Op tical a nd acoustic da mage det ection in la minated CFRP co mposite materials, Composites Science and Technology, 60(14), 2713, 2000. 7. Badcock, R.A. and Fernando, G.F., An intensity-based optical  bre s ensor f or fa tigue da mage det ection in ad vanced bre-reinforced composites, Smart Materials and Structures, 4(4), 223, 1995. 8. Mrad, N., Optical ber sensor technology: Introduction and evaluation a nd a pplication, in Encyclopedia of Smar t Materials, S chwartz, M., (e d.), J ohn Wiley a nd S ons, I nc., New York, 2, 715, 2002. 9. Murukeshan, V.M., Cha n, P.Y., On g, L.S., a nd Asundi, A., On-line heal th mo nitoring o f sma rt co mposite str uctures using  ber po larimetric se nsor, Smart M aterials a nd Structures, 8, 544, 1999. 10. Grossmann, B.G. and Huang, L.-T., Fiber optic sensor array for multi-dimensional strain measurement, Smart Materials and Structures, 7(2), 159, 1998. 11. Lo, Y.L., Sirkis, J.S., and Ritchie, K.T., A study of the optomechanical r esponse o f a dia metrically loade d hig h-birefringent optical ber, Smart Materials and Structures, 4(4), 327, 1995. 12. Udd, E., Nelson, D.V., Lawrence, C.M., and Ferguson, B.A., Ā ree axis stra in p lus t emperature  ber grating s ensor, in Proceedings of SPIE, 2718, 1996. 13. Sirkis, J.S., Fib er optic s ensors for smart str uctures: B asics and ap plications, i n Proceedings of SP IE’s F iĀh Annual International Symposium on Smart Structures and Materials, San Diego, CA, 1–5 March 1998. 14. Wosinski, L., D ainese, M., S willo, M., UV-photosensitivity and r efractive index en gineering in si lica-based pla nar lightwave cir cuits, in Proceedings o f 2005 S eventh International C onference on Transparent O ptical Networks, Barcelona, Spain, 3–7 July 2005, 1, 343. 15. Rizvi , N.H., G ower, M.C., G oodall, F .C., Arthur, G., a nd Herman, P., Excimer laser writing of submicrometre period bre Bragg gratings using p hase-shifting mask projection, Electronics Letters, 37(11), 901, 1995. 16. Brown, T., Wood, K ., C hilders, B ., C ano, R ., J ensen, B ., a nd Rogowski, R., Fiber optic sensors for health monitoring of morphing ai rcraft, in Proceedings o f S PIE S ixth Annual International Symposium on Smart Structures and Materials, Newport Beach, USA, 1999, 3674, 60. 17. Simpson, J.O., Wise, S.A., Bryant, R.G., Cano, R.J., Gates, T.S., Hinkley, J.A., Rogowski, R.S., and Whitley, K.S., Innovative materials f or a ircraft mo rphing, in Proceedings o f S PIE’S FiĀh Annual International Symposium on Smart Structures and Materials, San Diego, CA, 1998, 3326, 240.

43722_C014.indd 15

14-15

18. Anastasi, R .F. and L opatin, C., Application of a  ber optic distributed strain sensor system to woven E-Glass composite, NASA TM-2001-211051, 2001. 19. Prosser, W.H., Allison, S.G., Woodard, S.E., Wincheski, R.A., Cooper, E.G., Price, D.C., Hedley, M., Prokopenko, M., Scott, D.A., Tessler, A., and Spangler, J.L., Structural health management for future aerospace vehicles, in Proceedings of the Second Australasian Workshop on Str uctural H ealth Monitoring, M onash U niversity, M elbourne, Australia, 16–17 December, 2004, p. 1. 20. Kolokoltsev, O.V., Svyryd V.A., L lamas I.F., and Romero C.L.O., S ynthesis o f no nuniform lo ng-period o ptical gratings wi th a co ntinuous r efractive index p rofile, Investigation, Re vista M exicana D e F ’Isica, 51(6), 610, December 2005. 21. Wang, Y.P., Rao, Y.J., Ran, Z.L., Zhu, T., and Zeng, X.K., Bendinsensitive lo ng-period  ber gra ting s ensors, Optics a nd Lasers in Engineering, 41, 233, 2004. 22. James, S.W. and Tatam, R.P., Optical bre long-period grating s ensors: Cha racteristics a nd a pplication, Measurement Science and Technology, 14, R49, 2003. 23. Laffont, G. a nd F erdinand, P ., T ilted sho rt-period  breBragg-grating ind uced co upling t o c ladding mo des f or accurate refractometry, Measurement Science and Technology, 12, 765, 2001. 24. Allsop, T., Earthrowl, T., Reeves, R., Webb, D.J., and Bennion, I., Ā e interrogation and multiplexing of long period grating curvature sensors using a B ragg grating based, derivative sp ectroscopy te chnique, Measurement S cience a nd Technology, 15, 44, 2004. 25. Liu, Y., Zhang, L., and B ennion I., Fibre optic load sensors with high transverse strain sensitivity based on long-period gratings in B/G e co-do ped  bre, Electronics Le tters, 35(8), 661, 1999. 26. Chen, S., Tong, Z., Zhao, Q., Liu, Z., and Dong, X., A smart bending sensor with a novel temperature and strain-insensitive long-period grating, Sensors and Actuators A, 116, 103, 2004. 27. Araujo, F .M., F erreira, L.A., Sa ntos, J .L., a nd F arahi, F ., Temperature and strain insensitive bending measurements with d-type  bre Bragg gratings, Measurement Science and Technology, 12, 829, 2001. 28. Erdogan, T. and Sipe, J.E., Tilted ber phase gratings, Journal of Optical Society of America A, 13, 296, 1996. 29. Kang, S.C., Kim, S.Y., Lee, S.B., Kwon, S.W., Choi, S.S., and Lee, B ., T emperature-independent st rain s ensor system using a til ted  ber B ragg gra ting demo dulator, IEEE Photonics Technology Letters, 10(10), 1461, 1998. 30. Caucheteur, C., L homme, F ., M aakaroun, F ., Chah, K., Blondel, M., and Megret, P., Simultaneous strain and temperature s ensor usin g superimposed tilted B ragg gratings, in Proceedings of IEEE/LEOS Symposium, Benelex Chapter, Ghent, Belgium, 2004, p. 219.

10/16/2008 7:16:07 PM

14-16

31. Caucheteur, C. and Mégret, P., Demodulation technique for weakly til ted  ber B ragg gra ting r efractometer, IEEE Photonics Technology Letters, 17(12), 2703, 2005. 32. Allsop, T ., D obb, H ., M ezentsev, V., E arthgrowl, T ., Gillooly, A., Webb, D.J., and Bennion, I., The spectral sensitivity o f lo ng p eriod g ratings fa bricated in el liptical core D-shaped optical fibre, Optics Communications, 259, 537, 2006. 33. Zhao, D., Zhou, K., Chen, X., Zhang, L., Bennion, I., Flockhart, G., M acPherson, W.N., B arton, J .S., a nd J ones, J .D.C., Implementation of vectorial bend sensors using long-period gratings UV-inscribed in special shape bres, Measurement Science and Technology, 15, 1647, 2004. 34. Smith, K.H., I pson, B .L., L owder, T .L., H awkins, A.R., Selfridge, R.H., and Schultz, S.M., Surface-relief ber Bragg gratings f or s ensing ap plications, Applied O ptics, 45, 1669, 2006. 35. Allsop, T., Gillooly, A., Mezentsev, V., Earthgrowl-Gould, T., Neal, R., Webb, D.J., and Bennion, I., Bending and orientational cha racteristics o f lo ng p eriod g ratings wr itten in D-shaped o ptical  ber, IEEE Transactions o n I nstrumentation and Measurement, 53(1), 130, 2004. 36. Ghandehari, M., Materials health monitoring in smart cable structures, CANS MART 2006, in Proceedings of the International Workshop on Smart Materials and Structures, Toronto, Ontario, Canada, 2006, p. 213. 37. Chen, X., Zhou, K., Zhang, L., and Bennion, I., Simultaneous measurement o f tem perature a nd external r efractive index by use of a hybrid grating in D ber With enhanced sensitivity b y HF et ching, Applied O ptics, 44(2), 178, 2005. 38. Morey, W.W., Meltz, G., and Weiss, J.M., Evaluation of a ber Bragg grating hydrostatic pressure sensor, in Proceedings of the 8t h I nternational Co nference o n O ptical F iber S ensors, Monterey Marriott, Monterey, CA, January 29–31, 1992. 39. Kanellopoulos, S.E., H anderek, V.A., a nd Rog ers A.J., Simultaneous strain and temperature sensing with photogenerated in- ber gratings, Optics Le tters, 20(3), 333, 1995. 40. Cavaleiro, P.M., Araujo, F.M., and Lobo Ribeiro, A.B., Metalcoated bre Bragg grating sensors for electric current metering, Electronics Letters, 23(11), 1133, 1998. 41. Luo, F., Liu, J., and Chen, S., Fiber optic distributed sensing scheme f or mo nitoring str uctural stra in a nd def ormation, Optical Engineering, 36(5), 1548, 1997. 42. Lawrence, C.M., Nelson, D.V., Spingarn, J.R., and Bennett, T.E., Measurement of process-induced strains in composite ma terials usin g em bedded  ber opt ic s ensors, i n Proceedings o f t he S PIE I nternational S ociety f or O ptical Engineering, Fiber Optic Smart Structures and Skins, 1996, 2718, 60. 43. Levy, R .L. a nd S chwab, S.D ., M onitoring the co mposite curing process with a uorescence-based ber-optic sensor, Polymer Composites, 12(2), 96, 1991.

43722_C014.indd 16

Smart Materials

44. Land, H.B., III and Eddins, C.L., Optical pressure measurement: U sing  ber o ptic t ransducers in h ypersonic  ight vehicles, IEEE Instrumentation & M easurement M agazine, 7(3), 38, 2004. 45. Ecke, W., Latka, I., Willsch, R., Reutlinger, A., and Graue, R., Fibre optic s ensor netw ork f or space craft h ealth m onitoring, Measurement S cience a nd T echnology, 12, 974–980, 2001. 46. Troy, C.T., Fiber optic smart structures: A technology ahead of its time is  nally winning acceptance, Photonics Spectra, 31(5), 112, 1997. 47. Mckenzie, I. and Karafolas, N., Fiber optic sensing in space structures: Ā e experience of the European space agency, in Proceedings of SPIE 17th International Conference on Optic Fibre S ensors, Voet, M ., Willsch, R ., E cke, W., J ones, J ., a nd Culshaw, B. (eds.), 2005, 5855, p. 262. 48. Davis, C., Baker, W., Moss, S., Galea, S.C., and Jones, R., In situ health monitoring of bonded composite repairs using a novel ber Bragg grating sensing arrangement, inProceedings of SPIE Smart Materials II, 2002, 4934, p. 140. 49. Budtsev, Y., G orbatov, N., Tur, M., G reen, A.K., K ressel, I., Ben-Simon, U., Ghilai, G., Sharr, E., Berkovic, G., and Gali, S., S mart b onded co mposite r epairs f or a ging a ircraft, in Proceedings of the Second European Workshop on Structural Health and Monitoring, Munich, Germany, 7–9 J uly, 2004, p. 341. 50. Li, H.C.H., Dupouy, O., Herszberg, I., Stoddart, P.R., Davis, C.E., and Mouritz, A.P., Health monitoring of bonded composite r epairs usin g  bre opt ic s ensors, i n Proceedings of SPIE, S mart S tructural M aterials a nd ND E f or H ealth Monitoring and Diagnosis, 2006, 6174, p. 367. 51. Mrad, N., Progress towards bre optic smart structures, Ā e Department o f N ational D efence, D efence R&D C anada, Technical M emorandum, Publica tion N o. D RDC Atlantic TM 2003-200, 2004. 52. Li, H.C.H., Davis, C., Herszberg, I., Mouritz, A.P., Galea, S.C., and Ā omson, R.S., Application of bre optic strain sensors for the health monitoring of adhesively bonded composite ship j oints, in Proceedings of th e F ourth I nternational Workshop on Str uctural H ealth and M onitoring, S tanford University, CA, 2003, 1327. 53. Wang, G., Pran, K., Sagvolden, G., Havsgard, G. B., Jensen, A.E., J ohnson, G.A., a nd Vohra, S.T ., S hip h ull str ucture monitoring usin g  bre opt ic s ensors, Smart M aterial and Structure, 10, 472, 2001. 54. Davies, H.L., Everall, L.A., and Gallon, A.M., Application of smart optical  ber s ensors for str uctural load mo nitoring, in Proceedings o f S PIE S mart S tructure a nd M aterials, Industrial and Commercial Applications of Smart Structures Technologies, 2001, 4332, p. 114. 55. Udd, E., Schulz, W.L., Seim, J.M., Trego, A., Haugse, E., and Johnson, P.E., Use of transversely loaded ber grating strain sensors f or aer ospace applications, in Proceedings of SP IE, 2000, 3994, 96.

10/16/2008 7:16:07 PM

14-17

Fiber Optic Systems: Optical Fiber Sensor Technology and Windows

56. Okabe, Y., Mizutani, T., Yashiro, S., and Takeda, N., Detection of microscopic damage in composite laminates with embedded small-diameter  ber Bragg grating sensors, Composite Science and Technology, 62, 951, 2002. 57. Takeda, N., Okabe, Y., Kuwahara, J., Kojima, S., and Ogisu, T., Development o f sma rt co mposite str uctures wi th smalldiameter ber Bragg grating sensors for damage detection: Qualitative evaluation of delamination length in CFRP laminates usin g la mb wa ve s ensing, Composites S cience a nd Technology, 65, 2575, 2005. 58. Guemes, A., Pintado, J.M., Frovel, M., and Menendez, J.M., FBG bas ed S HM syst ems in mo nolithic co mposite str uctures for aerospace, in Proceedings of the S econd European Workshop on Str uctural H ealth and M onitoring, M unich, Germany, 7–9 July, 2004, 870. 59. Akhras, G., S mart ma terials a nd sma rt syst ems f or the future, Can. Military J., 25, 2000. 60. Zou, L., B ao, X., R avet, R ., a nd Chen, L., Distrib uted Brillouin ber sensor for detecting pipeline buckling in an energy pipe under internal pressure, Applied Optics, 45(14), 3372, 2006.

14.2 Smart Windows John Bell 14.2.1

Outline

Smart w indows en compass a w ide r ange o f te chnologies. Ā e common feature of all smart window technologies is the ability to s witch t he opt ical t ransmittance a nd re ectance o f t he window. W hile t his c haracteristic i s o f i nterest f or a r ange o f applications, including optical computing, display devices, and architectural glazing, the principal application considered here for sm art w indows i s a rchitectural g lazing. Ā i s application places a number of constraints on the technology under consideration, including two particularly important constraints: large area (up to 2 × 2 m); and long lifetime (in excess of 10 4 switching cycles for up to 30 years). Ā is chapter only considers technologies that target this application, although many of the technologies discussed also have applications in other areas, as outlined in the major recent reviews by Granqvist et a l. [1]. Ā e chapter a lso only considers technologies t hat a re a ctively u nder d evelopment for a rchitectural glazing applications. As a re sult, some technologies, such as photochromic materials, are not discussed despite signicant research in the 1980s. An introduction to smart windows is given in Section 14.2.2, followed by a discussion of the physics of building windows in Section 14.2.3, which is necessary to understand the differences between t he t ypes o f sm art w indows a nd sm art w indow te chnologies, w hich a re d iscussed i n S ection 1 4.2.4. Ā is chapter does not at tempt to en capsulate a ll t he i nformation re garding

43722_C014.indd 17

electrochromic materials, but focuses on t he key principles for smart window applications, and some of the key areas of critical importance f or t hese app lications. D etailed re cent re views o f electrochromic materials can be found elsewhere [1–3]. Inorganic ele ctrochromic smart w indows have b een s tudied for lo nger t han mo st ot her s ystems, a nd s till app ear to b e t he most likely candidates for commercial development in the near future. Ā ey h ave m any a dvantages ove r ot her t echnologies, either in the energy benets, which can result from their use, the ease of control of t he devices, a nd t he ability to i ndependently control the devices for a range of different environmental conditions. Ā ey a re a lso s pecularly t ransmitting i n a ll s tates. Consequently, Section 14.2.5 is devoted to a mo re detailed discussion of the materials, device structures, and control methods for ele ctrochromic sm art w indows. S ection 14.2.6 g ives a b rief consideration of future trends in smart window systems.

14.2.2

Introduction

Ā e term “smart window” was introduced in the mid-1980s by Professor Claes Granqvist to de scribe optically switchable electrochromic glazings. Ā ese devices exemplify the fundamental characteristic of all “smart windows”: controllable variation in the optical transmittance of the window. In the case of electrochromic smart windows, the variation in optical transmittance occurs through the double injection of electrons and ions (usually H + or L i+ ions) into an electrochromic material such as WO3. In WO3, t his leads to t he development of either a b road absorption band or a reection edge (depending on the details of t he m aterial p reparation), le ading to a lo w t ransmittance state. Ā e process is (ideally) reversible, with the extraction of the ions and electrons returning the material to a t ransparent state. Ā e control of the process is accomplished by the application of a small voltage (or the passing of a small current), which controls the ion injection and extraction process. Since the initial discovery of electrochromism in thin  lm WO3 by Deb et al. [4] in 1969, there has been an enormous amount of research directed toward the goal of large area switchable windows for architectural applications. Ā is initially focused on electrochromic systems, covering a w ide r ange o f m aterials a nd de vice s tructures. However during the 1980s and subsequently, several alternate optical switching systems have been developed, which fall into the gener al c ategory o f sm art w indows. Ā ese i nclude ot her electrically ac tivated s ystems suc h a s su spended pa rticle devices [5] and phase-dispersed liquid crystals [6], temperaturecontrolled s witchable de vices suc h a s t hermochromic [7] a nd thermotropic devices [8], and recently developed gasochromic devices, which are controlled by the use of reducing or oxidizing gas mixtures in a w indow unit [8]. Recent research in t he past 5 years has explored a vast range of new materials for electrochromism, including many organic materials, used in both devices [9–11], and exploring basic electrochromic properties of new polymer systems (see, e.g., Refs. [3,12–14], and further references in Ref. [1]).

10/16/2008 7:16:07 PM

14-18

Architectural Glazing Applications for Smart Windows

1

Ā e p otential r eductions in h eating, c ooling, a nd li ghting energy u se, which c an result f rom t he u se of switchable g lazings in buildings, has provided the impetus for the majority of this research. Ā ere have been over 700 U.S. patents lodged in this eld since 1976. It is now well established that reductions in energy consumption of up to 5 0% are possible with the use of electrochromic glazings in commercial buildings, with savings o f 2 0%–30% o btainable in m ost c limatic c onditions. Savings are highest in hot climates such as the Middle East (up to 49% energy reductions [ 15]), A ustralasia, a nd t he t ropics [16,17], and southern Europe [18]. However the energ y performance is very dependent on climate and building type, as well as t he p roperties o f t he ele ctrochromic g lazing. T wo re cent studies conclude that use of electrochromic smart windows is not justiable on economic grounds in Hong Kong [19] as only 6.6% energ y reduction could be achieved in a h igh-rise apartment building. Windows are also the primary application for thermochromic a nd t hermotropic w indows, w hich c hange t ransmittance as a f unction of temperature, a nd can t herefore reduce transmittance of infrared radiation (i.e., heat) when the temperature is high. Studies of thermotropic smart windows also show that signicant energ y s avings c an b e obtained i n appropriate c limates [20]. However, some of the other “smart” windows, such as the polymer-dispersed l iquid c rystal ( PDLC) de vices, do not provide s ignicant energy savings, but are used as privacy screens. In order to understand the role of switching in a smart window, i t i s ne cessary to u nderstand t he nat ure o f g lazings and h ow t he opt ical p roperties o f a w indow a ffect its performance and utility. 14.2.3.1

Physics of Windows

Common to a ll applications of glazing materials is the need for transmission of light, and in most applications, the reectivity is also very important. In order to quantify the light transmission properties, the transmittance and reectance spectra, T(λ) and R(λ) can be measured and used to dene several different average transmittance and reectance quantities. Ā e most important of these are the solar and visible transmittance and reectance: ∞

Tsol =

770

Tvis =

∞

∫0 T ( l )jsol ( l )dl ∫ R ( l )jsol ( l )dl ; Rsol = 0 ∞ (1 ∞ j ( l )d l ∫0 sol ∫0 jsol ( l )dl

4.1)

770

∫370 T ( l )j vis ( l )dl ∫ R ( l )j vis ( l )dl ; R vis = 370 770 (1 770 ∫370 j vis ( l )dl ∫370 j vis ( l )dl

4.2)

where j sol and jvis are the solar spectrum and visible response of the h uman e ye, re spectively. U sually t he a ir m ass 1 .5 s olar spectrum is used to dene the solar averages, although it is not necessarily the most appropriate at all locations, and there can be s ignicant differences bet ween d ifferent s olar s pectra [21]. The diff erence i n t he s pectral ra nges f or j sol, j vis, a nd j th,

43722_C014.indd 18

jvis(l)

0.8 Relative intensity

14.2.3

Smart Materials

0.6

jsol(l) (AM1.5)

0.4

jth(l)

0.2

T =300 K ⫻1000

0 1

10 Wavelength (µm)

FIGURE 14.22 Ā ree c omponents re lated to t he a mbient r adiation environment, which need to be considered for design of windows. The three distinct spectral regions correspond to the wavelengths 0.3 < λ < 2.5 µm (solar radiation j sol(λ)), λ > 3 µm (thermal IR), shown here as j th(λ) for T = 300 K, and 0.37 < λ < 0.77 µm corresponding to the visible response of the human eye, j vis(λ).

illustrated in Figure 14.22, immediately gives rise to the concept of spectral selectivity, which is central to many advanced window glazing systems. Ā is refers to t he ability of a w indow to t ransmit, for example, visible radiation (high Tvis), but to re ect heat (low Tsol and high Rsol). Ā is is f ully explained in Ref. [22]. Ā e selectivity Tvis/Tsol, which has a maximum value of approximately 2, can therefore be used as a gure-of-merit for glazings. Another important optical quantity is the solar absorption, Asol = 1 − Tsol − Rsol. A w indow w ith h igh a bsorption o f s olar r adiation w ill heat up c onsiderably, and this heat can be transmitted into the building, ne gating t he re duction i n i nsolation, w hich re sults from the low solar transmittance [23]. While these are the basic quantities of interest, there are several derived quantities, such as the total solar energ y transmittance (TSET) [21], which are commonly used to classify window glazings. In the case of switchable glazings, the key parameters of interest are the change in the visible and solar transmittance between the two states of the window. Ā e values usually quoted are t he normal–normal t ransmittance a nd re ectance, i .e., t he fraction of an incident beam normally incident on the window, which is specularly transmitted. Ā e spectral response typical of different t ypes of smart w indows is i llustrated i n Figure 14.23. While most windows are designed to have low diffuse scattering, many of the smart windows available or in development switch between a specularly transmitting and reecting state, and a diffusely transmitting and reecting state. In this case the normal– hemispherical or s ometimes t he he mispherical–hemispherical transmittance and reectance are quoted. Even for the purposes of minimizing energy use in buildings, there is no single  gure-of-merit for smart windows, which can be used to identify the optimum smart window. In general, the important characteristics, and those quoted in this work where

10/16/2008 7:16:07 PM

14-19

Fiber Optic Systems: Optical Fiber Sensor Technology and Windows

Type A

Type B

T

T

Type C

Type D T

T

Visible

Near IR

Near IR

Visible

l (nm)

l (nm)

FIGURE 14.23 Schematic illustration of t he different types of s witching of t ransmitted radiation, which are possible with different switchable window systems.

14.2.4 Survey of Smart Windows Ā e principal smart w indow systems a re su mmarized i n Table 14.4. Ā is table shows the switching mechanism, the method by which this mechanism is operated, whether the device exhibits memory and the type of spectral response (selectivity and contrast)

43722_C014.indd 19

that occurs. Ā e table also includes a control parameter, which in most cases is the same as the operational switching parameter. However, in the context used here, the control parameter is one which can be measured and provides a degree of control over the optical state of the device. Ā is is extremely important for practical

1.2 1 0.8 ∆OD

they are available, are the selectivity, dened above, and the contrast ratio, or amount of switching, in both the visible and solar clear/T colored and T clear/T colored respectively. Ā e spectral ranges: Tvis vis sol sol contrast ratio is frequently quoted as the change in optical density, ∆OD = log e Tclear/Tcolored. A c omplete de scription o f t he window behavior involves the variation of the selectivity or contrast a f unction of t he switching parameter, as is schematically illustrated f or s everal d ifferent de vices i n F igure 1 4.24. S ome authors prefer to u se the g-factor or TSET rather than the solar transmittance, as this include the effects of thermal conduction through the window and secondary heat gains which arise from absorption i n t he g lazing [21]. A go od si ngle number w hich i s frequently used to rate a smart window is the ratio of the maximum v isible t ransmittance to t he m inimum v alue o f T SET clear/T colored). Another approach to de ning the performance of (Tvis vis smart w indows i n b uildings i s t he energ y c ost f or a bu ilding when different types of windows are used. Selkowitz et al. [24] introduced t his c oncept t hrough a g raph, w hich c ompares t he cooling and lighting costs (the costs most affected by advanced glazing in commercial buildings) for diff erent t ypes of s witchable glazings. Ā ere are other important aspects of windows and the materials used to make windows, such as durability [25], cost [26], color [27] and haze, or the amount of radiation that is diffusely transmitted [28].

0.6 0.4 0.2 0

Control

FIGURE 14.24 Variation in opt ical density for fou r differently electrically activated smart window devices. ( ) A sol–gel deposited device manufactured by Sustainable Technologies Australia Ltd, and based on WO3; (×) a sputtered device manufactured by Asahi Glass Co, and based on a W O3-NiO c omplementary d evice ( see S ection 1 4.2.4.2); ( +) a n organic electrochromic from Gentex Corporation, based on v iologens; (▲) a P DLC d evice m anufactured by 3 M C orporation. Ā e different curves represent different spectra used i n calculating t he opt ical density: — — a bro adband t ransmittance; – – – t he v isible t ransmittance Tvis; and — — — the solar transmittance Tsol. Ā e divergence in the optical density for visible and solar transmittance for the Gentex viologenbased d evice i s a re ection of t he t ype of s pectral c hange ( Type D i n Figure 14.23).

•

10/16/2008 7:16:08 PM

43722_C014.indd 20

Reduction of electrochromic (e.g., WO3) by electrochemical ion insertion

Metal–insulator transition

Phase separation leading to multiple scattering

Alignment of liquid crystal particles in electric eld

Electrochromic (inorganic)

Ā ermochromic

Ā ermotropic

PDLC

Gasochromic

Reduction of gasochromic (e.g., WO3) by catalytic ion insertion

Suspended particle Alignment of suspended devices particles in electric eld

Reduction of organic species (e.g., viologen)

Mechanism of Switching

Electrochromic (organic)

System

Control Parameter

Applied voltage

Linear response 0–50 V, saturation by ∼100 V

Gas control concentration Concentration of H2 Nonlinear response to gas of H2 in window cavity in window cavity concentration

Power 2–10 W/m2

Electrical control (50–200 V AC)

Highly nonlinear response

Glazing temperature Gradual transition (in (°C) hydrogel, from 30°C to 65°C)

Type A and B

Type D

Type D

Type A

Type C

Glazing temperature Rapid change at Tc (°C) (transition point)

Electrical control (∼100 V Applied Voltage AC—voltage depends on system)

Temperature (Tswitch depends on material)

Temperature (Tswitch depends on material)

Type D

Spectral Response (See Figure 14.23)

Type A and B

Ā reshold voltage, then approximately linear response to saturation

Switching Behavior

Charge injected into Approximately linear electrochromic layer change in optical density (mC/cm2) with charge

Electrical control–voltage/ Voltage current (approximately 2–3 V DC; 50–1000 µA/cm2)

Operational Switching Parameter

TABLE 14.4 Characteristics of Smart Window Systems

Yes

No

No

N/A

N/A

Yes

No

Memory

17/19

2.8/2.5 specular

1.2–1.7/ 1.0–1.5 specular

Specular

N/A

4.2/4.9

14/-

1.6/1.5

15/16 (cloud gel)

1.1/1.2 (cloud gel) scattering/specular 0.9/1.0 scattering/ specular (with haze)

4.6/5 (hydrogel)

1.5–1.8/1.3–1.7 (hydrogel)

∼1/1.1–1.6

17/26

2.3/1.55 specular 1.0–1.2/<1 specular

3–10/3–11 depends on system

13/2.5

Visible/Solar

Contrast

1.4–1.9/1.3 specular

0.2/1.3 specular

Colored/Clear

Selectivity

[8,51,52]

[5]

[6,29]

[8,47]

[7,43]

(Sea green)

[30] (classic)

[29]

[29]

References

14-20 Smart Materials

10/16/2008 7:16:09 PM

14-21

Fiber Optic Systems: Optical Fiber Sensor Technology and Windows

applications of smart windows where it is essential to be able to control t he state of a de vice, as well as to b e able to s witch t he device. As can be seen in Figure 14.24, of the electrically activated devices, the inorganic electrochromic devices provide the most l inear re sponse w ith re spect to t he c ontrol pa rameter, which in this case is injected charge density.

and t his i s d iscussed i n S ection 1 4.2.4. Ā e mo st w idely u sed counterelectrode materials are cerium-titanium oxide (CeO2-TiO2) [32] v anadium o xide ( V2O5) [33] a nd der ivatives. I n a W O3-V2O5 device, the complementary reaction in the counterelectrode is illustrated in Equation 14.3b, with Li+ as the mobile ionic species [34]:

14.2.4.1 Electrochromic Smart Windows Electrochromic m aterials u ndergo a c hange i n t heir opt ical properties when an electric eld is applied. Ā ere is a wide range of b oth o rganic a nd i norganic m aterials t hat e xhibit ele ctrochromism. Inorganic electrochromic materials, most commonly transition me tal oxides, c hange f rom a c lear to a c olored s tate when ions and electrons are inserted into their lattice, under the application o f a sm all D C, le ading to opt ical a bsorption. Extraction of these ions and electrons is effected by a reversal of the ele ctric  eld a nd t he m aterial re turns to t he c lear s tate. Organic electrochromic materials i nvolve stoichiometric redox couples, w here t he o xidized a nd re duced f orms h ave d ifferent colors. Ā e app lication o f a n ele ctrical p otential su fficient to change t he o xidation s tate o f t he m aterial t herefore c auses a color change. 14.2.4.1.1 Inorganic Electrochromic Smart Windows Many inorganic electrochromic materials maybe produced in thin lms o n tr ansparent e lectrically c onducting s ubstrates, t hereby allowing their use in window fabrication. Tungsten oxide (WO3) is the most commonly used inorganic electrochromic material and is the c ompound i n w hich ele ctrochromism w as  rst observed. Tungsten o xide c an b e c olored b y t he i njection o f ele ctrons a nd small ions including H+, Na+, K+, and Li+, as schematically illustrated in the reaction scheme in Equation 14.3a. Intercalation of lithium and electrons into tungsten oxide forms the tungsten bronze LixWO3 (reversible for 0 < x < 0.4 [31]), which is deep blue in color. In order to fabricate an electrochromic device, a source of ions and electrons is required. Ā e most common device structure being used for inorganic electrochromic devices is illustrated in Figure 14.25. Āer e are a number of variants of this device structure with different components,

−V

⎯⎯⎯ → Li x WO3 Working electrode WO3 + xLi+ + xe − ←⎯ +V

transparent

Li+

deep blue

+V

⎯⎯⎯ → Li y − xV2O 5 + xLi+ + xe − Counterelectrode (14.3b) Li yV2O 5 ←⎯ −V

Ā e c hange i n opt ical den sity i s u sually p roportional to t he charge den sity i njected i nto t he ele ctrode [ 35,36], a nd t he coloration efficiency, CE = ∆OD/Qin, is frequently used to characterize t he performance of a  lm or a de vice. For WO3-based electrochromics, a t ypical c oloration e fficiency i s b etween 5 5 and 70 mC/cm2 at λ = 700 nm [37]. Ā e ele ctrochromic m aterial c hanges i ts opt ical c onstants, usually through the development of an absorption band [38], or though an increase in t he effective f ree electron density i n t he material [ 39]. Ā e se g ive r ise to opt ical c hanges k nown a s “absorption mo dulation” a nd “ reectance mo dulation” [ 22], respectively, according to w hether the change in transmittance results from an increase in optical absorption or reectance. Reectance modulation is the preferable optical change for glazing materials, as these materials do not absorb radiation signicantly in the colored state, minimizing secondary heat gain, and reducing t he T SET v alue of t he w indow. Ā e specic behavior depends on the microstructure of the material, with crystalline WO3 lms exhibiting reectance modulation, and amorphous WO3  lms showing absorption modulation. Ā e coloration efficiency spectra i n F igure 14.26 i llustrate t he c hange i n t he a bsorption characteristics of sputtered WO3 as the deposition temperature is increased. Despite the better performance that can be obtained with reectance modulation, t he majority of prototype devices

V

Glass

(14.3a)

Glass

Glass

Conducting ITO WO3 Ion conductor Counter electrode

FIGURE 14.25 Spectral c oloration e fficiency, w hich i s d irectly prop ortional to c hange i n opt ical d ensity i n s puttered WO3. Ā is s hows t he development of an absorption band in WO3 during ion injection. Note the shift in the absorption band peak position as the substrate temperature increases and the  lm becomes more crystalline, representing the change from absorption to reection modulation.

43722_C014.indd 21

10/16/2008 7:16:09 PM

14-22

Smart Materials

which will occur in architectural glazing applications of smart window. P olyaniline i s a p romising c andidate f or o rganic electrochromic sm art w indows d ue to i ts h igh s tability a nd ability to undergo multiple color changes [40,41].

300

Coloration efficiency (cm2/C)

250

14.2.4.2

200

150

100

47⬚C 200⬚C 300⬚C 350⬚C 400⬚C

50

0 300

800

1800 1300 Wavelength (nm)

2300

FIGURE 1 4.26 Schematic i llustration of t he s tructure of a t ypical inorganic electrochromic device, showing the ve-layer structure.

produced to date use absorption modulation, as it is hard to produce crystalline WO3 in conjunction with the other layers in the device. Devices ba sed o n i norganic ele ctrochromics a re u sually spectrally s elective to s ome de gree (type B b ehavior i n F igure 14.23), and it is possible to strongly modulate the solar transmittance of the window, while still maintaining a mo derate visible transmittance, thus allowing signicant daylighting in a building. Ā is m akes W O3 a n e xcellent c andidate f or sm art w indows where a lo w s olar t ransmittance i s a m ajor re quirement a nd daylighting is a direct advantage. 14.2.4.1.2 Organic Electrochromic Smart Windows Ā e absorption coefficients of organic electrochromic materials are gener ally e xtremely h igh w hen c ompared to i norganic materials, so it is possible to produce devices that undergo drastic optical changes involving very intense colors. Āe majority of o rganic ele ctrochromic m aterials a re c onducting p olymers. Some of these organic materials, such as polyaniline, have different colors for the various oxidation states, so it is possible to achieve multiple coloration by a stepwise oxidation or reduction [40]. Ā e t ransmittance o f o rganic ele ctrochromic m aterials depends o n t he s pecic c ompound; h owever, mo st o f t hese materials s witch p rimarily i n t he v isible re gion, w ith l ittle modulation of solar transmittance when compared to inorganic systems (type D spectral behavior). Ā is makes organic electrochromics less suitable for smart window applications than inorganic electrochromics. Ā e mo st w idely s tudied f amily o f o rganic ele ctrochromic materials i s t he v iologens, w hich a re c urrently u sed i n a c ommercial, va riable r eectance m irror for aut omobile re ar v iew mirrors ( Gentex C orporation). I n t his app lication, i t i s f ree from many of the problems associated with solar/UV degradation,

43722_C014.indd 22

Thermochromic Devices

Ā ermochromic materials change their optical properties when heated, a nd re turn to t he o riginal s tate w hen c ooled to t he initial tem perature [ 42]. Ā e m ost w idely st udied i norganic thermochromic material, and the best candidate for large area thermochromic smart window applications, is VO2 and various doped forms of VO2 [43]. Ā is material undergoes a metal–insulator phase transition at 6 8°C, exhibiting metallic properties at high temperature and associated infrared reectivity, as shown in Figure 14.27a. Ā e visible part of the spectrum is largely unaffected by the transition (type C spectral behavior), so the principal change that occurs at switching is in the solar transmittance. Ā e visible transmittance decreases slightly, but there is a moderately s trong a bsorption ba nd i n t he v isible s pectrum i n t he low-temperature i nsulating s tate, w hich me ans t hat t he v isible transmittance is relatively low in both states. Ā e major area of research for development of large area smart windows, which are useful for energy control in buildings, is the development of materials that have a higher visible transmittance, and, m ost im portantly, wi th a tr ansition t emperature c lose t o room temperature. Ā is means that the windows will switch to a low t ransmittance s tate to p revent o verheating i n t he b uilding. Both of these areas have been the subject of considerable research with n umerous e fforts to dope VO2 wi th v arious e lements t o achieve t hese goals. Fluorination of VO2 [7] increases t he v isible transmittance from 27% to 35% (in the low-temperature state), but also reduces the effect of switching, with a change in solar transmittance from 44% to 40% at the transition. However, the efforts to reduce the transition temperature have been much more successful, wi th a s ignicant r eduction in t he t hermochromic transition temperature from 68°C to approximately 10°C for W0.032V0.968O2 [43,44]. Ā is is illustrated in Figure 14.27b. Apart from the relatively poor spectral selectivity and contrast ratio of t hermochromic devices, t he i nability to i ndependently control the state of thermochromic devices makes it difficult to envisage wide ranging applications as smart windows. 14.2.4.3 Thermotropic Devices Ā ermotropic devices, originally classed with thermochromic devices in the literature [45,46], also change their optical properties w ith i ncreasing tem perature. H owever, t he me chanism o f the change is su fficiently different f rom t ypical i norganic t hermochromic materials that they are now recognized as a different class o f m aterial. Ā ermotropic m aterials c hange b etween a specularly t ransmitting t ransparent s tate to a w hite, d iffusely reecting state as their temperature rises. Ā e change is reversible. Ā is c hange o ccurs t hrough a ph ase s eparation, w hich occurs in the materials at elevated temperatures. Ā is transforms the material from a homogeneous optical material to a heterogeneous, t wo-phase material at high temperature. Ā e size of t he

10/16/2008 7:16:09 PM

14-23

Fiber Optic Systems: Optical Fiber Sensor Technology and Windows 100

100 W0.032V0.968O2

20⬚C

50

Transmittance (%)

Transmittance (%)

VO2

50 20⬚C

80

0 0

80

1 2 Wavelength (µm)

(a)

0 0

1 2 Wavelength (µm)

(b)

FIGURE 14.27 Transmittance of (a) VO2 and (b) W-doped VO2 at temperatures above (80°C) and below (20°C) the thermochromic transition. Ā e spectra are of t ype C i n both cases (see Figure 14.23). (Reprinted f rom Sobhan, M.A., K ivaisi, R .T., Stjerna, B., a nd Granqvist, C.-G., Solar Energy Mater. Solar Cells, 44, 451–455, 1996. With permission.)

particles of the two phases is such as to lead principally to backscattering of incident radiation [47]. Ā ere are two main types of thermotropic materials: hydrogels and polymer blends. Both systems have a si milar device structure, as illustrated in Figure 14.28. In both cases, the materials form h omogeneous t ransparent m ixtures at lo w temperatures, in which the two components are bound together at a molecular or molecular cluster (macromolecular) level. Ā e bonding is usually h ydrogen b onding a nd t hese b onds c an b e b roken a s t he temperature increases from room temperature. Once the bonds break, the materials separate into distinct domains with different refractive i ndices a nd c ause s cattering. I f t he pa rticle si ze i s appropriate ( close to t he w avelength o f l ight o ver t he s olar spectral range), then strong scattering occurs. 14.2.4.3.1

Hydrogels

Ā is is one class of thermotropic materials and the thermotropic material CloudGel, commercially available from Suntek, is of this

type. As the name suggests, a hydrogel is a water–polymer mixture. Another example of these materials is the TALD hydrogel, which i s a m ixture o f c olloidal pa rticles o f a p olyether–water mixture, emb edded i n a c arboxyvinyl p olymer a nd w ater gel [47]. Ā e colloidal pa rticles a re soluble i n t he gel at ro om temperature a nd b elow, b ut a s t he tem perature i ncreases, w ater i s expelled from the colloidal particles and the polyether particles become more dense and increase their refractive index. Ā e s witching temperature of t he hydrogel materials c an b e varied between 5°C and 60°C by changing the proportions of the components. T he m aterials s witch o ver a re asonably w ide temperature range, as illustrated in Figure 14.29. Ā e size of the particles i n t he ph ase-separated s tate, a nd t he t hickness o f the hydrogel l ayer ( typically 1 mm) en sure t hat t he m ajority of t he r adiation i s bac kscattered b y m ultiple s cattering i n t he high-temperature state (see Figure 14.29). Ā is means that these thermotropic m aterials h ave go od p otential a s energ y-efficient windows.

Homogeneous mixture

Glass cover

(a)

Low temperature state

Scattering layer

Glass cover

Glass cover

(b)

Glass cover

High temperature state

FIGURE 14.28 Structure of a t hermotropic l aminate. I n t he lo w t emperature s tate (a), t he d evice i s f ully t ransparent, ( b) but i n t he h ightemperature state the thermotropic state separates into discreet particles and the layer becomes scattering.

43722_C014.indd 23

10/16/2008 7:16:10 PM

14-24

Smart Materials

Normal-hemisph. reflectance r nh (%)

100

80

60

60⬚C 45⬚C 50⬚C

40

20

40⬚C 35⬚C

0

30⬚C 500

1000

1500 Wavelength (nm)

2000

2500

FIGURE 14.29 Normal-hemispherical re ectance of a t hermotropic h ydrogel l aminate at a r ange of t emperatures b elow ( 30°C) a nd a bove (>35°C) the thermotropic transition temperature. Ā e spectra are of type B, demonstrating a high degree of selectivity.

14.2.4.3.2 Polymer Blends Polymer blen d t hermotropic m aterials h ave t he adv antage o f not containing l iquid w ater. A n e xample o f a p olymer blen d hydrogel is a h igh refractive index (1.585) cross-linked matrix of polystyrol and polyhydroxy ethyl methacrylate, with a second, low re fractive i ndex (1.45) ph ase of p olypropylene oxide i nterpenetrating the matrix. At low temperatures, the lower refractive index component is hydrogen bonded to t he matrix. At higher temperatures, t hese b onds b reak a nd t he p olypropylene o xide component f orms sm all i nclusions i n t he m atrix, le ading to scattering. Ā e size of the scattering particles and the transition temperature depend on the proportions of the two components and the degree of cross-linking of the matrix.

which exhibit signicant ordering properties. Ā ere are several classes of electro-optic devices, which can be formed using liquid crystals, b ut o nly t wo c lasses, g uest–host s ystems a nd P DLCs, have b een e xtensively i nvestigated f or u se i n sm art w indows. Lampert [48] provides an excellent introduction to the physics of liquid crystals, and Basturk and Grupp [49] and Montgomery [6] provide good reviews of the basics of both types of liquid crystal devices.

14.2.4.3.3 Applications Ā e principal factors constraining the application of thermotropic devices a re t he i nability to i ndependently c ontrol t he de vice transmittance and the scattering state at high temperatures. Ā e latter means that thermotropic devices are unlikely to be used in view w indows, b ut t hey h ave app lications i n s kylights a nd f or overheating protection on brick facades. Ā ese applications have been extensively i nvestigated [47], a nd large a rea w indows c an be routinely produced. Figure 14.30 shows a number of thermotropic windows under partial shading, illustrating the sensitivity to temperature. Ā ese windows were part of a study to investigate energy savings and acceptance of the windows by building occupants. In this work, signicant energy savings were reported (up to 17%), and over 75% of building occupants reported improved comfort conditions [20]. 14.2.4.4 PDLC Devices Ā ere a re m any c lasses o f l iquid c rystals. M ost l iquid c rystals consist o f t hin, ne edle-like, o r ro d-shaped o rganic mole cules,

43722_C014.indd 24

FIGURE 14.30 Ā ermotropic windows (in various different congurations of i nner and outer glass in each separate pane), illustrating the transition between the transparent (low temperature, and in this case, unshaded state) and the opaque state. (From Inoue, T., Energy Build. 35, 463, 2003. With permission.)

10/16/2008 7:16:10 PM

14-25

Fiber Optic Systems: Optical Fiber Sensor Technology and Windows V Glass Glass

Glass

Conducting TCO Dielectric PDLC Layer

FIGURE 14.31 Structure of a PDLC device. Unlike the electrochromic device, it is not a conducting device, with the switching dependent on the electric eld across the PDLC layer.

Ā e principal liquid crystal devices currently available are the PDLC devices, which are commercially available from 3M under the name “Privacy Film.” Ā is device consists of a PDLC lm sandwiched between indium-doped tin oxide (ITO)-coated sheets of glass as illustrated in Figure 14.31. Ā e PDLC  lm consists of spherical liquid crystal droplets embedded in a polymer matrix. Ā e t ypical d roplet d iameter i s a bout 1 µm o r le ss, i .e., o f t he same order of magnitude as light wavelengths. On application of a voltage to t he ITO electrodes, the liquid crystal molecules are orientated i n t he  eld dir ection, r esulting in in dex m atching between t he d roplets a nd t he m atrix a nd t hus h igh t ransmittance for light propagating parallel to the applied eld. When the eld is sw itched o ff, t he l iquid c rystal mole cules re orient r andomly, t he ( unmatched) v alue o f t he e xtraordinary re fractive index also becomes effective, and causes scattering at the droplet surfaces; the  lm appears milky. While PDLC devices act very effectively as a p rivacy screen, with essentially zero specular transmittance in the zero voltage state, and good transmittance but w ith some diffuse scattering in t he o n s tate, t he de vices a re not pa rticularly e ffective for energy conservation as the majority of the scattered radiation is

Total transmittance (%)

100 100 V 10, 20, 30 V 0V

80

3 M PDLC film

60 40

20 0 0.25

0.5

1 Wavelength (µm)

2

4

FIGURE 14.32 Normal-hemispherical transmittance spectra for t he PDLC device for applied voltages of 0, 10, 20, 30, and 100 V. Ā e device is opaque (in the sense of being unable to perceive images through the device i n t he off s tate [ 0 V]), but s till t ransmits si gnicant energy. (Reproduced from Dr. Arne Roos. With permission.)

43722_C014.indd 25

forward s cattered. Ā is is illustrated in Figure 14.32, which shows t he tot al t ransmittance (including t he d iffuse transmittance). While the device is totally opaque in the “off ” state (0 V), the tot al t ransmittance i s s till sub stantial. A nother d rawback with the PDLC devices, and all liquid crystal systems, is the lack of me mory i n t he d evices. Ā is is i ntrinsic t o l iquid cr ystal systems, a s t he a lignment o f t he l iquid c rystal mole cules i s caused by dipolar polarization in the applied eld. Ā e curves in Figure 1 4.32 a lso s how t he h ighly no nlinear b ehavior o f t he switching with the applied voltage, which is the control parameter in Figure 14.23 for the PDLC device. 14.2.4.5

Suspended Particle Devices

Suspended particle devices, also termed light valves, consist of a suspension of  ne particles, which in the “off ” state (no applied voltage) are strongly absorbing, but which can be aligned by the application of a n e lectric  eld, le ading to a t ransparent “ on” state. Ā e de vice re quires t he su spension to b e c onstrained between c onducting ele ctrodes, a nd i n t he l arge a rea de vices currently b eing c ommercialized b y Ha nkuk G lass, t hese a re deposited on a PET lm, which is then laminated to glass [5]. Āe suspended pa rticles a re u sually ne edle-shaped pa rticles o f polyiodides o r pa rapathite app roximately 1 µm lo ng, su spended in an organic l iquid [45]. I n c urrent de vices, me thods have b een de veloped to en capsulate m icrodroplets c ontaining the suspended particles in a polymer matrix. Ā e device structure i s e ssentially identical to t he PDLC de vice (Figure 14.30), with the suspended particle layer replacing the PDLC layer. In t he a bsence o f a n app lied ele ctric  eld, t he pa rticles a re randomly arranged and absorb the incident radiation. Ā e absorption in all states (from zero applied voltage to the maximum applied vol tage) de pends o n t he c oncentration o f pa rticles i n the  lm, and the color can also be altered by the choice of particles. I ncreasing t he app lied vol tage i ncreases t he t ransmittance approximately linearly to about 50 V, and the transmittance then starts to s aturate, reaching saturation at a bout 100 V. Ā e major drawback with the devices for use as smart windows is that they do not prov ide si gnicant s pectral s electivity ( gures f or s olar transmittance are not available) or signicant reductions in solar transmittance. As with PDLC devices, they are principally privacy and v isual c omfort de vices, h owever, t hese de vices h ave a h igh viewing angle unlike liquid crystal devices.

10/16/2008 7:16:11 PM

14-26

14.2.4.6

Smart Materials

Gasochromic Devices

14.2.4.6.1 Gasochromic Technology

Gasochromic smart windows are devices that color by a complex process involving the adsorption and dissociation of a gas onto a catalyst, f ollowed b y a “ spill o ver” [50] o f t he d issociated ga s molecules i nto t he ga sochromic l ayer c ausing c oloration [8,51–53]. G asochromic de vices a re si milar to ele ctrochromic devices in that they both color when another species is inserted into the atomic network of the host material. Tungsten oxide is a good material for both electrochromic and gasochromic devices. Ā e major differences between electrochromic and gasochromic devices l ie i n t he d ifferent me chanisms o f c oloration a nd t he relative structural simplicity of the gasochromic devices. Simple gasochromic  lms have a laminar structure involving a substrate (usually glass), the gasochromic layer, and a catalyst layer. Ā e se  lms a re f abricated i nto ga sochromic sm art w indows by adding another pane of glass a short distance from the catalyst surface, so that a small cavity is formed inside the device, as s hown i n F igure 1 4.33. V arious ga s m ixtures c an t hen b e introduced into this interior cavity to color and bleach the window. G asochromic t ungsten o xide c olors ac cording to t he following simplied redox reaction, when hydrogen gas is used to induce coloration: xH + WO3
H x WO3 transparent

(14.4)

blue

Ā is reaction is essentially identical to t he reaction involved in the coloration of electrochromic WO3. Ā e principal difference is that the catalyst layer allows the dissociation of H2 gas into an activated species, which can diff use into and react with WO3. In an electrochromic system, the positive and negative charges are introduced via an electrical circuit and the polymer electrolyte and counterelectrode are used to s eparate t he charge and store cations. Pure hydrogen is not required to color gasochromic lms and very d ilute, no nexplosive m ixtures c an b e u sed. Ble aching o f gasochromic lms is usually accomplished by lling the window cavity with a dilute oxygen mixture; however, bleaching can also be achieved using argon or a vacuum [52]. Partial coloration or bleaching o f ga sochromic de vices c an b e ac hieved b y re placement of the gas in the window cavity with an inert gas such as argon once a suitable level of coloration has been attained.

Glass cover

Glass cover

Gas mixture

FIGURE 14.33 Typical structure of a gasochromic device.

43722_C014.indd 26

Gasochromic te chnology h as s temmed l argely f rom re search i nto electrochromism, and tungsten oxide is the most commonly reported gasochromic material [8,51,52,53]. Platinum is frequently used as the catalyst layer due to its ability to dissociate hydrogen and oxygen gas molecules. S ome ne w s ystems w ith a ga sochromic l ayer ba sed on alloys of lanthanides have been recently reported [54–56]. Catalystcoated lanthanum hydrides (LnHx) undergo extreme color changes when x is varied from 2 (lm reective in blue region/transmitting in red) to 3 (yellow/transparent) [54]. Exposure of catalyst-covered LnH2 lms to h ydrogen gas quickly results in the formation of the transparent LnH3 species. W hen exposed to a ir or a v acuum, the lm reverts back to the LnH2 form. Ā e function of the catalyst is to a llow t he d issociation of gas molecules, but a lso prevents t he underlying  lm f rom r apid o xidation i n a ir [56]. Ā e add ition o f magnesium reduces the absorption in the transparent state and the transmission in the opaque state, while increasing the opaque state reectivity, a nd a lso h as t he adv antage o f p roducing mo re c olorneutral windows [54,56]. A particularly impressive system involves a gadolinium-magnesium a lloy G d40Mg60 that is reported to switch from a silver white mirror with almost no transmission and high reection, to a fully transparent color-neutral state [54]. 14.2.4.6.2 Applications Ā e function of the glass pane used opposite the gasochromic layer is to contain the gas used for switching, and ideally has a high visible transmittance. Ordinary oat glass is suitable, however, glass with low-emittance coatings can be used to improve the thermal performance. Owing to the small number of layers required for gasochromic smart windows, the dynamic transmittance range is very high when c ompared to ele ctrochromics a nd ot her c hromogenic g lazings. Gasochromic smart windows have been reported with visible transmittances ranging from 76% (bleached state) to 18% (colored state). Ā e advantages associated with the relative structural simplicity of gasochromic smart windows are somewhat off set by the complexity of the infrastructure needed to hold and transfer the gas mixtures required for their operation. Ā e gas system must be capable of storing gas, preparing and  ltering gas mixtures, and delivering these to a ba nk o f m ultiple w indows. Re ference [8] off ers some innovative solutions to the problems of gas support. As for all smart windows, lifetime is a critical issue for gasochromic devices. Ā e principal degradation mechanism for gasochromics is catalyst poisoning and varying water content of the  lms [53]. Recent results are promising however, and one gasochromic smart window 1.1 × 0.6 m in size has been cycled 10,000 times with no signicant reduction in its performance [52].

WO3

14.2.5 Electrochromic Smart Windows

Catalyst

As noted in Section 14.2.2, one of the most promising candidates for a f uture c ommercial sm art w indows i s t he ele ctrochromic glazing based on inorganic electrochromic materials. Inorganic electrochromic smart windows are also the closest to commercial production f or l arge a rea a rchitectural app lications w ith sub stantial p otential to re duce building energ y u se a nd hence t he

10/16/2008 7:16:11 PM

14-27

Fiber Optic Systems: Optical Fiber Sensor Technology and Windows

discussion b elow i s ba sed o n t hose w indows w ith i norganic materials, and in particular tungsten oxide. Ā e method of w indow control is a lso very i mportant a s, i n order to optimize the energy performance, it is essential to be able to set specic optical states independent of temperature and irradiance incident on the window, and the lifetime of the electrochromic lms and entire devices can be reduced by inappropriate control methods [57]. Various diff erent t ypes o f ele ctrochromic m aterial a nd window testing and control [58] have been used: constant voltage control, trapezoidal voltage waveform [25], cyclic voltammetry (triangular voltage waveform) [59], and constant current control [60]. Ā ere have also been many theoretical analyses of the charge injection and extraction processes. It is widely recognized that the coloration process is based on double injection of io ns a nd ele ctrons f orming t he t ungsten b ronze M xWO3 [34]. Ā ere a re v arious t heories re garding t he me chanisms limiting t he c oloration c urrent, i ncluding t he p resence o f a barrier at t he io n-injecting i nterface (electrolyte/WO3 in terface) [34,61]; diffusion-limited motion of ions within the electrochromic  lms [62,63] a nd t he s eries re sistance o f t he c ell [64,65]. I mpedance spectroscopy has been u sed by a n umber of authors recently to s tudy ion i njection k inetics i n electrochromics [66,67], and this holds some promise as a useful tool for the analysis of degradation of smart windows in operation. 14.2.5.1 Electrochromic Smart Window Structures Ā e basic multilayer electrochromic device structure is shown in Figure 14.25. An example of a smart window is shown in Figure 14.34. It is possible to classify various electrochromic devices according to the structure and materials used to achieve the ion injection a nd e xtraction processes i n a de vice. Ā ere a re t hree principal generic device types, which are essentially all the same as that shown in Figure 14.25, but with different approaches to the ion t ransport a nd ion storage layers. Ā e materials u sed i n these devices are discussed in more detail in Section 14.2.5. 14.2.5.1.1

Type 1—Ion Conducting Layer and Passive Counterelectrode

Ā e most widely used device structure has an ion transport layer (also c alled t he ele ctrolyte l ayer) f ormed b etween t he w orking electrode and an ion storage electrode (also called the counterelectrode). Ā e fundamental requirement of the ion storage layer is c harge c apacity, i .e., t he a mount o f c harge (number o f io ns) that can reversibly be stored in the layer. Ā e charge capacity is normalized t o e lectrode a rea a nd m easured i n m C/cm2. Ā e charge capacity of the ion storage electrode should be in excess of charge required for sufficient coloration of the working electrode. Typical v alues a re b etween 10 a nd 3 0 mC/cm2. It i s a lso essential to ac hieve h igh t ransparency of t he counterelectrode, which is often in conict with the requirement for high charge capacity. A nother re quirement for t he ion s torage l ayer i n t his structure is a low coloration efficiency. A good ion storage electrode should have a v ery large charge capacity and be optically passive. An example of a device of this type is the near commercial window produced by Sustainable Technologies Australia. Ā is

43722_C014.indd 27

FIGURE 14.34 Large s cale protot ype e lectrochromic w indow. (From Zini, M., Build. Environ. 41, 1262, 2006. With permission.)

device u ses a mo died WO3 ele ctrochromic ele ctrode a nd a V2O5-based counterelectrode, with a lithium-doped cross-linked polymer electrolyte. Spectra of the colored and bleached states of this device are shown in Figure 14.35. 14.2.5.1.2

Type 2—Combined Ion Conducting Layer and Counterelectrode

In t he s econd app roach to de vice de sign, t he me ans o f c harge delivery a nd charge storage a re realized i n one layer. Ā is is t he ion-conducting layer, which contains dispersed redox couples, thus in t he dei ntercalated s tate io ns a re s tored i n t he ele ctrolyte. A n example of this type of device is based on WO3 as t he working electrode, w ith a s olid p olymer ele ctrolyte c ontaining l ithium polyorganodisulde redox salts [68]. Ā ese devices have the advantage of simplicity of manufacture, requiring deposition of one less layer than Type 1 devices. In addition, the redox salts are generally optically passive, so charge capacity of the counterelectrode is not an issue. However, the redox reaction only occurs at the surface of the layer a nd t ransport of t he electrons across t he polymer layer from t he t ransperent electrical conductor (TEC) on t he counterelectrode side of the device can limit the device. 14.2.5.1.3

Type 3—Ion Transport Layer and Complementary Counterelectrode

Ā e t hird de vice s tructure i s v ery si milar to t he  rst t ype a s there a re s eparate ion t ransport a nd c ounterelectrode l ayers.

10/16/2008 7:16:12 PM

14-28

Smart Materials 80

70

70

60

Bleached Transmittance (%)

Transmittance (%)

Bleached Tvis /Tsol = 1.19 50 40 30

60

5 mC/cm2

50

10 mC/cm2

40

15 mC/cm2

30 20

20

10

10

Colored Tvis /Tsol = 1.42

0

0

500 500

1000 l (nm)

1500

2000

FIGURE 14.35 Transmittance spectra for colored and bleached states of a n S TA e lectrochromic d evice. Ā e s electivity i n b oth s tates a re shown on the graph, indicating the increase in selectivity in the colored state. Ā e dynamic range for t he device is Tsol = 51.5% to Tsol = 2 0.6%, with injected charge of 10 mC/cm2.

However, t he co unterelectrode ma terial i s a el ectrochromic material, w hich i s c omplementary to t he w orking ele ctrode, i.e., it is an anodically coloring electrochromic material, which darkens upon extraction of ions from the counterelectrode and bleaches upon injection of ions [69]. In this way, both electrodes contribute to the coloration of the electrochromic device, enhancing the overall performance of the device. Ā is type of device has the advantage of requiring less overall charge transport to achieve the same level of coloration as a device of type 1. Hence the charge capacity requirements of the electrodes are lower than for type 1 devices. A device of this type has been developed by the Asahi Glass Company (Japan). Āis device consists of the following structure: Glass/TEC (200 nm)/NiOx (500 nm)/Ta 2O5 (5 00 nm)/WO3 (5 00 nm)/TEC (5 00 nm)/adhesive lm (250 µm)/glass. Ā e spectral response of this device is shown in Figure 14.36 for four different levels of charge injection. Note that the change in optical density per unit charge is signicantly larger for this device than for the STA device of type 1 (see Figure 14.35) owing to the coloration of both the working electrode (WO3) and the counterelectrode (NiOx). 14.2.5.2

Materials Used in Electrochromic Devices

14.2.5.2.1

Electrochromic Materials

Tungsten o xide i s c urrently t he mo st c ommonly u sed i norganic electrochromic material. Coloration during ion insertion (H+, Li+, Na+, a nd K+) a nd bleaching on ion extraction is termed cathodic coloration. In an electrochromic smart window, thin  lm electrodes must be in contact with the working and counterelectrodes and an electrolyte is required for ion transport. Use of electrochromic materials in window fabrication necessitates the transparency of all components and it is these constraints that limit the number of materials suitable for electrochromic device fabrication.

43722_C014.indd 28

1000 l (nm)

1500

2000

FIGURE 1 4.36 Transmittance s pectra of a n A sahi e lectrochromic device in the bleached state and for three colored states, with 5, 10, and 15 mC/cm2 injected into the WO3 layer at a constant potential of 1.5 V.

Numerous transition metal oxides are suitable to be used as the electrochromic layer in smart windows, and a de tailed discussion o f ele ctrochromism i n t hese m aterials i s g iven i n Re f. [70]. Ā e most widely studied electrochromic material is WO3 and dop ed WO3, w here t he dopa nt i s a nother t ransition me tal oxide such as TiO2 or MoO3. Tungsten oxide lms are commonly prepared b y t hermal e vaporation, s puttering, c hemical v apor deposition ( CVD), s pray de position, a nodization, a nd s ol–gel deposition. A detailed discussion of these deposition techniques and their effects on the properties of the deposited materials is given in Refs. [70,71] and references therein. On the basis of the amount of charge that can be inserted, the coloration efficiency and t he t ransmittance c hange o f t he  lm, t he ele ctrochromic performance of WO3 deposited by most techniques is very similar, with the most important effect being increased crystallinity in t he  lms s puttered o nto h eated sub strates ( above app roximately 350°C), or for sol–gel deposited lms heat treated at temperatures o f 3 50–400°C a fter deposition. Ā i s increase in crystallinity is usually accompanied by an increase in reectance modulation in the lms. Other inorganic electrochromic systems, which have received signicant a ttention, a re N iO, T iO2, I rO2, N b2O5, a nd mo re recently, SnO2 and Pr2O3. A variety of techniques have been used for the deposition of all of these coatings, but none of these materials show the electrochromic efficiency of WO3. 14.2.5.2.2

Counterelectrode Materials

NiO and V 2O5 have been extensively studied as potential counterelectrode m aterials f or ele ctrochromic de vices [ 70], w hile CeO2-TiO2, and more recently CeO2-ZrO2, have been extensively studied in the past few years [32]. NiO i s particularly attractive as it colors anodically (i.e., on the extraction of ions), and hence complements both the coloration and bleaching of the cathodically c oloring W O3 l ayer [ 69] ( see a bove). V 2O5 h as e xcellent charge capacity and can be produced relatively easily using

10/16/2008 7:16:12 PM

Fiber Optic Systems: Optical Fiber Sensor Technology and Windows

sol–gel deposition [33], but it has the disadvantage of an absorption e dge en croaching i nto t he v isible re gion o f t he s pectrum, resulting in a yellowish coloration in the  lms [72]. Initially (up to t he e arly 1 990s) s puttering a nd e vaporation do minated t he deposition of electrochromics, but sol–gel deposition is becoming very widespread except for NiO, although there has been one report of sol–gel-deposited NiO [73]. 14.2.5.2.3

Ion Transport Layer

Ā e f unction o f t he ele ctrolyte l ayer i s to a llow io ns to t ravel between the working and counterelectrodes, hence it is ionically conducting. Low electrical conductivity of the electrolyte layer is advantageous because it reduces the internal leakage current in a de vice a nd a llows de vices to s tay c olored for long p eriods of time without the need for an external power source. Electrolytes u sed i n ele ctrochromic sm art w indows c an b e classied as either polymeric or superionic conductors. Polymeric electrolytes are prepared as a liquid with some dissolved lithium salt, and are then used to laminate the working and counterelectrode l ayers toge ther. Ā e p olymer t herefore a llows io nic c onduction w hile g iving t he de vice i ts me chanical s trength. Superionic conductors such as M-β-alumina (where M is a monovalent metal such as Li, Na, or K) a llow ions to move through the host lattice of alumina. Ā ese electrolytes are advantageous because they are not prone to damage by factors such as UV irradiation a nd c an b e de posited d irectly o nto t he ele ctrochromic layer, however they may have a nonzero electrical conductivity, which leads to a reduced memory in the colored state. Ā e Asahi device, which uses Ta 2O5 as an ion conducting layer, shows poor memory c ompared to d evices t hat u se p olymer-based e lectrolytes (e.g., the STA device discussed above) [29]. 14.2.5.2.4

Transparent Electronic Conductors

Ā e tr ansparent e lectrical c onductors in m ost e lectrochromic devices are thin  lms of metal oxides such as indium-doped tin oxide (SnO2:In or ITO) and uorine-doped tin oxide (SnO2:F or FTO). Ā ese m aterials a re c ommercially a vailable p recoated onto glass substrates by spray pyrolysis methods [74]. Some electrochromic devices utilize conducting polymeric layers such as ITO-PET (I TO c oated o nto p olyethylene t erephthalate) a s t he transparent electrical conductors, which a llows for the fabrication of polymeric solid-state electrochromic devices [75]. 14.2.5.3

Control of Electrochromic Smart Windows

Ā e app lication o f ele ctrochromic sm art w indows i n a rchitectural glazings requires that the devices can perform tens of thousands o f s witching c ycles o ver a p eriod o f 2 0–30 ye ars, w ith minimal degradation i n opt ical performance. Ā e l ifetime of a smart window is measured in number of coloration/bleaching cycles w hen t he t ransmittance i n e ach c ycle i s mo dulated between speci ed m aximum a nd m inimum v alues. H owever, the t ransmittance o f t he w indow, pa rticularly o f a n i nstalled window, i s not e asily ac cessible. A s i llustrated i n Figure 14.24, the optical density of an electrochromic window is in a one-toone (and v ery ne arly l inear) rel ationship w ith i njected c harge

43722_C014.indd 29

14-29

density. Ā is relationship is also almost independent of temperature, s o a go od s witching a lgorithm f or ele ctrochromic sm art windows can be based on charge control. When considered as an electrochemical system, an electrochromic device is very similar to a rechargeable battery. Ā e energy in the battery is stored and released by me ans of re versible redox reactions. In electrochromic devices, these reactions result in optical changes. Ā e durability of batteries has been investigated extensively over the last 50 years and several empirical relationships between the lifetime and operational parameters have been found [76]. In particular, battery lifetime is found to be inversely proportional to the operational current, so the switching current must be small enough to provide a long lifetime. In the design of a c ontrol algorithm for electrochromic windows, two fundamental limits that restrict fast switching of the window must be considered: • Ā ermodynamic l imit (overvoltage): t he voltage must b e maintained below the potential at w hich destructive side reaction(s) occur. • Kinetic l imit (excessive charging c urrent): large c urrents cause i rreversible c ation t rapping o n t he su rface o f t he  lm [60] owing to charge “pile-up” at the surface. Ā is will usually be caused by slow cation diff usion into the electrochromic layer. Of course, if the voltage is too small, the desirable electrochromic reaction can never occur, and if the charging current is too small, t he s witching t ime of t he w indow i s u nacceptably long. Ā erefore, b oth c urrent a nd vol tage ne ed to b e m aintained within  xed l imits to ac hieve opt imum de vice l ifetime, w ith charge used as an independent parameter to control the optical state of the device. Both t hermodynamic ( redox p otentials) a nd k inetic ( resistance, diff usion coefficients) p roperties o f el ectrochromic window are temperature-dependent, thus they generally require adjustments in switching parameters for changing environmental conditions. For example, a c onstant applied voltage will not provide constant contrast ratio at different temperatures owing to the temperature dependence of the lm potential on temperature. At h igh temperatures, a vol tage su itable for low temperature could exceed the potential of side reactions (thermodynamic limit) and will lead to higher currents and possible kinetic device failure. Typical c ycles of electrochromic w indow performed at 50°C and 18°C are presented in Figure 14.37. Ā e charge intercalated into working electrode is the same in both cycles, therefore the changes in optical transmittance are almost identical. Āe charging c urrent den sity i s a lso iden tical at t he t wo tem peratures, however, at 50°C the window required signicantly lower coloring voltage si nce d iffusion coefficients a nd ele ctrolyte c onductance increase with temperature. Ā e control algorithm implemented to p roduce t hese re sults i ncludes c onstant c urrent c harging until t he i njected charge exceeds a g iven value Q in. If voltage exceeds a p redetermined s afe v alue Vmax, t he s ystem automatically switches to a c onstant voltage mode. In the bleaching part

10/16/2008 7:16:13 PM

14-30

Smart Materials

Device voltage

1.5

1

1 0.8 0.5 0.6 0 0.4

−0.5 −1

0

50

100

150 Time

200

250

Transmittance change (a.u.)

1.2

2

0.2 300

FIGURE 14.37 Electrochromic d evice p erformance at 5 0°C (×) a nd 18°C (+). Ā e s olid l ines re present t he d evice volt age a nd t he d ashed lines the relative transmittance of the device. In both cycles the operational parameters are: current density = 0.1 mA/cm2, Qin = 13 mC/cm2; Qout = − 13 mC/cm2, pre set s witching volt ages a re: Vmax = 1 .8 V, Vmin = −0.7 V. Note that at 50°C voltage does not reach Vmax. Ā e high-temperature cycle period is shorter and lower voltages are required, however, transmittance change in the two cycles is almost identical.

of a c ycle e xtracted c harge s hall not e xceed Qout, a nd vol tage shall b e lo wer i n m agnitude t han a p redetermined m aximum bleaching voltage Vmin. Methods for t he control of electrochromic de vices u sing t hese te chniques a re de scribed i n paten ts [77,78] a nd t he t heoretical bac kground i s o utlined i n Re fs. [57,60].

14.2.6 Future Directions Smart window prototypes of sufficient scale for full-scale testing are no w a vailable, a s e videnced b y t he e xamples i llustrated i n Figures 14.30 a nd 14.34 above. Ā ere is a g rowing body of evidence that these windows will provide signicant energy savings for commercial buildings in appropriate climates (not ably hot, arid c limates) p rovided app ropriate w indow c ontrol s trategies are employed. Smart windows based on sputtered tungsten oxide electrochromics are now [79] commercially available, a lthough previous commercial products released by Pilkington (see Pilkington C ompany H istory [ 80]) app ear to h ave b een w ithdrawn f rom t he m arket. Ā ere a re s till sig nicant questions about t he d urability o f suc h w indows, a nd t he m ajor re search efforts in the immediate future will be to improve window durability during cycling. Ā e advantage of the electrochromic windows in terms of independent control of the optical state (using injected charge density) is actually a disadvantage in this context as it allows overdriving of the device, while the gasochromic and thermally s witched de vices a re i ntrinsically s elf-limiting i n terms o f s witching t he w indows. H owever, ot her d urability issues exist with these devices.

43722_C014.indd 30

In o rder to i mprove t he p erformance o f sm art w indows i n architectural g lazing app lications, t wo i ssues ne ed to b e addressed: t he de velopment o f t ruly re ective sm art w indow devices (like crystalline electrochromic materials, thermotropic glazings, and some of the recently discovered gasochromic materials), which are fully specular and (relatively) color neutral; and controlling t he t hermal em ittance o f g lazings. Ā e l atter w ill broaden t he range of climates i n which smart w indows c an be usefully used, and will also open up a new range of nonarchitectural applications. Ā e f uture of smart w indows in architectural applications is very bright; the growing trend toward sustainable buildings and energy c onservation, c oupled w ith s ome re cent re views o f t he total l ife c ycle p erformance [ 81] a nd u ser ac ceptance [ 82] o f smart w indows, which show t hat t hese systems off er a gen uine technology for re ducing g reenhouse em issions a nd a re ac ceptable to users, suggest that widespread uptake of these technologies is likely within the next 10 years.

References 1. C.G. Granqvist, E. Avendaño, and A. Azens, Electrochromic coatings and devices: Survey of some recent advances, Thin Solid Films 442, 201–211, 2003. 2. E. Avendaño, L. B erggren, G.A. Niklasson, C.G. Granqvist, and A. Azens, Electrochromic materials and de vices: Brief survey and new data on optical absorption in tungsten oxide and nickel oxide lms, Thin Solid Films 496, 30–36, 2006. 3. P.R. Somani and S. Radhakrishnan, Electrochromic materials and devices: Present and future, Materials Chemistry and Physics 77, 117–133, 2002. 4. S.K. Deb, A Novel Electrophotographic system. Appl. Oppt. Supp. 3, 192–195, 1969. 5. B.-S. Yu, E.-S.K im, a nd Y.-W. L ee, D evelopments in suspended particle devices (SPD), SPIE 3138, 217–225, 1997. 6. G.P. M ontgomery, P olymer-dispersed a nd enca psulated liquid crystal  lms, in C.M. Lampert and C.G. Granqvist, eds., Large Area Ch romogenics: M aterials a nd D evices for Transmittance Control, SPIE, Washington, D C, 1998, p. 577. 7. K.A. Khan and C.G. Granqvist, Ā ermochromism of sputter deposited vanadium oxyuoride coatings, in C.M. Lampert and C.G. G ranqvist, e ds., Large A rea Chromogenics: Materials a nd D evices f or T ransmittance Co ntrol, S PIE, Washington, DC, 1998, p. 160. 8. A. G eorg, W. Graf, D. S chweiger, V. Wittwer, P. Nitz, and H. Wilson, Switchable glazing with a la rge dynamic range in to tal s olar ener gy t ransmittance (T SET), Solar E nergy 62(3), 215–228, 1998. 9. A. B essière, C. Duha mel, J .-C. B adot, V. L ucas, a nd M.-C. Certiat, Study and optimization of a exible electrochromic device ba sed o n po lyaniline, Electrochim. A cta 49, 2051– 2055, 2004. 10. L.-M. Huang, C.-H. Chen, and T.-C. Wen, Development and characterization of  exible ele ctrochromic de vices b ased

10/16/2008 7:16:13 PM

Fiber Optic Systems: Optical Fiber Sensor Technology and Windows

11.

12.

13.

14.

15. 16.

17. 18. 19. 20. 21.

22.

23.

43722_C014.indd 31

on polyaniline a nd p oly(3,4-ethylenedioxythiophene)poly(styrene sulfonic acid), Electrochim. Acta 51, 5858–5863, 2006. S.-W. Huang a nd K.-C. Ho, An all-thio phene ele ctrochromic de vice fa bricated wi th p oly(3-methylthiophene) a nd poly(3,4-ethylenedioxythiophene), Sol. E nergy M ater. S ol. Cells 90, 491–505, 2006. P. M anisankar, C. Vedhi, G. S elvanathan, a nd H.G. P rabu, Electrochemical syn thesis a nd c haracterization o f no vel electrochromic p oly(3,4-ethylenedioxythiophene-codiclofenac) wi th surfac tants, Electrochim. A cta 51, 2964– 2970, 2006. G.-S. Liou, N.-K. Huang, and Y.-L. Yang, New soluble triphenylamine-based amorphous aromatic p olymides for high performance b lue-emitting ho le-transporting a nd a nodically ele ctrochromic ma terials, Polymer 47, 7013–7020, 2006. G.-S. Lio u, S.-H. H siao, a nd T.-H. S u, S ynthesis, l uminescence and electrochromism of aromatic poly(amine-amide) s with pendent triphenylamine moieties, J. Mater. Chem. 15, 1812–1820, 2005. A.B.S. Bahaj et al ., Potential of emerging glazing technologies for highly glazed buildings in hot arid climates, Energy Buildings (2007), doi:10.1016/j.enbuild.2007.05.006 J.M. B ell, G.B. Smith, A.T. Baker, L.A. Evans, I.L. Skryabin, J. B arczynska, S.G. B osi, J .J. B right, D .R. M acfarlane, B .O. West, L. S piccia, G. Voelkel, A. K oplik, B .Greaves, G.E. Tulloch, S.M. T ulloch, S.M. J enkins, I.D . Watkins, M. Williams, D. F rost, G . Vogelman, and G . E vans, I ntelligent Glazing Materials for Energy Efficient Heating, Lighting and Cooling of Buildings, end of project report, Energy Research and Development Corporation, Canberra, May 1996. G.L. Yoong and G.E. Tulloch, Smart windows-major energy savings for the built environment in the tropics. Proceedings of the Eurosun, Copenhagen, Denmark, June 19–22, 2000. F. Gugliermetti and F. Bisegna, Visual and energy management of electrochromic windows in Mediterranean climate, Building Environment 38, 479–492, 2003. F. Yika a nd M. B ojic, Application o f swi tchable g lazing t o high-rise r esidential b uildings in H ong K ong, Energy an d Buildings 38, 463–471, 2006. T. Inoue, Solar shading and daylighting by means of autonomous r esponsive dimmin g g lass: Prac tical a pplication, Energy Buildings 35, 463–471, 2003. W.J. Platzer, Determination of key glazing parameters: Part 1—Total s olar ener gy t ransmittance. P roceedings o f Windows I nnovations C onference, WIC ‘95, p p. 22–31, Toronto, Canada, 1995. C.G. Granqvist. Introduction to materials science for solar energy conversion systems, in A.A.M. Sayigh, ed., Materials Science f or S olar E nergy Co nversion S ystems, P ergamon Press, New York, 1991, p. 1. ASHRAE H andbook o f F undamentals, America n Society of Heating, Refrigeration and Air Conditioning Engineers, Inc., Atlanta, 1981, Chapter 27.

14-31

24. S.E. Selkowitz and C.M. Lampert. Applications of large area chromogenics to architectural glazing, in C.M. Lampert and C.G. G ranqvist, e ds., Large Area Ch romogenics: M aterials and D evices f or Transmittance Co ntrol, S PIE, Washington, DC, 1998, p. 22. 25. A.W. Czanderna, J.-G. Zhang, C.E. Tracy, D.K. Benson, and S.K. Deb, Accelerated life testing of large area electrochromic devices for window application. Proc. SPIE, 3138, 68–77, 1997. 26. R. Sullivan, S. Selkowitz, P. Lyons, P.C. Ā omas, I. Heimonen, O. Andresen, H. Aschehoug, P. Simm ler, E ggimann, a nd T. F rank, Ener gy sim ulation st udies in IEA/S HC Task 18 advanced g lazings a nd ass ociated ma terials f or s olar a nd building applications. Proceedings of Windows Innovations Conference, WIC ‘95, pp. 635–662, Toronto, Canada, 1995. 27. P.K. N air, M.T .S. N air, A. F ernandez, a nd M. O campo, Prospects of chemically deposited chalcogenide thin lms for solar control applications. J. Phys. D. 22(6), 829–836, 1989. 28. J. F erber a nd W. Pla tzer, C orrect o ptical me asurement o f scattering samples. Proc. SPIE 2255, 708–717, 1994. 29. M.D. Rubin, IEA Task 18, Project B3: Chromogenic glazing, Final report, 1997. 30. Sage Electrochromics, http://www.sage-ec.com/pages/technol.html (accessed 15 October 2007). 31. J.-G. Zha ng, D .K. B enson, C.E. T racy, a nd S.K. D eb, Ā e inuence of microstructure on the electrochromic properties o f L ixWO3 lms: P art I I. L imiting me chanisms i n colouring a nd b leaching p rocesses. J. Ma ter. R es. 8(10), 2657–2667, 1993. 32. C.G. G ranqvist, A. Azens, A. H jelm, L. K ullman, G.A. Niklasson, D . Ro nnow, M.S. M attson, M. Veszelei, a nd G. Vaivars, Re cent adva nces in ele ctrochromics f or sma rt windows applications. Solar Energy 64(4), 199–216, 1998. 33. G.E. Tulloch, I.L. Skryabin, G. Evans, and J.M. Bell, Operation of ele ctrochromic de vices p repared b y s ol–gel metho ds. Proc. SPIE 3136, 426–432, 1997. 34. B.W. F aughnan a nd R .S. Cra ndall, Ele ctrochromic disp lay based on WO3 in J.I. Pankove, ed., Topics in Applied Physics V40, Dis play D evices, S pring-Verlag, N ew York, 1980, Chapter 5. 35. T. Kase, T. Miyamoto, T. Yoshimoto, Y. Ohsawa, H. Inaba, and K. N akase, P erformance o f t ungsten o xide/prussian b lue device, in C.M. L ampert a nd C.G. G ranqvist, e ds., Large Area Chromogenics: Materials and Devices for Transmittance Control, SPIE, Washington, DC, 1998, p. 504. 36. J.M. B ell, J . B arczynska, L.A. E vans, K.A. M acDonald, J.Wang, D.C. Green, and G.B. Smith, Electrochromism in s olgel deposited TiO2 lms, Proc. SPIE. 2255, 324–331, 1994. 37. A. Agrawal, J.P. Cronin, and R. Zhang, Review of solid state electrochromic coatings produced using sol–gel techniques. Solar Energy Mater. Solar Cells 31(1), 9–22, 1993. 38. C. Wang and J.M. Bell, Inuence of deposition temperature on electrochromic properties of sputtered WO3 thin  lms. Solar E nergy M aterials a nd S olar Ce lls 43(4), 377–391, 1996.

10/16/2008 7:16:13 PM

14-32

39. J.S.E.M. Svensson a nd C.G. G ranqvist, Modulated tra nsmittance and reectance in crystalline electrochromic WO3 lms: Ā e oretical limits. Appl. Phys. Lett. 45(8), 828–830, 1984. 40. S.C. Yang, Conducting polymer as electrochromic material: Polyaniline, in C.M. Lampert and C.G. Granqvist, eds., Large Area Ch romogenics: M aterials a nd D evices f or T ransmittance Control, SPIE, Washington, DC, 1998, p. 335. 41. R. Jiang and S. Dong, Chromatic reaction of polyaniline lm and its characterization, J. Chem. Soc., Faraday Trans. 85(7), 1585, 1989. 42. C.M. L ampert a nd C.G. G ranqvist. I ntroduction t o c hromogenics, in C.M. Lampert and C.G. Granqvist, eds., Large Area Ch romogenics: M aterials a nd D evices f or T ransmittance Control, SPIE, Washington, DC, 1998, p. 2. 43. M.A. S obhan, R .T. K ivaisi, B . S tjerna, a nd C.-G. G ranqvist, Ā ermochromism of sputter-deposited WxV1−xO2 lms. Solar Energy Materials and Solar Cells 44(4), 451–455, 1996. 44. G.V. Jorgenson and J.C. Lee. Ā ermochromic materials and devices: I norganic syst ems, in C.M. L ampert a nd C.G. Granqvist, e ds., Large Area Ch romogenics: M aterials a nd Devices f or Transmittance Co ntrol, S PIE, Washington, D C, 1998, p. 142. 45. C.M. Lampert, Science a nd Technology o f Elec trochromics: Lecture Notes, SPIE Short Course, 1993. 46. A. B eck, W. Körner, H . S cheller, J. Fricke, W.J. Platzer, and V. Wittwer, Control of solar insolation via thermo chromic light switching gels. Solar Energy Materials and Solar Cells 36, 339–347, 1995. 47. H.R. Wilson, P otential o f thermo tropic la yers t o p revent overheating—a review, Proc. SPIE 2255, 214–225, 1994. 48. C.M. L ampert. I ntroduction t o liq uid cr ystals, in C.M. Lampert and C.G. Granqvist, eds., Large Area Chromogenics: Materials a nd D evices f or T ransmittance Co ntrol, S PIE, Washington, DC, 1998, p. 550. 49. N. Basturk and J. Grupp. Liquid cr ystal guest host de vices and their us e as lig ht switches, in C.M. L ampert and C.G. Granqvist, e ds., Large Area Ch romogenics: M aterials a nd Devices f or Transmittance Co ntrol, S PIE, Washington, D C, 1998, p. 557. 50. S. K hoobiar, P article to p article mig ration o f h ydrogen atoms o n p latinum–alumina ca talysts f rom pa rticle t o neighbouring particles. J. Phys. Chem. 68(2), 411–412, 1964. 51. D. Schweiger, A. Georg, W. Graf, and V. Wittwer, Examination of the kinetics a nd performance of a ca talytically switching (gasochromic) device. Solar Energy Materials and Solar Cells 54, 99–108, 1998. 52. A. Georg, W. Graf, R. Neumann, and V. Wittwer, Stability of gasochromic WO3 lms. To be published in Proceedings of Eurosun 98, Portoroz, Slovenia, 1998. 53. A. Georg, W. Graf, R. Neumann, and V. Wittwer, Mechanism of t he gas ochromic colouration in WO3 lms. To b e published in Proceedings of the Ā ird International Meeting on Electrochromics (IME-3), London, England, 1998. 54. P. va n der S luis, N ew Ele ctrochromic ma terials bas ed o n metal hydrides. To be published in Proceedings of the Ā ir d

43722_C014.indd 32

Smart Materials

55.

56. 57.

58. 59.

60.

61. 62. 63. 64.

65.

66. 67.

68.

International Meeting on Electrochromics (IME-3), London, England, 1998. K. v on Ro ttkay, M. R ubin, F . Mic halak, R . Armitage, T. Richardson, and J. Slack, Effect of hydrogen insertion on the optical properties of Pd-coated magnesium lanthanides. To be published in P roceedings of the Ā ir d International Meeting o n Ele ctrochromics (IME-3), L ondon, En gland, 1998. P. va n der S luis, M. Ou werkerk, a nd P.A. Duine , Op tical switches b ased o n mag nesium la nthanide a lloy h ydrides. Appl. Phys. Lett. 70(25), 3356–3358, 1998. J.M. Bell, I.L. Skryabin, and G. Vogelmann, Towards a model for l arge are a el ectrochromic d evice, Pro ceedings of t he Ā ird S ymposium o n Ele ctrochromic M aterials, Ā e Electrochemical Society, 96-24, pp. 396–403, 1997. C.M. L ampert a nd C.G. G ranqvist, e ds., Large A rea Chromogenics: M aterials a nd D evices f or T ransmittance Control, SPIE, Washington, DC, 1998, Parts 5, 6, 7 and 9. J. G öttsche, A. H insch, a nd V. Wittwer, Ele ctrochromic mixed WO3-TiO2 thin lms produced by sputtering and the sol-gel te chnique: A co mparison. Solar Ene rgy M aterials and Solar Cells 31(3), 415–428, 1993. J. Wang, J.M. B ell, and I.L. Skryabin, Simulation model for electrochromic process at WO3 electrode by a current step and its application to the smart window system. Proc. SPIE 3138, 20–30, 1997. S.K. Mohapatra, Electrochromism in LixWO3. J. Electrochem. Soc. 125(2), 284–288, 1978. B. Reichman and A.J. Bard, A digital simulation model for electrochromic processes at WO3 electrodes. J. Electrochem. Soc. 127(3), 647–654, 1980. C. Ho, I.D. Raistrick, and R.A. Huggins, Application of A-C techniques to the study of lithium diffusion in tungsten trioxide thin lms. J. Electrochem. Soc. 127(2), 343–350, 1980. B. Vuillemin and O. Bohnke, Kinetics study and modelling of the electrochromic phenomenon in amorphous tungsten trioxide t hin  lms in acid a nd li thium ele ctrolytes. Solid State Ionics 68, 257–267, 1994. O. Bohnke, M. Rezrazi, B. Vuillemin, C. Bohnke, P.A. Gillet, and C. Rousselot, “In situ” optical and electrochemical characterization of ele ctrochromic phenomena in to tungsten trioxide thin  lms. Solar Energy Materials and Solar Cells 25, 361–374, 1992. M. S trømme, J . I sidorsson, G.A. N iklasson, a nd C.G. Granqvist, Impedance studies on Li insertion electrodes of Sn oxide and oxyuoride. J. Appl. Phys. 80(1), 233–241, 1996. C. Wang, J.M. B ell, and I.L. Skryabin, Degradation mechanisms in s ol–gel dep osited ele ctrochromic de vices. Proceedings o f the 2nd I nternational M eeting o n Electrochromism— Durability Forum, San Diego, CA, 1996. (To appear in Solar Energy Materials and Solar Cells). C.M. L ampert, S.J. Visco, M.M. D oeff, Y.P. M a, Y. H e, a nd J.-C. G iron, Cha racteristics f or la minated ele ctrochromic devices usin g p olyorganodisulphide ele ctrodes. Proc. S PIE 2017, 143–154, 1993.

10/16/2008 7:16:13 PM

Fiber Optic Systems: Optical Fiber Sensor Technology and Windows

69. X. H u, X. Chen, a nd Z. Li , All-solid st ate ele ctrochromic device with NiO/WO3 complementary structure. Proc. SPIE 3138, 58–62, 1997. 70. C.G. Granqvist, Handbook o f I norganic Elec trochromic Materials, Elsevier, Amsterdam, 1995. 71. J.M. B ell a nd J .P. M atthews, Glazin g M aterials. Materials Forum 22, 1–24, 1998. 72. S.F. C ogan, Ele ctrochromic Vanadium P entoxide, in C.M. Lampert and C.G. Granqvist, eds., Large Area Chromogenics: Materials a nd D evices f or T ransmittance Co ntrol, S PIE, Washington, DC, 1998, p. 313. 73. M. T akeski, K. Yoshimura, a nd S. T anemura, P reparation and cha racterization o f ele ctrochromic nick el o xide t hin lm by sol–gel method. Proceedings of Windows Innovations Conference, WIC ‘95, pp. 365–371, Toronto, Canada, 1995. 74. P.F. G erhardinger a nd R .J. M cCurdy, Floa t-line dep osited transparent conductors—implications for the PV ind ustry. Mater. Res. Soc. Symp. Proc. 426, 399–410, 1996. 75. M.A. De Paoli, G. Casalbore-Miceli, E.M. Girotto, and W. A. Gazotti Jr., All polymeric solid-state electrochromic devices.

43722_C014.indd 33

76. 77. 78. 79. 80. 81. 82.

14-33

To be published in P roceedings of the Ā ir d International Meeting o n Ele ctrochromics (IME-3), L ondon, En gland, 1998. V.S. B agotzky and A.M. Skundin, Chemical Power S ources, Academic Press, London, 1980. J.M. Bell and I.L. Skryabin, Method of control of an electrochromic device, Australian Patent PO3010, 1996. R.A. Batchelor, Electrochromic Devices, PCT/GB97/00199, 2/02/96. Sage Electrochromics Inc, http://www.sage-ec.com (accessed 15 October 2007). Pilkington Company History, 1950–1999, http://www.pilkington.com/about+pilkington/company+brieng/company+ history/1950+-+1999.htm (accessed 15 October 2007). E. S yrrakou, S. P apaefthimiou, a nd P . Yianoulis, E coefficiency evaluation of a smart window prototype, Science of the Total Environment 359, 267–282, 2006. M. Zinzi, Office worker preferences of electrochromic windows: a p ilot st udy, Building a nd E nvironment 41, 1262– 1273, 2006.

10/16/2008 7:16:13 PM

43722_C014.indd 34

10/16/2008 7:16:14 PM

15 Flip-Chip Underfill: Materials, Process, and Reliability 15.1 I ntroduction ...........................................................................................................................15-1 15.2 C onventional Underll Materials and Process ..................................................................15-2 15.3 Reliability of Flip-Chip Underll Packages ........................................................................15-4 15.4 New Challenges to Underll ................................................................................................15-5 15.5 N o-Flow Underll .................................................................................................................15-7

Zhuqing Zhang Hewlett-Packard Company

C.P. Wong Georgia Institute of Technology

15.1

Approaches of Incorporating Silica Fillers into No-Flow Under ll

15.6 M olded Underll ................................................................................................................. 15-11 15.7 Wafer Level Underll ..........................................................................................................15-12 15.8 S ummary ...............................................................................................................................15-15 References .........................................................................................................................................15-15

Introduction

Ā e brain of modern electronics is the integrated circuit (IC) on the s emiconductor c hip. I n o rder f or t he b rain to c ontrol t he system, i nterconnects ne ed to b e e stablished b etween t he I C chip and other electronic parts, power and ground, and inputs and outputs. Ā e rst-level i nterconnect u sually c onnects t he chip to a package made of either plastics and or ceramics, which in turn is assembled onto a printed circuit board (PCB). Ā re e main in terconnect t echniques ar e u sed: w ire-bonding, t ape automated b onding (T AB), a nd  ip-chip. In a w ire-bonded package, the chip is adhered to a carrier substrate using die-attach adhesive with the active IC facing up. A gold or aluminum wire i s b onded b etween e ach pad o n t he c hip a nd t he c orresponding b onding su rface o n t he c arrier a s s hown i n F igure 15.1. Ā e c hip a nd t he w ire i nterconnections a re u sually p rotected by encapsulation. TAB, on the other hand, uses a prefabricated lead f rame carrier w ith copper leads adapted to t he IC pads. Ā e c opper i s u sually gold-plated to p rovide a  nish for bonding to the IC chip pads. Ā e chip is attached onto the carrier a nd ei ther t hermosonic/thermocompression b onding o r Au/Sn bonding is used to establish the interconnect. Both wirebonding a nd T AB i nterconnects a re l imited to p eripheral arrangement a nd t herefore lo w i nput/output (I /O) c ounts. Flip-chip, h owever, c an ut ilize t he en tire s emiconductor a rea for interconnects. In a ip-chip package, the active side of an IC chip i s f aced do wn to ward a nd mo unted o nto a sub strate [1]. Interconnects, in the form of solder bumps, stud bumps, or

adhesive bumps, are built on the active surface of the chip, and are joined to t he substrate pads, in either a mel ting operation, adhesive joining, thermosonic, or thermocompression process. Figure 1 5.2 s hows a n e xample o f s older-bumped c hip su rface for  ip-chip i nterconnect. Si nce  ip-chip w as  rst developed about 40 years ago, many variations of the ip-chip design have been developed, among which, the controlled collapse chip connection ( also k nown a s C 4) i nvented b y I BM i n 1 960s i s t he most important form of  ip-chip [2]. Compared with conventional pac kaging u sing w ire-bonding te chnology,  ip-chip offers m any adv antages suc h a s h igh I /O den sity, s hort i nterconnects, s elf-alignment, b etter h eat d issipation t hrough t he back o f t he d ie, sm aller f ootprint, lo wer p ro le, a nd h igh throughput, etc. Ā e outstanding merits of ip-chip have made it one of the most attracting techniques in modern electronic packaging, including MCM modules, high-frequency communications, h igh-performance co mputers, p ortable el ectronics, and ber optical assemblies. Until t he late 1980s,  ip-chips were mounted onto si licon or ceramic sub strates. L ow-cost o rganic sub strates c ould not b e used due to the concern of the thermal–mechanical fatigue life of the C4 solder joints. Ā is thermal–mechanical issue mainly arises from t he c oefficient o f t hermal e xpansion ( CTE) m ismatch between t he s emiconductor c hip (typically Si, 2 .5 ppm/°C) a nd the substrate (4–10 ppm/°C for ceramics and 18–24 ppm/°C for organic FR4 board). As the distance from the neutral point (DNP) increases, t he s hear s tress at t he s older jo ints i ncreases a ccordingly. So with the increase in the chip size, the thermal–mechanical 15-1

43722_C015.indd 1

10/3/2008 8:50:13 PM

15-2

Smart Materials

FIGURE 15.2 FIGURE 15.1

Area array solder bumps for ip-chip interconnect.

First level interconnect using wire-bonding.

reliability becomes a critical issue. Organic substrates have advantages over ceramic substrates because of t heir low cost a nd low dielectric c onstant. But t he h igh C TE d ifferences bet ween t he organic substrates and the silicon chip exert great thermal stress on the solder joint during temperature cycling. In 1987, Hitachi rst demonstrated the improvement of solder fatigue life with the use of  lled resin to m atch solder CTE [3]. Ā is  lled r esin, l ater c alled “ under ll” w as o ne o f t he mo st innovative d evelopments to e nable t he u se of lo w-cost or ganic substrate in ip-chip packages. Under ll is a liquid encapsulate, usually epoxy resins heavily  lled with fused silica (SiO2) particles, w hich i s applied b etween t he c hip a nd t he sub strate a fter ip-chip interconnection. Upon curing, the hardened underll exhibits h igh mo dulus, lo w C TE m atching t hat o f t he s older joint, low mo isture a bsorption, a nd go od ad hesion toward t he chip and the substrate. Ā ermal stresses on the solder joints are redistributed a mong t he c hip, u nder ll, sub strate, a nd a ll t he solder joints, i nstead of concentrating on t he peripheral solder joints. It has been demonstrated that the application of underll can reduce t he a ll-important solder strain level to 0. 10–0.25 of the strain in joints, which are not encapsulated [4,5]. Ā er efore,

under ll c an i ncrease t he s older jo int f atigue l ife b y 10 to 1 00 times. I n addition, it provides an environmental protection to the I C c hip a nd s older jo ints. U nder ll b ecomes t he p ractical solution to e xtending t he app lication o f  ip-chip technology from ceramics to o rganic substrates, a nd f rom t he high-end to cost-sensitive p roducts. T oday,  ip-chip i s b eing e xtensively studied a nd u sed b y a lmost a ll m ajor ele ctronic c ompanies around w orld i ncluding I ntel, A MD, H itachi, I BM, D elphi, Motorola, Casio, etc.

15.2

Conventional Underfill Materials and Process

Ā e generic schematic of a  ip chip package is shown in Figure 15.3. C onventional u nderll is applied after t he  ip-chip interconnects a re formed. Ā e re sin  ows i nto t he gap b etween t he chip a nd t he sub strate b y a c apillary f orce. Ā erefore, it is also called “capillary underfill.” A typical capillary underfill is a mixture of liquid organic resin binder and inorganic  llers. Ā e organic binders are often epoxy resin mix, although cyanate ester or ot her re sin h as b een u sed f or u nder ll appl ication a s w ell. Figure 15.4 shows the chemical structure of some commonly used

IC chip Under bumping metallurgy

Solder bump Passivation

Underfill Solder mask

Contact pad Substrate

FIGURE 15.3

43722_C015.indd 2

Generic conguration of C4 with under ll.

10/3/2008 8:50:13 PM

15-3

Flip-Chip Underfi ll: Materials, Process, and Reliability CH3

O CH

CH2

CH2 ( O

OH

CH3 CH2 )O

O CH2 CH

C

O

C

n

O CH2 CH

CH2

CH3

CH3 Bis A epoxy resin O CH

CH2

CH2 ( O

H

OH

C

O CH2 CH

H CH2 )O

O

C

n

O CH2 CH

CH2

H

H Bis F epoxy resin

O O

O CH CH2 O

CH2

CH

CH2

CH2

CH

CH2

CH

CH2

N,N-diglycidyl-4-glycidyloxylaniline

FIGURE 15.4

CH

CH2

O

O O CH2

N

CH2

CH2

O

Naphthalene type epoxy resin

Typical epoxy resin structures used in under ll.

epoxy resins. In addition to epoxy resin, a hardener is often used to form a cross-linking structure upon curing. Sometimes a latent catalyst is incorporated to achieve long pot life and fast curing. Inorganic  llers t ypically u sed i n un der ll f ormulation ar e micron-sized s ilica. Ā e si lica  llers a re i ncorporated i nto t he resin binder to enhance the material properties of cured underll such a s low C TE, h igh modulus, a nd low moisture upt ake, e tc. Other a gents t hat c an b e f ound i n a n u nderll formulation include adhesion promoters, toughening agents, a nd d ispersing agents, et c. Ā ese c hemicals a re i ncorporated to h elp t he re sin mixing and enhance the cured underll properties. Figure 15.5 s hows t he process s teps of  ip-chip w ith conventional underll. Separate  ux d ispensing a nd c leaning s teps a re required before and after the assembling of the chip, respectively.

After the chip is assembled onto the substrate, the underll is usually needle-dispensed and is dragged into the gap between the chip and the substrate by a capillary force. Ā en a heating step is needed to cure the underll resin to form a permanent composite. Ā e ow of the capillary underll has been extensively studied since it is considered to be one of the bottlenecks for the ip-chip process. Ā e capillary ow is usually slow and can be incomplete, resulting in voids in the packages and also nonhomogeneity in the resin/ller system. Āe lling problem becomes even more serious as the chip size increases. Ā e ow modeling of ip chip underll is often approximated as the viscous ow of the underll adhesive between two parallel plates. One can use the Hele-Shaw model to simulate t he u nderll ow wi th th e a bove a pproximation. Ā e time required to ll a chip of length L can be calculated as [6]

Chip Heated

Flux

Bond pad Solder ball

Board

Fluxing dispensing

Chip placement

Solder reflow

Heated Underfill Solvent spray

Flux cleansing

FIGURE 15.5

43722_C015.indd 3

Underfill dispensing

Underfill cure

Flip-chip process using conventional under ll.

10/3/2008 8:50:14 PM

15-4

Smart Materials

β ⎛ ⎛ x⎞β⎞ ⎛ b⎞ ⎛ x⎞ f (x ) = ⎜ ⎟ ⎜ ⎟ exp ⎜ − ⎜ ⎟ ⎟ ⎝ x ⎠ ⎝q⎠ ⎝ ⎝q⎠ ⎠

(15.1)

⎛ 1⎞ MTTF = qΓ ⎜ 1 + ⎟ (1 ⎝ b⎠

where h is the under ll viscosity s is the coefficient of the surface tension q is the contact angle h is the gap distance It i s e asily s een t hat a l arger c hip w ith a sm aller gap d istance would require longer time to ll. Ā e a bove app roximation do es not t ake t he e xistence o f s older bumps into account. It is shown that the approximation breaks down when the spacing between bumps is comparable to the gap height [7]. Ā erefore, this model cannot apply to high-density area array ipchip applications. Using transparent quartz dies assembled onto different sub strates, N guyen e t a l. ob served t he  ow o f c ommercial underlls a nd used a 3 D PLICE-CAD to mo del t he underll ow front [8]. A comparison between the peripheral and area array chips showed that the bumps enhanced the  atness of t he  ow f ront by providing periodic wetting sites. A racing effect along the edges was observed. Voids c an be formed at t he merg ing of  ow f ronts. Ā e merging of the fronts also produced streaks, which are zones of noor slow-moving uids, leading to higher potential for ller settling. Recent development in underll ow models has also considered the effect o f t he c ontact a ngle o n s older a nd b ump ge ometry. A study by Young and Yang used a modied Hele-Shaw model considering the ow resistance in both the thickness direction between the chip and substrate, and the plane direction between solder bumps [9]. It was found that the capillary force parameter would approach a constant value at very large pitch for the same gap height. As the bump pitch is reduced, the capillary force will increase to a maximum as a result of underll wetting on the solder given the contact angle on the solder is small, and then quickly drops to zero as the pitch approaches the bump diameter. Ā eir study also showed that a hexagonal bump a rrangement i s more e fficient to en hance t he capillary force at critical bump pitch.

15.3

Reliability of Flip-Chip Underfill Packages

Ā e reliability of a  ip-chip package can be evaluated in a number of different methods, including thermal cycling, thermal shock, pressure-cook test, etc. Ā e lifetime of a solder joint interconnect during temperature cycling can often be described by statistical models suc h a s W eibull d istribution. Ā e p robability den sity function (PDF) of the Weibull distribution is given by β ⎛ ⎛ x⎞β⎞ ⎛ b⎞ ⎛ x⎞ f (x ) = ⎜ ⎟ ⎜ ⎟ exp ⎜ − ⎜ ⎟ ⎟ ⎝ x ⎠ ⎝q⎠ ⎝ ⎝q⎠ ⎠

where x is the thermal cycling life as a random variable q is the characteristic life b is the shape parameter

43722_C015.indd 4

Ā e me an t ime to f ailure (MTTF), w hich i s t he e xpectation of the time to failure, for the Weibull distribution is

(15.2)

5.3)

where Γ i s t he ga mma f unction. I t i s gener ally b elieved t hat fatigue of t he s older joints i s a m ajor re ason for s tructure a nd electrical f ailures. Ā e s older f atigue l ife c an b e de scribed a s a function of inelastic shear strain in the Coffin-Manson equation [10,11]: 1 ⎛ ∆g ⎞ Nf = ⎜ 2 ⎝ 2ε f′ ⎟⎠

1/c

(1

5.4)

where Nf is the number of cycles to fatigue failure ∆g is the inelastic shear strain e f ′ is the fatigue ductility coefficient c is the fatigue ductility exponent Other strain-based fatigue equations have been proposed, among which Solomon’s model is often used [12]: ⎛ θ ⎞ Nf = ⎜ ⎟ ⎝ ∆g p ⎠

1/a

(15.5)

where ∆gp i s t he p ercentage i nelastic s hear s train, q and a are constants. It has been shown that the use of the underll can increase the lifetime of the solder joints by at least an order of magnitude during thermal cycling [13]. It was found that in an underlled ip-chip package, the fatigue life is highly dependent on the material properties of the underll. Ā e analytic model by Nysaether et al. [14] showed that while an underll without ller increased the lifetime by a factor of 5–10, a lled underll with a lower CTE gave a 20–24fold increase in lifetime. For both lled and nonlled samples, the lifetime is nearly constant regardless of DNP, indicating that the underll effectively couples the stress among all the solder joints. Many n umerical mo dels h ave b een de veloped to s tudy t he solder fatigue life of a  ip-chip package with or without underll. Ā e polymeric nature of the under ll material requires careful c haracterization f or c orrect m aterial p roperty i nput to t he numerical models. Ā e modulus of a p olymeric material is not only a function of temperature, but also a function of time, i.e., it is a viscoelastic material. Ā ermal mechanical analyzer (TMA) and dy namical me chanical a nalyzer (DMA) a re t ypically u sed to characterize the viscoelastic properties of the underll material. Dudek et al. characterized four commercial electronic polymers and used nite element (FE) analyses to s tudy the effect of die size and under ll material properties on the thermomechanical reliability o f t he  ip-chip o n boa rd ( FCOB) p ackage [15]. Ā ey found that although the use of under ll can effective reduce the shear s train, it c an a lso c ause bump c reep s train i n t he t ransverse board direction due to s tretching and compressing of the bump during thermal cycling. This load is due to the CTE mismatch b etween t he s older a nd u nderfill/solder-mask layer. U nder ll w ith C TE t hat m atched t he s older m aterial

10/3/2008 8:50:15 PM

15-5

Flip-Chip Underfi ll: Materials, Process, and Reliability

(22–26 ppm/°C) gives the best thermal cycling life based on the creep strain criterion. Ā e function of the underll in a ip-chip package is stress redistribution, not stress reduction. A rigid underll material mechanically c ouples t he de vice a nd t he sub strate, c hanging pa rtially t he shear stress experienced by the solder joints into bending stress on the whole structure. Shrinkage of the underll during cure and the CTE mismatch during cooling after cure can generate large stress on the Si chip, resulting in die crack in some cases. Palaniappan et al. performed in-situ stress measurements in the ip-chip assemblies using a test chip with piezoresistive stress-sensing devices [16]. Ā e study concluded that the underll cure process generates large compressive stress on the active die surface, indicating a c omplex convex bending state in the  ip chip. Ā e level of stress measured can le ad to Si f racture. Ā e re sidual d ie s tress w as f ound to b e strongly dependent on underll CTE, modulus, and Tg. A FE analysis by Mercado et a l. on t he d ie edge cracking i n  ip-chip PBGA packages also concluded that the energy release rate for horizontal Si facture increases with underll modulus and CTE [17]. In addition to temperature-related thermomechanical failure, moisture-induced failures such as delamination and corrosion are common f or a  ip-chip underll package. Highly accelerated stress test (HAST) is often used to determine the temperature and moisture s ensitivity of t he p ackage. Ā e test uses hash environment c onditions s uch as hi gh t emperature, hi gh humidity, a nd high pressure. A typical test condition can be 121°C, 100% relative humidity (RH), and 2 atm pressure. It has also been known as the autoclave or pressure cooker test (PCT). Ā e moisture absorbed by the p olymeric m aterials c an h ydrolyze t he i nterfacial b onds between underll and the die, resulting in delamination starting from the corner of the die, which further promotes the moisture diffusion a long t he i nterface. Ā e mo isture at t he i nterface c an cause corrosion of the solder joints and metal traces on the substrate. Delamination decouples t he u nder ll with the Si die and can cause stress concentration on t he surrounding solder joints, leading to early fatigue failure of those joints. Ā e absorbed moisture a lso c auses h ygroscopic s welling. L ahoti e t a l. s tudied t he combined effect of moisture and temperature on the reliability of ip-chip ball grid array (FCBGA) packages using FE analysis. Ā e simulation re sults re vealed t he si gnicance of h ygroscopicinduced tensile stress on the under bump metallurgy (UBM) and interdielectric layer (ILD) [18]. Interfacial delamination of under ll to various materials such as d ie pa ssivation, s older m aterial, a nd s older m ask o n t he substrate i s a le ading c ause f or f ailure i n  ip-chip u nder ll packages. One way to improve the reliability under temperature humid aging is to incorporate adhesion promoters, or coupling agents, i nto u nder ll to increase adhesion of under ll t o th e surrounding materials. Luo e t al. studied six different coupling agents and their effect on underll. Ā e authors found that the incorporation of the coupling agents clearly affected the curing prole and bulk property of t he underll, such as Tg and modulus . Ā e effect of c oupling a gents on a dhesion a nd a dhesion retention after temperature moisture aging was highly dependent on coupling agent type and interacting surfaces. Ā e addition of

43722_C015.indd 5

TABLE 15.1 Desirable Under ll Properties for Flip-Chip Packages Curing Temperature Curing time Tg Working life (viscosity double @ 25°C) CTE (a1) Modulus Fracture toughness Moisture absorption (8 h boiling water) Filler contents

<150°C <30 min >125°C >16 h 22–27 ppm/°C 8–10 GPa >1.3 MPa*m <0.25% <70 wt 1%

titanate and zirconate coupling agents can improve adhesion of epoxy under ll with BCB-passivated silicon. However, with the addition of t he same coupling agents, t he ad hesion strength of under ll w ith p olyimide pa ssivation de creased a fter a ging at 85C/85% RH [19]. In summary, many studies have concluded that the underll material property is one of the key factors determining the reliability o f t he pac kage. Ā e gener al g uideline o n t he m aterial properties of under ll for ip-chip packages can be summarized in Table 15.1. However, one has to ke ep i n m ind t hat d ifferent failure mo des c oexist i n a rel iability te st, w hich s ometimes present c onicting r equirements o n u nder ll. F or i nstance, to effectively couple t he stress on t he solder joints, h igh-modulus under ll is desired. On the other hand, high underll modulus can lead to a high residual stress and therefore die crack. Another example is the  ller loading. Low CTE requirement on underll indicates high  ller loading. However, an under ll with higher  ller loading typically has a higher viscosity, causing difficulty in under ll dispensing. Ā e result might be under ll voids and nonuniformity, which would cause reliability issues. Ā er efore, the choice of under ll h ighly de pends on t he application, e .g., die size, passivation material, substrate material, type of solder, and en vironment c onditions t he pac kage w ill b e sub jected to during application, etc.

15.4 New Challenges to Underfill As si licon te chnology mo ves to sub -0.1 µm f eature si ze, t he demands for packaging also are involved as the bump pitch gets tighter, bump size smaller, die size larger for future ip-chips. As a re sult, t he c apillary u nder ll p rocess f aces t remendous challenges. As it was discussed previously, the underll ow problem is aggravated as the size of the chip becomes larger and the gap between the chip and substrate gets smaller. Among the emerging development of  ip c hip, le ad-free s older a nd lo w-K (d ielectric constant) ILD/Cu present new challenges to underll [20]. High-lead and lead-tin eutectic solders have been widely used for chip-package interconnections. Recent environmental legislations toward toxic materials and consumers’ demand for green electronics h ave p ushed t he d rive to wards le ad-free s olders. Alternatives have been proposed using multiple combinations of elements like tin, silver, copper, bismuth, indium, and zinc, most

10/3/2008 8:50:17 PM

15-6

Smart Materials

TABLE 15.2 Possible Lead-Free Alloys Alloy Sn96.5/Ag3.5 Sn99.3/Cu0.7 Sn/Ag/Cu Sn/Ag/Cu/X(Sb, In) Sn/Ag/Bi Sn95/Sb5 Sn91/Zn9 Bi58/Sn42

Melting Point (°C) 221 227 217 (ternary eutectic) Ranging according to compositions, usually above 210 Ranging according to compositions, usually above 200 232–240 199 138

of w hich re quire i ncreased re ow tem perature p roles during the soldering process relative to t he well-known tin-lead alloys. Table 15.2 shows some of the common lead-free solders. Among the several lead-free candidate solders, the near ternary eutectic Sn–Ag–Cu (SAC) alloy compositions, with melting temperatures a round 2 17°C, a re b ecoming c onsensus c andidates. Ā e opt imal c omposition 95.4 Sn /3.1Ag/1.5Cu h as p rovided a combination of good strength, fatigue resistance, and plasticity [21]. I n add ition, t he a lloy h as su fficient supp ly a nd ade quate wetting characteristics. Ā e use of the Sn/Ag/Cu solder presents two major challenges on the ip-chip assembly process. First, since the melting point of the alloy is more than 30°C higher than that of the eutectic Sn/Pb alloy, the process temperature is raised by 30°C–40°C. Ā e high process temperature has a g reat impact on the substrate since the conventional FR-4 material has a Tg at around 125°C and also subjects the attached components to a higher thermal stress. Higher warpage is introduced w hen t he b oard i s sub jected to h igher tem perature reow. Ā ere have been considerable research in the high Tg substrate f or t he le ad-free p rocess. Ā e s econd c hallenge c omes from the ux chemistry. Since the current uxes in use are usually designed for eutectic Sn/Pb solder, they either do not have high enough activity or do not possess sufficient thermal stability at high temperature. So generally, the wetting behavior of the lead-free solders is not as good as that of the eutectic Sn/Pb solder [22,23]. With the trends of a lead-free solder interconnect, the under ll for  ip-chip in package application faces new challenges of compatibility with higher reow temperature. High-temperature reow causes component damage due to the enhanced level of material degradation, moisture ingress, and mechanical expansion. Ā erefore, the thermal stability, adhesion to various interfaces, strength, and fracture toughness of the under ll need to be i mproved. Ā e SA C a lloy do es not p lastically de form a s much as the eutectic PbSn solder. Ā e creep deformation is less at a lower stress level and more at higher stress level compared to the PbSn solder, which indicates that the choice of the under ll w ould de pend o n t he app lication ne eds. A tem perature cycle w ith a l arge tem perature d ifference a nd lo wer dw elling times c ould i nduce mo re c reep a nd t herefore re quires mo re protection from the under ll [24]. An evaluation of under ll

43722_C015.indd 6

materials for lead-free application conducted by Intel Corporation [25] showed that the majority of the failures occurred during the moisture sensitive level (MSL) 3 followed by 260°C reow. Delamination seemed to be the common failure mechanism after the high-temperature reow. Ā is failure was also correlated to m aterials w ith lo w f iller c ontent a nd lo w c oupling agent content. In general, materials with high  ller content (and therefore lo w C TE, h igh mo dulus, a nd lo w mo isture uptake) and good adhesion were compatible with the lead-free process. As t he I C f abrication mo ves to ward sm all f eature a nd h igh density, the interconnect delay becomes dominant. Ā is calls out for new interconnect and ILD materials. Ā e Cu metallurgy and low-K ILD has been successfully implemented to increase device speed a nd re duce p ower c onsumption. Ā ese low-K ma terials tend to be porous and brittle, having high CTE and low mechanical strength, compared to the traditional ILD materials such as SiO2. Ā e CTE mismatch between the low-K ILD and the silicon die c reates a h igh t hermomechanical s tress at t he i nterface. Ā erefore, t he c hoice of u nder ll b ecomes c ritical si nce i t not only protects t he solder joints by stress redistribution, but a lso need to protect the low-K IDL and its interface with the silicon. Critical m aterial p roperties o f u nder ll to ac hieve rel iability requirement for the low-K ILD package include the Tg, CTE, and the modulus. However, the optimal combination of these properties is still controversial. Five under lls were evaluated for the low-K ip chip package by Tsao et al. [26]. Both modeling and experimental evaluation indicated that the low Tg and low stress-coupling index under lls yielded better reliability in the low-K ip-chip package. Two moderately l ow Tg un derlls (Tg b etween 7 0°C a nd 1 20°C) showed good potential in protecting both the solder joints and low-K i nterface. A n u nder ll w ith a v ery lo w Tg (l ower t han 70°C), o n t he ot her h and, f ailed to p rotect t he s older jo ints during the thermal cycling test. A study conducted by LSI Logic Corp and Henkel Loctite (now called Henkel) Corp [27], on the other hand, indicated that under lls with high Tg and low modulus is advantageous for the low-K ip chip. Ā e low modulus of the u nder ll e xerts lo wer s tress o n t he pac kage a nd t herefore reducing t he s tress on t he lo w-K layer, preventing underll delamination a nd d ie c racking. Ā e high Tg pre vents s older bump fatigue by maintaining a lo w C TE over t he temperature cycling. Āe high Tg, low modulus under ll developed by Henkel exhibited good manufacturability and reliability in the package qualication testing including JEDEC preconditioning, thermal cycling, biased humidity testing, and high-temperature storage. With a ll t he ne w c hallenges to t he  ip-chip a nd u nderll technology, c apillary u nderfill i s s till t he m ain pac kaging technology for  ip-chip de vices. H owever, t he c ontinuing shrinking of pitch distance and gap h eight will eventually post limitations o n t he c apillary  ow. Ā e i ndustry h as s tarted to look f or a lternatives to c apillary u nderll. Ā e f ollowing s ections de scribe t he ne w de velopments i n u nder ll material and processes.

10/3/2008 8:50:17 PM

15-7

Flip-Chip Underfi ll: Materials, Process, and Reliability

15.5

No-Flow Underfill

Ā e idea of integrated ux and under ll was patented by Pennisi and Papageorge i n Motorola back i n 1992 [28]. It t riggered t he research and development of t he no-ow under ll process. Ā e rst no-ow under ll p rocess w as p ublished b y W ong a nd Baldwin in 1996 [29]. Ā e schematic process steps are illustrated in F igure 1 5.6. I nstead o f u nder ll d ispensing after t he c hip assembly in the conventional process, in a no-ow underll process, t he u nder ll i s d ispensed o nto t he sub strate b efore t he placement of the chip. Ā en the chip is aligned and placed onto the substrate and the whole assembly goes through solder reow where t he i nterconnection t hrough s older ba lls i s e stablished while the solder melts. Ā is novel no-ow process eliminates the separate  ux d ispensing a nd  ux cleaning steps, avoids the capillary ow of under ll, and nally combines the solder bump reow and under ll cured into a single step, hence, improves the production e fficiency of the under ll process. It is a step forward for t he  ip-chip to b e c ompatible w ith su rface mo unt technology (SMT). Ā e key to the success of a no-ow under ll process lies in the under ll material. Āe rst patent on the no-ow underll material w as b y Wong a nd Sh i at G eorgia I nstitute o f Technology [30]. Ā e t wo c ritical p roperties o f no -ow under ll to enable this new process to h ave a l atent curing ability and the built-in uxing c apability. Ā e nat ure o f t he no -ow underll process requires that the under ll has enough reaction latency to maintain its low v iscosity u ntil t he solder joints a re formed. O therwise, g elled u nder ll w ould p revent t he mel ting s olders f rom collapsing onto the contact pads, resulting in a lo w yield of the solder joint. O n t he ot her hand, el imination of t he p ostcure i s desired si nce p ostcure t akes add itional off -line p rocess t ime, adding to the cost of this process. Many latent catalysts for epoxy resins have been explored for the application of no-ow underll. In the material system that Wong and Shi designed, Co(II) acetylacetonate was used as the latent catalyst [31,32], which gave enough c uring l atency for no -ow underll. Ā e adv antage o f metal chelates lies not only in its latent acceleration, but also in the w ide c uring r ange t hey off er. By e xploring d ifferent metal ions and chelates, the curing behavior of different epoxy resins could b e t ailored to t he app lication o f no -ow u nder ll f or

lead-free s older b umped  ip-chip [ 33]. Si nce le ad-free s olders usually h ave a h igher melting p oint t han eute ctic Sn Pb s older, no-ow under ll for lead-free bumped  ip-chip requires higher curing l atency to en sure t he wetting of t he le ad-free s older on the contact pad. Zhang et al. explored 43 different metal chelates and developed no-ow under ll compatible with lead-free solder reow [ 33]. Suc cessful le ad-free b umped  ip-chip o n a b oard package u sing t he no -ow under ll p rocess h as b een demo nstrated [34]. Despite t he i mportance o f t he c uring p rocess o f no -ow under ll, there is little study on the curing kinetics and its relation to t he re ow pro le. I n a n a ttempt to de velop s ystematic methodology to c haracterize t he c uring p rocess o f no -ow under ll, Zhang and Wong used an autocatalytic curing kinetic model w ith tem perature-dependent pa rameters to p redict t he evolution o f de gree o f c ure ( DOC) d uring t he s older re ow process [35]. Figure 15.7 shows the result of DOC calculation of a no-ow underll in eutectic SnPb and lead-free solder reow process. If t he DOC of t he under ll at t he solder melting temperature is lower than the gel point, the molten solder would be allowed to w et t he sub strate a nd m ake t he i nterconnection. Another approach is the in situ measurement of viscosity of noow under ll u sing m icrodielectrometry b y M organelli a nd Wheelock [36]. Since the viscosity is related to the ionic conductivity, t he d ielectric properties of t he u nder ll c an b e u sed for the in-situ analysis of the no-ow under ll in the reow process, which can be used to predict the solder wetting behavior. Ā e ot her ke y p roperty f or no -ow under ll i s t he  uxing capability. I n a c onventional  ip-chip p rocess,  ux is used to reduce and eliminate the metal oxide on the solder and to prevent it f rom b eing re oxidized u nder h igh tem perature. I nstead o f applying  ux, no-ow under ll i s d ispensed b efore t he c hip placement. He nce, t he s elf-uxing capability is required to facilitate solder wetting. To achieve this goal, research has been done to d evelop reow-curable polymer  uxes [37]. A c omprehensive study on the uxing agent of no-ow underll material was carried out by Shi et al. [38–40], which included the relationship between the surface composite on the Cu pad a nd the uxing capability of no-ow under ll, and also the effect of the addition of the uxing agent on the curing and material properties of no-ow underll.

No-flow underfill Heated

Underfill dispensing

Chip placement

Flux, solder reflow, and underfill cure

FIGURE 15.6 Flip-chip process using no-ow under ll.

43722_C015.indd 7

10/3/2008 8:50:18 PM

15-8

Smart Materials

Eutectic SnPb melting point Lead-free solder MP DOC in lead-free reflow

250

1

200

0.8

150

0.6

100

0.4

50

0.2

0 0

50

100 150 Time (s)

200

DOC

Temperature (C)

Eutectic SnPb reflow profile Lead-free reflow profile DOC in eutectic SnPb reflow

0 250

FIGURE 15.7 DOC evolution of a no-ow under ll in eutectic SnPb and lead-free reow process.

Ā e p rocess o f no -ow under ll h as a lways at tracted m uch attention i n t he a ssembly i ndustry. V oids f ormation i s often observed in many ip-chip no-ow under ll pac kages. Ā e origin o f t he vo ids c ould b e t he o ut-gassing o f t he u nderll, moisture in the board, and trapped voids during assembly, etc. Ā ey a re u sually t acked to a s older b ump o r i n b etween t wo bumps [41,42]. Figure 15.8 shows an example of underll voids in a no-ow under ll package observed with scanning acoustic microscope. V oids i n t he u nder ll, e specially vo ids ne ar t he solder b umps le ad to e arly f ailure t hrough a n umber o f w ays including stress concentration, under ll delaminate, and solder extrusion. Studies by Wang et al. have showed that solder bridging m ight re sult f rom t he s older b ump e xtrusion t hrough t he microvoids trapped between adjacent bumps [43]. Ā e material and process factors inuencing the voiding behavior are complicated and interacting. It has been shown that the outgassing of anhydride c ould c ause s evere vo iding, i f t he c uring l atency i s high and also the reow temperature is high; hence, the voiding becomes mo re p rominent i n a le ad-free re flow p rocess [44]. The important process parameters that affects under ll voiding in a no-ow process include the under ll dispensing pattern, the solder m ask de sign, t he p lacement f orce a nd s peed, a nd t he reow pro le, etc. [45,46]. Before a ssembly, t he PW B substrate needs to b e ba ked to d ry o ut a ny mo isture to p revent vo iding from the board [42]. It has been shown that in some cases a fast gelation of under ll is desired to minimize the voiding while in other cases, extending duration at high temperature can “push” out t he vo ids [ 42,47]. I n s hort, w ith t he r ight m aterial a nd process pa rameters, vo iding i n no -ow under ll c an b e m inimized. However, the process window is usually very narrow. An important p oint w as r aised b y Z hao a nd Wong [34] t hat f or a small circuit board where the temperature distribution is more

43722_C015.indd 8

homogenous, it is relatively easy to develop a “good” reow pro le w hile f or c omplex S MT a ssemblies i nvolving m ultiple components and signicant thermal mass difference across the board, t he opt imization o f t he re ow p rocess p resents a g reat challenge. Ā e rel iability o f a  ip-chip no -ow u nder ll p ackage h as been evaluated by many researchers. Discrepancies exist among these reports because the process and reliability of the no-ow

FIGURE 15.8

Example of voids in a no-ow under ll package.

10/3/2008 8:50:18 PM

15-9

Flip-Chip Underfi ll: Materials, Process, and Reliability

under ll pac kage de pends l argely o n t he pac kage de signs including the size of the chip, the pitch, the surface nish of the pad, e tc. A mong t he e arliest re porters o n no -ow under ll, Gamota a nd M elton c ompared t he rel iability a nd t ypical failure mode of a c onventional under ll package a nd no-ow under ll p ackages [ 48]. Ā ey f ound t hat i n a c onventional under ll assembly, the failure of the assembly mainly resulted from t he i nterfacial del amination b etween t he u nder ll a nd the chip passivation. However, with un lled materials in noow under ll, good interfacial integrity was observed, and the assembly failed mainly due to t he fracture through the solder interconnects ne ar t he P CB. Si nce t he no  ow u nder ll w as un lled, the CTE was high. Ā ey argued that the relative localized C TE m ismatch b etween t he c hip, t he u nder ll, a nd t he PCB resulted in a high local stress eld which initiated fracture in t he s older in terconnects. N o-ow u nder ll w ithout s ilica  llers or very low  ller loadings is not only high in CTE, but also low in fracture toughness [49]. Combined with high CTE mismatch, t he low f racture toughness le ads to e arly u nder ll cracking both inside the bulk and in under ll  llet. Fillet cracking c auses del amination b etween t he u nder ll a nd t he die pa ssivation a nd/or b etween t he u nder ll a nd t he b oard, while b ulk c racking c an i nitiate s older jo int c racking a nd solder b ridging [ 50]. Ā ese a ll b ecome t he c ommon f ailure modes for the  ip-chip no-ow under ll package. Efforts have been made to en hance t he toughness of t he no -ow under ll materials through the incorporation of toughening agents [51]. The e ffect o f t he g lass t ransition tem perature ( Tg ) o f t he no-ow u nder ll o n t he rel iability o f t he pac kage h as b een controversial. It is usually believed that the Tg of the underll should exceed the upper limit of the temperature cycling (125°C or 150°C) to ensure consistent material behavior during the reliability te st. H owever, s ome te sts h ave s hown t hat lo w Tg (~70°C) under ll material performed better in liquid-to-liquid thermal shock (LLTS) [52]. Ā e research by Zhang et al. on the development of nonanhydride based no-ow under ll [53] also showed t hat high Tg is not critical to reliability. Although the CTE of the underll above Tg is much higher than that below Tg, the mo dulus o f t he u nderfill de creases d ramatically; s o the overall stress in the under ll does not necessarily increase when the environment temperature exceeds its Tg. But high Tg might result in a higher residue stress inside the under ll after the m aterial c ools d own after c uring, w hich le ads to a n e arly crack in the under ll.

15.5.1 Approaches of Incorporating Silica Fillers into No-Flow Underfi ll Ā e previous research has shown that the correlation between the m aterial properties a nd pac kage rel iability i n t he c ase of  ip-chip under ll is very complicated. It is difficult to separate the effect of each factor since the material properties are often correlated with each other. However, it is generally agreed that low C TE a nd h igh mo dulus a re f avorable f or h igh i nterconnect rel iability [54]. Hence, t he i nclusion of si lica  llers i nto the u nder ll i s c ritical to en hance t he rel iability. H owever, since the under ll is predeposited on the board before the chip assembly in a no-ow process, the  llers are easily trapped in between t he s older b ump a nd c ontact pad a nd h inder t he interconnection [ 55]. Ā e rmocompression reow ( TCR) h as been used to exclude the silica  ller from the solder joint [56]. Ā e process step is illustrated in Figure 15.9. In a TCR process, the under ll is dispensed onto a preheated substrate. Ā e chip is then picked and bonded to the substrate and held at an elevated temperature under force for a certain period of time for solder joint formation. Ā e assembly is postcured after wards. It w as f ound t hat t he b onding f orce a nd tem perature w ere important f actors i nf luencing y ield. A de tailed s tudy w as carried o ut b y K awamoto e t a l. at N AMICS C orporation to determine the effect of  ller on the solder joint connection in a TCR-like no-ow under ll process [57]. Ā e study used two different sizes of silica  llers at different loading levels. It was found that a good solder connection can be made with under ll with up to 60 wt% ller loading without ller surface treatment. Sm aller  ller ten ded to i ncrease t he v iscosity o f t he under ll and more  llers remained at the solder interface due to t he larger number of  llers at t he s ame weight percentage loading. Ā e study also found that proper surface treatment of the  llers can lower the underll v iscosity and increase y ield at high  ller loading. Other appro aches h ave b een e xplored t o i ncorporate s ilica  llers into no-ow under ll. In a novel patented process, Zhang et a l. u sed a do uble-layer no -ow under ll [58], i n w hich t wo layers of no-ow under ll are applied. Ā e bottom layer under ll is relatively high in viscosity and is not  lled with silica llers. It is applied onto the substrate  rst; then the upper layer under ll, which is lled with silica  llers, is dispensed. Ā e chip is then placed onto the substrate and reowed, during which the solder joints a re formed a nd t he u nder ll is c ured or pa rtially c ured.

Heat Force

Underfill dispensing

FIGURE 15.9

43722_C015.indd 9

Heat

Heat

Heat

Chip placement and reflow

Underfill postcure

TCR for ip-chip.

10/3/2008 8:50:19 PM

15-10

Smart Materials

Dispense the bottom layer underfill without fillers

Dispense the upper layer underfill with fillers Heat

Place the chip

Flux, solder reflow, and underfill cure

FIGURE 15.10 Double-layer no-ow under ll process.

Ā e process ow chart is illustrated in Figure 15.10. It was demonstrated t hat h igh y ield w as ac hieved u sing a n upp er l ayer under ll of 65 wt% silica  lled [59]. Fu rther i nvestigations o n the process indicated t hat factors a ffecting the interconnection yield of t he double-layer no-ow under ll are complicated and interacting with each other [60]. Ā e process window is narrow and the thickness and the viscosity of the bottom layer underll are essential to the wetting of the solder bumps. And of course, it adds on a nother s tep i n t he  ip-chip process a nd has a h igher process cost. Ā e recent advances in nanoscience and nanotechnology have enabled innovative research in materials for electronic packaging. It was found that nanosized silica  llers with surface modication c an b e m ixed w ith t hermosetting re sins to p rovide a uniform d ispersion of non agglomerated p articles. Used a s no ow under ll, t he na nocomposite m aterials a llowed 5 0 wt%  ller lo ading w ith a go od i nterconnect y ield [ 61]. Ā is highperformance no-ow under ll developed by 3M used 123 nm silica ller. With ller loading of 50 wt%, the CTE of the material was 4 2 ppm/°C a nd t he go od i nterconnect y ield w as ac hieved using a PB10 die (5°5 mm, 64 peripheral bumps). A joint research study was conducted by 3M and Georgia Institute of Technology in t he p rocess a nd a rel iability e valuation o f t he na nosilicaincorporated no-ow under ll [62]. Figure 15.11 s hows a S EM picture o f t he s older jo int i n t he p resence o f no -ow underll with nanosilica llers. A 1.5× increase of characteristics life was observed in the air-to-air thermal cycling (AATC) reliability test with the nanosilica llers. Although the nanocomposite no-ow under ll m aterial s hows go od p otential f or a h ighly rel iable  ip-chip pac kage u sing a S MT-friendly no -ow under ll process, the fundamental mechanism of the nanosilica interaction w ith t he s older jo ints a nd t he u nder ll i s s till n ot w ell understood. Si nce na nosize pa rticles h ave a l arge su rface a rea and tend to form irregular agglomerations, which increase their difficulty to b e incorporated into a binder, surface treatment of

43722_C015.indd 10

nanosilica is of great importance in formulating an underll. A fundamental study on the surface modication of nanosize silica for under ll application was carried out by Sun et al. [63]. Ā ey found t hat t he t ype o f t he su rface t reatment w as t he p rimary factor affecting the property of the formulation. Using an epoxy silane, t he a uthors s howed t hat t he v iscosity o f t he c omposite under ll was greatly reduced. In su mmary, t he i nvention o f no -ow underll greatly simplies t he  ip-chip u nder ll p rocess a nd d raws  ip-chip toward S MT. A suc cessful no -ow u nder ll p rocess r equires careful i nvestigation i nto t he m aterials a nd p rocess pa rameters. A great deal of the research effort has been devoted to the materials, process, and reliability of  ip-chip no-ow under ll assembly. Si nce t he u nderll does not contain silica ller and hence b ehaves d ifferently from t he c onventional u nder ll, t he

+

+

Underfill with nanosilica

Da 2 = 138.8 nm Db 2 = 15131. nm2

Da 1 = 121.4 nm D b 1 = 11584. nm2

Solder joint 20 k magnification 1 µm

1.5 kV

FIGURE 15.11 SEM picture of a s older joint w ith nano-silica incorporated no-ow under ll.

10/3/2008 8:50:19 PM

15-11

Flip-Chip Underfi ll: Materials, Process, and Reliability

failure modes and reliability concerns are sometimes also different from the conventional ip-chip under ll assembly. Āer e are several ways to enhance the reliability of a ip-chip no-ow under ll package. One way is to enhance the fracture toughness of the under ll without degrading other material properties to prevent u nder ll c racking i n t he t hermal c ycling. O r, lo w Tg and lo w mo dulus m aterials h ave b een u sed to de crease t he stress in the under ll. However, this approach diminishes the role of t he underll as a stress redistribution layer, and although it does decrease the stress in under ll, it cannot prevent solder joint f atigue f ailure f rom t he t hermomechanical s tress, e specially in the case of the large chip, high I/O counts, and small pitch size applications. Another way is to add si lica  llers into the u nder ll a nd m atch t he p roperties o f a c onventional under ll. I n o rder to o vercome t he d ifficulty of  ller entrapment, different approaches have been explored. However, these approaches are less SMT transparent and diminish the low-cost purpose of a no-ow under ll process. Nanosilica-incorporated no-ow u nder ll h as s hown t he p otential of a h ighly rel iable  ip-chip pac kage w ith a S MT-friendly no -ow u nder ll p rocess. H owever, f undamental u nderstanding o f t he na nosilica and its interaction with under ll and solder is still lacking and further d evelopment i s ne eded to opt imize t he m aterials and processes.

15.6

Molded Underfill

Epoxy mold ing c ompounds ( EMCs) h ave b een p racticed i n component packaging for a lo ng time. Ā e novel idea of combining o vermolding a nd t he u nder ll toge ther re sults i n a molded u nder ll [64,65]. Molded under ll i s applied to a  ipchip in a pac kage via a t ransfer molding process, during which the mold ing compound not on ly  lls t he gap b etween t he chip and t he sub strate b ut a lso en capsulates t he w hole c hip [66]. It offers the advantages of combining the under lling and transfer molding i nto one s tep for re duced process t ime a nd i mproved mechanical stability [67]. It also utilizes EMCs, which have long been proven to provide superior package reliability. Compared with the conventional under ll, which is usually lled with silica at around 50–70 wt%, molde d u nder ll c an a fford a m uch higher  ller content up to 80 wt%, which offers a low CTE closely matched w ith t he s older jo int a nd t he b oard. A lso, c ompared with the conventional mold ing c ompound, mold ed u nder ll

Mold inlet

requires  llers i n sm aller si zes, w hich a lso c an c ontribute to lowering t he C TE o f t he m aterial [ 68]. M olded u nder ll i s especially suitable for ip-chip in package to improve production efficiency. It was reported that a f ourfold production rate increase can b e e xpected u sing molde d u nder ll v ersus a c onventional under ll process [69]. Molded u nder ll resembles t he pressurized u nderll encapsulation [70] in the mold design and process except that the materials i n u se a re not l iquid en capsulants t hat o nly  ll u p th e g ap between the chip and the substrate, but rather molding compounds that overmold the entire components. Figure 15.12 shows a design of the mold for FCBGA components using molded underll. Ā e de sign o f t he mold f aces t he c hallenge t hat t he  ip-chip geometry has a higher resistance to the mold ow so that air might be trapped under the chip. In fact, voids have been observed in the molded u nderll packages using an acoustic microscope [71]. Several molding processes can be used to minimize this geometry effect [72]. One way is to use mold vents as shown in Figure 15.12 and to u se a lso ge ometrical opt imization to c reate si milar  ow resistance o ver a nd u nder t he c hip. O ne c an a lso u se v acuumassisted molding to prevent air entrapment. Another approach is to design a cavity in the substrate as shown in Figure 15.12. Āou gh it re quires a s pecial de sign o n t he sub strate, t his me thod h as proved to be a robust process and is commonly adopted. Important process parameters in a molde d underll process include t he mold ing tem perature, c lamp f orce, a nd i njection pressure [ 56]. H igh-temperature mold ing i s f avored f or lo wer viscosity of the molding compound and hence better ow properties and less stress on the solder joint. However, the upper limit of the molding temperature is the melting point (Tm) of the solder material. Temperature ne ar Tm c ombined w ith hi gh in jection pressure m ight c ause t he solder to mel t a nd e ven t he d ie to b e “swept” away from the site. Also a low-Tg substrate is likely to be damaged at high molding temperature and high clamp force. Flash is affected by both the clamp force and the injection pressure. Ā e overow of the molding compound might contaminate other contact pads or testing pads on the substrate. Bump cracking and die cracking are likely to occur as a result of high injection pressure. In short, a successful molded under ll process requires a c ombined e ffort i n m aterial s election, mold de sign, a nd process opt imization. But t he p otential c ost re duction a nd reliability en hancement of molde d u nder ll is attracting great efforts in the industry.

Mold vent Vent hole

FIGURE 15.12 Design of ip-chip BGA with molded under ll.

43722_C015.indd 11

10/3/2008 8:50:20 PM

15-12

Smart Materials

15.7 Wafer Level Underfill The i nvention o f no -flow u nderfill el iminates t he c apillary flow a nd c ombines f luxing, s older re flow, a nd u nderfill curing i nto o ne s tep, w hich g reatly si mplifies t he u nderfill process. H owever, a s p ointed o ut p reviously i n t he c hapter, no-flow underfill has some inherent disadvantages including the u navailability of a h eavily f illed m aterial, w hich i s a b ig concern for high-reliability packages. Also, the no-flow process s till ne eds a n i ndividual u nderfill d ispensing s tep a nd therefore is not totally transparent to standard SMT facilities. An improved concept, wafer level underfill was proposed as a SMT-compatible f lip-chip p rocess to ac hieve lo w c ost a nd high reliability [73–76]. The schematic process steps are illustrated in Figure 15.13. In this process, the underfill is applied either onto a bumped wafer or a wafer without solder bumps, using a proper method, such as printing or coating. Then the underfill is B -staged a nd wafer is d iced i nto si ngle chips. In the c ase o f u nbumped w afer, t he w afer i s b umped b efore dicing w hen t he u nderfill c an b e u sed a s a m ask. T he i ndividual c hips a re t hen p laced o nto t he sub strate b y s tandard SMT assembly equipment. It is noted that in some types of WLCSP, a polymeric layer is a lso u sed o n t he w afer s cale to re distribute t he I /O o r to enhance the reliability. However, this polymeric layer usually does not glue with the substrate and cannot be considered as underll. Ā e wafer level under ll discussed here is an adhesive to g lue c hip a nd sub strate toge ther a nd f unctions a s a stress-redistribution layer rather than a stress-buffering layer. Ā e attraction of the wafer level under ll lies in the potential low-cost p otential ( since i t do es not re quire a sig nicant change in the wafer backend of the line process) and high reliability of the assembly enhanced with the under ll. However, the w afer le vel u nder ll f aces c ritical m aterial a nd p rocess challenges including uniform under ll  lm deposition on the wafer, B-stage process for the under ll, dicing, and storage of B-staged under ll, uxing capability, shelf-life, solder wetting

Apply underfill onto wafer

in the presence of under ll, desire for no postcure and reworkability, e tc. Si nce t he w afer le vel u nder ll p rocess su ggests a convergence of front-end and back-end of the line in package manufacturing, close cooperation between chip manufacturers, pac kage c ompanies, a nd material suppliers a re re quired. Several research programs have been carried out the cooperatively in this area [77–79]. Innovative ways of addressing the above issues and examples of wafer level processes are presented in this chapter. In most wafer level under ll processes, the applied underfill must b e B -staged b efore t he si ngulation o f t he w afer. T he B-stage process usually involves partial curing, solvent evaporation, or both, of the underfill. In order to facilitate dicing, storage, a nd h andling, t he B -staged u nderfill m ust app ear solid-like a nd p ossess enough me chanical i ntegrity a nd s tability after B-stage. However, in the final assembly, the underfill i s re quired to p ossess “ reflowability,” i .e., t he a bility to melt and flow to allow the solder bumps to wet the contacting pads a nd f orm s older jo ints. T herefore, t he c ontrol o f t he curing process a nd t he B -stage properties of t he u nderfill is essential for a successful wafer level underfill process. A study conducted at G eorgia I nstitute o f T echnology ut ilized t he curing k inetics mo del to c alculate t he D OC e volution o f different u nderfills d uring s older r eflow p rocess [ 80]. Combined w ith t he gel ation b ehavior o f t he u nderfills, t he solder w etting c apability d uring re flow w as p redicted a nd confirmed e xperimentally. B ased o n t he B -stage p rocess window and the material properties of the B-staged underfill, a suc cessful w afer le vel u nderfill m aterial a nd p rocess w ere developed. Full area array at 200 µm pitch flip-chip assembly with t he de veloped w afer le vel u nderfill w as a lso demo nstrated [81] as shown in Figure 15.14. Ā e above study shows that the control of the B-stage process of the wafer level underll is critical to achieve good dicing and storage p roperties a nd t he s older i nterconnect o n t he b oard-level assembly. O ne w ay to a void d icing i n t he p resence o f no nfully cured underll is presented in Figure 15.15, a w afer scale applied

Wafer bumping Flip chip assembly Wafer dicing

Wafer bumping

Apply underfill onto wafer

FIGURE 15.13 Process steps of wafer level under ll.

43722_C015.indd 12

10/3/2008 8:50:20 PM

15-13

Flip-Chip Underfi ll: Materials, Process, and Reliability

500 µm 03-Jun-03

WD28.7 mm

X60

500 µm

FIGURE 15.14 A 200 µm pitch wafer level under ll process: under ll coated wafer (left) and assembly on the substrate (right).

Pick and place

Flux

coating

Reflow

Bulk underfill

coating

FIGURE 15.15 Wafer scale applied reworkable uxing under ll process.

reworkable  uxing underll pro cess d eveloped by Motorol a, Loctite, and Auburn University [77]. Since uncured underll materials are likely to absorb moisture that leads to potential voiding in the assembly, in this process, the wafer is diced prior to u nderll coating. Two dissimilar materials are applied; the ux layer coating by s creen or s tencil printing a nd t he bulk u nderll coating by a modied screen printing to keep the saw street clean. Ā e separation of the ux from the bulk underll material preserves the shelf life of the bulk underll as well as prevents the deposition of llers on the top of the solder bump so as to ensure the solder joint interconnection in the ip-chip assembly. In this process, no additional ux d ispensing o n t he b oard i s ne eded a nd h ence t he u nderll needs to b e t acky i n t he  ip-chip b onding process to en sure t he attachment of the chip to the board, as discussed previously.

43722_C015.indd 13

Under ll de position o n t he w afer u sing l iquid m aterial v ia coating o r p rinting r equires s ubsequent B- staging, w hich i s often tricky and problematic. Ā e process developed by 3M and Delphi-Delco circumvents t he B -stage step using  lm lamination [82]. Ā e process steps are shown in Figure 15.16, in which the solid  lm comprised of thermoset/thermoplastic composite is laminated onto the bumped wafer in vacuum. Heat is applied under a vacuum to ensure the complete wetting of the lm over the w hole w afer a nd to e xclude a ny voids. Ā en a p roprietary process is carried out to expose the solder bump without altering the original solder shape. Ā e subsequent ip-chip assembly is carried out in a no-ow under ll-like process in which a curable ux adhesive is applied on the board and then the assembly is reowed.

10/3/2008 8:50:21 PM

15-14

Smart Materials

Pick and place

Vacuum

laminating

Reflow Bump

exposing

FIGURE 15.16 Wafer-applied under ll  lm laminating process.

Wafer level under ll can also be applied before the bumping process. F igure 1 5.17 s hows a m ultilayer w afer-scale u nderll process developed by Aguila Technologies, Inc. [83]. Ā e highly  lled wafer level under ll is screen printed onto an unbumped wafer and then cured. Ā en this material is laser-ablated to form microvias t hat e xpose t he b ond pad s. Ā e v ias a re  lled with solder paste and reowed. Bumps are formed on t he top o f t he  lled vias. Āe ip-chip assembly is similar to the no-ow under ll process again with a polymer  ux dispensed onto the board before chip placement. One similarity among all these three processes is the separation o f  ux m aterial f rom t he b ulk u nder ll. W afer le vel under ll p rocesses p rovide t he c onvenience o f s eparating different f unctionalities by using dissimilar materials so t hat “the one magic material that solves everything” is not required.

However, it is likely to create inhomogeneity inside the under ll l ayer, t he i mpact o f w hich o n t he rel iability i s not f ully understood. A novel photodenable material, which acts both as a photoresist an d an un der ll l ayer app lied o n t he w afer le vel, w as reported by the Georgia Institute of Technology [84]. In the proposed process as shown in Figure 15.18, the wafer level underll is applied on t he u nbumped w afer, a nd t hen i s e xposed to t he UV light with a mask for cross-linking. After development, the unexposed material is removed and the bump pads on the wafer are exposed for solder bumping. Ā e f ully c ured  lm is left on the wafer and acts as t he under ll during t he subsequent SMT assembly a fter de vice si ngulation. A p olymeric  ux is n eeded during the assembly for holding the device in place on the board and providing a  uxing capability, a p rocess similar to t he dry

Pick and place Underfill coating and cure

Laser-drill via Reflow

Via filling

Wafer bumping

FIGURE 15.17 Multilayer wafer-scale under ll process.

43722_C015.indd 14

10/3/2008 8:50:23 PM

15-15

Flip-Chip Underfi ll: Materials, Process, and Reliability

Wafer without bumps

Apply photo underfill

Mask

Fully photocuring underfill Development

Solder bump at opening

Water dicing and reflow

FIGURE 15.18 Photodenable wafer level under ll process.

 lm l aminated w afer le vel u nder ll. I n o rder to en hance t he material p roperty, t he add ition o f si lica  llers i s ne cessary. I n this case, nanosized silica was used to avoid UV light scattering, which hinders the photocross-linking process. It also resulted in an opt ical t ransparent  lm on t he wafer to f acilitate t he v ision recognition during d icing a nd a ssembly process. Ā e photodenable nanocomposite wafer level under ll presents a cost-effective way of applying wafer le vel u nder ll a nd has potentially a ne-pitch capability. Since wafer level under ll is a relatively new concept and most research is still in t he process a nd material development stage, there a re f ew re ports o n t he rel iability o f a  ip-chip package using wafer level under ll. Although there is no standard process for wafer level under ll yet, the nal decision might depend on t he wafer a nd chip size, bump pitch, a nd package t ype, etc. Like w afer le vel C SP, m ultiple s olutions c an c oexist f or w afer level underll process.

15.8

Summary

Flip-chip o ffers m any adv antages o ver ot her i nterconnection technologies and is practiced in many applications. Underll is necessary for a rel iable f lip-chip on an organic package but is process-unfriendly a nd becomes t he bottleneck of a h ighproduction  ip-chip a ssembly. A s si licon te chnology moves to nanometer nodes with feature size less than 65 nm, a shrinking pitch and gap distance, as well as new developments in lead-free solder and low-K ILD/Cu interconnect present new challenges to under ll materials and processes. Many variations of the conventional under ll have been invented that address the problem, a mong w hich, t he ne wly d eveloped no -ow underll, molded under ll, and wafer level under ll have attracted much

43722_C015.indd 15

attention. Ā e no-ow under ll process si mplies t he c onventional  ip-chip under ll p rocess b y i ntegrating  ux i nto th e under ll, el iminating c apillary  ow, a nd c ombining s older reow and under ll cure into one step. However, the predeposited under ll cannot contain high levels of silica  ller due to the interference of  ller w ith s older jo int f ormation. Ā e resulting high C TE o f t he u nder ll l imits t he rel iability of t he pac kage. Various w ays h ave b een e xplored to en hance t he rel iability through improved fracture toughness of the underll material, low Tg, a nd lo w mo dulus u nder ll, a nd t he i ncorporation o f  llers u sing ot her p rocess app roaches. M olded u nderll combines under ll and overmold together and is especially suitable for ip-chip in packaging to improve the capillary underll ow and t he production efficiency. Careful material selection, mold design, and process optimization are required to achieve a robust molded u nder ll process. Wafer le vel u nder ll presents a convergence of f ront-end a nd bac k-end i n pac kaging manufacturing and may provide a solution for low-cost and a highly reliable ip-chip process. Various material and process issues including under ll deposition, wafer dicing with under ll, shelf-life, vision recognition, chip placement, and solder wetting with underll, etc., have been addressed through novel material developments and different process approaches. Although the research is still in its early stages, there is no standard in the process yet developed. Additionally, there have been considerable successes in demonstrating the process, which looks promising for future packaging manufacturing. A ll o f t hese t hree app roaches re quire c lose cooperation between the material suppliers, package designers, assembly c ompanies, a nd m aybe c hip m anufacturers. A go od understanding in both the materials and the processes and their interrelationship is e ssential to ac hieve successful packages for the future success of these developments.

References 1. C.P. Wong, S. Lou, and Z. Zhang, Flip the chip, Science, 290, 2269, 2000. 2. E. Davis, W. Harding, R. Schwartz, and J. Coring, Solid logic technology: Verdatile hig h p erformance micr oelectronics, IBM Journal of Research & Development, 8, 102, 1964. 3. F. Nakano, T. Soga, and S. Amagi, Resin-insertion effect on thermal c ycle resistivity of  ip-chip mounted LSI de vices, the P roceedings o f the I nternational S ociety o f H ybrid Microelectronics Conference, p. 536, 1987. 4. Y. T sukada, S urface la minar cir cuit a nd  ip-chip attach packaging, Proceedings of the 42nd Electronic Components and Technology Conference, p. 22, 1992. 5. B. Han and Y. Guo, Ā ermal deformation analysis of various electronic pac kaging p roducts b y M oire a nd micr oscope Moire interferometry, Journal Elec tronic Packaging, 117, 185, 1995. 6. S. Han and K.K. Wang, Analysis of the  ow of encapsulant during un derll enca psulation o f  ip-chips, IEEE Transactions on Components, Packaging, and Manufacturing Technology, Part B, 20(4), 424–433, 1997.

10/3/2008 8:50:23 PM

15-16

7. S. Han, K.K. Wang, and S.Y. Cho, Experimental and analytical study on the ow of encapsulant during underll encapsulation of  ip-chips, P roceedings o f the 46th Ele ctronic Components a nd T echnology C onference, p p. 327–334, 1996. 8. L. Nguyen, C. Quentin, P. Fine, B. Cobb, S. Bayyuk, H. Yang, and S.A. Bidstrup-Allen, Underll of ip chip on laminates: Simulation and validation, IEEE Transactions on Components and Packaging Technology, 22(2), 168–176, 1999. 9. W.B. Young and W.L. Yang, Underll of ip-chip: Ā e effect of c ontact ang le and s older bu mp ar rangement, IEEE Transactions on Advanced Packaging, 29(3), 647–653, 2006. 10. Coffin, A st udy o f the ef fect o f c yclic thermal str esses on a d uctile met al, ASME T ransactions, 76, 931–950, 1954. 11. Manson, Fatigue: A complex subject-some simple approximations, Experimental Mechanics, 5, 193–226, 1965. 12. H.D. S olomon, Fatigue o f 60/40 s older, IEEE Transactions on c omponents, H ybrids a nd M anufacturing T echnology, CHMT-9(4), 423–432, 1986. 13. H. Bressers, P. Beris, J. Caers, and J. Wondergerm, Influence of ch emistry an d pro cessing of f lip ch ip u nderfills on reliability, 2nd I nternational C onference o n Adhesive Joining and Coating Technology in Electronics Manufacturing, Stockholm, Sweden, 1996. 14. J.B. N ysaether, P. L undstrom, a nd J . L iu, M easurements o f solder bump lifetime as a function of underll material properties, IEEE T ransactions o n Co mponents, P ackaging a nd Manufacturing Technology, Part A, 21(2), 281–287, 1998. 15. R. Dudek, A. S chubert, a nd B . Mic hel, Analyses o f f lip chip attach r eliability, P roceedings o f 4th I nternational Conference o n Adhesive Joining a nd C oating Technology in Electronics Manufacturing, pp. 77–85, 2000. 16. P. Palaniappan, P. Selman, D. Baldwin, J. Wu, and C.P. Wong, Correlation o f  ip c hip u nderll p rocess pa rameters a nd material p roperties wi th in-p rocess s tress g eneration, Proceedings o f the 48th Ele ctronic C omponents a nd Technology Conference, pp. 838–847, 1998. 17. L. Mercado and V. Sarihan, Evaluation of die edge cracking in ip-chip PBGA packages, IEEE Transactions on Components and Packaging Technologies, 26(4), 719–723, 2003. 18. S.P. L ahoti, S.C. K allolimath, a nd J . Zho u, Fini te elemen t analysis of thermo-hygro-mechanical failure of a  ip chip package, Proceedings of IEEE 6th International Conference on Electronic Packaging Technology, 2005. 19. S. Luo and C.P. Wong, Effect of coupling agents on underll material in  ip c hip pac kaging, P roceedings o f 2000 I nternational S ymposium o n Advanced P ackaging M aterials, p p. 183–188, 2000. 20. T. Chen, J. Wang, and D. Lu, Emerging challenges of underll f or  ip c hip a pplication, P roceedings o f the 54th Electronic C omponents a nd T echnology C onference, p p. 175–179, 2004. 21. J.S. Hwang, Lead-free solder: the Sn/Ag/Cu system, Surface Mount Technology, 18, July 2000.

43722_C015.indd 16

Smart Materials

22. B. Huang and N.C. Lee, Prospect of lead free alternatives for reow soldering, Proceedings o f S PIE—Ā e International Society for Optical Engineering, 3906, 771, 1999. 23. A. Buttereld, V. V isintainer, and V. Goudarzi, Lead-free solder past e  ux e valuation a nd im plementation in p ersonal co mmunication de vices, Pr oceedings o f t he 50t h Electronic C omponents a nd T echnology C onference, p. 1420, 2000. 24. S. Mahalingam, K. Goray, and A. Joshi, Design of under ll materials for lead f ree  ip chip application, Proceedings of 2004 IEEE I nternational S ociety C onference o n Ā ermal Phenomena, pp. 473–479, 2004. 25. C.K. Chee, Y.T. Chin, T. Sterrett, Y. He, H.P. Sow, R. Manepali, and D. Chandran, Lead-free compatible underll materials for ip chip application, Proceedings of the 52nd Electronic Components a nd T echnology C onference, p p. 417–424, 2002. 26. P. Tsao, C. Huang, M. Li, B. Su, and N. Tsai, Underll characterization for low-k dielectric/Cu interconnect IC ip-chip package r eliability, P roceedings o f the 54th IEEE Ele ctronic Components and Technology Conference, pp. 767–769, 2004. 27. S. Rajagopalan, K. Desai, M. Todd, and G. Carson, Underll for lo w-K silico n t echnology, P roceedings o f 2004 IEEE/ SEMI International Electronics Manufacturing Technology Symposium, 2004. 28. R. Pennisi a nd M. P apageorge, Adhesive a nd Enca psulant Material wi th Fl uxing P roperties, U .S. P atent 5,128,746, (July 7, 1992). 29. C.P. Wong and D. Baldwin, No-Flow Underll for Flip-Chip Packages, U.S. Patent Disclosure, April 1996. 30. C.P. Wong and S.H. Shi, No-Flow Underll of Epoxy Resin, Anhydride, Fl uxing Agent a nd S urfactant, U .S. P atent 6,180,696, (Jan. 30, 2001). 31. C.P. Wong, S.H. Shi, and G. Jefferson, High performance no ow underlls f or lo w-cost  ip-chip applications, Proceedings o f the 47th Ele ctronic C omponents a nd Technology Conference, p. 850, 1997. 32. C.P. Wong, S.H. Shi, and G. Jefferson, High performance noow underlls f or  ip-chip a pplications: M aterial cha racterization, IEEE T ransactions o n Co mponents, P ackaging, and M anufacturing T echnology, P art A: P ackaging Technologies, 21(3), 450, 1998. 33. Z. Zhang, S.H. Shi, and C.P. Wong, Development of no-ow underll materials for lead-free bumped  ip-chip applications, IEEE T rans. o n Co mponents, a nd P ackaging Technologies, 24(1), 59–66, 2000. 34. Z. Zhang and C.P. Wong, Development of no-ow underll for lead-f ree b umped  ip-chip ass emblies, Pr oceedings o f Electronics P ackaging Technology C onference, p p. 234–240, Singapore, 2000. 35. Z. Zhang and C.P. Wong, Study and modeling of the curing behavior of n o-ow underll, P roceedings o f the 8th International S ymposium a nd E xhibition o n Advanced Packaging M aterials Pr ocesses, Pr operties a nd I nterfaces, pp. 194–200, Stone Mountain, Georgia, 2002.

10/3/2008 8:50:24 PM

Flip-Chip Underfi ll: Materials, Process, and Reliability

36. P. Morganelli and B. Wheelock, Viscosity of a no-ow underll d uring r eow a nd i ts r elationship t o s older w etting, Proceedings o f the 51st Ele ctronic C omponents a nd Technology Conference, pp. 163–166, 2001. 37. R.W. J ohnson, M.A. C apote, S. Ch u, L. Zho u, a nd B . Gao , Reow-curable p olymer  uxes for  ip c hip e ncapsulation, Proceedings o f I nternational C onference o n M ultichip Modules and High Density Packaging, pp. 41–46, 1998. 38. S.H. Shi and C.P. Wong, Study of the uxing agent effects on the prop erties of n o-ow underll ma terials f or  ip-chip applications, Proceedings of the 48th Electronic Components and Technology Conference, p. 117, 1998. 39. S.H. Shi, D. Lu, and C.P. Wong, Study on the r elationship between the surface co mposition of copper pads a nd noow underll uxing ca pability, P roceedings o f the 5th International S ymposium on Advanced P ackaging Materials: P rocesses, P roperties a nd I nterfaces, p . 325, 1999. 40. S.H. Shi and C.P. Wong, Study of the uxing agent effects on the prop erties of n o-ow underll ma terials f or  ip-chip applications, IEEE T ransactions o n Co mponents a nd Packaging T echnologies, P art A: P ackaging T echnologies, 22(2), 141, June 1999. 41. P. P alm, K. Puhak ka, J . M aattanen, T . H eimonen, a nd A. Tuominen, Applicability of no-ow uxing encapsulants and ip chip technology in volume production, Proceedings of the 4th I nternational C onference o n Adhesive J oining and C oating T echnology in E lectronics M anufacturing, pp. 163–167, 2000. 42. K. Puhakka and J.K. Kivilahti, High density  ip chip interconnections produced with in-situ underlls and compatible s older coa tings, P roceedings o f the 3r d I nternational Conference on Adhesives Joining and Coating Technology in Electronics Manufacturing, pp. 96–100, 1998. 43. T. Wang, T .H. Che w, Y.X. Che w, a nd L. F oo, Relia bility studies of ip chi p p ackage w ith r eowable underll, Proceedings of the Pan Pacic Microelectronic Symposium, Kauai, Hawaii, February 2001, pp. 65–70. 44. Z. Zha ng a nd C.P. Wong, Assembly o f lead-f ree b umped ip-chip wi th no- ow underlls, IEEE T ransactions o n Electronics Packaging Manufacturing, 25(2), 113–119, 2002. 45. D. Miller a nd D.F. B aldwin, Eff ects o f subst rate desig n o n underll v oiding usin g the lo w cost, hig h thr oughput  ip chip assembly process, Proceedings of the 7th International Symposium o n Advanced P ackaging M aterials: P rocesses, Properties and Interfaces, pp. 51–56, 2001. 46 R. Zhao , R .W. J ohnson, G. J ones, E. Yaeger, M. K onarski, P. K rug, a nd L. Cra ne, P rocessing o f  uxing underlls for ip chip-on-laminate assembly, Presented at IPC S MEMA Council APEX 2002, Proceedings of APEX, San Diego, CA, pp. S18-1-1–S18-1-7, 2002. 47. T. Wang, C. L um, J . K ee, T.H. Che w, P. Miao , L. F oo, a nd C. Lin, Studies on a reowable underll for ip chip application, Proceedings of the 50th Ele ctronic C omponents and Technology Conference, pp. 323–329, 2000.

43722_C015.indd 17

15-17

48. D. Gamota and C.M. Melton, Ā e d evelopment of re owable materials systems to integrate the reow and underll dispensing p rocesses f or D CA/FCOB ass embly, IEEE Transactions o n Co mponents a nd P ackaging T echnologies, Part C, 20(3), 183, July 1997. 49. X. Dai, M.V. Brillhart, M. Roesch, and P.S. Ho, Adhesion and toughening me chanisms a t u nderll in terfaces f or  ip-chipon-organic-substrate pac kaging, IEEE T ransactions o n Components and Packaging Technologies, 23(1), 117–127, 2000. 50. B.S. Smith, R. Ā orpe, and D.F. Baldwin, A reliability and failure mode analysis of no  ow underll materials for low cost  ip chip assembly, Proceedings of 50th Electronic Components & Technology Conference, pp. 1719–1730, 2000. 51. K.S. M oon, L. F an, a nd C.P. Wong, S tudy o n the eff ect of toughening o f no- ow underll o n  llet cracking, Proceedings o f the 51st Ele ctronic C omponents a nd Technology Conference, pp. 167–173, 2001. 52. H. Wang and T. Tomaso, Novel single pass reow encapsulant f or  ip c hip a pplication, P roceedings o f the 6th International Symposium on Advanced Packaging Materials: Process, Properties, and Interfaces, pp. 97–101, 2000. 53. Z. Zhang, L. Fan, and C.P. Wong, Development of environmental f riendly no n-anhydride no- ow underlls, IEEE Transactions o n Co mponents a nd P ackaging T echnologies, 25(1), 140–147, 2002. 54. S.H. Shi, Q. Yao, J. Qu, and C.P. Wong, Study on the correlation of  ip-chip r eliability wi th me chanical p roperties o f no-ow underll ma terials, P roceedings o f the 6th International Symposium on Advanced Packaging Materials: Processes, Properties and Interfaces, pp. 271–277, 2000. 55. S.H. S hi a nd C.P. Wong, Re cent ad vances in the de velopment o f no- ow underll enca psulants—a p ractical approach t owards the ac tual ma nufacturing a pplication, Proceedings o f the 49th Ele ctronic C omponents a nd Technology Conference, p. 770, 1999. 56. P. Miao, Y. Che w, T. Wang, a nd L. Foo, Flip-chip ass embly development via mo died reowable underll process, Proceedings o f the 51st Ele ctronic C omponents a nd Technology Conference, pp. 174–180, 2001. 57. S. Kawamoto, O. Suzuki, and Y. Abe, Āe effect of ller on the solder connection for no-ow underll, Proceedings of the 56th Electronic C omponents and Technology C onference, pp. 479–484, 2006. 58. Z. Zha ng, J . L u, a nd C.P. Wong, P rovisional P atent 60/288, 246: A Novel Process Approach to Incorporate Silica Filler into No-Flow Underll, 5–2–2001. 59. Z. Zhang, J. Lu, and C.P. Wong, A novel approach for incorporating silica  llers into no-ow underll, Proceedings of the 51st Ele ctronic C omponents a nd T echnology Conference, pp. 310–316, 2001. 60. Z. Zha ng a nd C.P . Wong, N ovel  lled no-ow underll materials and process, Proceedings of the 8th nI ternational Symposium an d E xhibition on Advanced P ackaging Materials Processes, Properties and Interfaces, pp. 201–209, 2002.

10/3/2008 8:50:24 PM

15-18

61. K.M. G ross, S. H ackett, D .G. L arkey, M.J . S cheultz, a nd W. Ā ompson, New materials for high performance no-ow underll, Symposium Proceedings of IMAPS 2002, Denver, September 2002. 62. K. G ross, S. H ackett, W. S chultz, W. Ā ompson, Z . Z hang, L. F an, a nd C.P. Wong, N anocomposite under lls f or  ip chip application, Proceedings of the 53r d Electronic C omponents and Technology Conference, pp. 951–956, 2003. 63. Y. Sun, Z. Zhang, and C.P. Wong, Fundamental research on surface modication of nano-size silica for underll applications, P roceedings o f the 54th Ele ctronic C omponents and Technology Conference, pp. 754–760, 2004. 64. P.O. Weber, Chip Package with Molded Underll, U.S. Patent 6,038,136, March 14, 2000. 65. P.O. Weber, Chip Package with Transfer Mold Underll, U.S. Patent 6, 157,086, December 5, 2000. 66. K. Gilleo, B. Cotterman, and T. Chen, Molded underll for ip chi p in p ackage, H igh D ensity I nterconnection, p . 2 8, June 2000. 67. T. Braun, K.F. B ecker, M. Koch, V. B ader, R . Aschenbrenner, and H. Reichl, Flip chip molding—recent progress in ip chip encapsulation, P roceedings o f 8th I nternational Advanced Packaging Materials Symposium, pp. 151–159, March 2002. 68. F. Liu, Y.P. Wang, K. Chai, and T.D. Her, Characterization of molded underll material for ip chip ball grid array packages, P roceedings of the 51st Ele ctronic C omponents a nd Technology Conference, pp. 288–292, 2001. 69. L.P. Re ctor, S. G ong, T.R. Miles, a nd K. Ga ffney, Transfer molding encapsulation of  ip chip array packages, IMAPS Proceedings, pp. 760–766, 2000. 70. S. Han and K.K. Wang, Study on the p ressurized under ll encapsulation o f  ip chips, IEEE T ransactions o n Components, P ackaging, an d M anufacturing T echnology, Part B: Advanced Packaging, 20(4), 434–442, 1999. 71. L.P. Rector, S. Gong, and K. Gaffney, On the performance of epoxy mo lding co mpounds f or  ip ch ip t ransfer mol ding encapsulation, Proceedings of the 51st Ele ctronic Components and Technology Conference, pp. 293–297, 2001. 72. K.F. Becker, T. Braun, M. Koch, F. Ansorge, R. Aschenbrenner, and H. Reic hl, Advanced  ip ch ip e ncapsulation: t ransfer molding process for simultaneous underlling and postencapsulation, P roceedings o f the 1st I nternational IEEE Conference on Polymers and Adhesives in Microelectronics and Photonics, pp. 130–139, 2001. 73. S.H. Shi, T. Yamashita, and C.P. Wong, Development of the wafer-level c ompressive-ow underll p rocess a nd i ts required ma terials, P roceedings o f the 49th Ele ctronic Components and Technology Conference, p. 961, 1999.

43722_C015.indd 18

Smart Materials

74. S.H. S hi, T . Yamashita, a nd C.P . Wong, D evelopment o f the wa fer-level co mpressive-ow underll encapsulant, Proceedings o f the 5th I nternational S ymposium o n Advanced P ackaging M aterials: P rocesses, P roperties a nd Interfaces, p. 337, 1999. 75. K. G illeo a nd D. B lumel, Transforming Fli p c hip in to csp with re workable w afer-level u nderll, P roceedings o f the Pan Pacic Microelectronics Symposium, p. 159, 1999. 76. K. Gilleo, Flip Chip with Integrated Flux, Mask and Underll, W.O. Patent 99/56312, November 4, 1999. 77. J. Qi, P. Kulkarni, N. Yala, J. Danvir, M. Chason, R.W. Johnson, R. Zhao , L. Cra ne, M. K onarski, E. Yaeger, A. T orres, R. T ishkoff, a nd P. K rug, Assembly o f  ip c hips u tilizing wafer applied underll, Presented at IPC SMEMA Council APEX 2002, P roceedings o f APEX, Sa n Dieg o, CA, p p. S18-3 1–S18-3-7, 2002. 78. Q. Tong, B. Ma, E. Zhang, A. Savoca, L. Nguyen, C. Quentin, S. Lou, H. Li, L. Fan, and C.P. Wong, Recent advances on a wafer-level  ip chip packaging process, Proceedings of the 50th Electronic C omponents and Technology C onference, pp. 101–106, 2000. 79. S. Cha rles, M. K ropp, R . K inney, S. H ackett, R . Z enner, F.B. Li, R. Mader, P. Hogerton, A. Chaudhuri, F. Stepniak, and M. Walsh, P re-applied underf ill adhesi ves f or f lip chip a ttachment, IMAPS P roceedings, I nternational Symposium o n Micr oelectronics, B altimore, MD , p p. 178–183, 2001. 80. Z. Zhang, Y. Sun, L. Fan, and C.P. Wong, Study on B-st age properties o f wa fer le vel under ll, Journal of Adhesion Science and Technology, 18(3), 361–380, 2004. 81. Z. Zha ng, Y. S un, L. F an, R . D oraiswami, a nd C.P . Wong, D evelopment of w afer l evel u nderfill ma terial and p rocess, P roceedings o f 5th Ele ctronic P ackaging Technology C onference, S ingapore, p p. 194–198, December 2003. 82. R.L.D. Zenner and B. S. Carpenter, Wafer-applied under ll lm la minating, P roceedings o f the 8th I nternational Symposium on Advanced Packaging Materials, pp. 317–325, 2002. 83. R.V. B urress, M.A. C apote, Y.-J. L ee, H.A. L enos, a nd J .F. Zamora, A practical, ip-chip multi-layer pre-encapsulation technology for wafer-scale underll, Proceedings of the 51st Electronic C omponents a nd T echnology C onference, p p. 777–781, 2001. 84. Y. S un, Z. Zha ng, a nd C.P. Wong, P hoto-denable nanocomposite f or wa fer le vel pac kaging, P roceedings o f the 55th Electronic Components and Technology Conference, p. 179, 2005.

10/3/2008 8:50:24 PM

16 Dielectric Cure Monitoring of Polymers, Composites, and Coatings: Synthesis, Cure, Fabrication, and Aging 16.1 I ntroduction ............................................................................................................................16-1 16.2 I nstrumentation ......................................................................................................................16-2 16.3 Ā e ory .......................................................................................................................................16-2 16.4 Is othermal Cure ......................................................................................................................16-2 16.5 Monitoring the Cure Process in the Mold ..........................................................................16-4 Ā ick Laminates • Resin Infusion of Ā ree-Dimensionally Advanced Fiber Architecture Preforms

16.6 Automated Intelligent Closed Loop Control ......................................................................16-6 16.7 Rapid Cure by UV and Electron Beam Radiation .............................................................16-7 Monitoring Cure of Coatings • Monitoring Polyester Synthesis

David Kranbuehl College of William and Mary

16.8 L ife Monitoring .....................................................................................................................16-10 Acknowledgments ........................................................................................................................... 16-11 References ......................................................................................................................................... 16-11

16.1 Introduction Frequency-dependent electromagnetic sensors (FDEMS) dielectric measurements using in situ microsensors is a particularly useful technique for monitoring the changing state of a polymer during s ynthesis a nd c ure a nd a s a c omposite o r a s a c oating during f abrication a nd a ging i n t he app lication en vironment [1–5]. Measurements can be made i n t he laboratory a nd i n t he manufacturing en vironment to mo nitor t he p olymerization process and to monitor durability and aging in an environmental chamber or in the degradative environment. Dielectric in situ microsensor monitoring techniques can be used to monitor cure in production ovens and autoclaves on the plant  oor as well as in the  eld environment, for example, monitoring cure of coatings on the surface of a s hip in dry dock. Durability and aging can b e mo nitored w hile t he ob ject i s i n u se. E xamples a re a marine coating on a s hip, the protective coating on the liner of an ac id c ontaining t ank, a ro cket p ropellant, a n ad hesive i n a bond joint, or the polymer in a composite structure. Dielectric ( FDEMS) m icrosensing te chniques o ught to b e more widely used, particularly in monitoring cure and aging.

One reason the planar microsensor should be used more extensively is that it is ideally suited to monitoring changes in the chemical state at a pa rticular position or ply layer in a c omplex composite st ructure. Ā is co ndition is d iffic ult to duplicate in most rheological and calorimetric measurements. Furthermore, unlike some optical and calorimetric techniques, the sensitivity does not de crease when one t ries to de tect changes during t he late s tages o f c ure. Remote , p lanar, m icroelectric  eld sensors provide si mple, i n si tu, ac curate me asurements o f c hanges i n state throughout the fabrication process and with a high degree of sensitivity. A probable reason why FDEMS is not used more widely is that there is a l imited understanding by practitioners in industry of the physical principles upon which the instrumental technique is based. Often t here is a de sire to c ut short t he i mportance of understanding the science behind the frequency dependence of electric me asurements a nd to si mply i nterpret c hanges i n t he state of the instrument’s signal. Ā is simplied approach leads to misinterpretation and error. Ā e ke y to suc cessfully u sing ele ctric  eld s ensor m easurements to mo nitor changes in the state of a re sin system during 16-1

43722_C016.indd 1

10/3/2008 8:50:54 PM

16-2

cure a nd a ging i n t he u se en vironment, i s to u nderstand t he frequency dependence of t he complex permittivity e∗, a f orce– displacement parameter similar to t he complex modulus G∗ or compliance J∗ in rheological measurements. One might ask why doe one need in situ microsensors in addition to accurate composite models for fabrication. Modeling and individual thinking that lead to a procedure-driven cure cycle are beset w ith operating d ifficulties. Most not ably, a s has been described b y C iriscioli a nd S pringer [ 6], m odeling r equires extensive m aterial d ata c haracterization of re sin properties, a s well as fabric and tooling properties that are time consuming to measure and, most importantly, that will vary from day-to-day, batch, and lay up to lay up. Here the use of FDEMS microsensors to monitor the changes in processing properties in a variety of manufacturing situations is described. In one particularly complex example, t he buildup uses properties of an epoxy resin as a function of ply depth in the autoclave mold during cure of a 1 in. thick graphite composite as well a s u se of t his d ata to v erify a p rocess mo del i s de scribed. Monitoring the fabrication process during resin inltration and cure of an advanced ber a rchitecture composite a nd to v erify model predictions of that fabrication process is discussed. Use of the sensor to monitor rapid cure using UV or e-beam radiation will be described. Use of the sensor output to i ntelligently control t he fabrication process of a c omposite in conjunction w ith model predications will also be described. Ā en the application of FDEMS sensor to monitor ingress of degradative components, for example, into a c oating, a n ad hesive bond line, and a composite structure during use for the purpose of life monitoring and evaluating the time for replacement will be briey discussed.

16.2 Instrumentation Frequency-dependent complex dielectric measurements are made by u sing a n i mpedance a nalyzer c ontrolled by a m icrocomputer [1,2]. I n t he w ork d iscussed h ere, me asurements at f requencies from 5 to 5 × 106 Hz are taken continuously throughout the entire cure process at re gular intervals and converted to the complex permittivity, e∗ = e′ − ie″. Ā e measurements are made with a geometry-independent pl anar i nterdigitated microsensor. Ā is system is used with either a H ewlett Packard or a S chlumberger impedance bridge. Ā e system permits multiplexed measurement of several sensors. Ā e sensor itself is planar, 1 b y 0.5 in. in area and 3 mm thick. Ā is single sensor-bridge microcomputer assembly is able to ma ke continuous u ninterrupted measurements of both e′ and e″ over 10 decades in magnitude at a ll frequencies. Ā e sensor is inert and has been used at temperatures exceeding 400°C and pressures over 1000 psi.

16.3 Theory Frequency-dependent measurements of the dielectric impedance of a material as characterized by its equivalent capacitance, C, and conductance, G, a re u sed to c alculate t he c omplex p ermittivity,

43722_C016.indd 2

Smart Materials

e∗ = e′ − ie″, where i is the imaginary number, w = 2pf, f is the measurement frequency, and C0 is the equivalent air replacement capacitance of the sensor. C (w )material C0 G (w )material e ′′ (w ) = C0 e ′ (w ) =

(16.1)

Ā is calculation is possible when using the sensor whose geometry is invariant over all measurement conditions. Both the real and t he i maginary pa rts o f e * c an h ave d ipolar ( d) a nd io nic (i)-charge polarization components. e=d′ e+i′ e ′′ e d′′ = ε+i′′ (1 e′

6.2)

Plots of t he product of f requency (w) multiplied by t he i maginary component of the complex permittivity e″(w) make it relatively easy to visually determine the e″(w) permittivity when the low-frequency magnitude of e″ is dominated by the mobility of ions in the resin and when at h igher frequencies, the rotational mobility of b ound charge dominates e″. G enerally, t he magnitude of the low-frequency overlapping values of we″(w) can be used to me asure t he c hange w ith t ime o f t he io nic mob ility through the parameter s where s (ohm −1cm −1) = w 0 we i″ (w ) e 0 = 8.854 × 10 −14 C 2 J −1 cm −1

(1

6.3)

Ā e changing value of the ionic mobility is a molecular probe that can b e u sed to qu antitatively mo nitor t he v iscosity o f t he re sin during cure. Ā e dipolar component of the loss at higher frequencies can then be determined by subtracting the ionic component: e ″ (w )(dipolar) = e ″ (ω ) −

s (1 we 0

6.4)

Ā e peaks in e″ (dipolar) (which are usually close to the peaks in e″) can be used to de termine the time or point in the cure process when the mean dipolar relaxation time, t, has attained a specic value t = 1/w, where w = 2 pf is the frequency of measurement. Ā e dipolar mobility as measured by the mean relaxation time t can be used as a molecular probe of the buildup in Tg. Ā e time occurrence of a g iven dipolar relaxation time as measured by a peak in a particular high-frequency value of e″(w) can be quantitatively related to the attainment of a specic value of the resin’s Tg. Finally, t he tail of t he d ipolar relaxation peak a s monitored by the changing value of (de″/dt)/e″ can be used for in situ monitoring during processing the buildup in degree of cure and related en d-use p roperties suc h a s mo dulus, h ardness, e tc., during the nal stages of cure or postcure.

16.4 Isothermal Cure Ā e v ariation i n t he magnitude of e″ w ith f requency a nd w ith time f or t he d iglycidylether b isphenol-A a mine e poxy h eld at

10/3/2008 8:50:54 PM

Dielectric Cure Monitoring of Polymers, Composites, and Coatings: Synthesis, Cure, Fabrication, and Aging 10

150

Log e  w

1 MHz

6

500 kHz

120

90

250 kHz

50 kHz

4

25 kHz

60

5 kHz 500 Hz

2

Temperature (
C)

Temperature

8

30 50, 125, 250 Hz

0

0

80

160 240 Time (min)

320

0 400

FIGURE 16.1 Dielectric loss multiplied by frequency versus time over a range of frequencies during cure of an epoxide at 120°C.

121°C is shown in Figure 16.1. Ā e magnitude of e″ changes four orders o f m agnitude d uring t he c ourse o f t he p olymerization reaction. A plot of we″ is a particularly informative representation o f t he p olarization p rocess b ecause a s d iscussed f rom Equations 16.1 t hrough 16.4, o verlap o f we″(w) f requency f or differing f requencies i ndicates w hen a nd at w hat t ime t ranslational diffusion of charge, which can be used to monitor viscosity and reaction advancement, is the dominant physical process affecting the loss. Similarly, the peaks in we″(w) for individual frequencies indicate t he dipolar rotational diff usion relaxation time, a quantity which can be used to monitor Tg. Ā e f requency de pendence o f lo ss e″ is used to nd s by determining f rom a c omputer a nalysis, or a log –log plot of e″ versus frequency, the frequency region where we″ (w) is a constant. I n t his f requency re gion, t he v alue o f s i s de termined from Equation 16.3.

16-3

Figure 16.2 is a plot of log(s) versus log ( viscosity [h]) c onstructed from dielectric data and measurements on a dynamic rheometer. Ā e gure shows t hat at a v iscosity less t han 1 Pa s, (10 P), s is proportioned to 1/h because the slope of log(s) versus log(h) is approximately −1. Ā e gel p oint of the polymerization reaction occurs at 9 0 min based on t he crossover of G′ and G″ measured at 4 0 rad s −1. Ā is result is very close to the time at which h achieves 100 Pa s, which is also often associated with gel. Ā e region of gel marks the onset of a much more rapid change in viscosity than with s. Ā is change is undoubtedly because as gel o ccurs t he v iscoelastic p roperties o f t he re sin i nvolve t he cooperative motion of many chains while the translational diffusion o f t he io ns c ontinues to i nvolve mot ions o ver m uch smaller molecular dimensions. Ā e time of occurrence of a c haracteristic dipolar relaxation time c an b e de termined b y not ing t he t ime at w hich e″dipolar achieves a maximum for each of the frequencies measured where t = 1/2 p f at the time at which e″dipolar achieved a maximum for frequency f. Values of t can be measured over a range of frequencies and temperatures. Ā e time of occurrence of a given peak in e″ a nd t he c orresponding t is directly related to the time of achievement of a specic degree of reaction and the value of Tg at that point i n t ime. Ā is relationship between t he peak i n e″ at each frequency and degree of cure and Tg is made in the laboratory in c onjunction u sually w ith diff erential s canning c alorimetry (DSC) measurements versus time at the same temperature. Figure 16.3 i s a p lot s howing t he c orrelation o f s me asured with the electric  eld sensors versus degree of cure a as determined by DSC measurements at identical times during a 120°C cure. F igure 1 6.4 demo nstrates h ow t he c hange i n d e″/dt ca n detect t he  nal b uildup i n f ull c ure a s a app roaches 1.0 w ith considerably more precision than a DSC measurement. In summary, these in situ sensing measurements of e″, s de″/dt, and peaks in e″(w) can be used to mo nitor changes in essential cure processing properties, viscosity, degree of cure, and buildup in Tg, during fabrication at any point in the production mold.

1  10-7

s (ohm-1 cm)

1  10-8

1  10-9

1  10-10

1  10-11 0.01

1

100 h (Pa s)

10,000

FIGURE 16.2 Correlation of changes in the specic conductivity with the viscosity during cure of an epoxide.

43722_C016.indd 3

10/3/2008 8:50:56 PM

16-4

Smart Materials

10-7

s (ohm-1 cm)

10-8

10-9

10-10

10-11 0.00

0.20

0.40

0.60

0.80

1.00

a

FIGURE 16.3

Correlation of changes in the specic conductivity with the reaction advancement during cure of an epoxide.

1.0

0.4

0.6 0.2

a

Slope (de/dt) ε

0.8 0.3

0.4 0.1

0.2

0.0 0

100

200

300

0.0 400

Time (min)

FIGURE 16.4 an epoxide.

16.5

Correlation of changes in the rate of change of the dielectric loss with changes in the reaction advancement near the end of cure of

Monitoring the Cure Process in the Mold

16.5.1 Thick Laminates Figure 16.5 is a plot of the sensors output during a typical epoxy cure c ycle i n t he mold [ 7,8]. H ere, t he tem perature b oth l ags behind and overshoots the expected hold temperature due to the mass and heat transfer characteristics of a particular oven, autoclave, and mold. Yet, using the correlation relations developed in Figures 16.2 through 16.4 for a particular resin, the sensor output in Figure 16.5 can be used to monitor the actual state, changing viscosity, degree of c ure, a nd Tg at a particular position in the composite laminate. Figures 16.6 and 16.7 display the positions of sensors in a one inch thick 192 ply epoxy-photo graphite laminate.

43722_C016.indd 4

Figure 1 6.8 s hows t he s ensor mo nitored v iscosity. F rom this f igure, i t i s e vident h ow t he v iscosity i n t he c enter l ags behind t he v iscosity at t he surface until t he ramp approaching the nal hold temperature. At about 130 min, the viscosity at t he center ply catches up w ith t he outside ply and a m inimum similar viscosity is achieved at the same time uniformly throughout a ll 192 plys. This is a ke y f inal time for pressure consolidation o f t he 192 p ly l aminate. T he f act t hat a ll p lys achieve t he s ame v iscosity at t he s ame t ime i s u ndoubtedly one re ason w hy t his t ime–temperature c ure c ycle developed by trial and error is successful in creating a well-formed final product. Figure 16.9 shows how this sensor output can be used to verify model p redictions suc h a s f rom t he L oos–Springer p rocess model during fabrication on the plant oor [6–8].

10/3/2008 8:50:57 PM

16-5

Dielectric Cure Monitoring of Polymers, Composites, and Coatings: Synthesis, Cure, Fabrication, and Aging Bag

10

N-10 breather 9

Nonporous Teflon

Pressure plate Nonporous Teflon 8

Porous Teflon

6 32 ply 7 5 32 ply 6 4 32 ply Dielectric sensor 4 Dielectric sensor 3 Dielectric sensor 2 Dielectric sensor 1

5 3 32 ply 4 2 32 ply 3 1 32 ply

Porous Teflon

2

Nonporous Teflon Kapton Tool 1 Thick epoxy laminate autoclave run 192 PLY 3501−6 / AS4

FIGURE 16.5 Positions of the sensors in a thick 192 ply epoxy laminate.

Raw data for thick epoxy laminate autoclave run 200

8 Temperature

7

log e  w

6

1 MHz 500 kHz

5 4

FIGURE 16.6 Photograph of t he t hick 192 ply e poxy l aminate w ith the embedded sensors after cure.

16.5.2

Resin Infusion of Three-Dimensionally Advanced Fiber Architecture Preforms

Ā is is the second application where we have used electric eld sensors to verify model predictions and monitor the processing properties as a function of position. Figure 16.10 is a diagram of a bl ade s tiffened pa nel, w hich i s a p rototype f or a w ing. H ere resin i s p laced b elow t he pa nel. Ā en temperature is used to reduce the viscosity of the epoxy and a vacuum to draw the resin up into the panel. Sensors 2 through 9 were placed on top of the

43722_C016.indd 5

50 kHz 25 kHz

500 Hz 250 Hz 125 Hz 50 Hz 5 Hz

3 2

120

250 kHz

0

40

80

Temperature (
C)

160

5 kHz

40

80

120 160 Time (min)

200

240

Sensor output during cure of e   ω at the 64th ply of the composite

FIGURE 16.7 Dielectric loss multiplied by frequency versus time over a range of frequencies during cure of a 192 ply thick epoxy laminate at the 64th ply in an autoclave.

panel to mo nitor t he t ime at w hich t he v iscosity re ached t hat position and the resulting changes in viscosity. Ā e table below reports the wetout times at each sensor. Note sensor 1 was located at the bottom of the panel and its wetout time is the time when the resin began to ow. Sensors 3, 6, and 9 were at the top of the back stiffener and wetout last.

10/3/2008 8:50:57 PM

16-6

Smart Materials Comparison of FDEMS sensor and Loos–Springer model

192 ply laminate 3501–6 autoclave cure viscosity at the surface, 32nd, 64th, and center ply, determined

4

Determined from the frequency dependence of complex permittivity 200 5

3

Measured

150 Surface ply

3 100 2

Center ply

Temperature (
C)

Log viscosity (poise)

Log [h (Pa s)]

Temperature 4

Calculated 2

1

0

50 1

−1 0 240

0 0

60

120 Time (min)

180

0

40

80

120 Time (min)

160

200

240

Comparison of the viscosity at the 64th ply as predicted by the FDMS sensor and the Loos–Springer model for 3501-6

FIGURE 16.9 Comparison of t he s ensor me asured v iscosity at t he 64th ply versus the predictions of the Springer–Loos model.

FIGURE 16.8 Sensor measured v iscosity at e ach sensor ply location during cure in the autoclave.

9

2 0.81 (135 plies)

7 6

8

40

4

3 1 2

5

12 0.486 (81 plies)

FIGURE 16.10 Drawing of t he positions of t he sensors in the advanced  ber architecture blade stiffened panel. Ā e panel is impregnated with epoxy using vacuum assisted resin infusion from the bottom of the panel. 2 3 104 168

4 114

5 97

6 68

Figure 16.11 reports the viscosity of the resin underneath the panel at s ensor 1 a nd t he v iscosity o f p osition 3 a fter resin reaches this position. Further de tails o f mo del p redictions a nd t heir v erication during re sin i nfusion of t hree-dimensionally s titched c omplex ber preforms have been previously reported [9,10].

16.6 Automated Intelligent Closed Loop Control Using mo dels to p redict h ow t he c ure p rocessing p roperties should ideally change at each stage of fabrication, sensors can be

43722_C016.indd 6

105

7 8 9 110 100 163

200

Temperature

104

176 152

103 128

Position 1 102

101 140

Position 3

202

264 326 Time (min)

Temperature (
C)

1 63

h (Pa s)

Sensor location Wetout time (min)

104

388

80 450

FIGURE 1 6.11 Sensor me asured v iscosity of t he e poxide at t he bottom of the panel, sensor 1, and at position 3 after wetout.

10/3/2008 8:50:59 PM

Dielectric Cure Monitoring of Polymers, Composites, and Coatings: Synthesis, Cure, Fabrication, and Aging Processing tool

Frequency-dependent impedance measurement

Frequency/time molecular mobility models

Frequency/time molecular models data base

Impedance analyzer

Computer

Computer

Sensor

Heat pressure time

Molecular parameters: ionic mobility dipolar mobility

Impedance complex permittivity e and e

Processing parameters degree of cure viscosity Tg, modulus

Intelligent experience-based logic

Processing model

Computer

Computer

Controller

16-7

FIGURE 16.12 Schematic of the automated closed loop intelligent sensor controlled system.

used to monitor the fabrication process, verify when the appropriate s tage h as b een suc cessfully c ompleted suc h a s c omplete resin i nfusion o f t he g raphite p erform a nd t hen to ad just t he temperature and pressure for the next stage, not ba sed on time but rather the actual state of the resin–ber system [11–14]. Figure 16.12 i s a s chematic o f suc h a n i ntelligent c losed lo op system. Figure 16.13 displays the temperature and viscosity during an i ntelligent a utomated c ure o f t he 1 in. t hick e poxy g raphite panel de scribed e arlier. H ere f abrication r ules w ere de veloped based on mo del predictions a nd u ser e xpertise i n w hich t he a ir temperature increased and then would be held at 180°C until the viscosity at t he c enter p ly ac hieved t he v iscosity at t he o uter ply. Ā en the heat was turned off and the exotherm at the center ply increased the reaction advancement and viscosity fairly uniformly throughout the 192 plys. Twenty minutes later, after moving through the viscosity minimum under pressure, the heat was turned on to bring t he pa nel back to 1 80°C. Ā e heat was turned off after full cure was achieved as indicated by the sensors. Based on u ltrasonic C-scans a 1 in. t hick pa nel f ree of voids was produced in 180 min versus 240 min using the conventional cure cycle.

5 Temperature

3 120 2 80 1 Surface ply

0

40

Two center plies Air temperature

−1

0

40

80 120 Time (min)

160

200

FIGURE 16.13 Viscosity at v arious pl y p ositions of a t hick l aminate when cured using the closed loop intelligent rule based sensor controlled system.

43722_C016.indd 7

Rapid Cure by UV and Electron Beam Radiation

Over t he past decade, considerable at tention has been d irected to radiation-induced cure as it can be very rapid, more precisely controlled, a nd c an re duce m any p roblems a ssociated w ith thermal cure. Applications include rapid cure of coatings, lms, and p oints f or u ses suc h a s pa int o n a ntennas, r apid c ure o f repair patches on a composite, and improved control and speed in lament winding [15–18]. Electric  eld s ensors c an b e placed at a ny lo cation i n a pa rt and a re ideal for monitoring cure of a c oating when placed on the sub strates su rface a s t hey mo nitor c hanges i n s tate at t he bottom of the coating with one side o f the resin exposed to a ir and the other facing the substrates surface. Figure 16.14 di splays s ensor o utput d uring irr adiation o f a one second UV pulse and the subsequent rate of cure after the irradiation at t he 8t h s econd. App roximately n ine me asurements were made per second. Figure 16.15a and b d isplays the sensor’s a bility to demo nstrate t hat o ne e poxy f ormulation reaches f ull c ure i n 6 0 s w hile t he s econd w hile app earing to reach f ull c ure i n about 1 min, ac tually t akes 8 min to ac hieve full cure.

16.7.1

160 Temperature (
C)

Log [h (Pa s)]

4

200

16.7

Monitoring Cure of Coatings

In more traditional coating-cure situations, FDEMS output similar to Figure 16.16 can be used to monitor the variation in cure rate with temperature, humidity, airow, pigment loading, catalyst concentration, thickness, age, batch, etc. [15–18]. For example, it was thought epoxy–polyamide coatings would cure at temperatures as low as 10°C (50°F), even though the time required to reach a given stage is at least double that required at 22°C [4]. To examine t he e ffects of decreased temperature, t he polymerization p rocess w as mo nitored i n a n en vironmental chamber, h eld at 1 1°C a nd 3 0% R H f or 3 w eeks. F igure 16.17 shows a plot of lo g(e″ ∗ 2pf ) measured with the planar FDEMS sensor. Ā e t ime to s et-to-touch w as 4 20 min. D ry-to-hard occurred at 695 min. Ā e v alue o f log (e ° ∗ 2pf ) at s et-to-touch was 4.1 and at dry-to-hard was 2.9 compared to 4.2 and 2.8 in the

10/3/2008 8:51:00 PM

16-8

Smart Materials Monitoring effect on cure of a UV Pulse over a 10 s interval 6.5 6.25

Log (e   w)

6 5.75 5.5 5.25 5 4.75 4.5 5

6

7

8

9

10 Time (s)

11

12

13

14

15

FIGURE 16.14 Dielectric loss multiplied by frequency during rapid cure of a coating after a 1s exposure of UV radiation.

Comparing cure of three different epoxy formulations

Comparing cure of three different epoxy formulations Epoxy 2, 10 kHz

1300

1100

1275

1075 e   w

e   w

Epoxy 3, 10 kHz

1050

1250

1025

1225 1200

(a)

1000 0

2

4

6 Time (min)

8

10

0

12

(b)

2

4

6 Time (min)

8

10

12

FIGURE 16.15 Dielectric loss multiplied by frequency of two different coating formulations to determine which achieves full cure most rapidly after a one second UV dose.

24°C m easurement. Ā e d rop i n e″ ∗ 2pf is much slower than at 24°C. Ā e sample continued to c ure as seen by the continual drop in e″ e ven during t he t hird week si nce application. A fter 500 h at 11°C, the 50 kHz line still has not reached the degree of polymerization t hat t he same coating does i n 48 h at 2 4°C a nd the rate of cure as monitored by the change in e″ is relatively at, indicating a very slow cure rate at 11°C. Ā ese F DEMS re sults c learly i ndicate t hat b elow 1 0°C, t he curing is greatly retarded and full cure is not achieved. More important t his e poxy-polyamide w ill not re ach f ull c ure u ntil the temperature rises. A lthough t he pa rtly cured  lm may feel dry, poor resistance to abrasion, moisture, and chemicals result. It i s i nteresting to demo nstrate t hat t he s ensor c an monitor the e xtent to w hich t he  rst c oating o f t he S eaguard e poxy polyamide i s s oftened by t he s econd for v arying el apsed t imes since t he i nitial c oat’s app lication. F igure 1 6.18 d isplays t he softening o f t he  rst c oating b y t he s econd w hen t he s econd

43722_C016.indd 8

coating h as b een applied 144 min a fter t he  rst for t he e poxy– polyamide s ystem s hown i n F igure 16.16. Ā e i nitial c oating’s values o f l og(e″w) h ave o nly d ropped f rom 5. 5 to 4 .5. U pon application o f t he s econd c oating, t he v alues r ise to 6 .0 a fter 60 min as the second coat resoftens the rst layer. Next, the FDEMS sensor is used to examine the extent of softening of the  rst coat if 26 h elapse before the second coating is applied. F igure 1 6.19 s hows t hat a fter 2 6 h, t he  rst c oat h as decreased its low-frequency values of log (e″w) to around 2. Ā e application of the second coat on the next day resoftens the rst coat and log (e″w) returns back to 5.5. Hence, 24 h has little effect on the resoftening ability of the rst coat by a second coat for this paint system. Ā e ability of the FDEMS sensor to monitor cure in a proprietary latex coating is shown in Figure 16.20. Initially, there is a rather rapid decay in the log(e″w) overlapping ionic lines from 10 to 8.5 over 30 min. Ā is is followed by a rapid drop of over two

10/3/2008 8:51:01 PM

Dielectric Cure Monitoring of Polymers, Composites, and Coatings: Synthesis, Cure, Fabrication, and Aging A B

6.3

45 D

5.6

40 E

4.9 log (e   w)

50

1 MHz

C

F

35

5 kHz

4.2

30 Temperature

3.5

25 50 Hz

2.8 2.1

20 15

5 Hz

1.4

10

0.7

5

0

0

72

144

216

432 288 360 Time (min)

504

576

648

Temperature (
C)

7

16-9

720

7

50

6.3

45

5.6

40

4.9

35

4.2

30

3.5

25

2.8

20

2.1

15

1.4

10

0.7

5

0

0

84

168

252

336

420

504

588

672

756

Temperature (
C)

log (e   w)

FIGURE 16.16 Decrease in the log of the product of e″ times the frequency for a marine coating during cure at room temperature. Measurements of e″ were made at over a range of frequencies from 5 Hz to 1 MHz. Ā e series of peaks in e″ at the various frequencies are each associated with the attainment of a particular value of Tg. Ā e overlapping values of e″ times frequency values can be used to calculate s.

840

Time (min)

10

50

9

45

8

40

7

35

6

30

5

Temperature

25

4

20

3

15

2

10

1

5

0

0

36

72

108

144

180 Time (min)

216

252

288

324

Temperature (
C)

log (e   w)

FIGURE 16.17 Cure of the same coating at 11°C.

0 360

FIGURE 16.18 Similar measurement of coating cure where a second coat is applied at 144 min into cure.

43722_C016.indd 9

10/3/2008 8:51:01 PM

16-10 10

50

9

45

8

40

7

35

6

30

5

25

4

20

3

15

2

10

1

5

0 24 126 Hours +

252

378

504

630

756

882

1008

1134

Temperature (
C)

Log (e   w)

Smart Materials

0 1260

Time (min)

FIGURE 16.19 Similar measurements of coating cure where second coat is applied after 26 h.

decades from 8.5 to 6.2 in several more minutes. One can ascribe the initial 30 min to phase I packing of the latex spheres. Ā ere is a loss of volatiles to the point where the latex spheres touch. After this point, water is no longer the continuous medium, the latex spheres r ather th an th e s uspending  uid a re t he c onducting medium. Ā is transition point generated the large drop in e″w. Ā e rapid drop is then followed by a gradual decrease in log(e″w) which is ascribed to phase two deformation of the latex spheres. Ā ere app ears to b e a nother s hift i n t he s lope a round 72 min. Ā is shift in the rate of the drop of log(e″w) at 72 min is probably due to the transitions from what is described as a predominately phase I I de formation a nd pac king c ure p rocess to a ph ase I II autohesion and diff usion cure process.

16.7.2 Monitoring Polyester Synthesis Ā e polyester synthesized for this application is used as a precursor for a number of coating formulations. It is a condensation reaction 10

30

9

29

Phase I

8

Log (e   w)

7

1 MHz

6

27 26

ll

5

25 lll

4

24 50 Hz

3

Temperature

23

2

22

1

21

0

0

36

72

Temperature (
C)

28

108 144 180 216 252 288 324 360 Time (min)

FIGURE 16.20 Use of log(e″ × w) to monitor cure of a latex coating at room temperature.

43722_C016.indd 10

between two diols, neopentyl glycol and trimethylopropane, with two d iacids, ad ipic ac id a nd i sopthalic ac id. A d rying a gent w as added to remove residual water. Ā e reaction was carried out in the laboratory u sing a 3 L,  uted, round b ottom  ask. Ā e ask was outtted with a packed column condenser, an air-driven stir bar, a FDEMS sensor and a glass stopper on the ask’s ute for addition of i ngredients. Ā e s ame F DEMS s ensor w as l ater at tached to a steel pipe for insertion into the pilot plant’s reactor for monitoring production batches based on t he calibration of t he sensor output with extent of reaction from the laboratory experiments. Ā ere are two issues of industrial signicance which are usually mo nitored i n a p olyester s ynthesis: ac id n umber a nd t he refractive index of the distillate. Ā e acid numbers of the samples taken during t he s ynthesis a re ger mane b ecause t hey e stablish the criteria for the termination of the reaction. In an industrial setting, w hen a s ample h as a n ac id number of a c ertain v alue, then t he ne xt s tage of t he s ynthesis would b e u ndertaken. Ā e problem w ith t his te chnique i s i n t he t iming. C onsidering t he reaction is continuous and the acid number titration technique takes about 1 h to perform, any acid number data would be available at le ast a n h our b ehind t he ac tual re action. Ā e se timing problems are considerable. Ā ere is a need to establish an in situ technique for the determination of end point criteria. To accomplish this, samples were removed from the reactor and their acid numbers determined. Ā en, this acid number was correlated with the FDEMS sensor data at t he time of sampling. Ā e correlation of the absolute, online measurement of the ionic mobility, lo g(e″ ∗ w), a gainst ac id n umber i s s hown i n F igure 1 6.21. Figure 16.21 shows t hat t here i s a go od c orrelation of t he s ensor output with acid number. Ā e data were normalized by taking the highest log ( e″ ∗ w) v alue o f t he s ynthesis, c alled t he p eak, a nd dividing the changing log (e″ ∗ w) values by the value at the peak.

16.8

Life Monitoring

Aging, de gradation, a nd re duction i n p erformance p roperties during use in the  eld is another application [19]. Here sensors

10/3/2008 8:51:02 PM

Dielectric Cure Monitoring of Polymers, Composites, and Coatings: Synthesis, Cure, Fabrication, and Aging

changes occur much more slowly and are generally much smaller than during fabrication. Ā us, monitoring aging requires more precision a nd sensitivity t han in monitoring t he state of a pa rt during fabrication.

6.5 Run 1 Run 2 Log (e   w)

6

Acknowledgments

5.5

5

4.5

0

2

4

6

8 10 Acid number

12

14

16

FIGURE 16.21 Monitoring the polymerization process of a polyester in t he re action ve ssel by t he c orrelation of c hange i n e″ × w with the change in the acid number.

can b e emb edded i n w itness c oupons e xposed a long side t he actual composite structure to the same uctuations in the environment during use. Sensors can a lso be embedded a nd left in the structure itself during fabrication [20–22]. Figure 16.22 displays a c ommercially a vailable s ensor s ystem u sed to mo nitor aging of the polymer liner in a exible pipe used to transport oilgas from the ocean oor to a oating platform. Degradation ge nerally i nvolves c hemical pro cesses w hich produce changes at t he molecular level and which in one sense are the reverse of what happens during cure. Ā us, electric eld sensors emb edded i n t he m aterial c an a lso de tect a ging a nd changes in performance properties in situ during use. As in fabrication, the sensor output needs to b e calibrated with changes in mechanical properties or material properties such as degradation due to UV radiation, high temperature, atomic oxygen, or penetration of fuel oil, acid, or water. During aging, these

Instrumented polymer aging coupon system Service life monitoring for flexible pipes using FDEMS

FIGURE 1 6.22 Commercially av ailable e lectric  eld s ensor s ystem for monitoring aging of performance properties subsea of the polymer lining in pipes used to transport crude oil to oating platforms.

43722_C016.indd 11

16-11

Ā e a uthor t hanks n umerous u ndergraduate a nd g raduate students a nd P rof. A lfred C . L oos. Supp ort f rom N ASA Langley, National S cience A dministration, t he A ir F orce Edwards C alifornia L aboratory, B oeing-McDonnell D ouglas Aircraft C ompany, a nd ICI’s C oatings C ompanies i s g ratefully appreciated.

References 1. D. Kranbuehl, Dielectric monitoring of polymerization and cure, invited chapter in Dielectric Spectroscopy of Polymeric Materials, J. Runt and J. Fitzgerald, eds. American Chemical Society, Washington, DC, pp. 303–328, 1997. 2. D . K ranbuehl, I n-situ o nline me asurement o f co mposite cure wi th fr equency d ependent e lectromagnetic se nsors, Plastics, R ubber, a nd C omposite Pr ocessing, 16, 213–219, 1991. 3. S. Poncet, G. Boiteaux, H. Sautereau, G. Seytre, J. Rogozinski, and D. Kranbuehl, Monitoring phase separation and reaction ad vancement in si tu in thermo plastic/epoxy blends, Polymer, 40, 6811–6820, 1999. 4. J.M. B rown, S. S rinivasan, T. Ward, J. McGrath, A.C. L oos, D. H ood, an d D . Kr anbuehl, Pro duction of c ontrolled networks and mo rphologies in to ughened t hermosetting resins using real time in situ curve monitoring, Polymer, 37, 1691–1696, 1996. 5. D. Kranbuehl, In situ frequency dependent dielectric sensing of cure, Processing of Composites, Hansa, Munich, 137–157, 2000. 6. P. Ciris cioli a nd G. S pringer, Smart Autoclave Cur e of Composite, Technomic Publishing Co., 1990. 7. D. K ranbuehl, F requency dep endent diele ctric measur ement sensing: Monitoring in situ the changing state of polymeric ma terials d uring p rocessing a nd us e in the  eld, invited cha pter in Encyclopedia of Smar t M aterials, J ohn Wiley, New York, pp. 456–470, 2002. 8. D. Kranbuehl, In situ frequency dependent dielectric sensing of cure, Processing of Composites, Hansa, Munich, 137–157, 2000. 9. D. Kranbuehl and A. Loos, Monitoring resin position, reaction advancement and processing properties, in Resin Transfer Molding for A erospace A pplications, Cha pman & H all, London, pp. 412–432, 1999. 10. G. Hasko, H. Dexter, A. Loos, and D. Kranbuehl, Application of science based RTM for fabricating primary aircraft structural elements, Journal of Advanced Materials, 26, 9–15, 1994. 11. S. H art, D . K ranbuehl, D . H ood, A. L oos, J . K oury, a nd J. Harvey, FDEMS sensor-loss model QPALS intelligent cure processing of polyimides, International SAMPE Symposium Proceedings, 39(1), 1641–1651, 1994.

10/3/2008 8:51:03 PM

16-12

12. D. K ranbuehl, D . H ood, J . Rog ozinski, W. Lim burg, A. L oos, J . M acRae, a nd V. H ammond, I n si tu FD EMS sensing f or in telligent a utomated c ure in r esin t ransfer molding, International SAMP E S ymposium Pr oceedings, 40(1), 1995. 13. D. K ranbuehl, D. Hood, J. Rog ozinski, R . B arksdale, A. L oos, and D. MacRae, FDEMS sensing for automated intelligent processing o f p olyimides, Proceedings of ASME I nternational Mechanical Engineering Congress H1041A, 2, 1017–1030, 1995. 14. B.J. L epene, T .E. L ong, A. M eyer, a nd D . K ranbuehl, Moisture-curing kinet ics o f is ocyanate p repolymer ad hesives, Journal of Adhesion, 78(4), 297–312, 2002. 15. D. Kranbuehl and A. Domanski, In situ monitoring of coating polymerization, cure and aging using frequency dependent diele ctric monitoring Journal of C oatings Technology, 48–55, June 2004. 16. D. K ranbuehl, D . H ood, J . Rog ozinski, A. M eyer, a nd M. Neag, Monitoring the changing state of a polymeric resin coating during synthesis, cure and use, in Progress in Organic Coatings, Elsevier, Amsterdam, pp. 881–888, 1999. 17. D. Kranbuehl, D. Hood, C. Kellam, and J. Yang, In situ sensing for mon itoring mol ecular and ph ysical prop erty change s during  lm f ormation, in Film F ormation in Waterborne Coatings, e d., T . Pr ouder, American C hemical S ociety Symposium Series, Washington, DC, pp. 648, 96–117, 1996.

43722_C016.indd 12

Smart Materials

18. D. Kranbuehl, J. Rogozinski, A. Meyer, L. Hoipkemeier, and N. N ikolic, On-line in si tu s ensor mo nitoring o f ra pidly curing UV  lms in  lm formation in co atings, T. Provder and M. U rban, e ds., American Chemical S ociety, Washington, DC, pp. 141–156, 2001. 19. D. Kranbuehl, In situ monitoring of coating polymerization, Cure a nd Aging Coa tings T echnology, 1(6), J une 2004. 20. D. K ranbuehl, W. N ewby-Mahler, D . H ood, S. C use, K. Reifsnider, and A. Loos, Use of in si tu dielectric sensing for in telligent p rocessing a nd he alth mo nitoring, in Structural H ealth M onitoring, F.K. C hang, e d., Technomic, Penn, pp. 33–43, 1997. 21. D. K ranbuehl, D . H ood, L. M cCullough, M. Eriks en, a nd E. Hov, F requency dep endent ele ctromagnetic s ensing f or quality co ntrol a nd lif e mo nitoring o f p olymers, P roceedings of IC C C onference on I nspection of St ructural Composites, 1994. 22. D. Kranbuehl, D. Hood, J. Rogozinski, A. Meyer, E. Powell, C. Higgins, C. D avis, L. Hoipkemeier, C. Ambler, C. Elk o, and N. Olukeu, Monitoring and mo deling of durability of polymers us ed for c omposite off shore pipe, Recent Developments in Durability Analysis of Composites, 413–420, 2000.

10/3/2008 8:51:04 PM

17 Magnetorheological Fluids 17.1 Introduction to Magnetorheological Fluids ....................................................................... 17-1 17.2 MR Fluid History ................................................................................................................... 17-1 17.3 MR Fluid Composition ......................................................................................................... 17-3 17.4 17.5

J. David Carlson Lord Corporation

17.1

Polarizable Particles • L iquid Vehicle • Ad ditives

Properties of Typical MR Fluids ..........................................................................................17-4 Basic Operating Modes ......................................................................................................... 17-5 Valve Mode • Direct Shear Mode

17.6 MR Fluid Applications ..........................................................................................................17-6 References ........................................................................................................................................... 17-7

Introduction to Magnetorheological Fluids

Magnetorheological (MR)  uids are suspensions of magnetically responsive pa rticles i n a l iquid c arrier. Ā e e ssential feature of MR uids is their ability to c hange from a f reely owing liquid into a s emisolid when exposed to a n external magnetic eld as shown i n F igure 1 7.1. M ost c ommonly, M R  uids c onsist o f micron-sized, nearly pure elemental iron particles suspended in a h ydrocarbon-based o il. W hen a n ex ternal ma gnetic  eld is applied to M R  uid, t he  uid de velops a y ield s trength t hat i s more or less proportional to the strength of the applied magnetic eld. Ā is yield strength arises from the internal structure that is formed w hen t he i ron pa rticles p olarize a nd c hain to gether along the magnetic eld lines. Ā is magnetic eld-induced yield strength adds linearly to the off-state shear stress due to the uid viscosity such that the slope of the stress versus shear rate curve, i.e., plastic viscosity, remains largely unchanged. A simple, Bingham plastic model is effective at describing the basic  eld-dependent c haracteristics o f M R  uid [ 1]. I n t his model, the total yield stress ttotal is given by τ total = τ (H ) + ηpγ


(17.1)

where t(H) is the yield strength caused by the applied magnetic eld H . g is the shear rate hp is the magnetic eld-independent plastic viscosity dened as t he s lope o f t he s hear s tress v ersus s hear s train r ate relationship

Ā e b ehavior of M R  uid i s su mmarized b y t he g raph o f ttotal . versus g shown in Figure 17.2. MR fluids belong to a class of “smart” fluids that also include electrorheological (ER) fluids. Unlike ER fluids that respond to strong electric fields, MR fluids do not require the application high voltage. As a re sult, they are able to ac hieve very h igh y ield strengths w ithout encountering t he problem of dielectric breakdown. As a consequence, MR fluids are able to routinely reach magnetic f ield-induced yield strengths of 50 kPa or more with a magnetic field that can easily be generated by an electromagnet operated at low voltage and modest current. MR fluids should not be confused with colloidal ferrofluids i n w hich t he pa rticles a re about 1000 t imes smaller than t hose f ound i n t ypical M R f luids. W hile f errofluids experience a body force in a magnetic field gradient, they are incapable o f de veloping t he l arge y ield s trength f ound w ith MR fluids.

17.2

MR Fluid History

MR and ER uids both have histories that date from the 1940s. Jacob R abinow at t he U .S. N ational Bu reau o f St andards (now the National Institute of Standards and Technology) was responsible for t he i nitial d iscovery a nd e arly de velopment o f MR  uids [2,3]. He applied for his  rst MR  uid patent in 1947 [4]. MR uids made by Rabinow for use in clutches were not dissimilar to many of those made today and exhibited comparable yield strength. Rabinow’s pioneering work led to a brief period of interest in MR uids in the 1950s and early 1960s. However, by the late 1960s, interest in MR uids had waned and a g rowing interest in ER  uids b egan to emerge . By t he late 1980s, a 17-1

43722_C017.indd 1

10/3/2008 8:51:22 PM

17-2

Smart Materials

FIGURE 17.1 Demonstration of the operation of an MR uid. On the left, the magnetic eld is off. On the right, an electromagnet is energized to generate a magnetic eld between the two steel prongs of the MR probe.

large number of universities and companies were actively pursuing research on ER uids and devices [5]. Much of this work was mot ivated b y p otential a utomotive app lications suc h a s

H2

Yield strength (t )

hp t (H2) H1 Increasing magnetic field t (H1)

H=0

•

Shear rate (g )

FIGURE 17.2 Behavior of an idealized, Bingham model, MR uid in the presence of an applied magnetic eld H as a function of shear rate.

43722_C017.indd 2

real-time c ontrollable “ smart” s hock a bsorbers f or i mproved ride and handling and controllable torque transfer devices [6]. In spite of great improvements in ER uid f ormulations, E R uid f ailed to emerge i n a ny p ractical w ay. M aximum y ield strengths o f ER  uids rem ain to o lo w f or mo st app lications. Further, they require power supplies capable of many kilovolts as well a s e xpensive w ires a nd c onnectors r ated for suc h h igh voltages. Ā e inherent temperature dependence of their electrical polarization me chanisms a nd a s trong s ensitivity to mo isture and contamination makes their use outside of the laboratory difficult and costly. Beginning i n t he early 1990s, a re surgence of i nterest i n M R uids and applications emerged in response to a number of practical l imitations en countered w ith E R  uids [7,8]. MR uids offered substantially higher yield strength plus the ability to operate at h igher a nd lo wer tem peratures. M ost i mportantly, h igh voltages w ere not re quired to p rovide t he ne cessary m agnetic elds required to activate MR uids. Common, low-voltage power supplies and batteries could directly power the electromagnets in MR uid devices. Ā is was coupled with uid developments that led to h igh-durability M R  uids a nd t he adv ent o f mo dern,

10/3/2008 8:51:22 PM

17-3

Magnetorheological Fluids

microprocessor-based control systems. Ā e ultimate dream of a mass-produced “ smart” M R  uid a utomotive s hock a bsorber system was nally realized in early 2002 with the introduction of a real-time controlled suspension system as standard equipment on a Cadillac automobile [9].

17.3

MR Fluid Composition

Ā e composition of a typical MR uid can be broken into three parts: magnetically polarizable particles, a liquid vehicle, and an assortment of additives. While it is possible to create a suspension that will display a MR response with only a liquid and magnetically pa rticles, suc h su spensions a re h ighly u nstable and do not p erform s atisfactorily. A w ell-formulated add itive package i s gener ally ne cessary to ac hieve a s table, lo ng-lived MR uid. Much of the MR uid research activity in recent years has focused on de veloping proprietary add itive pac kages t hat have ena bled t he si multaneous re alization o f h igh o n-state force, low off-state forces and very long life in highly dynamic applications [10].

17.3.1 Polarizable Particles Ā e magnetically polarizable particles in an MR uid are almost always nearly pure elemental iron although a ny magnetically polarizable particle could be used to make a functioning, albeit usually weaker, MR  uid. Powdered iron, water- or gas-atomized i ron, n ickel a lloys, i ron/cobalt a lloys, m agnetic stainless s teels, a nd f errites a re a ll p ossible. Ā e maximum yield strength of MR uid scales approximately as the square of the s aturation m agnetization Js o f t he pa rticles [ 11]. I n t his regard, the best particles one could use would be alloys of iron and cobalt such as Permendur with a saturation magnetization of a bout 2 .4 T [ 12,13]. U nfortunately, t he v ery h igh c ost o f cobalt, w hich ac counts f or a bout 5 0% o f t hese a lloys, m akes them prohibitively expensive for most applications. By far, the most cost -effective M R  uid pa rticle i s p ure elemen tal i ron with a s aturation m agnetization o f 2 .1 T. Vi rtually a ll ot her metals, alloys, and oxides have saturation magnetizations signicantly lo wer t han t hat o f i ron, re sulting i n sub stantially weaker MR uids. Ā e most widely used material for MR uid particles is carbonyl iron. Carbonyl iron is the common name given to iron particles f ormed f rom t he t hermal de composition o f i ron pentacarbonyl. Key physical properties of carbonyl iron powder a re t he very spherical shape of t he pa rticles a nd t he fine p article si ze i n t he 1–10 µm r ange. O ther f orms o f ele mental i ron p owder, suc h a s w ater ato mized o r ele ctrolytic, are also possible. Typical iron particle volume fractions range between 2 0% a nd 4 5%. C hoosing t he ide al vol ume f raction for a pa rticular app lication i s a lways a t rade-off. W hile t he maximum on-state yield strength scales more or less proportionate to t he iron particle volume fraction, the off-state viscosity also increases at an even faster rate with increasing particle content.

43722_C017.indd 3

17.3.2 Liquid Vehicle Ā e most common liquid vehicles for MR uids are hydrocarbon oils. Ā ese c an b e m ineral o ils, s ynthetic o ils, o r m ixtures o f both. I n gener al, h ydrocarbon o ils a re t he v ehicle o f c hoice because of t heir good lubricity, durability, a nd availability of a wide range of well-behaved and well understood additives. Synthetic poly(alpha-olens) (PAOs) are often chosen because of their well-controlled properties over a b road range of temperatures. Other liquid vehicles for special-purpose MR uids are silicone oils and water. While generally inferior to hydrocarbon oils i n ter ms o f l ubricity, d urability, a nd a menability to m any additives, silicone oils do offer a somewhat broader temperature range w ith le ss v ariation i n v iscosity t han s ynthetic h ydrocarbons. Silicone oils are often ch osen i n si tuations w here a hydrocarbon vehicle may not be compatible with other materials exposed to t he MR  uid such as r ubber seals a nd d iaphragms. Care must be taken when using silicone oils to avoid conditions that might promote cross-linking of the oil, leading to viscosity increase and gum formation. Water-based MR uids offer the highest on-state yield strength and lowest off-state viscosity for a given particulate loading of any MR uid. However, the high vapor pressure of water means that signicant e vaporative l iquid lo ss m ust b e c onsidered. Water i s only used as a vehicle for MR uid in situations where evaporation is not a concern. Ā is generally means that water-based MR uid is o nly u sed i n s ystems t hat a re tot ally a nd a bsolutely s ealed. Water-based  uids a re u sually not app ropriate for s ystems c ontaining a dynamic, sliding seal such as a shock absorber since the lm of water invariably left on the surface of the exposed shaft will evaporate and lead to a progressive loss of uid.

17.3.3 Additives A wide variety of proprietary additives similar to those found in commercial lubricants are formulated into MR uids. Such additives a re used to d iscourage sedimentation prevent agglomeration, enhance lubricity, prevent oxidation, modify viscosity, and inhibit wear. Unlike colloidal ferrouids, the relatively large particle size and large difference between the specic gravity of the magnetic particles and the carrier liquid can cause rapid settling in a M R  uid [14]. Sedimentation is t ypically controlled by t he use of organic or i norganic t hixotropic agents a nd su rfactants. Ā e t hixotropic ne tworks t hat f orm i mpart a sm all, z ero-eld yield strength to the MR uid that inhibits ow at ultralow shear rates. Such thixotropic networks are very weak so that as the uid is purposely sheared at higher rates, the network collapses, allowing t he  uid to s hear t hin to a de sirable lo w off -state dynamic viscosity. Except f or v ery s pecial c ases, suc h a s d ampers de signed f or seismic damage mitigation in civil engineering structures, lack of c omplete s uspension s tability i s not a ne cessity. M R  uid devices suc h a s d ampers a nd s hock a bsorbers t hat a re u sed i n very dynamic applications are efficient mixing devices. As long as t he M R  uid do es not s ettle i nto a h ard s ediment, nor mal

10/3/2008 8:51:24 PM

17-4

Smart Materials

motion in these devices is usually adequate to remix any stratied MR  uid back to a h omogeneous condition [10]. In a w elldesigned MR  uid, damper remixing occurs within one or two strokes. While the MR  uid in these dampers will stratify over time, t he pa rticle-rich s ediment rem ains s oft a nd e asily remixed.

17.4 Properties of Typical MR Fluids While there are a number of “standard” MR uids commercially available today, MR uids for signicant applications are generally tailored to meet the specic requirements of the application. Ā e liquid type and viscosity, particle details, and volume fraction and additive package will all be chosen to optimize the MR uid f or t he s pecic c onditions of t he appl ication. More t han 99% of all commercial MR uids are based on synthetic hydrocarbon oils. Ā e most important property is the magnetic eld dependent yield strength. Ā is yield strength adds to the viscous-dependent stress at any given shear rate or speed. Maximum yield strengths are t ypically i n t he r ange o f 25 –100 kPa a nd de pend mo st strongly on the volume fraction of iron particles in the uid. For magnetic  elds b elow a bout 100 kA/m, t he  eld t he de veloped yield strength is nearly linearly proportional to magnetic eld. Above 1 00 kA/m, t he i ncrease i n y ield s trength p rogressively rolls-off and eventually completely saturates at a eld strength of about 400–500 kA/m. Ā e B–H curve for MR  uids throughout their useful range is nonlinear a nd of such a m agnitude a s to p lace t hem i nto a very u nique c ategory o f ma terials i ntermediate b etween l owsusceptibility m aterials l ike a luminum a nd f errous m aterials like steel. MR uids have initial (very low eld) relative magnetic

permeability in the range of 4–8. As the MR uid yield strength begins to saturate at higher magnetic elds so also the slope of the B–H curves roll-off a nd approaches a s lope of 1 at  elds in the r ange o f 4 00–500 kA/m. Ā e m aximum m agnetic  eld strength t hat i s u seful i n a p ractical s ense i s u sually a bout 250 kA/m. Ā e off-state property most critical for dynamic mechanical applications i s t he  eld-independent p lastic v iscosity hp. Ā is viscosity c reates s hear r ate o r v elocity-dependent f orces a nd torques t hat a re a lways p resent i n a M R de vice re gardless o f whether a magnetic eld i s applied or not. Ā is viscosity is also responsible for most of the temperature dependence observed in the force output of a device. Ā e magnitude of hp is controlled by the v iscosity of the liquid vehicle and the particle volume fraction. Plastic viscosities at room temperature typically range from 50 to 200 mPa s although much higher values are easily possible with higher viscosity carrier liquids. Ā e viscous off-state force in M R de vice c an b ecome qu ite large at t he shear r ates (104 to > 105 s−1) routinely encountered in MR shock absorbers, clutches, and b rakes. Gr aphs o f t he o n-state y ield s trength v ersus m agnetic eld and B–H curves for typical oil-based MR uids having a variety of iron particle loadings are given in Figures 17.3 and 17.4. Table 17.1 lists the key physical properties for several commercial, M R  uids a vailable f rom L ord C orporation [ 15–17]. More details about MR uid properties are discussed in the reference by Black and Carlson [14]. Ā e f ollowing pa ir o f em pirical e quations i s u seful f or describing the salient properties of most MR uids [18]. Ā e rst equation gives the magnitude of the induced yield strength t(H) as a f unction of t he applied magnetic  eld H while t he second gives t he relationship between magnetic  ux density B and t he applied magnetic eld intensity H in the MR:

70 40%

60

Yield stress (kPa)

50 32%

40 30

22%

20 10 0 0

100 200 300 Applied magnetic field strength, H (kA/m)

400

FIGURE 17.3 Typical yield stress versus magnetic eld curves for several oil-based MR uids having iron particle volume fractions ranging from 22% to 40% measured in parallel plate mode at a maximum shear rate of 262 s−1.

43722_C017.indd 4

10/3/2008 8:51:24 PM

17-5

Magnetorheological Fluids

m0 is the magnetic constant equal to 4π × 10−7 Henry/m constant C equals 1.0, 1.16, or 0.95, depending on whether the carrier uid is hydrocarbon oil, water, or silicone oil

1.20 40% Magnetic flux density, B (T)

1.00 32%

Ā ese equations have been developed to provide a convenient and p ractical de scription o f t he rh eological a nd m agnetic properties of virtually any MR uid.

0.80 22% 0.60

17.5 Basic Operating Modes

0.40

0.20

0.00 200 100 300 Magnetic field strength, H (kA/m)

0

400

FIGURE 1 7.4 Typical B–H c urves for s everal oi l-based M R  uids having iron particle volume fractions ranging from 22% to 40%.

t (H ) = C × 271,700 × f 1.5239 × tanh(6.33 × 10 −6 H ) (17.2) B = 1.91 × f 1.133[1 − exp( −10.97 m0 H )] + m0 H

(17.3)

where f is the volume fraction of iron particles t is in Pa B is in T H is in A/m

MR  uids a re t ypically u sed i n either a v alve mo de or a d irect shear mode as illustrated in Figure 17.5. Examples of valve mode devices i nclude s ervo-valves, d ampers, s hock a bsorbers, a nd actuators. Examples of direct shear mode devices include clutches, brakes, chucking and locking devices, some dampers, and structural composites. A less well-known squeeze-lm mode has been explored only minimally and will not be discussed here.

17.5.1 Valve Mode In the valve mode, MR uid ows through a  ow channel and a magnetic  eld i s app lied t ransverse to t he  ow d irection. Ā e yield s trength t hat de velops i n t he  uid e stablishes a p ressure threshold for a ny  uid ow. Varying t he m agnetic  eld allows one to vary this pressure threshold, thus creating a controllable and proportionate valve mechanism without the need for moving mechanical pa rts. A w ell-designed m agnetic c ircuit w ith ele ctromagnet a nd a ncillary h igh m agnetic p ermeability s teel  ux conduits and pole pieces dening the ow channel are all that is required. Valve mode operation is most commonly found in MR uid dampers and shock absorbers. Typically, in these devices, a

TABLE 17.1 Properties of Representative, Commercial Oil-Based MR Fluids Property

Normal Range

Carrier liquid Particle volume fraction, f Particle weight fraction Density (g/cm3) Yield strength (kPa) at 100 kA/m Yield strength (kPa) at 200 kA/m Yield strength (kPa) at saturation

.

MRF-122-2ED

MRF-132AD

MRF-140CG

— 0.20–0.45

Hydrocarbon 0.22

Hydrocarbon 0.32

Hydrocarbon 0.40

0.70–0.90 2–4 10–55 20–80 25–100

0.72 2.38 13 23

0.81 3.09 23 42

0.854 3.64 38 57

∼29 61

∼45 92

∼66 280

Plastic viscosity (mPa s) at 40°C, g > 500 s−1 Operating temperature (°C) Magnetic permeability, relative at low eld

50–200

Response time (s) Ā ermal conductivitya (W/m °C), at 25°C Specic heat (J/g °C), at 25°C Coefficient of thermal expansionb 0–50°C, (∆V/V)/°C

<0.001 — 0.6–1.0 2–7 × 10−4

— 3.5–10

−40 to 130

−40 to 130

−40 to 130

∼4 <0.001 0.21–0.81 0.94 6.5 × 10−4

∼6 <0.001 0.25–1.06 0.80 5.5 × 10−4

∼8 <0.001 0.28–1.28 0.71 5.0 × 10−4

So urces: Lord Corporation, MRF-122–2ED magneto-rheological uid, Lord Technical Data Sheet, DS7014, 2006; Lord Corporation, MRF-132DG magnetorheological  uid, Lord Technical Data Sheet, DS7015, 2006; Lord Corporation, MRF-140CG magneto-rheological  uid, Lord Technical Data Sheet, DS7012, 2006. a Values were calculated with and without magnetic eld applied. b Calculated values.

43722_C017.indd 5

10/3/2008 8:51:24 PM

17-6

Smart Materials Valve mode

Direct shear mode

L w

Force MR fluid

Pressure g

Velocity

MR fluid Flow

Applied magnetic field

Applied magnetic field

FIGURE 17.5 Two most common operation modes for MR uids—valve mode and direct shear mode.

piston moves back a nd forth i n a t ubular housing t hat is  lled with MR uid. Ā e MR uid is forced to pass through an orice in or around the piston. Ā e pressure drop developed by a v alve-mode device can be divided into two components, the pressure ∆Ph due to t he uid viscosity and ∆PMR due the magnetic  eld-induced yield. Ā ese pressures may be approximated by [6,8] ∆Pη = ∆PMR =

12η pQ L g 3w cτ (H ) L g

(17.4)

Direct Shear Mode

In a direct shear mode device, a layer of MR uid is constrained between the poles and a force is applied that causes one pole to move laterally relative to the other. Ā e amount of force needed to cause the uid to shear then depends on the applied magnetic eld strength and the developed shear strength in the MR uid. Again, a w ell-designed m agnetic c ircuit w ith ele ctromagnet and ancillary hi gh m agnetic p ermeability s teel  ux conduits and pole pieces are used to app ly t he magnetic  eld to t he MR uid. Direct shear devices are most commonly found in clutches and brakes. In these cases, the device is congured such t hat a layer of MR uid resists shear when one member is rotated relative to another. Common arrangements are disc and drum types of c lutches a nd b rakes t hat p roduce a c ontrollable a mount o f torque coupling.

43722_C017.indd 6

Fη =

ηpSLw g

(1

FMR = τ (H ) Lw

7.6) (17.7)

where S i s t he rel ative velocity. T he tot al force de veloped by the direct-shear device is the sum of Fh and FMR.

(17.5)

Q is t he volumetric  ow rate. Ā e parameter c has a v alue t hat ranges from a minimum value of 2 for ∆PMR /∆Ph less than ~1 to a maximum value of 3 for ∆PMR /∆Ph greater than ~100. Ā e total pressure drop in a valve-mode device is approximately equal to the su m o f v iscous a nd m agnetic  eld-induced contributions, ∆Ph and ∆PMR. Ā e tot al d amping f orce de veloped b y a v alvemode damper will thus be the total pressure drop multiplied by the piston area.

17.5.2

Ā e force de veloped by a d irect-shear de vice c an be d ivided into Fh the force due to the viscous drag of the uid and FMR the force due to magnetic eld-induced shear stress [6,8]:

17.6

MR Fluid Applications

MR uid technology has progressed to the point where it is routinely used on a commercial scale to provide semiactive control in a variety of automotive and industrial applications. MR uids have been used commercially since the mid-1990s. Today, the greatest d riving f orce b ehind M R  uid t echnology i s pr imary automotive suspension systems. In January 2002, General Motors Corporation introduced the Cadillac Seville STS automobile with a MagneRideÔ suspension system having real-time controllable MR  uid shock absorbers and struts as standard equipment [9]. Ā e MR uid shock absorbers w ere m ade b y D elphi C orporation a nd t he M R  uid was manufactured by Lord Corporation. MR uid-based suspension systems are now available on many models of automobile including C adillac ( SRX, X LR, X LR-V, S TS, a nd D TS), C hevrolet (Corvette), Bu ick ( Lucerne), F errari ( 599 GT B F iorano), A udi (TT a nd R 8), H olden (HS V C ommodore), a nd H onda A cura (MDX) with many more planned. All commercial M R f luid controlled primary suspension systems are based on a monotube shock absorber that utilizes a si ngle-stage, a xisymmetric M R v alve c ontained i n t he piston. Such a M R f luid piston/valve design efficiently provides high on-state and low off-state force in a very compact and robust package. The controllers for these systems process inputs from relative position sensors at each wheel. In addition,

10/3/2008 8:51:26 PM

17-7

Magnetorheological Fluids

Bearing and seal

Electric current

Annular flow channel Steel core

Magnetic flux

Coil

Steel flux ring MR fluid flow

Diaphragm and N2 accumulator

FIGURE 17.6

Basic MR uid damper with axisymmetric valve geometry.

inputs f rom a l ateral accelerometer yaw rate sensor, steering angle sensor, and speed sensor a ll feed by way of a C AN bus [19] i nto t he c ontroller. T he c ontrol a lgorithms si multaneously optimize a w ide range of factors that influence overall handling and ride comfort. System response time is less than 10 ms. Common to virtually all commercial MR uid shock absorbers is single-coil, a xisymmetric M R  uid v alve h aving a n a nnular ow path as shown in Figure 17.6. Ā is simple design has proven to be very reliable and is amenable to efficient, mass production. Ā e s traight  ow pat h w ithout a ny b ends o r c hanges i n  ow direction enables very low off -state forces w ithout compromising the on-state. Automotive MR uid shock absorbers have high durability with an operational life in excess of 150,000 km in a vehicle. Commercial MR uid dampers are available in a variety of c ongurations i ncluding: c oil-over s hock, a ir-lift sh ock, o r Macpherson strut. Other, c ommercially s ignicant, app lication a reas i nclude steer-by-wire ( haptic) s ystems a nd s econdary v ehicle su spension s ystems i n h eavy-duty v ehicles. A dditional de veloping areas that are poised to become commercially signicant include eng ine mounts, fan clutches, a ll-wheel d rive clutches, position locking and detent devices, crash protection systems, civil engineering infrastructure dampers, military vehicle suspensions, a nd p rosthetic de vices. Cu rrent p roduction o f M R uid i s o n t he o rder o f 100,000 L p er ye ar. I n 2 007, i t i s e stimated that more than half a million MR uid devices are in use worldwide [20].

43722_C017.indd 7

References 1. 2. 3. 4. 5. 6. 7. 8. 9.

10.

Phillips, R.W., Engineering Applications o f Fl uids w ith a Variable Yield Stress, P hD thesis, University o f C alifornia, Berkeley, CA, 1969. Rabinow, J ., Ā e ma gnetic  uid clutch. AIEE T rans., 67, 1308, 1948. Magnetic uid clutch, National B ureau of Stand ards Technical News Bulletin, 32(4), 54, 1948. Rabinow, J., Magnetic Fluid Torque and Force Transmitting Device, U.S. Patent 2, 575, 360, 1951. Carlson, J .D., S precher, A.F., a nd C onrad, H., Electrorheological Fluids, Technomic, Lancaster, 1990. Duclos, T.G., D esign o f de vices usin g ele ctro-rheological uids, S ociety o f Automotive En gineers, Technical P aper Series, 881134, Warrendale, 1988. Carlson, J .D., Ā e prom ise of c ontrollable  uids, in Proceedings of Actuator 94, Borgmann, H. and Lenz, K., Eds., AXON Technologie, Bremen, 1994, p. 266. Carlson, J .D., C atanzarite, D .M., a nd S t. Cla ir, K.A., Commercial magneto-rheological uid devices, Int. J. Mod. Phys. B, 10, 2857, 1996. Carlson, J.D., MR  uid t echnology—commercial st atus in 2006, in Proceedings of 10th International Conference on ER Fluids a nd MR S uspensions, G ordaninejad, F ., E d., World Scientic, Singapore, 389, 2007. Carlson, J .D., What mak es a g ood MR  uid, J. I ntelligent Mater. Syst. Struct., 13, 431, 2003.

10/3/2008 8:51:28 PM

17-8

11. Ginder, J.M., Davis, L.C., and Elie, L.D., Rheology of magnetorheological uids: Models and measurements, Int. J. Mod. Phys. B, 10, 3293, 1996. 12. Margida, A.J., Weiss, K.D ., a nd C arlson, J .D., Magnetorheological materials based on iron alloy particles, Int. J. Mod. Phys. B, 10, 3335, 1996. 13. Carlson, J.D. and Weiss, K.D., Magnetorheological Materials Based on Alloy Particles. U.S. Patent 5, 382, 373, 1995. 14. Black, T . a nd C arlson, J .D., M agnetizable  uids, in Synthetics, M ineral Oils, a nd B io-Based L ubricants, Rudnick, L.R ., E d., Taylor a nd F rancis, B oca R aton, FL, 2006, Chapter 35.

43722_C017.indd 8

Smart Materials

15. Lord C orporation, MRF -122–2ED ma gneto-rheological uid, Lord Technical Data Sheet, DS7014, 2006. 16. Lord C orporation, MRF -132DG ma gneto-rheological uid, Lord Technical Data Sheet, DS7015, 2006. 17. Lord C orporation, MRF -140CG ma gneto-rheological uid, Lord Technical Data Sheet, DS7012, 2006. 18. Carlson, J.D., MR uids and devices in the real world, Int. J. Mod. Physics B, 19, 1463, 2005. 19. International Organization for Standardization, ISO-11898, 2003. 20. Toscano, J.R., Lord Corporation, personal communication, 2007.

10/3/2008 8:51:28 PM

18 Intelligent Processing of Materials 18.1 Concept of Intelligent Materials Processing .....................................................................18-1 18.2 Processing of Composite Materials ....................................................................................18-2

J.A. Güemes Universita Politécnica Madrid

J.M. Menéndez AIRBUS

18.1

Cure of Ā ermosetting Resins • Sensors and Sensing Techniques for Composites Manufacturing • Fiber Optic Sensors Applied to RTM Process Monitoring

18.3 Intelligent Processing of Metallic Materials .....................................................................18-9 18.4 C onclusion ..............................................................................................................................18-9 References .........................................................................................................................................18-10

Concept of Intelligent Materials Processing

Intelligent m aterials p rocessing (I PM) i s ac hieved b y t he integration of online sensors, process models, and adaptive control s trategies. I ntelligent p rocessing me ans a c ontrol s ystem that uses prior knowledge and process models to manipulate the processing conditions in response to available online sensor data. In brief, intelligent processing is process control by objectives, rather than control following prescribed parameters. Ā e central elements of IPM are (1) process understanding expressed in ter ms of a p rocess model, (2) real-time i nformation on processing parameters and material condition obtained with online process s ensors, a nd ( 3) a mo del-based s ensing a nd c ontrol strategy to ac hieve t he de sired c haracteristics i n t he  nished product. Ā e traditional approach to material processing establishes a “process w indow,” na mely a r ange of ac ceptable v alues for t he main control pa rameters such as i nitial composition, temperature, and pressure. Ā ese parameters were established from previous experience, or by trial and error, and they may be readjusted to accord the results of the actual production. Metal welding is one such example of this off-line control procedure. Depending on t he t hickness, c omposition, a nd ot her c haracteristics, t he process p arameters a re se lected f rom sta ndard p rocessing c harts. Ā e quality of the execution relies on the operator’s skill. Final quality control through nondestructive procedures (NDE), like x-rays, m aybe ne eded o n c ertain s tructures to g uarantee t he compliance of the metal weld to the required properties. To make the metal welding process “intelligent,” one would need not o nly to h ave a c ontrol on t he electrodes voltage a nd the welding speed, as it is done in conventional processing, but

also to have sensors that collect data from the material in the molted a rea, i n t he same way a s t he operator’s eye gets i nformation a bout i ts s hape a nd c olor. A c omputer-based mo del would be used to predict the  nal microstructure based on the values o f t he mo nitored v ariables a nd t he s etting o f c ontrol parameters. Ā e s ystem t hen w ould p redict h ow to mo dify these c ontrol pa rameters d uring t he sub sequent s teps o f t he process i n o rder to ac hieve t he de sired  nal m icrostructure. Although our technology has not yet been able to produce such an “ intelligent welding machine” to re place the skilled operator, t here i s a n o ngoing p roject o n “W eld p rocess S ensing, Modeling a nd C ontrol” s ponsored b y N IST; de tails c an b e found in Ref. [1]. Ā e si mple w elding e xample h elp to iden tify t he t wo m ain issues that need to be addressed in order to implement “intelligent processing” on any material: 1. Availability o f rel iable s ensors to mo nitor t he p rocess variables. Such sensors must be able to work at the processing en vironment w ithout d isturbing t he p rocess itself. F or i nstance, w hile t hermal a nalysis i s a p owerful way to u nderstand t he material state and its transformations, so it is very useful for off-line control, this sensing principle cannot be applied for online control. 2. Full u nderstanding o f t he p rocess, t he i nterdependence among t he p rocess v ariables t hat c an b e s ensed a nd t he process pa rameters t hat c an b e c ontrolled, i n o rder to achieve p redictive c apability o n t he m aterial e volution. Ā e sensing and processing times must be short compared to the physical times. Ā e faster computers are, the higher will b e t he n umber o f p rocesses t hat c an b e re al-time controlled. 18-1

43722_C018.indd 1

10/3/2008 8:53:11 PM

18-2

Smart Materials

It must be emphasized that “intelligent processing” is not the same as “automated processing,” but a step ahead. Automated processing has t he c apability o f ke eping t he p rocess pa rameters i nside t he process window. It may even include some sensors to keep records of the process variables at intermediate steps or continuously. But automated processing cannot react to m ajor d isturbances, ot her than tr iggering an a larm, an d i t c annot p erform c ontinuous adjustments o f t he c ontrol pa rameters to ac hieve a p redened nal state. Differences are summarized in Table 18.1. Ā e main benet from intelligent processing is that it reduces the time spent in moving a design from project desks into production. Nowadays designers use soft ware tools to evaluate the process c onstraints a nd t he f easibility o f d ifferent concepts, known as “v irtual manufacturing.” However, several manufacturing tests are still required for experimental adjustment of the control parameters before production can begin. With IPM, these parameters would adjust by themselves, producing a highquality product at the  rst attempt. Signicant impact on costs would b e ac hieved b y re ducing t he t ime el apsed b etween t he conceptual design and the nal product. Structural components would not have to be mass-produced to be economical. To exemplify this assessment, let us think about the resin transfer molding (RTM) process for composite materials. Ā e softwa re simulating the resin  ow is a v alid tool for virtual manufacturing, but i t rel ies o n a ssumed d ata f or p erform p ermeability; s light differences i n t he c ompaction o f t he p erform i nduce s trong changes i n t he  ow, t he predictions a re only qu alitative. Only when combining the numerical model with local measurements, as w ill b e s hown l ater i n t his c hapter, t he p redictions m ay b e quantitative. A more ambitious benet is claimed in Ref. [2]: Ā e f ocus o f m aterials p rocess re search i s to e stablish theories and scientic methods for in situ self improvement of the design and control of material processing using selfdirected co mputer a ided s ystems. Ā e specic objectives are: [1] generate new methods and knowledge relative to the design process, [2] integrate, synthesize and generalize new knowledge i n t he f orm o f a xioms, a nd [ 3] a utomatically incorporate t his ne w k nowledge f or u se i n i mproving materials research and process control metrics (i.e., quality,

time and cost). Long term objective is to develop a language, principled in theories, for the development of virtual material processing systems and integrated, intelligent manufacturing systems ac ross a M aterials a nd P rocessing I nformation Highway. Ā ese general points apply to a ny eng ineering material, a nd industrial e xamples m ay b e found i n e very te chnological a rea, from polymers to metals, ceramics, and composites. Ā e different levels of “intelligence” correspond to the accuracy and predictive capability of the mathematical models of the process, the availability o f i n-situ re al-time s ensors a nd t he e fficiency o f t he control algorithms to nd the optimal control strategy. Ā e idea of IPM started around 1990 with the advent of highspeed, low-cost computers. Web sites of laboratories with specic activity on IPM are given in Refs. [3–7]. Ā e link to the University of Virginia deals with metal processing, while the next two are more s pecic f or c omposites. A l ink to t he I nternational Conferences o n “I ntelligent P rocessing a nd M anufacturing o f Materials” is available in Ref. [6]. Information on active or completed projects is available in Ref. [7].

18.2 Processing of Composite Materials Although ini tially r estricted t o mi litary air crafts, composite parts have been i ntensively used during t he last decade i n commercial airplanes to reduce weight and improve performances, the upcoming generation of aircrafts gives a clear indication of this continuing trend. Ā e new Airbus A380, the latest challenge of the European aerospace industry, will include more weight in composite materials than any other aircraft before. Ā is development was conditioned by t he need of substantial reductions of t he op erative c osts c ompared to p revious gener ations o f airliners. Ā e at tempts f or c ost re duction a re adv ancing i n t wo directions: • Design for manufacturing : Design concepts must take into account t he l imitations o f t he p rocess i n o rder to m ake designs better adapted to the manufacturing processes.

TABLE 18.1 Differences between Manual, Automated, and Intelligent Processing Manual Processing

Automated Processing

Ā e process parameters are selected from processing charts

A “process window” is preestablished for the raw material worst-case condition

Process execution relies on operator’s ability

Process parameters are kept inside the “process window” by online controllers Ā e process would not react to batch variation in the material composition, part thickness, environmental changes Example: Common autoclave cure of composite materials

Off-line quality control by NDE

Example: Manual soldering of metal sheets

43722_C018.indd 2

Intelligent Processing Model-based control algorithms calculate and modify in real-time the control parameters to achieve the goal product Real-time sensors monitor the process evolution A detailed math model correlates the required nal properties with the measured variables and the control parameters Example: Smart cure of composites

10/3/2008 8:53:11 PM

18-3

Intelligent Processing of Materials

• New m anufacturing p rocesses: Ā e d evelopment of i nnovative t echnologies i ncludes a utomated p rocedures a nd smart manufacturing. Current composite technologies have reached a go od level of automation. Cut ting, h ot d rape f orming, l aser p rojection, tape lay-up, tow placement, and finishing operations are fully automated. S everal aut omated pr oduction e quipment a nd machines were de veloped during t he 1990s a nd applied to a variety of components. Highly automated composites manufacturing facilities for primary composite aircraft structures are now in service. Most o f t he c omposites u sed f or a erospace s tructures a re made by stacking layers of prepreg material at prescribed directions onto a tool, then compacting and curing the stack to transform i t i nto a mo nolithic s olid o r l aminate. Ā e prepreg layers are done as a fabric or UD tape of high-strength/stiffness bers, emb edded i nto a p olymeric re sin. C omposite m aterial processing i n gener al, a nd a utoclave c uring i n pa rticular, i s a complex p rocess, w ith s ignicant batc h to batc h v ariability, where s ome o f t he c ritical qu ality v ariables, a s vo id c ontent, cannot be measured before the end of the cure, when they cannot be corrected. Ā is lack of access to t he quality variables during the cure is what is forcing to the autoclave cure process to be so conservative, ot herwise t he reje ction r ate o f c ured c omposite parts would be unacceptably high. Smart te chnologies h ave to i nclude t he f ollowing improvements: • Modeling te chniques o f ma terial b ehavior a nd p rocess parameters • Development a nd v alidation o f re al-time p rocess mo nitoring te chniques, suc h a s d ielectrometry, opt ical  bers, pressure sensors, and others • Development o f i ntelligent r egulation s ystems, w ith implementation o f mo deling a nd p rocess mo nitoring techniques Ā e re sults a re e xpected to b e a de crease o f de velopment a nd manufacturing costs a nd lead t imes, en hancement of quality and reliability of composite components, reduction of scraps, and reduction of qualication procedures due to better understanding and higher reliability of process as well as fast adaptation to the change in raw materials. Ā e process monitoring techniques range from simple pressure and temperature data logging to a d istributed sensor network on the p rocess v ariables at t he m anufacturing f acility, t he to oling, and the part itself. Ā e introduction of monitoring techniques is the basis for a subsequent real-time control of the process, whereby the process parameters are modied to optimize the product.

epoxies, bismaleimides (BMIs), a nd c yanate e ster. Ā e se resins undergo a c omplex ph ysicochemical c hange d uring t he c ure process. From t he c hemical p oint of v iew, t he small mole cules composing t he re sin ( molecular w eights r anging f rom 2 00 to 2000 AMU) start reacting each other when temperature is high enough, and, at the end of a process called “cure cycle,” a network of primary bonds is formed, transforming the resin into a no nmelting rigid material. Ā is evolution is measured by the degree of c ure, dened as t he percentage of reaction accomplished. As the chemical reaction is exothermic, the degree of cure is experimentally obtained as the ratio of the heat released up to a given time, to the total heat liberated at the full cross-linking. Under a physical point of view, heating liquees the uncured resin, so its behavior may be described by a v iscous ow. When the resin is reacting, its molecules i ncrease its leng th a nd t he viscosity r aises. Well b efore t he en d o f t he c ure, a gel s tate i s reached and the resin cannot longer ow. Ā e reacting mixture is not yet as stiff as it will become when the cure is completed and it b ehaves a s a n a morphous el astomeric s olid. Ā is important event is called gel transition, and it is the limit for the resin ability to adapt to t he shape of the mold a nd for the voids to c ollapse. Cure needs to be fully completed to attain the best properties of the re sin, pa rticularly i ts  nal g lass t ransition tem perature, which is related to the maximum in-service temperature. Ā e externally controlled process parameters are the autoclave temperature a nd p ressure, a nd t he v acuum le vel, ac ting a ll o f them a s t he b oundary c ondition o f t he u ncured l aminate. Currently, in t he automated procedure, t hey follow a p reestablished time pattern, called “cure cycle.” Ā e i nternal v ariables o f t he m aterial, suc h a s tem perature, pressure, resin viscosity, glass transition temperature, degree of cure, void size, void percentage and gas pressure inside the voids, are f unctions de pending on t he s patial lo cation a nd t ime. Ā e required  nal properties of t he m aterial a re ade quate s trength and stiff ness values. Both are correlated with the resin and void contents, a nd t he i nternal re sidual s tresses t hat app ear d uring the c ure c ycle. Ā e m aximum s ervice tem perature i s d irectly proportional to the nal Tg. Since 1 982, d ifferent a uthors h ave p roposed m athematical models with several levels of complexity, most of them based on phenomenological laws, trying to j ustify experimental ndings [8–10]. Essentially, all of them include ve submodels: 1. 2.

3.

18.2.1 Cure of Thermosetting Resins Even t here i s a n in creasing interest in t hermoplastic r esins as matrices f or c omposites, s till mo st o f t he h igh-performance composites m aterials a re do ne w ith t hermosetting re sins, a s

43722_C018.indd 3

Ā e rmochemical model: Predicts the degree of cure as a function of the temperature and time. Ā ermal transfer model: Calculates the internal temperature distribution as a balance between conductivities, thermal capacitances, and the thermal overshoot due to the internal exothermic reactions. Rheological m odel: P redicts t he v iscosity o f t he re sin a s a function of temperature and degree of cure. It also predicts the Tg and the gel time. It is usually connected to the resin ow mo del, a long t he  ber a nd t hrough t he t hickness. Ā is model calculates the internal pressure, the compaction of the laminate, and the nal resin contents.

10/3/2008 8:53:11 PM

18-4

4.

5.

Smart Materials

Void m odel: Establishes t he worst case scenario for voids that c ome f rom a ir i nitially t rapped b etween ad jacent plies, or later generated at high temperatures either from volatile substances dissolved in the uncured resin or from previous water condensation. Resin pressure should make these voids collapse before gel time is reached. Residual str esses: Ā e c hemical s hrinkage o f t he m atrix and t he a nisotropic e xpansion o f e ach l amina c reate a n internal stress eld, which may produce warping or other undesired geometrical distortions; these internal stresses may be particularly high at the free edges, decreasing the laminate failure loads.

Ā e details for these mathematical models are beyond the scope of this book. Interested readers may  nd a de tailed description in Ref. [10]; even these models rely on many empirical parameters, and in some cases, like in the void model, on heuristic assumptions, t he mo dels a re p owerful to ols f or u nderstanding h ow a change in the autoclave process parameters, or in the tooling and bagging material, may affect the nal part quality.

18.2.2

Sensors and Sensing Techniques for Composites Manufacturing

Conventional sensors are used currently to monitor the autoclave temperature a nd p ressure. H owever, t here i s m uch re search [11–17] being done to d evelop different k inds of sensors for insitu mo nitoring o f v ariables t hat s o f ar rem ain u nknown ( ply compaction p ressure) c an b e ro ughly e stimated (cure re sidual stresses), or can only be measured by laboratory techniques (Tg , degree o f c ure, et c.). Ā ere i s a nother c hapter i n t his b ook dealing with sensors for cure monitoring, written by T. Fukuda and T. Kosaka; to a void redundancies, only a b rief summary is done, together with a more detailed description on a  ber optic cure sensor [11] on which we pioneered the developments.

Dielectric sensors [12] are being used to measure the electrical capacitance a nd c onductivity o f t he re sin d uring t he c uring process. A d rop i n t he conductivity can be correlated w ith t he initiation of a t hree-dimensional structure a nd hence t he gelation and vitrication processes. Figure 18.1 shows a conductivity plot for t he 8552 re sin f rom w hich t he m inimum v iscosity, gel point, and vitrication point can be identied. Ultrasonic s ensors [ 13] u se a coustic pu lses i n e ither t he through-transmission o r t he p ulse-echo mo de. A s t he re sin cures and gets harder, the speed of the sound through it increases and t he at tenuation de creases, app roaching t he l imit v alues characteristic of t he completely cured composite. In t he pulseecho mode, the sensor sends a pulse and then receives the resulting echo f rom t he i nterface. A s t he re sin c ures, t he re turn t ime shortens compared to the original time when uncured. Fiber op tic-based t echniques ( evanescent w ave i nteraction, transmission spectrum analysis, Fresnel reection, uorescence, Raman spectrum measurement, etc.) are good candidates for the validation of cure cycles in autoclave. Ā is is because of the small size of the optical ber and because the optical ber can be easily brought i nside t he autoclave. A c omprehensive re view of more than 200 references on process monitoring of ber reinforced composites u sing opt ical  ber s ensors w as re cently do ne b y Fernando and Degamber [16]. Changes in the infrared spectrum of t he resin a re a go od i ndicator of t he cure evolution. Fourier transform infrared spectroscopy (FTIR) provides an attractive approach, since the information is spectrally encoded, and does not c hange w ith v ariation i n t he i ntensity o f t he l ight s ource. Standard opt ical  bers a re not t ransparent to m id-IR, but s till sensitive changes can be detected at t he near-IR ba nd (4000 to 9000 cm−1) f or e poxy a nd B MI re sins. Ā ere a re c ommercial probes available for IR spectroscopy, based on diffuse reection, but because of their large size, they cannot be embedded in composites. M easurements o f e vanescent wa ve i nteractions i s a promising development [16].

Process of 8552/T800 UD material 200 Minimum viscosity

150

−5.0

Gel point End of cure

−5.5

100

Temperature (⬚C)

log conductivity (S/m)

−4.5

−6.0 50 −6.5

20

60

100 140 Time (min)

180

220

FIGURE 18.1 Typical conductivity plot for 8552 resin.

43722_C018.indd 4

10/3/2008 8:53:11 PM

18-5

Intelligent Processing of Materials

Optical source

3 dB Coupler

Sensor

Photodetector

FIGURE 18.2 Schematic of a Fresnel-based reectometer.

Changes i n t he refractive i ndex of t he resin during t he c ure cycle are a simple but effective approach to monitor the cure cycle. Devices de veloped f or t his p urpose a re k nown a s  ber optic Fresnel reectometers. Ā e cleaved tip of an optical ber operates as a s ensor because the percentage of the back-reected light is proportional to t he refractive index of t he material outside t he ber. B asic s cheme i s s hown at F igure 18.2. N evertheless, t wo main issues have to be solved before a practical implementation of this concept: • Since these optical devices are intensity sensors, they are also sensitive to drifts i n t he opt ical power of t he source and p erturbations i n t he opt ical pat h ( e.g., m icro a nd macrobending of the optical ber). • Refractive index of the resin depends on both the degree of cure and the temperature. A temperature sensor has to be used together with the refractive index sensor. To overcome these issues, we proposed and successfully demonstrated [11] a new scheme (Figure 18.3), where the photodetector was sub stituted b y a n opt ical s pectrum a nalyzer a nd w hich includes two ber Bragg gratings (FBGs) as internal references. FBGs h ave p roven t heir v ersatility a nd f easibility i n m any applications as ber-optic-based smart structures. In this application, they are used as very accurate references to remove the undesirable p erturbations o n t he sig nal re ected bac k b y t he ber c ore–resin interface. Bragg gr atings operate in t his t echnique as narrowband mirrors, reecting a constant percentage of t he opt ical p ower t hat pa sses t hrough t hem. Ā er efore, the gratings a re u sed to de tect t he a mount of t he l ight p ower t hat reaches a c ertain point of t he optical pat h. One device, located

close to the optoelectronic system, operates as a main reference. Ā e other FBG at t he tip of the probe compensates the possible perturbations in the optical path affording also a me asurement of t he temperature. Figure 18.4 s hows t he re sults of a s eries of tests of isothermal cure cycles obtained by this method.

18.2.3

Fiber Optic Sensors Applied to RTM Process Monitoring

RTM is an adequate process for the production of medium to high series, l imiting t he human i nteraction to sup ervising a nd control tasks. A dry preform of bers is placed inside a closed mold, vacuum is applied at the venting ports, the resin gets into the mold through the inlet ports, and must spread out into the preform before reaching the exit or being gelled. To achieve a structural quality similar to that obtained with autoclave curing, a high ber volume is required, meaning that the preform has to be closely packed, which makes the resin’s ow difficult. Should dry spots occur and the resin gel before the end of the injection, the part would be rejected. Ā e s etup o f a R TM p rocess i s c omplex, i nvolving a l arge amount of variables in the tool design, the material, the injection equipment, and the process itself. Long development periods and a high number of initial scraps units are usual. Process modeling techniques are used during the development of successful molding strategies, reducing t he number of t rial a nd error tests. But modeling techniques are not a panacea; in practice, it is very difcult to incorporate all the process coefficients. Some coefficients introduced i n t he mo del a re o nly ro ugh app roximations. Variations in the raw materials, equipment performance, or environmental conditions c an re sult i n er ratic re sults for i njections

3 dB Coupler Optical source

Bragg gratings Probe

Optical spectrum analyzer

FIGURE 18.3 Schematic of a Bragg grating referenced cure monitoring system.

43722_C018.indd 5

10/3/2008 8:53:12 PM

18-6

Smart Materials

1.385 140⬚C 160⬚C

1.375

180⬚C 200⬚C

n

1.365

1.355

1.345

1.335 0

10

20

30

40

50 t (min)

60

70

80

90

100

FIGURE 18.4 Isothermal variation of the refractive index of the resin Hexcel 8552 obtained using an optical probe embedded in an 8552/T800 prepreg laminate at 140°C, 160°C, 180°C, and 200°C.

• Direct interpretability of the results • Small size of the sensor. A cleaved optical ber (0.125 mm of diameter), easily embeddable in the preform, reaching remote a reas o f t he c omposite pa rt u nder a nalysis w ith very low mechanical interference • Simplicity of the optoelectronic equipment (a light source, a photodetector, an optical coupler and an optical switch), data acquisition, and data processing hardware and software • Versatility o f t he c onguration o f t he s ensing ne twork that c an b e emb edded i nto t he p reform o r i ntegrated i n the mold

apparently i ncluded i n t he process w indows. K nowledge of t he resin  ow t hrough t he d ry p reform i s t he mo st c ritical p rocess variable. Some te chniques a re a vailable to mo nitor t he adv ance o f t he resin ow front inside the mold during RTM processes. Fiber optic sensors can easily be integrated in composite materials and so it makes this approach attractive. Ā e arrival of resin can be detected by measuring the change in the refractive index at t he interface. Fresnel reection-based sensors were chosen as the most adequate technique available due to their inherent advantages: • High s ignal-to-noise r atio d ue t o t he hi gh diff erence between the refractive indexes of the air and the resin (a refractive index of 1.3 to 1.5, typical of epoxy resins, promotes a change in the signal of two orders of magnitude) Optical source

Figure 18.5 p rovides a s chematic o f t he s ystem: l ight c oming out from an optical source is launched into an optical ber through a coupler. Ā e tip of the optical ber, which is cleaved

3 dB Coupler Optical switch

Sensors

Photodetector

Flow front

FIGURE 18.5 Schematic of a Fresnel-based RTM ow monitoring system.

43722_C018.indd 6

10/3/2008 8:53:13 PM

18-7

Intelligent Processing of Materials

FIGURE 18.6

Picture of the RTM injection system.

very precisely, is immersed in the medium under analysis, in this case, resin. Ā e ber tip interface acts as a sensor: most of the light that reaches this surface passes through it, but a small percentage of it (about 2%) is reected back according to Fresnel law; this percentage will drop to 0.1% when resin wets the interface . Ā e reected l ight t ravels back t hrough t he coupler and is detected by a p hotodetector. A n o ptical sw itch m ultiplexes several sensors, allowing for the interrogation of them at different preselected positions. Commercially available optical ber switches, from telecommunication applications, a llow to interrogating up to 32 optical bers, with a commutation time of some milliseconds, much better than required, having in mind the characteristic time of the RTM process. Figure 1 8.6 g ives a n e xample o n a  ow m onitoring s ystem implemented i n a c onventional RTM i njection equipment during the manufacturing of carbon ber reinforced plastic (CFRP) ribs of the ele vator le ading e dge o f t he A-340 6 00 a ircraft (Fi gure 18.7). Nine opt ical  bers w ere b rought i nside t he mold t hrough t he vacuum port using three Poly Tetrauroethylene (PTFE) tubes. Āe procedure to i ntroduce t he t ubes i nto t he mold w ithout c ausing vacuum leakage is straightforward, using a T tting in the vacuum port (Figure 18.8a). Ā e sensors are not at tached to t he mold, but located inside the preform, bonded with a small drop of cyanoacrylate, in different positions in the same interface, as shown in Figure 18.8b. Injection of the rib was carried out in special conditions to slow the ow of the resin in the mold (vacuum was applied). Every optical ber sensor survived the assembly of the mold and the injection process, showing a high signal-to-noise ratio. Six o ver eig ht s ensors de tected t he p resence o f re sin. T he time of transit of the resin through the mold was registered as indicated i n F igure 18.9. H ere, t he c omplex ge ometry o f t he

43722_C018.indd 7

part j usties t he u se o f c omputational mo dels to de sign t he mould a nd opt imize t he lo cation o f i nput a nd o utput p orts. Several in jection s trategies ( considering diff erent material permeability a t diff erent p ositions i n t he mold ) w ere p reviously te sted to c orrelate t he si mulated  ow t o th e r esults obtained in real injections. Experimental data of permeability were used as input in these models to obtain a map of times of resin transit. Ā ese maps were compared to t he data given by the ber optic monitoring system. Results presented in Figure 18.10 (numbers indicate the order of the activation of the sensors)

FIGURE 18.7 Composite part produced by RTM.

10/3/2008 8:53:14 PM

18-8

Smart Materials

7

8

out port

6

5

2

3

4 in port

1 RI

(a)

FIGURE 18.8 the preform.

(b)

(a) Detail of the T  tting with the vacuum port and the optical ber ingress. (b) Picture of the position of the sensors attached to

show t hat t he re al e volution o f t he re sin qu alitatively c orresponds to t he results obtained t hrough computational methods. However, besides important quantitative differences, the detection of dry spots i n t wo sensing points justify t he use of monitoring systems in RTM processes during the development stage: Minor variations on the preform condition bring about large c hanges i n t he re sults. Ā is d evelopment s upports t he use o f i ntelligent c ontrol s ystems to opt imize t he  lling o f molds with complex geometry. An advantage of using optical bers for RTM ow detection is that they can later be used to monitor the cure and to obtain the residual stresses. Ā is is simply done by adding a Bragg grating to the ow monitoring sensor, as explained in previous paragraphs. Ā us, the same device is used to monitor the arrival of the resin with one of the systems and, after this moment, the evolution of the re fractive i ndex w ith t he ot her s ystem. No ne gative e ffects over the sensors have been detected during the different steps of

the p rocess: m anufacture o f t he p reform, i ntegration o f t he sensors, assembly of the mold, heating of the mold, injection of the resin, and cure of the resin. Figure 18.11 shows the evolution of t he re fractive i ndex d uring t he c uring p rocess a nd t he temperature i n t he c oolest a nd t he h ottest p oint i n t he mold during the process. While the Bragg grating is very sensitive to the temperature, it is found that changes in the refractive index are due both to changes in the degree of cure and to changes in the tem perature, w hich p ose d ifficulties to t he a nalysis o f t he experimental results. It was found a low resolution of the technique after the gel time. It can be concluded t hat a  ber optic  ow monitoring system can detect t he resin f ront progression in real molding w ith very few limitations, and it has demonstrated as a valuable tool in the development stages of RTM process. Fiber optic cure monitoring systems a re i n a n e arly s tage o f de velopment, i mprovements i n their sensitivity, particularly at high degrees of cure, are required.

1.2 0

38,076

10,39

1

41,405

2,535

R

0.8 0.6 0.4

Sensor 1 Sensor 2 Sensor 3 Sensor 4 Sensor 5 Sensor 6 Sensor 7 Sensor 8

46,874

0.2 0 −20

−10

0

10

20 t (s)

30

40

50

60

FIGURE 18.9 A340-600 elevator rib: time of transit of the resin. Experimental values.

43722_C018.indd 8

10/3/2008 8:53:15 PM

18-9

Intelligent Processing of Materials

Out port

X

6

4

5

3

2

Inport

1 RI

FIGURE 1 8.10 A340-600 elevator rib: time of transit of the resin calculations.

18.3 Intelligent Processing of Metallic Materials

n

References [19–22] a re representative of t he work done i n t he eld of intelligent processing of metals and metal matrix composites ( MMC). M ost o f t he pap ers de al w ith i ntelligent h ot isostatic p ressing (H IP) a s a n adv anced p rocess to f abricate near-net shape parts with low ductility, difficult to melt intermetallic a lloys of aerospace interest, a s titanium a luminures. Ā e consolidation process includes solid state reactions, adding a difussional ow to the classical plastic yielding and creep. The p rocessing pa rameters d uring t he H IP p rocess a re temperature, pressure, and time. A practical way to understand the densication response for a p owder material is through the use of HIP maps,  rstly introduced by Ashby [23], for constant temperatures a nd p ressures. W hen de aling w ith M MC, a n additional p roblem app ears; h igh tem peratures a nd p ressures help t he c ompaction, b ut m ay de grade t he c omposite b y

18.4 Conclusion In o rder to m aintain t he te chnological le ad i n t he a rea o f advanced m aterials, i t i s e ssential to b uild t he qu antitative knowledge of the relationships between material properties and processes. Ā e development of advanced models is essential for the development of manufacturing systems must be done parallel to advances in sensor and control technology. Whereas a model provides a n u nderstanding o n t he ph ysics a nd c hemistry involved, and explains the relationship between dependent and

1,405

190

1,4

170

1,395

150

1,39

130

1,385

110

1,38

90

1,375

T (⬚C)

X

fiber–matrix reaction or ber breakage. Ā e nal objective is to achieve t he maximum densication at minimum time without causing ber degradation. Ā e laws for plastic deformation a nd creep  ow a re g iven in Ref. [ 22], toge ther w ith k inetic e quations f or d iffusion, ber– matrix re activity a nd p robability o f  ber f racture. Ā is model allows to si mulate t he o utcome to a dy namic h ipping l aw a nd coupled to a predictive control algorithm to nd the best process cycle for a g iven s et of i nitial c onditions, m aterials properties, machine dynamics, and desired product goal state. Eddy current probes were done using boron nitride preforms and p latinum w indings. Ā ese p robes me asure c hanges i n t he thickness of composite samples to a precision better than 20 µ m during HIP cycles up to temperatures in excess of 1000°C. Ā e authors conclude that the extension of IPM as a feedback control me thod w ould a llow to m aterial eng ineers to de sign processes to at tain a go al-state m icrostructure w ithout lo ng experimental p rograms. M aterials c ould b e b uilt opt imally adapted to its applications. Ā e main limitation comes from the availability o f no ninvasive s ensor te chniques, ba sed o n e ddy currents, l aser u ltrasonics, e tc., t hat m ay i nform o n e volving microstructure.

70 n Tmax Tmin

1,37

50

1,365 0

20

40

60

80

100

30 120

t (min)

FIGURE 18.11 Mold temperatures and refractive index of the resin during curing.

43722_C018.indd 9

10/3/2008 8:53:16 PM

18-10

control v ariables, s ensors p rovide i nformation rel ated to t he state of the material and process variables. Ā e control function maintains quality through the use of data provided by sensors as inputs to t he mo dels, a nd m anages t he p rocess to ac hieve t he desired product. Pioneering e xperiences h ave demo nstrated t he f easibility of t he I PM c oncepts, b oth f or adv anced c omposites a nd metallic materials. Among the benets for composite materials are t he re duction o f de velopment c osts a nd le ad t imes, t he enhancement o f qu ality a nd rel iability, a nd a si mplification of t he qu alification p rocedures. F or me tallic m aterials, most o f t he e xperiences a re c oncentrated o n c onsolidation o f intermetallic p owders, b ut ot her p rocesses a s f orging a nd soldering are also investigated. For both cases, the availability of affordable s ensors i s t he c ritical i ssue f or i ts i ndustrial implementation.

References 1.

www.msel.nist.gov. 2. LeClair-S-R and Jackson-A, Work Unit Directive (WUD54) Amendment: M aterials P rocess D esign. N ASA no . 19980069695. 1996. 3. www.ipm.virginia.edu. 4. www.ipc.northwestern.edu. 5. www.ccm.udel.edu. 6. www.ipmm.mining.ubc.ca. 7. www.ims.org. 8. A.C. Loos and Springer G., Curing of epoxy matrix composites, Journal of Composite Materials, 21:243–261, 1983. 9. S.G. Advani a nd P. S imacek, Modelling a nd sim ulation o f ow, heat transfer and cure (in r esin transfer molding). In Resin Transfer Moulding for Aerospace Sciences. Dordrecht, Netherlands a nd N orwell, MA: K luwer Academic, 1998, pp. 225–281. 10. R.S. D avé a nd A. L oos (E ds.), Processing of C omposites, Munich: Hanser Verlag, 2000. 11. F. Ro dríguez-Lence, P. M uñoz-Esquer, J .M. M enéndez, C. Pardo de Vera, S. Díaz, and J.A. Güemes. Smart sensors for resin  ow and composite cure monitoring. Proceedings of the 12th International Conference on Composite Materials, ICCM-12, Paris, France, 1999.

43722_C018.indd 10

Smart Materials

12. C.W. L ee a nd B .P. Rice , r esin tra nsfer p rocess mo nitoring and control, SAMPE Journal, 34(6), 48–55, 1998. 13. D.D. Shepard and K.R. Smith, A new ultrasonic measurement system for the cure monitoring of thermosetting resins and composites, Journal of Ā er mal Analysis, 49(1), 95–100, 1997. 14. G. M aistros, D . B olios, G. Crac knell, a nd A. Milb urn, Intelligent c uring o f co mposites: I ntegration o f diele ctric cure m onitoring wi th a kn owledge-based sys tem f or efficient co mponent r ecovery, P roceedings o f the 19th International SAMPE Europe Conference of the Society for the Advancement o f M aterials a nd P rocess En gineering, pp. 733–743, Paris, 1998. 15. S. C ossins, M. C onnell, and B. Cross et al ., In situ near-IR cure mo nitoring o f a mo del ep oxy ma trix co mposite, Applied Spectroscopy, 50(7), 900–905, 1996. 16. G.F. F ernando a nd B . D egamber, P rocess mo nitoring o f bre r einforced co mposites usin g o ptical  bre sensors, Inttnal Materials Reviews, 51(2), 65–106, 2006. 17. Y.M. Liu, C. Ganesh, J.P.H. Steele, and J.E. Jones, Fiber optic sensor development for real-time in-situ epoxy cure monitoring, Journal of Composite Materials, 31(1), 87–102, 1997. 18. J.L. K ardos, M.P . Dud ukovic, a nd R . D ave. Advances i n Polymer S cience, Vol. 80 E poxy r esin a nd co mposites. I n: Karel Dusek, ed. Berlin: Springer Verlag, 1986, pp. 101–124. 19. D.G. B ackman, E.S. R usell, D .Y. Wey, a nd Y. P ang, Intelligent prodessing f or met al ma trix co mposites, Proceedings of the S ymposium on Intelligent P rocessing of M aterials, Minerals, M etals a nd M aterials S ociety, Warrendale, PA, 1989. 20. T.F. Z ahrah, F. Cha rron, a nd N.M. Wereley, M odel-based consolidation f or nea r-net sha pe, AIAA/ASME S tructures, Structural Dynamics and Materials Conference, April 1994, pp. 2315–2323. 21. H.N.G. Wadley, I ntelligent p rocessing o f sma rt ma terials, AGARD LS-205 o n S mart S tructures a nd M aterials: Implications for Military Aircraft of New Generation, 1996. 22. H.N.G. Wadley a nd R . Vancheeswaran, Ā e intelligent processing of materials: An overview and case study, Journal of Materials, 19–30, January 1998. 23. W.B. Li , M.F. Ashby, a nd K.E. E asterling, On den sication and sha pe-change d uring ho t is ostatic p ressing, Acta Metallurgica, 35, 2831, 1987.

10/3/2008 8:53:17 PM

19 Magnets, Magnetic, and Magnetostrictive Materials 19.1 M agnets, Organic/Polymer ...................................................................................................19-1 V(TCNE)x · z(Solvent) Room Temperature Magnets • M(TCNE)2 · zCH2Cl2 (M = Mn, Fe, Co, Ni) High Room Temperature Magnets • Hexacyanometallate Room Temperature Magnets • Uses of Organic/Polymeric Magnets

Acknowledgments .............................................................................................................................19-7 References ...........................................................................................................................................19-7 19.2 Powder Metallurgy Used for a Giant Magnetostrictive Material Actuator Sensor.......................................................................................................................19-8 Introduction • Features of GMM • Characteristics and Physical Properties of GMM • Features of Powder Metallurgical GMM • Applications of GMM • Conclusion

References ......................................................................................................................................... 19-16

19.1

Magnets, Organic/Polymer

Joel S. Miller and Arthur J. Epstein Magnetism h as ena bled t he de velopment a nd e xploitation o f fundamental s cience r anging f rom qu antum me chanics, to condensed m atter c hemistry a nd physics, to m aterials s cience. Control of t he ma gnetic beha vior o f ma terials ha s enab led t he wide spread availability of low-cost electricity and electric motors, to the development of telecommunication devices (microphones, televisions, tele phones, e tc.) a s w ell a s t he m agnetic s torage f or computers. M agnets a re su itable c omponents f or s ensors a nd actuators a nd a re c rucial f or sm art m aterials a nd s ystems. Magnetic materials known from time immemorial are comprised of either transition or rare earth metals or their ions with spins residing in d- or f-orbitals, respectively, e.g., Fe, Gd, CrO2, SmCo5, Co17Sm2, a nd N d2Fe14B. Ā ese m aterials a re p repared b y h ightemperature metallurgical methods and generally are brittle, and chemically reactive. Ā e l atter h alf of t he t wentieth c entury h as witnessed the replacement of many metal and ceramic materials with l ightweight p olymeric m aterials. W hile t his h as b een p rimarily occurred for structural materials, examples in increasing numbers have also occurred for electrically conducting and optical materials. More recently, examples new of magnetic materials [1] have been reported and undoubtedly the twenty-rst century will see the commercialization of organic/polymeric magnets [2].

Magnetism i s a d irect c onsequence o f t he s pin o r e xchange coupling of unpaired electron spins. Noncoupled, independent electron spins (Figure 19.1a) lead to paramagnetic behavior, which c an be qu antitatively modeled by t he Cu rie law a nd t he Brillouin expression. Weak coupling (opposed or antiferromagnetic, or in alignment or ferromagnetic) leads to a modication of behaviors a nd is modeling w ith t he Curie-Weiss law. St rong coupling can lead to m agnetic ordering and alignment of spins (Figure 19.1b) and consequently a substantial magnetic moment and a ferromagnet, while that of opposed spins (Figure 19.1c) to an antiferromagnet as the net moments cancel. In contrast, the incomplete c ancellation o f s pins c an le ad to a ne t m agnetic moment and a ferrimagnet (Figure 19.1d) or a canted antiferromagnet (or weak ferromagnet) (Figure 19.1e). Ā e discovery of organic-based magnets with spins residing in p-orbitals have expanded the following properties already associated w ith m agnets to i nclude, e .g., s olubility, mo dulation o f t he properties via organic chemistry synthetic methods, and low-temperature ( nonmetallugical) p rocessing, en hancing t heir te chnological importance and value as part of “smart” materials/systems. A few organic nitroxides order ferromagnetically below a Tc of 1.5 K. Ā ese i nclude o ne p olymorph o f 4 -nitrophenyl n itronyl nitroxide reported by Kinoshita and coworkers with a Tc of 0.6 K (Figure 19.2a) [1c], a nd a p olymorph o f d initroxide re ported b y Rassat a nd c oworkers w ith a Tc of 1.48 K (Figure 19.2b) [3]. In addition, several canted antiferromagnets, e.g., one polymorph of 19-1

43722_C019.indd 1

10/3/2008 9:42:00 PM

19-2

Smart Materials

(a)

(b)

(c)

(d)

(e1)

(e2)

FIGURE 19.1 Two-dimensional spin alignment for (a) paramagnet, (b) antiferromagnet, (c) ferromagnet, (d) ferrimagnet, and (e) canted antiferromagnet behavior.

4-NCC6F4CNSSN (Tc = 3 6 K [4] t hat i s i ncreased to 65 K u nder 16 kbar pressure [4b]) (Figure 19.2c), and the electron transfer salt {[(H3C)2N]2C2[N(CH3)2)2]}+ [C 60] ( Tc = 16 K) [5], have been reported. Ā ese exa mples of organic-based magnets, a lbeit w ith very low Tc values, are crystalline solids, and not polymers. Reports of magnetic ordering in C60-based polymers, however, exist [6]. Organic-based magnets possessing unpaired electron spins in both p- and d-orbitals also have been reported [1a–1c]. Ā is includes i onic de camethylferrocenium te tracyanothanide, {Fe[C5(CH3)5]2}[TCNE] (Tc = 4.8 K; TCNE = tetracyanoethylene), Figure 19.3), exhibiting the rst evidence for magnetic hysteretic behavior in an organic-based magnet [7,8]. Ā is magnet has an alternating D + A−D+ A− { D = { Fe[C5(CH3)5]2}+; A = [ TCNE]−} solid state structure (Figure 19.4). Ā e observed 16,300 emu Oe/ mol s aturation m agnetization ( Ms) i s i n e xcellent a greement with the calculated value of 16,700 emu Oe/mol for single crystals aligned parallel to the chain (Fe···Fe) axis. Ā is value is 37% greater than that observed for iron metal on either a Fe or mole basis. Hysteresis loops with a coercive eld of 1 kOe are observed at 2 K (Figure 19.5) [1b,7,8]. Each D a nd A h as a si ngle s pin (S = 1 /2). A bove 16 K, t he magnetic su sceptibility b ehaves a s e xpected f or a o ne-dimensional (1D) ferromagnetically coupled Heisenberg chain with

J/k B = 27 K [ 8]. B elow t his tem perature, t he su sceptibility diverges a s ( T − Tc)−g a s a nticipated f or a 1 D H eisenberg-like system app roaching a 3 D m agnetically o rdered s tate. Sp ontaneous magnetization below the 4.8 K ordering temperature follows (Tc − T)b with b ∼ 0 .5. Hysteresis lo ops a re we ll d ened with coercive eld (Hcr) of 1 kOe at 2 K (Figure 19.5), indicating substantial domain walls pinning. Replacement of Fe by M n [9] a nd Cr [ 10] a nd substitution of TCNE with TCNQ [11–14] [7,7,8,8-tetracyano-p-quinodimethane (Figure 19.3c)], leads to fe rromagnets with Tc reduced for TCNQ substitution and enhanced when Mn is utilized (Table 19.1 and Figure 19.6). Partial substitution of spinless (S = 0) {Co[C5(CH3)5]2}+ for (S = 1/2) {Fe[C5(CH3)5]2}+ in {Fe[C5(CH3)5]2}+ [TCNE]− leads to a rapid decrease in Tc as a function of the fraction of spinless sites (1 − x) occurs, e.g., 2.5% substitution of {Fe[C5(CH3)5]2}+ sites by {Co[C5(CH3)5]2}+ decreases Tc by 43% [15]. Details of the magnetic properties can be found in reviews [16]. Gatteschi, Rey, and coworkers [1d] demonstrated and subsequently Iwamura a nd c oworkers [17] h ave s hown t hat c ovalent p olymers c omprised of bi s(hexuoroacetylacetonate)-manganese(II) (Figure 19.7) and nitroxides bound to the Mn(II) sites order as ferrimagnets. Using more complex nitroxides, Iwamura and coworkers have prepared related systems with Tc values ∼46 K. F

O

–

F N

O N•

+ N

•

N

O2N

N

N O (a)

•

•

O (b)

N

S

C

F

S

F

(c)

FIGURE 19.2 Structures of 4-nitrophenyl nitronyl nitroxide, which orders as a ferromagnet at 0.6 K (a), a dinitroxide, that orders as a ferromagnet at 1.48 K (b), and 4-NCC6F4CNSSN, which orders as a canted antiferromagnet at 36 K (c). Ā e unpaired electron spins are designated by a dot (·).

43722_C019.indd 2

10/3/2008 9:42:00 PM

19-3

Magnets, Magnetic, and Magnetostrictive Materials

N

N Fe

(a)

FIGURE 19.3

N

N

C

C

C

C

C

C

C

C

N (b)

N

N (c)

N

Molecular structures of Fe[C5(CH3)5]2 (a), TCNE (b), and TCNQ (c).

Using TC NE ele ctron-transfer s alts o f M nII(porphyrin)’s, e.g., [ MnTPP][TCNE] ( H2TPP = meso-tetraphenylporphyrin, Figure 19.8 [18], magnets with Tc < 28 K) [19] based upon metallomacrocycles can also be prepared. Both the [TCNE]− and the nitroxides each have one spin; however, the Mn(II) in the former 1D polymeric chain has ve spins (S = 5/2), while the Mn(III) in the latter polymeric chain has four spins (S = 2). In both cases, the Mn and organic spins couple antiferromagnetically, leading to ferrimagnetic ordering. Ā e solid state mot if is d istinctly d ifferent than th at f or { M[C5(CH3)5]2}+ [ TCNE] (Fi gure 1 9.4) a s

[TCNE]− does not coordinate to the M in the latter system. Āus, the bonding of [TCNE]− to Mn is a model for the bonding of [TCNE]− to V in the V[TCNE]x ·z(solvent) room temperature magnet. As a c onsequence of the alternating S = 2 a nd S = 1 /2 chain structure, t he [ MnTPP]+ [T CNE]− 1 D c hain s ystem, w hich forms a large family of magnets, is a model system for studying a n umber o f u nusual m agnetic ph enomena. Ā is i ncludes t he magnetic beha vior o f m ixed q uantum/classical sp in s ystems [20a] and the effects of disorder [20b], and the role of classical dipolar i nteration ( in c ontrast to qu antum me chanical exchange), a nd fractal interactions [21] in achieving magnetic ordering [20c]. Due to t he si ngle ion a nisotropy for [MnTPP]+ and t he large d ifference i n orbital overlaps a long a nd between chains, t his class of materials often u ndergo lattice- a nd spindimensionality cross-overs as a function of temperature [20d]. Metamagnetic-like behavior is noted for many members of this

Magnetization, M (emu Oe/mol)

15,000

10,000

5000

0 −5000 −10,000 −15,000

FIGURE 1 9.4 Crystal s tructure d epicting s egments of t wo p arallel out-of-registry c hains of { Fe[C5(CH3)5]2}[TCNE] s howing t he or bitals possessing the largest density of unpaired electrons.

43722_C019.indd 3

1000 −1000 0 Applied field, H (Oe)

FIGURE 19.5 2 K hysteresis loops with a coercive eld (Hcr) of 1 kOe for {Fe[C5(CH3)5]2}[TCNE].

10/3/2008 9:42:01 PM

19-4

Smart Materials

TABLE 19.1 Summary of the Critical Temperatures (Tc) and Coercive Fields (Hcr) for {M[C5(CH3)5]2}[A] (M = Cr, Mn, Fe; A = TCNE, TCNQ) [TCNE]•− M SD+ SA− Tc (K) q (K) Hcr (kOe) (K) References

FeC 1/2 1/2 4.8 +16.9

Mn 1 1/2 8.8 +22.6

1.0 (2) [7,8]

1.2 (4.2) [9]

[TCNQ]•− Cr 3/2 1/2 3.65 +22.2

Fe 1/2 1/2 3.0 +3.8

Mn 1 1/2 6.3 +10.5

Cr 3/2 1/2 3.3 +11.6

a

b

3.6 (3) [12]

a

[10]

[11]

[13]

Not observed. b Not reported. a

CF3

F3C 8

O

6

O

O

Tc (K)

Mn TCNE-based

4

F3C

TCNQ-based

O CF3

O N Et

2 N Fe

Mn

Cr

1

2

3

0 0

O

Number of spin (M )

FIGURE 19.6 Critical tem perature (Tc) for { M[C5(CH3)5]2}[A] ( M = Fe, Mn, Cr; A = TCNE, TCNQ).

FIGURE 19.8

43722_C019.indd 4

FIGURE 1 9.7 Structure o f M nII(hfac)2NITEt ( hfac = he xauoroacetylacetonate; NITEt = ethyl nitronyl nitroxide).

Segment of an 1D ···D+ A−D+ A−··· chain of [MnTPP][TCNE] ∙ 2C6H5CH3 showing [MnTPP]+ trans-µ-N-σ-bonding to [TCNE]−.

10/3/2008 9:42:02 PM

19-5

Magnets, Magnetic, and Magnetostrictive Materials

Magnetization, M (emu Oe/mol)

20,000

10,000

0

−10,000

−20,000 −40,000

−20,000 0 20,000 Applied magnetic field, H (Oe)

40,000

FIGURE 1 9.9 Metamagnetic a nd h ysteresis w ith 27 ,000 Oe c ritical, Hc, a nd c oercive, HcR,  elds at 2 K for meso-tetrakis(4-bromophenyl) porphinatomanganese(III) tetracyanoethanide. (From R ittenberg, D.K., Sugiura, K-i., Sakata, Y., Mikami, S., Epstein, A.J., and Miller, J.S., Adv. Mater., 12, 126, 2000.)

family at low temperature with unusually high critical elds (Hc) a nd u nusually h igh c oercive  elds (Hcr) a s l arge a s ~27,000 Oe and may have substantial remanent magnetizations (Figure 19.9) [20e].

19.1.1

V(TCNE)x ∙ z(Solvent) Room Temperature Magnets

Reaction of V(C6H6)2 [22a] or V(CO)6 [22b] and TCNE in a variety of solvents [22c], e.g., dichloromethane, leads to lo ss of the benzene l igands a nd i mmediate f ormation o f V (TCNE)x · zCH2Cl2 (x ~ 2; y ~1/2). Due to its extreme air and water sensitivities a s w ell a s in solubility, c ompositional inh omogenieties within a nd b etween p reparations o ccur, a nd t he s tructure remains e lusive. Ā is m aterial, h owever, is t he  rst ambient temperature or ganic- or p olymer-based m agnet ( Tc ~ 4 00 K). Ā e proposed structure has each V b eing octahedrally coordinated w ith up to si x l igands (N l igands f rom d ifferent TCNE ligands) and each TCNE is reduced and is either planar or twisted and bound to up to four V atoms. Recently, solvent-free thin magnetic  lms of V(TCNE)x composition on a v ariety of substrates, e .g., si licon, s alt, g lass, a luminum, h ave b een p repared [23]. Enhanced stability can be achieved via overcoating the V(TCNE)x  lms with Parylene [24]. Ā e inhomogeneities in t he s tructure i s e xpected t o lead t o variations in the magnitude, not sig n, of the exchange between the V II (S = 3/2) and [TCNE]− (S = 1/2), and the disorder should result in a small anisotropy [22].

TCNE = t etracyanoethylene) h ave b een pre pared [ 25]. Ā ese materials have Tc’ values of 97, 75, 44, and 44 K, respectively. Field cooled/zero eld cooled magnetization studies suggest that while the Mn [26] system is a reentrant spin glass, the Fe [27] system is a random anisotropy system. Both systems exhibit complex behavior below Tc. For example, hysteresis curves for the Fe compound, taken at 5 K, a re c onstricted, w ith a s pin-op shape, indicating ferrimagnetic behavior, and eld- and zero-eld-cooled magnetization studies reveal magnetic irreversibilities below Tc for both compounds. Static and dynamic scaling analyses of the dc m agnetization and ac susceptibility data for the Mn compound show that this system undergoes a transition to a 3D ferrimagnetic state at Tc, followed by a reentrant transition, to a s pin glass state at Tg = 2.5 K. For M = Fe, the results of static scaling analyses are consistent with a high-T transition to a correlated sperimagnet, while below ~20 K, there is a cross-over to sperimagnetic behavior. Solid solutions of VxM2−x[TCNE]2 (M = Fe, Co, Ni) composition have been prepared that enables control to the coercive eld and Tc [28]. In addition, photoinduced magnetic behavior, which is wa velength d ependent, is o bserved f or M n(TCNE)2 · zCH2Cl2) [29]. In addition to V(TCNE)2-based m agnets, s ubstitution o f TCNE with α-α¢-dicyanoperuorostilbene (DCNPFS), tetracyanopyrazine (TC NP), te tracyanobenzene (TC NB), a nd TC NQ form V(A)2 (A = D CNPFS [30], TCNP [31], TCNB [32]) with Tc ~ 2 00 K, and V(TCNQ)2 with Tc ~ 5 0 K [33]. Ā e reaction of TCNB and V(CO)6 also forms a magnet whose Tc exceeds room temperature (Tc ~ 325 K) [32].

19.1.2 M(TCNE)2 ∙ z(CH2Cl2) (M = Mn, Fe, Co, Ni) High Room Temperature Magnets

19.1.3

Additional memb ers o f t he f amily o f h igh-Tc or ganic-based magnets M (TCNE)2 · z(CH2Cl2), ( M = M n, F e, C o, Ni;

Although t echnically n ot o rganic-based m agnets, s everal materials ter med mole cule-based m agnets a re p repared b y

43722_C019.indd 5

Hexacyanometallate Room Temperature Magnets

10/3/2008 9:42:02 PM

19-6

Smart Materials

magnetic ordering with high Tc values. Examples include ferromagnetic CsNi II[CrIII(CN)6] ∙2H 2O (Tc = 90 K) [34], and ferrimagnetic Cs Mn II[CrIII(CN)6] ∙ H 2O (Tc = 9 0 K) [35], a s w ell as t hin f ilms ( ≤2 µm), either oxidized or reduced, of CrIII[CrIII(CN)6]0.93[CrII(CN)6]0.05 (Tc = 260 K) [36] that exhibit negative magnetization (Figure 19.11). Verdaguer and coworkers reported that ferrimagnetic V II 0.42V III I0.58[CrIII(CN) 6] 0.86 ∙ 2.8 H2O magnetically ordered above room temperature (Tc = 315 K) [37]. Further study of this class of materials has lead to an en hancement o f t he Tc to ~100°C (373 K) (Figure 19.12) [38]. Specific details on the properties of this class of magnets can be found in reviews [39].

19.1.4 Uses of Organic/Polymeric Magnets

FIGURE 1 9.10 Idealized s tructure o f F eIII4[FeII(CN)6]3 · zH2O, Prussian blue with ···FeIII-N≡C-FeII-C≡N-FeIII··· linkages along each of the three unit cell axes.

the s ame o rganic c hemistry me thodologies a nd h ave b een reported to m agnetically o rder i n m any c ases, a nd a bove room tem perature i n a f ew c ases. P russian bl ue, Fe III4 [FeII(CN)6]3 · zH 2O, p ossesses a 3 D ne twork s tructure, w ith ···FeIII–N ≡ C–FeII–C ≡N–FeIII··· linkages along each of the three identical unit cell axes of the cubic unit cell (Figure 19.10) and is a prototype st ructure t hat st abilizes bo th f erro- a nd ferrimagnetic ord er. R eplacement of i ron w ith ot her s pinbearing metal-ions leads to strong magnetic coupling and

The m agnetic e specially i n c onjunction w ith ot her physical properties m ay le ad to t heir u se i n sm art m aterials i n t he future [2]. This includes the next generation of electronic, magnetic, or photonic devices ranging from magnetic imaging to data storage and to static and low-frequency magnetic shielding a nd m agnetic i nduction. T he l atter appl ication i n static a nd lo w-frequency m agnetic s hielding a nd m agnetic induction i s p articularly p romising [ 2b], as r elatively hi gh initial p ermeabilities h ave b een re ported f or t he V (TCNE)x ∙ z(CH 2 Cl 2) materials. This, combined w ith t heir low density (~1 g/cm 3), re latively low re sistivity (~103 ohm-cm), a nd low power lo ss (as low a s ~2 erg/(cm 3 c ycle)) su ggest t hat f uture generations of these and related may be practical, especially for app lications re quiring lo w w eight. F easibility f or V(TCNE) x ∙ z(CH 2Cl 2) i n m agnetic s hielding appl ications i s illustrated i n F igure 1 9.13. O ther p otential app lications include photoinduced magnetic effects [29,40], and spintronic applications [41].

Oxidized, amorphous

Magnetization, M (emu Oe/mol)

200

150 100

100 50 0

0 −50

−100

−100 −150

−200 0

50

200 100 150 Temperature, T (K)

250

Magnetization, M (emu Oe/mol)

200

−200 300

FIGURE 19.11 Temperature-dependent magnetization of the amorphous lm of CrIII[CrIII(CN)6]0.98−[CrII(CN)6]0.02, zero eld cooled (+) and eld cooled in 5 Oe (°), −5 Oe (●), 10 Oe (ⵧ), −10 Oe (■) after 2 min oxidation at −0.2 V upon warming in a 5 Oe eld. (From Buschmann, W.E., Paulson, S.C., Wynn, C.M., Girtu, M., Epstein, A.J., White, H.S., and Miller, J.S., Adv. Mater., 9, 645, 1997; Chem. Mater., 10, 1386, 1998.)

43722_C019.indd 6

10/3/2008 9:42:02 PM

19-7

Magnets, Magnetic, and Magnetostrictive Materials

References

FIGURE 19.12 Sample of a V[Cr(CN)6]y ∙ zH2O m agnet (Tc = 3 72 K) being at tracted to a T eon coated magnet at room temperature in the air. (Hatlevik, O., Buschmann, W.E., Zhang, J., Manson, J.L., and Miller, J.S., Adv. Mater., 11, 914, 1999.)

FIGURE 19.13 Folded soft iron rods (staples) are shown attracted to a SmCo5 permanent magnet (left). With a pellet of V(TCNE)x ∙z(CH2Cl2) at ro om t emperature s hielding t he m agnetic  eld, t he s oft ir on r ods hang freely.

Acknowledgments Ā e authors gratefully acknowledge the extensive contributions of t heir c ollaborators, s tudents, a nd p ostdoctoral a ssociates i n the studies discussed herein. Ā e authors also gratefully acknowledge the continued partial support by the Department of Energy Division of Materials Science (Grant Nos. DE-FG03-93ER45504, DE-FG02-86ER45271, a nd D E-FG02-01ER45931), t he U .S. A ir Force Office of Scientic Research (Grant No. AFOSR FA955006-01-0175), a nd N ational S cience F oundation ( Grant N o. 0553573).

43722_C019.indd 7

1. (a) Blundell, S. J., Pratt F. L. J. Phys.: Condens. Matter, 2004, 16, R771. (b) Miller, J. S., Epstein, A. J., Angew. Chem. Int. Ed. 1994, 33, 385. Miller, J. S., Epstein, A. J., Chem. Eng. News, 1995, 73#40, 30. (c) Kinoshita, M., Jap. J. Appl. Phys. 1994, 33, 5718. (d) Gatteschi, D. Adv. Mat. 1994, 6, 635. Coronado, E., Gatteschi, D., J. Mater. Chem. 2006, 16, 2513. 2. (a) L andee, C. P., M elville, D., Miller J . S., in NATO ARW Molecular M agnetic M aterials, K ahn, O ., G atteschi, D ., Miller, J. S., Palacio, F., Eds. 1991, E198, 395. (b) Morin, B. G., Hahm, C., Epstein, A. J., Miller, J. S., J. Appl. Phys. 1994, 75, 5782. (c) Miller, J. S., Epstein, A. J., Chemtech 1991, 21, 168. (d) Miller, J. S., Adv. Mater. 1994, 6, 322. 3. Chiarelli, R., Rassat, A., Dromzee, Y., Jeannin, Y., Novak, M. A., Ā olence, J. L., Phys. Scrip. 1993, T49, 706. 4. (a) Palacio, F., Antorrena, G., Castro, M., Burrile, R., Rawson, J., Smith, N. B., Bricklebank, N., Novoa, J., Ritter, C., Phys. Rev. Lett. 1997, 79, 2336. (b) Mito, M., Kawae, T., Takagi, S., Matsushita, D. H., Rawson, J. M., Palacio, F., Polyhed. 2001, 20, 1509. 5. Omerzu, A., Arcon, D ., B linc, R ., Miha ilovic, D ., in Magnetism—Molecules to M aterials, J . S. Miller a nd M. Drillon, Eds., Wiley-VCH, Mannheim, Vol. 2, p. 123 (2001). 6. (a)Han, K.-H., Spemann, D., Höhne, R., Setzer, A., Makarova, T., Esquinazi, P., B utz, T., C arbon, 2003, 41, 785. (b) M akarova, T.L., Su ndqvist, B ., H öhne, R ., E quinazi, P., K apelvich, Y., Scharff, P., Davydov, V., Kashevarova, L.S., Rakhmania, A.V., Nature, 2006, 440, 707. 7. (a) Miller, J. S., Calabrese, J. C., Epstein, A. J., Bigelow, R. W., Zhang, J . H., Reiff , W. M., J. Chem. S oc. Chem. Co mmun. 1986, 1026. (b) Miller, J. S., Calabrese, J. C., Rommelmann, H., Chittipeddi, S. R., Zhang, J. H., Reiff, W. M., Epstein, A. J., J. Am. Chem. Soc. 1987, 109, 769. 8. Chittipeddi, S., Cr omack, K. R ., Miller, J. S., E pstein, A. J., Phys. Rev. Lett. 1987, 58, 2695. 9. Miller, J. S., McLean, R. S., Vazquez, C., Calabrese, J. C., Zuo, F., Epstein, A. J., J. Mater. Chem. 1993, 3, 215. 10. Yee, G. T., M anriquez, J . M., Dix on, D . A., M cLean, R . S., Groski, D. M., Flippen, R . B., Narayan, K. S., Epstein, A. J., Miller, J. S., Adv. Mater. 1991, 3, 309. 11. Broderick, W. E., Eichorn, D. M., Lu, X., Toscano, P. J., Owens, S. M., Hoffman, B. M., J. Am. Chem. Soc. 1995, 117, 3641. 12. Broderick, W. E., Ā ompson, J. A., Day, E. P., Hoffman, B. M., Science 1990, 249, 410. 13. Broderick, W. E., Hoffman, B . M., J. Am. C hem. S oc. 1991, 113, 6334. 14. Taliaferro, M. L., Palacio, F., Miller, J. S., J. Mater. Chem. 2006, 16, 2677. 15. Narayan, K. S., Kai, K. M., Epstein, A. J., Miller, J. S., J. Appl. Phys. 1991, 69, 5953; Narayan, K. S., Morin, B. G., Miller, J. S., Epstein, A. J., Phys. Rev. B 1992, 46, 6195. 16. Yee, G. T., Miller, J. S., in Magnetism—Molecules to Materials, J. S. Miller and M. Drillon, Eds., Wiley-VCH, Mannheim, Vol. 5, p. 223 (2005). Coronado, E., Galàn-Mascarós, J. R., Miller, J. S., in Comprehensive Organometallic Chemistry, Vol. 3 (2006).

10/3/2008 9:42:03 PM

19-8

17. Inoue, K., Hayamizu, T., Iwamura, H., Mol. Cryst. Liq. Cryst. 1995, 273, 67. Izoka, A., Murata, S., Sugawara, T., Iwamura, H., J. Am. C hem. S oc. 1985, 107, 1786; J. Am. C hem. S oc. 1987, 109, 2631. 18. Miller, J . S., E pstein, A. J ., J. Chem. S oc., Chem. Co mmun. 1998, 1319. 19. Brandon, E. J ., Ri ttenberg, D . K., Arif, A. M., Miller , J . S., Inorg. Chem. 1998, 37, 3376. 20. (a) Miller, J. S., Calabrese, J. C., McLean, R. S., Epstein, A. J., Adv. Mater. 1992, 4, 498. (b) Brinckerhoff, W. B., Morin, B. G., Brandon, E. J., Miller, J. S., Epstein, A. J., J. Appl. Phys. 1996, 79, 6147. (c) Wynn, C. M., Girtu, M. A., Brinckerhoff, W. B., Sugiura, K-I., Miller, J. S., Epstein, A. J., Chem. Mater. 1997, 9, 2156. (d) Wynn, C. M., Girtu, M. A., Miller, J. S., Epstein, A. J., Phys. Rev. B 1997, 56, 315. (e) Rittenberg, D. K., Sugiura, K-i., Sakata, Y., Mikami, S., Epstein, A. J., Miller, J. S., Adv. Mater. 2000, 12, 126. 21. Etzkorn, S. J., Hibbs, W., Miller, J. S., Epstein, A. J., Phys. Rev Lett. 2002, 89, 207201. 22. (a) Manriquez, J. M., Yee, G. T., McLean, R. S., Epstein, A. J., Miller, J. S., Science 1991, 252, 1415. Miller, J. S., Yee, G. T., Manriquez, J. M., Epstein, A. J., in the Proceedings of Nobel Symposium #NS-81. Conjugated P olymers a nd Re lated Materials: Ā e I nterconnection of C hemical and E lectronic Structure, Oxford University Press, Oxford, 1993, p. 461; La Chim. L a In d. 1992, 74, 845. (b) Zha ng, J ., Zho u, P ., Brinckerhoff, W. B., Epstein, A. J., Vazquez, C., McLean, R. S., Miller, J. S., ACS Sym. Ser. 1996, 644, 311. (c) Ā orum, M. S., Pokhodnya, K. I., Miller, J. S. Polyhedron 2006, 25, 1927. 23. Pokhodnya, K. I., Epstein, A. J., Miller, J. S., Adv. Mater. 2000, 12, 410. 24. Pokhodnya, K. I., Bonner, M., Miller, J. S., Chem. Mater. 2004, 16, 5114. 25. Zhang, J., Ensling, J., Ksenofontov, V., Gütlich, P., Epstein, A. J., Miller, J. S., Angew. Chem. Int. Ed. 1998, 37, 657. 26. Wynn, C. M., Girtu, M. A., Zhang, J., Miller, J. S., Epstein, A. J., Phys. Rev. B 1998, 58, 8508. 27. Girtu, M. A., Wynn, C. M., Zhang, J., Miller, J. S., Epstein, A. J., Phys. Rev. B 2000, 61, 492. 28. (a) Vickers, E. B., S enesi, A., Miller, J. S., Inorg. C him. Acta 2004, 357, 3889. (b) Pokhodnya, K. I., Vickers, E. B., Bonner, M., Epstein, A. J ., M iller, J . S ., Chem. Mater. 2004, 16, 3218. (c) Pokhodnya, K. I., Burtman, V., Epstein, A. J., Raebiger, J. W., Miller, J. S., Adv. Mater. 2003, 15, 1211. 29. (a) P ejakovic, D ., K itamura, C., Miller, J . S., E pstein, A. J ., Phys. Rev. Lett. 2002, 88, 57202. (b) Pejakovic, D., Epstein, A. J., Kitamura, C., Miller, J. S., J. Appl. Phys. 2002, 91, 7176. 30. Fitzgerald, J. P., Kaul, B. B., Yee, G. T., ChemComm, 2000, 49. 31. Vickers, E. B., Selby, T. D., Miller, J. S., J. Am. Chem. Soc. 2004, 126, 3716. 32. Taliaferro, M. L., Ā orum, M. S., Miller, J. S., Angew. Chem. 2006, 45, 5326. 33. Vickers, E. B., Selby, T. D., Ā orum, M. S., Taliaferro, M. L., Miller, J. S., Inorg. Chem. 2004, 43, 6414.

43722_C019.indd 8

Smart Materials

34. Gadet, V., Mallah, T., Castro, I.,Verdaguer, M., J. Am. Chem. Soc. 1992, 114, 9213. 35. Greibler, W. D., Babel, D., Z. Naturforsch 1982, 87b, 832. 36. Buschmann, W. E., Paulson, S. C., Wynn, C. M., Girtu, M., Epstein, A. J., White, H. S., Miller, J. S., Adv. Mater. 1997, 9, 645; Chem. Mater. 1998, 10, 1386. 37. Ferlay, S ., M allah, T., O uahes, R ., Veillet, P., Verdaguer, M ., Nature 1995, 378, 701. 38. (a) Hatlevik, O., Buschmann, W. E., Zhang, J., Manson, J. L., Miller, J . S., Adv. Ma ter. 1999, 11, 914. (b) H olmes, S. D ., Girolami, G., J. Am. Chem. Soc. 1999, 121, 5593. 39. (a) Verdaguer, M., Girolami, G. S., In Magnetism—Molecules to M aterials, Miller , J . S., Drillo n, M., E ds., Wiley-VCH: Weinheim, 2005, Vol. 5, p. 283. (b) Verdaguer, M., Bleuzen, A., Marvaud, V., Vaissermann, J., Seuleiman, M., Desplanches, C., Scuiller, A., Train, C., Ga rde, R ., G elly, G., L omenech, C., Rosenman, I., Veillet, P., Cartier, C., Villain, F., Coord. Chem. Rev. 1999, 190–192, 1023. (c) Verdaguer, M., B leuzen, A., Train, C., Ga rde, R ., F abrizi de B iani, F., D esplanches, C., Phil. Trans. R. Soc. Lond. A 1999, 357, 2959. (d) Hashimoto, K., Ohkoshi, S., Phil. Trans. R. Soc. Lond. A 1999, 357, 2977. 40. (a) Sato, O., Iyoda, T., Fujishima, A., Hashimoto, K., Science 1996, 272, 704. (b) Pejakovic, D. A., Manson, J. L., Miller, J. S., Epstein, A. J., J. Appl. Phys. 2000, 87, 6028. 41. (a) Prigodin, V. N., Raju, N. P., Pokhodnya, K. I., Miller, J. S., Epstein, A . J., Adv. Ma ter. 2002, 14, 1230. (b) R aju, N. P., Savrin, T., P rigodin, V. N., P okhodnya, K. I., Miller , J . S., Epstein, A. J., J. Appl. Phys. 2003, 93, 6799.

19.2

Powder Metallurgy Used for a Giant Magnetostrictive Material Actuator Sensor

Hiroshi Eda and Hirotaka Ojima 19.2.1 Introduction Magnetostrictive materials for actuator devices have used ferrite and n ickel m aterials i n t he pa st, h owever, t he a mount o f d isplacement of those materials has been only 30 ppm while piezoelectric materials whose amount of displacement is 800 ppm has been used for actuator devices. Currently, the giant magnetostrictive material (GMM) has at least 1 500 ppm i n t he a mount o f d isplacement a nd GM M h as generated s tress t hree t imes l arger t han t hat o f p iezoelectric materials and is as good or better than that of piezoelectric materials [1]. Ā erefore, the GMM is expected to be used for actuator devices in li eu o f p iezoelectric m aterials. B y c hanging t he process o f m anufacture f or GM M f rom t he c onventional Bridgman method to a p owder-metallurgical method, we have developed t he GM Ms at a lo w cost a nd w ith a h igh a mount of displacement.

10/3/2008 9:42:03 PM

19-9

Magnets, Magnetic, and Magnetostrictive Materials

: Mg,

: Cu

Crystal structure of Laves phase alloy

Laminated structure of Crystal orientation (111)

19.2.3

FIGURE 19.14 Crystal structure of Laves phase alloy.

At present, GMM are binary cubic compounds “RT2” of a lanthanoid element [2] (R), which has a large magnetic moment and magnetoelastic energy (Tb, Sm) and an iron group element (T), which covers Fe, C o, Ni, M n, e tc. or ps eudobinary c ubic compounds “RR¢T2” where R¢ a re D y (dysprosium), Ho (holmium) etc. [3]. However, the Curie temperature of RR¢ is very low and a high magnetic  eld i s ne cessary to d rive t he GM M f rom R T2. Ā us, ps eudobinary c ubic c ompounds T b0.3Dy 0.7T2 ha ve be en commercialized and SmR¢T2 has been marketed. RT2 has a Laves phase C15 structure shown in Figure 19.14.

19.2.2

and negative magnetostrictive materials while the former has a dimension increase with a driving magnetic eld (Tb0.3Dy 0.7T2) and the latter has a dimension decrease with a driving magnetic eld (SmT2). When an orthogonal driving magnetic eld is applied to t hese materials, these materials become kinked. Ā is phenomenon i s c aused b y m agnetizing t he m aterials w ith t he driving magnetic eld, and when the dimensions of the materials are c hanged b y e xternal s tress, t he m agnetization i ntensity changes. Ā at is, if the amount of change of the magnetization intensity is detected by a pickup coil, the GMM is potentially able to be used as a s ensor for stress and torsion. Moreover, since the Young’s mo dulus o f t he m aterial c hanges b y c hanges i n t he dimension o f t he m aterial, f or e xample, t hen t he m aterial whose LE e ffect i s a bove 2 0% c an b e c onsidered a s a su rface acoustic wa ve de vice [ 4]. Ā ese f eatures a re s ummarized i n Figure 19.15.

Features of GMM

Magnetostriction is caused by the interaction of elasticity and magnetoelasticity. Magnetostrictive materials are both positive

Characteristics and Physical Properties of GMM

Currently, t hin- lm s tudies a nd de velopments o f b ulk GM M have been actively researched in the United States and Europe, however, much of t he work on bulk GM M has been applied to practical app lications. GM M ac tuators w ere m ainly de veloped into the device, which is made with unidirectional solidication or anisotropic crystal forming in a magnetic eld [5]. Important attributes c onsist o f: m agnetostriction ( l), w hich i s l arge a nd hysteresis characteristics, which are low for a d riving magnetic eld (H), while frequency response and temperature characteristics reect a large-bandwidth. Figure 19.16 shows an H-l characteristic o f T b0.3Dy 0.7(Fe0.9Co0.05)1.92 made by the Bridgman method ( PS-2) a nd a p owder me tallurgical me thod ( PMS-1). From this gure, the magnetostriction of PMS-1 is equivalent to that of PS-2, however, the hysteresis characteristics of PMS-1 are about 10% lower than that of PS-2. Table 19.2 shows the characteristics a nd physical properties of GM M made by t he p owder metallurgical method.

Stress

Magnetic field H

Negative

Positive H

H=0

H = (a)

Positive and inverse magnetostriction effect

(b)

Twist, torsion detection

FIGURE 19.15 Positive and inverse magnetostriction effect of GMM.

43722_C019.indd 9

10/3/2008 9:42:03 PM

19-10

Smart Materials Magnetostriction ∆I/I (ppm) 1500

Powder-metallurgy compact: PMS-1 (pressure: 60 kgf/cm2) GMM by bridgman method: PS-2 (pressure: 100 kgf/cm2)

1000

500

−1000

−500

0

500

1000

Magnetic field, H (Oe)

FIGURE 19.16 Magnetostriction of GMM by Bridgman method (PS-2) and powder-metallurgy method (PMS-1).

TABLE 19.2 Powder Metallurgical Process Characteristics and Properties Material Name

PMS-1

PMH-1

Element Geometry

PMT-1

Magnetic eld H = 0.4 kOe = 1.0 kOe = 2.0 kOe = 3.0 kOe

600 ppm 1100 ppm 1300 ppm 1450 ppm

600 ppm 1100 ppm 1250 ppm 1400 ppm

600 ppm 1100 ppm 1300 ppm 1450 ppm

Generated stress (kgf/cm2)

Magnetic eld H = 0.4 kOe = 1.0 kOe = 2.0 kOe = 3.1 kOe

80(−) 200(−) 280(−) 390(−)

80(−) 200(−) 280(−) 390(−)

80(−) 200(−) 280(−) 390(−)

Mechanical property

Density (g/mL) Young’s modulus (N/m2) Compressive strength (kgf/cm2) Tensile strength (kgf/cm2)

7.95(−) 2.0 × 10 + E10 600 × 10 + E6

7.95(−) 2.0 × 10 + E10 600 × 10 + E6

7.95(−) 2.0 × 10 + E10 600 × 10 + E6

20 × 10 + E6

20 × 10 + E7 ←

Coefficient of thermal expansion (CTE) (ppm/°C) Specic heat (kJ/kg . °C) Ā ermal conductivity (W/m . °C)

43722_C019.indd 10

PMS-2

Board, Prism, Cylinder, etc.

Displacement (ppm) (preload: 60 kg/cm2)

Ā ermal property

PIS-1

12(−)

350 ppm 650 ppm — — — — — — 7.95(−) —

750 ppm 1250 ppm 1400 ppm 1550 ppm 100(−) 250(−) 395(−) 470(−)

←

8.50(−) 2.4 × 10 + E10 720 × 10 + E6

20 × 10 + E8

←

24 × 10 + E6

←

←

←

0.35(−)

←

←

←

←

10/11/2007

←

←

←

←

10/3/2008 9:42:04 PM

19-11

Magnets, Magnetic, and Magnetostrictive Materials

TABLE 19.2 (continued)

Powder Metallurgical Process Characteristics and Properties

Material Name

PMS-1

PMH-1

Element Geometry

PIS-1

PMS-2

Board, Prism, Cylinder, etc.

Electromagnetic property

Curie temperature (°C)

380(−)

←

←

←

←

Electrical resistance (W . m) Energy density(kJ/m3) Relative magnetic permeability Coercitivity (Hc; Oe)

600 × 10 − E8

←

←

←

←

Added characteristics

19.2.4

PMT-1

10(−) 8(−)

10(−)

50 Anisotropy and positive magnetostrictive standard material

20 Anisotropy and positive magnetostrictive low-hysteresis material

←

Features of Powder Metallurgical GMM

Displacement magnitude (ppm)

Anisotropic GM M m ade b y t he p owder me tallurgical me thod has b een p rocessed si milar to a p rocess u sed to p roduce r areearth m agnets, a nd h as b een f ormed i nto c ylindrical a nd C-forms by a near-net shape technique. However, the density of the device that was produced by the method above is low compared with the device made by the Bridgman method and the generated stress for t he device volume is low. W hen compared with the Bridgman method which utilizes a solid-state sintering or molten iron coagulation method, it is possible to add materials f or t he i mprovement o f t he m agnetism o f t he v arious ele ments and to realize the anisotropic GMM without Tb. Temperature characteristics, frequency response, and hysteresis characteristics of GMM are shown in Figures 19.17 and 19.18

1300 1200 1100 1000 900 800 700 600 500 400 300 200 100 0 −60

10(−) ≈6(−)

≈6(−)

—

70 Anisotropy and positive magnetostrictive large-bandwidth temperature material

50 Isotropy and positive magnetostrictive powder metallurgy material

16(−) ≈8(−) 50 Anisotropy and positive magnetostrictive high generation stress material

and Table 19.3, respectively. Magnetostrictive characteristics are directly in uenced b y f orming t he a nisotropic m aterial i n a magnetic  eld w ith a p owder me tallurgical me thod, a nd t he anisotropic o rientation c haracteristic i s i mportant. T able 1 9.4 shows the magnetostriction value of the GMM when made with metallic molding and rubber molding.

19.2.5 Applications of GMM Applications of GMM have been developed, and the bulk material of GMM has been applied to power instruments. Table 19.5 shows some examples of the applications of GMM. Ā e features of t his m aterial a re a lo w d riving vol tage, no ncontact d riving and lo w i mpedance. M oreover, t he adv antage o f GM M c ompared with the piezoelectric material has been to be able to drive

PMS-I PMT-I

−40

−20

0

20 40 Temperature (⬚C)

60

80

100

120

FIGURE 19.17 Temperature characteristic of GMM in magnetic eld H = 1 kOe.

43722_C019.indd 11

10/3/2008 9:42:04 PM

19-12

Smart Materials 0.2 0.18 0.16

∆Ip-p (µm)

0.14 0.12 0.1 PS-2

0.08

PMS-1

0.06 0.04 0.02 0 0.1

1

10

100

Frequency (kHz)

FIGURE 19.18 Frequency characteristic of PS-2 and PMS-1.

the p ower de vice i n a h igh-temperature r ange a nd h ave h igh fatigue endurance. However, e ven i f t he ele ctrical re sistance i s lo w a nd t he response is fast, these characteristics are effected by eddy-current loss. Ā us, in a h igh-frequency drive, the shape of the GMM is necessary to b e a t hin plate or have an acicular shape while the laminated shape is needed for a large driving force. However, in order to drive an actuator in a high-frequency area, the design of the g iant m agnetostrictive ac tuator (GMA) i s i mportant. O ne

must keep in mind the time constant of the coil of the GMA since t he dem agnetizing  eld b y t he e ddy c urrent i s re duced. Figure 19.19 shows the relationship between the size of the thin plate and the frequency characteristic. Ā e Curie temperature of GMM based on Tb and Sm is from 380°C to 420°C. Since the temperature characteristic of GMM is superior to that of piezoelectric materials, which is about 180°C, GMA has been considered as an automotive actuator especially since the coefficient of thermal expansion of GMM is near that

TABLE 19.3 Temperature Characteristics, Frequency Response, and Hysteresis Characteristics of GMM Characteristic (Elemental Substitution) Tb0.3Dy0.7Fe1.9 Co/Fe = 3% substitution Co/Fe = 6% substitution Ni/Fe = 6% substitution

Coercitivity Hc (Oe) 80 100 65 52

Magnetostrictive Hysteresis (∆λ)max 150 170 120 110

Magnetostrictive Value ∆l/l (ppm) H = 0.4 kOe

H = 1.0 kOe

580 680 620 660

1080 1150 1110 1150

TABLE 19.4 Magnetostriction Values of GMM Magnetostrictive Value: ∆l/l (ppm)

Magnetostrictive Characteristic (Mold and Methods) Metallic molding Rubber molding

43722_C019.indd 12

H = 0.4 kOe Molding of longitudinal magnetic eld Molding of transverse eld Molding of longitudinal magnetic eld Molding of transverse eld

450 680 600 740

H = 1.0 kOe 900 1200 1100 1300

10/3/2008 9:42:04 PM

19-13

Magnets, Magnetic, and Magnetostrictive Materials

TABLE 19.5

Examples of GMM Applications Actuator Large Displacement

Demands Typical application

Twist

Large Amplitude

Small Input

Valve opening control, table top control, clutch, servo valve, rudder face control, focus adjustment, etc.

Stepping motor, angle control, etc.

Pump, vibration exciter, sonar, traveling wave motor, ultrasonic disk cutter, ultrasound bath, active damper, etc.

Acouophone, various speakers, ultrasonic vibrator, linear motor, etc.

䊊

≥

䊊

䊊

×

≥

≥

䊊

×

≥

䊊

×

×

×

≥

䊊

䊊

≥

䊊

䊊

䊊

of gener al me tallic components. Ā us, si nce t he d imensions of GMA changes w ith i ncreased temperature by t he Joule heat of the coil, it is necessary to consider the design of the structure of GMA in order to reduce the change in dimensions by temperature. Many designs have been suggested. Ā e magnetomechanical coupling factor of GMM is from 0.75 to 0.8 [6] and equivalent to t hat o f a p iezoelectric m aterial. Ā e de vices w ith GM M a re discussed in the next section. 19.2.5.1

Sensor

Positioning

High speed on–off valve, printer head, shutter, fuel injection valve, high speed brake, driller, clutch, etc.

Standard material PMS-1 Wideband temperature PMT-1 Low-hysteresis material PMH-1

Oscillator

All-Around Vibration Sensor/Actuator

Ā is vibration sensor makes use of the Villari effect of GMM [7]. Ā e sensor has a low impedance, the transmission loss is low and

Vibration and Load

SAW, etc.

Vibration sensor, SAW, AE sensor, ultrasonic water level displacement indicator, etc. sensor, contact sensor, location sensor, magnetic eld sensor, magnetostrictive applied generation of electricity, weighing sensor, etc.

it is detectable i n a w ide temperature a rea. Figure 19.20 shows the sensor structure and Figure 19.21 shows the sensor’s characteristics. On the other hand, if current is input into the sensing terminal, the sensor is then able to be used as the oscillator, which i s app licable to c ontrolling t he t able top . H owever, t he oscillator is applicable and can be used as a traveling-wave motor in the ultrasonic range [8]. 19.2.5.2 Giant Magnetostrictive Torsional Actuator/Torque Sensor When we produced the isotopic cylindrical device of GMM and applied a cylindrical axial magnetic eld as well as a circumferential

0.2 0.18 0.16

D Ip-p ( µm)

0.14 0.12 0.1 0.08 0.06

10f

6f

1f

2f

(mm)

0.04 0.02 0 0.1

1

10

100

Frequency (kHz)

FIGURE 19.19 Frequency characteristic of GMM for material shape.

43722_C019.indd 13

10/3/2008 9:42:05 PM

19-14

Smart Materials Structure of multidirectional vibration sensor

Structure of unidirectional vibration sensor

A 60⬚

60⬚

A

A–A cross-sectional view Weight Weight GMM

Pressure bolt

Pressure bolt Pickup coil

Permanent magnet

Case

Pickup coil

Permanent magnet

Lowe yoke (a)

Size: Φ 18 ⫻ 25 mm

Size: Φ 12 ⫻ 25 mm

(b)

FIGURE 19.20 Structure of GMM vibration sensor.

magnetic eld to the device, the device was magnetized in the synthetic direction of those magnetic elds. Ā erefore, if the circumferential ma gnetic  eld i s  xed a nd t he a xial m agnetic  eld h as a n alternating-current m agnetic  eld, t he s ynthetic m agnetization direction is above 90° and a great amount of kinking occurs to the cylindrical device. On the other hand, if the circumferential magnetic eld is  xed and a pick up coil is equipped in the axial direction and the device is twisted, the voltage is induced in the pick up coil with a magnetic ux change. Figure 19.22 shows the structure of the torsional actuator device and Figure 19.23 shows the device’s characteristics. Other app lications, w hich a re u sed, i nclude i ndustrial instruments, mi crorobots, a nd a ctive da mpers, w hich h ave

been s tudied u nder t he a uspices o f t he J apan S ociety o f Mechanical E ngineers a nd t he J apan S ociety f or P recision Engineering in Japan. 19.2.5.3 GMA for Valve Ring Indentation From t he mo del o f F igure 1 9.24, t he f ollowing e quation w as derived. m d2 x = FB + F0 sin w x − m d FN dt 2 ⎡ m d2 x ⎤ + Ss y ⎢ 2 = FB + F0 sin wx − f r x − Fp ⎥ ⎣ dt ⎦

S/N ratio characteristic 40

150

30

(a)

Output (dB)

Output voltage (mV)

Acceleration characteristic 200

100

50

0

(19.1)

Vibration amplitude: 35 µm

20

10

0

2 4 Acceleration (G)

0

6 (b)

0

50

100

150

200

250

Frequency (Hz)

FIGURE 19.21 Characteristics of GMM vibration sensor.

43722_C019.indd 14

10/3/2008 9:42:05 PM

19-15

Magnets, Magnetic, and Magnetostrictive Materials Yoke

Pressure bolt (or spring) GMM

Solenoid coil

Permanent magnet

Search coil Toroidal coil

Yoke Housing

FIGURE 19.22 Structure of giant magnetostrictive torsional actuator.

Output voltage (mVp–p)

1000

750

500 ip–p: 1.2A Hp–p: 100 Oe H1DC: 200 Oe (vertical) H2DC: 200 Oe (circumferential)

250

0

0

300

600 900 Input frequency (Hz)

1200

FIGURE 19.23 Characteristic of giant magnetostrictive torsional actuator.

F B + F0 sinwt

Head

Valve sheet

Ssy (Fp)

d0 = 2r0

d = 2r

F B : Load of indentation m d: Coefficient of dynamic friction F N: Radial load (function of depth x) S: Area of indentation

A sinwt

m dFN (fr x)

x, n

(S = p (r 2 −r0 2)) sy : Yield stress fr : Coefficient of friction resistance (fr x = mdF N) Fp: Plastic deformation resistance (Fp= Ssy) w: Angular frequency of vibration F0sin wt: Vibrational load A: Amplitude of vibration x: Depth of indentation v: Velocity of indentation

FIGURE 19.24 Model of valve ring indentation.

43722_C019.indd 15

10/3/2008 9:42:08 PM

19-16

Smart Materials

Assuming that the equation of motion (19.1) has equality, since the r ight-hand side o f E quation 1 9.1 i s p ositive, t he f ollowing equation was derived: FB + F0 sin w x > m d FN + Ss , (1

9.2)

When Equation 19.2 is true, the valve rings are able to be pressed into place.

19.2.6 Conclusion It h as b een a bout 3 0 ye ars si nce GM M w ere d iscovered at t he Naval Surface Warfare Center (NSWC). During this time, scientists and researchers have commercialized GMM and the quality and c ost of GM M h as b een i mproved, ne vertheless t he c ost of GMM is still high for civil applications. Bad hysteresis characteristics for users of GMM have not improved sufficiently enough to warrant its full usage. We are keenly aware of our responsibility to resolve t hese current problems to ena ble t he use GMM, a nd this gives impetus for ourselves as well as other users to continue to pursue developments and improvements.

43722_C019.indd 16

References 1. J. L. Butler, Application Manual for the Design of ETREMA Terfenol-D Magnetostrictive Transducers, Image Acoustics, Inc., N . Manseld; T. Mori; I. Uni, Giant Magnetostriction Materials and Ā eir Applications, PhD dissertation, 1998. 2. A. E. Clark, Ferromag. Mater., 1, 531, 1980; A.E. Clark and H. Eda, G iant m agnetostrictive m aterials, Nikkan-kogyoshinbun (in Japanese), 3–80, 1995. 3. J. L. Butler, Application Manual for the Design of ETREMA TERFENOL-D™, Edge Technologies, p. 25, 1988. 4. D. C. Webb, K. L. Pavis, N. C. Koon, and A. E. Gangaly, Appl. Phys. Lett., 31(4), 1245, 1977. 5. G. F. Clark et al., Philosophical Mag. B, 1982, 146(4), 331–343. 6. Terfenol-D Notes, Edge Technologies Inc., Etrema Producys Division. March, 1990. 7. Y. Yamamoto, H. E da, T . M ori, a nd R . Amen, Ā e 8th Symposium on Electromagnetics and Dynamics, p. 79, 1996. 8. T. M ori a nd K. U shida, Ā e 8th S ymposium o n Electromagnetics and D ynamics, p. 85, 1996; T. Mori and H. Eda, A study on torque induction in the GMM, J. Alloys Compounds, 258, 93, 1997.

10/3/2008 9:42:09 PM

20 Shape-Memory Alloys and Effects: Types, Functions, Modeling, and Applications 20.1

Magnetically Controlled Shape Memory Alloys................................................................20-1 Activation of the Shape Memory Alloys with Magnetic Field • Production of MSMAs • Appl ications

References .......................................................................................................................................... 20-6 20.2 SMAs in Micro- and Nanoscale Engineering Applications ............................................ 20-8 Introduction • SMAs: A Brief Introduction from a Microsystem Design Prospective • SMA Microactuators: Design Principles • Intrinsic Methods: Tailoring the Microstructure • Extrinsic Methods • Monolithic Design: Laser Annealing of SMA (LASMA) • Summary of SMA Actuator Design Methods • SMA Flexures • Future Directions for SMA Applications in Microtechnologies

References .........................................................................................................................................20-16 20.3 Mathematical Models for Shape Memory Materials .......................................................20-17 Introduction • E xperimental Observations • O verview of the Literature • For mulation of SMM Models in the Framework of Ā ermomechanics with Internal Variables • Summary and Concluding Remarks

References .........................................................................................................................................20-27 20.4 S hape Memory Alloys ........................................................................................................ 20-28 SMAs and the Martensitic Transformation • SMAs Systems • Fu nctional Properties of SMAs References .........................................................................................................................................20-35 Bibliography .................................................................................................................................... 20-36 20.5 Smart Materials Modeling ................................................................................................. 20-36 Introduction • Conservation Laws, Equations of State, and Constitutive Equations • I nformal Classication of Modeling Techniques for Smart Materials • C onclusion

References ........................................................................................................................................ 20-41 20.6 On the Microstructural Mechanisms of SMEs ............................................................... 20-41 Microscopic Mechanisms • M acroscopic Effects • Su mmary

Acknowledgments .......................................................................................................................... 20-46 References ........................................................................................................................................ 20-46

20.1

Magnetically Controlled Shape Memory Alloys

Ilkka Aaltio, Oleg Heczko, Outi Söderberg, and Simo-Pekka Hannula

20.1.1

Activation of the Shape Memory Alloys with Magnetic Field

Shape me mory a lloys ( SMAs) a re me tallic c ompounds. Ā ey exhibit a c ubic high-temperature crystal structure, which transforms by martensitic reaction into another crystal structure (tetragonal, orthorhombic, hexagonal, etc.) when temperature decreases. Martensitic t ransformation ena bles l arge s hape c hanges a nd t he 20-1

43722_C020.indd Sec1:1

10/3/2008 9:04:22 PM

20-2

Smart Materials

shape memory effect (SME), but heating and cooling are needed to complete t he phase transformation. SME can produce large relative strains and high stress output, but usually small actuation frequency d ue to t he s low c ooling p rocess [1–4]. Ā e SMAs have typically twinned martensitic structure with mobile twin boundaries. Ā e multivariant twin structure can easily be deformed to the single-variant state to yield the maximum SME shape change. To trigger the SMAs with a magnetic eld, alloys must possess magnetic o rdering, i .e., to b e f erromagnetic o r f erri/antiferromagnetic. Ā ese a lloys m ay e xhibit ( 1) m agnetic  eld-induced strain (MFIS) by magnetic shape memory effect ( MSME) o r by m agnetic  eld-induced ma rtensite t ransformation (MF-SIM), (2) magnetic eld-assisted superelasticity (MFAS), (3) conventional magnetostriction ( MS), a nd ( 4) g iant m agnetocaloric e ffect (GMCE). Ā e latter two are not d iscussed here in detail, but the other phenomena are illustrated in Figure 20.1. A list of magnetically activated magnetic SMAs is presented in Table 20.1 [5–7]. 20.1.1.1 Magnetic Shape Memory Effect In M SME, t he m artensitic t win v ariant s tructure c hanges, resulting in giant strain changes of up to 10%. Ā e origin of the effect is f undamentally d ifferent f rom MS, where t he deformation is caused by magnetization rotation relative to crystal lattice (maximum 0. 25% s train). Ā e g iant M FIS o f m agnetic s hape memory a lloys (MSMAs) is based on t he rearrangement of t he martensite twin variants by the twin boundary motion [8–13], resulting in structure reorientation and changes in direction of magnetization. Ā e MSMA(or ferromagnetic SMA, i.e., FSMA) should have a thermoelastic t winned m artensite ph ase w ith no nsymmetric lattice structure (e.g., tetragonal or orthorhombic); high magnetocrystalline a nisotropy a nd h ighly mob ile t win b oundaries [14–16]. Furthermore, a single crystal with a single twin variant structure is necessary for the maximum shape change. High magnetic anisotropy and mobile twin boundaries ensure that the magnetic eld creates a magnetic stress smag that is larger than the twinning stress stw, i.e., the stress needed for reorientaMSME

tion of the martensite structure [17–21]. Ā e magnetic stress has its maximum in the magnetic saturation, given by smag = (1−c/ a)−1Ku where Ku is the magnetic anisotropy constant and c and a are l attice c onstants o f t he m artensite s tructure. Ā e detailed mechanism c an b e e xplained w ith t he 1 0M ( also c alled 5 M) Ni–Mn–Ga alloy. In its layered structure, the short c-axis (in the cubic parent phase coordinates) is simultaneously the easy axis of m agnetization [ 22,23]. W hen t he m agnetic  eld i s app lied along t he c rystallographic a xis a nd, t hus, p erpendicularly to magnetization d irection, t he m agnetization v ector rot ates toward t he  eld ( initial l inear i ncrease o f t he m agnetization curve, Figure 20.3a) and the magnetic energy increases. When this energ y e xceeds t he energ y le vel ne eded f or i nitiating t he twin b oundary mo vement ( detwinning), t he si ngle v ariant structure (A) begins to convert to a nother variant structure (B) having i ts e asy m agnetization a xis (the s hort c-axis) a long t he applied eld and the stored energy decreases. Ā is induced reorientation is observable on the sample surface as moving twin boundaries (Figure 20.2) and it is reected in the magnetization curve as a change of initial slope (Figure 20.3a) [16,21,24]. Figure 20.2 shows a schematic representation of the MSME in actuation and for clarity, the length change is exaggerated. Ā e c-axis o f t he v ariants A a nd B i s s hown b y s hort a rrows. Ā e magnetic  eld-driven actuation sequence (H as the dashed line arrows) together w ith t he returning prestress (sreturn) is shown from left to right. If no load or magnetic eld is present (the far left a nd r ight side s), t he t win s tructure o f t he M SM elemen t remains in its single variant state, long or short. With a low stw MSMA (< 1 MPa), the required triggering eld is also low. Increasing the eld i ncreases f urther t he smag, a nd the amount of the reoriented twin variants until the saturation point i s re ached a nd t he w hole t win s tructure h as c hanged. Further i ncrease of m agnetic  eld do es not re sult i n a l arger MFIS, i.e., maximum contraction of the sample is obtained when exchanging b etween t wo si ngle m artensite v ariant s tates. Ā e maximum obtainable strain e 0 is limited by the lattice distortion given by e 0 = 1−c/a. In 10M martensite, e 0 is about 0.06 (6%) and

Magnetic-field-induced phase transformation (MF-SIM)

Martensite Martensite single-variant A Changing single-variant B magnetic field

Magnetic field assisted superelasticity (MFAS)

Martensite Stress induced by Stress magnetic field Stress-induced single-variant A (H > H ) tr H(T > Af) Parent phase martensite Unloading

Removal of the field

Stress

Martensite single-variant B

Strain

Stress

Stress

Multivariant martensite

Stress

Magnetic field on

Magnetic field off Strain

Strain

FIGURE 20.1 Ā re e different actuation principles of magnetically controlled SMAs.

43722_C020.indd Sec1:2

10/3/2008 9:04:23 PM

20-3

Shape-Memory Alloys and Effects: Types, Functions, Modeling, and Applications TABLE 20.1 Magnetic Field Activated SMAs and Ā ei r Properties Maximum MFIS and Actuation Operation Temperatures Reported

Parent Phase /Martensite Structures/ Approximate Martensitic Transformation Temperatures

Ni–Mn–Ga

6% or 10% in SC depending on the martensite crystal structure, 6% up to 333 K

Ni–Mn–Al

0.17% in SC, 0.01% in PC at 253 K

Ni–Fe–Ga Ni–Fe–Ga–Co

0.02% in SC at ∼100 K 0.7% in SC at 300 K 0.06% in SC at 165 K 0.013% in PC at 293 K 0.011% in MS ribbons 0.003% in PC at RT

L21/10M or 14M 10M up to approximately 333 K, 14M up to approximately 368 K L21/2M usually above 373 K B2 (L21)/10M 12M 14M/around RT or below B2 (L21)/ 14M or 2M/above or well above RT L21/10M 6M 14M/above and below RT

Alloy

Co–Ni–Al Co–Ni–Ga Fe3Pt

∼0.6% in SC at 4.2 K 3.1% in SC at 77 K 0.01%–0.05% in PC 0.06–0.17 in MS ribbons

Fe–Pd

B2/L10 (=2M)/below, above and well above RT B2/L10 (=2M)/below, above and well above RT Ordered fcc/fct/∼100 K Disordered fcc/fct/around RT or below

Actuation Methods

Ductility

MS, MSM, MF-SIM, MFAS, GMCE

Depends on the martensite crystal structure and SC/PC

MSM

Improved by γ phase

MS

Improved by γ phase

MF-SIM

Improved by γ phase

MF-SIM

Improved by γ phase

MSM

Intrinsically better

MSM

Intrinsically better

Sources: From Pons, J., Cesan, E., S egui, C., M asdeu, F., Santamarta, R ., Ferromagnetic SMAs: Alternatives to Ni-Mn-Ga, Mat. S ci. Eng A 481, 57, 2008 presentation in ESOMAT2006, Bochum, Germany, September 16, 2006; Chernenko, V.A., Kokorin, V.V., and Vitenko, I.N., Scripta Metall. Mater., 33, 1239, 1995; Lanska, N., et al., Appl. Phys. 95, 8074, 2004. Note: Single crystal is denoted by SC and polycrystal by PC.

in 14M (7M) about 10% [16,25–28]. In reality, some parts of the MSMA do not change and, therefore, the measured maximum strains are somewhat lower. When the magnetic eld is removed, the s hape c hange rem ains ( Figures 2 0.2 a nd 2 0.3b) u nless t he original twin structure is brought back by applying a m agnetic

eld perpendicular to original one or by an external load such as a prestress spring load applied perpendicularly to the eld direction (Figure 20.2) [13,16,17,28–31]. Ā e applied external stress reduces the obtainable MFIS and increases the eld needed to start MSM as the eld has to work

Twin variant A unit cell

Twin variant B a

c
Twin boundary s return

b (= a) s return = 0

s return

s return

s return

s return

c b s return s return = 0

Length l

s return

H =0

H =0

H = H1

H = H2 > H1

H = H3 > H2

H = Hsat

H = H2

H =0

H=0

Actuation sequence

FIGURE 20.2 Schematic representation.

43722_C020.indd Sec1:3

10/3/2008 9:04:23 PM

20-4

Smart Materials 6 1.0

5

0.5

4

Saturation magnetization Ms = 65 emu/g

Strain (%)

Relative magnetization

Onset of MSME

0.0 −0.5

−1.0 −1.5 (a)

0.2 MPa 0.5 MPa 1 MPa 2 MPa 3 MPa

3 2 1

Magnetization rotation

0 −1.0

0.5 −0.5 0.0 Magnetic field m 0 H (T)

1.0

−1.0

1.5 (b)

−0.5

0.0 0.5 Magnetic field m0H (T)

1.0

FIGURE 20.3 (a) Magnetization curve showing MSME and (b) magnetic eld induced strain as a f unction of m agnetic eld H on d ifferent compressive stresses (Ni49.7Mn 29.1Ga 21.2).

against the combination of internal stw and external mechanical stress sreturn, i .e., i n s tatic s tate smag ≥ s tw + sreturn [31 ,32]. I n actuation, the optimum applied stress sreturn returns the MSM element to i ts i nitial s tate a fter t he remo val o f t he m agnetic eld. Too small sreturn causes the element to return only partially to t he i nitial s tate, a nd to o h igh a s tress re duces t he M FIS (Figure 20.3b). If the transversal stress exceeds magnetic stress smag, t he MSME is totally suppressed [16,24]. Ā is stress limit in a linear MSME mode is called blocking stress [33,34] and above it, only the conventional MS is observed [35]. It should be noted that the MFIS change can be also other than linear, such as bending [36]. MSM-fatigue and its mechanisms are presently not very well known—the desired actuation strain strongly affects them. Sometimes there is no signicant change, sometimes MFIS increases, and in the worst case the sample cracks [37,38]. Successful actuation of 200 × 106 and 100 × 106 cycles has been reported [29,39]. On nonoptimal stress modes, secondary twin variants may appear or s tepwise app earing l attice de fects m ay re duce t he M FIS o r produce microcracks [39]. Ā e service temperature of MSME is limited by the crystallographic and magnetic transitions. MSME does not occur in the paramagnetic state (above the Curie point) or in cubic austenite. It disappears also in phase transformation to another type of martensite (intermartensitic reaction). Furthermore, this effect is de pendent on t he c hanges of l attice pa rameters, el astic c onstants, t he t winning stress, magnetic a nisotropy, a nd magnetization with temperature [39–42]. Decrease of temperature increases stw, which eventually exceeds smag and MSME is blocked, for e xample, i n 10M Ni– Mn–Ga s omewhere b etween 1 53 a nd 253 K [41,43]. 20.1.1.2 Magnetic Field-Assisted Superelasticity Ā e magnetically assisted superelasticity (MFAS) of alloys occurs in the same temperature region as MSME as this effect is complementary to MSME [44]. By deforming single-variant martensite a long t he lo ng c rystallographic a xis, t he t win s tructure

43722_C020.indd Sec1:4

converts to the variant with the short axis in the stress direction. Application of a c onstant m agnetic  eld p erpendicular to t he stress direction tends to convert the stress-induced variant back to the initial state (short axis along the eld), when the external stress is removed. If the applied eld is strong and twin mobility high eno ugh, t he v ariant re orientation i s f ully re versible. I n a constant eld ≥0.6 T with the 10M Ni–Mn–Ga a strain of 6% can be obtained with less than 8 MPa external stress [31,44], while with 14M s tructure, 10% i s p ossible [19]. Ā is behavior c an be used f or de termining t he ma terial pa rameters n eeded i n t he MSM de vice de sign b y me asuring s tress–strain b ehavior i n a magnetic eld [21,31,45–49]. 20.1.1.3

Magnetically Activated Phase Transformation

When the stress-induced martensite formation (i.e., superelasticity) close to t he austenite–martensite t ransformation temperature i s induced by the magnetic  eld, it is called magnetically activated phase t ransformation ( MFSIM). Ā e effect o riginated f rom t he magnetic e nergy diff erence b etween a ustenite a nd m artensite phases in a magnetic  eld, which creates the stress triggering the SIM f ormation [1,48,50–55]. I n Ni– Mn–Ga, t his e ffect requires usually a large magnetic eld in the order of tens of teslas, but very close to the phase transformation temperatures where the needed eld may be smaller. Also, in some other alloys, such as Ni–Mn–In, 3–5 T i s eno ugh. A ctuation i s ob tained b y t he re peating ph ase transformations when applying and removing the magnetic eld. SIM-type behavior is typical for Co–Ni-based alloys (Co-Ni [52], Co–Ni–Ga [56], Co–Ni–Al [57], etc.). 20.1.1.4

Giant Magnetocaloric Effect

In addition to the dilatational changes triggered by the magnetic eld, some of the ferromagnetic SMAs also show a GMCE [58–60]. Ā is phenomenon c an b e u sed for h igh-temperature refrigeration [61]. To gain this behavior, the alloy should have the s tructural a nd m agnetic t ransitions v ery c lose to e ach other. In GMCE material produces or absorbs heat while being

10/3/2008 9:04:23 PM

20-5

Shape-Memory Alloys and Effects: Types, Functions, Modeling, and Applications

heated or cooled through the transition temperature region in a m agnetic  eld. Ā e m agnetocaloric e ffect c an b e e stimated from the isothermal magnetization curves at different temperatures [58,60,62–64].

MFI force: Fmag Push rod

Fext Freturn

20.1.2

Production of MSMAs

MSMAs ha ve t o m eet spec ic c haracteristics de scribed p reviously. Ā ere a re not m any a lloys s atisfying t hese c onditions. Microstructure and chemical composition must be correct as well as t he si ngle c rystalline s tructure. Fu rthermore, t he app lication specications re quire t ailoring o f t he M SMAs. Ā e b ulk M SM materials a re produced by t he si ngle c rystal g rowth te chniques, such as Bridgman [43,65,66] or Czochralski [67]. Optimum crystal o rientation o f t he i ngots i ncreases t he y ield. A fter casting, homogenizing and ordering heat treatments are needed. Ā en the annealed s amples a re c ut, g round, a nd p olished to t he de sired dimensions of the MSM element. In addition to the bulk samples, MSMAs have been produced in research scale in forms of thin lms [68–72], ribbons [73,74], and composites [75].

20.1.3 Applications MSMAs can be used for actuating applications, i.e., to produce movement, f or s ensor o r p ower gener ation app lications, a nd for v ibration d amping. P resently, t he mo st c ommonly u sed alloys a re 1 0M Ni– Mn–Ga; h owever, t he 1 4M Ni– Mn–Ga alloys have been studied to broaden MSM region to the higher temperatures. 20.1.3.1 Actuating Ā e main advantages of MSM actuators are their short response time, large relative strain and stroke, and proportional controllability. Furthermore, the MSM element itself does not have to be wired. S o f ar, t he d isadvantages a re t he l imited op erating temperature, relatively low force output and hysteresis. Basic parts of a t ypical linear MSM actuator are the piece of MSMA, i .e., t he M SM elemen t, t he m agnetic c ircuit a nd t he body. T he d irection o f m agnetic f ield i s o rthogonal to t he direction of the induced strain (Figure 20.2). Ā e variable eld, either a rotating magnetic eld (no returning force needed) or a transversal  eld ( assisted b y t he me chanical re turning f orce, Figure 20.4), is created by an electromagnetic circuit. Ā e magnetic circuit concentrates t he  eld to t he space of t he MSM element. For an actuator using 10M Ni–Mn–Ga, the necessary maximum H i s app roximately 5 00 k A/m [29]. Ā e p roduced f orce o f t he actuator de pends ba sically o n t he c ross-sectional a rea a nd the stroke on the length of the MSM element, both depending on the MSMA properties [76]. In a design similar to Figure 20.4, the s pring applies u sually a p restress of 0. 5–1 M Pa, w hich t he MSM element has to work against. MSM actuators are best driven with current, because there is a delay in magnetization of voltage-controlled coils [43]. By proper electromagnetic design, unipolar current can be used. With a bias magnetic  eld ob tained w ith p ermanent m agnets, sm aller p eak

43722_C020.indd Sec1:5

MSM element

H Coils

FIGURE 20.4 Cross-sectional scheme of a n MSM actuator, showing the b asic p arts a nd t he d irections of t he m agnetic  eld H, M FI forc e Fmag, returning spring force Freturn, and external load Fext.

power and smaller coils can be used. When the actuation frequency increases, more current is needed, due to, for example, skin effect and the generated eddy currents limiting the highest frequency performance. Overheating can lead to the reverse transformation, but the actuating element recovers when it cools back to martensite range; however, overheating is not recommended [77]. Ā e ac tuators e xhibit a h ysteresis due to t he t win b oundary motion, which causes a lag when reversing the actuation direction. Because stw is usually constant in actuator use, the hysteresis can be compensated by a suitable control system. Hysteresis prevents also the overshoot and drift of the actuator [24,29]. Ā e efficiency of M SM ac tuators de pend s trongly o n t he p roperties o f t he MSMA such as smag and stw [23,78] and on the electromagnetic design. M SM ac tuators c an b e u sed, f or e xample, to p roduce proportional l inear bac k-and-forth mot ion, i nchworm t ype linear moto rs, p roportional v alves, a nd p umps [36]. I n sm allsized M SM ac tuators, re sponse t imes a re s hort. A 0. 2 m s r ise time in small actuators can be reached [28]. Increase of the operating temperature of MSM alloys would extend the eld of applications, such as in automotive engines, where the fast response and relatively large strains provide advantages [79,80]. 20.1.3.2 Sensing and Power Generation MSMA sensors may be constructed using the inverse MSME: When the material shape changes, the effective magnetization curve of the material is altered [16,44]. Ā is effect c an b e u sed to mo nitor t he shape o f t he m aterial b y me asuring i ts m agnetization [ 81,82]. I n addition to s ensing, p ower gener ation b y t he i nverse M SME h as been dem onstrated [ 83]. W hen a n el ement o f N i50.1Mn28.4Ga21.5 MSMA was compressed at a v elocity of 1.5 m /s, a vol tage of over 60 V was generated using a simple electromagnetic circuit. 20.1.3.3

Damping

Vibration damping mechanisms that may operate in magnetically controlled S MAs a re ba sed o n t he h ysteretic mot ion o f t he

10/3/2008 9:04:24 PM

20-6

martensite twin boundaries, motion of magnetic domain walls, or stress-induced martensite formation. Ā e m ain da mping mechanism i s t he d issipative mot ion o f t he t win b oundaries. Due to t he phase transformations, the internal friction and the elastic mo dulus o f Ni– Mn–Ga i s tem perature de pendent [ 84– 86]. Ā e d amping elemen ts c an b e c onstructed a s c omposite overcoming intrinsic brittleness of the MSMA. Enhanced damping w as ob served i n c omposite u sing M SMA p owder [ 87]. However, d amping of MSMA is presently a sub ject of research and no commercial products use this technology.

References 1. Otsuka, K. a nd Wayman, C.M., Shape Me mory Ma terials. Cambridge, UK: Cambridge University Press, 1998, 284pp. 2. Duerig, T.W., Engineering Aspects of Shape Memory Alloys. London: Butterworth-Heinemann, 1990, 499pp. 3. Friend, C., S hape memo ry allo ys, in Encyclopedia of Materials: S cience a nd T echnology, Am sterdam: Elsevier Science, 2001, pp. 8445–8451. 4. Quandt, E., Actuator ma terials f or small s cale de vices, in Encyclopedia of M aterials: S cience and T echnology, Amsterdam: Elsevier Science, 2001, pp. 35–38. 5. Pons, J., et al ., Ferromagnetic shape memory alloys: Alternatives to Ni-Mn-Ga, Mat. Sci. Eng A 481, 57, 2008. 6. Chernenko, V .A., Kokorin, V .V., and V itenko, I.N., Ā e development o f ne w f erromagnetic shape memo ry allo ys in Ni-Mn-Ga system, Scripta Metall. Mater. 33, 1239, 1995. 7. Lanska, N., et al., Composition and temperature dependence of the cr ystal structure in Ni-Mn-Ga alloys, Appl. Phys. 95, 8074, 2004. 8. Rhyne, J.J., et al ., R are earth met al single cr ystals. I. Higheld properties of D y, Er, Ho, Tb, and Gd, J. Appl. Phys., 39, 892, 1968. 9. Liebermann, H.H. and Graham C.D., Plastic and magnetoplastic deformation of Dy single crystals, Acta Met., 25, 715, 1977. 10. Ullakko, K., Magnetically controlled shape memory alloys: A ne w class of ac tuator materials, J. Mat. Eng. Perform., 5, 405, 1996. 11. Ullakko, K., et al ., M agnetically co ntrolled sha pe memo ry effect in Ni2MnGa intermetallics, Scr. Mater., 36, 1133, 1997. 12. O’Handley, R .C. a nd Allen, S.M., S hape memo ry allo ys, magnetically activated ferromagnetic shape-memory materials, in Encyclopedia of Smar t M aterials, N ew York: J ohn Wiley and Sons, 2002, pp. 936–951. 13. Söderberg, O ., et al ., G iant m agnetostrictive m aterials, i n Handbook of Magnetic Materials, Vol. 16, Buschow, K.H.J., ed., Amsterdam: Elsevier, 2006, pp. 1–39. 14. Heczko, O., Sozinov, A., and Ullakko, K., Giant eld-induced reversible strain in magnetic shape memory NiMnGa alloy, IEEE Trans. Magn., 36, 3266, 2000. 15. Likhachev, A.A. and Ullakko, K., Magnetic-eld-controlled twin b oundaries mo tion a nd gia nt ma gneto-mechanical effects in Ni-Mn-Ga shape memory alloy, Phys. Lett. A, 275, 142, 2000.

43722_C020.indd Sec1:6

Smart Materials

16. Likhachev, A.A. a nd U llakko, K., Ā e mo del o f ma gneticeld-controlled shape memory effect in NiMnGa, J. Phys. IV Proc., 11, Pr8/293, 2001. 17. Sozinov, A., et al ., G iant ma gnetic-eld-induced stra in in NiMnGa seven-layered martensitic phase, Appl. Phys. Lett., 80, 1746, 2002. 18. Lik hachev, A.A., S ozinov, A., a nd U llakko, K., M agnetomechanical cycling and modeling the ext ernal stress effect on the ma gnetic-eld-controlled stra in r esponse in Ni-Mn-Ga, J. Phys. IV Proc., 115, 95, 2004. 19. Heczko, O ., S traka, L., a nd U llakko, K., Rela tion b etween structure, ma gnetization p rocess a nd ma gnetic sha pe memory eff ect o f va rious ma rtensites o ccurring in Ni-Mn-Ga alloys, J. Phys. IV Proc., 112, 959, 2003. 20. Heczko, O ., et al ., T emperature dep endence o f ma gnetic anisotropy in N i-Mn-Ga alloy exhibiting giant  eld-induced strain, J. Appl. Phys., 91, 8228, 2002. 21. Ge, Y., et al ., Crystal structure of three Ni-Mn-Ga alloys in powder and bulk materials, J. Phys. IV Proc., 112, 921, 2003. 22. Straka, L. and Heczko, O., Magnetic anisotropy in Ni-Mn-Ga martensites, J. Appl. Phys. 93, 8636, 2003. 23. Heczko, O., Magnetic shape memory effect and magnetization reversal, J. Magn. Magn. Mater, 290–291, 787, 2005. 24. S uorsa, I., P erformance a nd mo deling o f ma gnetic sha pe memory actuators and sensors, TKK Dissertations 4, Helsinki University of Technology, Espoo, Finland, 2005, 70pp. 25. James, R.D., Tickle, R., and Wuttig, M., Large eld-induced strains in ferromagnetic shape memory materials, Mat. Sci. Eng. A, 273–275, 320, 1999. 26. O’Handley, R.C., et al ., Phenomenology of giant magneticeld-induced strain in ferromagnetic shape-memory materials (invited), J. Appl. Phys., 87, 4712, 2000. 27. Likhachev, A.A., Sozinov, A., and Ullakko, K., Modeling the strain r esponse, ma gneto-mechanical c ycling under the external stress, work output and energy losses in Ni-Mn-Ga, Mech. Mater., 38, 551, 2006. 28. Aaltio, I. and Ullakko, K., Magnetic shape memory (MSM) actuators, B orgmann, H., e d., P roc. ACTUATOR 2000, Messe Bremen, Germany, 2000, pp. 527–530. 29. Tellinen, J., et al., Basic properties of magnetic shape memory actuators, B orgmann, H., e d., P roc. o f ACTUATOR 2002, Messe Bremen, Germany, pp. 566–569. 30. Murray, S.J ., et al ., 6% M agnetic-eld-induced stra in b y twin-boundary motion in f erromagnetic Ni-Mn-Ga, Appl. Phys. Lett., 77, 886, 2000. 31. Straka, L. and Heczko, O., Reversible 6% strain of Ni-Mn-Ga martensite using opposing external stress in static and variable mag netic  elds, J. Ma gn. Ma gn. Ma ter., 290–291, 829, 2005. 32. Likhachev, A.A., S ozinov, A., a nd U llakko, K., Op timizing work output in Ni-Mn-Ga and other ferromagnetic shapememory alloys, Lynch, C.S., ed., Proc. SPIE, 4699, 553, 2002. 33. Murray, S.J., et al ., G iant ma gnetic-eld-induced stra in in Ni-Mn-Ga cr ystals: E xperimental r esults a nd mo deling, J. Magn. Magn. Mater., 226–230, 945, 2001.

10/3/2008 9:04:24 PM

Shape-Memory Alloys and Effects: Types, Functions, Modeling, and Applications

34. Malla, A., et al., Eff ect of composition on the magnetic and elastic p roperties o f shape-memory Ni-Mn-Ga, L agoudas, D.C., ed., Proc. SPIE, 5053, 147, 2003. 35. Heczko, O., Determination of ordinary magnetostriction in Ni-Mn-Ga ma gnetic sha pe memo ry allo y, J. Ma gn. Ma gn. Mater., 290–291, 846, 2005. 36. Suorsa, I., et al ., Applications o f ma gnetic sha pe memo ry actuators, B orgmann, H., e d., P roc. ACTUATOR 2002, Bremen Messe, Germany, pp. 158–161. 37. Müllner, P., Chernenko, V.A., and Kostorz, G., L arge c yclic magnetic-eld-induced def ormation in o rthorhombic (14M) Ni-Mn-Ga martensite, J. Appl. Phys., 95, 1531, 2004. 38. Heczko, O ., et al ., M agnetic sha pe memo ry fa tigue, Armstrong, W.D., ed., Proc. SPIE 5761, 513, 2005. 39. Müllner, P., et al., Nanomechanics and magnetic structure of orthorhombic Ni-Mn-Ga martensite, Mat. Sci. Eng A 481, 66, 2008. 40. Glavatska, N., et al., Temperature dependence of martensite structure and its eff ect on magnetic-eld-induced strain in Ni2MnGa magnetic shape memory alloys, J. Phys. IV Proc., 112, 963, 2003. 41. Heczko, O . a nd S traka, L., T emperature dep endence a nd temperature limits of magnetic shape memory effect, J. Appl. Phys., 94, 7139, 2003. 42. Khovailo, V.V., et al., Magnetic properties and magnetostructural phase transitions in Ni2+ xMn1–xGa shape memory alloys, J. Phys.: Cond. Matter, 1–7, arXiv:cond-mat/0405134, 2004. 43. Pagounis, E. a nd Qua ndt, E., Re cent ad vances a nd c hallenges in magnetic shape memory materials, Borgmann, H., ed., P roc. ACTUATOR 2006, M esse B remen, G ermany, pp. 394–400. 44. Straka, L. and Heczko, O., Superelastic response of Ni-Mn-Ga martensite in mag netic  elds a nd a sim ple mo del, IEEE Trans. Magn., 39, 3402, 2003. 45. Müllner, P., Chernenko, V.A., and Kostorz, G., A microscopic approach t o the ma gnetic-eld-induced def ormation o f martensite (magnetoplasticity), J. Magn. Magn. Mater., 267, 325, 2003. 46. Chernenko, V.A., et al., Magnetic-eld-induced superelasticity of ferromagnetic thermoelastic martensites: Experiment and modeling, Phys. Rev. B, 69, 134410, 2004. 47. Suorsa, I. and Pagounis, E., Magnetic eld-induced stress in the Ni-Mn-Ga magnetic shape memory alloy, J. Appl. Phys., 95, 4958, 2004. 48. Sozinov, A., et al ., Stress- and magnetic-eld-induced variant rearrangement in Ni-Mn-Ga single crystals with sevenlayered ma rtensitic str ucture, Mater. S ci. E ng. A, 378, 399, 2004. 49. Likhachev, A.A., S ozinov, A., a nd U llakko, K., Diff erent modeling co ncepts o f mag netic sha pe memo ry a nd their comparison wi th so me e xperimental r esults o btained in Ni-Mn-Ga, Mater. Sci. Eng. A, 378, 513, 2004. 50. Kakeshita, T., Saburi, T., and Shimizu, K., Eff ects of hydrostatic pressure and mag netic  eld on mar tensitic transformations, Mater. Sci. Eng. A, 273–275, 21, 1999.

43722_C020.indd Sec1:7

20-7

51. Hamilton, R.F., et al ., Ā ermal and stress-induced martensitic t ransformations in N iFeGa sin gle cr ystals under tension and compression, Scripta Mat., 54, 465, 2006. 52. Liu, Y., et al., Magneto-shape-memory effect in Co-Ni single crystals, Appl. Phys. Lett., 78, 3660, 2001. 53. Li, Y.X., et al ., M agnetic  eld-controlled tw o-way sha pe memory in C oNiGa sing le cr ystals, Appl. P hys. L ett., 84, 3594, 2004. 54. Chernenko, V.A., et al., Ā ermal and magnetic properties of stress-induced martensites in Ni-Mn-Ga alloys, J. Phys. IV, 5, C2/95, 1995. 55. Vasil’ev, A., et al., Structural and magnetic phase transitions in sha pe memo ry a lloys N i2+xMn1–xGa, J. Ma gn. Ma ter., 196–197, 837, 1999. 56. Liu, J ., et al ., Micr ostructure e volution in C oNiGa sha pe memory alloys, J. Alloys Comp., 420, 145, 2006. 57. Hamilton, R .F., et al ., P seudoelasticity in C o-Ni-Al sin gle and polycrystals, Acta Mat., 54, 587, 2006. 58. Pakhomov, A.B., et al ., Magnetization a nd ma gnetocaloric effect in ma gnetic sha pe memo ry allo ys N i-Mn-Ga, IEEE Trans. Magn., 37, 2718, 2001. 59. Marcos, J., et al., Magnetic eld induced entropy change and magnetoelasticity in N i-Mn-Ga allo ys, Phys. Re v. B , 66, 224413, 2002. 60. Hu, F.-X., et al., Magnetic entropy change in Ni50.1Mn20.7Ga29.6 single crystal, J. Appl. Phys. 90, 5216, 2001. 61. Zhou, X., et al., A criterion for enhancing the giant magnetocaloric eff ect: (Ni-Mn-Ga)—a promising new system for magnetic refrigeration, J. Phys. Cond. Matter, 16, L39, 2004. 62. Zhou, X., et al., Inuence of the nature of the magnetic phase transition o n the ass ociated ma gnetocaloric eff ect i n th e Ni-Mn-Ga system, J. Magn. Magn. Mater., 293, 854, 2005. 63. Marcos, J ., et al ., M ultiscale o rigin o f the mag netocaloric effect in N i-Mn-Ga shape-memory alloys, Phys. Rev. B, 68, 094401, 2003. 64. Marcos, J., et al., M agnetic  eld ind uced en tropy c hange and magnetoelasticity in Ni-Mn-Ga alloys, J. Magn. Mater., 272–276, E1595, 2004. 65. Schlagel, D .L., et al ., Chemical s egregation d uring b ulk single cr ystal p reparation o f N i-Mn-Ga f erromagnetic shape memory alloys, J. Alloys Compounds, 312, 77, 2000. 66. Söderberg, O., et a l., R ecent bre akthrough development of the ma gnetic sha pe memo ry eff ect in N i-Mn-Ga allo ys, Smart Mater. Struct., 14, 223, 2005. 67. Liu, G., et al., Characterization of preferential orientation of martensitic va riants in a sin gle cr ystal o f N iMnGa, Solid State Commun., 130, 687, 2004. 68. Suzuki, M., et al ., Fabrication a nd c haracterization o f sp uttered Ni2MnGa thin lms, Mater. Trans. JIM, 40, 1174, 1999. 69. Ahn, J ., et al ., M agnetic p roperties, str ucture a nd sha pememory transitions in Ni-Mn-Ga thin  lms grown by ionbeam sputtering, IEEE Trans. Magn., 37, 2141, 2001. 70. Hakola, A., et al ., N i-Mn-Ga  lms on Si, G aAs and Ni-Mn-Ga single crystals by pulsed laser deposition, Appl. Surf. Sci., 238, 155, 2004.

10/3/2008 9:04:24 PM

20-8

71. Kohl, M., et al ., A novel actuation mechanism on the basis of ferromagnetic SMA thin lms, Sensors Actuators A Phys., A114, 445, 2004. 72. Dong, J.W., et al ., Shape memory and ferromagnetic shape memory eff ects in sin gle-crystal N i2MnGa t hin  lms, J. Appl. Phys., 95, 2593, 2004. 73. Heczko, O., et al., Magnetic properties of Ni-Mn-Ga ribbon prepared b y ra pid s olidication, IEEE T rans. M agn., 38, 2841, 2002. 74. Algarabel, P .A., et al ., M agnetic-eld-induced stra in in Ni2MnGa mel t-spun ribb ons, J. Ma gn. Ma ter., 272– 276, 2047, 2004. 75. Hosoda, H., et al., Material design and shape memory properties o f sma rt co mposites co mposed o f p olymer a nd ferromagnetic sha pe memo ry a lloy p articles, Sci. T echn. Adv. Mat., 5, 503, 2004. 76. Suorsa, I., et al., Design of active element for MSM actuator, Borgmann H., ed., Proc. ACTUATOR 2004, Bremen Messe, Germany, pp. 573–530. 77. Xiong, F., et al., Ā ermally induced fracture of single crystal Ni-Mn-Ga f erromagnetic sha pe memo ry allo y, J. A lloys Comp., 415, 188, 2006. 78. Tellinen, J., Efficiency of M SM ma terial, tech nical notes , Adaptamat Ltd., unpublished. 79. Pagounis, E., Magnetic shape memory actuators for motion control applications, Eur. Design Eng., 96, 2004. 80. Jokinen, T ., U llakko, K., a nd S uorsa, I., M agnetic sha pe memory ma terials—new p ossibilities t o cr eate f orce a nd movement b y mag netic  elds, P roc. I CEMS 2001, 1, 20, 2001. 81. Müllner, P., Chernenko, V.A., and Kostorz, G., Stress-induced twin rearrangement resulting in change of magnetization in a Ni-Mn-Ga ferromagnetic martensite, Scr. Mater., 49, 129, 2003. 82. Heczko, O . a nd S traka, L., M agnetization c hanges in Ni-Mn-Ga ma gnetic sha pe memo ry sin gle cr ystal d uring compressive st ress r eorientation, Scr. M ater., 54, 1549, 2006. 83. Suorsa, I., Tellinen, J., Ullakko, K., and Pagounis, E., Voltage generation ind uced b y me chanical st raining in mag netic shape memory materials, J. Appl. Phys., 95, 8054, 2004. 84. Cesari, E., et al., Internal friction associated with the structural p hase tra nsformations in N i-Mn-Ga allo ys, Acta Mater., 45, 999, 1997. 85. Chernenko, V.A., et al., Sequence of martensitic transformations in Ni-Mn-Ga alloys, Phys. Rev. B, 57, 2659, 1998. 86. Gans, E., Henry, G., and Carman, G.P., High energy absorption in b ulk fer romagnetic sha pe memo ry a lloys (Ni50Mn29Ga21), L agoudas, D .C., e d., P roc. S PIE, 5387, 2004. 87. Feuchtwanger, J ., et al ., Ener gy a bsorption in N i-Mn-Gapolymer composites, J. Appl. Phys., 93, 8528, 2003.

43722_C020.indd Sec1:8

Smart Materials

20.2

SMAs in Micro- and Nanoscale Engineering Applications

Yves Bellouard 20.2.1 Introduction Microsystems (also r eferred t o a s m icroelectromechanical systems [MEMS]) have ourished during the past decades to rapidly penetrate a b road r ange o f app lications l ike a utomotives, c onsumer products, and biomedical devices. Ā ey perform sophisticated tasks in a miniaturized volume. Common operations such as shaping or analyzing light signals, mixing, processing, or analyzing ultrasmall volumes of chemicals, sensing mechanical signals, probing gas, sequencing biomolecules are realized by these tiny machines. Ā ere are several advantages of the using microsystems, w hich i nclude t he re duction o f c onsumables ( for instance, t he u se o f le ss c hemicals i n l ab-on-a-chip), a f aster response time (like for airbag sensors), an enhanced portability (RF-MEMS), higher printing resolution (ink-jet head), higher efficiency for microchemical reactor, among others. For a ll t hese app lications, ne w ac tuating p rinciples t hat a re adapted to t he m icroscale de vices a re ne eded. I n t his c ontext, SMAs have been considered early on. Over the last decade, the fabrication of m icrodevices s uch a s m icrogrippers [ 1–4], microvalves a nd m icropumps [ 5–8], s pacers [ 5], ac tuated microendoscope [9], ner ves c lamp [11], t actile d isplays [10], to name a few, have been reported. Ā is chapter provides a critical overview on the use of SMA materials f or m icrosystems: A b rief i ntroduction o n S MA i s  rst p rovided f ollowed b y a re view o f S MA m icrosystems designing methods. Finally, t he chapter concludes w ith a d iscussion on future prospects. Specic issues related to material processing and in particular thin- lm processing are not addressed. Details on these particular aspects can be found, for instance, in Refs. [2,12–14].

20.2.2

SMAs: A Brief Introduction from a Microsystem Design Prospective

SMA a re i ntermetallics* t hat h ave t wo ( or s ometimes s everal) crystallographic phases† for which reversible transformations from one to t he ot her o ccur t hrough d iffusionless‡ tr ansformations (the so-called reversible martensitic transformations [15]). Figure 2 0.5 br iey i llustrates m ultiple c rystallographic phases in the case of a binary alloy like Ni–Ti. At low tempera-

* Intermetallics are alloys made of two or more metallic components forming a ne w c ompound t hat h ave d ifferent pr operties t hat it s c onstituents metals taken individually. † Each phase is thermodynamically stable for a certain domain of temperature and pressure conditions. ‡ During t he t ransformation, at oms c ollectively move s a long s ubatomic distances so that atomic liaisons remain unchanged although the crystal lattice is distorted.

10/3/2008 9:04:25 PM

20-9

Shape-Memory Alloys and Effects: Types, Functions, Modeling, and Applications

Critical stress to

induce slip (plas

ticity)

zo n

sit io

Austenite

sit io

n

an Tr

an

Tr

Critical stress for detwinning

e

n

(H

zo n

ea

e

tin

g)

(c oo

lin

Stress

g)

Martensite

T = Ms T = As

T = Mf

T = Af

T = Md

Temperature

FIGURE 20.5 Two crystallographic phases of a Shape Memory Alloy represented in a stress/temperature phase diagram.

Ā e SME refers to the ability of the material, initially deformed in i ts low-temperature ph ase (called m artensite), to re cover i ts original s hape up on h eating to i ts h igh-temperature ph ase (called a ustenite o r pa rent ph ase). Ā e S ME i s a m acroscopic effect o f t hermally i nduced c rystallographic ph ase c hanges. A very i mportant a spect i s t hat i t i s a o ne-way o ccurrence. Ā is phenomenon is illustrated in Figure 20.7. Let us consider an SMA material initially in its parent phase (b) (step 3 in Figure 20.7). On cooling, the material transforms to the martensitic phase (step 2). From a c rystallographic point of view, this phase is characterized by a lo wer symmetry t han t he pa rent phase and therefore has different possible crystallographic orientations (commonly called as variants). Multiple martensitic variants with diff erent crystallographic orientations (for instance A, B, C, and D i n F igure 2 0.7) n ucleate s o t hat t he de formation s train

Stress

ture a nd lo w p ressure, t he a lloy h as i ts lo w-crystallographic phase called martensite. At higher temperature, the alloy has a different crystallographic s tructure of h igher s ymmetry. Two remarkable effects are related with the phase change: the SME and superelasticity. A s t hese e ffects h ave b een e xtensively documented elsewhere (see for instance Refs. [15–17] as well as the paragraph 20.4 of this Chapter), we just brief ly summarize their main features f rom a m icrosystems de sign prospective. Superelasticity o ccurs w hen t he m artensitic ph ase t ransformation is stress-induced at a constant temperature. Ā e transformation i s c haracterized b y a p lateau a nd a h ysteresis up on unloading. A t ypical lo ading–unloading c urve i s s hown o n Figure 2 0.6. Ā e m agnitude o f re versible ps eudoelastic s train can be as high as 8% or even more for single crystals.

90

Mf

Ms As Af

1

Transformation temperatures

Temperature

2

3

Loading

Pa

Pa 10

Lo

0M

ad

Pa

0G

ing

0M 45

Austenite 4

Shape memory alloy

E-7

din loa

2 Martensite

Un

Stress

g

0M Pa

3

Stainless steel

E-200 GPa

Elastic limit

Unloading 4

1 1%

8% Strain

FIGURE 20.6 Left : Superelastic loading represented i n t he stress/temperature phase d iagram-note t hat t he loading a nd u nloading occur at constant stress during transformation, a characteristic feature not captured by the stress/temperature phase diagram. Right: Typical tensile test curve of a superelastic material compared to the one of a stainless steel.

43722_C020.indd Sec1:9

10/3/2008 9:04:25 PM

20-10

Smart Materials D B C A

A

A

A third method is based on monolithic design. Ā is method is a combination of intrinsic and extrinsic methods.

s D B

3

20.2.4 Intrinsic Methods: Tailoring the Microstructure

A C

e0

2

Martensite 1

20.2.4.1 Two-Way SME

e

4

e0

b T

Austenite

FIGURE 20.7 Illustration of the SME (one-way) (the small sketches illustrative of the material structure one adapted from [18].

energy is minimized (step 1 to step 2). Ā ese particular variants are called s elf-accommodating a nd form t win pa irs. From a mac roscopic point of view, at this stage, there is no change of shape as the volume globally remains the same although the crystallographic structure of the material has changed. If a stress is applied on the material (step 3 in Figure 20.7), variants that have a favorable orientation related to the stress grow at the expense of less favorably oriented ones. Ā is effect is called detwinning as self-accommodating variants ( that w ere s omehow de ning ma ting pa irs) d isappear. Once the stress is released (step 4), as the newly formed martensite variants a re t hermodynamically s table, t he m aterial re tains t he applied deformation (e0 in Figure 20.7). Upon heating above the martensitic phase transformation temperatures, the parent phase, which h as a h igher s ymmetry f rom a c rystallographic p oint o f view, nucleates and progressively replaces the martensite, causing the disappearance of the apparent deformation (e0). Ā is ability to recover i ts o riginal s hape ( i.e., t he s hape c orresponding to t he parent phase) is called the SME. Cooling again with no applied stress, self-accommodating variants will form (step 2) and no macroscopic change of volume will be observed. From a macroscopic shape change point of view, the SME is therefore not i ntrinsically reversible: A change of shape is only observed if nonself-accommodating variants have nucleated. Ā e design objective of an SMA actuator is to achieve a reversible m acroscopic s hape c hange: F rom a m icrostructure p oint o f view, it is equivalent to have nonself-accommodating martensitic variants nucleating on cooling. In the next section, we review various approaches that have been used to fulll this design objective.

20.2.3

SMA Microactuators: Design Principles

To provide t he necessary reversible SME, t wo approaches have been explored: intrinsic and extrinsic methods. Intrinsic me thods c onsist o f mo difying t he m aterial m icrostructure s o t hat c ertain m artensitic v ariant o rientations w ill preferably nucleate upon cooling. Extrinsic methods refer to the addition o f a n e xternal elemen t c oupled to t he S MA m aterial that provides the required stress to induce stress-oriented variants.

43722_C020.indd Sec1:10

Ā e m artensitic v ariants n ucleation i s i nuenced b y o riented precipitates and defects like dislocations present in the material lattice s tructures. O riented de fects le ad to w hat i s c ommonly called the “two-way shape memory effect” (TWSME). Different thermomechanical p rocesses i nducing a T WSME h ave b een identied [ 18,19] l ike s evere de formation o f t he m artensitic phase, t hermomechanical c ycling u nder c onstraints, a nd i sothermal mechanical c ycling in t he austenite phase. Ā e se thermomechanical processes that led to the TWSME are sometimes called t raining p rocesses to c onvey t he ide a t hat t he m aterial progressively memorizes a new shape. Ā is f ascinating b ehavior i s i llustrated i n F igure 2 0.8: A TWSME m icrogripper i s s hown [ 4,21]. Ā e millimeter-size device c onsists o f a si ngle p iece o f me tal l aser-cut f rom a 1 80 µm-thick Ni–Ti–Cu c old-rolled s heet. I t i s u sed b y a me dical instrument manufacturer to assemble submillimeter cylindrical lenses. Ā e gripper design consists of a  nger that can move in the plane to press an object against a surface. Ā e working principle is illustrated in Figure 20.8 (operating mode): Upon cooling, t he g ripper ja w op ens a nd c loses o n h eating. Ā e h eat i s provided b y a si mple ele ctrically re sistive l ayer de posited o n a ceramic substrate on which the gripper is glued to. Ā e gripper parent shape (the memorized shape) is the as-cut shape. At this point, t he de vice do es not s pontaneously ( i.e., w ithout a ny applied stress) undergo change of shape. To achieve the reversible  nger mot ion, t he g ripper is deformed a nd constrained so that it cannot recover its original shape (step 2 i n Figure 20.6). About 100 thermal cycles are applied between the austenite and martensite phases. During thermal cycling, stress builds up in the  nger as oriented defects are created. Once this sequence is complete, t he gripper is operational a nd a s pontaneous change of shape is observed. Ā e g ripper app lies f orces o f a bout 1 5 m N, w hich i s h uge compared to t he weight of the device. Time response for the jaw to op en a nd c lose de pends o n t he h eating a nd c ooling r ates. Typically, it varies from a few hundreds of milliseconds to a second. Ā e m icrogripper s hows e xcellent f atigue p roperties (>200,000 c ycles) [21] a nd s tability w ith re spect to me chanical perturbations [22]. Remarkable e ffects a re a ssociated w ith t he s tability o f t he TWSME [22,23]. After the training process, nonself-accommodating variants nucleate upon cooling and lead to a macroscopic shape c hange (the s econd memorized s hape). Yet, t he m aterial can still be deformed isothermally in martensite so that another metastable conguration of variants is introduced. Ā is martensite m icrostructure d oes not a ppear s pontaneously on c ooling but by mechanically deforming the material from its martensitic

10/3/2008 9:04:25 PM

20-11

Shape-Memory Alloys and Effects: Types, Functions, Modeling, and Applications

2

1

500 µm Acc.V

10.0 kV EPFL-ISR-IGA-10A (0.0197 Inch)

Cut shape

T

Operating mode 3

Heating

Cooling Martensitic shape

Austenitic shape

FIGURE 2 0.8 (Left) S canning e lectron m icroscope pic ture of t he T WSME m icrogripper for s ubmillimeter le ns m icromanipulation. (From Bellouard, Y., PhD dissertation, Lausanne, 2000.) (R ight) Ā e desired shape is laser-cut (1) and thermomechanically cycled under constraint (training process) (2). Ā e working principle is also shown (3) (scale bar is 500 µm).

shape. It is therefore a prepared state, articially introduced that puts the material in a sort of out-of-equilibrium state, i.e., outside the sequence of spontaneous shape changes introduce by the TWSME. Ā e material response to this perturbation is quite spectacular and is shown for a c antilever beam in Figure 20.9. Ā e cantilever beam is made out of a ternary alloy (Ni–Ti–Cu (5% at.) annealed at 515°C for 30 min) [23]. Ā e gure shows the relative displacement of t he c antilever. F irst, t he c antilever i s de formed i n m artensite (prepared state). Upon heating, the deformation slowly disappears until the phase transformation temperatures are reached. (We call this  rst seq uence a pproach-to-equilibrium.) A t t his po int, t he cantilever f ollows th e m otion p ath i ntroduced b y th e TW SME

25

Relative motion (µm)

0

4 2

1

−25 −50 −75 −100 −125

5 20

3

Ni–Ti–Cu (A) 40

100 60 80 Temperature (⬚C)

120

140

FIGURE 20.9 Out-of-equilibrium behavior in an Ni–Ti–Cu alloy.

43722_C020.indd Sec1:11

(sequence 2 –3–4–5). Ā e prepared state is not stable as it disappears a fter o ne c omplete h eating–cooling c ycle. N oticeably, t he introduced prepared state or perturbation is cancelled. A proposed interpretation [22,23] of this phenomenon is a two-step mechanism: First, upon heating, variants introduced by t he perturbation a re u nstable a nd a re reoriented w ithin t he stress eld introduced by the training process. Second, when the martensite to a ustenite transformation temperature is reached, some variants start to transform into austenite. Ā is phenomenon was also observed in a binary alloy of different composition. Specimens, although with diff erent chemical composition, d isplay a consistent similar behavior in response to mechanical perturbations. Although further studies are needed to better understand the mechanism rel ated to t his e ffect, t he ob servations c urrently show that TWSME is a robust effect. From an application standpoint, i n t he c ase of t he m icrogripper shown i n Figure 2 0.6, it demonstrates t hat t he ac tuator c an re cover f rom u nwanted deformation introduced in the martensite. At an even smaller scale, the use of the TWSME has recently been re ported f or tem perature-controlled su rface p rotrusions [24]. I n t hat pa rticular c ase, t he T WSME i s i nduced by s evere plastic de formation: Sph erical m icroindents a re m ade o n t he surface of a Ni–Ti thin  lm in its martensite phase followed by a planarization step to re store a  at surface. Protrusions spontaneously appear upon heating and disappear on cooling. 20.2.4.2 Oriented Precipitates (All-Around Effect) Oriented p recipitates w ere r eported t o i nduce a r eversible shape change [25,26]. Ā e principle is explained in Figure 20.10.

10/3/2008 9:04:26 PM

20-12

Smart Materials

R phase transformation

Tensile stress s>0 Compressive stress

Precipitate

s<0

M phase transformation

s<0

s>0

FIGURE 20.10 (Left) Stress distribution around precipitates. (Right) Schematic of t he a ll-around e ffect. Ā e speci men is a ged u nder co nstraint in a curved shape. (Adapted from Kainuma, R., Matsumoto, M., and Hon ma, T. i n Proc. I nt. C onf. Ma rtensitic Transform., Ā e Japan Institute of Metals, Nara, Japan, 1986, 717–722.)

Precipi tates (Ti3Ni4) w ith el lipsoid s hape f orm d uring c onstrained-aging o f Ni -rich T i–Ni a lloys. I n c ompression, t hese precipitates are roughly oriented perpendicular to the direction of applied stress while in tension, they are oriented parallel. For a beam loaded in bending, about half of t he precipitates a re in tensile stress while the rest of the precipitates experience a compressive s tress. I f t he m aterial i nitially  at is aged under constraint s o t hat i t rem ains b ent d uring t he h eat t reatment, precipitates with two opposing directions form across the material t hickness. D uring t he ph ase t ransformation, t he m aterial undergoes shape changes spontaneously: from curved to at and then to curved again but in the opposite direction (as illustrated in Figure 20.10). Numerous studies have been conducted to analyze the effects of c oherent Ti 3Ni4 p recipitates o n t he ph ase t ransformation. Multiple-step t ransformations w ere re ported a nd a nalyzed i n Ni-rich Ni– Ti [ 27,28]. I t w as ob served t hat t hese p recipitates produce an internal stress eld [27,29], which modies the thermodynamic e quilibrium lo cally. Ā e R -phase a nd m artensite morphologies are both affected by the stress eld, which induces the g rowth o f s pecic v ariants [ 27]. Fu rthermore, r ather lo w stresses [30] on t he order of a f ew MPa are reported to mo dify the precipitate distributions. Kuribayashi a nd c oworkers re ported t he u se o f t his p rocess for a n S MA ac tive jo int f or m iniature rob otarm [ 31,32]: A

5 µm-thick sputtered-deposited thin  lm is deposited on NaCl substrate. Ā e  lm is removed from the substrate and annealed at 1 073 K b etween t wo c rystal p lates f or 1 0 m in. F inally, t he  lm is aged in a quartz tube of 3.5 mm in diameter at 673 K for 6 h. A drawback o f t he p recipitation p rocess i s t hat i t re sults in an increase of Ni (due to the precipitation process during aging), which results in a de crease of reverse t ransformation temperatures [16].

20.2.5 Extrinsic Methods Ā ese methods consist in using an additional mechanical element to p rovide t he ne cessary f orce to p romote t he n ucleation o f stress-induced m artensitic v ariants. Ā is m echanical e lement can be either a bias spring, another SMA (antagonistic design), a deadweight, e tc. Ā e mo st c ommonly u sed elemen t i s t he b ias spring as it is rather easy to implement. A typical mechanical construction sketch is shown in Figure 20.11. An SMA coupled to a b ias spring is preloaded so that the system consisting of the SMA actuator and the spring is under stress: Martensite variants are then reoriented to m inimize the strain energy (detwinned, see stage 3 in Figure 20.7) inducing a net m acroscopic de formation. U pon h eating, t he m aterial transforms bac k to t he pa rent ph ase a nd lo ads t he s pring a s it tries to recover its original shape. Ā e output of the mechanism is taken between the SMA and the bias spring. A variety of small devices utilize this principle (e.g., microvalves [5–7], gripper [1,3], endoscope [8], and tactile displays [10]). As an illustration, a m icrovalve is shown [5] in Figure 20.12. It consists of an actuator die with a poppet controlled by a micropatterned Ni– Ti l ayer. A t lo w tem perature, t he b ias s pring ( a Cu–Be microfabricated layer) pushes the poppet toward the orice and de forms t he Ni–Ti r ibbon (detwinning p rocess). Re sistive heating c auses t he r ibbon to t ransform bac k to a ustenite (the parent phase) and to lift the bias spring up and open the valve. Another e xample o f c onstruction ut ilizing a b ias s pring i s shown in Figure 20.13. Ā is m icrodevice i s a m illimeter-sized microendoscope with ve degrees-of-freedom actuated by shape memory elemen ts t hat a re lo cally h eated o n c ertain p ortions. Sensors ba sed o n s train ga uges a re a lso i ncorporated o n t he structure. I n a nother v ersion, t actile s ensors h ave b een adde d Force

SMA spring

g in sprin SMA enite aust

Bias spring B Force max.

ring in A SMA sp martensite

Martensite

D

Austenite A

D B

Stroke

Bias spring

d

Displacement

FIGURE 20.11 General principle of a spring-based design.

43722_C020.indd Sec1:12

10/3/2008 9:04:26 PM

20-13

Shape-Memory Alloys and Effects: Types, Functions, Modeling, and Applications

TiNi ribbon Housing

Pressure vent

Bias spring

Si poppet

Actuator die

−

+

SST spacer

Pneumatic orifice Control pressure

Polymer membrane

Fluid in Fluid orifice

FIGURE 20.12 Microvalve with SMA actuator (Ti–Ni ribbon). Ā e valve is shown in its close position.

Tip of the manipulator 2nd SMA plate Flexible tube Unit combining Electrode sensing and heating MIF functions 3rd link 2nd link

5th link 4th link Unit combining sensing and heating functions

MIF Tube

1st link

Guiding element External electrode

1st SMA plate

FIGURE 20.13 Olympus Co. microendoscope. (L eft) Ā e endoscope w inding a round a m atch. (R ight) E xploded v iew of t he i nside structure. (From Benard, W.L., Kahn, H., Heuer, A.H., and Huff, M.A., J. MEMS, 7(2), 245, 1998; Kaneko, S., Aramaki, S., Arai, K., Takahashi, Y., Adachi, H., and Yanagisawa, K., J. Intell. Mater. Syst. Struct., 7, 331, 1996.)

43722_C020.indd Sec1:13

Ā is effect has been demonstrated in a microgripper [2]. Ā is device is shown in Figure 20.15. A 5 µm Ni–Ti–Cu thin  lm is deposited on a silicon substrate. Ā e jaws are fabricated by precision sawing and bulk-micromachining of Si.

Stress s

Phase transformation regime

Bimorph regime

Au

ste

ar te

ns

nit e

ite

s max

M

Temp.

Ta

ea lin g

Td sd Amorphous material

An n

that give reex functions to the catheter so that when the tube touches something, it automatically bends in the opposite direction. W hile p ossible f or m iniature de vices, t he p reload s tep becomes particularly challenging to implement for submillimeter de vices a s it re quires m icromanipulation or hybrid process integration at a smaller scale. For thin  lm-based devices deposited on a subst rate, an elegant method to solve this issue is to take advantage of the residual s tress th at d eveloped d uring th e th in  lm d eposition an d annealing. Ā e stress build-up mechanism during processing is schematically illustrated in Figure 20.14: A Ni–Ti  lm is deposited on a Si substrate. At the deposition temperature considered, the  lm is amorphous (no crystallographic structure is found). To be functional, an annealing step is required, which is realized by heating the  lm to typically 700–900 K. Upon h eating, d ue to t he d ifference bet ween t he coeffi cient of thermal expansion (CTE) between the lm and the substrate, a compressive stress builds up. Upon crystallization at a tem perature Ta, the stress is relaxed and the lm is under tensile stress on cooling. In the device operating mode, this biasing stress can be efficiently used to deect for instance, a Si-cantilever beam or a membrane.

Crystalline material

FIGURE 2 0.14 Mechanism of s tress ge neration for Ni– Ti d eposited on Si-substrate.

10/3/2008 9:04:27 PM

20-14

Smart Materials

SMA thin film Eutectic bonding interface 200 µm

Gripping jaws

Hollow channel

FIGURE 2 0.15 SMA/Si bi morph m icrogripper. ( Adapted f rom Krulevitch, P., Lee, A.P., Ramsey, P.B., Trevino, J.C., Hamilton, J., and Northup, M.A., J. MEMS 5, 270, 1996.)

Polyimide thin lm has also been proposed as a biasing mechanism to create bimorph-like device construction [33]. Along the same line, SMA composites consisting of Ti(Ni,Cu)/Mo have also been reported [14]. For this particular device, it was noted that t he w ork o utput i s a bout 1 0 t imes l arger t han t he w ork output of the corresponding bimetallic effect [14,27]. Although the residual stress generated during processing of bimorph structures is an efficient mechanism to create a biasing force effect, it limits the design space to out-of-plane motion.

20.2.6 Monolithic Design: Laser Annealing of SMA (LASMA) To bypass t he ne ed for a b ias spring a nd to en large t he de sign space, mo nolithic i ntegration h as b een su ggested [ 21,34]. Ā e key idea is to tailor spatially the microstructures across the SMA element. F igure 2 0.16 i llustrates t he ide a. L et u s c onsider a microactuator c onsisting o f t wo s prings opp osing e ach ot her. Ā e actuator is monolithic and symmetrical but yet has different

mechanical characteristics. Ā is can be achieved for instance by local annealing [35] instead of annealing the complete material. For t hin f ilms, u nless t he sub strate i s h eated su fficiently, as-deposited, the material is amorphous. Similarly, for cold-rolled sheets, a  nal annealing step is necessary to restore the crystallographic s tructure s everely de formed d uring t he c old-rolling process. Figure 2 0.17 s hows, f or i nstance, a ten sile te st e xperiment performed o n a b inary Ni– Ti te st s pecimen b efore a nd a fter annealing. B efore an nealing, the tr ansformation is suppressed and no sup erelasticity i s ob served d ue to t he l arge a mount o f defects t hat h as b een i ntroduced d uring t he f orming p rocess. After a nnealing, t he m aterial re crystallizes a nd i ts f unctional properties are restored. Local a nnealing c an b e p erformed b y d irect jo ule h eating [34,35]: An electrical current ow between t wo contact points dissipates he at. Ā is m ethod a ssumes th at th e e lectrical p ath matches t he de sired me chanical elemen t t hat ne eds to b e annealed. Another approach is to use laser annealing [35]. Ā is approach consists of scanning a laser beam over the specimen to lo cally h eat i t up to tem peratures w here a nnealing t akes place. Near-infrared laser (in t he 600–1000 nm range) is t ypically u sed. I n t his w avelength r ange, Ni– Ti a bsorbs a bout 30%–35% o f t he i ncident r adiation ( for a t hin  lm, i n ne ar normal incidence). Typical irradiance level, i.e., the power density, required to a nneal SMA microdevices is on the order of a few W /mm2. Ā ese le vels de pend o n t he m aterial t hickness, geometry, a nd sub strate. P rocess f eedback c an b e ac hieved using infrared sensors or by monitoring the change of electrical resistance d uring an nealing. Ā e l atter w as demo nstrated o n thin lm-based laser-annealed microswitch where a net decrease of resistivity was observed during annealing. In the annealing temperature ra nges, Ni–Ti oxidizes rapidly i n t he presence of oxygen. However, due to t he short high-temperature exposure time, t hese e ffects remain negligible for most cases. On thin

1000 900 800

Passive element

s Functional properties: active element

T e

e

700 Stress (MPa)

s

* Nonannealed

600 500 400 300 200

Annealed (515⬚C — 15 min)

100 0 0.00 Nonannealed

Annealed

FIGURE 20.16 Schematic of monolithic integration. By locally annealing the material, mechanical properties are distributed across the material.

43722_C020.indd Sec1:14

0.02

0.04 0.06 Strain (%)

0.08

FIGURE 2 0.17 Tensile c haracteristics on a bi nary Ni– Ti a nnealed and nonannealed.

10/3/2008 9:04:28 PM

20-15

Shape-Memory Alloys and Effects: Types, Functions, Modeling, and Applications

Bias spring Ceramic plate + −

Gripping jaws (a)

Actuator laser-annealed

(b)

FIGURE 20.18 Laser-annealed monolithic microgripper . Ā is example illustrates the integration of multiple functionalities in a same substrate. (From Zhang, H., Bellouard, Y., Burdet, E., Clavel, R., Poo, A.-N., and Hutmacher, D.W., Proceedings. ICRA ‘04. 2004 IEEE International Conference on Robotics and Automation, Vol. 5, 26 April–1 May 2004, pp. 4918–4924 [39].)

 lms, t he o xide t hickness o f l aser-annealed s pecimens i n normal air was measured for various laser irradiances and scanning speed [36]. It was found that the oxide thickness does not exceed 250 nm for the laser power density of interest. Wang e t a l. [ 37] p erformed a s ystematic a nalysis o f t he microstructure found for various scanning parameters and laser power. It was found that the microstructure is homogeneous in the laser spot region. In t he p revious s ection ( 20.2.5), i t w as s hown t hat f or  lm deposited on a S i subst rate, st ress generated during deposition and annealing can be used to achieve a reversible motion. Similarly, in the case of laser annealing of freestanding microdevice, stress induced d uring a nnealing r esulting fr om diff erential thermal expansion c an a lso b e e xploited to i ntroduce a t wo-way e ffect without the need to i ntroduce a prestrain. For instance, during

2.0 Normalized DSC signal (W g−1)

1.6 Cooling 1.2

Laser-annealed actuator

0.8 Laser-cut pull-back spring

0.4

Raw material (cold-rolled) 0.0 −0.4 −0.8 −1.2

Heating 40

60

80

100

120

140

Temperature (⬚C)

FIGURE 20.19 DSC of the microgripper main parts namely the pullback spring and the actuator element (laser annealed). Ā e signal for the raw material, as received, is shown for comparison. For each of the three DSC, the mass specimens are similar.

43722_C020.indd Sec1:15

local l aser e xposure, t he de vice c an lo cally b e put u nder c ompressive stress due to asymmetric thermal expansion; this stress is relaxed during annealing and transforms into a tensile stress on cooling [38]. A laser-annealed microgripper is shown in Figure 20.18 [38]. Ā is device was designed for the manipulation of scaffold parts used in bone reconstruction therapy. Ā e raw material is a cold-rolled Ni–Ti–Cu (5% at .) sheet. Ā e device is laser-cut using a Nd–YAG slab laser. Due to t he high level of cold-rolling, a s re ceived, t he r aw m aterial do es not e xhibit a ph ase transformation. Ā is can be seen on the differential scanning calorimetry ( DSC) re sults p resented i n F igure 2 0.19 ( raw material). Ā e laser-cut itself induced a partial annealing effect localized in the laser-affected zone. Ā is annealing effect is limited to the so-called heat-affected zone (typically 8–10 µm). Note that this annealing process resulting f rom t he machining process is not desirable: i t c an b e sig nicantly re duced b y c hoosing a le ss heat-producing micromanufacturing method (like femtosecond lasers or combined laser-machining with water jet). Ā e effect o f l aser a nnealing o n t he DS C sig nal i s c learly visible. A sharp, well-dened, transformation peak is observed. (Smaller secondary peaks are related to the DSC specimen preparation itself that required additional laser cutting.) From a de sign p oint o f v iew, t his de vice i s si milar to t he bias-spring/actuator c onguration. A s t he r aw m aterial i s a cold-rolled sheet, prestrain is achieved by introducing an offset between the contact pads during assembly. As mentioned earlier, this p restrain m ay not b e ne cessary f or l aser-annealed t hin l m-based microdevices.

20.2.7 Summary of SMA Actuator Design Methods Figure 2 0.20 su mmarizes t he v arious p rinciples a vailable to induce a re versible m acroscopic s hape c hange. Ā e most commonly f ound te chniques w ere b riey o utlined i n t he p revious

10/3/2008 9:04:28 PM

20-16

Smart Materials

Reversible shape change

20.2.9 Future Directions for SMA Applications in Microtechnologies 20.2.9.1 SMA on a Flexible Substrate

Extrinsic biasing mechanism

Intrinsic biasing mechanism

Mechanical construction (Deadweight load, bias spring, push–pull design)

Two-way shape memory effect (oriented defects) All-round effect (oriented precipitates) Localized annealing

FIGURE 2 0.20 Outlook of me thods t hat i nduce re versible s hape change in SMA.

sections. I n add ition to t hese te chniques, o ne s hould add io n bombardment as a means to induce localized amorphization or inhomogeneities, which, combined with unaffected zones, leads to an intrinsic two-way effect [40]. Special fabrication processes can a lso le ad to i ntrinsic t wo-way e ffects: Ā is i s t he c ase f or melt-spinning ( where mol ten me tal i s r apidly s olidied o n a high-speed spinning wheel) and thin lm fabricated by stacking sequence of thin layers of Ni and Ti [41]. Ā ese two-way effects essentially re sult f rom t he pa rticular m icrostructure f ound across the material thickness. Ā e high output force is almost systematically cited to support the use of SMA in microsystems. While it is an important argument, it has to be balanced with known limitations of SMA that have a poor bandwidth (<100 Hz) and the lack of intrinsic reversibility. Ā e l atter c an b e o vercome b y p roperly de signing t he S MA microsystems. Fu rthermore, lo cal a nnealing te chnique a lso offers a w ay to b ypass p rocess c ompatibilities i ssues re sulting from the need of high-temperature processing.

20.2.8

SMA Flexures

Flexures a re u sed i n precision eng ineering w here h ighly ac curate, w ear-free, smo oth, a nd re peatable mot ion i s de sired. Flexures a re ba sed o n de formation o f m aterial to ac hieve a motion between elastically joined parts. Ā ey are used in a variety of precision mechanisms such as high-resolution balances or high-accuracy optical positioning stages. SMA a re a n at tractive opt ion i n de signing  exures. Various examples o f S MA  exures we re re ported [42–47]: Sup erelastic exures can withstand larger deformations for the same weight as a conventional exure. In addition, the damping properties of SMA, controllable through the phase transformation, offer new design opp ortunities f or ad aptive c ompliant me chanisms [47]. Ā e martensitic phase t ransformation can a lso be used to s hift the nat ural f requency of  exures adding useful f unctionalities such as vibration rejection [47].

43722_C020.indd Sec1:16

Electronics on  exible sub strate a re de veloping at a f ast pac e. Ā e so-called polymer electronics or plastic MEMS a re rapidly emerging as a powerful platform for multifunctional integration in a v ariety o f pac kages a nd sub strates. Ā is is p articularly attractive for biomedical applications where uidics, optics, and electronics c an b e merge d on a s ame,  exible substrate (smart bandage, for instance). Ni–Ti alloys being biocompatible materials a re qu ite at tractive i n t hat context. A lthough early demonstrations o f Ni– Ti/polyimide ac tuator [ 43] w ere m ade, mo re work is needed to expand the process capabilities. 20.2.9.2

SME at the Nanoscale: Toward Nano-SMA Actuator

Observing SME effect at t he nanoscale has been of great interest recently both f rom a n application a nd t heoretical point of v iew. Nanoindentations experiments [48,49] have demonstrated nanoscale SME. Ā ese studies provide the direction for future nanoactuators based on SMA. However, shape memory transformation in nanopatterned SMA actuator is yet to be demonstrated.

References 1. K. Ikuta, Proc. IEEE Int. Conf. Robotics Autom., Cincinatti, 1990, pp. 215–216. 2. P. Krulevitch, A.P. Lee, P.B. Ramsey, J.C. Trevino, J. Hamilton, and M.A. Northup, J. MEMS 5(4), 1996, 270–282. 3. J. H esselbach, R . P ittschellis, E. H ornbogen, a nd M. M ertmann, P roc. S hape M emory a nd S uperelastic Technol., Pacic Grove, CA, 1997, pp. 251–256. 4. Y. B ellouard, R . Cla vel, J .-E. B idaux, R . G otthardt, a nd T. Sidler, Proc. Shape Memory Superelastic Technol., Pacic Grove, CA, 1997, pp. 245–250. 5. A.D. Johnson and V.V. Martynov, Proc. 2nd Int. Conf. Shape Memory S uperelastic T echnol., P acic G rove, CA, 1997, pp. 149–154. 6. J.D. Busch and A.D. Johnson, Shape-memory alloys microactuator, U.S. Patent 5,061–914, June 27, 1989. 7. M. Kohl, D. Dittmann, E. Quandt, B. Winzek, S. Miyazaki, and D.M. Allen, Mater. Sci. Eng. A, 230, 1999, 273–275. 8. W.L. Benard, H. Kahn, A.H. Heuer, and M.A. Huff, J. MEMS, 7(2), 1998, 245–251. 9. S. Kaneko, S. Aramaki, K. Arai, Y. Takahashi, H. Adachi, and K. Yanagisawa, J. Intell. Mater. Syst. Struct., 7, 1996, 331–335. 10. H. Fischer, R. Trapp, L. Schüle, and B. Hoffmann, J. Phys. IV France 7, 1997, C5–609–614. 11. S. Takeuchi and I. Shimoyama, J. MEMS, 19(1), 2000, 24–31. 12. S. Miyazaki and A. Ishida, Mater. Sci. Eng. A, 273–275, 1999, 106–133. 13. Y. F u, H. Du , W. H uang, S. Zha ng, a nd M. H u, Sensors Actuators A, 112, 2004, 395–408.

10/3/2008 9:04:30 PM

Shape-Memory Alloys and Effects: Types, Functions, Modeling, and Applications

14. B. Winzek, S. S chmitz, H. R umpf, T . S erzl, R . H assdorf, S. Tienhaus, J. Feydt, M. Moske, and E. Quandt, Mater. Sci. Eng. A, 378, 2004, 40–46. 15. L. D elaey, P hase tra nsformations in ma terials. In Material Science a nd T echnology, v ol. 5 , R .W. C ahn, P . H aasen, E.J. Kramer, eds. VCH, Weinheim, 1991. 16. K. Otsuka and X. Ren, Prog. Mater. Sci., 50, 2005, 511–678. 17. K. Otsuka and C.M. Wayman, eds., Shape Memory Materials. Cambridge University Press, Cambridge, U.K., 1998. 18. J. Perkins a nd D. Hogson, in Engineering Aspects of Sh ape Memory Alloys, T.W. Duerig, K.N. Melton, D. S töckel, a nd C.M. Wayman, eds. Butterworth-Heinemann, London, 1990. 19. H. S cherngell a nd A.C. K neissl, Acta M ater., 50(2), 2002, 327–341. 20. H. Scherngell and A.C. Kneissl, Script. Mater., 39(2), 205–212. 21. Y. Bellouard, PhD diss ertation, Ecole Poly technique Fe'de' rale de Lausanne, 2000. 22. Y. Bellouard, R. Clavel, and R. Gotthardt, J. Physique IV JP 11(8), 2001, Pr8583–Pr8588. 23. Y. Bellouard, R. Clavel, R. Gotthardt, and J. van Humbeeck, J. Physique IV JP 112(II), 2003, 765–768. 24. Y. Zhang, Y.-T. Cheng, and D.S. Grummon, Appl. Phys. Lett., 89, 2006, 41912. 25. M. N ishida a nd T . H onma, Script. M etallurg., 18, 1984, 1293–1298. 26. R. Kainuma, M. Matsumoto, and T. Honma, Proc. Int. Conf. Martensitic Transform., Ā e Japan Institute of Metals, Nara, Japan, 1986, pp. 717–722. 27. L. Bataillard, J.-E. Bidaux, and R. Gotthardt, Phil. Mag., A 78, 1998, 327–344. 28. J. Khalil-Alla, A. Dlo uhy, a nd G. E ggeler, Acta Mater., 50 2002 4255–4274. 29. W. Tirry and D. Schryvers, Acta Mater., 53 2005, 1041–1049. 30. J. Michutta, Ch. Somsen, A. Yawny, A. Dlouhy, and G. Eggeler, Acta Mater., 54, 2006, 3525–3542. 31. K. K uribayashi, P roc. MEMS, 1990, p p. 217–221. D OI 10.1109/MEMSYS.1990.110279. 32. K. K uribayashi, S. S himizu, T . N ishinohara, T . T aniguchi, M. Yoshitake, and S. Ogawa, Proc. IROS 1993, pp. 1697–1702. DOI 10.1109/IROS.1993.583865. 33. K. Kuribayashi and T. Fujii, Proc. Int. Symp. Micromechatronics Human Science (MHS), 1988, pp. 165–170. 34. Y. B ellouard, R . Cla vel, R . G otthardt, J .-E. B idaux, a nd T. Sidler, Actuators, Int. Conf. on New Actuators, H. Borgmann, ed., Bremen, Germany, June 17–19, 1998, pp. 502–505. 35. Y. Bellouard, T. Lehnert, J.-E. Bidaux, T. Sidler, R. Clavel, and R. Gotthardt, Mater. Sci. Eng., A273–275, 1999, 733–737. 36. F. Khelfaoui, Y. Bellouard, T. Gessmann, X. Wang, J. Vlassak, and M. Hafez, Proceedings of the International Conference on Shape Memory and Superelastic Technology (SMST-2004), Germany, ASM International, 2004. 37. X. Wang, Y. B ellouard, a nd J.J. Vlassak, Acta M ater., 2005, 4955–4961. 38. Y. Bellouard, T. Lehnert, R. Clavel, and R. Gotthardt, J. Phys. IV, JP 11(8), 2001, Pr8571–Pr8576.

43722_C020.indd Sec1:17

20-17

39. H. Zha ng, Y. B ellouard, E. B urdet, R . Cla vel, A.-N. P oo, and D.W. H utmacher, Rob otics a nd a utomation, 2004. Proceedings. ICRA ‘04. 2004 IEEE I nternational Conference on Robotics and Automation, Vol. 5, 26 April–1 May 2004, pp. 4918–4924. DOI 10.1109/ROBOT.2004.1302497. 40. D.S. G rummon a nd R . G otthardt, Acta M ater., 48, 2000, 635–646. 41. T. L ehnert, S. T ixier, P. B öni, a nd R . G otthardt, Mater. S ci. Eng., A273–275, 1999, 713–716. 42. K. Kuribayashi and T. Fujii, Proc. Int. Symp. Micromechatronics Hum. Sci., Piscataway, NJ, 1998, pp. 165–170. 43. J. Peirs, D. Reynaerts, and H. Van Brussel, Sensors Actuators A Phys 70(1/2), 1998, 135. 44. M. M ertmann, E. H ornbogen, J . H esselbach, a nd R. Pittschellis, Proceedings of SMST, A. Pelton, D. Hodgson, and S. R ussell, T. Duerig, e ds., P acic G rove, Sa nta Cla ra, March 2–6, 1997, p. 549. 45. J. H esselbach a nd A. R aatz, Mater. S ci. Fo rum 394(3), 2001, 79. 46. Y. B ellouard, R . Cla vel, R . G otthardt, J .E. B idaux, a nd T. S idler, P roceedings o f Actuators, H. B orgmann, e d., Bremen, 1998, p. 505. 47. Y. B ellouard and R. Clavel, Shape memory  exures, Mater. Sci. Eng. A, 378(25), 2004, 210–215. 48. G.A. S haw, D .S. S tone, A.D. J ohnson, A.B. Ellis, a nd W.C. Crone, Appl. Phys. Lett., 83, 2003, 257. 49. X.-G. M a a nd K. K omvopoulos, Appl. P hys. L ett., 83, 2003, 3773.

20.3 Mathematical Models for Shape Memory Materials Davide Bernardini and Thomas J. Pence 20.3.1 Introduction Ā e purpose of this chapter is to give a basic primer on the mathematical mo deling of s hape me mory m aterial b ehavior for t he purpose o f u nderstanding t he b ehavior o f s tructures a nd mechanical devices. It is assumed that the reader is familiar with the elementary aspects of shape memory materials (SMM) like pseudoelastic b ehavior, SME, a nd t heir u nderlying s olid ph ase transformations [27,33]. Attention is restricted to behavior that is observable at a macroscopic scale under thermal and mechanical lo ads i nvolving a si ngle c omponent o f s tress, e .g., ten sion versus compression, as is generally sufficient for describing wires. More advanced modeling issues, many associated with detailed metallurgical a spects, a re o utside o f t he s cope o f t he p resent exposition, i ncluding t he b ehavior u nder m ultidimensional stress states; low-cycle fatigue behavior induced by plastic deformation of the phases; effects associated with plasticity, viscoelasticity, or strain gradients; and the two-way shape memory effect. Ā e phenomenology that will be object of modeling is briey

10/3/2008 9:04:30 PM

20-18

Smart Materials

recalled in the following section, in order to set the terminology and point out the issues relevant for the subsequent sections.

20.3.2 Experimental Observations 20.3.2.1 Thermal Transformation In the absence of loads, the material is in the austenitic phase A at tem peratures T > Af, a nd i s i n t he m artensitic ph ase M at temperatures T < Mf. Between these temperatures, the crystallographic phase may involve a m ixture of martensite and austenite, i n w hich c ase i t i s c onvenient to i ntroduce t he f raction o f martensite x. Ā is fraction is dependent on the recent temperature history whereupon it is useful to introduce the additional material parameters Ms and As. Here Ms is the start temperature for the c ooling c onversion o f a ustenite to m artensite, t he f orward martensitic t ransformation ( FwT). Ā is kind of martensite is characterized b y t he a bsence o f m acroscopic t ransformation strain relative to austenite. Ā is is due to the fact that at a microscopic scale several variants occur in a way that the overall transformation strain vanishes (simple twinning is a s tandard case). Such thermal martensite is said to be self-accommodated. Similarly, As is t he start temperature for t he heating conversion of this martensite back to austenite, the reverse martensitic transformation ( RvT). Ā ese tem peratures a re de termined b y alloy c ontent a nd p rocessing h istory. Ā e s tandard ob served ordering of these temperatures is Mf < Ms < As < Af. In fact, the central inequality is not re quired by any thermodynamic principle, and the description naturally extends to the less restrictive conditions Mf ≤ Ms ≤ Af and Mf ≤ As ≤ Af. As is conventional for cooling transformations, the forward transformation is exothermic. C onversely, t he RvT i s en dothermic. Ā e thermodynamically reversible latent heat released and absorbed in these transformations is denoted by H.

20.3.2.2

Ā e i mposition of stress a lso leads to F wT whereas its removal facilitates RvT. I f c ycles o f me chanical lo ading a nd u nloading are applied at a sufficiently slow rate at temperatures T > Af, then the typical pseudoelastic outer loop is observed. Specically, so long a s plasticity is not pre sent, loading i n tension  rst g ives a linear el astic branch w ith mo dulus EA and x = 0, followed by a conversion plateau associated with FwT, followed by a second linear elastic branch with modulus EM and x = 1 (Figure 20.21). Here the conversion during FwT is no longer self-accommodated, a nd i nstead results i n a c rystallographic biased martensite M+ that provides transformation strain g + to the material. Since t he F wT i s e xothermic, i t w ould nat urally le ad to a n increase i n t he S MM tem perature a s de termined b y t he S MM specic heat, denoted by c. However, if the loading rate is sufficiently lo w, t hen h eat t ransfer p ermits t he tem perature o f t he SMM to remain close to its original value as determined by the overall thermal environment. In this isothermal setting, the FwT plateau b egins a nd  nishes a t t ransformation st resses sMs(T) and sMf(T). Ā ese t ransformation st resses a re t emperature dependent, as the same test carried out at a different temperature T′ yields sMs(T′) and sMf(T′) different from the former ones. A representation of the transformation stresses versus temperature is called either a phase diagram or a state diagram (Figure 20.21). Ā e de pendence o f sMs(T) a nd sMf(T) up on tem perature f or T > Af is essentially linear and t he common slope for t hese t wo curves, denoted by m, is also a material parameter. Āe Clausius– Clapeyron c ondition o f c onventional t hermodynamics g ives that the slope m is related to t he latent heat, the transformation temperature, and the transformation strain via a standard relation. In view of the multiple transformation temperatures associated w ith S MMs, t his rel ation c an b e t aken f or mo deling purposes as:

sMf

s

Isothermal Pseudoelastic Loading

s

sMs sAs sAf

EM

EA

sMf (T )

M+

sMs(T )

A

sAs(T ) g− Af

T T⬘

M+

sAf (T ) A

T

g+

e

M−

A M−

FIGURE 20.21 At a g iven temperature, u niaxial loading, a nd u nloading for c omplete FwT a nd complete RvT generates t he outer loop of t he stress–strain diagram. Changing the temperature alters the transformation stresses.

43722_C020.indd Sec1:18

10/3/2008 9:04:30 PM

Shape-Memory Alloys and Effects: Types, Functions, Modeling, and Applications

m=

4H (2 M M + ( f s + A s + A f )g +

0.1)

Since t he h igh d amping c apacity a ssociated w ith S MMs m ay complicate calorimetric measurement of H, Equation 20.1 is also useful for the determination of H from stress–strain data. 20.3.2.3 Isothermal Pseudoelastic Tensile Unloading Unloading at T > Af reverses the procedure discussed above with one essential difference, namely the plateau associated with RvT is below that of the FwT (Figure 20.21). Hence, even though the transformation s train g + is completely recovered, there is a stress–strain hysteresis loop. Ā e RvT is endothermic, but sufficiently slow unloading permits the temperature to remain at the ambient, whereupon the start and  nish of the lower plateau is at t ransformation s tresses sAs(T) a nd sAf(T). F or T > Af, t hese curves on the state diagram are also linear with slope m. 20.3.2.4 Isothermal Pseudoelastic Compressive Loading and Unloading Ā e application of the opposite sign load of compression induces a different assembly of martensite variants. Ā e overall assembly is denoted by M− and characterized by an overall transformation st rain g − < 0. F orward a nd re verse t ransformations i n compression are thus associated with the respective conversions A → M− and M− → A (Figure 20.21). Fu nctions a nalogous to sMs(T), sMf(T), sAs(T), a nd sAf(T) c an b e c onstructed i n t he compressive re gime o f t he s tate d iagram, a nd f or T > Af, t he slope relation (Equation 20.1) continues to hold provided that g + is replaced by g −. 20.3.2.5

Internal Subloops

At T > Af, if a tensile loading is arrested before the completion of FwT, then the material is a mixture of A and M+, say with martensite fraction value x = xa < 1. Unloading then activates RvT at a stress sAs(T,xa) that, in general, may be different from the one of t he o uter lo op. I f a sub sequent relo ading o ccurs b efore t he completion of the RvT, say when x = xb (so that 0 < xb <xa), then s

this reloading activates FwT at a stress sMs(T,xb), giving rise to an internal pseudoelastic subloop (Figure 20.22a). Unloading or re loading at d ifferent values of x yields different values of t he RvT activation stresses (Figure 2 0.22b). Ā e loci of potential ac tivation of t he t ransformations of i nternal lo ops a re called subloop activation curves a nd may be described by f unctions sAs(T,x0) and sMs(T,x0) where x0 denotes the fraction of martensite at the end of the last transformation process. Ā is expresses the fact that an internal resistance must be overcome in order to activate t he t ransformation, a nd t his re sistance de pends o n t he actual microstructure from which the transformation evolves. On the other hand, if internal transformations are brought to completion, then the completion stress is often the same as that for the outer lo op, i rrespective o f t he pa rticular i nitial m icrostructure. SMMs having a completion stress that is independent of the microstructure at transformation activation are said to obey the Ivshin condition for hysteresis completion (Figure 20.22b and c). 20.3.2.6

Nonisothermal Pseudoelasticity

Isothermal pseudoelasticity only takes place when the loading rate is sufficiently slow so as to permit the heat released and absorbed during transformation to be completely exchanged with the environment hence allowing the temperature of the material to remain unchanged. Conversely, if the loading rate exceeds the rate of heat exchange with the environment, then the latent heat of transformation gives rise to a change of temperature within the SMM. Since the transformation stresses depend on the temperature, it follows that the shape of the pseudoelastic loop may be signicantly diff erent w ith re spect to t he i sothermal c ase ( Figure 20.23). B eyond a c ertain loading r ate, t his temperature change process may be treated as adiabatic (no heat exchange at all). Of course, the adiabatic treatment does not pertain during any time intervals between load change wherein the now heated or cooled SMM exchanges heat with the environment at constant load. 20.3.2.7 Reorientation Ā e M+ type of martensite obtained by pseudoelastic FwT at T > Af can also be obtained at temperatures T < Af. For example,

s sMs ( xb )

20-19

s

x = xa

rF

rR

x = xb

sAs ( xa ) e

(a)

e (b)

e (c)

FIGURE 20.22 (a) Sublooping occurs when transformations do not go to c ompletion. (b) Elastic unloading from arrested FwT followed by the activation of RvT at the dashed line. (c) Elastic loading from arrested RvT followed by activation of FwT at the dashed line. Both (b) and (c) show subloop completion at a common point.

43722_C020.indd Sec1:19

10/3/2008 9:04:31 PM

20-20

Smart Materials s

T

Faster loading

Slower loading (nearly isothermal)

e

Time

FIGURE 20.23 Sufficiently slow loading g ives nearly isothermal behavior. For f aster loading, t he latent heat of t ransformation g ives SM M temperature change that in turn alters the stress–strain curve.

at T < Mf, this is accomplished by applying stress to the selfaccommodated M martensite obtained by cooling (i.e., purely thermal FwT). Ā e applied stress t hen gives a n M → M+ c onversion t hat i s c alled for ward re orientation (F wR). Ā is FwR also results in a transformation strain g + within the SMM. Ā e associated re orientation pl ateau b egins a nd  nishes at t ransformation stresses sRs+ and sRf+ (Figure 20.24). Unlike the activation s tresses for A → M+ t ransformation, t he re orientation stresses only display mild temperature dependence. Unloading does not ac tivate t he re verse t ransformation, s o t he t ransformation st rain g + re mains at z ero s tress. C ompressive lo ading can t hen i nitiate a c orresponding M+ → M− re verse re orientation ( RvR). I n add ition, c ompressive lo ading app lied to a self-accommodated martensite M gives M → M−, which is also classied as RvR. Ā e a ssociated re orientation pl ateau i n compression b egins a nd  nishes at t ransformation s tresses s Rs− and sRf−. Ā ere is no latent heat associated with reorientation, w hich i s c onsistent w ith t he m ild to v anishing e ffect of temperature on the threshold stresses sRs+, sRf+, sRs−, and sRf−.

s Mf

s

s Ms

s As

sRs Ms

As

Af

T

FIGURE 20.24 Extended state diagram showing outer loop activation and c ompletion l ines for F wT a nd Rv T a nd si milar l ines for F wR. Combined t ransformation a nd re orientation c an o ccur on t he h atched region for s uitably or iented ( s,T)-paths. Al though n ot i ndicated h ere, reorientation may cause the FwT and RvT lines to become curved in the region of combined transformation and reorientation.

43722_C020.indd Sec1:20

20.3.2.8 Shape Memory Effect Ā e S ME i s a c onsequence o f t he f act t hat i sothermal s hape change due to pure reorientation can be undone by thermal RvT. In particular, starting with austenite, consider a three stage cycle composed as: (sm1) thermal FwT (i.e., A → M by cooling in the absence of stress), followed by (sm2) i sothermal re orientation M → M+ d ue to lo ading which then persists upon unloading, followed by (sm3) thermal RvT (i.e., M+ → A by heating in the absence of stress).

s Af

sRf

Mf

Pure re orientation re fers to re orientation t hat o ccurs at constant ph ase f raction, f or e xample at T < Mf with x = 1 a s discussed above. C onversely, at tem peratures i n t he i nterval M f < T < Af, it is not unusual for reorientation to take place in conjunction with phase transformation. For example, cooling austenite to a tem perature b elow Ms b ut a bove M f gi ves a mixture o f a ustenite a nd s elf-accommodated m artensite. Isothermal tensile loading t hen converts both t he A a nd t he self-accommodated M to t he o riented m artensite M +, t he former c onversion b y ph ase t ransformation a nd t he l atter conversion by reorientation. An extended state diagram for which the temperature may range from below M f to above A f must therefore show multiple lines associated with the start and f inish o f b oth p hase tr ansformation an d r eorientation (Figure 20.24).

Stage (sm1) gives no shape change, whereas stage (sm2) provides a seemingly plastic deformation. However, stage (sm3) reverses the shape change by purely thermal means, giving rise to the SME. At the conclusion of (sm3), the material has returned to the shape associated with the original austenite and remains at that s hape d uring a ny tem perature c hange ( even a c ooling associated with FwT) so long as there is no further loading (Figure 20.25). Ā e t wo-way memory effect refers to c ertain processes w herein multiple cycles of this kind are performed in such a manner so as

10/3/2008 9:04:34 PM

20-21

Shape-Memory Alloys and Effects: Types, Functions, Modeling, and Applications s

s M+ M

sRf sRs M

e

Mf Ms

Heating

M+

e

M A

A

M+

As

Af

T

FIGURE 20.25 For T < Mf, loading and unloading reorients an originally self-accommodated martensite, resulting in an apparent plastic strain. Heating converts the reoriented martensite to austenite, returning the material to its original shape (the SME).

to c reate a de fect s tructure ( e.g., a d islocation pat tern) a fter reorientation that serves to stabilize the deformation after stage (sm2). Such “training” can eventually lead to a situation wherein the shape change associated with reorientation can be obtained by c ooling i n t he a bsence o f lo ad, at w hich p oint a s tress-free cooling and heating cycle gives a reversible shape change. Models for this two-way effect generally require some notion of internal defect stress, which is a concept that we do not treat here.

20.3.3 Overview of the Literature Ā e mo deling o f t he t hermomechanical b ehavior o f S MM h as received i ncreasing a ttention i n t he s cientic l iterature f or t he last two decades. Presently, the literature is vast and some review papers are available (e.g., Refs. [3,21]). In the following, a qu ick survey will be given, more details can be found in the abovementioned reviews. Roughly speaking, most of t he existing models can be classied into two broad categories. 20.3.3.1

Micromechanical Models

Such theories analyze the material at a scale in which the multiphase nat ure o f t he m aterial i s e xplicit. W hen t he m aterial i s modeled a s a c ontinuum, e ach p oint re presents a m aterial ele ment m ade o f a si ngle s olid ph ase r ather t han a m ixture o f phases. S ome k ind o f a veraging p rocedure m ay t hen s erve to recover the macroscopic behavior. In this category, there are several approaches, of which only two are mentioned here. One line of research was initiated by Patoor et al. [28] and then pursued by t he s ame g roup i n s everal pap ers ( e.g., Re f. [ 35]) a nd a lso developed b y Br inson a nd c oworkers ( e.g., Re f. [ 15]). Ā is approach describes the macroscopic behavior via the homogenization of a representative volume element of austenite and various variants of martensites. For single crystals, this produces a detailed analysis of the microstructure evolution. It can be extended also to polycrystals taking into account the grain structure although requiring a large computational effort.

43722_C020.indd Sec1:21

Another app roach i nitiated b y B all a nd J ames [2] a nd t hen pursued by m any authors (e.g., Re fs. [7,32]) a ims to de termine the actual microstructure formed under given thermomechanical c onditions v ia t he m inimization of a su itable energ y f unction. Ā is requires advanced methods of calculus of variations. 20.3.3.2

Macroscopic Models

Ā e detailed study of the microstructure may become unpractical both for the computational effort and for the lack of detailed knowledge about the microstructure of actual materials. Ā is mot ivates app roaches t hat d irectly mo del t he m acroscopic behavior without such detailed microstructural considerations. In this case, when the material is modeled as a continuum, each point represents a m aterial element made of a m ixture of phases, so that the features of the mixing has to be specied by suitable descriptors. Such models then differ on how the effect of the microstructure is included. • Transition models without internal variables. Such models describe hysteresis by considering minimization of macroscopic energy functions with multiple energy wells, each corresponding to a different microstructure, in conjunction w ith r ules t hat del ay t he t ransition b etween w ells. Such rules of delay, based in some sense on the idea that it is necessary to overdrive the system in order to transition between wells, have the effect of temporarily overruling the associated equation of state. Falk [14] initiated the development of such models in the context of SMM, while several adv ances w ere ob tained, a mong t he ot hers, b y Muller and Seelecke [25], and Puglisi and Truskinovsky [29]. Ā is type of modeling is not considered here in detail, although it is worthwhile to note that there are deep and fundamental connections between these models and models containing internal variables that are discussed below. • Phenomenological mod els of h ysteresis. Suc h mo dels o f hysteresis have been widely used in elds like, for example,

10/3/2008 9:04:35 PM

20-22

Smart Materials

that of magnetic materials, to model experimentally observed curves involving high nonlinearity and complicated looping. Ā is may, but often does not, involve explicit focus on their link with the physical phenomena of interest. Rel iability a nd t he rob ustness o f t he mo del a re o f paramount c onsideration. By f avorably m atching t he algorithm t o e xperiment, th ese al gorithms p ermit th e treatment o f a rbitrarily c omplex d riving i nput [ 18,23]. Preisach models for hysteresis can also be viewed as an example of this type of model [26]. • Models fo r th e e volution of inte rnal v ariables. Perhaps the most common macroscopic approach is based on the introduction of internal variables that describe the internal m icrostructure o f t he m aterial. Suc h v ariables a re included w ithin t he l ist of t he i ndependent variables of the standard constitutive functions. Ā e evolution of t he phase t ransformations i s t hen mo deled b y su itable equations o f e volution, t he s o-called t ransformation kinetics [3–6,9–13,17,19,34]. Connections can be forged between all of these approaches. In the f ollowing, o ur at tention w ill b e f ocused o n mo dels f or t he evolution of internal variables.

Ā e m ain c onceptual s teps ne eded to de velop a c onstitutive theory for SMM are usually (i) choice o f t he s et c o f t hermal a nd me chanical v ariables that may be controlled (ii) choice o f t he s et a o f i nternal va riables t hat r easonably describe the internal microstructure ( iii) specication of constitutive equations for t he remaining thermomechanical variables (i v) specication of a c onstitutive e quation for t he e nergetic variable consistent with (i) (v ) specication o f c onstitutive e quations f or t he in ternal variables For example, taking c = {s, T}, then steps (iii) and (iv) require the specication of functions {e(c,a), h(c,a)} and for the Gibbs free energy, y (c,a). Step (v) is generally accomplished by specifying a transformation kinetic, typically a differential equation in the form α
= f (c , α , c
) that prescribes the evolution of the internal variables. Ā ermodynamic c onsistency w ith t he a bove ba lance e quations re quires t he f ollowing e quations o f s tate f or s tress a nd temperature: e=−

20.3.4 Formulation of SMM Models in the Framework of Thermomechanics with Internal Variables In this setting, the state of the material element is described not only by a pa ir of me chanical v ariables (stress s and strain e) and a pair of thermal variables (temperature T and entropy h), but also by a list of suitable internal variables a. F or o nedimensional models, t he mechanical variables a re scalars; for multidimensional mo dels, m any o f t hese v ariables b ecome second-order tensors. Additional control variables, for example, electric or magnetic eld may be added, but this case is not considered here. Moreover, the thermomechanical formulation requires a n energe tic v ariable ( like t he i nternal energ y e or Gibb’s free energy y = e − Th − se), and the rate of energy dissipation Γ. All such variables are constrained by ba lance equations t hat may be stated i n va rious sl ightly d ifferent ways. Following, for example, Gre en a nd Naghdi [16], t here a re ba lance of i nternal energy and balance of entropy

. e
= se + r,

1 . h = (Γ + r ) (2 T

0.2)

where r is the external heat supply, as well as the Clausius–Duhem inequality Γ > 0 a ssociated w ith t he s econd l aw o f t hermodynamics. Since the constitutive theory is developed in the context of homogeneous processes there is no heat ow contribution to Equation 2 0.2. Ā e re sulting c onstitutive mo dels, h owever, would t ypically b e i mplemented i n t he c ontext o f e quations describing heat transfer effects a nd either static equilibrium or structural dynamics [20].

43722_C020.indd Sec1:22

∂y , ∂s

h=−

∂y (2 ∂T

0.3)

hence making step (iii) follow f rom step (iv). A lternatively, step (v) may be carried out by specifying a c onstitutive f unction for the r ate o f energ y d issipation Γ. C onsistency w ith t he ba lance Equation 20.2 then provides the appropriate transformation kinetic, w hich i n t his c ase, f ollows a s a log ical c onsequence, rather th an b eing s eparately p osed a t th e o utset [ 4,30]. Accordingly, each model is thus characterized by the choice of the i nternal variables a, t he Gibbs f ree energ y f unction y, a nd the dissipation function Γ. In the following, the above concepts will be discussed in some detail a nd exemplied a long t he l ines of Ref. [6]. To a id i n t he present exposition, we consider a number of simplifying assumptions on the material behavior. Specically, we consider a SMM that has been processed so that its thermomechanical behavior permits reasonable modeling under the following specializations: (s1) Tension–compression s ymmetry s o t hat g + = − g− = g, sRs+ = −sRs− = sRs, and sRf+ = −sRf− = sRf (s2) Similar elastic properties so that EA = EM = E (s3) Aforementioned Clausius–Clapeyron relation (Equ ation 20.1) (s4) Similar F wT a nd RvT tem perature i ntervals Ms − Mf ≅ A f − As (s5) Ivshin’s condition for hysteresis completion (s6) Negligible thermal expansion of the conventional kind (s7) No phases other than austenite and martensite that contribute to the SMM effect (s8) No slip plasticity or viscoelasticity None o f t hese si mplications i s c entral to mo del de velopment and the corresponding modeling generalizations associated with

10/3/2008 9:04:35 PM

Shape-Memory Alloys and Effects: Types, Functions, Modeling, and Applications

20-23

If only pseudoelasticity is of concern, then a single internal variable is sufficient for the basic modeling. Ā e martensite volume fraction x, w hich ne cessarily ob eys 0 ≤ x ≤ 1, i s therefore the standard choice in almost all such models. Increasing, decreasing, and unchanging x are then identied respectively with FwT, RvT, a nd no t ransformation ( nT). C lassical mo dels a re t hose developed by Tanaka et al. [36] and Liang and Rogers [22], which are distinguished by t he pa rticular form of t he k inetic r ule for the e volution of x. I n t his c ontext, Ivshin a nd Pence [19] were among t he  rst to mo del no nisothermal b ehavior a ssociated with heat transfer and the adiabatic limit. Several models have been proposed in this setting. Among the others, w e men tion t hose de veloped b y R aniecki e t a l. [ 31], Auricchio et al. [1] who pay special attention to issues related to numerical implementation, and by Matsuzaki and Kamita [24]. All such groups developed various generalizations of the model presented i n t he re ferenced pap ers, w hich t herefore re present just a p ointer to t he l iterature. O nce a gain, h owever, t here a re always exceptions, as internal variable models can be formulated with some type of proxy variable that accomplishes the same effect, for example, overall martensite transformation strain as used in the model by Bondaryev and Wayman [8].

range o f app lication to t he c ase o f si mple ps eudoelasticity. I f it is necessary to mo del either the SME or other effects associated wi th r eorientation, th en th ere i s th e n eed t o d istinguish between different martensite microstructures during the operation of the shape memory device, from which it follows that a more rened set of internal variables is required. Among the earliest proposals for such a renement was to distinguish b etween t hermally i nduced m artensite (which h as no net t ransformation s train) a nd p ure s tress-induced m artensite (with t ransformation strain) by splitting x into xthermal and xsim [10]. I n pa rticular, b y a llowing f or t he c onversion s equence: thermal martensite to stress induced martensite (by loading and unloading), to austenite (by heating), to thermal martensite (by cooling), suc h a f ramework p rovides t he si mplest re asonable model of the SME. In suc h mo dels, t he m icrostructure a ssociated w ith xsim restricts consideration to either tension or compression, but not both. I n o rder to t reat b oth ten sion a nd c ompression, f urther renement i s ne cessary. I n pa rticular, b y iden tifying t he mo st compressive a nd most tensile microstructures, indicated above by M− and M+, it is natural to introduce volume fractions x− and x +. Further, by identifying the particular combination of M− and M+ for which there is no net transformation, it is possible to dispense with xthermal [37]. Ā us the internal variable pair a = {x−, x +} provides more generality than the pair a = {xthermal, xsim} in that it continues to a llow for t he modeling of shape memory across the full temperature range, while also permitting compression– tension modeling. Properly formulated, such a = {x−, x +} models allow for tension–compression asymmetry [3]. It is, however, difficult to easily extend a = {x−, x+} models to two or three dimensions. One way to make this dimensional leap is to introduce a n e xtended v ariety o f m icrostructures, e .g., a = {x1, x2,……}. One may then seek to determine the specic fraction of each microstructure on the basis of the stress–temperature history [28,35]. A lternatively, t he t ype o f ma rtensite p resent c an b e described d irectly. I n a f ully t hree-dimensional s etting o ne m ay introduce a t ensor internal variable M de scribing t he deformation between austenite and a homogenization of the portion of the microstructure in the martensite phase. Modeling can then proceed with a = { x, M} i n either a  nite strain setting or in a setting in which the elastic strains are treated as in the conventional linear theory of elasticity [5]. Specialized to one dimension, the type of martensite is a scalar that c an b e iden tied w ith t ransformation s train, suc h a = {x, M} models also give rise to a broadly useful one-dimensional theory that is able to t reat pseudoelasticity, shape memory, and martensite reorientation. In this case, the scalar martensite internal variable M under simplifying assumption (s1) is required to obey −g ≤ M ≤ g. Increasing, decreasing, and unchanging M are then identied respectively with FwR, RvR, and nR.

20.3.4.1.2

20.3.4.2

dropping any of these simplications involve varying degree of additional mathematical detail. Such generalizations range from quite straightforward (e.g., for inclusion of conventional thermal expansion) to more complicated (e.g., additional phases, such as the R-phase, that contribute to the SMM effect). By rst addressing models in such specialized settings, it is also easier to understand the distinction between modeling effects that are energetic (i.e., stemming from y) and dissipative (stemming from Γ) so as to aid in the proper construction of more generalized models. In addition to standard treatments permitting dropping any or all of (s1)–(s8), there is also a clear path to a fully three-dimensional generalization f or t reating m ultiaxial s tress a nd s train. Ā e interested reader may consult Refs. [4,6] for details. 20.3.4.1

Choice of the Internal Variables (Step ii)

Ā e role of the internal variables is central to t he description of hysteresis, w hich i s a ba sic f eature o f t he t hermomechanical response. Ā is is because if y were only dependent on s and T, then t he e quations o f s tate ( Equation 2 0.3) w ould re quire t hat the s ame s train w ould b e su stained b y t he s ame s tress a nd temperature irrespective of whether this was obtained on loading o r u nloading. St ress–strain h ysteresis d uring lo ading a nd unloading is due to the presence of different microstructures, and this effect is captured by the internal variables a. Useful choices for a depend on whether the model of interest is one- or multidimensional and also by the kind of behaviors to be modeled. 20.3.4.1.1

Single Internal Variable

Two Internal Variables

Models i n which x is the only internal variable cannot distinguish between different kinds of martensite (thermal vs. stressinduced, t winned v s. un twinned) a nd s o t his r estricts t heir

43722_C020.indd Sec1:23

Choice of the Free Energy Function (Steps iii and iv)

For the purposes of providing an accurate and tractable model under (s1)–(s8), the case of two internal variables a = {x, M} is

10/3/2008 9:04:37 PM

20-24

Smart Materials

considered. A s d iscussed i n mo re gener ality i n Re fs. [ 4,6], a n appropriate form for the free energy (step iv) in this setting is

∧

M = M (s ) =

s b

∧

⎛ 1 T⎞ y = − s 2 − xM s + c ⎜ T − To − T ln ⎟ − hoAT 2E To ⎠ ⎝ (2 H 1 2 T − Teq x + e oA − Ωx (1 − x ) + bxM + Teq 2

(

)

x = x (s ,T , M ) 0.4)

with m aterial c onstants a s de scribed b elow. Ā is f orm m ay b e generalized in v arious w ays o ften by addition of terms. For example, addition of a term aCTEs(T – To) provides the standard treatment of conventional thermal expansion. We may also mention t he p ossibility to add ing ter ms t hat de pend o n t he s train gradient as in Ref. [34] so as to allow for the modeling of localization phenomena. Ā e constitutive equations for e, h (step iii) now follow from the Gibb’s free energy y(s, T, x, M) via Equation 20.3 as e=

1 T H s + xM , h = c ln − x + hoA (2 E To Teq

0.5)

As i ndicated earlier E is t he elastic modulus, c the specic heat, and H the latent heat of transformation. Morevoer To is any chosen reference t emperature, w hereas Teq is an “equilibrium temperature” at w hich t he pure phases of austenite a nd martensite have the same free energy. Āe term xM gives the transformation strain due to the presence of martensite. In particular, for thermal martensite, it is the case that M = 0 whereupon there is no transformation strain irrespective of the martensite fraction x. Ā e constants eoA and hoA give the SMM internal energy and the SMM entropy of the austenite phase at To when the stress vanishes. Ā e values of eoA and hoA are not of consequence, since the entropy and the internal energy only enter the treatment in a natural way via their changes. From a practical point of view, both of them can be ignored (e.g., set them to zero in any calculation). Ā ere are thus two remaining material constants in y, namely Ω and b. An indication of their role f ollows up on c onsideration o f t he h ypothetical s tate o f a n energy minimizing equilibrium as discussed next. 20.3.4.2.1

Energy Minimization

Because of hysteresis, there may be many equilibrium states consistent w ith a g iven s tress a nd temperature, mo st of w hich a re not energy minimizing. Ā e actual equilibrium state achieved in the operation of an SMM device depends on t he history of t he controllable variables a nd so w ill t ypically not c oincide w ith an energy m inimal s tate. E nergy m inimal s tates, h owever, l ie between the hysteresis extremes and so provide a nat ural reference state. Furthermore, the energy minimal states provide the basic form of the (s,T) state diagram. Ā e hypothetical energ y m inimizing e quilibrium s tate for a given ( s, T) i s de termined b y t he pa ir ( x, M) t hat minimize s y(s, T, x, M) under the simultaneous constraints 0 ≤ x ≤ 1 and −g ≤ M ≤ g. Requiring the vanishing of the appropriate derivatives and solving for the internal variables then gives that

43722_C020.indd Sec1:24

(

(20.6)

)

⎞ b 1 1 ⎛ H T − Teq = − − Ms + M 2⎟ ⎜ ⎟⎠ Teq 2 2Ω ⎜⎝ 2 Ā e rst equation shows that the orientation of the martensite, as given by M, depends solely on the stress, and is determined by proportionality factor 1/b. Hence b may be regarded as a reorientation modulus. On the other hand, it follows from the second equation in Equation 20.6 that the amount of martensite x is proportional to temperature and, for positive H an d Ω, t hat heating y ields mo re a ustenite w hereas c ooling p romotes mo re martensite. In particular, Ω is a coefficient of martensite forma∧ ∧ ∧ tion. S ubstitution o f M = M (s ) in to x s ,T , M (s ) g ives a 2 stress dependence i n t he for m s /4Ωb so that stress of either sign promotes martensite formation. Equation 20.6 does not ac knowledge the constraints 0 ≤ x ≤ 1 and −g ≤ M ≤ g. E nforcing t he c onstraints now g ives t he p recise description of the energy minimal values of x and M. Specically

(

∧ ⎧ −g , if M (s ) < −g ⎪ ∧ ⎪ ∧ M = ⎨ M (s ), if − g ≤ M (s ) ≤ g ⎪ ∧ g , if M (s ) > g ⎪ ⎩

)

(2

0.7)

Ā en, using this M, the energy minimal value of x is ∧ ⎧ 0, if x (s ,T , M ) < 0 ⎪ ∧ ⎪∧ x = ⎨x (s ,T , M ), if 0 ≤ x (s ,T , M ) ≤ 1 (2 ⎪ ∧ ⎪ 1, if x (s ,T , M ) > 1 ⎩

0.8)

In this energy minimal treatment, thermal martensite (x when s = 0) is formed on the temperature interval Teq − ΩTeq/H ≤ T ≤ Teq + ΩTeq H. Note also that a pure martensite microstructure holds at sufficiently low temperature (x = 1 if T ≤ Teq − ΩTeq/H for any s). However, at high temperatures, it is always possible to convert austenite to martensite with sufficiently large stress magnitude (for both positive and negative s). If s ≥ bg, so that M = g, it then follows that a constant value of x is maintained on lines in the (s,T)-plane with slope dT/ds = gTeq/H. Ā is is the Clausius– Clapeyron e quation, a re sult en sured b y t he o verall t hermodynamic consistency of the modeling framework. Ā e more general Clausius–Clapeyron relation is also true: dT/ds = MTeq/H. 20.3.4.3

Choice of the Dissipation Function (Step v) and Derivation of the Transformation Kinetics

Ā e energ y m inimal f ramework provides a ll t he correct t rends; it therefore rem ains to a ugment t he f ramework w ith a n appropriate

10/3/2008 9:04:37 PM

20-25

Shape-Memory Alloys and Effects: Types, Functions, Modeling, and Applications

description of hysteresis. Hysteresis causes the actual microstructure to lag behind that which is associated with energy minimization. A lternatively, to ac hieve pa rticular v alues o f t he i nternal variables a = {x, M}, i t i s gener ally ne cessary to d rive ( s, T) beyond that associated with energy minimization, because of the internal f riction, w hich opp oses m icrostructural c hange. Ā is amounts to a requirement that the thermodynamic forces driving the transformations must attain a nonzero threshold in order to activate a change in the microstructure. In particular, the derivatives of the free energy with respect to t he internal variables are precisely t hese d riving f orces, w ith Πx = − ∂y/∂x dri ving t he martensitic transformation and ΠM = −∂y/∂M driving reorientation. With the expression (Equation 20.4) for y, it follows that

Πx (s ,T , x , M ) = M s −

(

1 2 H T − Teq + Ω (1 − 2x ) − 2 bM Teq (2 0.9)

(

)

)

Π M (s , x , M )= s − bM x

It will be shown below that, whereas the energy minimal microstructures are formed under zero driving forces, in the presence of energ y d issipation t he m artensite f raction x c hanges i f and only if Πx both attains and sustains a threshold level denoted by Λx . Ā is t hreshold re presents t he i nternal f riction a ssociated with t he movement of t he phase boundaries at t he m icroscale. Similarly, m artensite or ientation M c hanges i f a nd o nly i f ΠM both attains and sustains a corresponding threshold L M . Ā is can be modeled by prescribing a constitutive function for the r ate o f energ y d issipation G. A s s hown i n Re f. [ 4], S MM models can be developed by a prescription in the form: . Γ = Λxx + Λ M M
, (2

0.10)

where Lx and L M are constitutive functions that, as shown below, express the abovementioned thresholds for the driving forces. In the most general case, such functions may depend on the current values of s, T, x, M as well as the previous history of microstructural states via past values of the internal variables. Any specication of such functions, combined with the free energy function y, yields a specic model. A simple choice that yields a model of pseudoelasticity, shape memory, and reorientation effects under the assumptions listed above, is the following: ⎧ ⎛ 1 − 2x0 + x0x ⎞ ⎪ Λx ↑ = k x ⎜ ⎟⎠ ≥ 0, if ⎝ 1 − x0 ⎪ Λx = ⎨ ⎪Λ = −k ⎛ x0x + x0 − x ⎞ ≤ 0, if x⎜ ⎟⎠ ⎪ x↓ x0 ⎝ ⎩ ⎧⎪ Λ = k M x ≥ 0, if M
> 0 ΛM = ⎨ M ↑
⎪⎩Λ M ↓ = −k M x ≤ 0, if M < 0

. x >0 . x < 0 (2

(Λ

0.11)

x

. − Πx x + (Λ M − Π M )M
= 0 (2

)

0.12)

One immediate consequence is that an assumption of zero dissipation (Λx = 0 and ΛM = 0) leads to t he t heory of energ y minimal m icrostructure d iscussed a bove. M ore gener ally, t he sig n conditions on Λx and ΛM ensure that hysteresis results from the need to overdrive the system so as to overcome frictional resistance. On this basis, one may now extract the conditions distinguishing b etween F wT, Rv T, o r n T a nd t he c onditions distinguishing between FwR, RvR, or nR. For FwT the conditions that must be satised are: x <1,

∂Πx . ∂Πx T
> 0, s + ∂s ∂T

and

Πx = Λ x ↑ (2 0.13)

Ā ese correspond respectively to requirements that there is (a) room for additional change to x, ( b) a c ontrol v ariable pat h direction o n t he s tate d iagram t hat i s c onducive to c hange, and (c) sufficient driving force to overcome the frictional dissipation. I n v iew o f t he e xpressions f or Λx and Πx , o. ne t hus .  nds t hat t he conditions for FwT a re x < 1 , Ms − HT/Teq > 0, along with Ms −

H 1 T − Teq + Ω (1 − 2x ) − bM 2 Teq 2

(

)

⎛ 1 − 2x 0 + x0 x ⎞ =kx ⎜ ⎟⎠ ⎝ 1−x

(2

0.14)

0

. . Ā e conditions for RvT are: x > 0, Ms −HT/Teq < 0, and Ms −

where, as in t he earlier sublooping discussion, x0 denotes t he fraction o f m artensite at t he en d o f t he l ast t ransformation

43722_C020.indd Sec1:25

process. Here κx and κM are material constants, which must be nonnegative s o a s to p rovide c onsistency w ith t he C lausius– Duhem s econd l aw i nequality. Ā e quot ient f orm o f Λx is a consequence of t he Ivshin hysteresis c ondition (s5), a s me eting this condition requires that Λx↑ is independent of x0 when x = 1, and that Λx↓ is independent of x0 when x = 0. Ā ere a re n ine c ombined k inetic d irection p ossibilities f or a = {x, M} corresponding to t he product of t he t hree transformation possibilities: FwT, RvT, nT, w ith t he t hree reorientation possibilities: F wR, Rv R, n R. E liminating t he h eat supp ly r between the balance equations for energy and entropy (Equation 20.2) and using Equation 20.3, Equation 20.10 leads to

H 1 T − Teq + Ω (1 − 2x ) − bM 2 Teq 2

(

)

⎛ x x + x0 − x ⎞ = −k x ⎜ 0 ⎟⎠ x0 ⎝

(20.15)

If there is neither FwT nor RvT, then there is nT. . Ā e condition for FwR is M < g, s > 0,and s − bM = k M (2

0.16)

10/3/2008 9:04:40 PM

20-26

Smart Materials

. Ā e condition for RvR is M > −g, s < 0, and s − bM = −k M

(20.17)

If there is neither FwR nor RvR, then there is nR. 20.3.4.3.1 Pseudoelasticity In this framework, pure pseudoelasticity corresponds to either (FwT, nR) or (RvT, nR). Ā e standard example involves loading at tem peratures a bove Af wi th e ither M = g or M = −g. I f t he inequality conditions for FwT are met, then the equality condition (Equation 20.14) determines the control variable pair (s, T) associated with activating FwT. Meeting Equation 20.14 serves to dene the function sMs(T,x0) that describes activation triggering via s = sMs(T,x0). Prior to activation x = x0, but after activation, x will increase while x0 remains xed. Similar considerations govern Rv T w ith E quation 2 0.15 re placing E quation 2 0.14. Triggering of either transformation requires a value Πx that is proportional t o t he a mount of t ransforming ( diminishing) phase, i.e., Πx = k x (1 − x0) for activation of FwT, and Πx = − k xx0 for a ctivation o f R vT. C ontinued tr ansformation r equires increasing |Πx |. Transformation completion then requires Πx to attain the value k x for FwT and −k x for RvT. During FwT, the increase of x occurs so as to maintain Equation 20.14. For x0 = 0, this gives the outer loop of the stress– strain curve, while a sublo op is obtained for 0 < x0 < 1. Ā e FwT evolution of x c an be obtained by solving E quation 20.14 for x algebraically. In numerical routines, it would be typical to compute this evolution via the rate form of Equation 20.14, that is −1

⎞ ⎛ ⎛ ξ ⎞ ⎞ ⎛ H ξ
= ⎜ 2Ω + ⎜ 0 ⎟ κ ξ ⎟ ⎜ − T
+ M σ
⎟ (2 ⎝ 1 − ξ0 ⎠ ⎠ ⎝ Teq ⎝ ⎠

0.18)

Reversing t he pat h of t he c ontrol v ariables, e .g., by u nloading, then ac tivates Rv T i f t he d riving f orce Πx at tains t he c orresponding t hreshold Λx↓. I t t hen f ollows f rom E quation 2 0.15 that (RvT, nR) occurs according to the kinetics −1

⎞ ⎛ ⎛ 1 − ξ0 ⎞ ⎞ ⎛ H
ξ
= ⎜ 2Ω + ⎜ κ ξ ⎟ ⎜ − T + M σ
⎟ (2 ⎟ ⎝ ξ0 ⎠ ⎠ ⎝ Teq ⎝ ⎠

0.19)

20.3.4.3.2 Reorientation Pure reorientation corresponds to (nT, FwR) or (nT, RvR). Ā e most important physical case of pure reorientation is that which occurs when the material is fully martensitic (x = 1), for example a conversion of thermal martensite M = 0 to t he maximal stressinduced martensite M = g at tem peratures below Mf. Observing, however, t hat t he conditions governing FwR a nd RvR a re i ndependent of temperature, it follows that the variable M can evolve at a ny temperature. W hen t his evolution takes place w ith x = 0, the c hange i s not e xpressed ph ysically ( this i s b ecause e ach appearance of M in y involves a multiplication by x), h owever such “virtual reorientation” is central to the mathematical model. For e xample at T > Af, t his v irtual re orientation s ets M = g for tensile loading and unloading, and sets M = − g for compressive

43722_C020.indd Sec1:26

loading a nd u nloading. Ā is s etting a nd re setting o f M occ urs during the elastic loading and unloading in the austenite phase, and so does not have an effect before the onset of FwT. In the temperature range Mf < T < Af, one may encounter the four cases (FwT, FwR), (FwT, RvR), (RvT, FwR), and (RvT, RvR). In these cases, one of the equality conditions, Equation 20.14 or Equation 20.15, occurs in tandem with one of the equality conditions, Equation 20.16 or Equation 20.17, so as to give two equations for x and M in terms of s, T, and x0. Once again, the rate form may be practical for numerical computation. For example, for (FwT, FwR), this kinetic rate form yields −1

. ⎛ ⎛ x ⎞ ⎞ ⎛ H s .⎞ x = ⎜ 2Ω + ⎜ 0 ⎟ k x ⎟ ⎜ − T
+ s ⎟ , b ⎠ ⎝ 1 − x0 ⎠ ⎠ ⎝ Teq ⎝

1 . M
= s (2 0.20) b

20.3.4.4 Material Parameter Identification Ā e model pa rameters E, g, c, a nd H a re t ypically available f rom measurement, whereas the model parameter T0 is chosen for convenience on the basis of a standard operating temperature, and model parameters h0A and e0A play no role. Ā is leaves the specication of model parameters: Teq, Ω, b, kM, and k x . Under the simplifying conditions (s1)–(s8) listed earlier, these may be taken as follows: 1 (A + As + M s + M f ) 4 f ⎛ A − As + M s − M f ⎞ Ω=⎜ f H ⎝ A f + As + M s + M f ⎟⎠ 1 b = (s Rf − s Rs ), k M = s Rs g

Teq =

(20.21)

⎛ A + As − M s − M f ⎞ kx = ⎜ f H ⎝ A f + As + M s + M f ⎟⎠ Observe t hat Ω and k x a re each a m ultiple of H > 0 whereupon the ordering Mf < Ms < As < Af ensures that Ω > 0 and k x > 0. It is i nstructive to c onsider t he more general t hermodynamically admissible conditions: Mf ≤ Ms ≤ Af and Mf ≤ As ≤ Af, which still yield Ω ≥ 0 and k x ≥ 0. Here we temporarily allow the possibility of equality in the various relations so as to remark on some special cases. Ā e extremely special case Mf = Ms = As = Af = Teq then corresponds to abrupt thermal transformations A → M and M → A without any hysteresis. In this case, both Ω = 0 and k x = 0. Ā e x versus T graph of an abrupt, hysteresis-free transformation is a s tep f unction at t he c ommon t ransformation t emperature. Ā is step can be modied, or “unfolded,” in two canonical ways: (u1) by making it a ramp, and (u2) by splitting it into two separate vertical steps one to b e t raveled for A → M a nd t he ot her to b e traveled for M → A. Ā e modication (u1) corresponds to a gradual but still hysteresis-free transformation and can be regarded as a pure coexistence unfolding. Ā e modication (u2) corresponds to t wo abrupt t ransformations separated by a h ysteresis interval and can be regarded as a pure hysteresis unfolding. Ā e pure hysteresis unfolding (u2) corresponds to taking Mf = Ms < As = Af in which c ase Ω = 0 and k x > 0. Ā e p ure c oexistence u nfolding (u1) correspond to taking Mf = As < Ms = Af in which case Ω > 0

10/3/2008 9:04:44 PM

Shape-Memory Alloys and Effects: Types, Functions, Modeling, and Applications

and k x = 0. In this sense, Ω is a d irect indicator of phase coexistence a nd k x is a direct indicator of hysteresis. In view of the observed ordering, Mf < Ms < As < Af, it then follows that k x > Ω > 0, indicating the central role of hysteresis in SMM behavior.

20.3.5

Summary and Concluding Remarks

In this contribution we have attempted to provide an overview of the basic modeling issues, and a ne cessarily simple review of the many modeling approaches that describe SMMs. We have also provided a t ractable mo del f or b oth ph ase t ransformation a nd reorientation appropr iate to u niaxial deformation. A s shown by Equation 20.21, this model provides a simple correspondence between physical parameters and modeling parameters. It is to be emphasized t hat t his model naturally generates a s tate diagram, whereupon the issue arises as to how to insure that such a diagram matches experiment. In this regard, the model construction was aided b y t he si mplifying a ssumptions ( s1)–(s8) a lthough, a s emphasized throughout, more general situations can be treated by a suitable extension of the basic formulae provided here. Ā e corresponding ge neralizations re quire v arying le vels of a dditional mathematical framework within the model. For example, replacing Equation 20.11 by a more involved expression allows one to t arbitrary transformation temperatures that are no longer bound by t he re quirement ( s4). Suc h re placements a lso a llow f or t he treatment of more subtle i ssues, suc h a s t he s pecication of t he slopes rr and rf appearing in Figure 20.22. By such means, one can construct su itably f aithful s tate d iagrams, re plicate c omplicated sublooping p rocedures, p redict d issipation a nd d amping, a nd more generally design shape memory devices and structures that perform as intended when incorporated into larger systems.

References 1. Auricchio, F., Taylor, R.L., and Lubliner, J. 1997. Shape-memory alloys: Macromodelling and numerical simulations of the superelastic behavior, Comp. Meth. Appl. Mech. Eng., 146, 281–312. 2. Ball, J.G. and James, R.D. 1987. Fine phase mixtures as minimizers of energy, Arch. Rat. Mech. Anal., 100, 13–52. 3. Bernardini, D. and Pence, T.J. 2002. Shape memory alloys: modeling, I n: Encyclopedia of Smar t M aterials, vol. 2 , pp. 964–980, Wiley, New York. 4. Bernardini, D. and Pence, T.J. 2002. Models for one-variant shape memo ry ma terials bas ed o n dissi pation f unctions, Int. J. Nonlinear Mech., 37, 1299–1317. 5. Bernardini, D. and Pence, T.J. 2003. A multield theory for the modeling of the macroscopic behavior of shape memory materials, I n: Advances in M ultield Theories of C ontinua with S ubstructure, C apriz G. a nd M ariano P .M. (E ds.), pp. 199–242, Birkhauser, Boston, MA. 6. Bernardini, D . a nd P ence, T.J. 2005. U niaxial mo deling o f multivariant shape-memory materials with internal sublooping using dissipation functions, Meccanica, 40, 339–364. 7. Bhattacharya, K. and Kohn, R.V. 1997. Elastic ener gy minimization and the recoverable strains of polycrystalline shapememory materials, Arch. Rat. Mech. Anal., 139, 99–180.

43722_C020.indd Sec1:27

20-27

8. Bondaryev, E.N. a nd Wayman, C.M. 1988. S ome str ess– strain–temperature relationships for shape memory alloys, Metallur. Trans. A, 19A, 2407–2413. 9. Boyd, J.G. and Lagoudas, D.C. 1996. A thermodynamical constitutive mo del f or sha pe memo ry ma terials. P art I. Ā e monolithic shape memory alloy, Int. J. Plasticity, 12, 805–842. 10. Brinson, L.C. 1993. One-dimensional constitutive behavior of shape memo ry a lloys: Ā ermomechanical d erivation wi th non-constant ma terial f unctions a nd r edened martensite internal variable, J. Intell. Mater. Syst. Struct., 4, 229–242. 11. Chang, B .C., S haw, J .A., a nd I adicola, M.A. 2006. Ā ermodynamics of shape memory alloy wire: Modeling, experiments and applications, Cont. Mech. Ther m., 18, 83–118. 12. De La Flor, S., Urbina, C., and Ferrando, F. 2006. Constitutive model of shape memory alloys; theoretical formulation and experimental validation, Mat. Sci. Eng A., 427, 112–122. 13. Elahina, M.H. and Ahmadian, M. 2005. An enhanced SMA phenomenological model: I. Ā e shortcomings of the existing models, Smart Mater. Struct., 14, 1297–1308. 14. Falk, F. 1980. M odel, f ree ener gy, me chanics a nd thermo dynamics of shape memory alloys, Acta Metallur., 28, 1773–1780. 15. Gao, X., H uang, M., a nd B rinson, L.C. 2000. A m ultivariant micromechanical mo del f or S MAs. P art 1. Cr ystallographic issues for single crystal model, Int. J. Plasticity, 16, 1345–1369. 16. Green, A.E. a nd N aghdi, P.M. 1977. On thermo dynamics and the na ture o f the s econd law, Proc. Ro y. S oc. Lo nd. A, 357, 253–270. 17. Ikeda, T ., N ae, F .A., N aito, H., a nd M atsuzaki, Y. 2004. Constitutive mo del for of shape memory alloys for unidirectional loading considering inner hysteresis loops, Smart Mater. Struct., 13, 916–925. 18. Ivshin, Y. a nd Pence, T.J. 1994a. A constitutive mo del f or hysteretic phase transitions, Int. J. Eng. Sci., 32, 681–704. 19. Ivshin, Y. and Pence, T.J. 1994b. A thermomechanical model for a one-variant shape memory material, J. Intell. Mat. Syst. Struct., 5, 455–473. 20. Lacarbonara, W., B ernardini, D ., a nd Vestroni, F . 2004. Nonlinear thermomechanical oscillations of shape memory devices, Int. J. Solids Struct., 41, 1209–1234. 21. Lagoudas, D.C., Entchev, P.B., Popov, P., Patoor, E., Brinson, L.C., a nd Gao , X. 2006. S hape memo ry allo ys, P art II: Modeling of polycrystals, Mech. Materials, 38, 430–462. 22. Liang, C. and Rogers, C.A. 1990. One-dimensional thermomechanical constitutive relations for shape memory materials, J. Intell. Mat. Syst. Struct., 1, 207–234. 23. Likhacev, A.A. a nd K oval, Y.N. 1992. On the diff erential equation des cribing the h ysteretic b ehavior o f sha pe memory alloys, Scrip. Metallur. Mater., 27, 223–227. 24. Matsuzaki, Y. and Kamita, T. 1998. One-dimensional pseudoelastic t heory o f sha pe memo ry a lloys, Smart M ater. Struct., 489–495. 25. Muller, I. and Seelecke, S. 2001. Ā ermodynamic aspects of shape memo ry a lloys, Math. C omputer M odelling, 34, 1307–1355. 26. Ortin, J. 1991. Preisach modeling of hysteresis for a pseudoelastic Cu-Zn-Al single crystal, J. Appl. Phys., 71, 1454–1461.

10/3/2008 9:04:47 PM

20-28

27. Otsuka, K. and Shimizu, K. 1986. Pseudoelasticity and shape memory effects in alloys, Int. Metal Rev., 31, 93–114. 28. Patoor, E., Eb erhardt, A., and B erveiller, M. 1988. Ā ermomechanical behavior of shape memory alloys, Arch. Mech., 40, 755–794. 29. Puglisi, G. a nd Truskinovsky, L. 2005. Ā er modynamics of rate-independent plasticity, J. Mech. Phys. Solids, 53, 655–679. 30. Rajagopal, K.R. and Srinivasa, A.R. 1999. On the thermomechanics of shape memory wires, Z. Angew. Math. Phys., 50, 459–496. 31. Raniecki, B., L excellent, C., and Tanaka, K. 1992. Ā ermodynamic mo dels f or ps eudoelastic b ehavior o f sha pe memory alloys, Arch. Mech., 44, 261–288. 32. Roubicek, T. 2004. M odels o f micr ostructure e volution in shape memory alloys, In: Nonlinear Homogenization and Its Applications to Composites, Polycrystals and Smart Materials, Ponte C asteneda P ., T elega J .J., a nd Ga mblin B . (E ds.), NATO Science Series, Math. Phys. Chem. 170, pp. 65–106, Kluwer. 33. Shaw, J .A. a nd K yriakides, S. 1995. Ā er momechanical Aspects of NiTi, J. Mech. Phys. Solids, 43, 1243–1281. 34. Shaw, J.A. 2002. A thermomechanical model for a 1-D shape memory a lloy wire with propagating in stabilities, Int. J. Solids Struct, 39, 1275–1305. 35. Siredey, N., Patoor, E., Berveiller, M., and Eberhardt, A. 1999. Constitutive eq uations f or po lycrystalline thermo elastic shape memory alloys. Part I. Intragranular interactions and behavior of the grain, Int. J. Solids Struct., 36, 4289–4315. 36. Tanaka, K., Kobayashi, S., and Sato, Y. 1986. Āer momechanics of transformation pseudoelasticity and shape memory effect in alloys, Int. J. Plasticity, 2, 59–72. 37. Wu, X. and Pence, T.J. 1998. Two variant modeling of shape memory materials: Unfolding a phase diagram triple point, J. Intell. Mat. Syst. Struct., 9, 335–354.

Smart Materials

transformation occurs is called the beta (b)-phase. Ā e temperature at which the transformation occurs can be chosen in a temperature r ange b etween −150°C a nd 2 00°C, de pending o n t he composition a nd t he m icrostructural c onstitution, t he l atter determined mainly by the thermomechanical processing. Ā e temperature-induced transformation is characterized by four tem peratures: Ms an d Mf d uring c ooling a nd As an d Af during he ating. Ms and Mf i ndicate t he temperatures at w hich the t ransformation f rom t he pa rent ph ase i nto m artensite respectively starts and nishes, as illustrated in Figure 20.26. A s and A f i ndicate respectively t he temperatures at w hich the reverse transformation (martensite to beta) starts and finishes. The overall transformation describes a hysteresis of the order of 10–50°C. At a temperature just above the A f temperature ( limited to a f ew ten s of ° C), t he m artensitic ph ase c an also b e i nduced b y s training t he s ample, a s i llustrated i n Figure 20.27. After re aching a c ritical s tress, c alled σp-m, t he s ample w ill start to t ransform to m artensite. Ā e v alue o f σp-m in creases linearly with temperature (averagely 5 MPa K−1 for polycrystalline material), starting from zero at the Ms tem perature. Ā e relation between the stress at which martensite is induced and the temperature is given by the Clausius–Clapeyron equation. During f urther straining, t he stress at w hich t he t ransformation occurs is almost constant until the material is fully transformed. Fu rther s training w ill no w le ad to el astic lo ading o f the m artensite, f ollowed b y p lastic de formation. W hen t he strain is limited to t he start of t he elastic loading of t he martensite a nd t he app lied s tress i s rele ased a gain, t he re verse transformation will occur at a lower stress level as during loading, leading to the reverse movement of the rst induced strain. Ā e material is said to be superelastic. Reversible strains up to 8% of the initial length can be obtained, compared to the 0.2% elastic strain of normal metallic material. Since this superelastic e ffect i s d ue to a ph ase t ransformation, t his e ffect is a lso

20.4 Shape Memory Alloys Jan Van Humbeeck

20.4.1 SMAs and the Martensitic Transformation [1–4,36] Ā e distinctive functional properties of SMAs are closely linked to a solid–solid phase transformation occurring in a metastable solid state of some specic alloys. It occurs diff usionless and is therefore c alled m artensitic. Ā e s olid ph ase ob tained d uring cooling i s c alled m artensite. Ā e pa rent ph ase i n w hich t he

43722_C020.indd Sec1:28

100 Volume transformed (%)

SMAs a re f unctional o r ad aptive m aterials w ith t he f ollowing properties: sup erelastic b ehavior, S ME ( reversible s trains o f several p ercent d uring h eating o r c ooling o ver a l imited tem perature ra nge), g eneration o f h igh r ecovery st resses, a w ork output with a high power and weight ratio, and a high damping capacity. All these properties are more extensively described in Part 3.

10–50 K Mf

As

Ms

0

Af

5–30 K

Temperature

FIGURE 20.26 Schematic representation of the volume transformed as a function of temperature.

10/3/2008 9:04:47 PM

20-29

Stress

Shape-Memory Alloys and Effects: Types, Functions, Modeling, and Applications

TABLE 20.2 Classication of the Nonferrous Martensites Group Austenite

Austenite

Martensite

Martensite

Alloy System

Terminal solid solutions based on an element having allotropic phases

Cobalt and its alloys Rare earth metals and their alloys Titanium, zirconium and their alloys Alkali metals and their alloys and thallium Others as Pu, Ur, Hg, and alloys

Intermetallic solid solutions with a bcc-parent phase

β-Hume–Rothery phases of the Cu-, Ag-, Au-based alloys

Alloys showing cubic to tetragonal trans. (incl. quasi-martensite)

β-Ni–Al alloys Ni-Ti-X alloys Indium-based alloys Manganese based alloys (paramagnetic ↔ antiferromagnetic) A15-compounds Others: Ru-Ta, Ru-Nb, Y-Cu, LaCd, LaAgx–In1−x

Strain

FIGURE 20.27 Superelastic behavior at constant temperature due to stress-induced transformation.

called m any t imes s uch as p seudoelasticity, p seudoelastic effect, or superelasticity. Ā e re verse de formation o ccurs at a lower stress plateau as during loading and shows thus an hysteresis, analogous to the temperature-induced transformation, but with a size of 50–300 MPa. It follows t hat t he volume transformed fraction due to tem perature- or stress-induced transformation cannot be characterized by a single value of temperature or stress. It is also important to not ice t hat many characteristics such a s Young’s modulus and electrical resistivity can change drastically during transformation. Ā e origin of the functional properties lies in the specic microstructure of the martensite. A single crystal of the parent phase ( i.e., a g rain) t ransforms to a c ollection o f m artensite crystals, c alled v ariants, w hich a re s eparated b y i nterfaces. Ā ey g row t hermoelastically i n a s elf-accommodating wa y. When this microstructure is deformed, reorientation of the variants occurs by displacement of the interfaces, in such a way that t he v ariants, w hich l attice s hear ac commodates, t he applied s train w ill b e f avored at t he e xpense of t he ot her le ss favored or iented v ariants. Ā is occ urs ( principally) w ithout plastic deformation. Ā e strain obtained during reorientation is thus limited. In the u ltimate c ase, a si ngle c rystal ( variant) o f m artensite i s obtained. Fu rther s training w ill le ad to p lastic de formation, hampering the functional properties. When a re oriented material is heated above its transformation temperatures, it will transform to the original parent phase orientation, giving rise to the so-called SME. During pseudoelastic deformation, the variants will be introduced, w hich ac commodate t he app lied s train u ntil t he w hole sample has been transformed into martensite, after which classic (elastic + plastic) deformation of the sample occurs.

20.4.2

SMA Systems [5,6]

Many s ystems e xhibit a m artensitic t ransformation. G enerally they are subdivided in ferrous and nonferrous martensites. Ā ey

43722_C020.indd Sec1:29

Source: F rom D elaey, L., Cha ndrasekaran, M., Andrade, M., a nd Van Humbeeck, J., Proceedings of the International Conference on Solid-Solid Phase Transformation, Met. Society of AIME, 1982, p. 1429.

are summarized in Tables 20.2 and 20.3. From all systems mentioned in both tables, in fact only one major system became successful in i ndustrial application: NiTi(X,Y) i n w hich X ,Y a re elements replaced by Ni or Ti. Besides the NiTi system, a lot o f attention has been given in earlier times to Cu-based alloys and to Fe–Mn– Si-based alloys. Furthermore, during recent years, special attention has been given to high-temperature SMAs (HTSMA). 20.4.2.1

Fe-Based Alloys [6–10,37]

In f errous a lloys, t he a ustenite ( FCC-γ p hase) ca n b e t ransformed to t hree k inds of martensites depending on composition o r s tress: γ–α′ (B CC), γ → ε (H CP), a nd γ → FC T martensite. A lthough a S ME h as b een ob served f or a ll t hree

TABLE 20.3 Ferrous Alloys Which Exhibit a Complete or Nearly Complete SME

Alloy Fe–Pt

Composition ≈ 25 at.% Pt ≈ 25 at.% Pt

Fe–Pd

Crystal Structure of Martensite

Nature of Transformationa

bct (α′) fct

TE

fct

TE

TE

≈ 30 at.% Pd 23%Ni–10%Co–4%Ti

bct (α′)

33%Ni–10%Co–4%Ti

bct (α′)

TE

Fe–Ni–C

31%Ni–0.4%C

bct (α′)

Non-TE

Fe–Mn– Si

30%Mn–1%Si

hcp (ε)

Non-TE

28%–33%Mn–4%– 6%Si

hcp (ε)

Non-TE

Fe–Ni– Co–Ti

—

Source: From M aki, T . a nd T amura, T ., Proc. I COMAT-86, J apanese Institute of Metals, 1986, p. 963. a TE, thermoelastic martensite; non-TE, nonthermoelastic martensite.

10/3/2008 9:04:47 PM

20-30

Smart Materials

types of transformation, most attention to develop a commercial alloy has been given to the alloys with a γ → ε transformation. Ā ese alloys have a low stacking fault energy of austenite (Fe–Cr–Ni, Fe-high Mn alloys). Ā e austenite to ε- martensite transformation p roceeds b y t he a/6 [ 112] S chockley pa rtial dislocations t railing a s tacking f ault r ibbon o n e very { 111} austenite plane and changing the crystal structure to m artensite. Ā e S ME, w hich i s o f t he o ne-way t ype, m ainly re sults from a re verse mot ion o f t he S chockley pa rtial d islocations during heating. Any factor impeding the reversibility of the motion of partial dislocations will lead to an incomplete recovery and in turn to a p oor S ME. O n a n a verage, a S ME o f a bout 3 % c an b e obtained with recovery stresses of a few 100 MPa. Ā e internal factors controlling the recovery are alloy composition, N éel tem perature, t ransformation tem perature, l attice defects, a mong ot hers. E xternal f actors a re app lied s tress a nd strain, deformation, recovery annealing temperature, and thermomechanical treatment. 20.4.2.2 Cu-Based Alloys [5,11–15] Copper-based S MAs a re m ainly der ived f rom Cu– Zn a nd Cu–Al binary systems. Both show a m artensitic t ransformation. T he c omposition r ange o f t hese a lloys c orresponds to that o f th e w ell-known β Hume–Rothery phase. In most SMAs, this phase has a disordered bcc structure at high temperatures b ut o rders to a B 2, D 03 or L 21 form at lower temperatures. The shear elastic constant of t he β-phase exhibits an anomalous behavior with decreasing temperature, i.e., it is lowered t ill t he l attice i nstability w ith re spect to { 110}<110> shears at some temperature transforms β-phase t o ma rtensite. T he tem perature o f t ransformation to ma rtensite, Ms , varies w ith t he a lloy c omposition. T he el astic a nisotropy o f the β-phase is much higher compared to normal metals and alloys and increases further as the martensitic transformation is approached. Cu–Zn and Cu–Al martensites are of three types: α′, β′, or γ ′ with subscript 1, 2, or 3 added to indicate the ordered lattices in β, viz. B2 (2) or D03 (1) or L2 (3). Some conversion from one martensite structure to another, e.g., β′ → γ ′ may also take place. Ā e net result is a coalescence of plates within a self-accommodating group a nd e ven c oalescence o f g roups. H eating t his de formed martensite microstructure transforms it to the β-phase with the SME accompanying the structural change.

Table 20.4 summarizes three systems that were once considered as potential SMA, but so far only with limited success. Ā e main p roblems en countered a re t he l imited d uctility d ue to a rather large grain size and large crystal anisotropy, stabilization of martensite as function of time and temperature, poor corrosion re sistance, u nderperformance for f unctional pr operties when compared to NiTi. 20.4.2.3 Ni–Ti Alloys [4,7,16] Ni50Ti50 a nd t he ne ar e quiatomic Ni– Ti a lloys a re t he b est explored system of all shape memory alloys and occupies almost the whole market of SMA. Ni50Ti50 is an intermetallic phase that has some solubility at higher temperature. Ā e science and technology of Ni–Ti is overwhelmingly documented. Ā e inuence o f c omposition a nd t hermomechanical processing on the functional properties is well understood and described i n l iterature. Ā e ba sic c oncept o f p rocessing Ni–Ti alloys is that in order to avoid plastic deformation during shape memory o r ps eudoelastic lo ading, t he m artensitic a nd t he β-phase have to be strengthened. Ā is occurs by classic methods: strain hardening during cold deformation, solution hardening, and p recipitation h ardening. Ni–Ti a lloys h ave t he sig nicant advantage that these techniques can be easily applied due to a n excellent d uctility a nd a v ery i nteresting b ut c omplicated p recipitation process (Table 20.5). Ā e c ompositions o f t he Ni– Ti–SMA a re app roximately between 48 and 52 at.% Ni and the transformation temperatures of t he B2 structure to t he martensitic phase w ith a mo noclinic B19′ structure are very sensitive to the nickel content (a decrease of about 150° for an increase of 1 at.% Ni). Ā e transformation temperatures can be chosen between −40°C and +100°C. Ni–Ti a lloys s how t he b est s hape memo ry b ehavior o f a ll SMAs. Even in polycrystalline state, 8% shape recovery is possible and 8% pseudoelastic strain is completely reversible above Af, while the recovery stress is of the order of 800 MPa. In some cases, the martensitic transformation is preceded by the s o-called R -phase t ransition. Ā e R -transition i s a B 2 ↔ rhombohedral t ransformation t hat h as a lso s econd-order characteristics. Ā e most specic characteristics of this R-phase transition is that it shows a clear one- and two-way memory effect of the order of 1% recoverable strain and that the hysteresis of the transformation is very small, only to the extent of a few degrees, which creates possibilities for accurately regulating devices.

TABLE 20.4 Actual (A) Industrial Cu-Based Alloys

Base Alloy

Composition Wt%

Ms (°C)

Hyst. (°C)

Other Alloying Elements in Solution (%)

Current Grain Rening Elements Producing Precipitates

Cu–Zn–Al

5–30 Zn 4–8 Al

−190 to +100

10 (β′)

Ni (−5%) Mn (−12%)

Co (CoAl); B (AlB2); Zr (?); B, Cr (CrxBy)

Cu–Al–Ni

11–14.5 Al 3–5 Ni

−140 to +200

10 (β′) 40 (γ′)

Mn (−5%)

Ti ((Cu, Ni)2 TiAl); B (AlB12); Zr (?)

Cu–Al–Be

9–12 Al 0.4–1 Be

−80 to +80

6 (β′)

Ni (−5%)

B (AlB2 or AlB12) Ti (Cu2TiAl)

43722_C020.indd Sec1:30

10/3/2008 9:04:48 PM

Shape-Memory Alloys and Effects: Types, Functions, Modeling, and Applications

TABLE 20.5

20.4.2.4

Property Values of Ni-Ti and Cu-Based SMAs

Property Physical Melting point Density Ā ermal conductivity of austenite or martensite Coefficient of thermal expansion of austenite or martensite Specic heat Transformation enthalpy Corrosion performance

Unit °C 103 kg m−3 W m−1 K−1 10−6 K−1

J kg−1 K−1 J kg−1

Wear resistance

Ni-Ti

Cu-Based

1300 6.5 18 8.6 11 6.6

1000 7.8 120 — — 16

500 20,000 Similar to 300 series stainless steel Good

400 10,000 Fair

1.0 0.8 <1.002 3 × 10−6

0.1 0.1 — — 80 — 800

—

Electromagnetic Resistivity of austenite or martensite Magnetic permeability Magnetic susceptibility

<1.002 emu g−1

Mechanical Young’s modulus (austenite) G (austenite) Ultimate tensile strength

GPa GPa MPa

Elongation at failure Fatigue strength N=106

%% MPa

80 27 700–1100 (annealed) 1300–2000 (not annealed) 30–50 350

Shape memory Transformation temperatures

°C

–100 to 100

°C %% °C

5–50 3–8 300

–200 to +200 5–20 4–6 200

%%SDC %% J g−1 MPa MPa K−1 J g−1

15 6–8 6.5 600–900 5–15 4

80 4 1.8 600 2–5 1

Hysteresis One-way memory strain Maximum temperature (short time) Damping capacity Superelastic strain Superelastic energy storage Maximum recovery stress Stress rate Work output

10−6 m

10–15 270

It should be noted that further cooling transforms the R-phase into B19′ martensite. During heating, generally only t he re verse martensitic t ransformation w ill be obser ved. It ha s be en shown that t he app earance o f t he R -phase de pends o n c omposition, alloying elemen ts, a nd t hermomechanical p rocessing. I n f act, the major common point is that all effects depressing the martensitic forward t ransformation below room temperature w ill favor the app earance o f t he R -phase t ransition, w hich i s qu ite s table near 30°C.

43722_C020.indd Sec1:31

20-31

Ternary Ni–Ti Alloy Systems [17–21]

Ā e addition of a third element opens up even more possibilities for adapting binary Ni–Ti alloys toward more specic needs of applications. Adding a third element implies a relative replacement of Ni or Ti. Ā erefore it must be always very well indicated which atom Ni or Ti, or both, is replaced by the third element. Alloying third elements will inuence not o nly the transformation temperatures but w ill a lso have a n e ffect on hysteresis, strength, d uctility, s hape memo ry c haracteristics, a nd a lso o n the B2→(R)→B19′ sequence. In more application-oriented cases, one can distinguish four purposes to add third elements: 1. To decrease (Cu) or increase (Nb) the hysteresis 2. To lower the transformation temperatures (Fe, Cr, Co, Al) 3. To increase the transformation temperatures (Hf, Zr, Pd, Pt, Au) 4. To strengthen the matrix (Mo, W, O, C) Some of t he ter nary a lloys have been de veloped for large-scale applications. We will only summarize the two most well developed: Ni–Ti–Cu and Ni–Ti–Nb. 20.4.2.4.1

Ti–Ni–Cu [22,23,38]

Ternary Ti–Ni–Cu alloys in which Ni is substituted by Cu are certainly as important as binary TiNi. Increasing the Cu content decreases the deformation stress for the martensite state and decreases also the pseudoelastic hysteresis without affecting signicantly the Ms temperature. However, more than 10% Cu addition embrittles the alloys, hampering the formability. It s hould a lso b e rememb ered t hat w hile T i–Ni t ransforms from a B2 into a monoclinic phase, Ti–Ni–Cu exceeding 15 at.% Cu transforms from a B2 into an orthorhombic phase. Ti–Ni–Cu with less than 15 at.% Cu transforms in two stages. 20.4.2.4.2

Ti–Ni–Nb [24,25]

Ā e i nherent t ransformation h ysteresis of Ni– Ti–Nb i s l arger than for binary Ni–Ti alloys. By the presence of a large dispersed volume fraction of deformable β-Nb particles, hysteresis can be further w idened b y a n o verdeformation o f t he s tress-induced martensite, generally between Ms and Md. Originally Ni–Ti–Nb (more specically Ni47Ti94Nb9) was developed by Raychem Corp. for clamping devices. Ā e large shift of the reverse transformation tem peratures, f rom b elow to ab ove ro om tem perature b y deformation, allows room storage for opened couplings. Recently, also pseudoelastic Ni–Ti–Nb alloys have been developed with three signicant differences relative to binary alloys • Stress rate is much lower. • Ā e σ p-m stresses are much higher. • Ā e superelastic window is much larger. 20.4.2.5 High-Temperature SMAs [26,27] Actual SMAs are limited to maximal Af temperatures of 120°C, Ms gener ally b eing b elow 1 00°C. H owever, si nce m arket

10/3/2008 9:04:48 PM

20-32

Smart Materials

demands on SMA have expanded greatly, the need for SMA transforming at higher temperatures t han presently available is increasing. Ā e main application areas of interest are actuators in automobile and oil industry and in safety devices. Ā er e is also an interest in robotics since SMAs with high transformation temperatures allow faster cooling, which would signicantly i ncrease t he ba ndwidth i n w hich t he rob ot s hould operate. In spite of the fact that many alloy systems show high transformation temperatures, no r eal large-scale applications have b een developed. No major breakthrough has been reported yet mainly due to the f ollowing problems: (much) lower performance than the successful Ni–Ti alloys, stabilization of the martensite, decomposition of the martensite or parent phase, brittleness due to high elastic an isotropy or du e to t he pre sence of br ittle ph ases or precipitates. Another condition for a good SME is that the stress to induce martensite or the stress to reorient martensite is (much) lower than t he critical stress for normal slip. Since t his critical stress for slip is generally decreasing with increasing temperature, the above condition is quite difficult to ful l, especially at high temperatures. A potential HTSMA should thus be designed at such a composition or thermomechanical treatment that strengthening mechanisms a re incorporated to i ncrease t he critical stress for slip. Ni–Ti–Pd, Ni–Ti–Hf, Ni–Ti–Zr, Cu–Al–Ni, a nd Cu–Zr alloys a re i n de creasing i mportance t he mo st p romising a lloy systems. 20.4.2.6 Other Types of SMAs [28,29,39] 20.4.2.6.1

β-Ti Alloys

In spite of the good biocompatibility of Ni–Ti alloys, doubts remain o n t he lo ng-term s tability o r o n t he d anger o f bad surface treatment, leading to Ni leaching. Since Ni is known for his high a llergic reaction, Ni-less SMAs could be attractive. Such alloys might be developed based on the allotropic transformation in Ti, a h ighly biocompatible material. Pure titanium shows an allotropic transformation from β (bcc) to α (hexagonal) at 1265 K. Transition elements (TM) stabilize the β-phase. T hus t he tem perature o f t he ( α + β)/β tr ansition decreases with increasing concentration of the alloying element. β-Phase Ti alloys can be martensitically transformed if they are quenched from the stable β-phase. Two types o f m artensite, re spectively α′ and α² ca n b e formed, depending o n t he c omposition a nd t he s olution t reatment conditions. Ā e α′-martensite is hexagonal, while α² has a n or thorhombic structure. It is the α²-martensite t hat s hows t he S ME. T i–Mo and T i–Nb a re t he ba sic s ystems t hat p resently re ceive mo st attention. Generally, a sha pe r ecovery in the o rder o f 3% ca n b e obtained bas ed o n stra in-induced ma rtensite a nd r ecovery stresses up to 170 MPa have been reported. Ā e disadvantage is that those alloys are very prone to stabilization and decomposition due to the fact that the β-phase is retained after quenching

43722_C020.indd Sec1:32

in i ts met astable st ate a nd co mpetes wi th ω-phase d uring quenching. Also sp inodal de composition o f α²-martensite i n Ti–Mo and Ti–Nb has been observed. Ā e sensitivity to decomposition at mo derate temperatures is less, if not, important at room tem perature. Ā er efore, pseudoelastic β-Ti allo ys co uld offer an interesting alternative to Ni–Ti alloys, for example, for orthodontic wires. 20.4.2.6.2 Magnetic Shape Memory Alloys [30] Magnetic shape memory (MSM) alloys are a new class of adaptive m aterials t hat c hange t heir s hape w hen e xposed to a m oderate ma gnetic  eld. Ā e adv antage o f M SM c ompared to “classic” SMAs i s t heir a bility to gener ate up to 1 0% s trains at f requencies w ell a bove 1 00 H z. Ā e M SM me chanism i s based on the martensite twin boundary motion driven by the external ma gnetic  eld w hen t he m aterial i s i n c omplete martensite state. Ā e martensite crystal, after transformation from cubic austenite, consists of a mixture of tetragonal martensite variants, having different c-axis orientation, separated by twin b oundaries. When the MSM material is exposed to an external magnetic  eld, the twins in a favorable orientation relative to the eld direction grow at the expense of other twins. Magnetic  eld i ncreases t he a mount o f t win v ariants o f “preferable” or ientation re lative to t he  eld, i .e., t he t win boundaries mo ve i n t he m aterial. A s a c onsequence, M SM exhibit giant MFIS b ut w ith a c lear lo wer w ork o utput volume ratio as “classic” shape memory a lloys. For t he t ime being, t he MSM effect was observed in Ni 2MnGa, Fe–Pd a nd Co–Ni–Al alloys. Among them, the near stoichiometric single crystalline N i 2MnGa a lloy h as s hown b y f ar t he b est s troke performance. Because o f t heir l arge energ y o utput, M SM m aterials s how enormous p otential i n h igh-stroke dy namic ac tuator app lications. O n t he ot her h and, w hen t he m aterial i s sub jected to mechanical s training, t he t win v ariants r eorient, t oo, w hich alters t heir magnetization a nd t he surrounding magnetic  eld. Ā is ph enomenon c an b e u sed f or s ensor app lications a nd f or voltage generation. One shortcoming of MSM materials in practical applications is their temperature range, which is limited by a maximum operating tem perature. Ā is tem perature is determine d b y st ructural and magnetic characteristics of the alloy.

20.4.3 Functional Properties of SMAs [31,32] Ā e temperature-induced transformation is the basis of the oneand t wo-way memo ry e ffect, of t he ge neration of re covery stresses, a nd o f w ork p roduction. Ā e s tress-induced t ransformation is the basis of the superelastic effect. When the material is martensite, it is said to be in its cold shape, when the material is in the beta phase, it is said to be in its hot shape. Ā e initial shape that the material obtains after processing i s a lways t he h ot s hape. Ā is m eans t hat a fter cooling

10/3/2008 9:04:48 PM

20-33

Shape-Memory Alloys and Effects: Types, Functions, Modeling, and Applications

D

A

T < M f T < M f T < M f T < Af F

F

B

T (a)

(b)

(c)

(d)

L C

FIGURE 20.28 One-way memory effect. Ā e sample is deformed (A → B) and unloaded (B → C) at a temperature below Mf. Ā e apparent plastic deformation i s re stored du ring he ating to a t emperature a bove Af (C → D). L ength c hange, lo ad a nd temperature a re i ndicated by L , F, a nd T, respectively.

into martensite the dened cold shape is equal to the initial hot shape. Ā is s hape c an no w b e f urther de formed to t he  nal cold shape ( limited to le ss t han 10% strain deformation of t he initial shape) to induce the specic properties as described now. (i)

One-way S ME: Ā e specic f unctional p roperties o f SMAs a re si xfold a nd a re i llustrated i n F igures 2 0.28 through Figure 20.32 by the example of a shape memory spring. I n t he m artensitic s tate, t he s hape memo ry element c an b e e asily de formed f rom a  xed s o-called “hot shape,” retained a fter c ooling to a lmost a ny “ cold shape” (Fi gure 2 0.28). Ā e o nly re striction i s t hat t he deformations may not exceed a certain strain limit (up to 8%) a nd should be covered solely by t he reorientation of the martensite variants, which leads to a rem nant deformation without inducing plasticity. Ā ese apparent plastic deformations can be recovered completely during heating, resulting in the original hot shape. Ā is effect is called the one-way memory effect, since only the hot shape is memorized. Ā e t ransition tem peratures b etween t he “ cold shape condition” and “hot shape condition,” As and Af are determined by the alloying and processing parameters.

(i i) Two-way m emory e ffect refers to the memorization of two shapes ( Figure 2 0.29). A c old s hape i s s pontaneously obtained d uring c ooling. D ifferent f rom t he o ne-way memory effect, no external forces are required for obtaining t he memorized cold shape. During subsequent heating, t he o riginal h ot s hape i s re stored. Ā e two-way memory effect is only obtained after a specic thermomechanical treatment, called training. (iii) When t he shape recovery f rom t he cold shape to t he hot shape i s i mpeded, h igh re covery s tresses a re g radually generated during heating, as illustrated in Figure 20.30. Stresses up to 800 MPa can be obtained. (iv) Shape recovery during heating can be biased by an external force (Figur e 20.31). Ā e r esulting mo tion a gainst a b ias force corresponds to work production, up to 5 J g−1. (v) SMEs described above require temperature changes. In contrast, t he superelastic eff ect (Figure 20.32) is isothermal i n nat ure a nd i nvolves t he s torage o f p otential energy. Isothermal loading of the shape memory element in t he h ot s hape c ondition re sults i n l arge re versible deformations (up to 8%) at nearly constant stress levels (see a lso F igures 2 0.27 a nd 2 0.32). T he de formations

A=C T < Af

T < Mf

T < Af

F T (a)

(b)

(c)

L

FIGURE 20.29 Two-way me mory e ffect. A s pontaneous s hape c hange o ccurs du ring c ooling to a t emperature b elow Mf (A → B). Ā is shape change is recovered during subsequent heating to a temperature above Af (B → C).

43722_C020.indd Sec1:33

10/3/2008 9:04:49 PM

20-34

Smart Materials

E T < M1 T < Mf T < Mf As
A F B

D C

(a)

(b)

(c)

(d)

T L

(e)

FIGURE 20.30 Generation of shape recovery stresses. Ā e sample is deformed (A → B) and unloaded (B → C) at a temperature below Mf. Recovery stresses are generated during heating (D → E) starting from the contact temperature Tc (D), situated between As and Af. T < Mf

T < Mf T < Mf

T < Mf

T > Af

F

W W E

(a)

(b)

(c)

(d)

(e)

D

A

F T B L C

FIGURE 20.31 Work output. Ā e sample is deformed at a temperature below Mf (A → B), followed by unloading (B → C), and repeated loading with a bias weight W (C → D). Shape recovery occurs at an opposing force W during heating to a temperature above Af (D → E), so work is done.

F

T > Af T > Af T > Af

A=C

F B

(a)

(b)

T

L

(c)

FIGURE 20.32 Superelastic effect. Ā e sample is strongly deformed (A → B) at a temperature above Af. During subsequent unloading a complete shape recovery occurs (B → C). Ā e area enclosed between the loading and unloading curves is a measure of the energy dissipated in one superelastic cycle.

43722_C020.indd Sec1:34

10/3/2008 9:04:49 PM

Shape-Memory Alloys and Effects: Types, Functions, Modeling, and Applications

are c ompletely re covered at a lo wer s tress le vel during unloading. Ā ese stress levels are alloy- and temperaturedependent. In general, the stress levels increase linearly (2–15 MPa K−1) with temperature. (vi) SMAs a lso have a h igh damping capacity in t he martensitic state and t wo-phase condition. Ā ese a lloys show in the c old s hape c ondition a s trong a mplitude-dependent internal f riction. For i mpact lo ads, t he s pecic damping capacity can be as high as 90%. In the hot shape condition, energy is dissipated during superelastic cycling as a result of t he s tress hysteresis b etween sup erelastic lo ading a nd unloading (Figure 20.27). Ā e functions and applications linked with those properties can be divided into ve categories. (i) One- a nd t wo-way memo ry e ffect c an b e u sed f or f ree recovery applications. Ā is refers to applications in which the single function of the SMA-element is to cause motions without any biasing stress. (ii) Generation of recovery stresses can be used for diverse clamping a nd  xation de vices, suc h a s S MA-couplings and SMA-connectors, ranging from very small diameters (<1 mm) to very large diameters (>1 m). (iii) Diverse actuation applications have been developed based on the work production capacity of SMAs. (iv) Pseudoelasticy is used in many biomedical and other superelastic appl ications w here h igh re versible s trains have to be combined with high-stress plateaus. (v) Damping applications can be developed based on the high damping capacity of SMAs. Ā e high damping capacity of the martensitic phase is i nteresting for pa ssive damping. Ā e superelastic hysteresis might be interesting for earthquake damping and isolation purposes. In the past several years, SMAs have found their specic niches in m any d omains o f in dustrial a ctivities. A s teadily gr owing amount of different SMA applications are now produced at large volumes. Detailed descriptions of many successful or potentially successful a pplications c an be found i n r ecent conference proceedings from the well-known conferences ICOMAT, ESOMAT, SMST, a nd s ome lo cally o rganized c onferences a s w ell a s o n many of the very instructive and informative Web sites. Ā e bibliography at t he end of the reference list provides a l ist of books related to martensite and SMAs.

4. 5.

6. 7. 8. 9. 10. 11.

12.

13. 14.

15. 16. 17.

References 1.

Funakubo H., Shape Memory Alloys, 1984, Gordon & Breach, London. 2. Delaey L., Diff usionless t ransformations, in Materials Science a nd T echnology, Vol. 5 : Phase T ransformations i n Materials (H aasen, P ., e d.), 1991, p p. 339–404, VCH Verlagsgesellschaft mbH, Weinheim, Germany. 3. Wollants P., Roos J.R., and Delaey L., Ā ermally and stress induced thermo elastic ma rtensitic t ransformations in the

43722_C020.indd Sec1:35

18.

19.

20-35

reference f rame o f e quilibrium thermo dynamics, Prog. Mater. Sci., 37(3), 1993, 2427–2488. Otsuka K. a nd Wayman C.M., Shape Me mory Ma terials, 1998, 284 pp., Cambridge University Press, Cambridge. Delaey L., Cha ndrasekaran M., Andrade M., a nd Van Humbeeck J., Non-ferrous martensites, classication, crystal structure, morphology, microstructure, Proceedings Int. Conf. S olid-Solid P hase T ransformation, M et. S ociety o f AIME, 1982, pp. 1429–1453. Maki T. a nd Tamura T., S hape memo ry eff ect i n f errous alloys, Proc. ICOMAT-86, Japanese Institute of Metals, 1986, pp. 963–970. Sato A., Sama K., Chishima E., and Mori T., Shape memory effect a nd me chanical b ehaviour o f a n Fe-30Mn-1Si alloy single crystal, J. Physique, 12 C4, 1982, 797–802. Sato A., Chishima E., Soma K., and Mori T., Shape memory effect in γ ↔ ε transformation in Fe-30Mn-1Si alloy single crystals, Acta Metal., 30 1982, 1177–1183. Gu Q ., Van H umbeeck J ., a nd D elaey L., A r eview o f the martensitic tra nsformation a nd sha pe memo ry eff ect in Fe-Mn-Si alloys, J. Physique IV, C3 4, 1994, C3-135–140. Maki T., Ferrous shape memory alloys, Chapter 5 in Shape Memory Ma terials (Otsuka K. a nd Wayman C.M., e ds.), Chapter 5, 1998, 284 pp., Cambridge University Press. Van Humbeeck J. and D elaey L., A comparative review of the (potential) shape memory alloys, The Martensitic Transformation in Science and Technology (Hornbogen E. and Jost N., eds.), 1989, pp. 15–25, DGM, Verlag, Oberursel, Germany. Van H umbeeck J ., Cha ndrasekaran M., a nd S talmans R ., Copper bas ed sha pe memo ry allo ys a nd the ma rtensitic transformation, P roc.-ICOMAT-92, 1993, p p. 1015–1025, Monterey Institute of Advanced Studies, Monterey, CA. Tadaki T., Cu-based shape memory alloys, Shape Memory Materials (Otsuka K. a nd Wayman C.M., e ds.), 1998, 284 pp., Cambridge University Press, Cambridge (Chapter 3). Ahlers M., Martensite and equilibrium phases in Cu-Zn and Cu-Zn-Al allo ys, Progress i n M aterials S cience, vol. 3 0 (Christian J.W., Haasen P., and Massalski T.B., e ds.), 1986, pp. 135–186. Wu S.K. a nd Min g H., C u-based sha pe memo ry allo ys, Engineering Aspects of Shape Memory Alloys (Duerig T. et al., eds.), 1990, pp. 69–88, Butterworths Scientic, London. Otsuka K. and Ren X., Physical metallurgy of Ti-Ni based shape memory alloys, Prog. Mater. Sci., 50, 2005, 511–678. Eckelmeyer K.H., Ā e effect of a lloying on t he s hape memory pheno menon in N itinol, Scripta M et. 10, 1976, 667–672. Honma T., Matsumoto M., Shugo Y., and Yamazaki I., Effects of addition of 3d transition elements on the phase transformation in TiNi compound, Proceedings ICOMAT-79, 1979, pp. 259–264. Huisman-Kleinherenbrink P atricia, On the M artensitic Transformation T emperatures of N iTi and Ā eir Dependence o n Alloying Elemen ts, P hD Ā esis, U niversity o f Twente, the Netherlands, 1991.

10/3/2008 9:04:50 PM

20-36

20. Kachin V.N., Martensitic transformation and shape memory effect in B2 intermetallic compounds of Ti, Rev. Phys. Appl. 24 1989, 733–739. 21. Kachin V.N., Voronin V.P., Sivokha V.P., and Pushin V.G., Martensitic tra nsformation a nd sha pe memo ry eff ect in polycomponent T iNi-based allo ys, I COMAT-95 p roceedings, P art I a nd II, L ausanne 1995 (G otthardt R . a nd Van Humbeeck J ., e ds.), E ditions de P hysique, J. Physique IV , 5(8), 1995, 707–172. 22. Saburi T., Takagaki T., Nenno S., and Koshino K., Mechanical behaviour of shape memory Ti-Ni-Cu alloys, MRS Intl Mtg. Adv. Mater., 9, 1989, 147–152. 23. Horikawa H. and Ueki T., Superelastic characteristics of NiTi-Cu-X alloys, Trans. Mat. Res. Soc. Jpn., 18B, 1993, 1113–1116. 24. Melton K.N., P roft J .L., a nd Duerig T .W., Wide h ysteresis shape memory alloys based on the N i-Ti-Nb system, MRS Intl Mtg. Adv. Mater. 9, 1989, 165–170. 25. Yang J.H. and Simpson J.W., Stress-induced transformation and su perelasticity in N i-Ti-Nb allo ys, I COMAT-95 Proceedings, P art I a nd II, L ausanne 1995, (G otthardt R . and Van H umbeeck J ., e ds.), E ditions de P hysique, J. Physique, IV, 5(8), 1995, 771–776. 26. Van Humbeeck J., High temperature shape memory alloys, Trans. ASME, 121, 1999, 98–101. 27. Firstov G.S., Van Humbeeck J., a nd Koval J.N., High t emperature shape memory alloys. Some recent developments, Mater. Sci. Eng. A, 378 2004, 2–10. 28. Maeshima T. and Nishida M., Shape memory properties of biomedical Fi-M o-Ag a nd T -Mo-Sn allo ys, Mater. T rans., 45(4), 2004, 1096–1100. 29. Kim H.Y., Ikehara, Y., Kim J.I., Hosoda H., and Miyazaki, S., Martensitic tra nsformation, sha pe memo ry eff ect and superelasticity of Ti-Nb binary alloys, Acta Mater., 54, 2006, 2419–2429. 30. O. S öderberg, O ., G e, Y., S ozinov, A., H annula, S.-P., a nd Lindroos V.K., Re cent b reakthrough de velopment o f the magnetic shape memory effect in Ni-Mn-Ga alloys, Smart. Mater. Struct. 14, 2005, 223–235. 31. Van Humbeeck J., Shape memory alloys: a ma terial and a technology. Adv. Eng. Mater., 3(11), 2001, 837–850. 32. Van Humbeeck J. and Kustov S., Active and passive damping of noise and vibration through shape memory alloys: application and mechanisms, Smart Mat. Struct., 14, 2005, 171–185. 33. Nishida M., Wayman C.M., a nd H onma T., P recipitation processes in ne ar-equiatomic T iNi sha pe memo ry a lloys, Met. Trans. 17A, 1986, 1505–1515. 34. Todoriki T. a nd Tamura H., Eff ect o f h eat tr eatment a fter cold w orking o n the p hase tra nsformation in T iNi allo y, Trans. Jpn. Inst. Metals, 28, 1987, 83–94. 35. Wayman C.M., P hase tra nsformations in N iTi typ e sha pe memory alloys, Proc. Int. Conf. On Mart. Transf. 1986, Ā e Japanese Institute of Metals, pp. 645–652. 36. De 89.

43722_C020.indd Sec1:36

Smart Materials

37. Sat 89. 38. Sa 89b. 39. an H 97.

Bibliography Z. Nishiyama, Martensitic Transformation, Academic Press, London, 1978. 2. H. Funakubo, Shape M emory Alloys, G ordon & B reach, London, 1986. 3. G.B. Olson and W.S. Owen, Martensite, ASM-international, 1992. 4. K..Otuska a nd C.M. Wayman, Shape Me mory Ma terials, Cambridge University Press, Cambridge, 1998. 5. V.E. Gunther, Delay La w a nd N ew Clas s o f M aterials and Implants in M edicine, SS T-Publishing, T omsk, Northampton,MA, 2000. 6. L’ Hocine Yahia, Shape Memory Implants, S pringer, B erlin, 2000. 7. M. Kohl, Shape Memory Actuators, Springer, Berlin, 2004. 1.

20.5 Smart Materials Modeling Manuel Laso 20.5.1 Introduction Ā e most salient feature of modeling work in the area of smart materials i s i ts g reat d iversity. M aterials c onsidered a s sm art span a s taggeringly wide range. Smart materials run the gamut from th e i norganic, m onolithic c rystalline m aterials, t o th e organic, polymeric, semicrystalline ones. Composites, polycrystalline materials, hydrated gels, magnetostrictive/ferromagnetic tagged composites, electrochromic materials, etc. to mention but a f ew, f urther e xpand t he r ange o f sm art m aterials to b e modeled. Ā e c omplexity t hat a rises f rom t his g reat v ariety of material types is compounded with the wide range of interesting properties they display. Finally, the question of the time and the length scales at which the modeling is to be implemented adds an extra level of complexity to the eld: Even when applied to the very s ame m aterial a nd t he v ery s ame p roperty, i t f requently happens t hat diff erent sm art m aterial mo delers (i) lo ok at t he material at vastly different s patial o r tem poral s cales, ( ii) u se completely unrelated modeling techniques, and (iii) even come to conclusions and modeling results, which can be unrelated for all practical purposes. Polyvinylidene  uoride (PVDF) is often considered a sma rt material and provides a good illustration of this situation. (a) At the quantum chemical (QC) level, appreciable effort has been devoted to t he determination of t he electronic st ructure of PVDF via QC methods with a view to estimate the

10/3/2008 9:04:51 PM

20-37

Shape-Memory Alloys and Effects: Types, Functions, Modeling, and Applications TABLE 20.6

Representative Modeling Levels and Goals for Major Types of Smart Materials

Type of Material

Application

Piezoelectric ceramic functional gradient Piezoelectric ceramic-adaptive composites

Actuators Smart structures

Lithium insertion compounds

Smart rotors Structural damping Aerodynamic control Smart batteries

Shape memory alloys

Smart polymers

Electrostrictive ceramic

Piezoelectric ceramic

Smart coatings Magnetostrictive/ferromagnetic tagged composites Electrorheological uids

Actuators

Mechanical Biomedical Microrobotics Microdevices Actuators Biomedical sensing Drug delivery Immobilization Ultrasonic transducers Actuators MEMS Pressure sensors Accelerometers Gyroscope Resonators Filters Stress visualization

Organic gelators

Composite cure and health monitoring Actuators Brakes Mechanical couplers Shock absorbers Dampers Brakes Distributed sensors Bragg gratings Fabry–Perot sensor Optical signal processing Intelligent transportation Optical multiplexing Molecular recognition

Gels/hydrogels

Drug delivery

Magnetostrictive materials

Transducers Sensors Actuators Motors Magnetometers

Magnetorheological uids

Fiber sensors

Modeling Level/Technique

Modeling Goal

Micromechanical CEs, composites (c), FE (d) Micromechanical CEs, composites (c) FE (d)

Electromechanical coupling Mechanical response Electromechanical coupling Mechanical response

QC (a)

Structure prediction Structure stability Energy diagrams Jahn–Teller distortion Shape memory effect Superelasticity Hysteresis

Micromechanical CEs, plasticity, composites (c) FE (d)

FE (d)

Electromechanical coupling Mechanical response

FE (d)

Electromechanical coupling Mechanical response

QC (a)

Mechanism of mechanoluminescence, band calculation Nondestructive materials testing

Micromechanical CEs (c) FE (d) Colloid dynamics (c), Viscoelastic ow calculations (d) Colloid dynamics (c), Viscoelastic ow calculations (d) FE, analytical (d)

QC (a) Atomistic MD and MC (b) Mesoscopic (c) Coarse-grained polymer MC (b) Micromechanical CEs (c) FE (d)

Structure–property relationship Rheological CE Device design Structure–property relationship Rheological CE Device design Optothermomechanical behavior

Crystal structure Molecular complementary Gel structure, transport behavior Magnetomechanical design

(continued)

43722_C020.indd Sec1:37

10/3/2008 9:04:51 PM

20-38 TABLE 20.6 (continued)

Smart Materials Representative Modeling Levels and Goals for Major Types of Smart Materials

Type of Material Giant magnetostrictive materials

Application

Modeling Level/Technique

Modeling Goal

Sensors Actuators Positioning devices Dampers Intelligent synthesis

Micromechanical CEs (c) FE (d)

Magnetomechanical design

Chemical thermodynamics, chemical kinetics (d)

Reaction kinetics

Vibrational sensor Atmospheric sensor Human skin mimicry Structural active elements Strain sensors Transducers Active and passive vibration control Damping Articial tissue Medical applications Nonvolatile memories electromechanical conversion fuel cells

Micromechanical CEs (c) FE (d)

Reactor design and control Electromechanical coupling Mechanical response

Atomistic MD and MC (b) Micromechanical CEs (c) FE (d)

Crystal structure prediction Electromechanical coupling Polarization response

Micromechanical CEs (c) FE (d)

Electromechanical coupling Polarization response

QC (a) atomistic MD and MC (b)

Smart skins

Sound control

Ā ermoresponsive inorganic materials

Temperature-sensing responsive devices

Micromechanical CEs (c) FE (d) QC (a) Atomistic MD and MC (b)

Crystal structure Piezoelectricity Pyroelectricity Ionic conductivity Electromechanical coupling

Electrochromic materials

Smart windows Architectural glazing Ā ermochromic devices Ā e rmotropic devices

Smart ceramics

Smart paints

Piezoelectric polymers

Functionally graded polymer blends

Smart perovskites

monomeric dipole. Very frequently, QC methods are applied to quite small fragments of the polymeric chains, sometimes even in vacuo. Ā is applies especially to those most advanced and sophisticated QC methods available today. (b) Somewhat si milar work has a lso been performed at t he atomistic le vel a lthough t he a ssignment o f ele ctronic distribution to individual atoms is not directly based on a f undamental Q C approach. At t his le vel, t he de scription of PVDF is based on a c lassical picture of atoms as sites i nteracting v ia a n em pirical f orce  eld. Typical modeling goals at this level are the correct calculation of crystalline p olymorphs, a nd a n e stimation o f p iezoelectric, d ielectric, a nd el astic p roperties v ia mole cular dynamics ( MD) a nd M onte C arlo ( MC) me thods. P rediction of electric properties is frequently limited by the use o f pa rtial c harges to de scribe t he s patial ele ctron distribution. (c) At the mesoscopic level, questions such as the stacking of folded PV DF l amellae, t he p rediction o f s emicrystalline morphology, a nd t he p rediction o f p iezoelectric mo duli based on homogenization are addressed.

43722_C020.indd Sec1:38

QC (a)

Crystal structure Phase transitions Band structure Transmittance Ion conduction

(d) At t he macroscopic level, modeling PVDF often r efers to the si mulation a nd de sign o f a pa rticular ge ometry o r device based on a purely continuum-mechanical description of the material via partial differential equations. All smart material information is condensed in a few macroscopic parameters, like piezoelectric moduli, elastic compliances, etc. At this level, modeling goes well beyond the material itself and is intimately linked to the design of a specic device for a specic function. In spite of the general propensity to consider one’s own specialty eld as slightly more central and momentous than others elds, or to consider, say, quantum mechanics more fundamental than nite elements (FE), scientists and engineers working at a ny of the f our le vels j ust de scribed a re j ustly a nd e qually entitled to claiming their work as smart materials modeling. For some specic materials, it also happens that modeling efforts at the different levels develop more or less independently of each other, with little if any communication among them. Ā is state of affairs is easily detected by the corresponding sets of cited literature being often disjointed or having minimal overlap.

10/3/2008 9:04:51 PM

20-39

Shape-Memory Alloys and Effects: Types, Functions, Modeling, and Applications

As a natural consequence, the description of smart materials and t heir equilibrium a nd nonequilibrium properties makes it necessary to re sort to mo st of t he modeling techniques, which are considered basic and fundamental in materials science, and, in addition, to some rather specialized ones. Ā er efore, this chapter c annot be more t han a t abular overview of techniques and a ba sic de scription o f t he gener al mo deling s etting w here these techniques are applied. Ā e very modest goal of this chapter is thus to provide a roadmap of the modeling landscape in the context of smart materials and a set of pointers to the fundamental literature. Specialists in a g iven area will  nd this necessarily cursory treatment of t heir  eld very supercial a nd l acking i n de tail. Specialists and generalists both may, however, nd more useful the general directions pertaining to ot her  elds and t he summary overview of ot her modeling smart materials techniques in use.

20.5.2 Conservation Laws, Equations of State, and Constitutive Equations Although a t  rst s eemingly remo ved f rom t he de scription o f specic modeling techniques, it is quite useful to ge t started in the area by describing the general setting in which most modeling e fforts a re emb edded. Ā e t hree blo cks de picted i n F igure 20.33 correspond to t he three main types of equations required to formulate a calculation, design, or modeling task in the most general sense. 1.

Macroeld ( conservation) e quations ( Block 1 ) e xpress very f undamental l aws, l ike t he c onstancy o f m ass, momentum, energ y, electric charge, mole n umber i n t he absence of chemical reaction, etc. In the present context, it is largely irrelevant whether we t hink of t hem as a xiomatic or a s or iginating f rom d eeper s ymmetry pr inciples. Conservation e quations a re o f v ery gener al v alidity a nd are t ypically formulated as pa rtial d ifferential equations. Ā ey describe t he relationship a mong macroscopic  elds via t heir t ime a nd space der ivatives. Ā ey hence contain

basic dy namic i nformation f or t he e volution o f m acroscopic  elds o f v ariables l ike i nternal energ y, en tropy, stress, velocity, temperature, electric charge, electric eld, density, c omposition, e tc. Ā ey a re o ften ba sed o n t he assumption of local equilibrium, i.e., deviations from local equilibrium are vanishingly small, and a few, macroscopically smooth elds suffice to describe the state of the material. Ā ese e quations do not f orm a c losed s et, i .e., t hey cannot be solved on their own. Most importantly, they contain no i nformation w hatsoever a bout t he me dium (material) in which these elds exist, so they are not material specic. 2. Equations o f s tate ( EOS) ( Block 2) de scribe t he e quilibrium behavior (properties) of a specic material. EOS are not universal laws, but statements of relationships between some o f t he  elds i n Blo ck 1 t hat app ear to b e a v alid description o f t he s tatic ( equilibrium) t hermodynamic properties f or a g iven m aterial. Ā ese rel ationships a re very frequently formulated as algebraic equations. One of the simplest examples of a Block 2 EOS is the assumption of constant mass density for a pa rticular material. While some EOS can be derived from more basic principles, like equilibrium statistical mechanics, many practical EOS are empirical correlations of adequate exibility and containing a sufficient number of parameters to be adjusted to the material under consideration. 3. Constitutive equations (CEs (Block 3) describe the nonequilibrium behavior of a specic material. Again, they are not universal laws, but more or less plausible statements of relationships b etween s ome o f t he  elds i n Blo ck 1 t hat appear to de scribe the dynamic behavior of a g iven material when driven out of equilibrium by an external inuence. Examples of well-known CEs are the assumption that the lo cal ele ctric c urrent den sity i n a m aterial i s p roportional to t he lo cal ele ctric  eld, t he p roportionality c onstant b eing a ( scalar o r ten sorial) ele ctrical c onductivity (microscopic Ohm’s law),* or Hooke’s relationship between stress a nd s train. C Es often appear in the form of local

1 Macrofield equations (conservation laws)

Discretization techn. for partial diff. eqs.

2 Equations of state (equilibrium thermodynamics)

Composite theory homogenization

Brownian/ stochastic dynamics

Atomistic, classical

3 Constitutive equations (nonequilibrium thermodynamics)

QM, QC

FIGURE 20.33 General formulation of modeling in smart materials; levels of description.

43722_C020.indd Sec1:39

10/3/2008 9:04:51 PM

20-40

relationships b etween gener alized  uxes a nd generalized forces ( spatial g radients, d riving f orces) i n t he s pirit o f linear i rreversible t hermodynamics. I n ot her c ases, C Es are formulated as differential, integral, or integro differential equations. Material properties appear as characteristic magnitudes r elating several  elds, e .g., a N ewtonian v iscosity relates the rate of strain tensor and the stress tensor. In t he gener al c ase, o nly t he c omplete s et o f e quations t hat appear in the three blocks is solvable. Its solution consists of the macroscopic  elds and their time dependence, possibly including an asymptotic approach to a steady state. In the PVDF example, a set of stress, strain, electric eld, and electric d isplacement  elds ( i.e., t heir v alues at e very p oint i n space a nd t heir v ariation i n t ime) c an b e ob tained b y w riting down a nd solving a s et of equations i n Blocks 1 to 3 . Si nce t he previous d iscussion i s ba sed o n a p urely m acroscopic le vel o f description (level (d) in the list in Section 20.5.1), it would be perfectly adequate and self-contained if the modeling goal were to achieve a quantitative description of the electric polarization of a PVDF piezoelectric of a given equilibrium shape when subjected to a g iven s tress. I n o rder to p erform mo deling at t his m acroscopic l evel, e lastic c ompliances, di electric p ermittivities, a nd piezoelectric moduli must be available, coming either from an experiment or possibly from a lower level modeling technique. Modeling at levels (a), (b), and (c) are however quite different: now the goal is not to c ompute the solution of a c omplex (heterogeneous) p roblem f or a pa rticular m aterial ge ometry. I t i s rather the prediction under very simple (e.g., homogeneous) conditions and starting from some basic principles (the meaning o f ba sic b eing no w le vel-dependent) o f s ome o r a ll o f t he physical, chemical, and thermodynamic properties that appear in the EOS and in the CEs, so that the modeler at le vel (d) can take them as input. In this sense, one of the goals of levels (a), (b), and (c) is to focus on Blocks 2 and, above all, 3 and pass on the information, which is required at le vel (d) to de al with the complete set of Blocks 1, 2, and 3. It i s h owever a lso t rue t hat i nformation pa ssing t akes p lace across these three rst levels (a), (b), and (c): • Some o f t he pa rameters app earing i n t he f orce  eld for classical MD modeling (level b) of PVDF may come from semiempirical or ab initio QC calculations (level a). • Atomistically computed crystal and amorphous matrix compliances via Parrinello–Rahman MD (level b) can be used to predict overall PVDF elastic compliance via homogenization methods borrowed from the theory of composites (level c). * As a side comment, EOS and CEs can be detected by their carrying a surname: On the one hand, conservation laws are universal and one would be hard pressed to associate a person’s name with their discovery. However, many EOS and all CEs are only plausible postulates but by no means universal laws. Ā e i nventor of s uch a pl ausible postulate seems to be naturally e ntitled t o a ppend h is/her n ame t o t he c orresponding e quation, hence Fourier’s, Fick’s, Newton’s, and Ohm’s laws for heat, mass, momentum, and electric charge transport, respectively.

43722_C020.indd Sec1:40

Smart Materials

Ā us, what is an output at a given level represents a necessary input at the next higher level. In this sense, it is now possible to assign modelers (a), (b), and (c) to t he EOS and CE blocks. Modeler ( d) m ust i nclude a ll t hree blo cks i n t he mo deling work, see Table 20.6. Ā e reader looking for an entry point in the eld of modeling smart materials is urged to  rst devote some effort to c arefully consider at what level or levels his or her modeling should take place. A go od deal of frustration can be avoided by a j udicious choice of the correct description level. It should also be mentioned that multiscale modeling methods exist a nd a re ga ining i ncreasing p opularity. Suc h h ierarchical modeling te chniques a re u ndergoing r apid de velopment a nd are bound to have a signicant impact on the modeling of smart materials i n t he ne ar f uture. Ā e ke y ide a b ehind m ultiscale modeling is to link in a single modeling approach two or more techniques residing at differential time/length scales ((a) through (d) above). As a prototypical example, the possibility of linking QC (density fu nctional th eory [DFT], a ctually) wi th c lassical MD and FE modeling techniques in a single calculation [1] has already been demonstrated for crack propagation in Si.

20.5.3

Informal Classification of Modeling Techniques for Smart Materials

is a f airly c omprehensive l ist o f m aterials c onsidered a s sm art together with their applications and suitable modeling techniques. As expected, modeling goals range f rom t he very microscopic, low-level ones ( band c alculations, c rystal s tructure prediction, etc.) to t he c ompletely ph enomenological a nd m acroscopic (design of transducers, actuators, etc.). Furthermore, at the largest s patial s cales ( level d a bove) t he mo deling o f t he m aterial itself, i.e., its equations of state and CEs, blends into the overall problem formulation, i.e., the design of a smart structure. Since the goal of the present chapter is the modeling of smart materials, it would be desirable to draw as clear a dividing line as possible in order to separate the modeling of the material itself from that of the structure. Ā is separation is closely linked to the idea of homogeneity: Except f or m aterials w hose sm art f unction de pends o n b eing intrinsically g raded, t he mo deling of t he m aterial itself i s b est addressed in a homogeneous setting. It is also possible that the goal o f t he mo deling i s p recisely t he de termination o f a verage (homogenized) properties. Unfortunately, the line that separates modeling a material from modeling a structure is often blurred. We c an g roup t he mo deling me thodologies t hat app ear i n Table 2 0.6 i n a f ew r ather well-dened f amilies. Ā e following list refers to some of the most basic literature in each area. Ā e list and the references at the end of the chapter are by no means exhaustive, but they represent useful entry points: • Quantum chemistry methods [2] · Ab initio QC, including Car–Parrinello MD [3] · S emiempirical [4] • Classical atomistic methods

10/3/2008 9:04:52 PM

Shape-Memory Alloys and Effects: Types, Functions, Modeling, and Applications

MD [5] MC and hybrid MC [6] Mesoscopic methods · Brownian/stochastic dynamics [7] · Lattice Boltzmann [8] Coarse-graining methods: · Micromechanical modeling [9] · Composite theory [10] Discrete techniques for partial differential equations: · Galerkin a nd its v ariants (Petrov–Galerkin, generalized Galerkin) [11] · Co llocation [12] · Meshless methods [13] Nonequilibrium thermodynamics · Linear irreversible thermodynamics [14] · GE NERIC [15] ·

7.

·

• • •

•

As a general introduction to the eld of smart materials, and not only their modeling, Ref. [16] is highly recommended.

20.5.4 Conclusion Ā e great diversity and the range of smart materials make it very difficult to draw general conclusions about their modeling. Only a few general guidelines can be suggested: • Strive f or a c lear s eparation b etween mo deling a sm art material, a nd mo deling a sm art de vice or s tructure. B e aware that in some cases, this separation is not possible. • Prior to any modeling work, carefully consider the general modeling si tuation de picted. D ecide at w hich le vels t he smart material is to be modeled. • Clearly d istinguish e quilibrium a nd no nequilibrium modeling goals. • Clearly distinguish between homogeneous and nonhomogeneous situations. • If multilevel modeling is involved, pay extra at tention to consistent coarse-graining a nd t hermodynamic admissibility [15].

References 1. G. Lu, E.B . Tadmor, a nd E. K axiras, Physical Re view B 73, 024108, 2006. 2. J.P. Lowe and Peterson, K.L., Quantum Chemistry, Academic Press, New York, 2005. 3. D. Marx and Hutter, J., in Modern Methods and Algorithms of Qu antum Ch emistry, e d. J . G rotendorst, J ohn v on Neumann I nstitute f or C omputing, J ulich, NI C S eries, Jülich, 2000, Vol. 1, p. 301. 4. M.C. Zerner, Semiempirical M olecular O rbital M ethods, VCH Publishers, Weinheim, 1991. 5. D. Frenkel and Smit, B., Understanding Molecular Simulation, Academic Press, London, 2001. 6. A. Leach, Molecular Modeling: Pr inciples a nd Applications, Prentice Hall, Englewood Cliffs, NJ, 2001.

43722_C020.indd Sec1:41

8. 9. 10. 11. 12. 13. 14. 15. 16.

20-41

H.C. Öttinger, Stochastic Processes in Polymeric Fluids: Tools and E xamples f or D eveloping S imulation Algorithms, Springer, Berlin, 1996. S. Succi, The Lattice Boltzmann Equation: For Fluid Dynamics and Beyond, Clarendon Press, Oxford, 2001. J. Aboudi, Mechanics o f Co mposite M aterials: A U nied Micromechanical Approach, Elsevier, Amsterdam, 1991. T.W.C. D erek H ull, D .R. Cla rke, a nd S. S uresh An I ntroduction to Composite Materials, Cambridge University Press, Cambridge, UK, 1996. A. Qua rteroni a nd Valli A., Numerical A pproximation of Partial Di fferential Equations, S pringer-Verlag, B erlin a nd Heidelberg, 1994. B.L. Nicola Bellomo, Revelli, R., and Ridol, L., Generalized Collocation Methods, Birkhauser Verlag AG, 2007. Y.C. James Lee and Azim Eskandarian, Meshless Methods in Solid Mechanics, Springer-Verlag, New York, 2006. S.R.D.G.A.P. Mazur, Non-equilibrium Ther modynamics, Dover, 1984. H.C. Ottinger, Beyond E quilibrium Ther modynamics, J ohn Wiley, Chichester, 2005. M. Schwartz, Encyclopedia of Smar t Materials, John Wiley, Chichester, 2002.

20.6

On the Microstructural Mechanisms of SMEs

Monica Barney and Michelle Bartning SMAs h ave t he u nique a bility to re cover t heir e xact original shape e ven a fter l arge m acroscopic d isplacements. T ypical elastic–plastic m aterials b egin to p ermanently de form at strains less than 1%, while in SMAs, this can be as high as 8% [1]. Ā e “magic” behind this property is rooted in the martensitic phase transformation that is driven by either a change in temperature or a change in stress [2,3]. While the existence of this solid-state transformation is necessary for SME to occur, it i s not a s ufficient c ondition ( martensitic t ransformations were rst seen in steels, where SMEs are not observed [4]). Ā e property of thermoelasticity is critical in determining whether a m aterial t hat u ndergoes t his t ype o f ph ase t ransformation will e xhibit s hape memo ry. I t i s ne cessary to d iscuss t he microscopic de tails o f t his r eaction to de velop a s ystematic understanding o f t he m acroscopic e ffects o f t hermal sha pe memory ( commonly c alled s hape me mory) a nd me chanical shape me mory (commonly c alled s uperelasticity) t hat a re of interest to eng ineers de veloping p ractical app lications. Ā e more common descriptions of SMAs discuss these two effects separately, giving the impression that they are only arbitrarily related and just happen to o ccur in the same material. In the second half of this chapter, we will show that these two macroscopic e ffects l ie o n a c ontinuum t hat i s go verned b y te st temperature [ 5,6]. Wi thin t his tem perature sp ectrum, t he

10/3/2008 9:04:52 PM

20-42

material ra nges f rom sta ble, n ontransforming a ustenite t o stable, nontransforming m artensite. A lthough mu ch o f t he following discussion applies to a ll SMAs, some of the details are s pecic to t he mo st w idely u sed a lloy: Ni -rich Ni Ti (Nitinol) t hat h as b een p rocessed w ith s olution t reatment followed by aging.

20.6.1 Microscopic Mechanisms The a bility o f S MAs to not o nly re cover l arge a mounts o f deformation but also return to the exact same shape lies in the details of the rst-order martensitic phase transformation. It is important to understand this reaction to be able to fully exploit the re sulting m acroscopic p roperties, s o w e w ill s pend s ome time describing it here. Ā e easiest way to organize the details of this reaction in the proper context is to review the thermal SME rst. We rst start with the parent phase, called austenite, which has t he u nusual combination of being a h igh-temperature and high-symmetry phase. As the specimen is cooled to a characteristic tem perature, t he ma terial sudden ly cha nges to the lower-symmetry phase, called martensite. Ā is change happens suddenly, as it is displacive (i.e., diffusionless) in nature, with the microstructure changing from a high-symmetry crystal structure (specically, simple cubic (B2) in NiTi [7]) to a lower symmetry one (monoclinic (B19’) i n NiTi [7]), w ith only very small shift s required of the atoms. Ā e crystallographic arrange-

Smart Materials

ments of t he t wo structures a re shown i n Figure 20.34a, w ith the black circles representing the nickel atoms [8]. As martensite n ucleates w ithin t he a ustenite, t hese t wo c ongurations must nd a way to t together in order to form an interface free from m acroscopic s tresses a nd d islocation f ormation i f t he material i s to e xhibit re versibility o f t his ph ase c hange. A lso, the new phase must  t into the space (volume and shape) that the a ustenite f ormerly o ccupied. Ā is ab ility is c alled s hape accommodation [4], and can be achieved by either twinning or slip [9,10]. Ā e ability to accomplish this by twinning is one of the mo st c ritical i ngredients o f m aterials t hat c an u ndergo a reversible m artensitic re action, w hich g ives r ise to s hape memory behavior. Ā ough twin boundaries are highly mobile, twinning c an ac commodate o nly sm all vol ume c hanges a s compared to slip [9]. Ā is limitation places strict constraints on the diff erences i n l attice pa rameters o f t he t wo ph ases, suc h that t he austenite is required to h ave a c ubic crystal structure [4]. In this way, the interface is coherent and the volume change between the two is very small. If this criterion is not met, plastic de formation ac companies t he ph ase t ransformation, destroying reversibility. To maintain this required compatibility, the martensite forms a twinned structure, where the different o rientation o f e ach t win i s o nly a s ymmetry-related crystallographic d ifference, c alled a v ariant [ 3,4]. Ā e twin boundaries are low energy, since they can move by nucleating new, t hough e nergetically eq uivalent, va riants. To see this

100

010

001

FIGURE 2 0.34 (a) S chematics of t he t wo b asic c rystal s tructures i n Nit inol. Ā e upp er s tructure i s m artensite (monoclinic, B19’), w hile t he simple cubic (B2) austenite is drawn in the lower left hand corner and 100 family of planes (b) Ā e three members of the (100) family of planes are drawn for the martensite (left-hand side) and austenite (right-hand side). Note the degeneracy of the cubic symmetry causing only one variant to be possible. All black circles represent the nickel atoms. (From Bronfenbrenner, D., 2006, with permission.)

43722_C020.indd Sec1:42

10/3/2008 9:04:52 PM

Shape-Memory Alloys and Effects: Types, Functions, Modeling, and Applications

more clearly, t he t hree members of t he {100} family of planes are drawn for each of t he phases in Figure 20.34b. Because of the cubic symmetry, all of the {100} planes of austenite appear exactly the same (i.e., it is univariant [3]). For the low-symmetry m artensite, t hey lo ok d ifferent, b ut a re a ll o f t he s ame energy. A s a re sult o f t his d ifference, w hen t he ma rtensite transforms back to the austenite upon heating (the last step of the t hermal SME), a ll of t he martensite variants t ransform to the only austenite variant has, causing it to remember the exact original shape.

20.6.2

Macroscopic Effects

We a re no w i n a b etter p osition to a nswer t he s eemingly straightforward question, “If you pull on NiTi, what happens?” Ā ere i s not , h owever, a si ngle a nswer to t his que stion, a s i t depends on the relative relationship between the ambient temperature while pulling and the transition temperatures of the material. Figure 20.35 de nes four of the important transition temperatures: ma rtensite st art ( Ms), martensite  nish (Mf ), austenite st art (As), a nd austenite  nish (A f ). Ā ese temperatures c an b e de termined b y DS C, a te chnique t hat me asures the heat ow into (endothermic reaction) or out of (exothermic reaction) t he s ample w hile v arying t he tem perature [ 11,12]. Ā e s ample s tarts off i n t he austenite ph ase at h igh temperature (upper right-hand corner of Figure 20.35); upon cooling to Ms, m artensite s tarts to f orm. B etween Ms and Mf, t he t hermally i nduced m artensite i s o nly pa rtially s table a nd s ome austenite is still present. Once the Mf temperature is reached, the ph ase t ransformation i s c omplete a nd f urther de creasing the tem perature w ill o nly c ause t hermal c ontraction o f t he martensite. Upon increasing the temperature to As, the material

30

Heat flow (mW)

20 10 Mf

Ms

0 As

−10

Af

−20 −30 −150

−100

−50 0 50 Temperature (⬚C)

100

150

FIGURE 20.35 Plot of t he thermal transitions typical in Nitinol as measured by DSC, with the characteristic temperatures labeled. Note that the transitions upon cooling do not occur at the same temperature as heating (i.e., Ms ¹ A f, Mf ¹ As), a sign of hysteresis.

43722_C020.indd Sec1:43

20-43

transforms back to the austenitic phase, completing when the A f is reached. After this point, more heat supplied only causes thermal e xpansion o f t he a ustenite. Ā is m ethod de termines the t hermal t ransitions i n t he a bsence of me chanical d riving force, making it easier to understand which type of driving force i s b ehind t he ob served m acroscopic ( stress–strain) mechanical re sponse. Ā ese t ransition t emperatures a re strongly a ffected b y t he h eat t reatment p erformed. S olution treating a nd a ging Ni -rich Ni Ti g ives a  ne a rray o f p recipitates needed to stabilize the martensite and raise the A f into a range suitable for medical devices. Ā ese defects also raise the stress ne eded f or s lip [ 13], a n i mportant c onsideration f or de ning the window over which superelasticity can occur. Ā is will be discussed further in the next section. Usually, the mechanical properties of NiTi are described by two l imiting c ases o f t he s tress–strain rel ationships, c alled superelasticity a nd s hape memo ry. H owever, to c apture t he full range of the mechanical performance possible for SMAs, stress, strain, and temperature must all be displayed together in some fashion, either by a three-dimensional plot of all these parameters (see Figure 20.36) or in a series of stress–strain curves, as displayed in Figure 20.37a-h. In this  gure, the test temperature decreases from a-h [14], and the left-hand column shows u nloading b ehavior w hile t he r ight-hand side s hows strain-to-failure cu rves. Ā e p lot at t he h ighest tem perature (Figure 20.37a) shows the response typical of an elastic–plastic material; i n t his case, it is stable, nontransforming austenite. Because high temperature destabilizes the martensite [15], it is possible to increase temperature to a point where stress-inducing martensite is no longer possible; this temperature is called the martensite deformation temperature Md [16]. It represents the point at which the critical stress for forming martensite is greater than that required to move a dislocation [17]. Āer efore, the top c urve (Figure 20.37a) shows t he expected behavior at temperatures above Md . Below Md , it becomes possible to stress-induce martensite. This effect is shown in Figure 20.37b-c and 20.37d-e, where upon re aching t he el astic l imit o f t he a ustenite, a c onstantstress p lateau t akes p lace b efore el astic–plastic de formation of m artensite o ccurs. I n t his c onstant-stress re gion, s tressinduced martensite (SIM) is created and this new phase forms in the detwinned state (unlike when it is thermally induced) since t he lo ading d irection b iases m artensite f ormation toward one variant [9]. It is increasingly easier to stress induce martensite as the temperature is lowered, since less mechanical d riving f orce f or t ransformation i s re quired up on approaching the region where martensite becomes thermally stable ( starting a t M s , b ecoming f ully s table at M f ). T his is evidenced b y t he d rop i n t he SI M p lateau w ith de creasing temperature, a s s een i n Figure 2 0.38 (from Re f. [9], s ee a lso Figure 2 in Ref. [6]). Upon u nloading, SI M i s no lon ger s table a nd t ransforms back to a ustenite a long t he lower plateau. Fu ll re covery up on unloading only happens if the temperature is above the Af. Ā is

10/3/2008 9:04:53 PM

20-44

Smart Materials

600 Stress, s

200

400 300 200 100

100

e, T

tur

a per

2

2

4 Strain, e

As

Af 0 0

100 0 0

200 Stress, s

Stress, s (MPa)

500

Below Mf

4 Strain, e

m

Te

0

0

2

4

Strain, e (%)

FIGURE 20.36 Ā ree-dimensional plot of stress, strain, and temperature, with temperature increasing into the page (from Wayman, C.M. and Duerig T.W., Engineering Aspects of SMAs, But terworth-Heinemann, L ondon, 1990, p. 3). Ā e c urve c losest to t he re ader shows elastic–plastic deformation of austenite, which occurs if the temperature is above Md. Ā e middle curve displays the behavior that takes place below Md, but above Af, commonly referred to as superelasticity. Ā e nal part of t he schematic displays the entire path taken for t he limiting case of shape memory. Following the arrows in the diagram, the sample is cooled to Mf, deformed, unloaded, and heated through As to a temperature above Af for strain full recovery.

thermal driving force to recover some types of permanent set. A better description is to refer to the permanent set as residual strain that can be made up of thermally recoverable strain, plastic strain, or a combination of both. If t he operating temperature is lowered below Mf, t he martensite is f ully stable and deforms by  rst detwinning (as evidenced by a constant stress plateau, see Figure 20.37f and g,

is t he l imiting c ase o f sup erelasticity ( see F igure 2 0.39 [18]), which o ccurs i n t he tem perature r ange o f Af < T < Md. I f t he SIM is formed below Af, not all of the strain is recovered upon unloading, leaving what is commonly called permanent set [17]. Ā is term can be misleading, since it is used to describe all unrecovered s train le ft in the material after r emoving t he mechanical driving force. However, heating above Af supplies

s

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

e

FIGURE 20.37 Alternate way of displaying the thermoelastic interdependence on stress, strain, and temperature. Temperature decreases from a–h in the gure, while strain is on the horizontal axis and stress on the vertical. Ā e plots represent characteristic mechanical behavior under the following temperature conditions: (a) T > Md, stable, nontransforming austenite elastically t hen plastically deforms; ( b) Af
43722_C020.indd Sec1:44

10/3/2008 9:04:54 PM

20-45

Shape-Memory Alloys and Effects: Types, Functions, Modeling, and Applications

Ms = −125⬚C −66⬚C 15

100

−75⬚C (ksi)

Stress (MPa)

125

−88⬚C

75

10

−99⬚C

50

5 25 0

0

2

6

4

8

10

0

Strain (%)

FIGURE 20.38 Stress–strain curves of a SMAs, showing a decrease in the stress plateau as the Ms temperature is approached. Martensite becomes thermally stable at that temperature, requiring less mechanical input to drive the transformation.

20.6.3

Summary

By covering some of the microstructural mechanisms that occur in NiTi, a systematic description of the observed macroscopic effects can be summarized in a context that allows for a better understanding of what is possible for this complex material. A lthough mo st appl ications t ypically e xploit e ither the special case of superelasticity or shape memory, covering the add itional de tails o f a ll c ombinations o f t hermal a nd mechanical S ME c an i lluminate add itional p ossibilities o r potential limitations, especially in highlighting the inuence of temperature.

s

s

which h appens at a lo wer s tress t han t he SI M p lateau ( i.e., Figure 2 0.37b-e)) fol lowed by e lastic–plastic d eformation (Figure 20.37g). Ā is mechanism is utilized in the limiting case of shape memory (see Figure 20.40 [18]). Ā ough not shown in this gure, the original shape will be fully restored upon heating above Af (see Figure 20.36 [9]). At the most extreme end, it is possible to cool so far below Mf that the twin boundaries are no lon ger mobi le, pre venting d etwinning. I n t his c ase, t he material o nly dem onstrates ela stic–plastic def ormation ( see Figure 2 0.37h). Ā e tem perature r ange o ver w hich a ll t hese phenomena occur is typically from about 50°C below to about 100°C above Ms [19].

e e

FIGURE 20.39 Stress–strain curve showing fully superelastic behavior. (From Barney, M., 2007, unpublished data. With permission.)

43722_C020.indd Sec1:45

FIGURE 20.40 Stress–strain curve showing the  rst two steps in the shape me mory me chanism. U pon u nloading, t he m aterial w ill f ully restore its original shape if it is heated above its A f temperature. (From Barney, M., 2007, unpublished data. With permission.)

10/3/2008 9:04:54 PM

20-46

Acknowledgments Ā e authors would like to t hank Drs T.W. Duerig, A.R. Pelton, and S.W. Robertson for the many helpful discussions.

References 1. AL McKelvey and RO Ritchie, 1999, Fatigue-crack propagation in ni tinol, a sha pe-memory and superelastic endovascular stent material, Journal of Biomedical Materials Research A, 47(3), 301. 2. K Mukherjee, S Sircar, and NB Dahotre, 1985, Āermal effects associated wi th str ess-induced ma rtensitic tra nsformation in a Ti-Ni alloy, Materials Science and Engineering, 74, 75. 3. JA Shaw and S Kyriakides, 1995, Ā ermomechanical aspects of NiTi, Journal of the Mechanics and Physics of Solids, 43, 1243. 4. K Bhattacharya, 2003, Microstructure of Martensite: Why It forms a nd H ow I t G ives R ise t o t he Sh ape-Memory E ffect, Oxford University Press, Oxford. 5. K Otsuka and X Ren, 2005, Physical metallurgy of Ti-Ni based shape memory alloys, Progress in Materials Science, 50, 511. 6. AR Pelton, J DiCello, and S Miyazaki, 2001, Optimization of processing a nd p roperties o f me dical-grade N itinol wir e, SMST-2000: Ā e International Conference on Shape Memory and Superelastic Technologies, Pacic Grove, CA p. 361. 7. GM Mic hal a nd R S inclair, 1981, Ā e str ucture o f T iNi martensite, Acta Crystallographica B, B37, 1803. 8. D Bronfenbrenner, 2006, unpublished work. 9. CM Wayman and TW Duerig, An introduction to martensite and sha pe memo ry, Engineering Aspects o f Sh ape M emory Alloys, Butterworth-Heinemann, London, 1990, p. 3.

43722_C020.indd Sec1:46

Smart Materials

10. H-S Koo, L Chang, H Chen, and K-L Shieh, 1992, An apparatus f or extin guishing a lig hted ciga rette, S hape-Memory Materials a nd P henomena—Fundamental Aspects a nd Applications, MRS P roceedings, 246, , Warrendale, PA, p . 421. 11. ASTM Designation: F 2004–05, Standard Test Method for Transformation Temperature of Nickel-Titanium Alloys by Ā er mal Analysis. 12. ASTM Designation: F 2005–05, Standard Terminology for Nickel-Titanium Shape Memory Alloys. 13. S Mi yazaki, Ā ermal a nd st ress c ycling eff ects a nd fa tigue properties of N i-Ti a lloys, Engineering Aspects o f Sh ape Memory A lloys, B utterworth-Heinemann, L ondon, 1990, p. 394. 14. AL M cKelvey, 1999, P hD diss ertation, U niversity o f California, Berkeley. 15. S Miyazaki, K Otsuka, and Y Suzuki, 1981, Transformation pseudoelasticity and deformation behavior in a Ti-50.6 at% Ni alloy, Scripta Metallurgica, 15, 287. 16. AL M cKelvey a nd R O Ri tchie 2000, On the t emperature dependence of the superelastic strength and the prediction of the theoretical uniaxial transformation strain in Nitinol, Philosophical Magazine A, 80(8), 1759. 17. TW Duerig a nd R Z adno, An en gineer’s p erspective o f pseudoelasticity, Engineering Aspects o f Sh ape M emory Alloys, Butterworth-Heinemann, London, 1990, p. 369. 18. M Barney, 2007, unpublished data. 19. J Van H umbeeck, R S talmans, M Cha ndrasekaran, a nd L D elaey, On the st ability o f sha pe memo ry allo ys, Engineering Aspects of Shape Memory Alloys, B utterworthHeinemann, London, 1990, p. 96.

10/3/2008 9:04:55 PM

21 Current Developments in Electrorheological Materials 21.1 I 21.2 21.3 21.4 21.5 21.6

Frank E. Filisko The University of Michigan

ntroduction ........................................................................................................................... 21-1 High-Yield Strength Electrorheological Fluids.................................................................. 21-1 ER under Compression/Squeeze Flow................................................................................ 21-2 Giant ER Effect Materials ..................................................................................................... 21-2 Lamellar Particle Structures—Effect of Coupled Electric and Shear Fields .................. 21-3 Alternative Model for Yield Strength of ER Materials ......................................................21-4 Ā e oretical Basis

References ...........................................................................................................................................21-5

21.1 Introduction Ā e h istory a nd b ackground of e lectrorheological (E R)  uids are c overed i n t he 2 000 v ersion o f Smart M aterials. T his chapter f ocuses o n t hose a reas w hich m ight b e c onsidered a signicant advancement i n t he  eld. Some of these were covered i n t he 2 000 e dition but at t hat t ime t hey may have b een just speculation but have e volved to b e more ac cepted w ithin the a rea a nd mo re o f a gener al c oncern to i nterested p eople. For the sake of space, many references were not included because they maybe redundant. However, main ones are given, which will allow interested parties to search for other relevant articles. It should be added that this includes information, which was presented at t he 10th I nternational C onference on ER/MR held i n July 2006 a nd a s such may not a s yet be published except in the proceedings of the meeting. It should also be noticed that even though the title of this chapter is “current developments,” m any o f t he re ferences a re not c urrent. Ā e purpose of this is to give adequate recognition to those with insight who may have recognized things years ago that are just beginning to be considered.

21.2 High-Yield Strength Electrorheological Fluids As recognized over the years as the various problems with ER uids (ERF) were re solved, i .e., w ater, c aking, h igh c urrents, temperature changes, etc. as discussed in the 2000 issue of this book, o ne o f t he m ajor i ssues rem aining i s t hat o f t he y ield strengths b eing to o lo w, i .e., a round 3 –5 kPa gener ally. Ā us

major attempts have been made to increase the strengths of ER materials, most commonly by looking at various components and/or compositions of t he  uids. Ā is has been tried for t he last 50 ye ars or s o a nd de spite t he a lmost i n nite number of possibilities h as re sulted i n v irtually no c onsistent suc cess, until possibly now as discussed in the giant electrorheological (GER) section. Ā is is one of the reasons that some feel that the failures may be due to an inadequate or wrong understanding of t he me chanisms i nvolved i n t he phenomenon, s pecically that yield occurs by chains breaking consecutively between the electrodes at high elds or strengths, which are of concern here. To f urther supp ort t his, s tudies p erformed to me asure the mo dulus o f E RF p rior to y ield h ave p roduced mo duli i n the M Pa r ange [1,2]. Ā us a n a lternate mo del h as b een p roposed, which is that at high elds o r st rengths, y ield o ccurs not by shear between the particle chains or columns between the electrodes but by slippage of the dense lamellar particulate structures at the electrode surfaces. Ā ere are at least two published ways that this has been tested. One method is by increasing t he no rmal s tress o n t he pa rticulate s tructures, t hus increasing t he “ frictional” f orce b etween t he s tructures a nd the electrodes [3]. A second is by designing electrodes, which will prevent slippage, something which is not as obvious to do as it may seem [4]. It is not altogether obvious that increasing normal s tresses b etween t he ele ctrodes or on t he pa rticulate structures c an b e do ne w ithout mo ving t he ele ctrodes to o close together since t he pa rticulate chains or columns (as we are generally led to v iew them) would apparently readily collapse or buckle. However, it was su rprisingly d iscovered t hat the structures under E elds, columns if no shear or lamellae 21-1

43722_C021.indd 1

10/3/2008 8:56:23 PM

21-2

if under shear, could carry very large loads as discussed in the section on compression loading/squeeze  ow. A similar study was performed by Tang et al. [5] on magnetorheological uids (MRF) and large increases to a round 800 kPa were recorded. It was suggested in this article that something similar should happen with ERF but we were unable to  nd any published data on this. Ā e s econd i nvolved a r ather i ngenious de sign of de vices to prevent s lippage. Ā ere a re t wo suc h de vices, w hich were u sed and b oth a re de scribed i n t he P roceedings o f t he 1 0th International Conference on ER/MR held in Lake Tahoe, CA in July 2006. In both cases, yield stresses in the 200 kPa range were recorded at a bout 4 kV/mm  eld s trengths on TiO2/silicone oil ER formulations. One e xplanation f or t he i ncreased s trengths b y i ncreasing normal stresses is based upon the alternative surface slip model of y ield strength [6]. A nother is ba sed upon a n explanation by Tang et al. [5] who also suggest surface friction but as well relate the effect to i ncreases t hickness o f t he c olumns o f pa rticles a s they are compressed, i.e., increasing the force to initiating shearing a s c overed i n t he l ast s ection o f t his c hapter. H owever, their studies d id not i nvolve shear, i.e., t he photographs t hey took of the expanding columns, which if they did would have observed t he l amellae. Ā erefore, t he e xplanation t hey p ropose, increasing of the cross-sectional area, is identical to that proposed in the alternative model. Ā e third type of high yield stress ERF involve the specic formulation of particles and the suspending uids and this is separately discussed in the section of GER ERF.

21.3 ER under Compression/Squeeze Flow ERF h ave not re ceived a lot o f at tention i n c ompression o r squeeze  ow until somewhat recently for reasons mentioned in this c hapter. W hile t hese m aterials h ave b een mo stly s tudied under shear, it was p ointed out by Sproston e t a l. [7] t hat t hey had i nteresting a nd v aluable p roperties u nder c ompression o r squeeze ow with applications in vibration damping. Ā e potential of ER materials in squeeze ow or compression was recognized by Stanway et al. [8] for damping control and by Williams et al. [9] who attempted to model squeeze ow for application to automotive engine mounts. In this case they considered the uid as homogenous, modeling it as a biviscous uid and did not consider the particle structures as acting independently to support the load although the displacements they were concerned with were small which is satisfactory for large electrode gap. Monkman [ 10,11] s tudied E RF u nder u niaxial c ompression, which extended beyond squeeze ow to higher amounts of compression a nd f ound t hat c ertain E RF a s a f unction o f E eld and gap width could support compressive stresses as high as 140 kPa, a p roperty w hich h e ter med h ardness a nd note d i t was t he s ame re gardless o f E eld m agnitude b ut at d ifferent electrode gaps . Ha rquing a nd K ing [12] s tudied a n E RF u nder static a nd dy namic lo ading u nder c ompression a nd ten sion at

43722_C021.indd 2

Smart Materials

constant st rain ra tes a nd r eported co mpressive st resses abo ve 200 kPa. Vieira and Arruda [13] did a very comprehensive study on the properties o f s ome E RF u nder c ompression a nd s hear b ut terminated their studies at low strains and thus stresses, before they re ached t he s tage e xperienced by Monkman [10,11]. Ā ey reported an initial elastic region, which occurs at very low strains and which they explain as being due to compaction of the initial particle columns, but was more likely due to t he “set up” of the uids i n t he ble ed off channels of their apparatus. Ā is is f ollowed by a no nlinear region i n t he stress–strain c urves, which initiates at some critical stress, which is due to the uid’s beginning to ow. Ā e highest critical stresses they report are around 30 kPa. Ā ey, however, very i mportantly recognize a nd d iscuss the c onsideration t hat t he pa rticles a re  owing in/ out fr om between the plates of the device during the tests. Nilsson and Olson [14] reported compressive forces of 400 N and as well derived equations to model the behavior at low and medium elds under dynamic loading conditions. Recently Tian et al. [15] report compressive stresses near 250 kPa and point out that t he b ehavior u nder h igh c ompressive s trains r ises mo re rapidly than predicted by any theory. Ā ey suggest this may be due to the fact that the uids do not act as homogenous materials under all loading conditions. More recently an extensive study by Lynch et al. [16] concentrates only on the high stress aspects of these studies as a f unction of particle concentration and viscosity of dispersing oil [16]. Ā ey reported compressive stresses above 1000 kPa at compressive strains of only around 30%. Ā ey nd that the data generally follows a p ower l aw b ehavior b ut v ery i nterestingly t he c oefficients depend strongly on the viscosity of the dispersing oil. Ā e reason for t his is related to t he oils pushing pa rticles out f rom between the plates as the uid is compressed. An example of such data is shown in Figure 21.1. Whereas mo st t heories t reat E lectroheological M aterials (ERM) as homogenous liquids, and model them as Bingham or biviscous uids [9,17,18], in reality under high electric elds, the materials c onsist o f t wo ph ases, o ne ph ase b eing o f a s eries o f columns o r l amellae o f c losely pac ked pa rticles e xtending between t he ele ctrodes i n a mo re o r le ss re gular f ashion [ 6] separated b y a s econd ph ase o f e ssentially pa rticle-free d ispersion oil. Depending upon particle size, concentration, and magnitude o f E, t he c olumns c an b e si ngle-particle c hains a f ew microns in diameter (very dilute suspensions) or most commonly a few hundred microns in diameter for ERF with concentrations around 10–20 vol. %. Ā us under squeeze ow, stacks of particles are b eing c ompressed a nd a  uid ph ase i s b eing s queezed o ut. Ā is suggests the rheological behavior is not simple, though it is modeled as such in most cases as referenced above.

21.4

Giant ER Effect Materials

One of the more signicant practical reported discoveries in the eld of ER since the discovery of water-free particulate systems in t he 1980’s i s c alled t he g iant E R e ffect (GER) [19–21]. Ā ese

10/3/2008 8:56:24 PM

21-3

Current Developments in Electrorheological Materials 1200 3000 V 2000 V 1000 V

Normal stress (kPa)

1000

800

600

400

200

0 0

0.05

0.1

0.15

0.2 0.25 0.3 Compressive strain

0.35

0.4

0.45

0.5

FIGURE 21.1 Normal stress vs. gap for a 40 wt% zeolite/silicone oil ERF under three electric elds.

dispersions, w hich c an c onsist o f a w ide v ariety o f pa rticulate materials, and apparently dielectric uids as well, treated in the fashion in the published papers, show yield stresses in the 100– 200 kPa range consistently. Although at the time of this writing, the materials have not been licensed by any company for production o r a re not b eing p roduced c ommercially no r a re s amples available from the researchers. Ā is is understandable considering the potential of these materials. In addition to t he high yield stresses, there are at le ast two aspects of these results, which are interesting. Ā e  rst is that the re sults do not f ollow t he r ather elemen tary ye t w idely ascribed to views of the ER effect, i.e., an initial E2 qualitative dependence of yield stress on electric eld. S econd, th at the ER eff ect go to z ero as particles get smaller, i.e., in t he nanoparticle si ze r ange d ue to Bro wnian mot ion d ue to t hermal effects overcoming the effects of particulate dipolar coulombic interactions. Ā e rst reported GER effect [19–21] involved barium titanyl oxalate particles (actually nanoparticles in the 50–70 nm diameter range) coated with a thin urea layer with a dielectric constant of about 6 0 at 1 0 Hz, w hich a re a bout 3 –10 nm t hick. Ā e yield stresses me asured o n a 3 0 vol .% su spension o f t he pa rticles i n silicone oil was about 130 kPa at 5 kV/mm  eld strength. Other material systems have been reported by this group but in general the GER uids are dispersions of 50–100 nm particles consisting of a core–shell structure, the core being a hard material consisting of v arious i norganic s alts a nd t he s hell of s oft u rea mole cules. Ā e h ard-core m aterials a re gener ally o f t he f orm M 1xM22-x TiO(C2O4)2 w here M 1 i s B a o r Sr a nd M 2 i s R b, C s, o r L i. A common e xample u sed m uch i n t heir i nitial s tudies b eing barium ti tanyl o xalate. Ā e p rocedures f or s ynthesizing t hese materials are given in detail in the publications [22]. Ā e explanation for this behavior they relate to a high degree of alignment

43722_C021.indd 3

of p olar mole cules w ithin t he c oating a nd not to t he b ulk particle. An extension of this behavior, which was presented at E R 10 in J uly 2 006, w as to ad sorb p olar mole cules o nto t he pa rticle surfaces, m aterials w hich were c alled PM-ER m aterials. A gain very high yield stresses are reported in the range of 200 kPa [23]. Ā e e xplanation for t his b ehavior i s rel ated to p olar mole cules adsorbed onto the particles. Ā ese p olar mole cules f orm v ery strong bonds at t he interface between adjacent particles, which are very strong due to the greatly attenuated E eld between the particles [24]. Although on ly s peculated on by t hat g roup, d irect e vidence does exist for this in the literature, but involving liquid crystalline solutions with sulfonated polystyrene particles [25]. It is clearly shown involving birefringence t hat t he molecules are collected and oriented at t he c ontact p oint of t he pa rticles. A d ifference being t hat t he L C s tudies i nvolved p olymeric si ze pa rticles i n solution w hereas t he ad sorbed mole cules i n t heir mo del a re much shorter [25].

21.5

Lamellar Particle Structures— Effect of Coupled Electric and Shear Fields

Ā ere have been substantial reports of the appearance of unique textures or higher ordered structures (i.e., beyond strings or columns) being formed in dispersions simultaneously exposed to an ele ctric a nd a s hear  eld. Bossis et al. [26] reported the occurrence of stripes of particles aligned in the ow direction for a d ispersion ex posed t o a n e lectric  eld a nd u nder o scillatory shear. Ā ey men tion t hat t hese s tripes, w hich o ccur i n s hear elds alone, may actually be sheets due to the imposed E eld.

10/3/2008 8:56:24 PM

21-4

Smart Materials

In another s tudy, t hey s how t hat u nder v arious a mplitudes o f oscillation and under an E eld, these sheets may be perpendicular or parallel to the  ow direction. Cutillas et al. [27] compared different structures observed in ERF and MRF under  eld and oscillatory s hear. Ā is g roup re fers i n t heir pap ers a s t he ph enomenon being a phase separation induced by the eld, which it is, of course, in t he sense t hat t he dispersed phase, normally uniformly dispersed, is caused to associate into unique structures by either/or E/M and shear elds. Another study involving steady s hear a nd  ow t hrough a s lit, c oncentric c ylinders, a nd between parallel rotating disks, Tang et al. [28] report structures consisting of parallel lines of particles in the couette and concentric rings of particles b etween t he pa rallel d isks. Tao a nd J iang [29] a lso re ported h igher o rder pa rticulate s tructure f ormation for ERF under  eld and shear. A sub sequent extensive study by Henley and Filisko [6] using parallel disks and concentric cylinders, clearly demonstrate t he el ectric  eld a nd s hear  eld induced phase separation (assembly) of the dispersed particles as well as the particulate structures resulting from the coupled E and shear elds. Ā e occurrence of particle aggregation and structuring within dispersions exposed to electric and shear elds has not only been observed experimentally but has been shown in computer simulations [30–32]. Ā e si mulations p redict t he o ccurrence o f l ayered structures ( of pa rticles) i n t he ide alized si tuation c onsidering point d ipole i nteractions, Bro wnian, a nd s hear f orces b ut neglecting m any-body i nteractions. I n a mo re re cent s eries o f excellent articles, K lingenberg and coworkers [33] not o nly did

(a)

extensive w ork o n ob serving t he s tructures, b ut de veloped a theory that shows that these structures must form under certain circumstances from free energy considerations (Figure 21.2)[33].

21.6

Alternative Model for Yield Strength of ER Materials

As pointed out in a previous section of this chapter, the particles in E RF a nd M RF a s w ell [ 28] at app ropriately h igh c oncentrations and under shear, upon application of an electric eld, rapidly organize or regiment themselves into numerous tight packed lamellar formations, which a re more or less pa rallel a nd periodic in relationship to each other. Tight packed means that the particles a re c rowded toge ther a s c losely a s p ossible, i f not e ven compressed together under the inuence of the E eld, consistent with the irregular size and shape of the particles. Because the particles a re not mo nodisperse s pheres o r a ny re gular ge ometric shape, they cannot be packed into any regular lattice arrangement such as simple cubic or hexagonal close packed. Under zero shear conditions the particles agglomerate together into columns under the in uence of t he E eld while add itionally u nder shear t hese particles pack together into these continuous lamellar structures more or less equally spaced and of similar thickness. Ā e relative arrangements and characteristics of t hese structures are dependent in a not as yet well determined way upon magnitude of electric eld, shear rate, particle concentration, time of shear, and shear prole. Ā ese structures take t he form of walls between pa rallel plates, short cylinders between parallel disks, and stacked washers

(b)

FIGURE 21.2 Examples of wall and ring particulate structures, i.e., the dark line and rings are particles. Ā e dark spots in the circular gure are the ends of columns.

43722_C021.indd 4

10/3/2008 8:56:24 PM

21-5

Current Developments in Electrorheological Materials

between c oncentration c ylinders. I n a ll c ases, t he t hickness dimension of the walls, the walls of the cylinders, and the thickness of the washers is the direction normal to plane formed by the E eld and shear directions. A major consequence of the tight packing of the particles into these structures is that it results in cooperative strengthening of the structure to failure under shear in the long dimension of the structure due simply to t he increased cross-sectional area. Ā is is b ecause t he a rea o f t he s hear p lane i ncreases w ith l amellar length and thickness as well as the lack of an easy path for slip planes to de velop w ithin t he s tructures d ue to t he i rregular packing and irregular size and shapes of the particles. Ā us it is hypothesized that the lamellar structures remain intact for the most part under ow (i.e., in the postyield region) and t he mo st p robable pat h o f s lippage u nder  ow is at the interface b etween t he pa rticle s tructures a nd t he ele ctrodes. Ā is being the case, it is proposed that the structures themselves do not b reak or shear. Because polarization forces a re responsible for holding the structures together, if the structures do not fail, then these forces are irrelevant to the postyield behavior, as long as t hey ar e s trong e nough t o m aintain t he s tructure. Thus the p ostyield b ehavior i s de termined p rimarily b y t he mechanics of s lippage b etween t he ends of t he s tructures a nd the adjacent electrode. Shear or slippage between the ends of the lamellae a nd t he ele ctrodes i s t he mo re do minant s ource o f energy dissipation in ERF under eld and the source of the yield stress a s w ell a s t he de termining f actor f or t he y ield s tress (Figure 21.3)[3].

21.6.1

Theoretical Basis

Consistent w ith t he Bi ngham model approach, t here is critical yield s hear s tress, Ty, f or t he o nset o f  ow, w hich i s o f c ourse independent of sample dimensions, and a corresponding critical yield shear strain gy. Ty = G (E ) γ y = Fy /(Lt ) where G(E) is the electric eld-dependent shear modulus L is the length of lamellae t is the thickness of lamellae Fy is the shearing force to initiate yield

L t

Gap

FIGURE 2 1.3 shear.

43722_C021.indd 5

F

G(E )

Model s howing a p ortion of a p article w all u nder

However, a s t he s hear p lane a rea ( Lt) i ncreases, t he f orce(Fy) necessary to re ach t hat y ield s train i ncreases w ith a rea a nd must be balanced by frictional (adhesive) forces between the lamellae and the electrodes. Ā e frictional (adhesive) forces (Fa) between the lamellae ends and the electrodes may be partially electrical in nature or mechanical. For lack of a better understanding o f th e p arameters an d i n an alogy wi th s tandard frictional phenomena, t he su rface f rictional (adhesive) forces could b e de scribed b y a s tatic f rictional o r ad hesive c oefficient, mS, a nd a no rmal f orce b etween a l amellae en d a nd t he electrode, N: Fa = µS N Ā e y ield f orce, h owever, de pends up on s hear p lane a rea a nd eld-dependent modulus: Fy = G (E )γ y (Lt ) Ā us, in order for a lamellae to shear internally, the surface adhesive forces must be greater than the yield force. However, if the force required for yield is greater than the frictional or adhesive forces at t he su rface, t han s lip w ill o ccur at t he su rface b efore internal shear occurs. In other words, for internal shear to occur, Fa > Fy. Slip at the wall occurs when Fy > Fa. From the above discussion, it can be seen that Fy i ncreases wi th th e l ength an d thickness of t he lamellae at a g iven  eld, whereas t he ad hesive force may be independent of the contact area. Ā us it can easily be seen what these lamellar structures can accomplish increasing the force necessary for yield by virtue of the increased crosssectional or shear plane areas.

References 1. S.O. O yadiji, in Ele ctrorheological  uids, in Magnetorheological Suspensions and Associated Technology, Ed. W.A. Bullough, W orld Scientic, H ong K ong, p p. 419–428, 1995. 2. D.R. Gamota and F.E. Filisko, Dynamic mechanical studies of ER materials under moderate frequencies, J. Rheol., 25(3), 399–427, 1991. 3. F.E. Filisko, An alternative vie w for the yield str ess of ER/ MR ma terials, in Proceedings o f t he 10t h I nternational Conference on E lectrorheological F luids and M agnetorheological Su spensions, E ds. F. G ordaninejad, O.A. Gr aeve, A. Fuchs, Lake Tahoe, CA, July 2006. 4. R. Shen, X. Wang, Y. Lu, W. Wen, and K. Lu, Proceedings of the 10th I nternational C onference on E lectrorheological F luids and Magnetorheological Su spensions, E ds. F. G ordaninejad, O.A. Graeve, and A. Fuchs, Lake Tahoe, CA, July 2006. 5. X. Tang, X. Zha ng, a nd R . Tao, S tructure enha nced yield strength of MR uids, J. Appl. Phys., 87(5), 2634–2638, 2000. 6. S. Henley and F.E. Filisko, Flow proles of electrorheological suspensions: An alternative model for ER activity, J. Rheol., 43(5), 1323–1336, 1999.

10/3/2008 8:56:25 PM

21-6

7. J.L. S proston, R . S tanway, E.W. Wiliams, a nd S. Rigb y, Ā e electrorheological automotive engine mount, J. Electrostat., 32(3), 253–259, 1994. 8. R. S tanway, J .L. S proston, M.J . P endergast, J .R. C ase, a nd C.E. Wilne, ER uids in the squeeze mode: An application to vibration isolation, J. Electrostat., 28, 89–94, 1992. 9. E.W. Williams, S.G. Rigb y, J .L. S proston, a nd R . S tanway, Electrorheological  uids a pplied to an a utomotive eng ine mount, J. Non-Newtonian Fluid Mech., 47, 221–238, 1993. 10. G.J. M onkman, Ā e e lectrorheological e ffect under co mpressive stress, J. Phys. D, 28(3), 588–593, 1995. 11. G.J. M onkman, E xploitation o f co mpressive str ess in ER coupling, Mechatronics, 7(1), 27–36, 1997. 12. G. Haiqing and L-M. King, Experimental investigations on tension and compression properties of an ER ma terial, J. Intell. Mater. Syst. Struct., 7, 89–96, 1996. 13. S.L. Vieira a nd A.C.F. Arruda, ER  uids r esponse under mechanical test ing, J. In tell. M ater. S yst. S truct., 9, 44–59, 1998. 14. M. N ilsson a nd N. Ohls on, An ele ctrorheological  uid in squeeze mo de, J. In tell. M ater. S yst. S truct., 11, 545–554, 2000. 15. Y. Tian, Y. Meng, and S. Wen, Electrorheological uid under elongation, co mpression, a nd she aring, Phys. Re v. E 65, 031507, 2002. 16. R. Lynch, Y. Meng, and F.E. Filisko, Compression of dispersions to high stress under electric elds: Effect of concentrations and dispersing oil, J. Coll. Interf. Sci., 297, 322–328, 2006. 17. J. Stefan, Akad. Wiss. Math. Natur. Wien, 69, part 2713–2735, 1874. 18. J.R. Scott, Trans. Inst. Rubber Ind, 7, 1931, 169–186; 481–493, 1935. 19. W. Wen, X. Huang, S. Yang, K. Lu, and P. Shing, Ā e giant ER e ffect in susp ensions o f na noparticles, Nat. Ma ter., 2, 727–730, 2003, 20. W. Wen, X. Huang, and P. Sheng, Particle size scaling of GER, Appl. Phys. Lett., 85(2), 299–301, 2004. 21. K. Lu, R. Shen, X. Wang, G. Sun, and W. Wen, ER uids with high shear stress, Int. J. Mod. Phys, B, 19, 1065–1070, 2005. 22. K. L u, R . S hen, X. Weng, G. S un, W. Wen, a nd J . Li u, Polar molecule typ e ER  uids, in Proceedings o f t he 10t h

43722_C021.indd 6

Smart Materials

23. 24. 25. 26.

27.

28.

29. 30. 31. 32. 33.

International C onference on E lectrorheological F luids and Magnetorheological Su spensions, E ds. F . G ordaninejad, O.A. Graeve, and A. Fuchs, Lake Tahoe, Nevada, July 2006. W. Wen, X. H uang, a nd P. S hing, Appl. P hys. L ett., 85(2), 299–301, 2004. Y. Chen, A. Sprecher, and H. Conrad, J. Appl. Phys., 70(11), 6796, 1991. G.P. Quist a nd F.E. Filisko, ER susp ensions with ER ac tive matrix liq uids, Int. J . M od. Ph ys B , 13(14–16), 1675–1681, 1999. G. Bossis, Y. Grasselli, E. Lemaire, L. Petit, and J. Persello, Yield stress and  eld induced str ucture in ele ctro and magnetorheological susp ensions, in Proceedings of the I nternational Conference on E R  uids, M echanisms, P roperties S tructure, Technology, a nd Applications, E d. R . T ao, p g. 7 5, World Scientic, Singapore 1992. S. Cutillas, G. Bossis, A. Cebers, E. lemaire, and A. Meunier, Flow a nd f ield ind uced str uctures in r elation wi th the rheology o f ER a nd MR  uids, in Proceedings of the International Conference on ERF, MRS, and Āei r Applications, Eds. K oyama, K. a nd N akano, M., Yonezawa, J apan, p . 43, July 22–25, 1997. X. Tang, W.H. Li, X.J. Wang, and P.Q. Zhang Structure evolution o f e lectrorheological  uids under  ow conditions, Electrorheological F luids, M agnetorheological Su spensions and t heir Applications, E ds. M. N akano a nd K. K oyama, World S cientic, Singapore, pp. 164–171, 1998. R. Tao a nd Q . J iang, S tructure f ormation in ER  uids, in Progress in Electrorheology, Eds. F.E. Filisko and K. Havelka, Plenum Press, New York, pp. 325–333, 1995. J.R. Melrose and D.M. Heyes, Simulation of ER and particle mixture suspensions: Agglomeration and layered structures, J. Chem. Phys., 98(7), 5873–5886, 1993. M. Whittle, C omputer sim ulation o f a n ER  uid, J. No n Newtonian Fluid Mech., 37, 233–263, 1990. J.R. M elrose, B rownian dyna mics sim ulation o f di pole suspensions under she ar: Ā e p hase dia gram, Mol. P hys., 76(3), 635–660, 1992. K. vonPfeil, M.D. Graham, D.J. Klingenberg, and J.F. Morris, Structure evolution in ER a nd MR susp ensions from continuum perspective, J. Appl. Phys., 93(9), 5769–5779, 2003.

10/3/2008 8:56:27 PM

22 Carbon Microtubes and Conical Carbon Nanotubes Santoshrupa Dumpala University of Louisville

Gopinath Bhimarasetti University of Louisville

Suresh Gubbala

22.1 I ntroduction ............................................................................................................................22-1 Carbon Microtubular Morphologies • Conical Carbon Nanotube Morphologies

22.2 Synthesis and Morphological Control of CMTs ................................................................22-3 Synthesis Methods and Growth Mechanisms • Control of Morphology

University of Louisville

22.3 Synthesis and Morphological Control of CCNTs ..............................................................22-8

Praveen Meduri

22.4 App lications ...........................................................................................................................22-11

University of Louisville

Silpa Kona University of Louisville

Mahendra K. Sunkara University of Louisville

Synthesis Methods and Growth Mechanisms • Morphological Control

New Electrode Materials • Templates for Nanoelectrode Ensembles • Field Emission Applications • Porous Carbons • Drug Delivery, Fluid Flow, and Other Miscellaneous Applications

22.5 Su mmary ................................................................................................................................22-14 Acknowledgments ...........................................................................................................................22-14 References .........................................................................................................................................22-14

One-dimensional a nd conical c arbon structures were reported as early as 1960 [1,2], however, no evidence of hollowness was presented at t hat t ime. Ā e c arbon n anotubes we re d iscovered much later in 1991 [3], consisting of single or multiple layers of cylindrical graphene sheets. Ā e morphology of a ll t he derived structures of nanotubes was primarily cylindrical in nature with internal diameters ranging from 1 to 25 nm [4]. In the last ve years, a number of new morphologies of carbon tubular structures with hollow cores have been discovered. Ā ese include the conical carbon morphologies tapering from micronsized base to a nanometer scale tip [5–7], with a hollow core not exceeding few tens of nanometers, and micron-scale morphologies with internal diameters exceeding several hundred nanometers [8–10]. All these morphologies are unique compared to c arbon nanotubes in that they provide a clearer connection between the micrometer and nanometer scales, at t he same time facilitating newer applications [6]. Ā e promising application areas include electrochemistry, b iosensors,  eld em ission, h igh su rface a rea catalysts/supports, a nd p ossibly d rug del ivery. Ā is ch apter is devoted to p roviding a de tailed d iscussion o n t heir s ynthesis, possible growth mechanisms, and applications.

22.1

Introduction

In t his c hapter, a ll t he ne w mo rphological f orms o f c arbon are considered and are divided in to t wo main categories for

ease o f d iscussion: ( a) c arbon m icrotubes ( CMTs), w hich include a ll t he s tructures c ontaining i nternal d iameters greater t han 1 00 nm a nd ( b) ( CCNTs), w hich i nclude a ll tapered mo rphologies w ith h ollow c ores a nd i nternal diameters of only a few nanometers.

22.1.1 Carbon Microtubular Morphologies Typically, carbon nanotubes are formed as a result of self-assembly of g raphene s heets i nto t ubular s tructures on t he su rface of t he nanometer-sized c atalyst c luster [3]. H owever, suc h s elf-assembly processes c annot t ake p lace re adily o n m icron-sized c atalyst clusters. Many of the earlier processes using micron-sized catalyst particles y ielded s olid g raphite/carbon w hiskers but not t ubular structures [11]. So, a set of different process conditions than those used for typical self-assembly processes are required for making carbon tubular structures with large internal diameters. Ā e main types of morphologies of CMTs achieved are illustrated with electron microscopic images in Figure 22.1: (a) Graphite t ubes w ith i nternal d iameters r anging f rom 7 0 to 1300 nm having inner obstructions reminiscent of bamboo styling; (b) thin-walled, straight carbon tubular structures with internal d iameters r anging f rom f ew h undred na nometers to f ew microns; (c) t hin-walled, c onical m icrotubular s tructures; a nd (d) micro-tubular carbonaceous tubes made by the pyrolysis of polymeric bers. 22-1

43722_C022.indd 1

10/15/2008 4:27:56 PM

22-2

Smart Materials

(a)

(b)

Internal closure Gas Liquid

Liquid

Gas

200 nm

1 µm

100 nm

(c)

1 µm

(d)

18 µm

FIGURE 22.1 Various carbon microtubular morphologies. (a) CMTs with encapsulated liquids and catalyst particles using a hydrothermal method. (From Libera, J. and Gogotsi. Y., Carbon, 39, 1307, 2001. With permission.) (b) Straight CMTs with diameters in the range of a few microns obtained using low-melting metals such as gallium. (c) Tapered CMTs with thin walls using low-melting metals such as gallium with a different set of conditions than (b). (d) Carbon tubes made by pyrolizing polymeric bers. (From Han, C.C. et al., Chem. Mater., 11, 1806, 1999. With permission.)

The wall structures of the CMTs, shown in Figure 22.1, are all different from each other depending upon their morphology a nd me thod o f s ynthesis. T he g raphite t ubes s hown i n Figure 2 2.1a e xhibit c rystalline, g raphitic w alls b ut c ontain bamboo-styled i nternal o bstructions r eminiscent o f th e nickel catalyst used for their synthesis. The walls of the microtubes resulting from the pyrolysis of polymeric fibers typically contain disordered carbon. The thin-walled, CMTs shown in both Figures 22.2b and c contain clear internal channels and their walls are made of graphene sheets. However, the straight morphologies with uniform diameter contain walls of parallel graphene sheets similar to that of multiwalled carbon nanotubes (see Figure 2 2.2a). Even a s light c onical a ngle m akes the w alls e xhibit d iscontinuous re gions w ith pa rallel g raphene s heets. T he w alls w ithin t he c onical m icrotubular structures cannot maintain any sort of ordered arrangement and so break down into an ordered arrangement of small graphitic domains within a disordered carbon matrix as shown in Figure 22.2c. Almost all of the microtubular structures with their thin walls are electron transparent even at modest accelerating voltages of 15 keV or more.

43722_C022.indd 2

22.1.2 Conical Carbon Nanotube Morphologies Ā eoretical calculations predict that a single graphene sheet can be f olded t hrough a s pecic a rrangement o f p entagonal a nd hexagonal rings to produce a set of conical morphologies with a limited n umber o f d iscrete c onical a ngles [ 12]. S ome o f t he structures synthesized to-date include fullerene cones [13] and nanohorns [ 14]. A ll t he a bove s tructures a re t rue na noscale morphologies. H owever, t here i s a nother i mportant c lass o f conical morphologies that extend from micron to nanometer scale. By denition, we refer to them as CCNTs in which the wall thickness tapers from micron to nanometer scale but contains a uniform hollow interior with a diameter range similar to that of multiwalled carbon nanotubes. Several morphologies of t he conical ca rbon n anotubes w ere r eferred t o e arlier as tu bular graphite c ones (TG Cs) [ 6], c arbon na nopipettes ( CNP) [ 5], conical graphite tubes [15], and conical nanocarbon (CNC) [7]. Ā e m orphologies o f sev eral o f t hese C CNT st ructures a re illustrated in Figure 22.3. All of them have micron-scale bases extending conically to form nanoscale tips similar to that shown

10/15/2008 4:27:56 PM

22-3

Carbon Microtubes and Conical Carbon Nanotubes

500 nm

17 nm

Tube wall

(a)

3 nm

5 nm

can themselves be tapered at t he tips to p roduce conical shape (see F igure 2 2.3c). I n ot her c ases, c onical s tructures h aving a hollow core with the catalyst particle closing the tip have been observed (see Figure 22.3d). Although these morphologies look similar from the exterior, the wall structures could be completely different. Ā e CCNTs can be divided into three main types based on their wall structure as shown in Figure 22.4. One type of structure consists of a central carbon na notube w ith g raphene s heets helically rol led over t he tube g iving t he conical structure [5] (Figure 22.4a). Ā e second type a lso consists of a c entral na notube, but its wall is made of concentric cylinders of graphene sheets with gradual reduction in the thickness along the length to y ield conical shape [6] (Figure 22.4b). Neither of these structures contain any metal contamination o r o bstructions w ithin t heir in ternal c hannels. Ā e third structure h as s tacked g raphene l ayers ( similar to m ultiwalled carbon na notubes) a nd t apers a fter a c ertain leng th towards its tip. Ā e internal channel of the third type of structure may not be as perfect as that of the rst two types and also contains several types of internal obstructions resulting from the catalyst particle assisted growth for extended periods of time [17] (Figure 22.4c).

22.2 Synthesis and Morphological Control of CMTs 22.2.1 500 nm (b)

5 nm

(c)

FIGURE 2 2.2 Variations o bserved i n t he w all s tructures for C MTs synthesized u sing lo w-melting me tals: ( a) C MT w alls m ade of lon g continuous concentric graphite sheets. (From Hu, J. e t al., Adv. Mater., 16, 153, 2004. With permission.) (b) Ā e walls made up of p arallel, but discontinuous g raphene s heets. ( From B himarasetti, G. et a l., Adv. Mater., 1 5, 1629, 2 003. W ith p ermission.) (c) T EM photo graph of t he wall s tructure s howing a ligned n anocrystals of g raphite. (From Bhimarasetti, G., Cowley, J.M., and Sunkara, M.K., Nanotechnology, 16, S362, 2005. With permission.)

in Figure 22.3a. In some cases, as seen in Figure 22.3b, carbon nanotubes w ith e xtended leng ths c an b e found at t he t ip o f a CCNT [16]. In some cases, thick multiwalled carbon nanotubes

43722_C022.indd 3

Synthesis Methods and Growth Mechanisms

Ā e synthesis of CMTs requires totally different process conditions a nd c atalysts c ompared to t hat o f m ulti-walled c arbon nanotubes (MWNTs). CMTs with thick walls, as shown in Figure 22.1a, were produced using micron-sized nickel catalyst particles by a hydrothermal synthesis method at temperatures and pressure of ∼800°C a nd ∼100 MPa, respectively [9]. Ā e p recipitation o f carbon at high pressures onto larger nickel catalyst particles suspended in water leads to the formation of highly crystalline graphite t ubes w ith l arger i nternal d iameters. Ā e bamboo-styled internal ob structions ob served a re si milar to t hose s een i n t he MWNTs grown using nickel catalysts. Ā e mo st p redominant mo rphologies o f C MTs, i .e., t hinwalled CMTs with no internal obstructions are made only using low-melting metals such as Ga, Sn, Zn, and In. See Table 22.1 for a concise review of all the synthesis studies reported. Ā e rst set of studies used carbothermal reduction of gallium oxide powder to produce thin-walled, straight CMTs half- lled with Ga [8,18]. Several other reports showed minor variations to produce CMTs with encapsulated metals with a range of internal diameters [19–23]. With subtle variations in the gas phase chemistry along with lo w-melting me tals, a n umber o f mo rphological s hapes ranging from conical to s traight tubes have been reported [10]. As s hown i n Table 2 2.1, a ll me thods e xcept the hydrothermal synthesis utilize chemical vapor deposition of carbon onto lowmelting metals such as Zn, Sn, In, and Ga. Ā ese studies utilized various o xide, n itride, a nd s ulde s ources f or me tals suc h a s ZnS, SnS, GaN, Ga 2O3, CdS, SnO, In, and Ga along with different gases and solid carbon sources.

10/15/2008 4:27:57 PM

22-4

Smart Materials

(a)

(b)

100 nm

(c)

1µm Channel (Carbon nanofiber)

(d)

100 nm

Outer wall (precipitatied C) 200 nm

100 nm

1µm

FIGURE 22.3 Various morphologies of C CNT morphologies: (a) CCNT structures termed as CNP with constant hollow cores and open tips. (From Mani, R.C. et al., Nano Lett., 3, 671, 2003. With permission.) (b) A CCNT structure termed as “tubular graphitic cone” with a nanotube at the tip. (From Shang, N., Milne, W.I., and Jiang, X., J. Am. Chem. Soc., 129, 8907, 2007. With permission.) (c) Long carbon nanotubes with conical tips and the inset shows the high resolution image of the conical tip. (d) A CCNT with the catalyst encapsulated at the tips. (From Merkulov, V.I. et al., Appl. Phys. Lett., 79, 1178, 2001. With permission.)

All me thods t hat em ployed su ldes, n itrides, a nd o xides f or metal s ources p roduced s traight, t hin-walled C MTs. A nother method using a mixture of activated carbon, ZnS and SnS, produced both straight and conical carbon tubes using different process conditions [21]. Ā e reaction was carried out in low-pressure N2 atmosphere at 1350°C. It was suspected that the Sn nanostructures with different morphologies templated the growth of CMTs with a similar variation in morphology. When Ga 2O3 was used as the me tal s ource, t he ga llium d roplets formed during d issociation assisted the growth of carbon wall around them [8,18]. In the case w here Mg a nd Fe c atalyst were u sed a s t he me tal s ources, growth of carbon tube at t he interface of MgO and Fe was proposed [23]. Different me chanisms h ave b een p roposed f or C MT g rowth using lo w-melting me tals. H owever, o nly t wo o f t he su ggested mechanisms are generic and can explain the growth of a variety of CMT morphologies. One of t he suggested mechanisms i nvolves vapor-liquid-solid t ype g rowth, i .e., p referential d issolution o f carbon into low-melting metals such as Zn followed by precipitation a s a t hin w all f orming t he t ube a round t he m icron-sized droplets (see the schematic in Figure 22.5a). Here, ZnS was used as the source for Zn metal along with activated carbon [19]. Ā e synthesis was carried out in N2 atmosphere at 200 mtorr base pressure and 1400°C. However, there is no t hermodynamic data available to su ggest t he p ossibility of c arbon d issolution i nto molten ga llium and other low-melting metals. On the other hand, the hydrocarbon adsorption onto the surface of low-melting metal melt is more likely (similar to that of bonding in trimethyl gallium). Ā e

43722_C022.indd 4

adsorbed h ydrocarbons co uld co ntribute to t he g rowth o f g raphene wall around the molten metal droplet, leading to the growth of t he CMT. Ā e molecular level aspects of carbon precipitation on molten metal surface still need to be explored. Ā e d roplet-led me chanism by it self do es not prov ide a ny insight into the growth of various CMT morphologies. In order to understand the low-melting metal droplet led growth of CMTs, it is proposed that the addition of carbon at the molten metal– carbon i nterface o ccurs i n suc h a w ay t hat t he c ontact a ngle between t he me tal a nd c arbon i s c onstantly m aintained. Ā is mechanism i s s chematically i llustrated i n F igure 2 2.5b. A s t he carbon tube grows in length, the interface lifts the metal droplet maintaining a steady contact angle between the meniscus and the growing carbon wall, thus setting up the conical angle of the overall structure. Ā e meniscus angle and the conical angle of the resulting carbon tube are related by the following relationship: φ = 2θ − 180
(2

2.1)

where f is the conical angle q is the meniscus angle Ā e l atter i s d irectly rel ated to t he c ontact a ngle b etween t he liquid metal and the carbon wall. For example, a meniscus angle of 90° should lead to a straight microtube [10]. The meniscus angle is primarily controlled by the wetting behavior of molten metal with carbon. The wetting or contact

10/15/2008 4:28:00 PM

22-5

Carbon Microtubes and Conical Carbon Nanotubes

00.4 00.2

20.1 10.0

Spiral graphene planes

100 nm (a)

(b)

Concentric graphene planes

5 nm

m

0n

11

Stacked graphene layers

.5⬚

14

002 Lattice tringes

(c)

Herring-bone tube

200 nm

FIGURE 22.4 Different wall structures observed for CC NT structures: (a) TEM image of a C CNT with inner hallow core and diffraction pattern showing spiral graphene sheets. (From M ani, R.C. et a l., Nano Lett., 3, 671, 2003. With permission.) (b) HRTEM image of a C CNT with central hallow core and wall made of c oncentric cylinderical graphene sheets. (From Z hang, G., Jiang, X., and Wang, E., Science, 300, 472, 2 003. W ith p ermission.) (c) T EM i mage of a C CNT s howing t he s tacked gaphe ne l ayers for ming a t apered i nner h allow c ore. (F rom Gogotsi,Y., Dimovski, Y.S., and Libera, J.A., Carbon, 40, 2263, 2002. With permission.)

TABLE 22.1 Review of All Synthesis Methods Reported to-Date for CMTs Carbon Source

Metal Source

Carrier Gas

Graphite powder

ZnS, SnS

N2

C2H2

GaN

Ar

Activated carbon Activated carbon

Ga2O3 CdS, SnO

N2 N2

Polyethylene

Ni powder

—

CH4

Ga

H2

43722_C022.indd 5

Other Gas Phase Components

Process Conditions

S (from dissociation of ZnS Ā ermal CVD, 1400°C, and SnS) 20 Pa base pressure N (from dissociation of Ā ermal CVD, 1150°C, GaN) 200 torr O from dissociated Ga2O3 S and O from dissociated CdS and SnO — Varying amounts of O2 and N2

Resulting Morphology Straight tube, conical horn

Tubes with very low conical angles with bulb like morphology on the top Ā ermal CVD, 1360°C Straight morphology Ā ermal CVD, 1150°C, Straight tubes with very small 1 atm conical angle Pyrolysis, 730°C–800°C, Straight tubes 90–100 MPa PECVD, 600–800°C, Straight tubes, cones, tune40 torr on-cones, funnels, n-staged morphologies capsule, dumbbells

Reference [19] [8]

[9,17] [20] [54] [22]

10/15/2008 4:28:01 PM

22-6

Smart Materials

Carbon microtube growth Hydrocarbon dissolution into low melting metal

(a)

Metal + carbon

Metal

Carbon tube

Metal + carbon Carbon tube

Oxygen dosing leading to straight tube growth

Hydrocarbons adsorption on molten metal surface

Formation of conical CMTs

β Metal droplet

Metal–carbon interface

(b) φ

FIGURE 22.5 Schematics illustrating various suggested mechanisms for growth of CMTs using low-melting metals. (a) A schematic illustrating a mechanism in which carbon dissolves into the molten metal droplet and precipitates as CMT. (b) A schematic illustrating another mechanism, which explains the molten metal-carbon wall interface and its inuence on the morphology of the resulting CMT.

angle o f mol ten me tal o n c arbon su rfaces c ould v ary b oth with t he ga s ph ase c omposition a nd t he tem perature. I f t he contact a ngle q, b etween c arbon a nd t he lo w mel ting me tal droplets, can be varied in-situ during growth, it is possible to control t he morphology of t he g rowing c arbon structure. In the case of gallium, it is known that the contact angle between gallium and carbon can be reduced in the presence of oxygen or ni trogen [ 24,25], r esulting in car bon tu bular s tructures with very small conical angles or near straight tubes as predicted by Equation 22.1. It is also likely that the presence of sulfur i n t he gas phase could modify t he su rface of t he lowmelting metal melt reducing its surface tension and, in turn, improving i ts w etting o n c arbon su rface [ 26]. T his me chanism of “wetting angle controlled growth” can also be applied to a lmost a ll t he methods reviewed in Table 22.1. For example, t he u se of ga llium oxide p owder provides t he ne cessary dosing of oxygen in the gas phase and can explain the results of straight CMTs with encapsulated gallium when using Ga2O3 and carbon powder [8]. Similarly, straight CMT structures were obtained when u sing Z nS [21], which is a lso consistent with the proposed mechanism of the gas-phase sulfur reducing the surface tension. However, the wetting properties also depend on the temperature with contact angles being higher at lo wer tem peratures. T he ob servation o f c onical C MTs i n the lo w-temperature re gions o f t he f urnace le ad to b etter understanding about t he wetting properties a s a f unction of temperature.

43722_C022.indd 6

22.2.2

Control of Morphology

According to E quation 22.1, the conical angle of the CMTs can be v aried b y c hanging t he w etting b ehavior o f mol ten me tals with carbon, which in turn, can be controlled by changing t he gas phase composition or temperature during their g rowth. In the a bsence o f o xygen o r n itrogen, e xperiments do ne at 600–1100 W m icrowave p ower, 4 0–90 torr p ressure o ver molten gallium with molybdenum as the promoter and CH4/H2 plasma en vironments, y ielded o nly c onical mo rphologies w ith conical angles of +20° (Figure 22.6a). Similar experiments with intentional g as p hase d osing o f 5 sccm o f oxygen h ave y ielded straight morphologies (Figure 22.6b). L ater experiments i n t he absence of oxygen and then followed by oxygen dosing yielded a predictable s hape o f t ube-on-cone mo rphology a s s hown i n Figure 22.6c. As predicted, the reverse sequence during CMT growth, i.e., initial dosing of oxygen followed by no dosing of oxygen produced a cone-on-tube or funnel-shaped morphology (Figure 22.6d). Repeating these dosing sequences “n” times, one can obtain “n-staged” morphologies as shown in Figure 22.6e. Even sm all a mounts of O 2 (5 sccm) leads to reduced surface tension of Ga, in turn reducing the conical angle of the resulting morphology signicantly. However, there is a limitation to how much O2 can be used before gallium oxidizes, resulting in Ga2O3 nanostructures [24]. Nitrogen on the other hand, does not readily form G aN at t he process temperatures u sed, a nd t herefore, can be used to f urther  ne-tune the conical angles. In fact, the

10/15/2008 4:28:03 PM

22-7

Carbon Microtubes and Conical Carbon Nanotubes

(a) (b)

1 µm 1 µm

(c)

(e)

(d)

With O2

500 nm

No O2

2 µm

1 µm

FIGURE 2 2.6 SEM i mages s howing v arious mor phologies of C MTs s ynthesized by c ontrolled, i n-situ o xygen do sing: ( a) f unnels (From Bhimarasetti, G., Cowley, J.M., and Sunkara, M.K., Nanotechnology, 16, S362, 2005. With permission), (b) cone-on-tube, (c) straight tubes, (d) tube-on-cone, and (e) six-stage morphology. (Figures c, d, e taken from Bhimarasetti, G. et al., Adv. Mater., 15, 1629, 2003. With permission.)

resulting morphologies, ultimately leading to nanocapsules with high n itrogen dosi ng ( Figure 22 .7a). Ā e a bility to c reate c onverging c ones c ould b e u sed to c reate p inched mo rphologies using a t hree-step s equence o f n itrogen do sing: t he  rst step involves growth for few minutes without the presence of nitrogen; t he second step involves high a mounts of nitrogen during the growth for few minutes to create a converging cone; and the

experiments using nitrogen dosing with amounts ranging from 0 t o 3 5 sccm y ield m orphologies w ith co nical a ngles ra nging from +20° to −15° (see Figure 22.7a and b). At higher nitrogen dosing, the conical angles of the resulting structures were negative, le ading to c onverging c ones c ompared to t he ob served diverging c ones w ith no o r l ittle a mounts o f do sing ( Figure 22.7c). Ā e converging conical geometry limits the length of the

1

(a)

0 sccm (b)

10 sccm

0.8 Arb units

N2 dosing

20 sccm

0.6

35 sccm

0.4 0.2 0 −40

−20

No N2 500 nm

0 20 Conical angle

40

(d) No N2 N2 dosing

(c)

No N2

1 µm

500 nm

FIGURE 22.7 SEM images and a plot showing the control of conical angle of the resulting microtube morphology using nitrogen dosing during growth: (a) capsule, (b) a plot showing the effect of increasing nitrogen concentration as a function on the conical angle of the resulting CMT, (c) horn structures with converging conical angles, and (d) dumbbell-shaped morphologies. (From Bhimarasetti, G., Cowley, J.M., and Sunkara, M.K., Nanotechnology, 16, S362, 2005. With permission.)

43722_C022.indd 7

10/15/2008 4:28:03 PM

22-8

Smart Materials

0.5 µm

(a)

(b)

l1

ne

n ha

C

Main

nel

chan

Cha

nne

l2

0.5 µm

2 µm

FIGURE 2 2.8 TEM i mages of br anched mor phologies: ( a) Y-junctions for med by s pontaneous c oalescence of l iquid me tals up on ph ysical impingement of two CMTs. (b) A CMT structure with multiple branches resulting through coalescence of several CMTs during growth.

third step involves growth without the presence of nitrogen. Ā e net morphology w ill be a d umb-bell-shaped pinched structure as shown in Figure 22.7d in which the lengths of each portion are controlled by the durations used for each step. One can also control the internal diameters of the resulting structures by varying the t imes at w hich n itrogen i s i ntroduced d uring t he C MT growth [26]. Ā e w alls o f t he c onical C MTs c ontain na nocrystals o f graphite w hose o rientation w ith re spect to t he w all de pend very much on the conical angle of the tube. The specific orientation o f e ach na nocrystal c an b e d irectly rel ated to t he contact a ngle b etween c arbon a nd ga llium d uring t he t ube formation [26]. The molten metal droplets on the tips of two or more CMTs could coalesce quickly into one larger droplet during growth upon physical impingement leading to Y- and multiple-junction C MTs for m icrofluidic appl ications a s shown in Figure 22.8a and b. The as-synthesized CMT structures are partially filled with the low-melting metal used. As the CMTs have open channels, it is fairly easy to remove the low-melting metals using either acids to dissolve the metals or by heating them in vacuum. Ā e availability of CMT structures in large quantities will nd uses a s me soporous m aterials, l ithium io n bat teries, a nd f or other ele ctrochemical energ y c onversion app lications. Cu rrent efforts in the laboratories of the authors and others are aimed at large-scale production and purication of these structures from metal contamination. In addition to large-scale production, the integration of these structures into microuidic devices will also need to be addressed.

22.3

Synthesis and Morphological Control of CCNTs

22.3.1

Synthesis Methods and Growth Mechanisms

Ā e synthesis of the CCNT structures differs extensively from that of thin-walled, carbon microtubular structures. Firstly,

43722_C022.indd 8

the synthesis me thods do not i nvolve lo w-melting me tals and secondly, t he C CNT s tructures a re s ynthesized at m uch higher t emperatures. M ost o f t hese C CNT st ructures a re formed with the help of transition metal catalyst particles and characteristically c ontain a na nometer s cale hollow c ore or a central nanotube. Ā e plasma discharge assisted growth of multiwalled carbon nanotubes u sing c atalysts re sulted i n t he f ormation o f c onical carbon nanotubular structures or termed as CNC [7]. Ā ese structures contained t he c atalyst particle encapsulated at t heir tips w hile t he t hickness t apered f rom m icrons at t he ba se to nanometers at the tip. Ā e s ynthesis w as p erformed u sing a PECVD system with acetylene and ammonia in the gas phase at 700°C and pressures of 2–5 torr. Ā e growth of conical morphology is explained by the growth of central carbon nanober from the c atalyst pa rticle, a nd f urther c arbon de position f rom t he plasma-assisted h ydrocarbon de composition. Ā e ne t re sult i s the growth in two directions, vertically (inner carbon ber) and laterally (the deposition on t he outer wall) forming t he carbon nanocone. F igure 2 2.9a de picts t he v arious s teps i n t he a bove mechanism [7]. Two si milar s ynthesis s tudies w ere re ported f or C CNT structures w ith op en-ended t ips a nd i nner c hannels f ree of any internal obstructions: One of the studies produced conical morphologies termed as TGC by placing iron needles vertically i n a m icrowave p lasma d ischarge w ith me thane a nd nitrogen, at 750 W, 600°C, 15 torr pressure [6]. In this case, the iron from the needle acted as the catalyst to promote the central carbon nanotube growth. In the other report, similar CCNT morphologies termed as CNPs were synthesized on platinum w ires placed vertically i n a m icrowave plasma d ischarge c ontaining me thane a nd h ydrogen, at 95 0 W a nd 25 torr pressure [5]. Both types of structures are very similar, having op en t ips w ith no me tal c atalyst p articles a nd no obstructions within the internal channel. In both reports, it was su ggested t hat sm all a mounts o f me thane d iluted i n either hydrogen or argon resulted in high surface temperature at the wire substrate. In both cases, a catalyst particle

10/15/2008 4:28:04 PM

22-9

Carbon Microtubes and Conical Carbon Nanotubes

2C + H2

C2H2

H2

C

Deposition Growth

Plasma Discharge Deposition

Nickel (a) Vertical growth versus basal nucleation

Basal thickening of carbon tube walls Nucleation and growth carbon tubes C, Cx Hy

(b)

Iron Steel substrate

Steel substrate

Growth of CNT and etching Carbon deposition (c)

Steel substrate

Evaporation of platinum

Additional deposition and etching

Deposition and etching

Steel substrate Etching and deposition Etching

Deposition

Growth at edge sites

Platinum

FIGURE 22.9 Schematics illustrating diff erent g rowth mechanisms suggested for c onical c arbon nanotubular g rowth: (a) t he c atalyst a ssisted growth of central carbon nanotube along with uneven plasma deposition/etching enables the observed growth of conical graphite structure, (b) the catalyst-assisted basal growth of a central carbon nanotube followed by continuous nucleation of new graphene walls for thickening the base while tapering vertically in thickness forming a c onical structure, and (c) catalyst assisted tip-led growth of a c entral carbon nanotube followed by the evaporation of the catalyst particle and simultaneous deposition and etching of the helical graphite edge planes leading to the conical morphology.

(either i ron or pl atinum) he lped w ith t he n ucleation a nd growth of a multiwalled carbon nanotube. However, there are two c ontrasting su ggestions o n h ow t apered g rowth o ccurs: The f irst su ggested me chanism i mplies t hat t he ba sal ac cumulation of iron catalyst initiates nucleation of new graphene walls thickening at the base and its competition with vertical growth leads to tapering (Figure 22.9b). The second suggested mechanism s tates t hat a s t he c entral m ultiwalled na notube continues to grow axially with catalyst at its tip, it acquires a tapered morphology in the early stages as an effect of the shrinkage of the catalyst particle due to evaporation. Because of the presence of plasma discharge containing methane and hydrogen, f urther g rowth c an o ccur w ith i nitiation o f ne w scrolling g raphene s heets a nd g rowth at t he op en-ended t ip without the presence of catalyst (Figure 22.9c). In addition to deposition, the plasma exposure leads to greater etching near the t ip, p reserving t he c onical mo rphology e ven at lo nger synthesis timescales. Experiments i nvolving hydrocarbon de composition o n to presynthesized carbon fibers also resulted in sparse growth of conical-shaped morphologies on carbon fiber substrates [27]. The underlying mechanism is not clear but may be similar to that shown in Figure 22.9b. Conically shaped tips on multiwalled carbon nanotubes were also observed with fur-

43722_C022.indd 9

ther h ydrocarbon d ecomposition o nto i ron c atalyst l oaded, presynthesized carbon fibers [28]. Similar structures of conical sha ped t ips o n m ultiwalled c arbon na notubes sh own i n Figure 22.4c were synthesized using a s ynthesis temperature of 1325°C and bubbling of ferrocene in toluene, using hydrogen a nd a rgon ga s m ixtures [15]. O ne o f t he su ggestions i s that t he c onical s hape i s ac quired f rom t he m ultiwalled carbon nanotube growth basally on a c onically faceted catalyst in the early stages. Irrespective of the underlying growth mechanism, it is essential to grow CCNT arrays on to large area, f lat substrates for applications. H owever, i t i s d ifficult to re produce t he e xact conditions on the f lat substrates with those e xisting on the wire substrates immersed i n p lasma d ischarges f or f orming the CCNT structures. For example, t he temperature a nd t he radical species density t hat exist on t he i mmersed w ire substrates are much higher t han t he t ypical conditions t hat can be obtained for flat substrates. Only one to two reports beyond the a uthors’ labo ratory ex ist o n t he growth of CCNTs on planar sub strates a nd t he re sults i ndicated limited control on t he re sulting mo rphology a nd t he a rray den sity [16,29]. Much more work needs to be done in obtaining similar CCNT array s tructures o n l arge a rea, f lat sub strates f or v arious applications.

10/15/2008 4:28:06 PM

22-10

Smart Materials

(a)

(c)

(b)

5 µm

5 µm

1 µm

FIGURE 22.10 SEM images illustrating different morphologies of CCNT structures grown using platinum wire substrates placed vertically in the m icrowave plasma d ischarge: (a) t he CCNTs w ith large bases containing he avy a mount of lo ose c arbon a round t heir bases, ( b) the t ypical morphology of a high density of CCNTs grown in an array fashion, and (c) low-density arrays of CCNTs.

22.3.2

Morphological Control

The shape of the conical carbon nanotube structures grown on planar substrates with tip-led growth were mostly affected by t he ga s ph ase c omposition [7]. I n t he c ase o f t he C CNT structures grown using direct immersion of substrates inside plasma d ischarges, t he leng th, t apering, a nd t he a rray density are affected by varying different process conditions such as pressure, microwave power, and the wire substrate’s placement inside the plasma discharge [5,6,16,29]. The latter condition seems to be the most effective in terms of controlling the t apering ge ometry a nd t he s tructure o f t he re sulting tapered morphology with simultaneous etching and growth. Rapid de position o f lo ose c arbon at ba ses re sults i n lo w aspect r atio c onical s tructures as s hown in F igure 2 2.10a. The high aspect ratio CCNTs are grown where growth/etching is optimized as shown in Figure 22.10b. The aspect ratios can a lso b e c ontrolled b y c hanging t he c omposition o f H 2 and CH4 [30]. In a similar way, the density of the CCNTs can be va ried b y cha nging t he p rocess co nditions a s sh own i n Figure 22.10c.

(a)

500 nm

(b)

500 nm

Postsynthesis s trategies c ould a lso b e u sed f or a ltering t he tapered m orphology o f t he p resynthesized C CNT st ructures. Ā e H2 plasma etching can be used to remo ve the loose carbon akes a nd p rolonged e tching c an le ad to s traight m ultiwalled CNT a s s hown i n F igure 2 2.11a a nd b [ 30]. H ydrogen e tching can also result in carbon nanotube on CCNT tips as shown in Figure 22.11c. Ā e oxidation of conical carbon tubular structures grown on silicon wafers [16] modied the CCNT structure to a stepped cylindrical morphology as shown in Figure 22.11d. Ā e tip size and inner diameter can be altered by changing the initial c atalyst pa rticle si ze a s t he l atter p lays a ke y role i n determining t he si ze o f t he M WNT. It i s a lso a ssumed t hat t he spinning of the catalyst droplet formed on the surface promotes the growth of helical graphite sheets surrounding the core CNTs [29]. Ā e tem perature (controlling t he si ze o f c atalyst pa rticle) and the methane concentration should be sufficient for the growth on  at substrates, where conditions are different from the wires immersed into the plasma. Ā ough the temperature of substrate was low, self-bias a round t he i ron needles i ncluding t he strong electric  eld i n t he p lasma re sulted i n c onical s tructures [6].

(c)

500 nm

(d)

100 nm

FIGURE 22.11 SEM images of the CCNTs after postsynthesis etching treatments: (a) an as-synthesized CCNT; (b) the CCNT after fully etched using hydrogen plasma exposure at 800°C–900°C temperature; (c) the CCNT structure after partial etching with hydrogen plasma exposure, and (d) the telescopic structure of the CCNT structure formed with thermal oxidation using oxygen. (From Shang, N., Milne, W.I., and Jiang, X., J. Am. Chem. Soc., 129, 8907, 2007. With permission.)

43722_C022.indd 10

10/15/2008 4:28:07 PM

22-11

Carbon Microtubes and Conical Carbon Nanotubes

Continuous supply of metal suppressed the growth of CNTs, encouraging t he g rowth of TG Cs [29]. Ā e f ormation o f t ubes having a c ontinuous t apered mo rphology, a s s hown i n F igure 22.3c, ca n b e e xplained by catalyst as sisted g rowth, w ith c ontinuous e vaporation o f t he c atalyst pa rticle, f orming ∼200 nm long c onical s tructures, b efore t he c atalyst pa rticle e vaporates completely. A s teady supp ly o f c atalyst pa rticles le ads to t he growth of new tapered tubes on top of the already existing ones from t he new catalyst pa rticles, u ltimately forming long t ubes. Similar structures, as shown in Figure 22.4b can also be explained by the same growth mechanism.

and ascorbic acid compounds occur at nearly the same potential (∼200 mV vs. Ag/AgCl). Ā e as-synthesized conical carbon tube array e lectrodes a fter o ne-time o xidation t reatment at a nodic potential of 2.0 V vs. Ag/AgCl in 0.5 M H2SO4 shifts the ascorbic acid p eak w hile le aving t he dopa mine p eak i ntact. S ee F igure 22.12 showing well resolved peaks of ascorbic acid and dopamine [31]. In essence, the CCNT arrays can become an important class of electrochemical materials for sensing neurotransmitters and other important biological species.

22.4

Conical c arbon n anotubular s tructures ar e a lso o f in terest for na noelectrode app lications d ue to t heir na noscale t ips. Nanoelectrodes o ffer sev eral a dvantages co mpared t o t heir micron-sized c ounterparts d ue to t heir sm all c ritical d imensions. Nanoelectrodes exhibit radial diff usion, rapid attainment of steady state c urrents, i ncreased sig nal to bac kground ratios, higher sensitivities, and decreased effects from an ohmic potential drop compared to large electrodes [31]. However, an ensemble o f na noelectrodes i nstead o f a si ngle na noelectrode i s required for obtaining t he i ncreased sig nals. I n order to m ake such an ensemble, the spacing between each nanotip should be in t he o rder o f m icrons to a void o verlapping c oncentration boundary layers. Conical CNP arrays, due to their conical geometry, contain tips spaced f rom each ot her by few m icrons depending upon their base widths. However, the as-synthesized conical carbon nanopippette arrays behave as microelectrodes. Ā is is because the conical carbon structures touch each other at t heir bases.

Applications

22.4.1 New Electrode Materials All conical carbon tubular structures have unique surface structures c ompared to b oth si ngle w alled a nd m ultiwalled c arbon nanotubes. Ā e surfaces of the conical carbon tubular structures are completely made up of highly reactive graphene edge sites [5]. So, t he ele ctrochemical p roperties o f a ny c onical c arbon material will be different from that of traditional carbon-based materials suc h a s h ighly o riented py rolytic g raphite, c arbon nanotubes, a nd d iamond. T ypically, s everal re dox re actions involving neurotransmitters such as dopamine exhibit irreversibility on a number of carbon-based materials with peak separations a s h igh a s 1000 mV. H owever, t he k inetics o f t he s ame redox reaction on the as-synthesized conical carbon tube arrays are fast and reversible with peak separations less than 150 mV. In addition, dopamine needs to be detected in the presence of high amounts of ascorbic acid. But, the oxidation peaks of dopamine

22.4.2

Templates for Nanoelectrode Ensembles

−3 −8 −13 Current (µA)

−18 −23 −28

Dopamine

−33 −38 −43 Ascorbic acid

−48 −53 1000

800

600

400 200 E (mV vs. Ag/AgCI)

0

−200

FIGURE 22.12 A cyclic voltammogram using CNP electrode at 100 mV/s scan rate in a 0.1 mM dopamine and 10 mM ascorbic acid solution in 0.1 M KCl in 0.2 M phosphate buffer solution, pH 7, after a single anodic oxidation treatment. (From Lowe, R.D. et al., Electrochem. Solid State Lett., 9, H43, 2006. With permission.)

43722_C022.indd 11

10/15/2008 4:28:08 PM

22-12

Smart Materials Independent boundary layers CNP Polymer coating

5 mm Pt (b)

(a)

Current (µA)

5 3

11

400 200 50 20 10

Current (nA)

7

1 −1 −3 −5 500

(c)

9 7 5

400 200 50 20 10

3 1

300

100

−2 500

−100

E (mV vs. Ag/AgCI)

(d)

300 100 E (mV vs. Ag/AgCI)

−100

FIGURE 22.13 Fabrication and electrochemical testing of nanoelectrode ensembles using the CCNT arrays: (a) a schematic illustrating the effect of dip coating of CCNT array using a polymer for the separation of diff usion boundary layers for each nanotip exposed. Polymer coating helps to space the electrodes; (b) the SEM of C NP NEE after polymer coating showing the well-spaced nanotips; (c) the cyclic voltammograms taken at 10–400 mV/s scan rates for an uncoated CNP array exhibiting planar electrode behavior with peak-shaped responses; and (d) the cyclic voltammograms u sing p olymer-coated C NP N EE s howing t he si gmoidal re sponse i ndicating t he n anoelectrode b ehavior. (From L owe, R .D. e t al., Electrochem. Solid State Lett., 9, H43, 2006. With permission.)

A simple dip coating of CNP arrays using a p olymer exposes only the top portion of well-separated, conical tips to electrolyte as shown in Figure 22.13a and b. Electrochemical studies performed on t hese ele ctrode a ssemblies s how a v ery c lear s hift from m icroelectrode be havior t o na noelectrode be havior when coated with the polymer. Cyclic voltammetry performed on these electrodes is shown in Figure 22.13c and d. Ā e plots show t hat t he u ncoated ele ctrodes e xhibit a p lanar ele ctrode behavior w ith a p eak-shaped re sponse a nd t he c oated ele ctrodes exhibit a sigmoidal response, indicative of steady-state, diff usion-limited currents [31]. Even though carbon itself is an interesting material for several applications b ased on n anotips, c onical-shaped n anotip a rrays of other materials are also interesting. In this regard, the nanoelectrode ensembles of other materials could easily be fabricated by depositing materials of interest onto the conical carbon tube arrays. Ā e d iamond na notip a rray ele ctrodes a re f abricated easily by depositing diamond on to conical carbon tube arrays (see Figure 22.14a and b) [30].

22.4.3

Field Emission Applications

Ā e CCNT structures with conical geometry and nanoscale tips are i nteresting f or  eld em ission app lications. A re cent s tudy

43722_C022.indd 12

[32] reported very low threshold voltages of 0.27 V/µm a nd a current density of 1 µA/cm2. A s table emitting current density of 1 .9 mA/cm2 at only 0.6 V/µm w as re ported. I n add ition to having lo w w ork f unction, t he c onical ba ses a llow m aximum contact with the substrate for high  eld emission currents over long p eriods o f t ime. Ā e t hreshold vol tage s hown b y t hese structures w as t hree t imes lo wer t han t hat s hown b y c onical nanobers [33]. Ā e re sults s hown i n Re f. [32] a re a ffected to some extent by the presence of encapsulated metal catalyst particles at t heir t ips. H owever, t he t rue w ork f unction a nd  eld emission properties of c onical carbon nanotube structures are yet to be understood.

22.4.4

Porous Carbons

Large quantities of CMTs with controlled internal diameters and conical angles could be useful as mesoporous carbons for high su rface a rea a nd c atalyst supp ort app lications. I n gen eral, t he applications of p orous c arbons c an b e g rouped i nto two categories: (i) based on the architecture of their pores for structural and thermal applications and their use as templates for m aking c eramics a nd (ii) a s ac tivated c arbons. Activated carbons are used extensively as catalysts and catalyst supports for w ater a nd a ir p urication, remo ving c olor f rom v arious

10/15/2008 4:28:08 PM

22-13

Carbon Microtubes and Conical Carbon Nanotubes

microneedles present a su itable a lternative as t hey a re pa inless and can be used for highly localized drug targeting. Many different de signs o f m icroneedle-based d rug del ivery s ystems h ave been demonstrated. In many cases, these micropipettes are fabricated r ather t han s ynthesized. O ne limitation in microfabricating vertical microneedles is their maximum achievable length [42]. A lso, t he m aterials w hich h ave b een u sed f or f abricating these needles are mainly limited to silicon [43], glass [44], metal, and b iodegradable p olymer [ 45]. C arbon ba sed na nopipettes were fabricated using chemical vapor deposition of carbon onto catalyst coated, interior surfaces of pulled quartz capillary (or nanopipette) tubes, and removing the quartz exterior [46]. Such carbon-based na nopipettes situated at t he end of a lo ng quartz capillary will be useful in both drug delivery and sensing applications. Ā e synthesized carbon based micro/nanopipette arrays and needles are also of immense interest to t he above type of transdermal drug delivery applications if they can be integrated into macroscale s ystems. M oreover, re cent s tudies o n  uid ow through carbon nanotubes have shown interesting  uid behavior due to t he hydrophobicity of t he i nterior su rface. Ā e uid ow through 7 nm diameter nanotubes was shown to be four to ve orders of magnitude faster than that predicted by conventional  uid dy namics [ 47]. Ā is w as at tributed to t he a lmost frictionless i nterface at t he C NT w all. Si milar re sults w ere obtained using even smaller nanotubes (∼2 nm) for liquids and gases [ 48]. Fu rther s tudies s howed t hat b y at taching d ifferent end groups to nanotubes, selective liquids can be own through them [49,50]. With open-ended tips and the mechanical rigidity of t he c onical s tructure, t he  uid ow t hrough C CNT a rrays should nd several applications including drug delivery.

types of sugar syrups, separating amino acids from their solution acetic acid, separating aromatic acids, catalysts, and catalyst supp orts [34]. A ctivated c arbon m aterials a re a lso b eing investigated for possible applications in hydrogen storage [35]. Microtubular materials can offer highly interconnected porous structures enabling facile diff usion of species into a nd out of the c arbon s tructures. Ā e no ncarbon-based m icrotubular materials a re a lso o f i nterest. F or e xample, t he a luminum borate microtubular materials are used in structural applications [36,37], zinc oxide microtubes show promise in gas sensing [38], a nd Z nO na no/microtubes a re go od c andidates f or safe biological cavities and template materials due to their innocuity a nd chemical characters. High surface area silicon carbide (SiC) is used as an important high-temperature structural material, such as catalyst support and hot-gas  lter, and offers many advantages due to its unique properties, e.g., low thermal e xpansion c oefficient, go od t hermal shock resistance, and chemical stability at elevated temperature [39] Similarly, TiO2 microtubes for photocatalytic applications [40], diamond microtubes a s rei nforcement a gents [41] a re k nown app lications. Ā e CMTs can serve as templates for producing microtubes of other materials. For example, diamond can easily be deposited o n to C MTs t ransforming them into diamond microtubes with shapes of the original CMTs as shown in Figure 22.14c and d.

22.4.5

Drug Delivery, Fluid Flow, and Other Miscellaneous Applications

Conventional syringes used for drug delivery are often pa inful and l imit t argeted d rug del ivery. I n t his re gard, na no a nd

1332.8

500 nm Intensity (a.u.)

(a)

1584.3 1616

1148

2795

Raman shift (cm−1)

(b) (c) Intensity (a.u.)

1584

5 µm

1332

1300

(d)

1400

1500

1600

1700

−1 Raman shift (cm )

FIGURE 22.14 (a) SEM image of CCNT arrays covered with diamond, (b) visible Raman spectra obtained on the diamond coated CCNT arrays, (c) SEM image of a diamond coated CMT, and (d) UV Raman spectra obtained on the diamond coated CMT indicating the presence of ultrananocrystalline diamond.

43722_C022.indd 13

10/15/2008 4:28:09 PM

22-14

Micropipette structures are also of interest for many new types of applications suc h a s fountain p en c hemistry w hen i ntegrated with atomic force microscope’s cantilever [51]. Nanopipette syringes, w hich c an d ispense  uids w ith a p recise c ontrol o f volume, a re of t remendous i nterest. Two k inds of s yringes were reported, a double barrel nanopipette [52] and an electrochemical attosyringe [53]. In the double barrel type nanopipette, an electrochemical potential is applied to two different uid compartments, which sets up an electroosmotic ow of the uid out of the nanopipette. By controlling parameters, such as the applied voltage and pulse duration, the volume of uid coming out of the syringe can be controlled to attoliter precision. In the electrochemical attosyringe [54], a n applied potential between electrodes i mmersed i n the uid inside and the uid surrounding the nanopipette induced the uid ow into or out of the pipette with attoliter precision. Another i nteresting application for t he ele ctron t ransparent micro t ubes, pa rtially  lled w ith ga llium i s t heir u se a s na nothermometers [8]. Gallium exists as liquid from 29.8°C to 1050°C while exhibiting low vapor pressure. Interestingly, the expansion coefficient of ga llium i nside t he na notube w as t he s ame a s t he expansion coeffi cient of ga llium i n t he macroscopic s tate. Ā is was u nlike t he mel ting p oint f or na noquantities o f m aterials which are inuenced by surface effects.

22.5 Summary Ā e s ynthesis, s tructure, a nd app lications o f v arious m icronscale a nd c onical mo rphologies o f c arbon t ubes h ave b een described and discussed.

Acknowledgments Ā e a uthors w ould l ike to ac knowledge supp ort f rom U .S. Department of Energy (DE-FG02–05ER64071 and DE-FG02– 07ER46375) a nd U .S. A rmy Spac e M issile D efense C ommand (W9113M-04-C-0024). Ā e a uthors a lso ac knowledge t he Chemical Engineering department at the University of Louisville for support, Dr. Zhiqiang Chen for help with TEM analysis, and the I nstitute for Advanced Materials a nd Renewable Energy at University of Louisville for allowing us to access to the materials characterization f acilities suc h T EM, S EM, a nd R aman spectroscopy.

References 1. Bacon, R ., G rowth, str ucture, a nd p roperties o f gra phite whiskers, J. Appl. Phys., 31, 283, 1960. 2. Haanstra, H.B., Verspui, G., and Knippernb, W.F., Columnar growth of carbon, J. Cryst. Growth, 16, 71, 1972. 3. Iijima, S., Helical microtubules of graphitic carbon, Nature, 354, 56, 1991. 4. Ajayan, P.M., Nanotubes from carbon, Chem. Rev., 99, 1787, 1999. 5. Mani, R.C. et al., Carbon nanopipettes, Nano Lett., 3, 671, 2003.

43722_C022.indd 14

Smart Materials

6. Zhang, G., Jiang, X., and Wang, E., Tubular graphite cones, Science, 300, 472, 2003. 7. Merkulov, V.I. et al., Shaping carbon nanostructures by controlling the syn thesis p rocess, Appl. P hys. L ett., 79, 1178, 2001. 8. Gao, Y. and Bando, Y., Carbon nanothermometer containing gallium, Nature, 415, 599, 2002. 9. Libera, J. and Gogotsi. Y, Hydrothermal synthesis of graphite tubes using Ni catalyst, Carbon, 39, 1307, 2001. 10. Bhimarasetti, G. et al., Morphological control of tapered and multi-junctioned carbon tubular stuctures, Adv. Mater., 15, 1629, 2003. 11. Iwanaga, H., K awaguchi, M., a nd M otojima, S., G rowth mechanisms an d prop erties of c oiled w hiskers of s iliconnitride and carbon, Jpn. J. Appl. Phys., 32, 105, 1993. 12. Krishnan, A. et al ., G raphitic co nes a nd the n ucleation o f curved carbon surfaces, Nature, 38, 451, 1997. 13. Ge, M. a nd Sa ttler, K., Obs ervation o f f ullerence co nes, Chem. Phys. Lett., 220, 192, 1994. 14. Iijima, S., Yusasaka, M., Yamada, R ., B andow, S., Suenaga, K., Kokai, F., and Takahashi, K., Nano-aggregates of single-walled graphitic carbon nano-horns, Chem. Phys. Lett., 309, 165, 1999. 15. Kona, S., M.S. thesis, University of Louisville, 2007. 16. Shang, N., Milne, W.I., and Jiang, X., Tubular Graphite Cones with single crystal nanotips and their antioxygenic properties, J. Am. Chem. Soc., 129, 8907, 2007. 17. Gogotsi, Y., Dimovski, S., and Libera, J.A., Conical crystals of graphite, Carbon, 40, 2263, 2002. 18. Gao, Y. and Bando, Y., Nanothermodynamic analysis of surface eff ect o n exp ansion cha racteristics o f Ga in ca rbon nanotubes, Appl. Phys. Lett., 81, 3966, 2002. 19. Hu, J. et al., Growth and eld-emission properties of crystalline, thin-walled carbon micro tubes, Adv. Mater., 16, 153, 2004. 20. Pan, Z.W . et al ., Galli um-mediated gr owth o f m ultiwall carbon nanotubes, Appl. Phys. Lett., 82, 1947, 2003 21. Shen, G. et al ., Tubular carbon nano-/microstructures synthesized f rom g raphite p owders b y a n in si tu tem plate process, J. Phys. Chem. B, 110, 10714, 2006. 22. Hu, J . et al ., T apered ca rbon na notubes f rom ac tivated carbon powders, Adv. Mater., 18, 197, 2006. 23. Sun, Z. et al., Synthesis of tubular graphite cones through a catalytically thermal reduction route, J. Phys. Chem. B, 108, 9811, 2004. 24. Sharma, S. and Sunkara, M.K., Direct synthesis of gallium oxide tubes, nanowires, and nanopaintbrushes, J. Am. Chem. Soc., 124, 12288, 2002. 25. Graham, U.H. et al ., Morphological control of tapered and multi-junctioned ca rbon t ubular str uctures, Adv. Fu nct. Mater., 13, 576, 2003. 26. Bhimarasetti, G., Cowley, J.M., and Sunkara, M.K., Carbon microtubes: t uning in ternal dia meters a nd co nical a ngles, Nanotechnology, 16, S362, 2005. 27. Xia, W. et al., Conical carbon laments with axial cylindrical channels and open tips, Adv. Mater., 17, 1677, 2005.

10/15/2008 4:28:09 PM

Carbon Microtubes and Conical Carbon Nanotubes

28. Muradov, N. and Schwitter, A., Formation of conical carbon structures on vapor-grown carbon laments, Nano Lett., 2, 673, 2002. 29. Shang, N.G. a nd J ianga, X., L arge-sized t ubular gra phite cones w ith nano tube t ips, Appl. P hys. L ett., 87, 163102, 2005. 30. Chernomordik, B. et al ., Nanodiamond tipped and coated conical ca rbon t ubular st ructures, chem. va p. D eposition., 14, in press, 2008. 31. Lowe, R .D. et al ., N anoelectrode en sembles usin g ca rbon nanopipettes, Electrochem. Solid State Lett., 9, H43, 2006. 32. Li, J.J. et al., Field emission from high aspect ratio tubular carbon cones grown on gold wire, Appl. Phys. Lett., 87, 143107, 2005. 33. Tanemura, M. et al ., Field ele ctron emission f rom sputterinduced ca rbon na nobers g rown at ro om t emperature, Appl. Phys. Lett., 86, 113107, 2005. 34. Monacha, S., Sadhana, Parts 1&2, 335, 2003. 35. Ā o mas, K.M., Catal. Today, 120, 389, 2007. 36. Yang, W. et al ., P olygonal sin gle-crystal al uminum b orate micro tubes, J. Am. Ceram. Soc., 88, 485, 2005. 37. Ma, R . et al ., S ingle-crystal Al18B4O33 micr o t ubes, J. Am . Chem. Soc., 124, 10668, 2002. 38. Fu, M. et al ., L., Tetrapod-shaped ZnO microtubes synthesized f rom Z n/C mixt ures Mater. Res. B ull., do i:10.1016/j. materresbull.2007.04.035, 2007. 39. Kim, J .W. et al ., S ynthesis o f S iC micr otubes wi th radial morphology using biomorphic carbon template, Mater. Sci. Eng. A, 434, 171, 2006. 40. Motojima, S. et al., Preparation of helical TiO2/CMC microtubes a nd p ure helical T iO2 micr otubes, J. Ma ter. S ci., 39, 2663, 2004. 41. Dua, A.K. et al ., Formation of s elf-supporting ho llow diamond helix and diamond sieve using jet-ow HFCVD, Solid State Commun., 93, 759, 1995. 42. Reed, M.L., Microsystems for drug and gene delivery, Proc. IEEE, 92, 56, 2004.

43722_C022.indd 15

22-15

43. Stoeber, B . a nd L iepmann, D ., D esign, fa brication, a nd Testing o f a MEMS S yringe, P roceedings o f S olid-State Sensor a nd Actuator Workshop. 2–7 J une, 2002, H ilton Head I sland, SC, Transducers Res. F ound (TRF C at. N o. 00TRF-0001). 44. Martanto, W. et al ., Minco infusion usin g ho llow micr o needles, Pharmaceut. Res., 23, 104, 2006. 45. McAllister, D.V. et al., Microfabricated needles for transdermal deliv ery o f macr omolecules a nd na noparticles: Fabrication methods and transport studies, Proc. Natl Acad. Sci., 100, 13755, 2003. 46. Kim, B .M., M urray, T., a nd B au, H.H., Ā e fa brication o f integrated ca rbon p ipes wi th sub micron dia meters, Nanotechnology, 16, 1317, 2005. 47. Majumder, M. et al ., Nanoscale hydrodynamics: Enhanced ow in carbon nanotubes. Nature, 438, 44, 2005. 48. Holt, J. K. et al., Fast mass transport through sub-2-nanometer carbon nanotubes. Science, 312, 1034, 2006. 49. Majumder, M., Chopra, N., and Hinds, B.J. Effect of tip functionalization on t ransport t hrough ve rtically or iented carbon nanotube membranes. J. Am. Chem. Soc., 127, 9062, 2005. 50. Babu, S. et al., Y. Guiding water into carbon nanopipes with the aid of bipolar electrochemistry. Microuid. Nanouid., 1, 284–288 2005. 51. Lewis, A. et al., Fountain pen nanochemistry: Atomic force control of chr ome etching , Appl. P hys. L ett., 75, 2689, 1999. 52. Rodolfa, T. et al., Nanoscale pipetting for controlled chemistry in small a rrayed wa ter dr oplets usin g a do uble-barrel pipet kit, Nano Lett., 6, 252, 2006. 53. Laforge, F.O. et al., Proc. Natl. Acad. Sci., 104, 11895, 2007. 54. Han, C.C., Lee, J.T., Yang, R.W., Chang, H., and Han, C.H., A new and easy method for making well-organized micrometer-sized ca rbon t ubes a nd their r egularly ass embled structures, Chem. Mater., 11, 1806, 1999.

10/15/2008 4:28:09 PM

43722_C022.indd 16

10/15/2008 4:28:09 PM

23 “Smart” Corrosion Protective Coatings 23.1 I ntroduction ...........................................................................................................................23-1 23.2 Smart Coating Technologies for Corrosion Protection ....................................................23-1

Patrick J. Kinlen Crosslink

Martin Kendig Teledyne Scientific Company

23.1

Chromate and Partially Soluble Pigments · Environmentally Responsive Corrosion Protection

23.3 I mportant Opportunities ....................................................................................................23-14 23.4 S ummary ...............................................................................................................................23-15 References .........................................................................................................................................23-15

Introduction

“Smart” corrosion protective coatings have gained recent attention a s a re sult o f ne w de velopments gener ally i n t he  eld of smart environmentally responsive materials. For the purpose of this c hapter, sm art c orrosion p rotection re fers t he a bility o f a protective coating not only to provide a barrier to corrosive compounds, but also to s ense one of the following: (1) the environment, (2) changes in the environment, (3) its own condition, or (4) c hanges i n i ts c ondition, a nd, o n t he ba sis o f t hese i nputs, generate a n app ropriate c orrosion-protective re sponse. H ighperformance c orrosion pr otective c oatings h ave a lways f unctioned more than by acting as a barrier to water, ions, and corrosive chemical species. In fact, f rom a n h istorical perspective, the best corrosion protective coatings have provided some environmentally s timulated p rotective re sponse. F or e xample, zinc and tin coatings have been utilized for more than 150 years as protective coatings providing galvanic protection of defects in the p resence o f a n ele ctrolyte en vironment, a nd t hey gener ate corrosion-inhibiting s pecies w hen a ga lvanic c urrent  ows. Mayne [1] pointed out that coatings serve as more than barriers since permeation by oxygen exceeds the rates necessary to su stain corrosion reactions. He also showed the reaction of leadbased p igments w ith a n o il-based re sin c oating a nd t he environment p roduces a zelaic ac id, a demo nstrated c orrosion inhibitor for iron. It has long been known that sparingly soluble hexavalent chromium inhibitors (e.g., Sr, Ca, and Zn chromate) release i nhibiting c hromates w hen t he c oating b ecomes w et. Results have shown that conversion coatings formed by hexavalent c hromate (Cr(VI) ) re tain u nreacted Cr (VI) t hat b ecomes active up on t he f ormation o f a p it t hat c athodically p olarizes

adjacent su rfaces, rele asing c hromate a s a re sult o f t he c onsequent increase in pH [2]. It is clear then that smart coatings have always provided the best inhibition. W ith re cent d evelopments i n n anotechnology, i nherently conducting polymers and composite materials, more precise engineering of smart behavior in corrosion protective coatings has started and will benet environmental concerns about the historic use of hazardous a nd c arcinogenic materials, such a s hexavalent chromium, common to existing smart coatings. In addition, smart coatings for corrosion prevention can use t he energ y inherent in the me tal i tself ( an a erated en vironment p rovides ne arly 2 V o f electrochemical energy) to drive a protective process such as release of an inhibitor, buffer, or passivator [3]. The scope of this chapter (corrosion protective smart materials) includes a critical discussion of recent advances where environmentally r esponsive ma terials ha ve be en co nsidered a nd engineered f or sm art c orrosion p rotection. W e w ill c onsider only corrosion protective coatings and exclude self-lubricating, optical, nontumescent, nonskid, and tamper-resistant coatings, important top ics o f t heir o wn b ut b eyond t he s cope o f t his review. Although related to corrosion protective coatings, corrosion indicating coatings such as t hose described by Zhang a nd Frankel [ 4] a lso w ill not b e c overed i n g reat de tail i n t his chapter.

23.2

Smart Coating Technologies for Corrosion Protection

A number of technologies currently exist for rendering protective coatings smart or responsive to the stimulus provided by 23-1

43722_C023.indd 1

10/3/2008 8:57:44 PM

23-2

Smart Materials

corrosion. G enerally t he s timulus re sults f rom a c hemical o r electrochemical potential, but may also be provided by mechanical, thermal, or optical processes. Biological processes generally fall u nder t he h eading o f c hemically s timulated re sponse. A s stated in the introduction, our focus is on the intelligent response of materials to corrosion processes.

23.2.1

Chromate and Partially Soluble Pigments

In t he e arly 1 950s, a s men tioned i n t he i ntroduction, M ayne showed that paints containing basic lead and zinc pigments released an inhibitor in the form of heavy metal soaps to inhibit corrosion. I ndeed w ater t hat h ad b een i n c ontact w ith suc h paints w as demo nstrated to i nhibit t he c orrosion o f s teel. Chromates for paints include the alkaline earth chromates such as Ba, Ca, and Sr that have variable solubility as summarized by Sinko [5]. H owever, a c ritical c oncentration o f c hromate f or a given p H a nd c hloride ac tivity i s re quired f or i nhibition. Formulation o f t he c hromate-based p igments i nto a c oating requires a matrix that allows their slow dissolution. Formulation also requires an optimum solubility of the pigment since excessive solubility results in rapid dissolution to the detriment of the barrier property of the coating. Too little solubility or too strong binding by the matrix results in an activity of chromate below a critical concentration at defects. Often, Sr chromate exists as the pigment of choice due to its intermediate solubility [6]. In 2000, work at Ohio State University showed that hexavalent chromium adsorbed on the trivalent oxides of conversion coatings o n a luminum a lloys [ 7]. W hile t his ob servation w as not entirely ne w, t he re search c learly demo nstrated t hat t he h exavalent chromate deadsorbed as a function of pH and ionic strength. Elevated pH, as encountered when a pit cathodically polarized the adjacent surface, desorbs the hexavalent chromate species. In addition, the increase in ionic strength also enhanced the release of hexavalent chromium species adsorbed on the trivalent oxide. Clearly both of these chemical phenomena contribute to a sm art or re sponsive b ehavior b ased on e nvironmental f actors. This model behavior for a t raditionally, and exceptionally performing coating shows the importance of a sm art response for protective coatings and has guided the way for recent investigations.

23.2.2 Environmentally Responsive Corrosion Protection In recent years, effort has focused on using environmental conditions to stimulate release of corrosion inhibitors or imposition of a physical response (passivation) to slow corrosion. The se can be divided roughly into the categories of (1) response to electrochemical p otential, ( 2) r esponse to ch emical p otential, ( 3) response to me chanical s tress, a nd (4) re sponse s timulated b y living systems (electrochemical, chemical, mechanical, biological). Recognize t hat m aterials c an re spond to add itional ph ysical phenomena suc h a s ele ctromagnetic r adiation a nd m agnetic elds as a basis for the formation of “smart coatings.” However,

43722_C023.indd 2

since this chapter focuses on corrosion protection, these stimuli will be of secondary importance. 23.2.2.1 Electrochemical Potential 23.2.2.1.1

Importance of Electrochemical Responsive Coatings

Response to electrochemical potential presents the most important stimulus for smart, corrosion protective coatings since corrosion i s a n ele ctrochemical p rocess re quiring a ga lvanic potential b etween a c athode, t ypically t he lo cus o f o xygen o r proton reduction, and an anode, an anodically dissolving metallic surface. A coating that connects electrically to this process to drive a corrosion-protective response represents the ultimate in smart c orrosion p rotection si nce c orrosion i s (1) a n ele ctrical phenomenon and (2) provides sufficient electrical energy making an ele ctrically re sponding c oating ide al f or sm art c orrosion inhibition. The need for a nticorrosion coatings, which a re pinhole and scratch tolerant, coupled with growing environmental concerns involving heavy metals, such as hexavalent chromium, has le d to c oating s trategies em ploying i nherently c onductive polymers (ICPs) as a main focus. To date signicant progress in this d irection h as b een m ade p rimarily w ith I CPs a s t he p latform for directing the signal provided by the conditions for corrosion to activate corrosion protection. 23.2.2.1.2 Smart Release of Inhibitors by ICPs Smart c oating s ystems a re eng ineered to re spond to ele ctrochemical processes responsible for corrosion by providing a selfrepairing system. When corrosion events are sensed, the coating automatically releases anodic and/or cathodic corrosion inhibitors to a lleviate t he c orrosion p rocess. S everal sm art c oating system st rategies employ ICPs such a s polyaniline (PANI) a nd polypyrrole ( PPY) [ 8]. I CPs h ave t hree u nique p roperties t hat enable t heir use a s coating materials for ga lvanically d riven or electrochemical potential stimulated release of corrosion inhibitors. First, they are highly stable and inert to dissolution. Second, they are conducting, and, third, they hold and release ionic species depending on their state of charge as illustrated by the equilibria shown in Figure 23.1. The switching of the ICP redox state is triggered by local electrochemical reactions occurring on the surface of a metal. Through proper design, ICPs may be synthesized to c ontain i nhibitor mole cules o r io ns a s dopa nts t hat release when a corrosion process is sensed. The drug release literature provides considerable information for the potential development of smart protective coatings that release corrosion inhibitors. Much of this effort originated with the w ork p erformed at t he University o f M innesota u nder t he direction o f M iller [ 9] a nd h as c ontinued o ver t he ye ars w ith many other important contributions [10]. 23.2.2.1.3 Background and History of ICPs In 1862, Letheby [11] reported that a blue substance (now known to be PANI base) was produced upon electrolysis of aniline on platinum. Twenty-nine years later, Goppelsroeder [12] discovered that a solution of aniline hydrochloride may be electrochemically oxidized rst to emeraldine and then to aniline black. Othmer states that ani-

10/3/2008 8:57:44 PM

23-3

“Smart” Corrosion Protective Coatings

H N

H N +

H N

Reduction

H N +.

Oxidation

n

Emeraldine salt (green)

H + N H Leuco salt (colorless)

−2H+

H N

2n

−2H+

Reduction

H N

H N

N

H N

N

H N

Oxidation n

Emeraldine base (blue)

2n

Leuco base (colorless)

−2e, −2H+

N

N

N

N n

FIGURE 23.1 Redox and pH equilibria for PANI.

line bottoms were used as pigments in paints. (Note: these materials would most likely be similar to aniline black.) Hans Kuhn (formerly of Milliken) used to start his presentations with the story that crude PANI ( still b ottoms f rom p reparation o f a niline b y re duction o f nitrobenzene in Fe lings a nd HCl) was used to c oat tank cars i n Germany at t he t urn o f t he c entury to p revent c orrosion. Ha ns remembers seeing the green/black painted rail cars passing on the tracks. Although Hans was told that the crude PANI was used, there are no published records on the subject. As will be discussed below, the rich redox and ion exchange chemistry of ICPs, as exemplied by PANI, governs their corrosion protection capability. The rst documented observations of corrosion protection of metals (carbon steel and titanium) using ICPs were reported by Schreiber i n 1 978 [13]. I n S chreiber’s w ork, p olymer c oatings were prepared by exposing the metals to thiophene or aniline in a large volume microwave plasma discharge. The polythiophene and PANI coatings were found to provide long-term protection from solvents a nd vapors. I n t he early 1980s, Mengoli’s g roup [14] began a series of pioneering experiments t hat established de nitively t he appl icability of e lectropolymerized I CPs a s viable c orrosion p rotection c oatings. Si nce t hen, n umerous papers h ave b een p ublished o n t he c orrosion p rotection o f carbon s teel [15], s tainless s teel [16], i ron [17], t itanium [18], copper [19], and aluminum alloys [20] with ICPs such as PANI, PPY, polythiophenes, and others. Two comprehensive review articles have b een p ublished [21]. I n t he e arly 1980s, w hen I CPs w ere rst considered for corrosion protection, investigators concentrated on the mechanism of corrosion protection as an anodic inhibition process. The anodic inhibition (or passivation) interpretation re sulted f rom e arly e xperiments b y D eBerry [ 22], who showed that conducting polymers in the oxidized form

43722_C023.indd 3

anodically p rotect s tainless s teel i n d ilute su lfuric ac id. According to this model, ICPs stabilize the potential of the metal in a passi ve regime, to maintain a protective oxide layer on t he me tal. O xygen re duction o n t he p olymer  lm t hen replenishes the polymer charge consumed by metal dissolution, thereby s tabilizing t he p otential o f t he e xposed me tal i n t he passive re gime a nd m inimizing t he r ate o f me tal d issolution. Equations 23.1 and 23.2 depict this concept: ICP Re duction b y me tal to f orm me tal o xide a nd re duced ICP: (1/n)M + (1/m )ICPm + + ( y/n)H2O
(1/n)M(OH) (yn - y ) o

+ (1/m )ICP + (y/n)H

+

(2

3.1)

+

ICP oxidation by oxygen from air to form oxidized ICP: (m /4)O2 + (m /2)H2O + ICPo
ICPm + + m OH − (2 3.2) Unfortunately, t he o xide s tabilization a rgument a s u sed b y DeBerry cannot be generalized, particularly for steel and aluminum in neutral chloride environments. As a re sult, not u ntil the late 1990s w as t he i mportance o f t he dopa nt a nion f or m aking conducting polymers corrosion protective noticed. K inlen et a l. [23] for example, doped PANI with phosphonate corrosion inhibitors with successful inhibition of steel in acid chloride, an environment that does not lead to passivity in steel. Kendig et al. [24] used the inhibitor release mechanism to explain the inhibition of corrosion o f C u-containing a luminum a lloy in n eutral s odium chloride. Cook and colleagues [25] showed that release of inhibit-

10/3/2008 8:57:45 PM

23-4

ing anions as a re sult of the reduction of ICPs could account for the inhibition of steel in chloride containing media where electrochemical pa ssivation f or m ild s teel gener ally do es not o ccur. A recent review outlines the arguments for both mechanisms [26]. As far as cathodic d isbonding, t he ICP should qu ite reasonably d isplace t he o xygen re duction re action (ORR) a way f rom the polymer/metal interface [27] where it does the most damage as a re sult of the formation of hydroxide and free radical intermediates to t he ORR. However, a report using scanning Kelvin probe analysis [28] showed that this effect cannot readily be sustained, and once the ICP is reduced to a nonconducting state the disbonding proceeds at a faster rate. The report concludes that a conducting polymer interface can actually passivate sufficiently small defects. Rohwerder [29] has also reported that this mechanism only works if extended percolation networks of the conducting polymer are avoided in the coating. In recent years, a rather considerable literature has developed around corrosion protection by ICPs. Table 23.1 provides examples of some of the more recent reports on the subject. The citations in Table 23.1 are not exhaustive, but serve to illustrate the breadth of the rather extensive work. Much of the work remains disappointing since it reports corrosion protection by ICPs usually claiming a passivation mechanism (see Table 23.1) with no data on the performance by a realistic blank, a coating with the same barrier properties as the ICP but neither conductivity nor oxidizing potential. Ba re metal ex posed to t he macroenvironment are often c onsidered in t hese reports as t he blank (Table 23.1). U nfortunately, b are me tal h ardly re presents a re alistic blank w hen c onsidering s pecial p rotection at m icroenvironments of coating defects due to ele ctrical a nd ele ctrochemical properties of the ICP coating. Much of the work cited in Table 23.1 considers an apparent “enoblement” of steel by ICPs as evidence for the “passivation” mechanism (Table 23.1). The enoblement app ears a s a rel atively h igh op en c ircuit p otential of t he coated metal that eventually drops as the coated metal begins to rapidly corrode, explained to be a result of discharge of the electrochemically o xidizing c oating. H owever, t his a rgument becomes le ss c onvincing i f o ne c onsiders re ports d ating f rom the p re-Internet er a t hat s how no noxidizing, no nconducting coatings such as polybutadiene on steel in NaCl to exhibit similar b ehavior [30]. Fu rthermore s everal c omparative s tudies o f Fe or steel coated with materials containing the oxidized emeraldine salt (ES) or emeraldine base (EB) in chloride environments s how t he nonc onducting, nono xidizing E B to e xhibit superior performance [31] (see a lso Table 23.1), a re sult hardly supportive of the passivation model. Readers who delve into the technical l iterature o n t he sub ject o f c orrosion p rotection b y conducting polymers should ask the following questions: 1. Does the report consider a relevant blank? 2. Does the report rely on data other than OCP transients to explain the protection mechanism? While most of the reports in Table 23.1 propose the passivation model, as inspired from the DeBerry work for stainless steel in

43722_C023.indd 4

Smart Materials

a nonchloride environment, a number of reports focus on other aspects o f t he I CPs, suc h a s i nhibitor rele ase, a s ke y to t heir corrosion protection, a nd of prime i mportance to sm art protective coatings. 23.2.2.1.4 Visual Example of Electrochemically Stimulated Release A v isual e xample i llustrating t he sm art c oating ph enomena can be demonstrated by incorporating a dye in an ICP  lm. In this case the dye, phenol red, is incorporated into PPY through electrochemical p olymerization o f py rrole i n t he p resence o f phenol re d, w hich ac ts a s a dopa nt. The re action s cheme i s shown i n F igure 23 .2. W hen t he ph enol re d dop ed P PY is electrochemically reduced, by connecting it to a freely corroding aluminum panel, phenol red is expelled from the PPY as shown by the photographs in Figure 23.3. Clearly, the electrochemical potential of the corroding aluminum (ca. −0.5 V vs. A g/AgCl) f urnishes su fficient d riving f orce to re duce t he PPY and release the dopant (in this case phenol red). Without the driving force, phenol red is not released. In conclusion, an on-demand release system results; release will not occur unless the p olymer i s ga lvanically c oupled to t he c orroding me tal, a situation that will occur when a c oating is scratched to ba re metal. 23.2.2.1.5

Case Studies in the Role of Dopants

23.2.2.1.5.1 PANI on Steel with Phosphonate Dopants One o f t he most promising uses of ICPs for corrosion protection is PANI whose structure and oxidation states are shown in Figure 23.1. PANI in its various oxidation states mediates t he a nodic current between t he passivated metal surface and oxygen reduction on the ICP lm. In an acidic environment, the conductive emeraldine form of PANI is reduced reversibly to the leuco form, while in a basic environment, the EB form is reversibly reduced to the leuco base form. The reactions depicted in Equations 23.1 and 23.2 and Figure 23.1 do not take into account the nature of the dopant ion. In most c ases, t he e arly w ork c ited a bove ut ilized su lfonic ac ids, probably because commercially available conductive PANI salts are doped with monofunctional sulfonic acids such as p-toluene sulfonic ac id. A m ajor d rawback to t he u se o f su lfonic ac iddoped PANIs is that sulfonic acid itself is corrosive to carbon steels w ith p assivation on ly ob served at h igh c oncentrations (30% w/w), which may have contributed to the somewhat contradictory results reported in the literature. In contrast to sulfonates, phosphonates, widely used in applications for corrosion mitigation in aqueous systems, were found to be very effective a s c orrosion i nhibitive dopa nts f or PANI. For example, PANI doped with aminotri(methylenephosphonic acid) (ATMP) w as f ound to b e v ery e ffective for pr otection of carbon steel in salt fog [32]. Scanning reference electrode technique ( SRET) s tudies h ave s hown ( Figure 23 .4) t hat i nitially, PANI-ATMP-coated steel panels exhibit anodic activity in pinholes and cathodic activity on the conductive polymer surface (oxygen reduction). After ca. 92 h exposure to a corrosive environment, the c ontrol, a no nconducting PV B-coated s teel pa nel, s till

10/3/2008 8:57:46 PM

43722_C023.indd 5

AA 2024 AA 2024 Fe Fe

Fe Fe

Fe st

st Zn Fe SS304 St St SS Fe

PANI

PANI

PANI

PANI

PANI

PANI

PANI

PANI

PANI

PANI and iodo PANI

PANI and Poly anisidines

PANI DBSA doped

PANI, Polythiophene

PANI, PPY

PANI, PPY

Substrate

PANI

Numerous reviews

Numerous reviews

Polymer

3.5% NaCl

0.5 M HCl

0.5M HCl

0.1 M NaCl, 0.1 M HCl 3.5% NaCl

3.5% NaCl, 0.1 M HCL

various

NaCl

various

Environment

TABLE 23.1 Summary of ICP Corrosion Literature

EIS, OCP, polarization EIS

EIS

OCP, EIS polarization EIS, OCP, polarization EIS, polarization

Polarization

Polarization

Dissolved [Fe], galvanic couple EIS, galvanic couple, polarization

EIS, raman

SVET

Method

Adhesion promotion but no effect of passivation

50 nA/cm2

Increase corrosion resistance with coating

Inhibition eff. >75% after 48 h Passivation mechanism

Dopant is varied. Compared to epoxy, EB performs best Passivation mechanism

Inhibitor release

PANI is compared to non-conducting org. coating; EB performs better than other forms Fe reduced ES; EB is superior to ES Passivation

Controlled release of inhibitor Extraction of galvanically active Cu from alloy

Mentions inhibitor release

Imporant Observation

Iroh, JO; Gajela, P; Cain, R; Nelson, T; Hall, S Rammelt, U; Nguyen, PT; Plieth, W

Nguyen, PT; Rammelt, U; Plieth, W

Subrahmanya, S; Hung, VH; Holze, R

Bereket, G; Huer, E; Sahin, Y

Bereket, G; Huer, E; Sahin, Y

Wei, Y; Wang, JG; Jia, XR; Yeh, JM; Spellane, P Zhao, YP; Yin, RH; Cao, WM; Yuan, AB

Dominis, AJ; Spinks, GM; Wallace, GG

Cook, A; Gabriel, A; Laycock, N

Spinks, GM (Reprint); Dominis, A; Wallace, GG Lu, WK; Elsenbaumer, RL; Wessling, B

Epstein, AJ; Smalleld, JAO; Guan, H; Fahlman, M Bernard, MC; Deslouis, C; ElMoustad, T; HugotLeGoff, A; Joiret, S; Tribollet, B Mirmohseni, A; Oladegaragoze, A

Spinks, GM; Dominis, AJ; Wallace, GG; Tallman, DE Tallman, DE; Spinks, G; Dominis, A; Wallace, GG Kinlen, PJ; Graham, CR; Ding, Y

Authors

(continued)

J. Electrochem. Soc., 154(2), C67–C73 (2007) Macromolecular Symposia, 187, 929–938 (2002) International SAMPE Symposium and Exhibition (2005) Electrochim. Acta, 48(9): 1257–1262 Apr 20 (2003)

Prog. Org. Coat., 54(1), 63–72, (2005).

Act. Metallur. Sin. (English Letters), 17(6), 849–855 (2004). Appl. Sur. Sci., 252(5), 1233–1244 (2005)

Polymer, 36(23), 4535–4537 (1995)

J. Electrochem. Soc., 151(9), B529–B535 (2004) Prog. Organic Coatings, 48(1): 43–49 Nov 2003

Synth. Met., 71(1–3), 2163–2166 (1995)

Corrosion, 59(1), 2–31 (2003)

Synth. Met., 114(2), 105–108 (2000)

Synth. Met., 102(1–3), 1381–1382 (1999)

J. Solid State Electrochem., 6(2), 85–100 (2002) J. Solid State Electrochem., 6(2), 73–84 (2002) ACS Symp. 22–26, 2004, POLY-239 (2004) Synth. Met., 102(1–3), 1374–6 (1999)

Reference

“Smart” Corrosion Protective Coatings 23-5

10/3/2008 8:57:46 PM

43722_C023.indd 6

2M H2SO4 3.5% NaCl

Fe

st AA 2024

AA7075 Cu St,AA2024

Cu ss304

SS mild Steel SS,st Fe Fe

PANI/PMMA Blends

PANI/PMMA/CSA

PANI/polyanion double strand

PANI-co-POA

PDOT

Poly mercaptobenzimidazole

Poly o-Phenylene diammine/ phosphate

poly(4-vinylpyridine) with hexacyanoferrate poly(m-toluidine)

poly-2,5-dimethoxyaniline

PPY

PPY

NaCl

NaCl

0.5 M HCl

0.4 HCl, 0.1 NaCl

NaCl, dil Harrison’s Soln.

3.5% NaCl

0.01 M H2SO4

3.5% NaCl

AA 2024

PANI/montmorillonite clay, (MMT)/polyimide PANI/Ni hexacyanoferrate dopant

Environment 5% NaCl

st

Substrate

Method

Linear polarization, EIS, and wt. loss

EIS

Polarization, OCP, wt. loss CV, raman, OCP

Potentiodynamic

CV EQCM

EIS, OCP, polarization SVET

OCP, raman, scanning probe

EIS, OCP, polarization RQCM, OCP

Kelvin probe

Summary of ICP Corrosion Literature

PANI/dispersed in paint

Polymer

TABLE 23.1(continued) Imporant Observation

Passivation mechanism

Passivation mechanism

Redox active anion

Same protection as when inhibitor is in soln. Increase in OCP and Increase in exhange current density

Substrate anodic to ICP cathode

Passivation mechanism proposed Passivation mechanism

Passivation enhanced by anion release Suppression of H2 evolution at defects

Displace ORR from interface, observed dopant effect Rp = 10E9 ohm cm2

Authors

Van Schaftinghen, T; Deslouis, C; Hubin, A; Terryn, H Bazzaoui, M; Martins, JI; Costa, SC; Bazzaoui, EA; Reis, TC; Martins, L

D’elia, Luis, F; Reynaldo, L; Orti′Z; Ma’Rquez, ZOP; Ma’Rquez, J; Marti′nez, Y Galkowski, MT; Kulesza, PJ; Miecznikowski, K; Chojak, M; Bala, H Hegazy, HS; Youssef, EAM; Abd El-Ghaffar, MA Yin, P; Kilmartin, PA

Oezyilmaz, A; Tuncay; Colak, N; Sanguen, MK; Erbil, M; Yazici, B Jie, H; Victoria Johnston, G; Tallman Dennis, E; Bierwagen Gordon, P; Wallace Gordon, G Trachli, B; Keddam, M; Takenouti, H; Srhiri, A

Iroh, JO; Kottarath, S; Shah, K; Rajamani, D Kulesza, PJ; Miecznikowski, K; Malik, MA; Galkowski, M; Chojak, M; Caban, K; Wieckowski, A Pereira Da Silva, JE; Córdoba De Torresi, SI; Torresi, RM Jesse, C; Seegmiller, A; Pereira, JE; Da Silva, B; Buttry, DA; Susana, I. Córdoba De Torresi, B; Torresib, RM

Williams, G; Gabriel, A; Cook, A; Mcmurray, HN

Reference

J. Solid State Electrochem., 8(6), 430–434 (2004) Egypt. J. Textile Polym. Sci. Technol., 8 77–88 (2004) Current Appl. Phys., 4(2–4), 141–143 (2004) Electrochim. Acta, 51(8–9), 1695–1703 (2006) Electrochim. Acta, 51(21), 4516–4527 (2006)

J. Electrochem. Soc., 148(4), C297–C300 (2001)

Prog. Org. Coat., 44(1), 17–23 (2002)

J. Electrochem. Soc., 147(10), 3667–3672 (2000)

Synth. Met., 1997, 85(1–3) 1263–1264, (1997) Prog. Org. Coat., 54(4), 353–359 (2005)

J. Electrochem. Soc., 152(2), B45–B53 (2005)

Prog. Org. Coat., 58(1), 33–39

International SAMPE Symposium and Exhibition (2005) Electrochim. Acta, 46(26–27), 4065–4073 Aug 24 (2001)

J. Electrochem. Soc., 153(10), B425–B433 (2006)

23-6 Smart Materials

10/3/2008 8:57:46 PM

43722_C023.indd 7

Fe

Fe Ni-plated Cu SS304 st

PPY

PPY

PPY

PPY

PPY

1Cr18Ni9Ti stainless steel

PPY, SDS anion

Fe

Cu

PPY, PTh

Review

St

PPY, PPY-PDAN

0.3M HCl

Local EIS Raman

EIS, OCP, polarization Polarization

OCP, raman

Varied dopants. EB performs best

Increase in pitting potential

ICP gives 1/200 corrosion rate of bare steel

St

PPY PMo12O403- and PO43-

3.5% NaCl pH 5.3 pH 1.9. 0.1 M K2SO4, pH 4 3.5% NaCl

Fe

PPY Oxide Composite

OCP

Reaction of dopant with interface

CV, XPS

SS

PPY doped with DBSA

EIS

Short-term protection/No long term protection Repels chloride ion

Polyanion doped, cation selective coating Passivation mechanism

ICP not completely discharged before depassivation

Critical experiment on enobling mechanism

3% NaCl

0.1M H2SO4

3.5% NaCl

EIS, OCP, polarization Polarization curves XPS, AAS

Fe disk PPY ring

0.1 M K2SO4, pH 4 3% NaCl

OCP, EIS

NaCl

PPY

PPY

Fe

PPY

Dominis, AJ; Spinks, GM (Reprint); Wallace, GG Nguyen, TD; Nguyen, TA; Pham, MC; Piro, B; Normand, B; Takenouti

Zhang, T; Zeng, CL

Nguyen, TD; Pham, MC; Piro, B; Aubard, J; Takenouti, H; Keddam, M

Prissanaroon, W; Brack, N; Pigram, PJ; Liesegang, J; Cardwell, TJ Garcia, B; Lamzoudi, A; Pillier, F; Nguyen Thi Le, H; Deslouis, C Ohtsuka, T; Iida, M; Ueda, M

Lamprakopoulos, S; Yfantis, DK; Depountis, S; Yfantis, CD; Schmeisser, D; Yfantis, AD Van Schaftinghen, T; Deslouis, C; Hubin, A; Terryn, H Michalik, A; Rohwerder, M

Garcia, MA; Lucio; Smit, Mascha, A

Le, Hnt; Garcia, B; Deslouis, C; Le Xuan, Q Tueken, T; Yazici, B; Erbil, M

Le HNT (REPRINT); Garcia, B; Deslouis, C; Le Xuan, Q Nguyen, Td; Keddam, M; Takenouti, H

J. Electroanal. Chem., 572(2), 225–234 (2004)

Prog. Org. Coat., 48(1), 43–49 (2003)

Electrochim. Acta, 50(24), 4721–4727 (2005)

J. Electrochem. Soc., 151(6), B325–B330 (2004)

J. Electrochem. Soc., 149(12), B560–B566 (2002) J. Solid State Electrochem., 10(9), 714–720 (2006)

Electrochim. Acta 51(8–9), 1695–1703 (2006) ZeitschriĀ fuer Physikalische Chemie (Muenchen, Germany), 219(11), 1547–1559 (2005) Surf. Interf. Anal., 33(8): 653–662 (2002)

WSEAS Trans. Environment Develop, 2(6), 742–746 (2006)

J. Power Sources, 158(1), 397–402, (2006)

J. Appl. Electrochem., 32(1): 105–110 (2002) Prog. Org. Coat., 54(4), 372–376 (2005)

Electrochim. Acta, 46(26–27), 4259–4272 (2001) Electrochem. Solid State Lett., 6(8), B25–B28 (2003)

“Smart” Corrosion Protective Coatings 23-7

10/3/2008 8:57:46 PM

23-8

Smart Materials O

H

SO−3

N +

O

SO−3

HO H

HO

N

N

Oxidative electrochemical polymerization

H

N

N

n

H

H SO−3

HO

O

H

H N

N

N

N

H

H

lon release upon reduction n

+

2

O

SO−3

HO

FIGURE 23.2 Reaction scheme for the release of phenol red by PPY.

No contact

Contact to aluminum

FIGURE 23.3 Photograph of phenol red-doped PPY before and after contact to a reducing aluminum surface.

43722_C023.indd 8

10/3/2008 8:57:46 PM

23-9

Displacement (µm)

19 K

17 K

15 K

13 K

11 K

9.50 K

7.60 K

5.70 K

3.80 K

Displacement (µm) Displacement (µm)

19 K

17 K

15 K

13 K

11 K

9.50 K

7.60 K

5.70 K

3.80 K

19 K

17 K

15 K

13 K

11 K

9.50 K

7.60 K

5.70 K

3.80 K

500 µ

300 µ

200 µ

100 µ

0

−100 µ

−200 µ

−300 µ

−400 µ

−500 µ

7.13 K 5.70 K 4.28 K 2.85 K 1.43 K 0 −1.43 K −2.85 K −4.28 K −5.70 K −7.13 K

Displacement (µm)

(f)

400 µ

Displacement (µm)

1.90 K

0

Displacement (µm)

19 K

17 K

7.13 K 5.70 K 4.28 K 2.85 K 1.43 K 0 −1.43 K −2.85 K −4.28 K −5.70 K −7.13 K

7.13 K 5.70 K 4.28 K 2.85 K 1.43 K 0 −1.43 K −2.85 K −4.28 K −5.70 K −7.13 K

Displacement (µm)

(d)

15 K

13 K

11 K

1.90 K

0

Displacement (µm)

19 K

17 K

15 K

13 K

11 K

9.50 K 9.50 K

7.60 K

5.70 K

3.80 K

1.90 K

1.90 K

0

Displacement (µm)

19 K

17 K

15 K

13 K

11 K

9.50 K

7.60 K 7.60 K

5.70 K

3.80 K

1.90 K

0

7.13 K 5.70 K 4.28 K 2.85 K 1.43 K 0 −1.43 K −2.85 K −4.28 K −5.70 K −7.13 K

7.13 K 5.70 K 4.28 K 2.85 K 1.43 K 0 −1.43 K −2.85 K −4.28 K −5.70 K −7.13 K

Displacement (µm)

(b)

Displacement (µm)

(c)

0

7.13 K 5.70 K 4.28 K 2.85 K 1.43 K 0 −1.43 K −2.85 K −4.28 K −5.70 K −7.13 K

Displacement (µm)

(a)

(e)

5.70 K

3.80 K

1.90 K

0

“Smart” Corrosion Protective Coatings

(V)

FIGURE 23.4 SRET scans over PVB coatings on carbon steel in tap water with three 1 mm diameter pinholes equidistant at 1 cm: (a) PVB rst scan, (b) PVB after 92 h, (c) ATMP/PVB after 68 h, (e) PANI-ATMP/PVB,  rst scan, and (f) PANI-ATMP/PVB after 140 h.

43722_C023.indd 9

10/3/2008 8:57:47 PM

23-10

Smart Materials

Al → Al3+ + 3e−

FIGURE 2 3.5 Schematic of a n a luminum a lloy s urface c ontaining intermetallic phases that catalyze the ORR.

600 ⫻ 10−6

Current density (A/cm2)

3e− + 3/4 O2 + 3/2 H2O → 3OH−

5% NaCl Air exposed Cu RDE −0.7 V vs Ag/AgCl

500 400

No inhibitor 300

200

100 2,5-Dimercapto-1,3,4-thiadiazole, potassium salt 0

exhibited ga lvanic ac tivity whereas t he PANI-ATMP exhibited no galvanic activity. It is postulated that PANI-ATMP coating shifts t he potential of t he steel i n a p ositive or noble d irection, creating a c ombination o f f errous a nd f erric io ns. The se ions form a pa ssivating i ron/dopant c omplex at t he c oating/steel interface. For a scribed coating, Fe2 + is proposed to react initially with the ATMP dopant anion (supplied by the leuco-ES equilibrium in Figure 23.1 in an anodic regime). The i nsoluble i ronATMP salt formed effectively pa ssivates t he e xposed me tal surface. 23.2.2.1.5.2 PANI on High-Strength Aluminum Containing ORR Inhibitors as Dopants Alloy 2024-T3 is an inhomogeneous aluminum/copper a lloy containing intermetallic particles t hat are enriched in alloying elements compared to the homogeneous alloy matrix. In contrast to pure aluminum, which is resistant to pitting c orrosion, t he i ncreased su sceptibility o f t he 2 024-T3 aluminum alloy to pitting corrosion is centered upon the unique electrochemistry of t he intermetallic inclusions, specically on the ORR as illustrated in Figure 23.5. Chromate c oatings i nhibit t hrough a sm art rele ase o f h exavalent chromate when called upon by the presence of a corrosive environment [33]. Chromate primarily inhibits the ORR at coating defects by a n i rreversible electrosorption at si tes ot herwise catalytic to oxygen reduction [34]. Based on these results, a class of organic compounds has been found to act as inhibitors for the ORR to an extent comparable to that for chromate [35]. Application o f t hese c ompounds i n re al c oatings re quires a mechanism for stabilizing them in the coating for release when needed i n a f ashion a nalogous to c hromate. F ortunately t he chemistry available from conducting organic polymers provides a mechanism for performing this function. The u ltimate go al i s to i ncorporate o rganic O RR i nhibitors into P ANI c oating s ystems t hat a re rele ased up on dem and. 2,5-Dimercapto-1,3,4-thiadiazole (D McT) a nd 1 -pyrrolidine dithiocarboic acid [36] derivatives are well-known ORR inhibitors for copper and its alloys. Figure 23.6 shows the ORR currents at a Cu RDE as a function of rotation rate in the presence and absence of DMcT. DMcT virtually eliminates the ORR current.

43722_C023.indd 10

0

10

20

30 (rpm)½

40

50

60

FIGURE 23.6 Oxygen re duction c urrent at a Cu rot ating d isk electrode as a f unction of rot ation rate in a s odium chloride electrolyte in the presence and absence of t he di-potassium salt of DMcT. As can be seen, the addition of a low level of DMcT causes a substantial inhibition of this diffusion-limited reaction.

DMcT a nd poly(DMcT) (the polydisulde formed by oxidative polymerization of DMcT) have been extensively studied in combination with PANI as reversible redox electrodes for lithium ion batteries. PANI  lms c ontaining t he ba ses o f 1 -pyrrolidine d ithiocarboic acid have enabled the release of an ORR inhibitor apparently in response to the galvanic inuence of a coating defect. A PANI-coated Al 2024-T3 doped in a c ircular region with the 1-pyrrolidine dithiocarboic anion before exposing to the ASTM B117 salt fog environment for 48 h demonstrated corrosion inhibition around a scribe in the doped region (Figure 23.7) [37]. The SVET scan indicates a n en hanced cathodic current for t he scribe through the undoped region. To su mmarize, c onducting p olymers c an p rovide p rotection to aluminum alloys in chloride environment at breaks in the c oating b y a ga lvanic me chanism sub stantially d ifferent from anodic protection. Protection of base metals such as aluminum in chloride environments by conducting polymer coatings entails a mechanism that is similar to that provided by chromate. Rather than requiring the conducting polymer film to bias exposed substrates to a passive oxidized state, our hypothesis c onsiders t hat e xposed de fects b ias t he c onducting p olymer, t hereby c ausing it to ( a) re duce to t he nonconducting f orm a nd ( b) rele ase i nhibiting s pecies t hat s top corrosion in the defect. 23.2.2.1.5.3 Other Polymer Dopant Combinations Many o ther conducting polymer/dopant combinations have been considered and appear in Table 23.1. A typical analysis used in many investigations entails following the open circuit potential of the coated metal as a f unction of time. The ICP maintains the surface at a

10/3/2008 8:57:49 PM

23-11

“Smart” Corrosion Protective Coatings

As red’D dopant

L-Pyrrolodine dithiocarbothioic acid

(a) Approximate boundary of doped region

Cathodic current

1000 0 −1000

0

CD

x-micron

−500 0

20

y-micron 500

0 Scribe

CD

20

Doped

(b) Approximate boundary of doped region Scanning vibrating probe results

FIGURE 23.7 (a) Scribed salt fog exposure test for a portion of a PANI coating on AA 2024-T3 that was doped in a local region with 1-pyrrolidine dithiocarboic acid anion. As can be seen the anion released at the region of the scribe inhibits corrosion of the scribe. (b) SVET data for a similar specimen.

high potential until the metal depassivates with the reduction of the potential closer to t he reversible potential of t he metal dissolution. Bare metal exhibits a more rapid drop in potential. 23.2.2.1.6

Redox Polymers

PANI-DMcT i s e asily p repared b y o xidative p olymerization o f aniline and DMcT in water using ammonium peroxydisulfate as the o xidizing a gent. Si ze e xclusion c hromatograms o f P ANIDMcT show that NMP-soluble fractions of PANI-DMcT are heterogeneous p olymer blen ds o f P ANI a nd p oly(DMcT), w here poly(DMcT) i s t he p roduct o f t he o xidative p olymerization o f DMcT (Figure 23.8a) [38]. The c yclic volt ammogram i n Figure

43722_C023.indd 11

23.8b provides evidence that poly(DMcT) is reversibly reduced, expelling an anion of DMcT in the process [39]. Thu s, upon electrochemical re duction, b oth PANI-DMct a nd p oly(DMcT) a re expected to release DMcT anions, an excellent ORR inhibitor as shown by the data in Figure 23.6. SVET for an AA 20224-T3 surface coated with the PANI– PolyDMcT composite with pinholes appears in Figure 23.9.The three pinholes i nitially e xhibit c athodic ac tivity but w ith no c orrosion. After 1 h exposure to aerated tap water, the potential of the pinholes indicates the cathode reaction has been inhibited as expected. After 42 h exposure, however, anodic activity is observed to occur. Thus although t he O RR i n t he p in h oles h as b een el iminated, a nodic

10/3/2008 8:57:49 PM

23-12

Smart Materials N N NH2

+ S

HS

SH

H2O, (NH4)2S2O8

H

H

H

H

N

N + −S

N

N + −S

N

S

n N

S

N

N HS

HS + N N −S

S

N N

N N S

S

S

S

S

S

N N S S

(a)

S

S− m

20.00 µA 10 mM DMcT, 0.10 M NaCl, H2O Depolymerization

Current

10.00 µA 1 eq OH− 0.000 A

Polymerization

2 eq OH− 1.000 V (b)

FIGURE 23.8

500.0 mV

−500.0 V

−10.00 µA

E (mV vs. Ag/Agcl)

(a) Oxidation of DMcT to PolyDMcT and (b) cyclic voltammogram for the oxidation and reduction of DMcT to PolyDMcT.

processes in the pin hole are not completely eliminated, as might be expected si nce DMcT i s a n ORR i nhibitor on c opper. A p otential approach to el iminate t his problem would b e to i ncorporate b oth anodic and ORR inhibitor as dopants into the ICP system. More recently, the reduction of poly(dimercapto thiadiazole) (PolyDMcT) to 2 ,5-dimercapto t hiadiazole contained in a c onducting organic matrix provided a direct demonstration of galvanic rele ase o f a n i nhibitor, P olyDMcT, h eld w ithin a c arbon paste as schematically shown in the inset of Figure 23.10 [40]. A Cu R DE c athode b iased at − 0.7 V v ersus A g/AgCl a nd p laced above t he c arbon pa ste ele ctrode de tected t he rele ase o f t he ORR-inhibiting mo nomer a s a de crease i n t he c athodic O RR current. Upon coupling the carbon paste to a freshly polished Al surface, t he rele ased mo nomer o f p olyDMcT w as ob served to lower the ORR current at the Cu RDE as shown in Figure 23.10. This observation demonstrates the feasibility of an organic coating formed from the electroactive polymer to release the inhibitor upon galvanic coupling to a defect.

43722_C023.indd 12

0.000 V

23.2.2.2 Chemical Potential For a n umber of years, ion exchange or host/guest compounds have s hown p otential a s i nhibiting p igments f or pa ints. F or example, t he c ompound c ommercially k nown a s Sh ieldex (Grace-Davison and Zin et al.) can release calcium ions as corrosion inhibitors as a result of the mass action of a high ionic strength, high corrosive ion concentration. Often these pigments have been used synergistically with other inhibitors for the corrosion i nhibition o f ga lvanized s teel [41]. I n ot her w ords suc h systems release corrosion i nhibiting species a s a re sult of mass action or a chemical potential of an ion. More re cently, M cMurray a nd Wi lliams [ 42] a nd m any others [43] have considered the ion exchange properties of layered minerals such as hydrotalcite as sources of anionexchangable corrosion inhi bitors in cluding c hromate. B uchheit and M ahajanam u sed t hese s ame m aterials a s a n io nicstrength re sponding re servoir for v anadate i nhibitors [ 44],

10/3/2008 8:57:54 PM

23-13

“Smart” Corrosion Protective Coatings 7.50 K 6.00 K 4.50 K

1.50 K 0 −1.50 K

Displacement (µm)

3.00 K

−3.00 K −4.50 K −6.00 K

0

2.00 K 4.00 K 6.00 K 8.00 K

10 K

12 K

14 K

16 K

18 K

−7.50 K 20 K

20 µ

25 µ

Displacement (µm)

(a) 25 µ −20 µ −15 µ −10.00 µ −5.00 µ 0

5.00 µ 10.00 µ

15 µ

V

7.50 K 6.00 K 4.50 K

1.50 K 0 −1.50 K

Displacement (µm)

3.00 K

−3.00 K −4.50 K −6.00 K

0

2.00 K 4.00 K 6.00 K 8.00 K 10 K

12 K

14 K

16 K

18 K

−7.50 K 20 K

20 µ

25 µ

Displacement (µm)

(b) 25 µ

−20 µ −15 µ −10.00 µ −5.00 µ 0

5.00 µ 10.00 µ

15 µ

V

FIGURE 23.9 (a) First SVET scan of UV-cured PANI-DMcT/poly(DMcT) formulation. Circles depict pin holes (anodic areas are positive volts, cathodic negative). (b) Final SVET scan of UV cured PANI-DMcT/poly(DMcT) formulation after 48 h of exposure.

and a number of research reports have described the use of the hydrotalcites to p rovide c ontrolled rele ase o f o rganic a nion that a re qu ite e ffective f or s lowing t he O RR o n Cu su rfaces [45]. Similarly, ion exchange minerals can also hold exchangeable c ations a s i n t he c ase o f Sh ieldex f or C a 2+, b ut a lso t he effective rare-earth cations [46]. pH represents an important chemical potential to activate the release o f c orrosion i nhibitors f or a n umber o f re asons. M any

43722_C023.indd 13

structural materials are amphoteric. They corrode or their passivating oxides become unstable with respect to dissolution under conditions of either high pH or low pH. As a result, local variations of pH resulting from corrosion can accelerate corrosion. For example, the alkalinity developed around noble i ntermetallics on Cu-containing Al alloys have been considered to promote depassivation of the alloy in the vicinity of the intermetallic. L ocalized c revice o r p itting c orrosion i n s tainless s teels

10/3/2008 8:57:54 PM

23-14

Smart Materials

have not ye t been applied to c orrosion-protective coatings, but the pH responsiveness appears promising [49].

ORR current density (A/cm2)

−10 ⫻ 10−3

23.2.2.3

−0.8

−0.6 Cu RDE Ag/AgCl ref. Pt counter

Al −0.4

C paste with active inhibitor

−0.2 Galvanic couple 0.0

0

2

4

6 8 Time (s)

10

12 14 ⫻ 103

FIGURE 2 3.10 Demonstration of t he ga lvanic re lease of t he O RR inhibitor, D McT f rom c arbon p aste e lectrode ( inset) c ontaining PolyDMcT.

involves the lowering of the pH in occluded regions as a result of the h ydrolysis o f a nodic p roducts to en hance t he a nodic dissolution. M
Mn + + ne − (2 +

n M + + H 2O
M(OH)(n −1) + H + (2

3.3) 3.4)

A second consideration regarding the importance of pH recognizes t hat t he o nset o f lo calized c orrosion w here t he a node i s separated from the cathode, results in localized changes in pH. In the case of coated steel, the cathodic reduction of oxygen: H2O + 1/2 O2 + 2e −
2OH − (2

3.5)

at the coating metal interface around a defect tends to raise the pH at the interface, leading to propagation of coating failure. In such c ases, t he rele ase o f a c orrosion i nhibitor i n re sponse to increase in p H w ould b e b enecial. Re search i s o ngoing f or devising me thods f or rele ase o f c orrosion i nhibitors a s s timulated by the increase or decrease in pH. For example, Li and Calle have described an approach for microcapsules containing corrosion inhibitors that degrade to release a corrosion inhibitor as a result of alkaline hydrolysis. So far they have demonstrated nominally 10 µm microcapsules that breakdown under alkaline conditions to release pH indicators with no adverse inuence on other paint properties [47]. Cook et al. reported early results on carboxylate-linked inhibitors on Boehemite. Alkaline hydroysis of the carboxylate linkage above pH 9 w ill release the inhibitor [48]. p H-controlled p orosity o f o rganic/inorganic c omposites

43722_C023.indd 14

Mechanical

Smart coatings that respond to mechanical stress resulting from abrasion, impact, and cyclic fatigue have particular engineering importance. The main contribution in this area has considered inhibitor-containing microspheres as potential or actual paint pigments. Inspired by the invention by Silver in 1970 of acrylatecopolymer m icrospheres t hat c ontain ad hesive [50]. B ailin a nd Agarwala [51] developed microspheres to contain dichromate, nitrite, borate, and molybdate inhibitors as a corrosion-inhibiting pigment f or pa int. U .S. A rmy re search h as a lso demo nstrated corrosion inhibition by stress release from microcapsules [52]. A recent re view men tions t he p ossibility o f en vironmentally responsive (in t he mechanical sense) polymers for smart corrosion inhibiting coatings [53]. A group at the University of Illinois has taken this generic approach several steps further with a coating t hat releases a t wo-component polymer-forming agent. The stress re sponsive re lease c omes not f rom m icrocapsules, but rather from a 3D biomimetic microvascular system developed in the coating using a d irect-write assembly method [54]. The vascular approach has an important advantage over t he microcapsule-bourne h ealing a gents. I t ena bles re peated h ealing o f a damaged region drawing material from the entire coating via the vascular system and enables use of a two-component system. 23.2.2.4

Bioproduced Inhibitors and Protective Films

Often microbial growth accelerates corrosion and general environmental de gradation o f ma terials. H owever, N agiub a nd Mansfeld found that Shewanella algae and Shewanella ana prevented pitting of Al 2024, tarnishing of brass and rusting of mild steel i n a rticial seawater [55] and others report that Bacillus subtilus prevents pitting of Al 2024 (2002) [56]. A living material capable of inhibiting corrosion could potentially be placed in a paint or primer matrix in the form of seeds or spores only to be awakened to growth in response to adverse conditions. Kus and Mansfeld recently reported that Shewanella oneidensis inhibited a number of metals. Notably Cu-bearing brass and Cu appear to form a p orous but somewhat insulating layer at t he surface for these materials [57]. Microbes bioengineered to p roduce aspartate and polyphosphate inhibitors were indeed shown to inhibit corrosion i n Lu ria B ertani g rowth me dia [58]. A lthough t here have b een no p ractical c oatings f ormed u sing b iotech-based smart inhibitors, this appears to be an important area for future development.

23.3

Important Opportunities

Protective coatings must provide a barrier to corrosive reagents, but should also be able to respond to environmental stresses in a manner to l imit corrosion without depleting their reactivity or without using environmentally hazardous materials. Release of inhibitors in response to d amage to t he coating remains a v ery

10/3/2008 8:57:56 PM

“Smart” Corrosion Protective Coatings

attractive p roperty f or p rotective c oatings. I deally t he sm art response will either not be depleted or can be continuously and inherently renewed. Galvanic stimulation remains a n i mportant, i f not t he most important, sig naling me chanism, g iven t he ele ctrochemical nature of corrosion. Therefore, ele ctroactive a nd c onducting polymers will continue to play a signicant role in the development of smart corrosion protective coatings. Coupling the electrochemical processes responsible to corrosion to the opening of colloidal microspheres embedded in paint represents an important avenue for research. Regarding conjugated inherently conducting polymers, there remains a ne ed to de termine t he role p layed by t heir high oxidization potential. It remains hard to believe that these materials can protect base alloys in environments where they are susceptible to pitting by anodic protection alone. Microbiological ( genetic eng ineered) app roaches rely ing o n the growth of a living  lm have shown early feasibility and represent a promising future approach for certain applications. This area deserves further investigation. To render these coatings smart, however, a mechanism must exist to activate or deactivate the inhibiting process as corrosive conditions demand.

23.4

Summary

In summary, new developments in nanotechnology, biotechnology, biomimetic materials, and functional materials hold great promise for smart corrosion protective  lms. In all cases, the application of corrosion p rotection o f h ighly energe tic su rfaces ( high p otential energy for oxidation) suggests the use of materials that exhibit some type of galvanic response or galvanic sensitivity to the bias provided by the metallic substrate in a corrosive environment. To date, conducting polymers appear to h ave the most promise since they can hold or release anion inhibitors depending on their oxidation state. In general, the ability of the energy released by the corrosion process itself (nearly 2 V coulomb for steel and aluminum in neutral water) is sufficient to drive corrosion-mediating processes. Thei r semiconductivity allows them to direct the galvanic forces that drive corrosion to a process that inhibit corrosion.

References 1.

J.E.O. Mayne, J. Oil Colour Chemists’Assoc., 34, 473, 1951. 2. L. X ia, E. Akiyama, G. F rankel, a nd R . M cCreery, J. Electrochem. Soc., 147, 2556, 2000. 3. M. Kendig, M. Hon, L. Warren, Prog. Org. Coatings, 47, 169, 2002. 4. J. Zhang and G.S. Frankel, Corrosion, 55, 957, 1999. 5. J. Sinko, Prog. Organic Coatings, 42, 267, 2001. 6. J. Sinko, Prog. Organic Coatings, 42, 267, 2001. 7. L. X ia, E. Akiyama, G. F rankel, a nd R .J. M cCreery, J. Electrochem. Soc., 147, 2556 2000. 8. P.J. Kinlen, Y. Ding, and D.C. Si lverman, Corrosion, 58, 490 2002.

43722_C023.indd 15

23-15

9. R.L. B lankespoor a nd L.L. Miller , J. Chem. S oc., Chem. Commun., 90, 1985. 10. Reynolds et al., J. Chem. Soc., Chem Commun., 620, 1987. 11. H. Letheby, J. Chem. Soc., 15, 161, 1862. 12. F. G oppelsroeder, I nternationale E lektrotechniche Ausstellung No. 19, 1047; No. 18, 978, 1891. 13. H.P Schreiber et al., Ind. Eng. Chem. Prod. Res. Dev., 17, 27, 1978. 14. G. Mengoli, M. Munari, P. Bianco, and M. Musiani, J. Appl. Polymer Sci., 26, 4247, 1981. 15. P.J. Kinlen, Y. Ding, and D.C. Silverman, Corrosion, 58, 490, 2002. 16. R. Gasparac and C.R. Martin, J. Electrochem. Soc., 148, B138, 2001. 17. F. Beck, Metalloberaeche, 46, 177 1992; F . B eck, R . Michaelis, F. S chloten, a nd B . Z inger, Electrochim, Acta, 39, 229, 1994. 18. Z. D eng, W.S. S myrl, a nd H.S. White, J. Elec trochem. S oc., 136, 2152, 1989. 19. V. B rusic, M. Angelopulos, a nd T. G raham, J. Elec trochem. Soc., 144, 436, 1997. 20. V.J.Gelling, M.M. Wiest, D .E. T allman, G.P. B ierwagen, a nd G.G. Wallace, Prog. Organic Coatings, 43, 149, 2001; P. McCarthy, W. Li, and S.C. Yang, Polym. Mater. Sci. Eng., 83, 315, 2000. 21. T. P age M cAndrew, Trends P olym. S ci., 5, 7, 1997, G.M. Spinks, A.J. Dominis, and G.G. Wallace, J. Solid State Electrochem., 6(2), 85, 2002; G.M. Spinks, A.J. Dominis, and G.G. Wallace, J. Solid State Electrochem., 6(2), 73–84, 2002. 22. D. DeBerry, J. Electrochem. Soc., 132, 1022, 1985. 23. P. Kinlen, V. Menon V, and Y. Ding, J. Electrochem. Soc., 146, 3690, 1999. 24. M. Kendig, M. Hon, and L. Warren, Prog. Org. Coatings, 47, 169, 2002. 25. A.Cook, A. Gabriel, D. Siew, and N. Laycock, J. Current Appl. Phys., 4,133, 2004; A. C ook, A. Ga briel, a nd N. L aycock, J. Electrochem. Soc., 151(9), B529–B535, 2004. 26. Michalik and Z, Rohwerder. Phys. Chem., 219, 1547, 2005. 27. P. Kinlen, D. Silverman, and C.R. Jefferys, Synthetic Metals, 85, 1327, 1997; T . S haauer, A. J oos, L. Dulo ng, a nd C.D. Eisenbach, Prog. Org. Coating, 33, 20, 1998. 28. G. P aliwoda-Porebska, M. S tratmann, M. Ro hwerder, K. Potje-Kamloth, Y. Lu, A. Z. Pich, and H.-J. Adler, Corrosion Sci., 47, 3216, 2005. 2 9. http://www.mpg.de/bilderBerichteDokumente/dokumentation/jahrbuch/2007/eisenforschung/forschungsSchwerpunkt/index.html (2007). 30. M. Kendig a nd H. L eidheiser, Jr., J. Elec trochem. S oc., 123, 982, 1976. 31. A. Mirmo hseni a nd A. Oladega ragoze, Synthetic M etals, 114, 105, 2000; G. S pinks, A. Do minis, a nd G. Wallace, Corrosion, 59, 2, 2003; A. Dominis, G. Spinks, and G.Wallace, Prog. Organic Coatings, 48, 43, 2003. 32. P.J. Kinlen, Y. Ding, and D.C. Silverman, Corrosion, 58, 490, 2002.

10/3/2008 8:57:57 PM

23-16

33. L. X ia, E. Akiyama, G. F rankel, a nd R .J. M cCreery, J. Electrochem. Soc., 147, 2556, 2000. 34. W.J. Cla rk a nd R .L. M cCreery, J. Elec trochem. S oc., 149, B379, 2002. 35. M. K endig, un published r esults, Ro ckwell S cientic Company LL C Res earch S tudy su pported b y B oeing Independent Research and Development. 36. J. S inko, U .S. P atent Application 20020197468, 2002; M.W. Kendig, et al ., U.S. Patent Application 20040005478, 2004; P .J. K inlen, U .S. P atent Application 20040035498, 2004; M. K endig, M. H on, a nd J. S inko, Electrochem. S oc. Trans., 1, 119, 2006. 37. M. Kendig, M. Hon, and L. Warren, Prog. Org. Coat., 47, 169, 2002. 38. P.J. Kinlen, C.R. Graham, and Y. Ding, Abstracts of Papers, 228th ACS N ational M eeting, P hiladelphia, P A, August 22–26, 2004. 39. J .M. P ope, T . Sa to, E. S hoji, N. O yama, K.C. White, a nd D.A. Buttry, J. Electrochem. Soc., 149, A939, 2002. 40. M. Kendig and P. Kinlen, J. Electrochem. Soc., 154, C195, 2007. 41. M. Zubielewicz, W. Gnot, Prog. Org. Coatings, 49, 358, 2004; I.M. Zin, S.B. Lyon, V. Pokhmurskii, Corrosion Sci., 45, 777, 2003; F. Deorian, I. F elhosi, S. Rossi , L. F edrizzi, a nd P.L. B onora, Macromolecular Symposia, 187, 87–96, September 2002. 42. H.N. McMurray and G. Williams, Corrosion, 60, 219, 2004; H.N. McMurray and G. Williams, Electrochem. Solid State Lett., 6, B9–B11, 2003; S. B ohm, H.N. M cMurray, S. M. P owell, a nd D .A. Worsley, Materials C orrosion, 52, 896, 2001. 43. A.N. K hramov, N.N. Voevodin, V.N. B albyshev, a nd R.A. Mantz, Thin Solid Films, 483,191, 2005; M. Kendig and M. Hon, Corrosion 60, 1024, 2004. 44. R. Buchheit and S. Mahajanam, J. Electrochem. Soc., 502, 290, 2006.

43722_C023.indd 16

Smart Materials

45. G. Williams and H.N. McMurray, J. Electrochem. Solid-State Lett., 7, B13, 2004; M. K endig a nd M. H on, Electrochem. Solid-State Lett., 8, B10, 2005. 46. D. A. Worsley, N. McMurray, and D. Williams, ECS Trans., 1(9), 153, 2006. 47. W. Li a nd L. C alle, P roceedings o f the Fir st I nternational Conference o n S elf H ealing M aterials, 18–20 April 2007, Noordwijk aan Zee, The Netherlands. 48. R. C ook, R .L., C ook, J . Ellio tt, a nd A. M yers, T ri-Service Corrosion Conference 2005, Orlando, FL, November 14–18, 2005. 49. G. Garnweitner, B. Smarsly, R. Assink, W. Ruland, E. Bond, and C.J. Brinker, J. Am. Chem. Soc., 125, 5626, 2003. 50. S.F. Silver, U.S. Patent 3,691,140, 1972. 51. L.J. B ailin a nd V.S. Agarwala, Micr oencapsulated D NBM Quaternary Ammonium C orrosion I nhibitors, U .S. Government Report, ADD443022, Lockheed Missiles And Space Co Inc., Palo Alto, CA, 1989. 52. S. Sa rangapani, A. Kumar, C. Thies, a nd L.D. S tephenson, U.S. Pat, 7, 192, 993. 53. M. Urban, J. M acromol. S ci., P art C: P olym. Rev ., 46, 329, 2006. 54. K. Toohey, N. S ottos, J. Lewis, J. Moore, and S. White, Nat. Mater., June 2007. 55. A. N agiub a nd F . M ansfeld, Electrochim. A cta, 47, 2319, 2002. 56. D. Örnek, A. Jayaraman, T.K. Wood, Z. Sun, C.H. Hsu, and F. Mansfeld, Corros. Sci., 43, 2121, 2001. 57. Corrosion Protection of Different Metals by Shewanella Oneidensis MR -1 Esra K us, K en N ealson, a nd Flo rian Mansfeld M eet. Abstr. Elec trochem. S oc., 601, 1192, 2006. 58. F. M ansfeld, H. H su, D . Ornek, T . Wood, a nd B . S yrett, J. Electrochem. Soc., 1, B130–B138, 2002.

10/3/2008 8:57:57 PM

24 Smart Polymers for Biotechnology and Elastomers 24.1 C onducting Polymers ............................................................................................................24-1 Introduction • C onducting Polymer Synthesis • S ensing • Ac tuating • E nergy Conversion and Storage • Polymer Processing and Device Fabrication • Future Developments

References ...........................................................................................................................................24-7 24.2 Piezoelectricity in Polymers................................................................................................24-10 Introduction • P roperties • D evelopment of New Materials • C oncluding Remarks

References .........................................................................................................................................24-12 24.3 P olymers, Biotechnology, and Medical Applications .....................................................24-13 Introduction • Smart Polymers Used in Biotechnology and Medicine • Applications

24.4 C onclusion ............................................................................................................................ 24-26 References ........................................................................................................................................ 24-26

24.1 Conducting Polymers Gordon G. Wallace and C.O. Too 24.1.1 Introduction Traditional polymers have been engineered to provide materials that are low density, mechanically strong, and environmentally stable. Ā ey have been made to be nonresponsive—in fact, lightweight dumb materials. H N

Intelligent polymers are engineered from the molecular level to re spond to en vironmental s timuli i n a w ay t hat en hances performance. I magine, f or e xample, a p olymer t hat c hanges color in sunlight to stop transmission of light, or that becomes stronger in response to s tress. Consider a p olymer coating that releases a c orrosion inhibitor in response to t he onset of corrosion or even a medical device that releases anti-inammatory drugs in response to the onset of infection. Polymers such as polypyrrole (PPy) (see Equation 24.1) conduct electricity and can be reversibly oxidized and reduced according to the processes described by the following equation.

H N N H

H N N H

N H

n

m

Contracted state in reduced form + 2e− − 2A−

− 2e− + 2A−

(2

Expanded state in oxidised form H N

H N N H

A−

4.1)

H N N H

A−

N H

n

m

[A− represents anions, e-electrons] n determines degree of doping, m determines molecular weight

PPy a dynamic (electroactive) polymer structure.

24-1

43722_C024.indd 1

10/3/2008 8:58:19 PM

24-2

Smart Materials

Unlike traditional polymers they are not electronic insulators; t hey c onduct ele ctricity a nd a lso, u nlike t raditional polymers, they are not i nert, but are dynamic (electroactive) structures. Ā e chemical, biological, electrical, and mechanical properties of t hese c onducting p olymers a re c ritically de pendent o n t he oxidation s tate o f t he s tructure a nd h ence o n t heir w orking environment. If this dynamic behavior can be determined by appropriate mole cular eng ineering, t ruly i ntelligent p olymer structures will emerge. This ability of conducting polymers to respond is being coupled w ith a n a bility to s ense t he en vironment a nd h ence produce truly intelligent systems. A simple example has involved the de velopment o f a c onducting p olymer t rilayer s ystem t hat responds to o xygen le vels b y op ening a v alve to mo nitor t he environmental c onditions ne eded f or f ood pac kaging at a n optimal level during transport [1]. As we develop our ability to molecularly eng ineer t hese s ystems, s ome fa scinating s ystems will emerge. Intelligent polymer systems possess the ability to s ense, process i nformation, a nd ac tuate re sponses. E nergy i s u sually required to i mplement t hese f unctions a nd s o energ y c onversion/storage capabilities are desirable if the system is to be truly autonomous. Ideally these functions would be integrated at t he molecular level. It h as b een c learly demo nstrated o ver t he pa st 2 0 ye ars that inherently c onducting p olymers ( ICPs) a re c apable o f providing all of the above functions and as such they have a critical role to play in the development of intelligent polymer systems. In this chapter, we briey review the properties of ICPs and their a bility to f unction a s s ensors, ac tuators a nd energ y conversion/storage systems. To illustrate the ability of ICPs to provide the range of functions required for intelligent polymer systems, we will draw on examples that utilize PPy or polythiophenes. W hile p olyanilines a re a lso o f i nterest, t hey a re less amenable to use in a wide range of environments due to the need to retain protonation to ensure conductivity. Consequently, most of the examples here will focus on PPy and polythiophenes.

Application of a p ositive potential to a n inert electrode substrate u sually re sults i n de position o f a n i nsoluble c onducting material on the electrode surface. Simple PPy and poly thiophenes are i ntractable a nd not a menable to sub sequent p rocessing and device fabrication. A dopant counterion (A−) is incorporated during electrosynthesis to ba lance t he charge on t he polymer backbone. A w ide range of dopants can be incorporated using this approach. Ā e choice of dopant is critical as it is inextricably linked to the mechanical, ele ctronic, a nd c hemical/biological p roperties o f the p olymer. F or e xample, t he u se o f a romatic su lfonates i s known to give rise to materials with high conductivity [3], the use o f su lfonated p olyethers re sults i n e xcellent me chanical properties [4], and a myriad of chemical molecular recognition units such as metal complexing agents [5] or biological receptors [6] such as antibodies or biological polyelectrolytes have been incorporated. M onomer o xidation a nd sub sequent p olymerization may also be induced chemically. Common chemical oxidants such a s F eCl 3 an d ( NH4)2S2O8 , m ay a lso b e u sed a nd these provide the anion from the oxidant as the dopant (A−). A critical feature of ICPs is that they are amenable to oxidation/ reduction processes that can be initiated at moderate potentials (Equation 24.1). It is this electrochemical behavior that is the basis of their ability to sense and to respond as well as to convert and s tore energ y. Ā e o xidized f orms e xhibit go od ele ctrical conductivity (s = 1 –400 S cm−1), w hile t he re duced forms have very low conductivity (s ∼ 10−8 Scm−1). If the dopant anion (A−) is small and mobile (e.g., Cl−) and the polymer h as a h igh su rface a rea to vol ume r atio, t hen up on reduction, the anion will be efficiently ejected from the polymer. However, if the dopant is large and immobile (e.g., if A− is a polyelectrolyte suc h a s p olystyrene su lfonate), t hen a n ele ctrically induced c ation e xchange p rocess o ccurs, w here t he c ation i s incorporated f rom t he s upporting e lectrolyte s olution a s t he polymer is reduced and then expelled when the polymer is oxidized [7]. Ā e exact nature of this redox process has a d ramatic effect o n t he physical a nd c hemical p roperties o f t he p olymer. Ā ese c hanges a re i mportant i n de termining t he s ensing a nd actuation capabilities of the systems as discussed below.

24.1.3 Sensing 24.1.2

Conducting Polymer Synthesis

Each of these materials may be produced via either chemical or electrochemical oxidation of t he appropriate monomer [2]. For PPy, the electrodeposition process can be described simplistically as in the following equation:

n

Oxidation N H

A−

Nature has de veloped recognition s ystems t hat a re able to d iscriminate o n t he ba sis o f h ighly s pecic molecule–molecule interactions generating a unique signal. Alternatively nature utilizes arrays of less specic sensors to collect information that is deciphered u sing pat tern re cognition p rocesses c arried o ut i n

H N

H N N H

A−

H N N H

A−

N H

n

m (2

4.2)

A− molecular dopant, n determines degree of doping, m determines molecular weight

Synthesis of PPy.

43722_C024.indd 2

10/3/2008 8:58:20 PM

24-3

Smart Polymers for Biotechnology and Elastomers

the brain. Both approaches have also been pursued with ICPbased sensor systems. Chemical sensors specicity can be induced by using molecular recognition components from nature. For example, the ICP may be used as an immobilization platform for enzymes [8–11], antibodies [12–16], oligonucleotides [17–20], or even whole living cells [ 21,22]. Ā e b ioactive c omponent m ay b e i ncorporated during t he p olymerization p rocess. Ā e m ajority o f en zymecontaining ICP s ensors gener ate a sig nal due to t he en zymatic generation o f a n ele ctroactive p roduct (e.g., H 2O2) o r t he c onsumption of an electroactive product (e.g., O 2). Ā e mechanism of signal generation with antibody-containing conducting polymer sensors appears to be associated with the modication of cation movement into and out of the polymer upon oxidation–reduction in the presence of the antigen [23,24]. Ā e selective detection of metal ions has also been achieved by the covalent attachment of molecular recognition moieties to the ICP backbone. Ā e usual approach has been to u se a mo nomer or d imer c ontaining t he app ropriate re cognition g roup [ 25]. Polythiophenes c ontaining c rown-ethers a nd c alixarenes covalently bound to the bithiophene repeat units, which exhibit controllable s electivity toward L i+, Na+ a nd K + ions, h ave b een used [26]. Ā e incorporation of metal complexing agents such as sulfonated 8 -hydroxyquinoline a s dop ants i n PPy prov ides a simple route to metal ion-selective ICPs [27]. ICPs have also been assembled as the active components in microsensing arrays with a view to collecting less specic data subsequently de ciphered u sing pat tern re cognition s oftwa re. Ā e so-called “electronic noses” are based on this principle [28]. A r ange o f c onducting p olymers w ith d iffering molecular selectivity re spond ( by c hanges i n ele ctronic re sistance o f t he polymer) to a c omplex mixture to p roduce a u nique pat tern of responses. Ā is approach has been used to differentiate beers [29], de tect m icroorganisms [30] f or w aste w ater m anagement [31,32] and wine characterization [33]. Ā e array approach has also been developed for amperometric sensing when used in solution. Change in amperometric responses in the presence of different ions are used as the signal transduction method. Ā is has been used to discriminate between simple ions [34,35] and even proteins [36]. Ā e approach used is similar to the “electronic nose” in that none of t he sensing elements is specic, however, each polymer has a different selectivity series, giving rise to a unique pattern of responses for any given protein.

24.1.4

Actuating

ICPs are capable of undergoing transformations at the molecular through to the macromolecular level. Ā e former can be used to build re sponsive mole cular rele ase s ystems, w hile t he l atter forms the basis of articial muscles. A change in oxidation state of the polymer (see Equation 24.1) is accompanied by dramatic changes in chemical and physical properties. At least when the dopant (A−) is small and mobile, reduction decreases the anion exchange capacity of t he polymer a nd increases t he hydrophobicity of the polymer backbone.

43722_C024.indd 3

Using a memb rane s etup ( Figure 2 4.1) t his c hange i n io n exchange capacity can be continually oscillated v ia application of a p ulsed ele ctrical s timulus. P rovided a c oncentration o r electrical  eld gradient is imposed, t he net result is a t riggered and controlled transport of ions across the membrane structure. Ā e t ransport o f si mple c ations (e.g., K +, N a+, C a2+) o r a nions (e.g., Cl−, NO3−, SO42−) [37,38], small organics [39], and even protein molecules [40,41] can be regulated in this way. Ā e selective transport of prot eins i s a chieved by i ncorporating appropr iate complementary b iomolecules i nto t he c onducting p olymer membrane. Upon application of an appropriate potential, selective p rotein b inding o ccurs a nd t his t hen c an b e re versed b y applying a different potential. Ā is can be repetitively induced to pump proteins across the membrane as illustrated in Figure 24.1. Ā is behavior forms the basis of controlled release devices based on conducting polymers. For b iomedical app lications, t he c ontrolled rele ase o f d rugs such a s de xamethasone ( an a nti-inammatory) [ 42] o r th e anticancer d rug  ouracil [ 43] a nd e ven g rowth f actors t hat stimulate mammalian cell growth [44] has been demonstrated. Massoumi a nd E ntezami [ 45,46] ut ilized c onducting p olymer bilayers to demonstrate controlled release of sulfosalicylic acid and 2-ethylhexyl phosphate. Ā e latter is a mo del for a number of phosphate drug systems and has anti-inammatory properties. Ā e i ncorporation a nd rele ase o f ol igonucleotides [ 47] h as b een achieved using polyethylene dioxy thiophene (Pedot). Pernaut and Reynolds [ 48] demo nstrated t he c ontrolled rele ase o f adeno sine triphosphate (ATP) (of interest for use as a cardiovascular therapeutic) from a PPy membrane in response to electrical stimulation or even to changes in the chemical environment (such as pH). Ā e i ncorporation o f ei ther a nions o r c ations d uring re dox cycling re sults i n p olymer s welling. Ā ese incorporation/ expulsion events at the molecular level also result in changes in the o verall vol ume ( dimensions) o f t he p olymer. I t w as t hese volume changes that led Baughman et al. [49] to t he concept of electromechanical ac tuators ( artificial m uscles) ba sed o n conducting polymers. By p roducing si mple l aminated memb rane s tructures (Figure 2 4.2) ele ctromechanical ac tuation c an b e i nduced. When one side o f t his membrane structure acts as t he anode (oxidation), anion incorporation occurs causing swelling, and the other acts as the cathode (reduction), with anion expulsion causing contraction. Ā e net result is movement due to electrical stimulation (Figure 24.2). Ā e molecular dopant incorporated at the time of synthesis will play a critical role in determining the extent of this behavior as well the anions/cations present in the electrolyte solution. Using ions such as TFSI (see below), strains of up to 26% have recently been reported [50].

F

F

O

C

S

F

O

N

O

F

S

C

O

F

F

TFSI

10/3/2008 8:58:20 PM

24-4

Smart Materials

PPy membrane & connections +

N H

n

Biological molecule

Electrical stimulator Off

Feed solution

Receiving solution

PPy membrane & connections +

N H

n

Biological molecule

Electrical stimulator −

Feed solution

Receiving solution

PPy membrane & connections +

N H

n

Biological molecule

Electrical stimulator +

Feed solution

Receiving solution

FIGURE 24.1 Stimulated transport of proteins across ICPs. Schematic of membrane transport experiment.

We have utilized electrolytes containing the TFSI anion to produce extremely fast responding actuators that operate up to 40 Hz [51]. Interestingly, suc h a nions a re a lso a c ommon c omponent o f ionic liquid electrolytes [52]. Appropriate combinations of organic anions a nd c ations p roduce s alts mol ten at ro om tem perature (considered an ionic liquid if the melting temperature is <100 °C). Ā ese s alts h ave lo w v apor p ressure ( nonvolatile) a nd h ave extremely wide electrochemical potential windows. Ā ese benets translate into the operation of conducting polymer-based devices (such a s electromechanical ac tuators) t hat u se t hem [53,54] w ith increased lifetime over multiple actuation cycles observed. In fact, not only improved stability (increasing cycles length from hundreds to hundreds of thousands of cycles) but also increased per-

43722_C024.indd 4

formance is observed. Ā e latter feature arises from the inclusion of the i midazolium c ation d uring re duction, a process t hat app arently preserves the inherent modulus of the materials [55]. Another interesting form of these electrolytes involves their use as solid electrolytes. By per forming polymerization in the ionic liquid, a solid can be produced [56]. A highly stable, easily handled actuator can be constructed using this electrolyte. Various applications of ICP actuators have been demonstrated. For example, conducting polymer  aps have been used as valves to open containers containing active molecules [57]. Ā is provides a route to a novel drug delivery system wherein the drug needs to be highly water soluble as it does not need to be incorporated into the conducting polymer. In addition, using a c oncentric hollow ber a rrangement, w e ha ve c reated a m icropumping s ystem

10/3/2008 8:58:20 PM

24-5

Smart Polymers for Biotechnology and Elastomers

PPy PPy

Polymer in ionic liquid electrolyte within the pores of PVDF

(a)

Flaps open

(b)

Flaps closed

FIGURE 24.2 Cross-section of a PPy double-sided PVDF electromechanical actuator with the pores of PVDF  lled with electrolyte.

(TITAN) based on the ability to produce a peristaltic action using conducting p olymer ac tuators [ 58]. Ā is a lso m ay p rovide a n avenue for c ontrolled d elivery of a ctive mole cules. C onducting polymer ac tuators c an a lso b e c ongured to op erate a xially [59,60] using appropriate ber or polymer strip geometries. Another recent advance involves the use of carbon nanotube additives to i ncrease the strength of conducting polymer  bers [61]. Ā e se bers have been shown to be highly impressive articial muscles capable of operating at stress levels in excess of 100 MPa [62]. All of the above responses (actuators) are initiated by creating an app ropriate ele ctrical p otential t hat c auses t he p olymer to change form. Ā is is usually achieved by imposition of a potential from an external source although coupling to an autonomously triggered sensor is the ultimate goal.

24.1.5

Energy Conversion and Storage

Given t he a utonomous nat ure re quired o f i ntelligent p olymer systems, they must also be capable of converting energy from a natural source, such as su nlight, a nd storing it u ntil required. ICPs and devices containing them are capable of achieving this [63,64]. Upon irradiation of the ICP, electron–hole pairs (excitons) a re gener ated. I f t hese e xcitons a re t ransported to a n interface where they can be split before recombination occurs, then electrical energy is generated. Both Schottky devices [65] and ph otoelectrochemical c ells [ 66–69] h ave b een u sed. Photovoltaic responses were obtained from a variety of ICPs in Schottky devices with quantum efficiencies up to 1% [68,70,71]. However, t he best results have been obtained using interpenetrating blends of functionalized polyphenylenevinylenes (PPVs)

43722_C024.indd 5

[72,73] a nd/or p olythiophenes (P Ā s) [74] with power conversion efficiencies up to 2%. Ā e addition of dyes [75–77] and electron acceptors [76,78,79] to such devices has also increased the overall power conversion e fficiency to around 3%. Ā e attachment of a light-harvesting porphyrin functional group to polyterthiophene h as a lso b een c onsidered to i mprove p ower conversion efficiency [80]. All-polymer batteries where conducting electroactive polymer comprises both the anode and cathode have been developed [81] (Figure 24.3). A specic charge capacity of 22 mA h g −1 at a c ell potential of 0.4 V was achieved. No loss in capacity was observed after 100 cycles. More re cently a n a ll-polymer bat tery ba sed o n der ivatized polythiophenes su pported o n g raphite-coated su pports wa s described [82]. In this instance, polythiophene functioned more effectively in the n-doping region and provided an improved cell discharge voltage of 2.4 V but lower capacity. Recently, we have produced conducting polymer bers for use in exible electrode structures [83,84] (Figure 24.4). Capacities in the order of 10 mA h g−1 were observed. An alternative energy storage approach based on conducting polymers is the use of material that is more capacitive in nature. Conducting polymer composite electrodes have proven to be extremely u seful. F or e xample, P Py h as b een de posited o nto activated c arbon to p roduce sup ercapacitor ele ctrodes w ith a specic c apacitance o f 3 54 F g −1 [ 85]. P oly(3-methylthiophene) grown within a porous PVDF structure has been shown to have a c apacitance o f 6 16 F g −1 [86]. We recently produced an electrode s tructure c ontaining P Py, c arbon na notubes, a nd MnO2, w ith a s pecic c apacitance of 2 80 F/g a nd t his m aterial was stable to electrochemical cycling [87].

10/3/2008 8:58:20 PM

24-6

Smart Materials Anode: pseudo n-dopable polypyrrole/polystyrene sulfonate on graphite fiber substrate

350 µm 1 mm

22 mAh g−1

100 µm 350 µm

Cathode: PPy electropolymerized on graphite fiber substrate

Solution-cast polymer gel electrolyte

FIGURE 24.3 Schematic illustration of an assembled all-polymer battery showing the PPy cathode and the PPy-polystyrenesulfonate (PSS) anode sandwiched with a layer of solution-cast polymer gel electrolyte. Ā e total dimensions of the assembled cell are 1 mm × 2 cm × 10 cm (8 cm of electroactive material). (From Killian, J.G., Coffey, B.M., Gao, F., Poehler, T.O., and Searson, P.C., J. Electrochem. Soc., 143, 936, 1996. With permission.)

24.1.6 Polymer Processing and Device Fabrication Ā e integration of energy conversion and storage within a structure containing spatially a rranged s ensing a nd ac tuating u nits requires the development of fabrication/patterning systems with nanometer to m icrometer re solution. A mongst t he approaches to have emerged in recent years are wet spinning of bers, inkjet

printing, a nd s creen p rinting. Wi th a ll t hree app roaches, i t i s necessary to have the polymer in solution or be able to form stable dispersions containing small particles. Ā ese solutions or dispersions are the  rst necessary steps in wet spinning (Figure 24.5) or inkjet printing (Figure 24.6). Solubility has been induced in PPy by attaching alkyl [88,89] or a lkyl su lfonate [90,91] g roups to p roduce 3 -substituted py rrole monomers prior to polymerization. Ā is results in markedly enhanced s olubility i n o rganic o r aque ous me dia re spectively. Polythiophenes c an a lso b e ren dered ei ther o rganic s olvent soluble [ 92] o r w ater-soluble [ 93] u sing t hese der ivatization approaches. Solubility h as a lso b een i nduced i n P Py b y t he u se o f appropriate dopants (see structure below) [94]. We have recently shown that the wet spinning technique can be used to p roduce continuous lengths of conducting polymer ber down to 150 µm in d iameter, w ith c onductivities i n t he o rder o f 4 97 Scm−1 and tensile strength 170 MPa [95].

H N

H N

H N

N H

N H

−O S 3 O O

43722_C024.indd 6

n m

−O S 3

O

FIGURE 24.4 Photograph of knotted ber battery.

N H

O

O

O O O

Schematic of interaction between Na+ DEHS− and oxidised PPy

10/3/2008 8:58:21 PM

24-7

Smart Polymers for Biotechnology and Elastomers

X

Spinning solution reservoir

Spool spool

Spinneret

Wash bath

Take-up reel

Coagulation bath

Series of rollers

Nitrogen

FIGURE 24.5 Schematic showing wet spinning process.

Cartridge Ink

Printed features

Plattern

FIGURE 24.6

Jets

Substrate

Vacuum

Schematic showing inkjet printing setup.

Due to the ability to print at higher resolution, inkjet printing has emerged as an attractive patterning technique for conjugated polymers in areas such as light-emitting diodes [96]. Patterning of organic molecules by inkjet printing can be achieved either by printing a s olution o f t he mole cule ( in o rganic s olvent) o r b y printing dispersions of the molecule in water or organic solvents. Sirringhaus e t a l. [ 97] de veloped a me thod to p rint h ighresolution all-polymer transistor circuits utilizing hydrophobic surface patterning, which connes the spreading of water-based conducting p olymer i nk d roplets up on i nkjet p rinting. O ther workers h ave p rinted P Py, P edot, a nd p oly[2-methoxy-5-(2ethylhexyloxy)-1,4-(1-cyano-vinylene) phenylene] combinations to produce diodes [98]. Preparation of PPy nanodispersions and issues in volved in t he inkj et p rinting o f t hese m aterials h ave been discussed [99]. Recently Yoshioka and Jabbour used a desktop inkjet printer to produce a patterned Pedot anode for use in an organic light-emitting diode [100]. As w ell a s s oluble c onducting p olymer, pro cessable d ispersions containing conducting polymer colloids or nanoparticles have also been produced for printing purposes. Emulsion polymerization i s t he most appropriate me thod for producing submicron ( nano) pa rticles d ue to i ts v ersatility a nd rel ative simplicity. F ormation o f I CP na nodispersions c onsisting o f highly conducting nanoparticles dispersed in either organic or aqueous solution has been achieved using polyaniline [101], PPy [102], and regioregular polythiophene [103]. We have successfully p repared p olyaniline na nodispersions a nd ob tained rheological prop erties t hat m ake t hem a menable to i nkjet printing [104].

43722_C024.indd 7

Screen printing is a simple and environmentally friendly way to produce electronic circuitry and make interconnections [105]. Bao et al. [106] used screen printing to develop organic transistors with high eld-effect m obilities (ca. 0 .015–0.045 cm 2/V s) using regioregular poly(3-hexylthiophene). Garnier et al. [107] demonstrated t he usefulness of printing i n organic  eld effect transistor f abrication, w hile Gu stafsson e t a l. [108] s howed i t was p ossible to p rint p olyaniline to f orm t he gate ele ctrode. While screen printing is a h ighly efficient and relatively simple way to fabricate intricate electronics and circuits, this technique suffers from low resolution. In most instances, screen printing is used to pat tern the parts of the circuits that do not dem and high resolution [109].

24.1.7

Future Developments

Advances in the application of ICPs in sensing, actuating, energy conversion a nd energ y s torage h ave b een re viewed. Fut ure developments could be as follows: (1) In t he i ncorporation o f s ensors w ith ac tuators w here a chemical or biochemical e vent i s de tected, le ading to a n actuation response, for example, in the controlled release of a drug. ( 2) Ā ese combined sensors and actuators require power and this can be supplied by energy conversion devices such as plastic solar cells and/or energy storage devices such as all-polymer batteries. (3) For these future developments to occur, advances in polymer processing and device fabrication will be crucial in order to amalgamate all these components into a single device structure. For many future applications, these new devices would need to be exible, wearable, and even implantable.

References 1. Andrews, M.K., Jansen, M.L., S pinks, G.M., Zho u, D., a nd Wallace, G.G. Sensors and Actuators A 2004, 114, 65. 2. Wallace, G.G., S pinks, G.M., K ane-Maguire, L.A.P ., a nd Teasdale, P.R. Conductive Electroactive Polymers: Intelligent Material S ystems 2nd e dn., CR C P ress, B oca R aton, FL, 2003.

10/3/2008 8:58:23 PM

24-8

3. Price, W.E., Wallace, G.G., a nd Zhao , H. Journal o f Membrane Science 1994, 87, 47. 4. Ding, J., Price, W.E., Ralph, S.F., and Wallace, G.G. Synthetic Metals 2002, 129, 67. 5. Reece, D .A., R alph, S.F ., a nd Wallace, G.G. Journal o f Membrane Science 2005, 249, 9. 6. Kane-Maguire, L.A.P. a nd Wallace, G.G. Advanced M aterials 2002, 14(13–14), 953. 7. Ren, X. and Pickup, P.G. Journal of Physical Chemistry 1993, 97, 5356. 8. Umana, M. a nd Waller, J . Analytical C hemistry 1986, 58, 2979. 9. Foulds, N.C. and Lave, C.R. Analytical Chemistry 1988, 60, 2473. 10. Adeloju, S.B ., B arisci, J .N., a nd Wallace, G.G. Analytica Chimica Acta 1996, 332, 145. 11. M a, M., Qu , L., a nd S hi, G. Journal o f Applied P olymer Science 2005, 98, 2550. 12. Sadik, O.A. and Wallace, G.G. Analytica Chimica Acta 1993, 279, 200. 13. Barnett, D., Sadik, O.A., John, M.J., and Wallace, G.G. Analyst 1994, 119, 1997. 14. Bender, S. a nd Sadik, O .A. Environmental S cience & Technology 1998, 32, 788. 15. Sargent, A. a nd Sadik, O .A. Analytica Ch imica Acta 1998, 376, 125. 16. Sargent, A. a nd Sadik, O .A. Electrochimica A cta 1999, 44, 4667. 17. Bidan, G., B illan, M., Galass o, K., Li vache, T., M athis, G., Reget, A., T orres-Rodriguez, L.M., a nd Vieil, E. Applied Biochemistry (in press) (and references cited therein). 18. Garnier, F., Karri-Youssou, H.K., Srivastava, P., Mandrand, B., and Delair, T. Synthetic Metals 1999, 100, 89. 19. Wang, J., Jiang, M., and Kawde, A.-N. Electroanalysis 2001, 13, 537. 20. Xu, Y., Ye, X ., Yang, L ., H e, P., a nd F ang, Y. Electroanalysis 2006, 18(15), 1471. 21. Campbell, T .E., H odgson, A.J., a nd Wallace, G.G. Electroanalysis 1999, 11(4), 215. 22. Deshpande, M.V . a nd H all, E.A.H. Biosensors an d Bioelectronics 1990, 5, 431. 23. Barisci, J .N., H ughes, D ., Minet t, A.I., a nd Wallace, G.G. Analytica Chimica Acta 1998, 371, 39. 24. Gooding, J ., Wasiowych, C ., B arnett, D ., H ibbert, D .B., Barisci, J.N., and Wallace, G.G. Biosensors and Bioelectronics 2004, 20, 260. 25. Higgins, S.J. Chemical Society Review 1997, 26, 247. 26. Swager, T.M. Acc Chem Re s 1998, 31, 201 (a nd r eferences cited therein). 27. Misoska, V. Chemistr y H onors Ā esis, U niversity o f Wollongong, 1998. 28. Lonergan, M.C., S everin, E.J ., D oleman, B .J., B eaber, S.A., Grubbs, R.H., and Lewis, N.S. Chemistry of Materials 1996, 8, 2298.

43722_C024.indd 8

Smart Materials

29. Shurmer, H.V., Corcoran, P., and Gardner, J.W. Sensors and Actuators B 1991, 4, 29. 30. Namdev, P .K., Alroy, Y., a nd S ingh, V. Biotechnology Progress1998, 14, 75. 31. Stuetz, R .M., Fenner, R .A., Hall, S.T., Stratful, I., and Loke, D. Water Science and Technology 2000, 41, 41. 32. Bourgeois, W., Hogben, P., Pike, A., and Stuetz, R.M. Sensors and Actuators B 2003, 88, 312. 33. Guadarrama, A., Fernandez, J.A., Iniguez, M., Souto, J., and de Saja, J.A., Analytica Chimica Acta 2000, 411, 193. 34. John, R ., On garato, D ., a nd Wallace, G.G. Electroanalysis 1996, 8, 623. 35. Nguyen, T.A., Barisci, J.N., Partridge, A., and Wallace, G.G. Synthetic Metals 2003, 137, 1445. 36. Lu, W., Nguyen, T., and Wallace, G.G. Electroanalysis 1998, 10, 1101. 37. Zhao, H., Price, W.E., Too, C.O., Wallace, G.G., and Zhou, D. Journal of Membrane Science 1996, 119, 199 (and references cited therein). 38. Price, W.E., Wallace, G.G., a nd Zhao , H. Journal o f Electroanalytical Chemistry 1992, 334, 111. 39. Zhao, H., Price, W.E., and Wallace, G.G. Journal of Membrane Science 1995, 100, 239. 40. Zhao, H., Too, C.O., and Wallace, G.G. Journal of International Material Systems and Structures 1997, 8, 1052. 41. Zhou, D ., T oo, C.O ., H odges, A.M., M au, A.W.H., a nd Wallace, G.G. Reactive a nd F unctional P olymers 2000, 45, 217. 42. Wadhwa, R ., L agenaur, C.F ., a nd C ui, X.T . Journal o f Controlled Release 2006, 110, 531. 43. Huang, H., Li u, C., Li u, B ., Chen g, G., a nd D ong, S. Electrochimica Acta 1998, 43, 999. 44. Ā ompson, B .C., M oulton, S.E., Din g, J ., Ric hardson, R ., Cameron, A., O ’Leary, S., Cla rk, G.M., a nd Wallace, G.G. Journal of Controlled Release 2006, 116, 285. 45. Massoumi, B. a nd Entezami, A. European P olymer Journal 2001, 37, 1015. 46. Massoumi, B. and Entezami, A. Polymer International 2000, 51, 555. 47. Piro, B., Pham, M.C., and Ledoan, T. Journal of Biomedical Materials Research 1999, 46, 566. 48. Pernaut, J.M. a nd Re ynolds, J.R. Journal o f Physical Chemistry B 2000, 104, 4080. IPRI/7145/1.8.07 49. Baughman, R .H., S hacklette, L.W ., Els enbaumer, R .L., Plichta, E., a nd B echt, C. in Conjugated P olymeric Materials: O pportunities in Elec tronics, O ptoelectronics and M olecular E lectronics K luwer Academic, D ordrecht, 1990, pp. 599–682. 50. Hara, S., Zama, T., Takashima, W., and Kaneto, K. Journal of Materials Chemistry 2004, 14, 1516. 51. Wu, Y., Alici, G., Spinks, G.M., and Wallace, G.G. Synthetic Metals 2006, 156, 1017. 52. Seddon, K.R ., Stark, A., a nd Torres, M.J. Pure and Applied Chemistry 2000, 72, 2275.

10/3/2008 8:58:24 PM

Smart Polymers for Biotechnology and Elastomers

53. Lu, W., Fadeev, A.G., Qi, B., Smela, E., Mattes, B.R., Ding, J., Spinks, G.M., Mazurkiewicz, J., Zhou, D., MacFarlane, D.R., Forsyth, S.A., Forsyth, M., and Wallace, G.G. Science 2002, 297, 983. 54. Ding, J ., Zho u, D ., S pinks, G., F orsyth, S., F orsyth, M., MacFarlane, D., and Wallace, G. Chemistry of Materials 2003, 15, 2392. 55. Ding, J ., Zho u, D ., S pinks, G., F orsyth, S., F orsyth, M., MacFarlane, D., and Wallace, G. Chemistry of Materials 2003, 15, 2392. 56. Zhou, D., Spinks, G.M., Tiyapiboonchaiya, C., MacFarlane, D.R., Forsyth, M., Sun, J., and Wallace, G.G. Electrochimica Acta 2003, 48, 2355. 57. Xu, H., Wang, C., Wang, C., Z oval, J ., a nd M adou, M. Biosensors and Bioelectronics 2006, 21, 2094. 58. Wu, Y., Zhou, D., Spinks, G.M., Innis, P.C., Megill, W.M., and Wallace, G.G. Smart M aterials a nd S tructures 2005, 14, 1511. 59. Ding, J ., Li u, L., S pinks, G.M., Zho u, D ., G illespie, J ., a nd Wallace, G.G. Synthetic Metals 2003, 138, 391. 60. Yamato, K. and Kaneto, K. Analytica Chimica Acta 2006, 568, 133. 61. Lu, W., Smela, E., Adams, P., Zuccarello, G., and Mattes, B.R. Chemistry of Materials 2004, 16, 1615. 62. Mottaghitalab, V., Spinks, G.M., and Wallace, G.G. Polymer 2006, 47, 4996. 63. Mottaghitalab, V., X i, B ., S pinks, G.M., a nd Wallace, G.G. Synthetic Metals 2006, 156, 796. 64. Dastoor, P ., Offi cer, D .L., T oo, C.O ., a nd Wallace, G.G. Chemical Innovation 2000, 15. 65. Wallace, G.G., T oo, C.O ., Offi cer, D .L., a nd D astoor, P.C. MRS Bulletin 2005, 30, 46. 66. Brabec, C.J., Sariciftci, N.S., and Hummelen, J.C. Advanced Functional Materials 2001, 11, 15. 67. Gazotti, W.A., N ogueira, A.F., G rotto, E.M., Gallazi , M.C., and De Paoli, M.A. Synthetic Metals 2000, 108, 151. 68. Yohannes, T., Solomon, T., and Inganas, O. Synthetic Metals 1996, 82, 215. 69. Cutler, C.A., Burrell, A.K., C ollis, G.E., Offi cer, D.L., Too, C.O., and Wallace, G.G. Synthetic Metals 2001, 123(2), 225. 70. Kunugi, Y., Harima, Y., and Yamashita, K. Journal o f t he Chemical Society. Chemical Communications 1995, 787. 71. Shirakawa, H. a nd Ik eda, S. Kobunshi R onbunshu (in Japanese) 1979, 28, 369. 72. Halls, J.J.M., Walsh, C.A., Greenham, N.C., Marseglia, E.A., Friend, R .H., Moratti, S.C., and Holmes, A.B. Nature 1995, 376, 498. 73. Yu, G. and Heeger, A.J. Journal of Applied Physics 1995, 78, 4510. 74. Granstrom, M., P etritsch, K., Arias, A.C., L ux, A., Andersson, M.R., a nd F riend, R .H. Nature 1998, 395, 257. 75. Armstrong, N.R. Journal o f P orphyrins Ph thalocyanines 2000, 4, 414.

43722_C024.indd 9

24-9

76. Yoshino, K., T ada, K., F ujii, A., C onwell, E.M., a nd Zakhidov, A.A. IEEE T ransactions o n Ele ctron D evices 1997, 44, 1315. 77. Yamaue, T., Kawai, T., Onoda, M., and Yoshino, K. Journal of Applied Physics 1999, 85, 1626. 78. Halls, J .J.M. a nd F riend, R .H. Synthetic M etals 1997, 85, 1307. 79. Yu, G., Gao, J., Hummelen, J.C., Wudl, F., a nd Heeger, A.J. Science 1995, 270, 1789. 80. Chen, J ., B urrell, A.K., C ampbell, W.M., Of ficer, D .L., Too, C.O., and Wallace, G.G. Electrochimica Acta 2004, 49, 329. 81. Killian, J .G., C offey, B .M., Gao , F ., P oehler, T .O., a nd Searson, P.C. Journal o f t he Ele ctrochemical S ociety 1996, 143, 936. 82. Gofer, Y., Sa rker, H., K illian, J.G., G iaccai, J., Poehler, T.O., and Searson, P.C. Biomedical Institute and Technology 1998, 32, 33. 83. Wang, J ., T oo, C.O ., a nd Wallace, G.G. Journal o f P ower Sources 2005, 150, 223. 84. Wang, C.Y., Mottaghitalab, V., Too, C.O., Spinks, G.M., and Wallace, G.G. Journal of Power Sources 2007, 163, 1105. 85. Muthulakshmi, B ., K alpana, D ., P itchumani, S., a nd Renganathan, N.G. Journal o f P ower S ources 2006, 158, 1533. 86. Fonseca, C.P., Benedetti, J.E., and Neves, S. Journal of Power Sources 2006, 158, 789. 87. Sivakkumar, S.R., Ko, J.M., Kim, D.Y., Kim, B.C., and Wallace, G.G. Electrochimica Acta (in press). 88. Masuda, H. and Kaeriyama, K. Journal of Materials Science 1991, 26, 5637. 89. Ashraf, S.A., Chen, F., Too, C.O., and Wallace, G.G. Polymer 1996, 37, 2811. 90. Patil, A.O., Ikenoue, Y., Wudl, F., and Heeger, A.J. Journal of the American Chem ical S ociety 1987, 109, 1858. IPRI/7145/1.8.07 91. Havinga, E.E., ten Hoeve, W., Meijer, E.W., and Wynberg, H. Chemistry of Materials 1989, 1, 650. 92. Bryce, M.R ., Chiss el, A., K athirgamanathan, P., P arker, D ., and Smith, R.M. Journal of the Chemical Society. Chemical Communications 1987, 466. 93. Barisci, J .N., M ansouri, J ., S pinks, G.M., Wallace, G.G., Kim, C.Y., K im, D .Y., a nd K im, J .Y. Colloids a nd S urfaces 1997, 126, 129. 94. Foroughi, J., Spinks, G.M., Whitten, P.G., and Wallace, G.G. Synthetic Metals (submitted for publication). 95. Spinks, G.M., M ottaghitalab, V., B ahrami-Samani, M., Whitten, P.G., and Wallace, G.G. Advanced Materials 2006, 18, 637. 96. Hebner, T.R., Wu, C.C., Marcy, D., Lu, M.H., and Sturm, J.C. Applied Physics Letters 1998, 72, 519. 97. Sirringhaus, H., K awase, T ., F riend, R .H., S himoda, T ., Inbasekaran, M., Wu, W., and Woo, E.P. Science 2000, 290, 2123.

10/3/2008 8:58:24 PM

24-10

Smart Materials

98. Liu, Y. and Cui, T. Macromolecular Rapid C ommunications 2005, 26, 289. 99. Winther-Jensen, B., Clark, N., Subramanian, P., Helmer, R., Ashraf, S.,Wallace, G., Spiccia, L., and MacFarlane, D. Journal of Applied Polymer Science 2007, 104, 3938. 100. Yoshioka, Y. a nd Jabbour, G.E. Synthetic M etals 2006, 156, 779. 101. Ghosh, P., Siddhanta, S.K., Haque, S.R., and Chakrabarti, A. Synthetic Metals 2001, 123, 83. 102. Omastova, M., T rchova, M., K ovarova, J ., a nd S tejskal, J . Synthetic Metals 2003, 138, 447. 103. B rustolin, F ., G oldoni, F ., M eijer, E.W., a nd S ommerdijk, N.A.J.M. Macromolecules 2002, 35, 1054. 104. Ngamna, O., Morrin, A., Killard, A., Moulton, S., Smyth, M., and Wallace, G.G. Langmuir (in press). 105. Gilleo, K. Polymer Ā ic k Films, Van N ostrand Reinho ld, New York, 1996. 106. Bao, Z. and Lovinger, A.J. Chemistry of Materials 1999, 11, 2607. 107. Garnier, F., Hajlaoui, R., Yassar, A., and Srivastava, P. Science 1994, 265, 1684. 108. Gustafsson, G., Cao, Y., Treacy, G.M., Klavetter, F., Colaneri, N., and Heeger, A.J., Nature 1992, 357, 477. 109. Rogers, J.A. and Bao, Z. Journal of Polymer Science: Part A: Polymer Chemistry 2002, 40, 3327. IPRI/7145/1.8.07.

24.2

Piezoelectricity in Polymers

Aleksandra M. Vinogradov 24.2.1 Introduction In the past decades, rapid advances in synthetic polymer science have led to an explosive growth in the development of electroactive polymers with capabilities to detect changes in the loading or en vironmental c onditions, de cide r ationally o n a s et o f t he respective actions, and implement such decisions in a controlled manner. Ā ese special qualities of electroactive polymers can be compared wi th b iological fu nctions th at i nvolve t ransformations of the sensed information into a reactive response. Ā e g roup o f ele ctroactive p olymers i ncludes p iezoelectric and electrostrictive polymers, the response of which is observed either in the form of an electric charge or voltage produced by applied mechanical forces or, conversely, in the form of mechanical d eformation i nduced by a n appl ied e lectric  eld. Ā ese electromechanical eff ects have been dened, re spectively, a s “direct” and “converse.” It i s i mportant at t his p oint to d raw a d istinction b etween electrostriction and piezoelectricity. Essentially, electrostriction, which is a property of all dielectrics, is characterized by a quadratic relation between mechanical deformations and the applied electric  eld, whereas in piezoelectric materials, this relation is linear. A s a re sult, p iezoelectric e ffects e ntail a r eversal o f mechanical deformations in response to reversed electric elds.

43722_C024.indd 10

Ā ere a re m any p olymers k nown to p ossess p iezoelectric properties including, in particular, polypropylene, polystyrene, poly(methyl me thacrylate), v inyl ac etate, a nd o dd n umber nylons. However, strong piezoelectric effects have been observed only i n p olyvinylidene  uoride ( PVDF o r PV F2) a nd PV DF copolymers. At present, t hese polymers comprise t he principal commercially produced group that plays a le ading role i n practical applications.

24.2.2

Properties

PVDF is a thermoplastic uoropolymer rst patented in 1948 [1]. In the 1960s, PVDF was used primarily as an electric insulator or as a base for durable long-life coatings for exterior nishes [1]. Strong p iezoelectric e ffects in PVDF discovered by Kawai in 1969 [2] have attracted attention to t he functional capabilities of t he p olymer s timulating qu alitatively ne w te chnological applications [3–6]. PVDF is a s emicrystalline polymer with typical crystallinity of approximately 50% and a molecular structure consisting of a repeated monomer unit (–CF2–CH2–)n. Ā e amorphous phase of the p olymer h as t he p roperties o f a sup ercooled l iquid w ith the g lass t ransition temperature of about −50°C a nd melting temperature i n t he r ange of 155–180°C. Ā e p olymer do es not absorb v isible l ight a nd i s t ransparent, a lthough i ts c larity i s often c ompromised b y l ight s cattering i n t hicker s ections a s a result of crystallinity. Permanent dipole polarization of PVDF is obtained through a technological p rocess t hat i nvolves s tretching a nd p olling o f extruded thin sheets of the polymer. Typically, PVDF is produced in th e f orm o f th in  lms w ith t hicknesses r anging f rom 9 to 800 µm (10−6 m). A thin layer of nickel, silver, or copper is deposited on both material surfaces to provide electrical conductivity when electric  eld i s applied, or to a llow me asurements of t he charge produced by mechanical deformations. Ā e ele ctromechanical p roperties o f PV DF a re c ommonly represented by the constitutive equations of linear piezoelectricity [7]. Re spectively, t he p iezoelectric re sponse o f t he p olymer i s characterized b y  ve pie zoelectric c oefficients d3i (i = 1 ,2,3), d15, and d24, where the rst subscript denotes the direction of the applied e lectric  eld a nd t he s econd identies t he d irection o f mechanical deformation. Piezoelectric properties of PVDF are also characterized by the hydrostatic coefficient dh = d31 + d32 + d33, which represents the electric charge generated by hydrostatic pressure. It is important to note that the values of all piezoelectric coefficients of PVDF depend on the polarization conditions of the p olymer c haracterized b y p olarization t ime, tem perature, and polarizing eld [8]. Ā e constitutive equations of linear piezoelectricity imply that the mechanical response of PVDF is dened by the generalized Hooke’s law. Ā is formulation involves nine independent elastic constants reecting the anisotropic nature of the polymer. Experiments indicate that the mechanical properties of PVDF strongly de pend o n t he o rientation o f t he mole cular c hains o f the polymer aligned in the stretch direction [4]. Āe stress–strain

10/3/2008 8:58:24 PM

24-11

Smart Polymers for Biotechnology and Elastomers

diagram o f PV DF i n t he d irection o f t he a ligned mole cular chains is characterized by continuous increase of stresses, culminating in sudden failure. Ā is type of behavior is usually exhibited b y b rittle m aterials. I n c ontrast, i n t he m aterial direction no rmal to t he o rientation o f mole cular c hains, t he stress–strain response of the polymer is similar to that of ductile materials. It is characterized by an increase of stresses up to a certain maximum and followed by a sharp decrease in the loadcarrying capacity of the material. Since t he ele ctromechanical re sponse o f PV DF de pends o n such f actors a s p olarization c onditions, s tress/strain r ates, temperature and hydrostatic pressure, the piezoelectric and elastic constants o f t he p olymer c annot b e iden tied w ith abso lute precision. Nevertheless, based on well documented experimental studies [4,5,9] it is possible to iden tify t he t ypical values of t he material characteristics of PVDF as summarized in Table 24.1, which is adopted from Ref. [4]. It i s i mportant to em phasize t hat t he c onstitutive l aw o f linear piezoelectricity t ends to ne glect e nergy d issipation, time-dependent e ffects, a nd v arious m aterial no nlinearities i n the ele ctromechanical re sponse o f p iezoelectric p olymers. I n reality, however, t he electromechanical properties of PVDF are characterized b y st ress-dependent n onlinearities, h igher o rder

electromechanical couplings such a s electrostriction, a nd nonlinear s train–displacement rel ations at l arge d riving vol tages [10–12]. Besides, the polymer tends to demonstrate piezoelectric and mechanical relaxation as well as measurable energy losses under c yclic conditions [3,8,13–15]. Ā ere a re clear i ndications that the same molecular relaxation mechanisms that give rise to mechanical relaxation are also likely to give rise to dielectric relaxation [14,16]. Further, it has been observed that the dynamic response of PVDF under superimposed static and cyclic loading conditions is essentially nonlinear since it does not amount to a superposition of the responses to static and fully reversed cyclic loads app lied s eparately [ 17]. A t p resent, t hese e ffects i n t he behavior o f p iezoelectric p olymers a re not c learly u nderstood and require further investigations.

24.2.3 Development of New Materials To date, piezoelectric polymers have been used as active elements in transducers, sensors, and actuators. Ā e key factors that dene their f unctional p erformance i nclude suc h pa rameters a s t he maximum achievable strain, stiffness, spatial resolution, and the frequency ba ndwidth. I n t he pa st de cades, co nsistent e fforts have been made to en hance t hese characteristics by modifying

TABLE 24.1 Properties of PVDF Property Range of lm thicknesses

Value 9–800 µm (10 m) 1.78 × 103 kg/m3 −6

Mass density, ρ Water absorption Operating temperature range Glass transition temperature, Tg Melting temperature, Tm Ā ermal conductivity Maximum operating voltage

0.02% −40°C to 80°C −60°C to −20°C 170°C to 178°C 0.18 Wm−1 K−1

Breakdown voltage

100 V/µm (2000 V/mil)

Capacitance

380 pF/cm2 for 28 µm lms @1 kHZ d31 = 21.4 × 10−12 C/N; d32 = 2.3 × 10−12 C/N; d33 = −31 × 10−12 C/N; d24 = −35 × 10−12 C/N; d15 = −27 × 10−12 C/N k31 = 12%; k32 = 3%; k33 = 19% @1 kHZ (106–113) × 1012 F/m 12–13 2.1 D Y1 = 2.56 × 109 Pa; Y2 = 2.6 × 109 Pa

Piezoelectric coefficients Electromechanical coupling factors Permitivity Relative permitivity Dipole moment Young’s moduli Poisson ratio

30 V/µm (750 V/mil)

ν21 ∼ 0.1, ν31 ∼ 0.8

Yield stress

(sy)1 = 45 MPa; (sy)2 = 39 MPa

Yield strain

(ey)1 = 1.8%; (ey)2 = 1.4%;

Ultimate stress

(su)1 = 350 MPa; (su)2 = 50 MPa

Ultimate strain

(eu)1 = 16.9%; (eu)2 = 2.5% 0.015–0.25 0.10 PVC, PTFE, PVDF–TrFE

Dielectric loss factor Mechanical loss tangent Related compounds

Source: Adapted f rom Vinogradov, A.M., in S chwartz, M. (E d.), Encyclopedia of Smar t M aterials, Vol. 2, John Wiley & Sons, New York, 2002, pp. 780–792. With permission.

43722_C024.indd 11

10/3/2008 8:58:24 PM

24-12

Smart Materials

the prop erties of pie zoelectric p olymers or c reating nove l polymer-based composite systems with superior properties. In particular, strong electrostrictive response of PVDF– triuoroethylene (PVDF–TrFE) h as b een obtained a s a re sult of material irradiation [5,18]. Large electrostrictive strains in PVDF– TrFE have been also achieved by chemically increasing the degree of polymer cross-linking [19]. Certain changes in manufacturing processes h ave pr ovided s ignicant improvements in the electrostrictive r esponse o f P VDF–hexauoropropylene (PVDF–HFP) copolymer [20]. Similar effects have been observed in PVDF–HFP as a result of special thermal treatment [21]. Ā e use of conductive polymers i n p lace o f t he t raditional me tallic su rface l ayers h as improved the functional performance of PVDF–TrFE over a wide range of temperatures and frequencies [22]. Enhanced performance of el ectroactive t hermoplastic ela stomers ha s b een ach ieved b y subjecting t he m aterials to t hermal a nnealing t reatment. A s a result, a g raft-elastomer (G-elastomer) h as been developed w ith improved functional response in terms of large electromechanical strains and a signicant increase in the electromechanical output power density [23]. To date, various approaches have been developed to produce novel electroactive polymer systems by combining the properties of diff erent m aterials. I n pa rticular, adv anced p iezoelectric composites have been produced by integrating ceramic bers or particles o f le ad z irconate t itanate ( PZT) o r c alcium-modied lead titanate (PbTiO3) into a polymer matrix [24,25]. Electroactive composite s ystems h ave b een a lso s ynthesized u sing h igh dielectric o rganic c ompounds a nd ele ctroactive p olymers. A promising composite material of this type represents a system of CuPc powders used as  ller a nd t he ele ctrostrictive p olymer PVDF–TrFE as the matrix. Similarly, an all-polymer composite material h as b een s ynthesized u sing t he c onductive p olymer polyaniline ( PANI) a s t he d ielectric en hancement c omponent and t he ter polymer PV DF-TrFE-CTFE a s t he m atrix. B oth composite systems are characterized by high elastic modulus, high elastic energy density, and tend to produce large strains [26,27]. A c ombination o f a n ele ctrostrictive g raft elastomer (G-elastomer) described above with a piezoelectric poly(vinylidene uoride–triuoroethylene) c opolymer h as p roduced s everal types of fe rroelectric–electrostrictive mole cular c omposite systems with dual functionality involving two types of responses, piezoelectric and electrostrictive. Such materials can form a basis for various multifunctional smart systems with the advantage of using a single lm for both sensor and actuator functions [28].

24.2.4

Concluding Remarks

To date, t he inherent intelligence of piezoelectric polymers has stimulated rapid progress in many modern technological elds, including a viation, s pace e xploration, ele ctronics, a utomated control, a nd biomedical eng ineering. I n t he i mmediate f uture, continuing progress in this  eld will depend on the intensity of research efforts directed toward the development of electroactive polymers with enhanced functional capabilities. Further, it is important t hat i nnovative te sting me thodologies b e de veloped in order to obtain consistent material characterization of existing

43722_C024.indd 12

and emerging piezoelectric polymer systems. Renewed efforts in computational materials science are required to facilitate accurate material modeling of piezoelectric polymers as an integral part of new developments [29]. It is clear t hat only on t his basis t he unprecedented o pportunities o ffered b y p iezoelectric po lymer systems w ill c ontinue to s timulate f urther te chnological progress.

References 1. Sperati, C.A. Fl uoropolymers, P oly(vinylidene  uoride) (PVDF), in Handbook of Plastic M aterials and Technology, I.I. Rubin, ed., John Wiley & S ons, New York, pp. 131–136, 1990. 2. Kawai, H. Ā e Piezoelectricity of Poly(vinylidene uoride), Jpn. J. Appl. Phys., 8, 975–976, 1969. 3. K epler, R .G. a nd Anderson, R .A. F erroelectric p olymers, Adv. Phys., 41(10), 1–57, 1992. 4. Vinogradov, A.M. P iezoelectricity in p olymers, in Encyclopedia of Smar t M aterials, Vol. 2, M. S chwartz e d., John Wiley & Sons, New York, pp. 780–792, 2002. 5. Zhang, Q.M., B harti, V., a nd K avarnos, G. Poly(vinylidene uoride) (PVD F) a nd i ts co polymers, in Encyclopedia of Smart Materials, Vol. 2, M. Schwartz ed., John Wiley & Sons, New York, pp. 807–825, 2002. 6. Harrison, J .S. a nd Ouna ies, Z. P olymers, p iezoelectric, in Encyclopedia of Smar t M aterials, Vol. 2, M. S chwartz e d., John Wiley & Sons, New York, pp. 860–873, 2002. 7. Ikeda, T. Fundamentals of P iezoelectricity, Ox ford S cience, Oxford, 1990. 8. Hilczer, B. and Malecki, J. Electrets, Elsevier, Amsterdam, 1986. 9. Tasaka, S. a nd Mi yata, S. Ā e Orig in o f p iezoelectricity in poly(vinylidene uoride), Ferroelectrics 32(1), 17–23, 1981. 10. Sessler, G.M. 1982, Polymeric electrets in Electrical Properties of P olymers, D .A. S eanor e d., Academic P ress, N ew York, pp. 241–284, 1982. 11. Tiersten, H.F. Ele ctroelastic e quations f or ele ctroded thin plates subject to large driving voltages, J. Appl. Phys., 74(5), 3389–3393, 1993. 12. McGrum, N.G., Reid, B.E., and Williams, G. Anelastic and Dielectric E ffects i n P olymeric S olids, D over Publica tions, New York, 1991. 13. Hahn, B.R. Studies of the nonlinear piezoelectric response of polyvinylidene uoride, J. Appl. Phys., 57(4), 1294–1298, 1985. 14. Vinogradov, A.M. et al . D amping a nd ele ctromechanical energy loss es in the p iezoelectric p olymer PVD F, Mech. Mater., 36, 1007–1016, 2004. 15. Vinogradov, A.M. a nd F. H olloway, F. Ele ctro-mechanical properties of the piezoelectric polymer PVDF, Ferroelectrics, 226, 169–181, 1999. 16. Lakes, R. Viscoelastic Solids, CRC Press, Boca Raton, FL, 1998. 17. Vinogradov A.M. a nd S chumacher, S.C. C yclic cr eep o f piezoelectric po lymer po lyvinylidene  uoride, AIAA J ., 39(11), 2227–2229, 2001. 18. Zhang, Q .M. a nd S cheinbeim, J .I. Ele ctric EAP in Electroactive Polymer (EAP) Actuators as Articial Muscles.

10/3/2008 8:58:25 PM

24-13

Smart Polymers for Biotechnology and Elastomers

19. 20.

21.

22.

23. 24. 25.

26. 27. 28. 29.

Reality, P otential and C hallenges, Y. B ar-Cohen e d., S PIE Press, Washington, DC, pp. 89–120, 2001. Casalini R. and Roland, C.M. Electromechanical properties of p oly(vinylidene  uoride-triuoroethylene) networks, J. Polymer Sci., 40, 1975–1982, 2002. Jayasuriya, A.C., S chirokauer, A., a nd S cheinbeim, J .I. Crystal-structure dep endence o f ele ctroactive p roperties in differently p repared p oly(vinylidene  uoride/haxauoropropylene) copolymer lms, J. Polymer Sci., Part B, Polymer Phys., 39(22), 2793–2799, 2001. Lu, X., S chirokauer, A., a nd S cheinbeim, J .I. G iant ele ctrostrictive response in poly(vinylidene uoride-hexauoropropylene) c opolymers, IEEE T rans. Ul trasonics Ferroelectrics Frequency Control, 47(6), 1291–1298, 2000. Xu, H.-S. et al . All-polymer ele ctromechanical syst ems consisting o f e lectrostrictive po ly(vinylidene  uoridetryuoroethylene) a nd co nductive p olyaniline, J. Ap pl. Polym. Sci., 75, 945–957, 2000. Su, J., Harrison, J.S., a nd S t. Cla ir, T. Ele ctrostrictive G raft Elastomers, U.S. Patent, 6,515,077, 2000. Chan, H.L.W. and Unsworth, J. Mode coupling in modied lead titanate/polymer 1–3 co mposites, J. Appl. Phys., 65(4), 1754–1758, 1989. Marra, S.P., Ramesh, K.T., and Doudlas, A.S. Ā e mechanical and ele ctromechanical p roperties o f ca lcium-modied lead-titanate/poly(vinylidene  uoride-triuoroethylene) 0–3 composites, Smart Mater. Struct., 8, 57–63,1999. Zhang, Q.M. et al. An all-organic composite actuator material wi th h igh di electric co nstant, Nature, 419, 284–287, 2002. Huang, C., Zhang, Q.M., and Su, J. High dielectric constant all-polymer p ercolative co mposites, Appl. P hys. L ett., 82, 3502–3504, 2003. Su, J., Harrison, J.S., and St. Clair, T. Novel polymeric elastomers f or ac tuation, Proc. IEEE I nt. S ymp. Application o f Ferroelectrics, IEEE, 2000. Smith, R .C. S mart ma terial syst ems. M odel de velopment, Frontiers i n Applied Mathematics, S IAM, Phi ladelphia, PA, 2005.

24.3 Polymers, Biotechnology, and Medical Applications Igor Yu Galaev and B. Mattiasson 24.3.1 Introduction Life is polymeric in its essence. Ā e most important components of l iving c ell, p roteins, c arbohydrates, a nd n ucleic ac ids a re polymers. Even lipids, which have lower molecular weights, can be re garded a s me thylene ol igomers t hat h ave a p olymerization degree around 20. Nature uses polymers as constructive elements a nd pa rts o f c omplicated c ell m achinery. Ā e salient feature o f f unctional b iopolymers i s t heir a ll-or-nothing o r at least h ighly no nlinear re sponse to e xternal s timuli. Sm all

43722_C024.indd 13

changes h appen i n re sponse to v arying pa rameters u ntil t he critical point is reached; then a t ransition occurs in the narrow range of t he varied pa rameter, a nd a fter the transition is completed, t here i s no sig nicant f urther re sponse o f t he s ystem. Such nonlinear response of biopolymers is warranted by highly cooperative interactions. Despite the weakness of each particular interaction in a separate monomer unit, these interactions, when summed t hrough hundreds a nd t housands of monomer u nits, provide s ignicant d riving f orces f or t he p rocesses i n suc h systems. Not surprisingly, understanding the mechanism of cooperative interactions in b iopolymers h as o pened t he  oodgates for attempts to m imic t he c ooperative b ehavior o f b iopolymers i n synthetic s ystems. Re cent de cades w itnessed t he app earance o f synthetic f unctional p olymers, w hich re spond i n s ome de sired way to a change in temperature, pH, electric, or magnetic elds or some other parameters. Ā ese polymers were nicknamed stimuliresponsive. Ā e na me “ smart p olymers” w as c oined d ue to t he similarity of the stimuli-responsive polymers to biopolymers [1]. We have a s trong belief t hat nature has a lways striven for smart solutions in creating life. Ā e goal of scientists is to mimic biological processes, and therefore understand them better, and also to create novel species and invent new processes. Ā e app lications o f sm art p olymer i n b iotechnology a nd medicine a re d iscussed i n t his a rticle. Ā e h ighly no nlinear response o f sm art p olymers to sm all c hanges i n t he e xternal medium is of critical i mportance for t he successful f unctioning of a system. Most applications of smart polymers in biotechnology a nd medicine i nclude biorecognition a nd/or b iocatalysis, which t ake place principally i n aque ous s olutions. Ā us, only water-compatible smart polymers are considered; smart polymers in organic solvents or water/organic solvent mixtures are beyond the scope of the chapter. Ā e systems discussed in the chapter a re ba sed o n ei ther s oluble/insoluble t ransition o f smart polymers in aqueous solution or on the conformational transition of macromolecules physically attached or chemically grafted to the surface. Systems that have covalently cross-linked networks of macromolecules, c alled smart hydrogels, are not considered. One could dene smart polymers used in biotechnology and medicine a s m acromolecules t hat u ndergo f ast a nd re versible changes f rom hydrophilic to h ydrophobic m icrostructure t riggered by small changes in their environments. Āe se microscopic changes a re appa rent at t he m acroscopic le vel a s p recipitate formation i n s olutions o f sm art p olymers o r c hanges i n t he wettability o f a su rface to w hich a sm art p olymer i s g rafted. Ā e changes are reversible, and the system returns to i ts initial state when the trigger is removed.

24.3.2

Smart Polymers Used in Biotechnology and Medicine

Ā e highly nonlinear transitions in smart polymers are driven by diff erent f actors, f or e xample, neut ralization o f c harged groups by either a pH shift [2] or the addition of an oppositely charged p olymer [ 3], c hanges i n t he e fficiency of hydrogen

10/3/2008 8:58:25 PM

24-14

Smart Materials

24.3.2.1

pH-Sensitive Smart Polymers

The first group of smart polymers consists of polymers whose transition b etween t he s oluble a nd i nsoluble s tate i s c reated by decreasing the net charge of the polymer molecule. The net charge can be decreased by changing the pH to neutralize the charges on the macromolecule and hence to reduce the hydrophilicity (increase the hydrophobicity) of the macromolecule. Copolymers of me thylmethacrylate ( hydrophobic pa rt) a nd methacrylic acid (hydrophilic at high pH when carboxy groups are deprotonated but more hydrophobic when carboxy groups are protonated) precipitate from aqueous solutions by acidification to pH around 5, and copolymers of methyl methacrylate ( hydrophobic p art) wi th d imethylaminethyl methacrylate (hydrophilic at low pH when amino groups are protonated b ut mo re h ydrophobic w hen a mino g roups a re deprotonated) are soluble at low pH but precipitate in slightly alkaline c onditions [ 6]. H ydrophobically mo dified c ellulose derivatives t hat h ave p ending c arboxy g roups, f or e xample, hydroxypropyl me thyl c ellulose ac etate suc cinate a re a lso soluble i n ba sic c onditions b ut p recipitate i n s lightly ac idic media [7]. Ā e pH-induced precipitation of smart polymers is very sharp and usually requires a change in pH of not more than 0.2–0.3 units ( Figure 2 4.7). W hen s ome c arboxy g roups a re u sed to couple a biorecognition element, for example, noncharged sugar, the i ncreased h ydrophobicity o f t he c opolymer re sults i n p recipitation at a higher pH [8]. Ā e copolymerization of N-acryloyl sulfametazine w ith N, N-dimethylacrylamide re sults i n a p Hsensitive polymer whose reversible transition is in the physiological range of pH 7.0–7.5 [9]. Ā e charges on the macromolecule can also be neutralized by adding a n e fficient c ounterion, f or e xample, a lo w mole cular weight c ounterion o r a p olymer mole cule o f opp osite c harges. Ā e l atter s ystems a re c ombined u nder t he na me o f p olycomplexes. Ā e cooperative nature of interaction between two polymers of opposite charges makes polycomplexes very sensitive to changes i n p H o r io nic s trength [10]. Ā e c omplex f ormed b y poly(methacrylic ac id) ( polyanion) a nd p oly(N-ethyl-4-vinylpyridinium bromide) (polycation) undergoes reversible precipitation f rom aque ous s olution at a ny de sired p H v alue i n t he range 4.5–6.5 that depends on the ionic strength and polycation/ polyanion ratio in the complex (Figure 24.8) [11]. Polyelectrolyte complexes formed by poly(ethylene imine) and poly(acrylic acid) undergo s oluble–insoluble t ransition i n a n e ven b roader p H range from pH 3–11 [12]. Ā e p H o f t he t ransition o f p H-sensitive p olymers suc h a s poly(methylmethacrylate-co-methacrylic acid) or poly (N-acryloyl sulfametazine-co-N,N-dimethylacrylamide) is strictly xed for

43722_C024.indd 14

0.8

Absorbance 470 nm

bonding and an increase in temperature or ionic strength [4], and critical phenomena in hydrogels and interpenetrating polymer networks [5]. Ā e polymer systems that have highly nonlinear re sponse c an b e d ivided i nto t hree gener al g roups: pH-sensitive smart polymers, thermosensitive smart polymers, and reversibly cross-linked networks.

0.6

0.4

0.2

0.0 3.5

4.0

4.5

5.0

5.5

6.0

6.5

7.0

7.5

8.0

pH

FIGURE 2 4.7 pH-induced pre cipitation of a r andom c opolymer of methacrylic acid and methacrylate (commercialized as Eudragit S 100 by Röhm Pharma GMBH, Weiterstadt, Germany) (open squares) and p-amino-phenyl- α-d-glucopyranoside-modied copolymer (open circles) me asured a s t urbidity at 47 0 nm. S ome d ecrease i n t urbidity at lower pH values is caused by occulation a nd s edimentation of t he polymer precipitate. (Redrawn from Hilczer, B. and Malecki, J. Electrets, Elsevier, Amsterdam, 1986.)

the g iven c omposition o f c omonomers. Ā us, a ne w p olymer should be synthesized for each desired pH value. Ā e advantage of polyelectrolyte c omplexes i s t hat b y u sing o nly t wo d ifferent polymers and mixing them in different ratios, reversible precipitation can be achieved at a ny desired pH value in a r ather broad pH range. 24.3.2.2 Thermosensitive Smart Polymers Ā e reversible solubility of t hermosensitive smart polymers is caused by changes in the hydrophobic–hydrophilic balance of uncharged p olymers i nduced b y i ncreasing tem perature o r ionic strength. Uncharged polymers are soluble in water due to hydrogen b onding w ith w ater mole cules. Ā e efficiency of hydrogen bonding lessens as temperature increases. Ā e phase separation of a polymer occurs when the efficiency of hydrogen bonding beco mes i nsufficient f or t he s olubility o f macromolecule. When the temperature of an aqueous solution of a smart polymer is raised above a certain critical temperature (which is often referred to as the transition temperature, lower critical solution temperature ( LCST), o r c loud p oint), ph ase s eparation t akes place. A n aque ous ph ase t hat c ontains p ractically no p olymer and a p olymer-enriched phase a re formed. Both phases can be easily separated by decanting, centrifugation, or  ltration. Ā e temperature o f t he ph ase t ransitions de pends o n t he p olymer concentration and molecular weight (MW) (Figure 24.9) [13,14]. Ā e ph ase s eparation i s c ompletely re versible, a nd t he sm art polymer d issolves i n w ater w hen t he tem perature i s re duced below the transition temperature.

10/3/2008 8:58:25 PM

24-15

Smart Polymers for Biotechnology and Elastomers 7

7 0.01 M NaCl

pH

0.25 M NaCl

pH

6

6

5

5 Insoluble

Insoluble

4

4

3

3

(a)

(c) 0.01 M NaCl

0.5 M NaCl

6 5 5

Insoluble 4

Insoluble

4

3

3

(b)

2

0.1

0.2

0.3

0.4

(d)

2

0.1

0.2

0.3

0.4

Ratio Poly(N-ethyl-4-vinyl-pyridinium bromide)/ poly(methacrylic acid)

FIGURE 24.8 Phase diagram for the polyelectrolyte complex formed by poly(N-ethyl-4-vinyl-pyridinium bromide) (polymerization degree 530) and p oly(methacrylic acid) (polymerization degree 1830). Ā e dot s pre sent pH v alues at w hich t he t urbidity of t he p olymer s olutions w as  rst observed at 47 0 nm. Ionic strength was (a) 0.01 M NaCl, (b) 0.1 M NaCl, (c) 0.25 M NaCl, a nd (d) 0.5 M NaCl. Dashed area represents pH/ composition range where the complex is insoluble. (Reproduced from tiertsen, H.F., J. Appl. Phys., 74(5), 3389, 1993. With permission.)

Two groups of thermosensitive smart polymers are most widely studied and used: • Poly(N-alkyl sub stituted ac rylamides) a nd t he mo st w ellknown o f t hem, p oly(N-isopropyl a crylamide) (p oly (NIPAAM) ), whose transition temperature is 32°C [14] • Poly(N-vinylalkylamides) such as poly(N-vinylisobutyramide) whose transition temperature is 39°C [15] or poly(N-vinyl caprolactam) whose transition temperature is 3 2°C–33°C ( depending o n t he p olymer’s mole cular weight) [13]

Temperature (⬚C)

50

40

30 0.1

0.2 0.3 0.4 Wt. fraction polymer

0.5

FIGURE 24.9 Phase diagram for poly(NIPAAM) in aqueous solution. Ā e a rea u nder t he bi nodal c urve pre sents t he r ange of t emperatures/ polymer concentrations for homogeneous solution. Separation into polymer-enriched and polymer-depleted phases takes place for a ny polymer concentration/temperature above t he binodal curve. (Reproduced from Vinogradov, A.M. et al., Mech. Mater., 36, 1007, 2004. With permission.)

43722_C024.indd 15

A variety of polymers that have different transition temperatures from 4° C–5°C f or p oly(N-vinyl p iperidine) to 1 00°C f or poly(ethylene glycol) are available at present [16]. pH-sensitive smart polymers usually contain carboxy or amino groups that can be used for covalent coupling of biorecognition or biocatalytic elements (ligands). Ā ermosensitive polymers, on the contrary, do not h ave inherent reactive groups, which could be used for ligand coupling. Ā us, copolymers that contain reactive groups can be synthesized. N-Acryloylhydroxysuccinimide [17] or glycidyl methacrylate [18] have often b een used as active comonomers i n c opolymerization w ith N IPAAM, a llowing further c oupling o f a mino-group-containing li gands t o t he

10/3/2008 8:58:25 PM

24-16

Smart Materials

synthesized c opolymers. Ā e u se of a n i nitiator of p olymerization [19] or chain transfer agent [20] that has an active group results in a p olymer modied only at t he end of t he macromolecule. A n a lternative st rategy i s to i ncorporate a p olymerizable double bond i nto t he ligand, for example, by modication with acryoyl g roup, a nd t hen to c opolymerize t he mo died ligand with NIPAAM [21,22]. An increase in the hydrophilicity of the polymer-accompanied incorporation of hydrophilic comonomers or coupling to hydrophilic l igands i ncreases t he t ransition tem perature, w hereas hydrophobic comonomers and ligands have the opposite effect [4]. Ā e pH-induced change in ligand hydrophobicity could have a dramatic effect on the thermoseparation of the ligand– polymer conjugate. A copolymer of NIPAAM and vinyl imidazole pre cipitates at a bout 3 5°C at p H 8 .0 w here i midazole moieties a re no ncharged a nd rel atively h ydrophobic, b ut no precipitation occurs even when heating the polymer solution to 80°C at p H 6 w here imidazole groups are protonated and very hydrophilic [23]. Ligand–ligand in teractions in a li gand–polymer c onjugate also h ave a s ignicant effect o n t he t hermoseparation. Ā e precipitation temperature for the previously mentioned copolymers of NIPAAM and vinyl imidazole increases as the imidazole content in t he copolymer increases. On t he contrary, t he precipitation tem perature de creases a s t he i ncrease o f i midazole content i ncreases, w hen t he p olymer f orms a Cu(I I)-complex [23]. Each Cu(II) ion interacts with two to three imidazole groups to c ross-link t he s egments o f t he p olymer mole cule [ 24]. Ā e restricted mob ility o f t he p olymer s egments re sults i n a lo wer precipitation temperature. Block c opolymers t hat h ave a t hermosensitive “ smart” pa rt that consists of poly(NIPAAM) form reversible gels on an increase i n temperature, whereas ra ndom copolymers separate from aque ous s olutions b y f orming a c oncentrated p olymer phase [ 25]. Ā us, t he p roperties o f sm art p olymers t hat a re important for biotechnological a nd medical applications could be controlled by the composition of comonomers and also by the polymer architecture. Ā e phase transition of thermosensitive polymers at increased temperature re sults f rom h ydrophobic i nteractions b etween polymer mole cules. B ecause hydrophobic i nteractions a re promoted by high salt concentrations, the addition of salts shifts the cloud point to lower temperatures. When the transition temperature i s b elow ro om tem perature, p olymer p recipitation i s achieved by just a salt addition without any heating. Ā e addition of organic solvents, detergents, and chaotropic agents increases the transition temperature because these compounds deteriorate hydrophobic interactions. 24.3.2.3

Reversibly Cross-Linked Polymer Networks

Systems that have reversible noncovalent cross-linking of separate polymer molecules into a p olymer network belong to t he t hird group of smart polymers. When formed, reversibly cross-linked polymers either precipitate or form a physical gel. Polymers that

43722_C024.indd 16

have su gar l igands c ross-linked b y le ctins w ith m ultibinding sites [26] and boronate-polyols [27–29] are the most widely used systems of this type. Ā e reversible response in these systems is achieved b y add ition o r remo val o f a lo w mole cular w eight analog of the polymer. For example, small sugars added at high concentrations co mpete w ith s ugar-containing po lymers f or binding to le ctin a nd de stroy i ntrapolymer c ross-links t hat result in disengagement of the network. 24.3.2.4

Heterogeneous Systems Using Smart Polymers

A s olid su rface ac quires ne w p roperties w hen mo died by adsorption or chemical graft ing o f sm art p olymers. Sm art polymers that have terminal (only single-point attachment possible) o r r andom ( multipoint at tachment p ossible) c ould b e covalently coupled to the respective active groups on the surface [30]. Single-point attachment could also be achieved by covalent modication of the surface using an initiator of polymerization and t hen c arrying o ut p olymerization o f mo nomers i n t he solution t hat su rrounds t he supp ort. Ā e g rowth o f p olymer chains occurs only at t he sites where initiator was coupled [31]. Alternatively, t he s olid supp ort i s i rradiated b y l ight [32] o r a plasma beam [33] when monomer is in the surrounding solution. Active r adical sites on t he su rface, w hich app ear a s a re sult of irradiation, initiate the growth of polymer macromolecules. As a rule, irradiation methods give a h igher density of grafted polymer, but polymerization is less controlled as in covalent coupling or using a covalently coupled initiator. Irradiation, especially at high mo nomer c oncentrations, c ould p roduce a c ross-linked polymer gel attached to the solid support [34]. A separate group of smart polymers is represented by particulate systems. Liposomes that reversibly precipitate on salt addition and removal were prepared from a s ynthetic phospholipid t hat had a diacetylene moiety in the hydrophobic chain and an amino group in the hydrophilic head of the phospholipid, followed by polymerization of diacetylene bonds [35]. Lattices composed of thermosensitive polymers or a layer of thermosensitive polymer at the surface represent another example of insoluble but reversibly suspended pa rticulate s ystems t hat re spond to i ncreasing o r decreasing temperature [31].

24.3.3 Applications Ā ere are numerous potential applications for smart polymers in biotechnology a nd medicine. Ā e main commercial application of smart polymers is the production of “smart” pills where the shell of the smart polymer protects the pill from the harmful action of the stomach contents but a llows the pill to dissolve in the intestine. Ā ere is not yet any other product in the m arket t hat app lies sm art p olymers, b ut t he i nterest i n these applications is growing in both the academic community a nd i ndustry. Ā e f ollowing app lications a re c onsidered in this chapter:

10/3/2008 8:58:26 PM

24-17

Smart Polymers for Biotechnology and Elastomers

• Smart pills that have an enteric coating • Smart polymers for affi nity precipitation of proteins • Aqueous t wo-phase p olymer s ystems f ormed b y sm art polymers and their application for protein purication • Smart surfaces for mild detachment of cultivated mammalian cells • Smart c hromatographic m atrices t hat re spond to temperature • Smart p olymers f or c ontrolled p orosity o f s ystems— “chemical valve” • Liposomes that trigger the release of their contents • Smart polymers for bioanalytical applications • Reversibly soluble biocatalysts 24.3.3.1 Smart Pills That Have an Enteric Coating It is common knowledge that peroral introduction of medical preparations is the most convenient method compared to subcutaneous or intravenous injection and even to nasal sprays or eye droplets. Ā e absorption of a swallowed pill takes place predominantly i n t he i ntestine a nd to re ach t he i ntestine, t he medicine must pass unharmed through the stomach that has a very low pH value of 1.4 and abundant hydrolytic enzymes that can de grade a b road v ariety o f c hemical s tructures. M any medicines a re susceptible to d amage i n t he stomach environment. Ā e ideal condition for peroral introduction is to have a smart pill, which is insoluble in the stomach and hence passes through t he s tomach u naffected b ut e asily d issolves at t he higher p H i n t he i ntestine w here t he me dicine i s a bsorbed. Smart p olymers p rovide t he s olution. H ydrophobic p olymers such a s po ly(methylmethacrylate) o r h ydrophobically m odied celluloses a re i nsoluble i n water per se, but t he i ntroduction o f c arboxy g roups ( either b y pa rtial h ydrolysis o f e ster groups in methylmethacrylate or modication of cellulose HO groups by dicarboxylic acids such as succinic or phthalic acid) endows th e p olymers wi th p H-dependent s olubility. Ā e pill covered by a shell of such a polymer (enteric coating) is insoluble at lo w p H w hen t he c arboxy g roups a re p rotonated a nd uncharged, b ut e asily s oluble at a p H a bove 6 w hen c arboxy groups a re pr otonated a nd c harged. I ndustrially pr oduced

polymers for enteric coating belong to t wo main groups, synthetic copolymers of methylmethacrylate and methacrylic acid and modied derivatives of cellulose, a natural polymer (Table 24.2). Ā e  rst group of polymers is used mainly by European and U.S. manufacturers, and the second group is more popular in Japan. Whenever t he c harge-bearing c omonomer h as a n a mino group instead of a c arboxy group, the solubility of the polymer acquires opposite pH dependence. Ā e polymer is soluble at low pH values but insoluble in neutral and alkaline media. Poly(diet hylaminoethylmethacrylate-co-methylmethacrylate) ( commercialized as Eudragit E) is an example of such a polymer. Ā e shell that is composed of this polymer protects the tablet against dissolution in the neutral saliva, and the mouth is not affected by the unpleasant taste of bitter medicine, but the polymer dissolves readily in the stomach. 24.3.3.2

Bioseparation—Affinity Precipitation

All b ioseparation p rocesses i nclude t hree s tages: p referential partitioning o f t arget sub stance a nd i mpurities b etween t wo phases (liquid–liquid or liquid–solid), mechanical separation of the phases (e.g., separation of the stationary and mobile phases in a c hromatographic column), a nd recovery of t he target substance f rom t he en riched ph ase. B ecause sm art p olymers c an undergo phase transitions, they could facilitate the second and the third stages of bioseparation processes. Ā e a bility o f sm art p olymers to f orm i n si tu h eterogeneous systems is exploited i n a ffinity precipitation (Figure 2 4.10). Ā e technique is based on using a conjugate of a smart polymer that has a c ovalently coupled biorecognition moiety, that is, a l igand specic for a target protein. Ā e conjugate forms a complex with the t arget p rotein b ut not w ith t he ot her p roteins i n t he c rude extract. P hase s eparation o f t he c omplex i s t riggered b y sm all changes in the environment, resulting in transition of the polymer backbone into an insoluble state. Ā e target protein specically coprecipitates with the smart polymer and the impurities in the crude remain in solution. Ā en, the target protein is either eluted f rom t he i nsoluble m acroligand–protein c omplex o r t he precipitate is dissolved. Ā e protein is dissociated from the macroligand and the ligand–polymer conjugate is precipitated again

TABLE 24.2 Industrially Manufactured Smart Polymers for Producing Smart Pills

43722_C024.indd 17

Polymer

Trade Name

Manufacturer

Country

Poly(methacrylic acid-co-methylmethacrylate) 1:1 monomer ratio, MW 135,000 Poly(methacrylic acid-co-methylmethacrylate) 1:2 monomer ratio, MW 135,000 Carboxymethylcellulose Cellulose acetate phthalate Hydroxypropylmethyl-cellulose phthalate Hydroxypropylmethyl-cellulose acetate succinate Poly(diethylaminoethyl methacrylate-comethylmethacrylate) MW 150 000

Eudragit L

Röhm Pharma GmBH

United States, Germany

Eudragit S

Röhm Pharma GmBH

United States, Germany

CMEC CAP HP-50, HP-55 ASM, AS-H Eudragit E

Freund Sangyo Co., Ltd. Wako Pure Chemicals Ltd. Shin-Etsu Chemical Co., Ltd. Shin-Etsu Chemical Co., Ltd. Röhm Pharma GmBH

Japan Japan Japan Japan United States, Germany

10/3/2008 8:58:26 PM

24-18

Smart Materials

Impurities in supernatant

Addition of crude extract, complexing of protein with macroligand

Polymer precipitation, separation of pellet Polymer dissolution

Polymer precipitation, separation of pellet

Dissociation of the complex

Purified protein in supernatant Keys

Product Ligand

Polymer in

Polymer in

Impuritie

FIGURE 24.10 Schematic of affi nity precipitation technique for protein purication.

now without the protein that remains in the supernatant in puried f orm. A v ariety o f d ifferent l igands suc h a s t riazine dye s, sugars, p rotease i nhibitors, a ntibodies, n ucleotides, do uble-or single-stranded DNA, and chelated metal ions were successfully used for affinity precipitation [36]. After elution of the target protein, the ligand–polymer conjugate could be recovered and used in the next purication cycle [37]. Triazine d yes, r obust a ffi nity l igands f or m any n ucleotidedependent enzymes, were successfully used in conjugates with the p H-sensitive c opolymer o f me thacrylic ac id a nd me thylmethacrylate, which precipitates when pH decreases (Eudragit S 100), for the purication of dehydrogenases from various sources by affinity precipitation [38,39]. Sugar ligands constitute another attractive a lternative a nd h ave b een u sed i n c ombination w ith Eudragit S 100 for bioseparation of lectins [40]. Restriction endonuclease Hind III was successfully isolated using the thermosensitive c onjugate o f po ly(NIPAAM) w ith p hage λ D NA [ 21]. Human Ig G w as s pecically p recipitated w ith a c onjugate o f protein A a nd ga lactomannan. G alactomannan p olymer w as reversibly precipitated by adding tetraborate [41]. Ā e efficient precipitation of Cu(I I)-loaded poly(N-vinylimidazole-co-NIPAAM) by high salt concentrations at mild temperature is very convenient for metal affin ity precipitation of proteins t hat have i nherent h istidine re sidues at t he su rface or for recombinant proteins articially provided with histidine tags (usually four to six residues). High salt concentration does not interfere with protein–metal ion–chelate interaction, and, on the

43722_C024.indd 18

other hand, it reduces t he possibility of nonspecic binding of foreign proteins to t he polymer both in solution and when precipitated [23]. Ā e exibility of polymer chains in solution allows several imidazole ligands on a p olymer molecule to c ome close enough to interact with the same Cu(II) ion and thus to provide sufficient strength of polymer–Cu(II) interactions to purify a variety of histidine-containing proteins [37]. Polyelectrolyte complexes that have pH-dependent solubility were s uccessfully u sed i n d ifferent b ioseparation p rocedures. When an antigen, inactivated glyceraldehyde-3-phosphate dehydrogenase, f rom rabbit was covalently coupled to a p olycation, the resulting complex was used to purify monoclonal antibodies specic to ward i nactivated g lyceraldehyde-3-phosphate deh ydrogenase [11]. Ā e successful affinity precipitation of antibodies using g lyceraldehyde-3-phosphate deh ydrogenase b ound to a polyelectrolyte c omplex i ndicates t hat t he l igand i s e xposed to the solution. Ā is fact was used to develop a new method for producing monovalent Fab f ragments of a ntibodies. Traditionally, Fab fragments are produced by proteolytic digestion of antibodies in solution followed by isolation of Fab fragments. In the case of monoclonal antibodies against inactivated subunits of glyceraldehyde-3-phosphate deh ydrogenase, d igestion w ith papa in resulted i n signicant damage of binding sites of t he Fab f ragment. P roteolysis o f mo noclonal a ntibodies i n t he p resence o f the a ntigen–polycation c onjugate f ollowed b y (1) p recipitation induced by adding polyanion, poly(methacrylic) acid, and a pH shift from 7.3 to 6.5 and (2) elution at pH 3.0 that resulted in 90%

10/3/2008 8:58:27 PM

24-19

Smart Polymers for Biotechnology and Elastomers

immunologically c ompetent F ab f ragments. M oreover, t he papain concentration required for proteolysis was 10 times less for a ntibodies b ound to t he a ntigen–polycation c onjugate compared to that for free antibodies in solution [42]. Active glyceraldehyde-3-phosphate dehydrogenase from rabbit muscle was separated from the inactivated enzyme by using monoclonal antibodies specic for the inactivated enzyme covalently coupled to the polyanion component of the polyelectrolyte complex. Ā is system can be regarded as a simplied model of chaperone action in living cells that assist in separating active protein molecules from misfolded ones [43]. Apart from specic interactions between a target protein and a ligand–polymer conjugate, nonspecic interactions of protein impurities w ith t he p olymer bac kbone c ould t ake p lace. Ā e nonspecific i nteractions l imit t he e fficiency o f t he a ffinity precipitation te chnique, a nd sig nicant efforts were made to reduce t hese i nteractions. Ā e adv antage o f p olyelectrolyte complexes as carriers for affi nity precipitation is low nonspecic coprecipitation of proteins when the polymer undergoes a soluble– insoluble transition [10]. Smart particles capable of reversible transition between aggregated and dispersed states were used for affin ity precipitation of proteins. Ā ermosensitive [44] or pH-sensitive lattices [45] or salt-sensitive liposomes that have polymerized membranes [35] are examples of such systems. Two elements are required for successful affin ity precipitation. Ā e backbone of a smart polymer provides precipitation at the desired c onditions ( temperature, p H, io nic s trength), a nd the biorecognition element is responsible for selective binding of

the protein of interest. By proper choice of a smart polymer, precipitation c ould b e ac hieved p ractically at a ny de sired p H o r temperature. F or e xample, p oly(N-acryloylpiperidine) ter minally modied with maltose has an extremely low critical temperature (soluble below 4°C and completely insoluble above 8°C) and was used to purify thermolabile α-glucosidase [46]. 24.3.3.3 Bioseparation—Partitioning in Aqueous Polymer Two-Phase Systems Two aqueous polymer solutions become mutually incompatible when t he t hreshold c oncentrations o f p olymers a re e xceeded. Both of the polymer phases formed contain about 90% water and hence present a very friendly environment for proteins and other biomolecules. Proteins partition selectively between two phases depending on their size, charge, hydrophobicity, nature, and the concentration of the phase-forming polymers. Ā e partitioning could be also directed by adding some salt or coupling an affinity ligand specic for a g iven protein to o ne of t he phase-forming polymers [47]. Ā e selective partitioning of proteins between the two phases formed has proven to be an efficient tool for purifying proteins a nd some low molecular weight substances. Ā e main problem of the method—how to separate the target protein from the phase-forming polymer—has not yet been completely solved. Smart p olymers provide a n ele gant s olution to t his problem— simple precipitation of the phase-forming polymer leaves protein in the supernatant (Figure 24.11): (1) Ā e crude protein extract is mixed w ith t he aque ous t wo-phase p olymer s ystem a nd t he conditions are selected so that the protein of interest partitions into a phase formed by a smart polymer (for example by coupling

Discarded Keys: Phase separation Addition of crude

Protein of interest Impurities

Concentrated polymer phase reused after dissolution in cold buffer 1. Heating 2. Phase separation

Purified protein in buffer solution

FIGURE 24.11 Schematic presentation of protein partitioning in aqueous two-phase polymer system formed by a smart (thermosensitive polymer). (Reproduced from Galaev, I.Y. and Mattiasson, B., Trends Biotechnol., 17, 335, 1999. With permission.)

43722_C024.indd 19

10/3/2008 8:58:27 PM

24-20

Smart Materials

affinity ligand to the smart polymer), and the impurities concentrate in the other phase. (2) the phases are separated mechanically and t he ph ase f ormed b y t he sm art p olymer i s sub jected to conditions (pH o r tem perature) w here t he p olymer u ndergoes phase s eparation; (3) t wo ne w ph ases a re formed, a p olymerenriched ph ase o f h igh p olymer a nd lo w w ater c oncentration, which contains practically no p rotein, a nd a p olymer-depleted aqueous ph ase t hat c ontains mo st o f t he p uried p rotein a nd minute amounts of the polymer left after phase separation. pH-sensitive acrylic copolymers [48] or thermoresponsive polymers, po ly(ethylene o xide-co-propylene o xide) [ 49,50] or poly(N-vinyl c aprolactam-co-vinyl imidazole) [51], form twophase s ystems f rom rel atively h ydrophilic p olymers suc h a s dextran or modied starch and have been successfully used for protein pu rication. Ā e p H- o r t hermoprecipitated po lymer opposite de xtran c ould b e re generated by d issolution at a lo wer temperature. Q uite re cently, a n aque ous t wo-phase p olymer system w as de veloped w here b oth ph ase-forming p olymers, poly(N-isopropylacrylamide-co-vinyl imidazole) and poly(ethylene oxide-co-propylene o xide) e nd mo died b y h ydrophobic C 14H29 groups, are thermoresponsive and could be recycled [52]. 24.3.3.4

Smart Surfaces—Cell Detachment

Ā e driving force behind phase separation of smart polymers is a sharp increase in hydrophobicity after a small change in environmental c onditions. Ā e h ydrophobic “ collapsed” p olymer aggregates form a separate phase. When grafted to t he surface, macromolecules of the smart polymer cannot aggregate, but the conformational t ransition f rom t he h ydrophilic to t he h ydrophobic state endows the surface with regulated hydrophobicity: the surface is hydrophilic when the smart polymer is in the expanded “soluble” conformation and hydrophobic when the polymer is i n t he c ollapsed “ insoluble” c onformation. Ā e c hange o f hydrophobicity of the surface by grafted poly(N-isopropylacrylamide) w as demo nstrated b y c ontact a ngle me asurements [53] and water absorbency [54]. Ā e t ransition tem perature f or ad sorbed ( presumably v ia multipoint attachment) poly(NIPAAM) molecules is lower than that in bulk solution, and the properties of the layer of collapsed macromolecules f ormed a bove t he t ransition tem perature depend strongly on the speed by which the temperature increases. At a low speed of temperature increase, the “liquid-like” polymer layer i s formed, w hereas at h igh s peeds, t he p olymer l ayer h as more “ solid-like” prop erties [55]. W hen c ooling, t he c ollapsed polymer molecules return to t he initial loopy adsorbed conformation via transitional extended conformation. Ā e relaxation process f or t he e xtended-to-loopy ad sorbed c onformational transition occurs slowly and depends on the temperature observance of a n A rrhenius l aw. K inetic c onstraints, it i s proposed, play an important role in this transition [56]. Ā e change of surface properties from hydrophobic above the critical temperature of the polymer grafted to hydrophilic below it h as b een suc cessfully u sed f or de taching m ammalian c ells. Mammalian c ells a re no rmally c ultivated o n a h ydrophobic solid substrate and are detached from the substrate by protease

43722_C024.indd 20

treatment, which often damages the cells by hydrolyzing various membrane-associated p rotein mole cules. Ā e poly(NIPAAM)grafted surface is hydrophobic at 37°C because this temperature is above the critical temperature for the grafted polymer and that cells t hat a re g rowing w ell o n i t. A de crease i n tem perature results in transition of the surface to the hydrophilic state, where the cells can be easily detached from the solid substrate without any damage. Poly(NIPAAM) was grafted to polystyrene culture dishes using an electron beam. Bovine hepatocytes, cells that are highly sensitive to enzymatic treatment, were cultivated for 2 days at 37°C and detached by incubation at 4°C for 1 h. Nearly 100% of t he hepatocytes was detached a nd recovered f rom t he poly(NIPAAM)-grafted d ishes b y l ow-temperature t reatment, whereas only about 8% of the cells was detached from the control dish [57]. Ā e technique has been extended to different cell types [58,59]. It i s noteworthy t hat hepatocytes re covered by cooling retained t heir nat ive form had numerous bulges a nd d ips, a nd attach well to the hydrophobic surface again, for example, when the temperature was increased above the conformational transition o f p oly(NIPAAM). O n t he c ontrary, en zyme-treated c ells had a smooth outer surface and had lost their ability to attach to the su rface. Ā us, c ells re covered b y a tem perature s hift from poly(NIPAAM)-grafted s urfaces h ave a n in tact s tructure a nd maintain normal cell functions [58]. Ā e molecular machinery involved in cell-surface detachment was i nvestigated u sing tem perature-responsive su rfaces [ 60]. Poly(NIPAAM)-grafted a nd n ongrafted s urfaces sh owed n o difference in attachment, spreading, growth, conuent cell density, o r mo rphology o f b ovine ao rtic en dothelial c ells at 3 7°C. Stress bers, peripheral bands, and focal contacts were established in similar ways. When the temperature was decreased to 2 0°C, the c ells g rown o n p oly(NIPAAM)-grafted supp ort lo st t heir attened morphology and acquired a rounded appearance similar to t hat of c ells immediately after plating. Mild agitation makes the cells oat free from the surface without a trypsin treatment. Neither c hanges i n c ell mo rphology no r c ell de tachment occurred on ungrafted surfaces. Sodium azide, an ATP synthesis inhibitor, and genistein, a t yrosine kinase inhibitor, suppressed changes in cell morphology and cell detachment, whereas cycloheximide, a p rotein s ynthesis i nhibitor, s lightly en hanced c ell detachment. P halloidin, a n ac tin  lament s tabilizer, a nd i ts depolymerizer, c ytochalasin D, a lso i nhibited cell detachment. Ā e se ndings suggest that cell detachment from grafted surfaces is mediated by intracellular signal transduction and reorganization of the cytoskeleton, rather than by a si mple changes in the “stickiness” of the cells to t he surface when the hydrophobicity of the surface is changed. One c ould i magine p roducing a rticial o rgans us ing temperature-induced detachment of cells. A rticial sk in could be produced as the cells are detached from the support not as a suspension ( the u sual r esult o f p rotease-induced de tachment) but preserving their intercellular contacts. Fibroblasts were cultured o n t he p oly(NIPAAM)–collagen supp ort u ntil t he c ells completely covered the surface at 37°C, followed by a decrease in temperature to a bout 1 5°C. Ā e s heets o f  broblasts detached

10/3/2008 8:58:27 PM

Smart Polymers for Biotechnology and Elastomers

from t he d ish a nd w ithin a bout 1 5 min  oated i n t he c ulture medium [57]. Ā e detached cells could be transplanted to another culture surface without functional and structural changes [34]. Grafting of poly(NIPPAM) onto a polystyrene surface by photolitographic te chnique c reates a s pecial pat tern o n t he su rface, and by decreasing temperature, cultured mouse broblast STO cells a re de tached o nly f rom t he su rface a rea o n w hich poly(NIPAAM) wa s g rafted [ 61]. Li thographed  lms o f sm art polymer p resent supp orts f or c ontrolled i nteractions o f c ells with su rfaces a nd c an d irect t he at tachment a nd s preading o f cells [62]. One could envisage producing articial cell assemblies of complex architecture using this technique. 24.3.3.5

Smart Surfaces—Temperature-Controlled Chromatography

Surfaces th at h ave th ermoresponsive h ydrophobic p roperties have been used in chromatography. High-performance liquid chromatography (HPLC) columns with grafted poly(NIPAAM) have been used for separating steroids [63] and drugs [64]. Ā e chromatographic re tention a nd re solution o f t he s olutes w as strongly dependent on temperature and increased as temperature increased f rom 5° C to 5 0°C, w hereas t he re ference c olumn packed with nonmodied silica displayed much shorter retention times t hat de creased a s tem perature de creased. H ydrophobic interactions dominate in retaining solutes at higher temperature, and t he preferential re tention of hydrogen-bond ac ceptors was observed at low temperatures. Ā e effect of temperature increase on t he re tention b ehavior o f s olutes s eparated o n t he poly(NIPAAM)-grafted si lica c hromatographic m atrix w as similar to t he add ition o f me thanol to t he mob ile ph ase at constant temperature [65]. Ā e temperature response of t he poly(NIPAAM)–silica matrices de pends d rastically o n t he a rchitecture o f t he g rafted polymer molecules. Surface wettability changes dramatically as temperature changes across the range 32°C–35°C (corresponding to t he ph ase-transition tem perature f or N IPAAM i n aque ous media) for surfaces where poly(NIPAAM) is terminally grafted either directly to the surface or to the looped chain copolymer of NIPAAM and N-acryloylhydroxysuccinimide, which was initially coupled to t he su rface. Ā e w ettability c hanges f or t he lo opgrafted surface itself were relatively large but had a slightly lower transition tem perature (∼27°C). Ā e r estricted co nformational transitions for multipoint grafted macromolecules are probably the re ason f or t he re duced t ransition tem perature. Ā e largest surface f ree energ y c hanges a mong t hree su rfaces w ere observed for t he c ombination o f b oth lo ops a nd ter minally grafted chains [30]. Ā e i ntroduction o f a h ydrophobic c omonomer, b uthylmethacrylate, i n t he p olymer re sulted i n a de creased t ransition temperature o f a bout 2 0°C. Re tention o f s teroids i n p oly (NIPAAM-co-buthylmethacrylate)-grafted columns increases as column temperature increases. Ā e capacity factors for steroids on the copolymer-modied silica beads was much larger than that on poly(NIPAAM)-grafted columns. Āe effect of temperature on steroid re tention o n p oly(NIPAAM-co-buthylmethacrylate)-

43722_C024.indd 21

24-21

grafted st ationary p hases wa s m ore p ronounced co mpared t o supports modied with poly(NIPAAM). Furthermore, retention times for steroids increased remarkably as the buthylmethacrylate content increased in the copolymer. Ā e temperature-responsive elution of steroids was strongly affected by the hydrophobicity of the grafted polymer chains on silica surfaces [63]. Ā e mixture of polypeptides, consisting of 21–30 amino acid residues (insulin chain A, β-endorphin fragment 1–27 and insulin chain B) could not be separated at 5°C (below the transition temperature) on copolymer-grafted matrix. At this temperature, the copolymer is in an extended hydrophilic conformation that results in decreased interactions with peptides and hence short retention t imes i nsufficient to re solve t hem. Ā e m ixture h as been easily separated at 30°C, when the copolymer is collapsed, hydrophobic i nteractions a re mo re p ronounced, a nd re tention times s ufficiently lo ng f or re solving p olypeptides [ 66]. L arge protein molecules such a s immunoglobulin G demonstrate less pronounced changes in adsorption above and below the transition temperature. O nly a bout 2 0% o f t he p rotein ad sorbed o n poly(NIPAAM)-grafted silica at 37°C (above the LCST) were eluted after decreasing temperature to 24°C (below the transition temperature) [67]. Quantitative elution of proteins adsorbed on the m atrix v ia h ydrophobic i nteractions h as not ye t b een demonstrated, although protein adsorption on poly(NIPAAM)grafted matrices could be somewhat controlled by a temperature shift. A suc cessful s trategy f or tem perature-controlled p rotein chromatography proved to b e a c ombination of temperatureresponsive p olymeric g rafts a nd b iorecognition elemen t, for example, affinity ligands. The ac cess of t he protein mole cules to t he l igands on t he surface of the matrix is affected by the transition of the polymer macromolecule g rafted o r a ttached t o t he ch romatographic matrix. Triazine dyes, for example, Cibacron Blue, are often used as ligands for dye-affinity chromatography of various nucleotidedependent enzymes [68]. Poly(N-vinyl caprolactam), a t hermoresponsive p olymer w hose c ritical tem perature i s a bout 3 5°C, interacts e fficiently w ith t riazine dye s. P olymer mole cules o f 40,000 MW are capable of binding up to seven to eight dye molecules hence, the polymer binds via multipoint interaction to the dye ligands available on the chromatographic matrix. At elevated temperature, polymer molecules are in a c ompact globule conformation t hat c an b ind o nly to a f ew l igands o n t he m atrix. Lactate deh ydrogenase, a n en zyme f rom p orcine m uscle, h as good access to the ligands that are not occupied by the polymer and b inds to t he c olumn. P oly(N-vinyl c aprolactam) m acromolecules undergo transition to a more expanded coil conformation a s temperature decreases. Now, t he polymer molecules interact with more ligands and begin to compete with the bound enzyme for the ligands. Finally, the bound enzyme is displaced by the expanded polymer chains. Ā e temperature-induced elution w as q uantitative, and rst re ported i n t he l iterature w hen temperature change was used as the only eluting factor without any changes i n buffer composition [69]. Small changes i n temperature, as the only eluting factor, are quite promising because there is no ne ed in this case to s eparate the target protein from

10/3/2008 8:58:27 PM

24-22

an eluent, usually a competing nucleotide or high salt concentration in dye-affin ity chromatography. 24.3.3.6 Smart Surfaces—Controlled Porosity, “Chemical Valve” Environmentally c ontrolled c hange i n m acromolecular si ze from a c ompact h ydrophobic g lobule to a n e xpanded h ydrophilic coil is exploited when smart polymers are used in systems of en vironmentally c ontrolled p orosity, s o-called “ chemical valves.” W hen a sm art polymer is g rafted to t he su rface of t he pores i n a p orous memb rane o r c hromatographic m atrix, t he transition in the macromolecule affects the total free volume of the pores available for the solvent and hence presents a means to regulate the porosity of the system. Membranes of pH-sensitive permeability were constructed by graft ing sm art p olymers suc h a s p oly(methacrylic ac id) [ 70], poly(benzyl g lutamate), p oly(2-ethylacrylic ac id) [ 71], p oly(4vinylpyridine) [72], which change their conformation in response to pH. Ā ermosensitive chemical valves have been developed by grafti ng poly(N-acryloylpyrrolidine), p oly(N-n-propylacrylamide), o r p oly(acryloylpiperidine) [ 73], p oly(NIPAAM) a lone [33,74] or in copolymers with poly(methacrylic acid) [74] inside the pores. For example, grafted molecules of poly(benzyl glutamate) at high pH are charged and are in extended conformation. Ā e efficient pore size is reduced, and the ow through the membrane is low (“off-state” of the membrane). As pH decreases, the macromolecules a re protonated, lose t heir charge, a nd adopt a compact conformation. Āe efficient pore size and hence the ow through the membrane increases (“on-state” of the membrane) [71]. Ā e uxes of b igger mole cules ( dextrans of mole cular weights 4 400–50,600) ac ross a tem perature-sensitive, poly(NIPAAM)-grafted membrane were effectively controlled by tem perature, en vironmental io nic s trength, a nd de gree o f graft ing of t he membrane, w hile t he  ux of sm aller mole cules such as mannitol was not a ffected by temperature even at high degree of membrane grafting [75]. Ā e on–off permeability ratio for different molecules (water, Cl− ion, choline, insulin, and albumin) ranged between 3 and 10 and increased as molecular weight increased [76]. A n even more abrupt change of t he on–off permeability r atio w as observed for a memb rane t hat had na rrow pores f ormed b y h eavy io n b eams w hen p oly(NIPAAM) o r poly(acryloyl-l- proline methyl ester) were grafted [77]. Different s timuli c ould t rigger t he t ransition o f t he sm art polymer, making it possible to p roduce membranes whose permeabilities re spond to t hese s timuli. W hen a c opolymer o f NIPAAM w ith t riphenylmethane le ucocianide w as g rafted to the memb rane, i t ac quires ph otosensitivity—UV i rradiation increases permeation through the membrane [78]. Fully reversible, pH-switchable permselectivity for both cationic and anionic redox-active probe molecules was achieved by depositing composite  lms formed from multilayers of amine-terminated dendrimers a nd p oly(maleic a nhydride-co-methylvinyl et her) o n gold-coated silicon [79]. When t he sm art p olymer i s g rafted i nside t he p ores o f t he chromatographic m atrix f or ge l p ermeation c hromatography,

43722_C024.indd 22

Smart Materials

the transition of grafted macromolecules regulates the pore size and a s a re sult, t he el ution p ro le of s ubstances of d ifferent molecular weights. As the temperature is raised, the substances are eluted progressively earlier, indicating shrinking of the pores of the hydrogel beads composed of cross-linked poly(acrylamideco-N-isopropylacrylamide) [80] o r p orous p olymer b eads w ith grafted poly(NIPAAM) [81]. When u sing a s pecic biorecognition element, which recognizes specic substances and translates the signal into a change of ph ysicochemical p roperties, f or e xample, p H, a sm art mem brane t hat c hanges i ts p ermeability i n re sponse to pa rticular substances c an b e c onstructed. Sp ecic i nsulin rele ase i n response to increasing glucose concentration, that is, an articial pa ncreas, p resents a n e verlasting c hallenge to b ioengineers. O ne o f t he p otential s olutions i s a “ chemical v alve” (Figure 2 4.12). Ā e en zyme, g lucose o xidase, w as u sed a s a biorecognition element, capable of specic oxidation of glucose accompanied b y a de crease i n pH. Ā e en zyme w as i mmobilized o n p H-responsive po ly(acrylic aci d) g raft on a p orous polycarbonate memb rane. I n neut ral c onditions, p olymer chains a re den sely c harged a nd h ave e xtended c onformation that p revents i nsulin t ransport t hrough t he memb rane b y blocking the pores. Under exposure to glucose, the pH drops as the result of glucose oxidation by the immobilized enzyme, the polymer c hains adopt a mo re c ompact c onformation t hat diminishes t he blo ckage o f t he p ores, a nd i nsulin i s t ransported t hrough t he memb rane [ 82]. S ystems suc h a s t his could be u sed for effi cient drug delivery that responds to the needs of the o rganism. A memb rane t hat c onsists o f p oly(2hydrohyethyl a crylate-co-N,N-diethylaminomethacrylate-co-4trimethylsilylstyrene) u ndergoes a s harp t ransition f rom a shrunken state at pH 6.3 to a swollen state at pH 6.15. Ā e transition between t he t wo states changes t he membrane permeability to insulin 42-fold. Copolymer capsules that contain glucose oxidase and insulin increase insulin release vefold in response to 0. 2 M g lucose. A fter g lucose remo val, t he r ate o f insulin release falls back to the initial value [83]. Alternatively, reversible cross-linking of p olymer macromolecules c ould b e u sed to c ontrol t he p orosity i n a s ystem. Two polymers, po ly(m-acrylamidophenylboronic a cid-co-vinylpyrrolidone) a nd p oly(vinyl a lcohol) f orm a gel b ecause o f s trong interactions between boronate groups and the hydroxy groups of poly(vinyl a lcohol). W hen a lo w mole cular w eight p olyalcohol such as glucose is added to t he gel, i t competes with poly(vinyl alcohol) for boronate groups. Ā e boronate–poly(vinyl a lcohol) complex changes to a b oronate–glucose complex that results in eventual dissolution of the gel [84]. In addition to a glucose oxidase-based articial pancreas, the boronate–poly(vinyl alcohol) system has been used for constructing glucose-sensitive systems for i nsulin del ivery [29,85–87]. Ā e g lucose-induced t ransition from a gel to a sol state drastically increases the release of insulin from t he gel . Ā e re versible re sponse to g lucose h as a lso b een designed using another glucose-sensitive biorecognition element, Concanavalin A , a p rotein t hat c ontains f our si tes t hat c an bind glucose. Polymers that have glucose groups in the side

10/3/2008 8:58:28 PM

24-23

Smart Polymers for Biotechnology and Elastomers

Insulin

(a)

Insulin H+ Glucose

Glucose H+ Glucose oxidase (b)

FIGURE 24.12 Schematic of a “chemical valve.” Glucose oxidase is immobilized on a pH-responsive polyacrylic acid grafted onto a porous polycarbonate membrane: (a) poly(acrylic acid) is in an expanded conformation that blocks insulin transport; (b) the oxidation of glucose is accompanied by a decrease in pH and the transition of poly(acrylic cid) into a compact conformation that results in opening of the pores and transport of insulin. (Redrawn from Imanishi, Y. and Ito, Y., Pure Appl. Chem., 67, 2015, 1995.)

chain such a s p oly(vinylpyrrolidone-co-allylglucose) [ 26] o r poly(glucosyloxyethyl me thacrylate) [ 88], a re re versibly c rosslinked by Concanavalin A a nd form a gel . Ā e addition of glucose results in displacing the glucose-bearing polymer from the complex with Concanavalin A and dissolving the gel. Reversible gel-formation of thermosensitive block copolymers in response to temperature could be utilized in different applications. P oly(NIPAAM) b lock co polymers w ith po ly(ethylene oxide), which undergo a temperature-induced reversible gel–sol transition, were patented as the basis for cosmetics such as depilatories a nd ble aching a gents [ 89]. Ā e c opolymer s olution i s liquid at room temperature and easily applied to the skin where it f orms a gel w ithin 1 min. C ommercially a vailable ethyl(hydroxyethyl)celluloses t hat h ave c loud p oints o f 65° C– 70°C have been used as redeposition agents in washing powders. Adsorption of t he precipitated p olymer on t he l aundry during the initial rinsing period counteracts readsorption of dirt when the detergent is diluted [90]. 24.3.3.7 Liposomes that Trigger Release of the Contents When a smart polymer is attached somehow to a lipid membrane, the transition in the macromolecule affects the properties of the membrane a nd ren ders t he s ystem s ensitive to en vironmental changes. To attach a smart polymer to a lipid membrane, a suitable “anchor,” which could be incorporated in the membrane, should be introduced into the macromolecule. Ā is could be achieved

43722_C024.indd 23

by copolymerizing poly(NIPAAM) with comonomers that have large h ydrophobic t ails suc h a s N,N-didodecylacrylamide [91], using a l ipophilic r adical i nitiator [ 92] mo difying c opolymers [93], or polymers that have terminally active groups [94] with a phospholipid. A lternatively, s mart po lymers ha ve bee n co valently coupled to t he ac tive g roups i n t he hydrophilic heads of the lipid-forming membrane [95]. Interesting and practically relevant materials for studying the behavior o f sm art p olymers at tached to l ipid memb ranes, a re liposomes, self-assembled 50–200 nm vesicles that have one or more (phospho)lipid bilayers, which encapsulate a fraction of the solvent. Liposomes are stable in aqueous suspension due to t he repulsive forces that appear when two liposomes approach each other. L iposomes a re w idely u sed f or d rug del ivery a nd i n cosmetics [96]. Ā e results of a tem perature-induced conformational transition of a smart polymer on the liposomal surface depend signicantly o n t he  uidity o f t he l iposomal memb rane. W hen t he membrane i s i n a  uid s tate at tem peratures b oth a bove a nd below t he p olymer t ransition tem perature, t he c ollapse o f t he polymer molecule forces anchor groups to move closer together by lateral diffusion within the membrane. Ā e compact globules of c ollapsed p olymer c over o nly a sm all pa rt o f t he l iposomal surface. Such liposomes have a low tendency to aggregate because the most of their surface is not c overed by the polymer. Naked surfaces contribute to t he repulsion between liposomes. On the other hand, when the liposomal membrane is in a s olid state at

10/3/2008 8:58:28 PM

24-24

temperatures b oth a bove a nd b elow t he p olymer t ransition temperature, the lateral diff usion of anchor groups is impossible, and the collapsed polymer cannot adopt a compact globule conformation b ut s preads o ver t he mo st o f t he l iposomal su rface [97]. Liposomes whose surfaces are covered to a l arge degree by a collapsed polymer repel each other less efficiently than intact liposomes. Ā e s tability o f a l iposomal su spension i s t hereby decreased, and aggregation and fusion of liposomes takes place, which is o ften ac companied b y t he rele ase o f t he l iposomal content into the surrounding medium [98]. When t he liposomal membrane is perturbed by t he conformational t ransition of the polymer, both the aggregation tendency and liposomal permeability for incorporated substances are aff ected. Poly(ethacrylic acid) u ndergoes a t ransition f rom an e xpanded to a c ompact c onformation i n t he ph ysiological pH r ange o f 7 .4–6.5 [ 99]. Ā e p H-induced t ransition o f poly(ethacrylic acid) covalently coupled to t he surface of liposomes f ormed f rom ph osphatidylcholine re sults i n l iposomal reorganization in to m ore c ompact mi celles a nd c oncomitant release of the liposomal content into the external medium. Ā e temperature-induced t ransition o f p oly(NIPAAM-co-N,Ndidocecylacrylamide) [ 100] o r p oly(NIPAAM-co-octadecylacrylate) [ 101], i ncorporated i nto t he l iposomal memb rane, enhanced the release of the uorescent marker, calcein, encapsulated in copolymer-coated liposomes. Liposomes hardly release any marker at temperatures below 32°C (the polymer transition temperature), w hereas t he l iposomal c ontent i s rele ased c ompletely within less than 1 min at 40°C. To increase the speed of liposomal response to tem perature change, the smart polymer was attached to the outer and inner sides of the lipid membrane. Ā e p olymer b ound only to t he outer su rface i f t he l iposomes were treated with the polymer after liposomal formation. When the l iposomes w ere f ormed d irectly f rom t he l ipid–polymer mixture, the polymer was present on both sides of the liposomal membrane [91]. Changes of liposomal surface properties caused by polymer collapse a ffect l iposomal i nteraction w ith c ells. L iposomes modied by a pH-sensitive polymer, partially succinilated poly(glycidol), deliver calcein into cultured kidney cells of the African green monkey more efficiently compared to liposomes not treated with the polymer [102]. Polymeric micelles formed by smart polymers and liposomes modied by smart polymers could b e u sed f or t argeted d rug del ivery. P olymeric m icelles have been prepared from amphiphilic block copolymers of styrene (forming a h ydrophobic c ore) a nd N IPAAM (forming a thermosensitive outer shell). Ā e polymeric micelles were very stable in aqueous media and had long blood circulation because of small diameter, unimodal size distribution (24 ± 4 nm), and, a lo w c ritical m icellar c oncentration o f a round 10 µg/mL. At temperatures above the polymer transition temperature (32°C), the polymer chains t hat form a n outer shell collapse, become more h ydrophobic, a nd a llow a ggregation b etween m icelles and favoring binding interactions with the surface of cell membranes. Ā us, h ydrophobic mole cules i ncorporated i nto t he micelles are delivered into the cell membranes. Ā ese micelles

43722_C024.indd 24

Smart Materials

are capable of site-specic delivery of drugs to the sites as temperature changes, for example, to inammation s ites o f increased temperature [103]. 24.3.3.8 Smart Polymers in Bioanalytical Systems Because sm art p olymers c an re cognize sm all c hanges i n en vironmental p roperties a nd re spond to t hem i n a p ronounced way, they could be used directly as sensors of these changes, for example, a series of polymer solutions that have different LCSTs could be used as a simple thermometer. As salts promote hydrophobic interactions and decrease the LCST, the polymer system could “sense” the salt concentration needed to decrease the L CST b elow ro om tem perature. A p oly(NIPAAM)-based system that can sense NaCl concentrations above 1.5% was patented [104]. Ā e response of the polymer is controlled by a ba lance o f h ydrophilic a nd h ydrophobic i nteractions i n t he macromolecule. U sing a re cognition elemen t t hat c an s ense external stimuli and translate the signal into the changes of the hydrophilic balance of the smart polymer, the resulting system presents a s ensor for t he stimulus. If t he conjugate of a sm art polymer and a recognition element has a transition temperature T1 in t he absence a nd T2 in the presence of stimuli,  xing the temperature T i n t he r ange T1 < T < T2 al lows achieving the transition o f a sm art p olymer i sothermally b y t he e xternal stimulus [105]. A n e xample of suc h a s ensor w as c onstructed using trans–cis isomerization of the azobenzene chromophore when irradiated by UV light. Ā e transition is accomplished by an increase in the dipole moment of azobenzene from 0.5 D (for the trans-form) to 3.1 D (for the cis-form) and hence a signicant decrease o f hydrophobicity. I rradiation w ith U V l ight re sults in increasing the LCST from 19.4°C to 26.0°C for the conjugate of t he chromophore w ith poly(NIPAAM). Ā e solution of t he conjugate is turbid at 19.4°C < T < 26.0°C, but when irradiated, the conjugate dissolves because the cis-form is below the LCST at this temperature. Ā e system responds to UV light by transition from a t urbid to t ransparent solution. Ā e termination of U V irradiation re sults i n a s low re turn of t he s ystem to i ts i nitial turbid s tate [ 105]. A f ew ot her l ight-sensitive s ystems w ere proposed that use different chromophores: triarylmethylcyanide [106] and leuconitriles [107]. Ā e h ydrophobicity o f t he re cognition mole cule w as a lso changed b y ch emical s ignals. P oly(NIPAAM) co ntaining 11.6 mol% of crown ether 9 has a LCST of 31.5°C in the absence of Na+ or K+ ions, 32°C in the presence of Na+, and 38.9°C in the presence of K+. Ā us, t he introduction of both Na+ a nd K+ ions leads to the dissolution of the insoluble polymer at that temperature. At 37°C, this effect is achieved only by K+ ions [108]. From b etter un derstanding o f li gand–host in teractions a nd development of new highly selective binding pairs (e.g., by using combinatorial l ibraries to f ind l igands o f h igh a ffinity f or particular biomolecules), one could expect t hat smart polymer systems will be used as “signal ampliers” to visualize a physicochemical event, which takes place in a recognition element, by a pronounced change in the system—conversion of a transparent solution into a turbid one or vice versa.

10/3/2008 8:58:28 PM

24-25

Smart Polymers for Biotechnology and Elastomers

Antibody–antigen in teractions p resent n early i deal a nalytical s electivity a nd s ensitivity de veloped b y nat ure. N ot surprisingly, they are increasingly used for a broad variety of bioanalytical applications. D ifferent a nalytical formats h ave been developed. The common feature of t he most of t hem is the requirement for separating an antibody–antigen complex from a nonbound antibody or antigen. Traditionally, the separation i s ac hieved b y c oupling o ne o f t he c omponents o f antibody–antigen pair to a solid support. The binding step is followed by washing nonbound material. Interactions of the soluble partner of t he binding pair with t he partner coupled to t he supp ort a re o ften ac companied b y u ndesired d iffusional l imitations, a nd he nce, i ncubation t imes of s everal hours are required for analysis. Because smart polymers can undergo t ransition f rom t he s oluble to t he i nsoluble s tate, they a llow combining t he advantages of homogeneous binding a nd, a fter t he phase t ransition of t he smart polymer has taken place, easy separation of t he polymer precipitate f rom the sup ernatant. T he e ssential f eatures o f a n i mmunoassay that uses smart polymers (named PRECIPIA) are as follows. The co valent co njugate o f po ly(NIPAAM) w ith m onoclonal antibodies to the κ-chain of human immunoglobulin G (IgG) are incubated for 1 h at room temperature (below the LCST of the conjugate), a nd t he Ig G solution is a nalyzed. T hen plain poly(NIPAAM) ( to f acilitate t hermoprecipitation o f p olymer–antibody c onjugates) a nd f luorescently l abeled mo noclonal antibodies to the γ-chain of human IgG are added. The temperature is raised to 45°C, the precipitated polymer is separated by c entrifugation, a nd  uorescence is measured i n t he supernatant [109]. Immunoassay systems that use temperatureinduced precipitation of poly(NIPAAM) conjugates with monoclonal antibodies are not inferior in sensitivity to the traditional heterogenous immunoassay methods, but because the antigen– antibody interaction takes place in solution, the incubation can be shortened to about 1 h [110,111]. Ā e limitations of PRECIPIA as an immunoassay technique are essentially the same as those of a ffinity precipitation, namely, nonspecific coprecipitation of analyzed pr otein w hen p oly(NIPAAM) pre cipitates. Polyelectrolyte complexes that have a low degree of nonspecic protein c oprecipitation h ave a lso b een suc cessfully u sed a s reversibly s oluble c arriers f or P RECIPIA-type i mmunoassays [112]. Ā e conjugate of antibody and polyanion poly(methacrylic acid) binds to the antigen within a few minutes, and the polymer hardly exerts any effect on the rate of antigen–antibody binding. Subsequent addition of a polycation, poly(N-ethyl-4-vinyl-pyridinium b romide) i n c onditions w here t he p olyelectrolyte i s insoluble, results i n quantitative precipitation of t he a ntibody– polymer conjugate within 1 min. Ā e total assay time is less than 15 min [10]. In p rinciple, P RECIPIA-type i mmunoassays c ould b e used f or si multaneous a ssay o f d ifferent a nalytes i n o ne sample, provided that conjugates specific toward these analytes are coupled covalently to different smart polymers that have d ifferent p recipitating c onditions, f or e xample, p recipitation of one conjugate by adding a polymeric counterion

43722_C024.indd 25

followed by thermoprecipitation of the second conjugate by increasing temperature. 24.3.3.9

Reversibly Soluble Biocatalysts

Ā e transition between the soluble and insoluble state of stimuliresponsive polymers has been used to develop reversibly soluble biocatalysts. A re versibly s oluble b iocatalyst c atalyzes a n en zymatic re action i n a s oluble s tate a nd h ence c ould b e u sed i n reactions of insoluble or poorly soluble substrates or products. As soon as the reaction is completed and the products are separated, the conditions (pH, temperature) are changed to promote precipitation of the biocatalyst. Ā e precipitated biocatalyst is separated and can be used in the next cycle after dissolution. Ā e reversibly soluble b iocatalyst ac quires t he adv antages o f i mmobilized enzymes (ease of separation from the reaction mixture after the reaction is completed and the possibility for biocatalyst recovery and re peated u se i n m any re action c ycles) but at t he s ame t ime overcomes the disadvantages of enzymes immobilized onto solid matrices such a s d iffusional l imitations a nd t he i mpossibility of using them in reactions of insoluble substrates or products. Biocatalysts t hat a re re versibly s oluble a s a f unction o f p H have been obtained by the covalent coupling of lysozyme to alginate [113]; of trypsin to poly(acrolein-co-acrylic acid) [114]; and of c ellulase [115]; a mylase [115]; α-chymotrypsin, a nd papa in [116] to poly(methylmethacrylate-co-methacrylic acid). A reversibly soluble cofactor has been produced by the covalent binding of NAD to alginate [117]. Reversibly soluble α-chymotrypsin, penicillin acylase, and alcohol dehydrogenase were produced by coupling to t he p olycation c omponent o f p olyelectrolyte c omplexes f ormed b y po ly(methacrylic aci d) a nd po ly(N-ethyl-4vinyl-pyridinium bromide) [118]. Biocatalysts t hat a re re versibly s oluble a s a f unction of temperature have been obtained by the covalent coupling of α-chymotrypsin a nd p enicillin ac ylase to a pa rtially h ydrolysed poly(N-vinylcaprolactam) [ 119], a nd o f t rypsin [ 120], a lkaline phosphatase [121], α-chymotrypsin [122], and thermolysin [123,124] to NIPAAM copolymers that contain active groups suitable for covalent co upling o f b iomolecules. L ipase wa s co upled t o a g raft copolymer composed of NIPAAM grafts on a poly(acrylamide-coacrylic acid) copolymer [125]. No signicant differences in biocatalytic properties were found for α-amylase coupled to poly(NIPAAM) via single-point or multipoint mode. Both enzyme preparations demonstrated increased t hermostability and t he absence of diffusional limitation when hydrolyzing starch, a high molecular weight substrate [126]. Ā e temperature of a p rotein–ligand interaction was controlled by site-directed coupling of terminally modied poly(NIPAAM) to a specically constructed site (close to a biotin binding site) on a genetically modied streptavidin [127]. Biocatalysts, w hich a re re versibly s oluble a s a f unction o f Ca 2+ concentration, were produced by covalent coupling of phosphoglyceromutase, enolase, peroxidase, and pyruvate kinase to αs1-casein. Ā e enzyme casein conjugates are soluble at a C a 2+ concentration below 20 mM but precipitate completely at a Ca 2+ concentration above 5 0 mM. Ā e precipitate re dissolves when EDTA, a strong Ca 2+-binding agent is added [128].

10/3/2008 8:58:28 PM

24-26

Ā e reversible occulation of lattices has been used to produce thermosensitive re versibly s oluble ( more p recisely re versibly dispersible) b iocatalysts u sing t rypsin [129], papa in [130], a nd α-amylase [131]. Lattices sensitive to a magnetic eld have been used to immobilize trypsin and β-galactosidase [132]. Liposomes that have a polymerized membrane, which reversibly aggregates on changing salt concentration, have been used to immobilize α-chymotrypsin [133]. Ā e most attractive application of reversibly soluble biocatalysts is repeated use in a reaction, which is difficult or even impossible to carry out using enzymes immobilized onto insoluble matrices, for e xample, h ydrolysis of w ater-insoluble ph lorizidin [ 134]; hydrolysis of high molecular weight substrates such as casein [123,130] and starch [115]; hydrolysis of insoluble substrates such as cellulose [135] and raw starch (corn our) [7,134,136–138]; production of insoluble products such as peptide, benzyloxycarbonyll -tyrosyl-Nω -nitro-l -arginine [116] and phenylglycine [139]. Ā e hydrolytic cleavage of corn our to glucose is an example of successfully using a reversibly soluble biocatalyst, amylase coupled to po ly(methylmethacrylate-co-methacrylic acid), in an industrially interesting process [136]. Ā e reaction product of the process, glucose, i nhibits t he hydrolysis. Ā e u se of a re versibly s oluble biocatalyst i mproves t he e fficiency o f t he h ydrolysis, w hich i s carried out at p H 5, at w hich t he a mylase–polymer conjugate is soluble. A fter each 24 h, the pH is reduced to 3.5, the unhydrolyzed solid residue and the precipitated conjugate are separated by centrifugation, the conjugate is resuspended in a fresh portion of the substrate at pH 5, and the hydrolysis is continued. Ā e conversion achieved after ve cycles is 67%, and the activity of the amylase after the fth cycle was 96% of the initial value [136].

24.4 Conclusion In the future, one looks forward to further developments and the commercial introduction of new smart polymers whose transition tem peratures a nd p H a re c ompatible w ith ph ysiological conditions or conditions for maximal stability of target biomolecules, suc h a s tem peratures o f 4 –15°C a nd p H v alues o f 5 –8. Additional prospects w ill stem f rom a b etter u nderstanding of the me chanism o f c ooperative i nteractions i n p olymers a nd increasing k nowledge o f s tructure–property c orrelations t o enable rational synthesis of smart polymers that have predened properties. D ue to t he p ossibility o f c ombining a v ariety o f biorecognition or biocatalytic systems and the unique features of smart p olymers, e xpectations a re r unning h igh i n t his a rea. Only t ime a nd mo re e xperimentation w ill de termine w hether smart polymers will live up to their generous promises.

Smart Materials

3. 4.

5. 6. 7. 8. 9.

10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29.

References 1.

R. Dagani, Chem. Eng. News, pp. 30–33, 1995. 2. H. B rønsted a nd J . K opecek, in Polyelectrolyte G els. P roperties, Pr eparation, a nd Applications, R .S. H arland a nd R.K. P rud’homme, e ds., American Chemical S ociety, Washington, DC, 1992, pp. 285–304.

43722_C024.indd 26

30. 31.

V.A. Kabanov, Polym. Sci. 36: 143–156, 1994. H.G. Schild, Prog. Polym. Sci. 17: 163–249, 1992. E.Y. Kramarenko and A.R. Khokhlov, Polym. Gels Networks 6: 46–56, 1998. Eudragit, Rö hm P harma GMBH I nformation M aterials, 1993. K. H oshino, H. Yamasaki, C. Chida, S. M orohashi, M. Taniguchi, a nd M. F ujii, J. C hem. E ng. J pn. 30: 30–37, 1997. E. Linné-L arsson a nd B . M attiasson, Biotechnol. T ech. 8: 51–56, 1994. S.Y. P ark a nd Y.H. B ae, Macromol. R apid C ommun. 20: 269–273, 1999. V.A. I zumrudov, I.Y u. Galae v, a nd B . M attiasson, Bioseparation 7: 207–220, 1999. M.B. Dainiak, V.A. Izumrudov, V.I. Muronetz, I.Yu. Galaev, and B. Mattiasson, Bioseparation 7: 231–240, 1999. U. Dissing and B. Mattiasson, J. Biotechnol. 52: 1–10, 1996. A.A. T ager, A.P. Sa fronov, S.V. S harina, a nd I.Y u. Galae v, Colloid Polym. Sci. 271, 1993. M. H eskins a nd J .E. G uillet, J. M acromol. S ci. Chem . A2: 1441–1455, 1968. K. Suwa, K. Morishita, A. Kishida, and M. Akashi, J. Polym. Sci. Part A: Polym. Chem. 35: 3087–3094, 1997. I. Yu. Galaev and B. Mattiasson, Enzyme Microb. Technol. 15: 354–366, 1993. F. Liu, F.H. Liu, R.X. Zhuo, Y. Peng, Y.Z. Deng, and Y. Zeng, Biotechnol. Appl. Biochem. 21: 257–264, 1995. S. Mori, Y. Nakata, and H. Endo, Protein Expression Purif. 5: 151–156, 1994. F.M. Winnik, A.D. D avidson, G.H. H amer, a nd H. K itano, Macromolecules 25: 1876–1880, 1992. G. Chen and A.S. Hoffman, 373: 49–52, 1995. M. M aeda, C. N ishimura, A. I nenaga, a nd M. T akagi, Reactive Functional Polym. 21: 27–35, 1993. D. U meno, T . M ori, a nd M. M aeda, Chem. C ommun. 1433–1434, 1998. B. Mattiasson, A. Kumar, and I.Yu. Galaev, J. Mol. Recognition 11: 211–216, 1998. A. Kumar, I.Yu. Galaev, and B. Mattiasson, Biotechnol. Bioeng. 59: 695–704, 1998. A. Durand and D. Hourdet, Polymer 40: 4941–4951, 1999. S.J. Lee and K. Park, J. Mol. Recognition 9: 549–557, 1997. K.-Y.A. Wu a nd K.D . Wisecarver, Biotechnol. Bi oeng. 39: 447–449, 1992. E. Kokufuta and S. Matsukawa, Macromolecules 28: 3474– 3475, 1995. S. K itano, I. H isamitsu, Y. K oyama, K. K ataoka, T. Oka no, M. Yokoyama, a nd Y. Sak urai, P roc. 1st I nternational Conference on Intelligent Materials, 1993, pp. 383–388. T. Yakushiji, K. Sakai, A. Kikuchi, T. Aoyagi, Y. Sakurai, and T. Okano, Langmuir 14: 4657–4662, 1998. H. Kawaguchi, in Biomedical Functions and Biotechnology of Natural and Articial Polymers, M. Yalpani, ed., ATL Press, 1996, pp. 157–168.

10/3/2008 8:58:29 PM

Smart Polymers for Biotechnology and Elastomers

32. H. Kubota and N. Shiobara, Reactive Functional Polym. 37: 219–224, 1998. 33. Y.M. Lee and J.K. Shim, Polymer 38: 1227–1232, 1997. 34. H.A. VonRecum, S.W. Kim, A. Kikuchi, M. Okuhara, Y. Sakurai, and T. Okano, J. Biomed. Mater. Res. 40: 631–639, 1998. 35. Y. Sun, K. Yu, X.H. Jin, and X.Z. Zhou, Biotechnol. Bioeng. 47: 20–25, 1995. 36. I. Yu. Galae v, M.N. G upta, a nd B . M attiasson, Chemtech 19–25, 1996. 37. A. Kumar, I.Yu. Galaev, and B. Mattiasson, Bioseparation 7: 185–194, 1999. 38. D. Guoqiang, M.A.N. Benhura, R. Kaul, and B. Mattiasson, Biotechnol. Prog. 11: 187–193, 1995. 39. D. G uoqiang, R . B atru, R . K aul, M.N. G upta, a nd B. Mattiasson, Bioseparation 5: 339–350, 1995. 40. E. Linné-L arsson, I.Y u. Galae v, a nd B . M attiasson, Bioseparation 6: 283–291, 1996. 41. A.P. Bradshaw and R.J. Sturgeon, Biotechnol Tech. 4: 67–71, 1990. 42. M.B. Dainiak, V.A. Izumrudov, V.I. Muronetz, I.Yu. Galaev, and B. Mattiason, Anal. Biochem. (submitted). 43. M.B. Dainiak, V.A. Izumrudov, V.I. Muronetz, I.Yu. Galaev, and B . M attiason, Biochim. B iophys. Acta 1381: 279–285, 1998. 44. A. Kondo, T. Kaneko, and K. Higashitani, Biotechnol. Bioeng. 44: 1–6, 1994. 45. C.S. Chen, C.K. Lee, C.Y. Chen, and M.J. Yeh, Colloids Surf. B: Biointerfaces 6: 37–49, 1996. 46. K. H oshino, M. T aniguchi, T . K itao, S. M orohashi, a nd T. Sasakura, Biotechnol. Bioeng. 60: 568–579, 1998. 47. P.-Å. Albertson, Partition o f C ell P articles a nd M acromolecules. Wiley, New York, 1986. 48. P. H ughes a nd C.R . L owe, Enzyme M icrob. T echnol. 10: 115–122, 1988. 49. G. Johansson and F. Tjerneld, in Highly Selective Separations in Bi otechnology, G. S treet, e d., Black ie Academic & Professional, London, 1994, pp. 55–85. 50. M. L u, P .-Å. Albertsson, G. J ohansson, a nd F . T jerneld, J. Chromatogr. B 680: 65–70, 1996. 51. T.T. F ranko, I.Y u. Galae v, R . H atti-Kaul, N. H olmberg, L. Bülow, and B. Mattiasson, Biotechnol. Tech. 11: 231–235, 1997. 52. J. Persson, H.-O. Johansson, I.Yu. Galaev, B. Mattiasson, and F. Tjerneld, Bioseparation 9: 105–116, 2000. 53. Y.G. T akei, T . Aoki, K. Sa nui, N. Oga ta, Y. Sak urai, a nd T. Okano, Macromolecules 27: 6163–6166, 1994. 54. T. Kondo, M. Koyama, H. Kubota, a nd R . K atakai, J. Ap pl. Polym. Sci. 67: 2057–2064, 1998. 55. F.-J. S chmitt, C. P ark, J . S imon, H. Rin gsdorf, a nd J. Israelachvili, Langmuir 14: 2838–2845, 1998. 56. P.W. Zhu and D.H. Napper, J. Phys. Chem. B 101: 3155–3160, 1997. 57. N. Yamada, T. Okano, H. Sakai, F. Karikusa, Y. Sawasaki, and Y. Sakurai, Makromol. Chem. Rapid Commun. 11: 571–576, 1990.

43722_C024.indd 27

24-27

58. T. Okano, N. Yamada, M. Okuhara, H. Sakai, and Y. Sakurai, Biomaterials 16: 297–303, 1995. 59. T. Oka no, A. K ikuchi, Y. Sak urai, Y.G. Takei, T. Aoki, a nd N. Ogata, J. Controlled Release 36: 125–133, 1995. 60. M. Yamato, M. Okuhara, F. Karikusa, A. Kikuchi, Y. Sakurai, and T. Okano, J. Biomed. Mater. Res. 44: 44–52, 1999. 61. Y. I to, G. Chen, Y. G uan, a nd Y. I manishi, Langmuir 13: 2756–2759, 1997. 62. T. B ohanon, G. Elender , W. K noll, P. K oeberle, J . L ee, A. Offenhaeusser, H. Rin gsdorf, E. Sac kmann, a nd J . S imon, J. Biomater. Sci. Polym. Ed., 8: 19–39, 1996. 63. H. K anazawa, Y. K ashiwase, K. Yamamoto, Y. M atsushima, A. Kikuchi, Y. Sakurai, and T. Okano, Anal. Chem. 69: 823– 830, 1997. 64. K. Hosoya, K. Kimata, T. Araki, and N. Tanaka, Anal. Chem. 67: 1907–1911, 1995. 65. H. Go, Y. Sudo, K. Hosoya, T. Ikegami, and N. Tanaka, Anal. Chem. 70: 4086–4093, 1998. 66. H. K anazawa, Y. M atsushima, a nd T. Oka no, Trends A nal. Chem. 17: 435–440, 1998. 67. H. Yoshioka, M. Mikami, T. Nakai, and Y. Mori, Polym. Adv. Technol. 6: 418–420, 1995. 68. B. M attiasson a nd I. Yu. Galae v, in High R esolution Chromatography, P .A. Millner , e d., IRL P ress, 1996, pp. 153–190. 69. I.Yu. Galaev, C. Warrol, and B. Mattiasson, J. Chromatogr. A 684: 37–43, 1994. 70. Y. Ito, Y.S. Park, and Y. Imanishi, Macromol. Rapid. Commun. 18: 221–224, 1997. 71. Y. Ito, Y. Ochiai, Y.S. Park, and Y. Imanishi, J. Am. Chem. Soc. 119: 1619–1623, 1997. 72. Y. Okaha ta, K. Ozaki , a nd T . S eki, J. Chem. S oc. Chem. Commun. 519–521, 1984. 73. Y. Okaha ta, H. N oguchi, a nd T. S eki, Macromolecules 19: 493–494, 1986. 74. T. Nonaka, K. Hashimoto, a nd S. Kurihara, J. Appl. Polym. Sci. 66: 209–216, 1997. 75. S. Akerman, P. Viinikka, B. Svarfvar, K. Putkonen, K. Jarvinen, K. Kontturi, J. Nasman, A. Urtti, and P. Paronen, Int. J. Pharm. 164: 29–36, 1998. 76. R. Spohr, N. Reber, A. Wolf, G.M.Alder, V. Ang, C.L. Bashford, C.A. P asternak, O . H ideki, a nd M. Yoshida, J. C ontrolled Release 50: 1–11, 1998. 77. H. Omichi, M. Yoshida, M. Asano, N. Nagaoka, H. Kubota, R. Katakai, R. Spohr, N. Reber, A. Wolf, G.M. Alder, V. Ang, C.L. Bashford, and C.A. Pasternak, Nucl. In strum. Methods Phys. Res. Sect. B 131: 350–356, 1997. 78. S. Kurihara, Y. Ueno, and T. Nonaka, J. Appl. Polym. Sci. 67: 1931–1937, 1998. 79. Y. Liu, M. Zhao, D.E. B ergbreiter, and R .M. Crooks, J. Am . Chem. Soc. 119: 8720–8721, 1997. 80. T.G. P ark a nd A.S. H offman, Biotechnol. Pr og. 10: 82–86, 1994. 81. K. Hosoya, E. Sawada, K. Kimata, T. Araki, N. Tanaka, and J.M.J. Frechet, Macromolecules 27: 3973–3976, 1994.

10/3/2008 8:58:29 PM

24-28

82. Y. I manishi a nd Y. I to, Pure A ppl. Chem. 67: 2015–2021, 1995. 83. K. Ishihara and K. Matsui, J. Polym. Sci. Polym. Lett. e d. 2 4: 413–417, 1986. 84. S. Kitano, K. Kataoka, Y. Koyama, T. Okano, and Y. Sakurai, Makromol. Chem., Rapid Commun. 12: 227–233, 1991. 85. K. K ataoka, H. Mi yazaki, B . Mi yazaki, T . Oka no, a nd Y. Sakurai, J. Am. Chem. Soc. 120: 12694–12695, 1998. 86. S. Kitano, Y. Koyama, K. Kataoka, T. Okano, and Y. Sakurai, J. Controlled Release 19: 162–170, 1992. 87. D. Shiino, K. Kataoka, Y. Koyama, M. Yokoyama, T. Okano, and Y. Sakurai, J. Controlled Release 28: 317–319, 1994. 88. K. N akamae, T . Mi yata, A. J ikihara, a nd A.S. H offman, J. Biomater. Sci. Polymer Ed. 6: 79–90, 1994. 89. Y. M ori, M. Yoshioka, a nd T. M ukoyama, C osmetic bas es and cosmetics containing polymers with sol–gel transition temperature, Jpn. K okai T okkyo K oho JP 09227329 A2, Sep. 2, 1997. 90. K. Lindell , D octoral Ā esis, L und U niversity, S weden, 1996. 91. H. Hayashi, K. Kono, and T. Takagishi, Bioconjugate Chem. 10: 412–418, 1999. 92. H. Ringsdorf, J. Simon, and F.M. Winnik, in Colloid–Polymer Intercations. P articulate, Amphiphilic, a nd B iological S urfaces, P .L. Dub in a nd P . T ong, e ds., American Chemical Society, Washington DC, 1993, pp. 216–240. 93. X.S. Wu, A.S. Hoffman, and P. Yager, Polymer 33: 4659–4662, 1992. 94. K. Kono, R. Nakai, K. Morimoto, and T. Takagishi, Biochim. Biophys. Acta 1416: 239–250, 1999. 95. D.A. T irrell, in Pulsed a nd S elf-Regulated Dr ug D elivery, J. Kost, ed., CRC press, Boca Raton, FL, 1990, pp. 110–116. 9 6. D.D. Lasic, Trends Biotechnol. 16: 307–321, 1998. 97. H. Ringsdorf, J. Venzmer, and F.M. Winnik, Angew. Chem. Int. Ed. Engl. 30: 315–318, 1991. 98. H. Hayashi, K. Kono, and T. Takagishi, Bioconjugate Chem. 9: 382–389, 1998. 99. S. Sugaj, K. Nitta, N. Ohno, and H. Nakano, Colloid Polym. Chem. 261: 159–165, 1983. 100. K. K ono, A. H enmi, H. Yamashita, H. H ayashi, a nd T. Takagishi, J. Controlled Release 59: 63–75, 1999. 101. H. H ayashi, K. K ono, a nd T. Takagishi, Biochim. B iophys. Acta 1280: 127–134, 1996. 102. K. Kono, T. Igawa, and T. Takagishi, Biochim. Biophys. Acta. 1325: 143–154, 1997. 103. S. Cammas, K. Suzuki, C. Sone, Y. Sakurai, K. Kataoka, and T. Okano, J. Controlled Rel. 48: 157–164, 1997. 104. Y. Tomita, S ensors f or co ncentrations o f s alts, Jpn. K okai Tokkyo Koho JP 63206653 A2 Aug. 25, 1988. 105. M. Irie, in Responsive Gels: Volume Transitions II, K. Dusek, ed., Springer-Verlag, Berlin, 1993, pp. 50–65. 106. R. K röger, H. M enzel, a nd M.L. H allensleben, Macromol. Chem. Phys. 195: 2291–2298, 1994. 107. A. M amada, T . T anaka, D . K ungwatchakun, a nd M. I rie, Macromolecules 23: 1517–1519, 1990.

43722_C024.indd 28

Smart Materials

108. M. Irie, Y. Misumi, and T. Tanaka, Polymer 34: 4531–4535, 1993. 109. N. Monji and A.S. Hoffman, Appl. Biochem. Biotechnol. 14: 107–120, 1987. 110. Q.-Z. Zhu, F. Liu, J.-G. Xu, W. Su, and J. Huang, Anal. Chim. Acta 375: 177–185, 1998. 111. Y.G. T akei, M. M atsukata, T . Aoki, K. Sa nui, N. Oga ta, A. Kikuchi, Y. Sakurai, and T. Okano, Bioconjugate Chem. 5: 577–5582, 1994. 112. B.B. Dzantiev, A.N. Blintsov, A.F. Bobkova, V.A. Izumrudov, and A.B. Zezin, Dokl. Biochem. (in R ussian) 342: 549–552, 1995. 113. M. Charles, R.W. Coughlin, and F.X. Hasselberger, Biotechnol. Bioeng. 16: 1553–1556, 1974. 114. E. VanLeemputten and M. Horisberger, Biotechnol. Bi oeng. 18: 587–590, 1976. 115. M. Taniguchi, K. Hoshino, Y. Netsu, and M. Fujii, J. C hem. Eng. Jpn. 22: 313–314, 1989. 116. M. F ujimura, T. M ori, a nd T. Tosa, Biotechnol. Bi oeng. 29: 747–752, 1987. 117. R.W. C oughlin, M. Aizawa, a nd M. Cha rles, Biotechnol. Bioeng. 18: 199–208, 1976. 118. A.L. Margolin, V.A. Izumrudov, V.K. Svedas, and A.B. Zezin, Biotechnol. Bioeng. 24: 237–240, 1982. 119. Y.E. K irsh, I. Yu. Galae v, T.M. Karaputadze, A.L. Margolin, and V.K. Svedas, Biotechnology (Russian), 184–189, 1987. 120. H. Lee and T.G. Park, Biotechnol. Prog. 14(3): 508–516, 1999. 121. A.L. N guyen a nd J .H.T. L uong, Biotechnol. Bi oeng. 34: 1186–1190, 1989. 122. J.-P. Chen, J. Chem. Technol. Biotechnol. 73: 137–143, 1995. 123. F. Liu, G. Tao, and R. Zhuo, Polymer J. 25: 561–567, 1993. 124. K. H oshino, M. T aniguchi, H. K awaberi, Y. T akeda, S. M orohashi, a nd T . Sas akura, J. F erment. B ioeng. 83: 246–252, 1997. 125. S. T akeuchi, I. Omo daka, K. H asegawa, Y. M aeda, a nd H. Kitano, Makromol. Chem. 194: 1991–1999, 1993. 126. J .-P. Chen, D .-H. Ch u, a nd Y.-M. S un, J. Ch em. T echnol. Biotechnol. 69: 421–428, 1997. 127. P.S. S tayton, T . S himoboji, C. L ong, A. Chilk oti, G. Chen, J.M. Harris, and A.S. Hoffman, Nature 378: 472–474, 1995. 128. K. Ok umura, K. Ik ura, M. Yoshikawa, R . Sas aki, a nd H. Chiba, Agric. Biol. Chem. 48: 2435–2440, 1984. 129. H. Kitano, C. Yan, and K. Nakamura, Makromol. Chem. 192: 2915–2923, 1991. 130. A. K ondo, K. I mura, K. N akama, a nd K. H igashitani, J. Ferment. Bioeng. 78: 241–245, 1994. 131. T. K ondo, T . U rabe, a nd K. Yoshinaga, Colloids S urf. A: Physicochem. Eng. Aspects 109: 129–136, 1996. 132. A. Kondo and H. Fukuda, J. Ferment. Bioeng. 84: 337–341, 1997. 133. Y. S un, X.H. J in, X.-Y. D ong, K. Yu, a nd X.Z. Zho u, Appl. Biochem. Biotechnol. 56: 331–339, 1996. 134. K. Hoshino, M. Taniguchi, H. Ueoka, M. Ohkuwa, C. Chida, S. M orohashi, a nd T . Sas akura, J. F erment. B ioeng. 82: 253–258, 1996.

10/3/2008 8:58:29 PM

Smart Polymers for Biotechnology and Elastomers

135. M. Taniguchi, M. Kobayashi, and M. Fujii, Biotechnol. Bioeng. 34: 1092–1097, 1989. 136. K. Hoshino, M. Taniguchi, Y. Netsu, and M. Fujii, J. C hem. Eng. Jpn. 22: 54–59, 1989. 137. K. H oshino, M. T aniguchi, H. M arumoto, a nd M. F ujii, Agric. Biol. Chem. 53: 1961–1967, 1989.

43722_C024.indd 29

24-29

138. K. H oshino, M. T aniguchi, H. M arumoto, a nd M. F ujii, J. Ferment. Bioeng. 69: 228–233, 1990. 139. Yu.V. Galaev, Prikl. B ioxim. M ikrobiol. 30: 167–170, 1994. 140. I.Yu. Galae v a nd B . M attiasson, Trends B iotechnol. 17: 335–340, 1999.

10/3/2008 8:58:29 PM

43722_C024.indd 30

10/3/2008 8:58:29 PM

25 Vibration Control for Smart Structures 25.1 V ibration Control...................................................................................................................25-1 Introduction • Vibration Control of Flexible Structures • Vibration Control of Dynamic Systems

References .........................................................................................................................................25-14 25.2 Smart Materials for Sound and Vibration Control..........................................................25-16 Introduction • L inear Ā eory of Electrodynamics • Methods for Laminated-Type Smart Structures • Methods for Discrete Type Smart Structures • Potential Research Areas in Smart Structures • Closure

References .........................................................................................................................................25-21

25.1

Vibration Control

Seung-Bok Choi and Young-Min Han 25.1.1 Introduction Ā e goal of vibration control is to suppress unwanted vibration of various dy namic systems. Traditional v ibration control uses passive element to i ncrease st iffness or d amping. Howe ver, t he insatiable dem and f or h igh p erformance qu antied b y h ighspeed operation, h igh control accuracy, a nd lower energ y consumption h as t riggered v igorous re searches o n ac tive a nd semiactive v ibration c ontrol o f d istributed  exible structures and d iscrete s ystems. Numerous control st rategies for conventional electromagnetic actuators have been proposed and implemented to suppress unwanted vibration. However, the successful realization of electromagnetic actuators maybe sometimes very difficult u nder c ertain c onditions d ue to h ardware l imitations such a s s aturation a nd re sponse s peed. Ā is difficulty can be resolved b y em ploying sm art m aterial ac tuators i n v ibration control. A s i s we ll-known, sm art m aterial t echnology fe atures actuating capability, control capability, and computational capability [1]. Ā erefore, these inherent capabilities of smart materials can execute specic f unctions autonomously in response to changing environmental stimuli. A mong many smart material candidates, e lectrorheological (E R)  uids, magnetorheological (MR) uids, and piezoelectric materials are effectively exploited for vibration control in various engineering applications.

A viable vibration control algorithm can be optimally synthesized by integrating control strategies, actuating technology, and sensing technology, as shown in Figure 25.1. Ā e design philosophy p resented i n F igure 25. 1 c ontains a v ery l arge n umber o f decisions a nd design pa rameters for t he characteristics of controllers, actuators, and sensors. Furthermore, the designer seeking a g lobal opt imal solution for t he synthesis of a c losed-loop smart structure system must also address other crucial decisions concerning t he t ime del ay o f a h igh-voltage/current a mplier, the speed of the signal converter, and the microchip hardware of the c ontrol s oftware. Ā is a rticle i ntroduces v ibration c ontrol methodology of exible structures and various dynamic systems using smart material actuators. In Section 25.1.2, vibration control techniques for the exible structure are presented by utilizing smart mount systems fabricated from ER  uids, MR  uids, and piezoelectric materials. In Section 25.1.3, v ibration control of seat suspension and vehicle suspension systems under various road conditions is given by adopting ER damper and MR damper, respectively. In addition, piezoelectric shunt technique is presented to suppress vibration of a CD-ROM drive.

25.1.2 Vibration Control of Flexible Structures 25.1.2.1 ER Fluid-Based Mount Utilizing ER uids, signicant progress has been made in vibration control of exible structures. Ā is is typically accomplished by t wo diff erent me thods: E R  uid-based sm art s tructure a nd ER uid-based mount. Choi et al. initiated ER-uid-based smart 25-1

43722_C025.indd 1

10/3/2008 10:32:35 PM

25-2

Smart Materials

Start

Characteristics of desired control performance

Control strategies PID control Optimal control Sliding mode control Adaptive control

Actuation technologies

Sensing technologies

Electrorheological fluid Magnetorheological fluid Pieoelectric material Shape memory alloy

Piezoelectric material Fiber optics Accelerometer Strain gage

Control performance and characteristics Optimal combination and best performances Control accuracy and fastness Energy consumption and const effectiveness Easy implementation and practical feasibility Robustness to unstructured uncertainties

Acceptable performance

No

Yes Stop

FIGURE 25.1

Flow chart for synthesizing vibration control system using smart material actuators and sensors.

structure, which can control the stiff ness and energy dissipation characteristics o f t he s tructure [2]. Ā ey c ompleted a n e xperimental study of a variety of shear congurations based of sandwich beam structures. So far, numerous researches have been undertaken in this way [3–5]. Ā e idea of applying the ER uidbased mount to v ibration control has been initiated in automotive eng ineering app lications. P etek e t a l. p roposed ac tively controllable damping characteristics using ER engine mount [6]. Choi et al. experimentally analyzed performances of ER engine mount [7]. Ā is idea can be easily exploited in structural vibration control. Duclos proposed externally tunable hydraulic mount f eaturing E R  uid [ 8]. C hoi e t a l. adopte d E R mo unt system to vibration control of frame structure [9]. When the ER uid-based mount is used for vibration control, operating mode of t he mo unt c an b e c lassied b y t hree diff erent t ypes:  ow mode, s hear mo de, a nd s queeze mo de. I n t he  ow mo de, i t i s assumed that two electrodes are  xed, and hence vibration control is achieved by controlling the ow motion between two xed electrodes [10]. In the shear mode, it is usually assumed that one of two electrodes is free to translate or rotate relative to the other, and h ence v ibration c ontrol i s ac hieved b y c ontrolling s hear force b etween t he t wo ele ctrodes [11]. U nlike t he f ormer t wo modes, the electrode gap is varied in the squeeze mode and the ER uid is squeezed by a normal force. Ā e ER uid-based mount using this mode is very effective for vibration control of a exible

43722_C025.indd 2

structure sub jected to sm all a mplitude a nd h igh-frequency excitations [12]. Figure 25. 2 s hows t he s chematic c onguration an d p hotograph o f t he s queeze mo de E R  uid-based mo unt. Ā e lower electrode is xed to the base plate, while the upper electrode is to be moved up a nd down. Ā us, the squeeze-mode motion of the ER  uid o ccurs i n t he v ertical d irection. Ā e c oil s pring i s attached to support a static mass, which is the mass of a exible beam structure to be controlled. Ā e total force of the proposed ER mount can be obtained as follows: F (t ) = kh (t ) + c f (t )h
(t ) + f er (t )

(2

5.1)

In the above, k is the stiff ness constant of the coil spring, h(t) is exciting d isplacement, cf(t) i s t he d amping c oefficient in the absence o f t he ele ctric  eld. fer(t) i s t he c ontrollable d amping force o wing to t he ele ctric .  eld of E a nd c an b e e xpressed a s fer(t) = 4 pR 3 ty (E)sgn(h(t))/3(ho + h(t)). R and ho are radius of the circular electrode and initial gap between lower and upper electrodes, respectively. ty(E) is the eld-dependent yield stress, which is given by a E b . Here, a and b are intrinsic values of the ER uid to be experimentally determined. Figure 25. 3 s hows t he s chematic c onguration for v ibration control o f a  exible b eam s tructure u sing t he E R  uid-based mount. A c ontinuous a nd u niform b eam i s supp orted b y t wo

10/3/2008 10:32:35 PM

25-3

Vibration Control for Smart Structures

Electrodes Coil spring

Housing

ER fluid

(a)

FIGURE 25.2

(b)

Schematic configuration

Photograph

Squeeze mode ER uid-based mount.

Exciter ER mount

Flexible beam

Spring mount

Input field Rigid base

Force displacement acceleration

Dynamic signal analyzer

FIGURE 25.3

High-voltage amplifier

Computer (controller)

Experimental conguration for vibration control of a exible beam structure using ER uid-based mount.

spring mo unts a nd t wo E R  uid-based mo unts. W hen t he Bernoulli-Eulor b eam t heory i s app lied, t he k inetic energ y, potential energ y, a nd no nconservative w ork a re ob tained. By employing energy equations and Hamilton’s principle, the governing equations of mot ion for t ransverse deection y(x,t) a nd associated b oundary c ondition c an b e ob tained [12]. By u sing the Lagrange equation, a de coupled ordinary equation for each mode is derived by q

i (t ) + 2ς iω i q
i (t ) + ω i2qi (t ) = Q i (t )/I i + Q exi (t )/I i (2

43722_C025.indd 3

5.2)

where Ii is the generalized mass Qi(t) is the generalized force including controllable damping force Qexi(t) is the generalized force including exciting force qi, Vi, and w i are modal coordinate, damping ratio, and the natural frequency of the ith mode Considering the boundary conditions of the proposed structural system, t he  rst a nd second modes a re chosen to b e dominant for v ibration characteristics. Ā e control purpose is to re gulate

10/3/2008 10:32:38 PM

25-4

Smart Materials

unwanted v ibrations o f t he b eam s tructure w ith app ropriate control input voltage. Among m any p otential c ontroller c andidates, a n opt imal controller, which is very effective for vibration control is adopted [13]. In order to implement the controller, an experimental apparatus i s e stablished a s s hown i n F igure 25. 3. Ā e exible steel beam is ex ternally ex cited b y t he e lectromagnetic ex citer, a nd displacement sensor (noncontact proximitor) and accelerometer are used to catch the vibration signals. Ā e signals are fed back to the m icroprocessor t hrough a nalog/digital ( A/D) c onverter. Depending on t he sig nal i nformation, c ontrol voltage i s de termined by means of the optimal controller. Ā e control voltage is then applied to t he ER uid-based mount through high-voltage ampliers. On the other hand, the force transmissibility is measured by two force transducers: one for the input force and the other for the output force at mount position (spring or ER uidbased mount position). Figure 25.4a presents time responses of the vibration signals when the beam structure is excited by the rst mode natural frequency of 31.5 Hz [12]. It is clearly seen that displacement at the top of ER uid-based mount is substantially reduced b y ac tivating t he p roposed opt imal c ontroller. I t i s remarked that uncontrolled response is obtained in the absence of the control voltage. Acceleration levels at the top of ER mount is me asured a nd p resented i n F igure 25.4 b i n t he f requency domain [ 12]. I t i s c learly ob served t hat ac celerations at t he resonance f requencies a re remarkably at tenuated by ac tivating the control voltage to the ER-uid based mount. 25.1.2.2

MR Fluid-Based Mount

MR  uids h ave ve ry s imilar c haracteristics. Ā e y are suspensions consisting of magnetizable particles in a low-permeability base  uid i nstead o f p olarized pa rticles i n no nconducting o il. Ā eir rheological properties a re changed by applying magnetic eld to t he  uid domain i nstead of electric  eld. Furthermore, MR  uids u se lo w vol tage a nd h igh c urrent t hrough a c oil to generate a magnetic eld whereas ER uids use low current and high voltage to generate an electric eld. Ā e rheological change

of MR  uids are primarily observed as a sig nicant increase of the y ield s hear s tress o f t he  uid. It c an be co ntinuously co ntrolled by the intensity of applied magnetic eld. By choosing an appropriate operation mode, controllable pressure or shear force can be effectively applied to v arious dy namic systems. Like ER uid, MR  uid h as b een w idely app lied to v ibration c ontrol elds. Choi et al. investigated the vibration control performance of M R  uid dampers for vehicle applications [14]. Wereley a nd Pang studied hysteresis and plug ow models of MR uid dampers using approximate parallel plate models [15]. Stanway et al. studied v arious op eration mo des o f M R  uid appl ications for structural vibration control [16]. In this chapter, a compact MR uid-based mount, which consists o f M R d ashpot u nder m ixed mo de (ow m ode a nd s hear mode) and rubber element, is proposed for vibration control of a exible beam structure [17]. Its schematic conguration and photograph are presented in Figure 25.5. MR uid is lled in the gap between piston (or plunger) and outer housing. Ā e electromagnetic coil is wired inside of the outer cylindrical housing to produce magnetic eld. Ā e housing can be xed to the base structure, and the plunger is attached to the top end of the rubber element. Ā is rubber element has a role to support a static mass and isolate the v ibration t ransmission at t he no nresonant re gions. D uring the rel ative mot ion o f t he i nner c ylinder to o uter c ylindrical housing, t he M R  uid ows t hrough a nnular gap . Ā e crosssectional a rea o f t he p lunger i s re garded a s t he e ffective piston area. If a certain level of magnetic eld is applied through the gap, the MR uid-based mount produces an additional damping force due to the yield stress of the MR uid. Figure 25.6 shows the measured damping forces in time domain according to applied current [17]. Ā e MR  uid-based mount is placed between the load cell and electromagnetic exciter to me asure t he  eld-dependent damping force. When the shaker table moves up a nd down by a command sig nal gener ated f rom t he e xciter c ontroller, t he M R uid-based mount produces the damping force and it is measured by the load cell. It is seen from the results that the damping force increases as the input current increases.

Time domain

Frequency domain 50

100

Uncontrolled Controlled

40

50

Acceleration (m/s2)

Acceleration (m/s2)

Uncontrolled Controlled

0

−50

30 20 10 0

−100 2.0 (a)

2.1

2.2 Time (s)

2.3

−10

2.4 (b)

40

60 Frequency (Hz)

80

100

FIGURE 25.4 Vibration control responses of a exible beam structure using ER uid-based mount. (From Jung, W. J., Jeong, W. B., Hong, S. R., and Choi, S. B., J. Sound Vibration, 273, 185, 2004. With permission.)

43722_C025.indd 4

10/3/2008 10:32:39 PM

25-5

Vibration Control for Smart Structures

Rubber Flux guide Electromagnet Housing

MR fluid

Piston (a)

(b)

Schematic configuration

Photograph

FIGURE 25.5 Mixed mode MR uid-based mount.

uncontrolled a nd controlled acceleration a re measured at t he position of t he exciting mass about t he M R mount a nd beam structure b elow t he M R mo unt. Ā e c ontrol i nput i s de termined o n t he ba sis o f t he L QG c ontroller. Ā e a mplitude of excitation force is set by 1 N, and the frequency range focused

30 0.0 A

0.4 A

0.8 A

1.2 A

20

Force (N)

10 0

Acceleration at the mass above the MR mount

−10

3

0.1

0.2 Time (s)

0.3

0.4

FIGURE 2 5.6 Field-dependent d amping forc e of M R  uid-based mount. (From Hon g, S. R. and Choi, S B ., J. Intell. Mater. Syst. Struct., 16, 931, 2005. With permission.)

Ā is d irectly i ndicates t hat a c ertain de sired d amping f orce can be achieved by controlling the input current. Ā e manufactured MR uid-based mount is installed in a exible beam structure, w hich i s supp orted b y t wo r ubber mo unts at b oth en ds (refer to Figure 25.3). At the upside of the exible beam, the MR uid-based mount is placed between the beam and the exciting mass. I n order to supp ress v ibration of t he beam structure, a n optimal c ontroller i ntegrating w ith K alman  lter ( LQG) i s synthesized as follows: ∧

∧

i

∧

i

∧

∧

u(t ) = −(k1 q1 (t ) + k2 q1 (t ) + k3 q 2 (t ) ∧

i

+ k 4 q 2 (t ) + k 5 q 3 (t ) + k 6 q 3 (t ))

(25.3)

where ki is the .element of the state feedback gain matrix qˆi and qˆ i a re e stimated s tates, w hich c orrespond to t he modal displacement Figure 25.7 presents the measured acceleration responses of the s tructural s ystem w ith M R f luid-based mo unt [17]. Ā e

43722_C025.indd 5

2

1

0 0

20

(a)

40 Frequency (Hz)

60

80

Acceleration at the beam below the MR mount 3 Uncontrolled Controlled Accelerance (m/s2 N)

−30 0.0

Accelerance (m/s2 N)

Uncontrolled Controlled

−20

2

1

0 (b)

0

20

40

60

80

Frequency (Hz)

FIGURE 25.7 Vibration control responses of a exible beam structure using MR uid-based mount. (From Hong, S. R. and Choi, S. B., J. Intell. Mater. Syst. Struct., 16, 931, 2005. With permission.)

10/3/2008 10:32:40 PM

25-6

Smart Materials

is chosen from 5 to 80Hz. It is clearly observed that the acceleration levels of  rst, second, a nd t hird modes a re at tenuated by controlling the damping force of MR uid-based mount. 25.1.2.3 Piezoelectric Actuator-Based Mount Many types of passive mounts have been developed to support static lo ad a nd i solate u nwanted v ibration of  exible structure systems a nd d iscrete s ystems. Ā e r ubber mo unt i s o ne o f t he most p opular a nd e ffective pa ssive mounts applied for v arious vibrating systems. It is generally known that it has low damping, and hence shows efficient vibration isolation performance in the nonresonant a nd h igh-frequency e xcitations [ 18]. H owever, the passive mo unt h as a n i ntrinsic p erformance l imitation, which leads to t he study on active mounts. As well-known, the piezoactuator is featured by fast response time, small displacement, and low power consumption. Using these salient features, we can accomplish very effective vibration control performance of a f lexible s tructure sub jected to sm all m agnitude a nd high-frequency resonant excitations. Active vibration control of exible s tructures a nd d iscrete s ystems ut ilizing p iezoelectric materials has been studied in recent years by many researchers. Premont et al. p roved t he rob ustness a nd e ffectiveness of active damping of space s tructure u sing p iezoelectric ac tuator [ 19]. Choi et al. analyzed the performance of the hybrid mount, which is the combination of efficient mounts of different types, associated with semiactive ER uid and active piezoelectric actuator [5]. In this chapter, a new type of hybrid mount featuring the passive r ubber elemen t a nd ac tive p iezoactuator i s p resented i n order to ac hieve sup erior v ibration c ontrol p erformance o f a exible bea m st ructure a t bo th r esonant a nd n onresonant regions [20]. Figure 25.8 presents a schematic diagram and photograph o f t he p roposed p iezoelectric ac tuator-based mo unt. Ā e piezostack ac tuator is a b ipolar t ype a nd connected to t he rubber mount through the intermediate mass. Ā e intermediate mass acts like a re action mass for t he piezoelectric actuator on

the beam structure. In order to e valuate v ibration control performance of the proposed piezoelectric actuator-based mount, a exible structure is established as shown in Figure 25.9. A s teel beam is supported by two rubber mounts at both ends while the piezoelectric-based mo unt i s p laced b etween t he t wo r ubber mounts below the  exible beam. Ā e exible beam is excited by the electromagnetic exciter. Among numerous control strategies, a s liding mo de control (SMC) scheme, which has inherent robustness to system uncertainties a nd e xternal d isturbances, i s adopte d to i solate t he vibration of the  exible beam structure. As a  rst step, the sliding surface s(t) is dened as follows: s(t ) = Gx (t ) (2

5.4)

where G is the sliding surface gradient x(t) is state variable Ā e existence condition of the sliding mode motion is given by 1 d 2 s (t ) ≤ −η s(t ) 2 dt In this equation, h is a strictly positive constant. Ā is condition allows the state variable x(t) converge to t he sliding surface s(t). The s liding mo de c ontroller, w hich s atisfies t he e xistence condition of the sliding mode motion, is obtained by [20] u(t ) = −(GB)−1 (GAx (t ) + k sgn(s(t ))) (2

5.5)

where k is the discontinuous control gain sgn (·) is the sign function A and B are system matrix and input matrix, respectively, which come from state space realization of t he governing equation of

Housing Piezostack actuator

Prestress spring

Rubber Intermediate mass

(a)

FIGURE 25.8

43722_C025.indd 6

Schematic configuration

(b)

Photograph

Piezoelectric actuator-based mount.

10/3/2008 10:32:41 PM

25-7

Vibration Control for Smart Structures

Exciter

Rubber mount

Piezoelectric actuator-based mount

Flexible beam

Input field

Force displacement acceleration

High-voltage amplifier

Dynamic signal analyzer

FIGURE 25.9

Rigid base

Computer (controller)

Experimental conguration for vibration control of a exible beam structure using piezoelectric actuator-based mount.

Transmitted force 3

20

2 Transmitted force (N)

Displacement (µm)

Displacement 30

10 0 −10 −20 −30

(a)

0

1

2 Time (s)

3

1 0 −1 −2 −3

4

0

(b)

1

2 Time (s)

3

4

FIGURE 25.10 Time re sponses of a  exible b eam s tructure u sing pie zoelectric a ctuator-based mou nt. (From K im, S . H ., C hoi, S . B ., a nd Hong, S. R., Int. J. Mech. Sci., 46, 143, 2004. With permission.)

motion. F igure 25.10 p resents t he me asured t ime re sponses at the third mode frequency of 180 Hz [20]. We clearly observe substantial reductions of the displacement and transmitted force by activating the controller-associated piezoelectric actuator-based mount. Figure 25.11 shows the corresponding frequency response [20]. Ā e uncontrolled displacement of the beam at mount position i s 7.9 µm at the fourth mode (298 Hz). By activating the controller, the displacement is reduced to 2.1 µm. It is clear from the results that the proposed piezoelectric actuator-based mount is very effective for vibration control of a  exible structure. Ā e force transmissibility is also evaluated as shown in Figure 25.11b

43722_C025.indd 7

[20]. From the remarkable reduction of the force transmission to the ba se, i t i s a ssured t hat t he i mposed v ibration o f t he b eam structure has been effectively isolated by applying the piezoelectric actuator-based mount.

25.1.3 Vibration Control of Dynamic Systems 25.1.3.1

Seat Suspension Using ER Damper

Heavy c ommercial v ehicles a re t ypically e xposed to e xtended driving hours and severe working environment. Although many

10/3/2008 10:32:43 PM

25-8

Smart Materials Transmitted force

Displacement 3 Uncontrolled Controlled

20

Force transmissibility

Displacement (µm)

30

10

0

−10 160

200

(a)

240 Frequency (Hz)

280

Uncontrolled Controlled

2

1

0

−1 160

320 (b)

200

240 Frequency (Hz)

280

320

FIGURE 25.11 Frequency responses of a  exible beam structure using piezoelectric actuator-based mount. (From K im, S. H., Choi, S. B., and Hong, S. R., Int. J. Mech. Sci., 46, 143, 2004. With permission.)

efforts to improve ride quality have been made through primary and s econdary ( cabin) su spensions [ 21,22], d rivers s till su ffer from t he t ransmission o f lo w-frequency a nd h igh-amplitude vibration, which is a major factor in the ride quality [23,24]. Ā e alleviation of seat vibration is a signicant problem in the heavy commercial vehicle because its ride quality is directly related to health d isorders, work ing e fficiency, a nd s afety. A si mple a nd effective method for attenuating unwanted v ibration in a c ommercial v ehicle i s a s eat su spension s ystem t hat i s pa rticularly effective i n t he low-frequency range i n which vehicle v ibration energy is concentrated and a human body is sensitive [25]. Recently, a s emiactive s eat su spension s ystem f eaturing E R uids has been investigated for vibration isolation. Ā is type of the

semiactive seat suspension system has several advantages such as fast response time, continuously controllable damping force, and low energy consumption compared with active seat suspension systems. Ā e s chematic d iagram a nd photograph of t he E R s eat damper is shown in Figure 25.12. Ā e ER seat damper is divided into the upper and lower chambers by the piston, and is fully lled with the ER uid. Ā e ER  uid ows t hrough t he duct between inner and outer cylinders from one chamber to t he other by the motion o f t he p iston. A p ositive vol tage i s p roduced b y a h ighvoltage supply unit connected to the outer cylinder. Ā e gas chamber located outside of the lower chamber acts as an accumulator. If electric eld is applied, the ER damper produces a damping force caused only by uid resistance. However, if a c ertain level of the

ER fluid

Gas chamber Outer cylinder

Inner cylinder

Piston V

Voltage source (a)

Schematic configuration

(b)

Photograph

FIGURE 25.12 Controllable ER seat damper.

43722_C025.indd 8

10/3/2008 10:32:44 PM

25-9

Vibration Control for Smart Structures

electric  eld is supplied to the ER damper, it produces an additional damping force owing to the yield stress of the ER uid. Ā is damping force of t he E R d amper c an b e c ontinuously t uned by controlling the intensity of the electric eld. Ā e damping force of the ER damper can be obtained as follows [26]: F = k e x P + c e x
p + FER (2

Accelerometer LVDT 2

ER damper

5.6)

Voltage amplifier

where k e is the effective stiff ness due to gas pressure ce is the effective damping due to the uid viscosity xp is the excitation displacement FER is the eld-dependent damping force, which is tunable by a function of applied electric eld E that is typically Bingham model On t he ba sis of t he si ze a nd required d amping force level of a conventional passive oil seat damper for a commercial truck, an ER seat damper is designed and manufactured as shown in Figure 25.12b. Its eld-dependent damping force is experimentally evaluated that it can be increased up to 89 N. Ā e most signicant feature in the seat suspension system is that the seat suspension dynamics includes a seated human body. However, the seated human body exposed to vibration is a complex dynamic system. So far, various lumped parameter models have b een de veloped to de scribe t he h uman v ibration mot ion [27]. Ā ese mo dels c onsider t he h uman b ody a s s everal r igid bodies, s prings, a nd d ampers. I n t his c hapter, a h uman b ody model is derived and integrated with the governing equations of the E R s eat su spension s ystem. F or t he si mplication o f t he dynamic modeling, it is assumed that there exists only the vertical motion of the vehicle. Ā e seat suspension system consists of the E R s eat d amper, l inear s pring, s eat f rame, a nd c ushion on which the driver is seating. Ā e spring constant is held to be constant a nd t he d amping force varies according to c ontrol i nput. Ā e seat cushion is modeled by a set of linear springs and dampers whose characteristics are constant. A d river is modeled as a multisegment m ass, w hich i nclude h ead, upp er to rso, lo wer torso, and thighs [28]. Ā e arms and legs are combined with the upper torso and thigh, respectively. Because of uncertainties in the human vibration model, one of the potential controller candidates for t he ER seat suspension system is sliding mode controller, which is very effective f or no nlinear a nd pa rameter uncertain system [29]. By imposing semiactive actuating conditions, the SMC is experimentally realized with a state observer. Figure 25.13 shows an experimental setup to evaluate control performance of the seat suspension system using the ER damper [30]. In the experimental conguration, a driver directly sits on the controlled seat. Ā e oor under the seat is excited from two types of pro les; bump and random road conditions. Ā e excited oor displacement is calculated and then the displacement signal is c onverted to t he h ydraulic c ontrol u nit f or c ontrolling t he excitation input to the platform. When the vibration is transmitted to t he E R s eat su spension, t he dy namic re sponse sig nals, which are acquired from the linear variable differential transformer

43722_C025.indd 9

LVDT 1

Computer (controller) Hydraulic power unit

FIGURE 25.13 Experimental conguration for vibration control of a seat suspension using ER seat damper.

(LVDT) and accelerometer, are then converted to digital signals via A /D c onverter. Ā e S MC i s t hen ac tivated to re duce t he vibration level at t he driver seat. Figure 25.14 presents the measured transmissibility of the proposed ER seat suspension system [30]. The natural frequency of the ER seat suspension system without electric  eld is 2.1 Hz, which is the same as for the passive seat suspension system for commercial truck. It is clearly observed that the peak of transmissibility is substantially reduced b y ac tivating t he ele ctric  eld de termined f rom t he SMC, and its frequency is slightly increased. Figure 25.15 presents the measured bump responses of the seat suspension system in the time domain [30]. As clearly observed from the results, the maximum d isplacement a nd ac celeration a re re duced b y 4 0% and 3 0%, re spectively. Ā ese re sults clearly i ndicate sig nicant improvement of the ride quality. 25.1.3.2

Vehicle Suspension Using MR Damper

Recently, a great attention on the vibration control of a vehicle system has been signicantly increased. Ā e vehicle vibration needs to b e at tenuated f rom v arious ro ad c onditions. Ā is is normally ac complished b y em ploying su spension s ystem. S o far, t hree t ypes o f su spensions h ave b een p roposed a nd suc cessfully i mplemented; pa ssive, ac tive, a nd s emiactive. Ā e passive su spension s ystem f eaturing o il d amper p rovides design si mplicity a nd c ost-effectiveness. Ho wever, p erformance limitations are inevitable. On the other hand, the active suspension system provides high control performance in wide frequency range. However, the active suspension requires high power s ources, m any s ensors, s ervo v alves a nd s ophisticated control log ic. O ne w ay to re solve t hese re quirements o f t he active suspension system is to use the semiactive suspension

10/3/2008 10:32:44 PM

25-10

Smart Materials Seat displacement

Head acceleration 2.0

3.5 Uncontrolled Controlled Transmissibility

1.5 Transmissibility

Uncontrolled Controlled

3.0

1.0

0.5

2.5 2.0 1.5 1.0 0.5

0.0

0

1

2 Frequency (Hz)

(a)

3

0.0

4

0

1

(b)

2 Frequency (Hz)

3

4

FIGURE 25.14 Transmissibility of a seat suspension using ER seat damper. (From Han, Y. M., Jung, J. Y., Choi, S. B., and Wereley, N. M., Int. J. Mod. Phys. B, 19, 1689, 2005. With permission.)

Vertical displacement at seat frame

Vertical acceleration at driver’s head

Uncontrolled Controlled

30 15 0 −15 −30

(a)

6 Head acceleration (m/s2)

Seat displacement (mm)

45

0

1

2 Time (s)

2 0 −2 −4

3 (b)

Uncontrolled Controlled

4

0

1

2

3

Time (s)

FIGURE 25.15 Bump responses of a seat suspension using ER seat damper. (From Han, Y. M., Jung, J. Y., Choi, S. B., and Wereley, N. M., Int. J. Mod. Phys. B, 19, 1689, 2005. With permission.)

system. Ā e s emiactive su spension s ystem off ers a de sirable performance gener ally en hanced i n t he ac tive mo de w ithout requiring l arge p ower s ources a nd e xpensive h ardware. Recently, ve ry at tractive a nd e ffective s emiactive s uspension system M R  uid h as b een p roposed b y m any i nvestigators. Carlson e t a l. [ 31] p roposed a c ommercially a vailable M R damper, w hich i s app licable to o n-and-off-highway vehicle suspension s ystem. Ā ey e xperimentally dem onstrated t hat sufficient le vels o f d amping f orce a nd a lso sup erior c ontrol capability of the damping force by applying control magnetic eld could be achieved. Spencer Jr. et a l. [32] proposed dy namic model for prediction of d amping force of a M R d amper. Ā ey compared the measured damping forces with the predicted ones in time domain. Kamath et al. [33] proposed a semiactive M R l ag mo de d amper. Ā ey proposed dynamic model and veried its validity by comparing the predicted damping force with the measured one.

43722_C025.indd 10

In this chapter, the hardware-in-the-loop simulations (HILS) are u ndertaken to e valuate t he p erformance of t he f ull-vehicle MR suspension system [34]. It generally takes a long time and a high cost to suc cessfully develop t he new t ypes of components for a c omplete s ystem. To s ave t he cost a nd t ime, a t heoretical analysis using computer simulation method is widely used. However, si nce m any re al si tuations, w hich a re d ifficult to be modeled a nd, e ven, c annot b e mo deled b y a nalytical me thod, are often neglected or approximated by linearization, the theoretical method cannot precisely predict the performance of the system to occur in real eld. Ā erefore, in order to overcome the limit o f t he c omputer si mulation me thod, H ILS me thod h as been proposed recently. Ā e HILS method has major advantages such a s e asy m odication o f s ystem pa rameters a nd r elatively low-cost te st f acilities. I n add ition, a w ide r ange o f op erating conditions to emulate the practical situations can be investigated in the laboratory.

10/3/2008 10:32:45 PM

25-11

Vibration Control for Smart Structures

Electromagnet

Gas chamber Outer cylinder

MR fluid Inner cylinder

(a)

Piston

(b)

Schematic configuration

Photograph

FIGURE 25.16 Controllable MR damper.

Ā e schematic conguration and photograph of the proposed MR damper are shown in Figure 25.16. Ā e MR damper is divided into the upper and lower chambers by the piston, and it is fully lled with the MR uid. By the motion of the piston, the MR uid ows through the orice at both ends from one chamber to the other. Ā e gas chamber located outside acts as an accumulator for absorbing sudden pressure variation of lower chamber of the MR

damper i nduced by t he r apid mot ion of t he piston. I n order to simplify the analysis of the MR damper, it is assumed that the MR uid is incompressible and that pressure in one chamber is uniformly distributed. Furthermore, it is assumed that frictional force between oil seals and uid inertia is negligible. Figure 25.17 shows t he s chematic c onguration o f t he p roposed H ILS. I t i s divided into t hree parts: interface, hardware, and software. Ā e

Software simulation

Hardware implementation

Force

Full-vehicle suspension model

MR damper

Displacement

Current amplifier

Hydraulic power unit

Force

Excitation

Control input

Displacement

Interface

FIGURE 25.17 Schematic diagram for vibration control of a vehicle suspension using MR damper via HILS.

43722_C025.indd 11

10/3/2008 10:32:46 PM

25-12

Smart Materials

interface part is composed of computer on which digital signal processing (DSP) board is mounted. Ā e hardware part is composed of the MR damper, current amplier, and hydraulic damper tester. Ā e soft ware part consists of the theoretical model for the full-vehicle M R su spension s ystem a nd c ontrol a lgorithm p rogrammed in the computer. Ā e vehicle body itself is assumed to be rigid and has degrees of freedom in the vertical, pitch, and roll directions. It is connected to f our r igid bodies representing t he wheel u nsprung m asses i n w hich e ach h as a v ertical de gree o f freedom. In practice, the sprung mass is varied by the loading conditions such as the number of riding persons and payload. In order to take into account for these structured system uncertainties, t he lo op s haping de sign p rocedure ( LSDP) ba sed o n H∞ robust s tabilization p roposed b y M cFarlane a nd G lover [35] i s adopted in this article. By choosing a suitable g, the suboptimal controller is constituted as follows [34]: ⎡ A cse + γ 2 Wh∗−1ZC∗se (C se + Dse Fh ) γ 2 Wh∗−1ZC∗se ⎤ u = W⎢ ⎥ y (2 5.7) B∗se X −D∗se ⎢⎣ ⎥⎦ A cse = A se + Bse Fh ,

Wh = I + (XZ − γ 2I)

Fh = −S −1 (D∗se C se + B∗se X ),

S = I + D∗se D se

where A se, Bse, Cse and Dse are the state space system matrices of the shaped plant Gse. X and Z are the positive denite solutions of generalized c ontrol a lgebraic R iccati e quation (GCARE) a nd generalized ltering algebraic Riccati equation (GFARE) W is the weighting function

lic d amper te ster, w hich app lies t he d isplacement ( suspension travel) to t he M R d amper ac cording to t he dem and sig nal obtained from the computer. It is also connected to t he current amplier, which applies control current determined from control algorithm to the MR damper. Subsequently, the damping force of the M R d amper i s me asured f rom t he hydraulic d amper te ster and t he measured damping force is fed back i nto t he computer simulation. In short, the computer simultaneously runs both the hydraulic damper tester and current amplier during simulation loop w hile t he c omputer si mulation i s p erformed ba sed on t he measured data. Figure 25.18 presents time responses of t he MR suspension s ystem for t he bu mp e xcitation [34]. I t i s ge nerally known t hat t he d isplacement a nd ac celeration o f s prung m ass and ti re d eection a re u sed to e valuate r ide c omfort a nd ro ad holding of the vehicle, respectively. We clearly see that the vertical displacement and acceleration of sprung mass are reduced by activating the control eld determined from the H∞ controller. In addition, t he p itch a ngular d isplacement a nd ac celeration, a nd tire deection are also well reduced by applying the control input. Figure 25. 19 p resents c ontrolled f requency re sponses u nder random ex citation [ 34]. Ā e f requency re sponses a re ob tained from power spectral density (PSD) for the vertical acceleration of the sprung ma ss a nd t ire de ection. A s e xpected, t he PSDs for the v ertical ac celeration a nd t ire de ection a re r espectively reduced about 20% and 40% in the neighborhood of body resonance by applying input current. In addition, it is seen that the vertical acceleration is also reduced about 30% at the wheel resonance. Ā e c ontrol re sults presented i n Figures 25.18 a nd 25.19 indicate t hat b oth r ide p erformance a nd s teering s tability o f a vehicle can be improved by employing the proposed semi-active MR suspension system. 25.1.3.3

As a  rst s tep f or t he H ILS, t he c omputer si mulation o f t he full-vehicle MR suspension system is performed with initial value. Ā is computer simulation is incorporated with the hydrau-

CD-ROM Drive Featuring Piezoelectric Shunt

Ā e opt ical d isk d rives a s t he i nformation s torage de vice a re classied i nto CD -ROM, CD -R/RW, D VD-ROM/RAM/RW, and so on. Especially, the CD-ROM drive has been widely used Front tire deflection

Vertical displacement 0.08

0.008

0.04

0.00

−0.04 (a)

Tire deflection (m)

Displacement (m)

Uncontrolled Controlled

0

1

2

3 4 Time (s)

5

6

0.004

0.000

−0.004

−0.008

7 (b)

Uncontrolled Controlled

0

1

2

3 4 Time (s)

5

6

7

FIGURE 25.18 Bump responses of a ve hicle suspension using MR damper. (From Choi, S. B., Lee, H. S., and Park, Y. P., Vehicle Syst. Dyn., 38, 341, 2002. With permission.)

43722_C025.indd 12

10/3/2008 10:32:47 PM

25-13

Vibration Control for Smart Structures

Uncontrolled Controlled

6

4

2

0 0.1 (a)

0.6 Tire deflection PSD (m2/Hz)

Acceleration PSD (m2/s4 Hz)

Vertical acceleration

1

10

Front tire deflection Uncontrolled Controlled

0.5 0.4 0.3 0.2 0.1 0.0 0.1

100

Frequency (Hz)

⫻10−4

1

10

100

Frequency (Hz)

(b)

FIGURE 25.19 Random responses of a vehicle suspension using MR damper. (From Choi, S. B., Lee, H. S., and Park, Y. P., Vehicle Syst. Dyn., 38, 341, 2002. With permission.)

as a secondary storage device for computer peripherals. Ā e typical CD-ROM drive consists of the disk loading system, the feeding s ystem i ncluding opt ical p ick-up a nd s pindle, t he p rinted circuit b oard, a nd t he d rive ba se. Ā e t rack den sity a nd l inear density of t he d isk a re a bout 16,000 t racks-per-inch (T PI) a nd 43,000 (BPI), re spectively. I n order to ac hieve h igh c apacity of the C D-ROM d rive, ac curate p osition c ontrol o f t he opt ical pick-up head and vibration suppression of the feeding system are demanded. Ā e v ibration o f t he f eeding s ystem i s gener ally affected b y u nbalanced  exible disk with high rotating speed and e xternal e xcitation to t he m ain ba se. A ntivibration o f t he feeding system is an urgent problem to be solved to achieve high capability of the drive [36]. Normally, conventional drives adopt passive rubber mounts to prevent the feeding system from external excitation and the vibration of the spindle. In addition, auto ball balancer is often used [37], and a semiactive mount using ER uid has been also studied in order to overcome the limit of the passive rubber mounts [38]. Ā e CD-ROM drive base, which has a role of supporting the feeding system, is easily exposed to environmental v ibration s ources suc h a s u ser’s h andling a nd h igh speed rotating disk. If the vibration of the main base is not effectively re duced, t he rob ust s ervo c ontrol o f t he opt ical p ick-up cannot be guaranteed. Choi et al. initiated the effective vibration suppression methodologies for the CD-ROM drive by using piezoelectric damping i ntegrated w ith s hunt c ircuit [ 39]. I t i s k nown t hat e ach resonant mo de o f t he sm art s tructure i ntegrated w ith t he piezoelectric s hunt c ircuit c an b e mo deled b y me chanical vibration absorber that is one of very effective vibration reduction methods [40]. Ā e mechanical v ibration absorber is t hen modeled b y a n ele ctrical i mpedance mo del v ia me chanical– electrical c ircuit a nalogy. F igure 25. 20 p resents a s chematic diagram of the proposed system that consists of the structure (CD-ROM drive base) with piezoelectric material and resonant shunt circuit [39]. A piezoceramic (PZT) is incorporated to the rear pa rt o f t he d rive ba se, w hich i s e quipped w ith v ibration isolation work station by  xture. Ā e d rive ba se i s e xternally

43722_C025.indd 13

excited by another PZT positioned to the  xture. Ā e vibration of the drive base is measured by an accelerometer and analyzed by dynamic signal analyzer. Ā e target mode and the position of the PZT can be identied by modal analysis of the CD-ROM base using  nite element method (FEM). Ā e mo dal a nalysis h as b een u ndertaken t hrough t wo phases. Firstly, we investigate the modal analysis results in working frequency region for the original CD-ROM drive base, and choose a target mode shape, which exhibits the most severe vibration characteristics. In second phase, the PZT considering the deformation of the target mode shape is integrated to the drive base and modal analysis is performed again. From the analyzed mode shapes of the CD-ROM d rive ba se, t he t arget m ode i s de termined to b e t he second mo de (336 Hz) b ecause t he re ar pa rt o f t he ba se i s mo re severely deformed t han ot her pa rts at t he second mode. Prior to CD-ROM drive base

Shunt circuit Current Resistor

Inductor

Acceleration

Signal analyzer

Excitation

High-voltage amplifier

Function generator

FIGURE 25.20 Experimental conguration for vibration suppression of a CD-ROM drive that features piezoelectric shunt.

10/3/2008 10:32:49 PM

25-14

Smart Materials Frequency domain

Time domain 30

40 Open Shunted

Shunt on 20 Acceleration (m/s2)

Magnitude (dB)

30 20 10 0

(a)

0 −10 −20

−10 −20 250

10

275

300

325

350

375

−30 0.25

400

0.30

(b)

Frequency (Hz)

0.35

0.40

0.45

0.50

Time (s)

FIGURE 25.21 Vibration responses of CD-ROM drive that features piezoelectric shunt. (From Park, J.S., Lim, S. C., Choi, S. B., Kim, J. H., and Park, Y. P., J. Sound Vibration, 269, 1111, 2004. With permission.)

integrating the shunt circuit to the drive base, design parameters of the shunt circuit are tuned optimally by [39] x γ 2 + δ 2r γ + δ 2 = (2 x ST (γ 2 + 1)(γ 2 + δ 2rγ + δ 2 ) + K ij2 (γ 2 + δ 2rγ ) g = s/w nE , r = RiC spiw nE , d = w e w n , w e = 1

5.8)

LiC spi

where Li and Ri are the parameters of inductor and resister of the shunt circuit to be desired Cspi is the inherent capacitance of the piezoelectric shunted at constant strain wn and w e are mechanical and electrical natural frequency of the inherent structure and shunt circuit w En is the natural frequencies of t he s tructural mo de of interest with short circuit piezoelectric material Kij is the generalized electromechanical coupling coefficient xST is the static displacement of the system s is the Laplace variable In o rder to i nvestigate p erformance o f v ibration re duction using the piezoelectric shunt damping in target frequency region, the optimal parameter values such as inductance and resistance of the shunt circuit are determined, and then set for empirical realization. Ā e amount of dissipated energy is determined directly from the resistance, and thus a proper choice of resistance is essential. Hence, the optimal resistance value is experimentally retuned. Ā e analytical values of optimal resistance and inductance are different f rom t he e xperimentally o btained v alues. Ā e optimal values f or t he re sistance a nd i nductance i n s hunt c ircuit a re directly affected by natural frequency. In addition, the inductor in the shunt circuit can affect the optimal resistance value and vice versa. Ā is coupling effect between the resistor and the inductor is caused by electrical structure of synthetic inductor consisting of OP amps and resistors [41]. Figure 25.21a presents the frequency

43722_C025.indd 14

response in region of target frequency of 356 Hz [39]. It is clearly observed that the vibration reduction of 6.7 dB is achieved at t he resonant peak by activating the shunt circuit. Ā is result at the target mode is represented in time domain as shown in Figure 25.21b [39]. When the shunt circuit is short, the magnitude of the acceleration is reduced from 21.7 to 10.1 m/s2. In terms of the displacement, the vibration magnitude of 4.4 µm is reduced to 2.1 µm by activating the piezoelectric shunt circuit. Consequently, these experimental re sults demo nstrate t hat t he p iezoelectric s hunt damping can be effectively appl ied to v ibration reduction of t he CD-ROM drive base. It is expected that vibration reduction will give a high capability of the CD-ROM drive.

References 1 S. B. Choi, B. S. Ā ompson, and M. V. Gandhi, An exp erimental investigation on the active-damping characteristics of a class of ultra-advanced intelligent composite materials featuring ele ctro-rheological  uids, P roceedings o f Damping’89 C onference, 1, CA C.1–CAC.14. West P alm Beach, FL, February 1988. 2 M. V. Ga ndhi, B . S. Ā ompson, S. B . Cho i, a nd S. S hakir, Electro-rheological-uid-based a rticulating r obotic systems, ASME J ournal o f M echanisms, T ransmissions a nd Automation in Design, 111(3), 328–336, 1989. 3 J. P. Coulter and T. G. Duclos, Applications of electrorheological ma terials o n vib ration co ntrol, P roceedings o f the Second I nternational C onference o f ER Fl uids, 300–325, Raleigh, NC, August 1989. 4 Y. Choi, A. F. Sprecher, and H. Conrad, Response of electrorheological uid- lled laminate composites to forced vibration, Journal of Intelligent Material Systems and Structures, 3, 17–29, 1992. 5 S. B . Cho i, Y. K. P ark, a nd C. C. Che ong, Active vib ration control of intelligent composite laminate structures incorporating a n ele ctro-rheological  uid, Journal of I ntelligent Material Systems and Structures, 7(4), 411–419, 1996.

10/3/2008 10:32:50 PM

25-15

Vibration Control for Smart Structures

6 N. K. Petek, R. J. Goudie, and F. P. Boyle, Actively controlled damping i n e lectrorheological  uid- lled e ngine m ounts, SAE Technical Paper Series 881785, 1988. 7 S. H. Cho i, Y. T . Cho i, S. B . Cho i, a nd C. C. Che ong, Performance a nalysis o f a n en gine mo unt f eaturing ER uids and piezoactuators, International Journal of Modern Physics B, 10(23), 3143–3157, 1996. 8 T. G. Duclos, An externally tunable hydraulic mount which uses ele ctro-rheological  uid, SAE T echnical P aper S eries 870963, 1987. 9 S. B. Choi, S. R. Hong, and M. S. Han, Vibration control of a frame str ucture usin g ele ctro-rheological  uid mounts, International Journal of Mechanical Sciences, 44, 2027–2045, 2002. 10 N. K. Petek, D. J. Romstadt, M. B. Lizell, and T. R. Weyenberg, Demonstration o f a n a utomotive s emi-active susp ension using e lectrorheological  uid, SAE T echnical Paper S eries 950586, 1995. 11 S. B. Choi, Vibration control of a exible structure using ER dampers, Journal o f D ynamic S ystems, M easurement a nd Control, 121, 134–138, 1996. 12 W. J. Jung, W. B. Jeong, S. R. Hong, and S. B. Choi, Vibration control o f a  exible b eam str ucture usin g s queeze-mode ER Mount, Journal of Sound and Vibration, 273, 185–199, 2004. 13 S. B. Choi and W. K. Kim, Vibration control of a semi-active suspension fe aturing ele ctrorheological  uid dampers, Journal of Sound and Vibration, 234, 537–546, 2000. 14 S. B. Choi, H. S. Lee, and Y. P. Park, H-innity control performance of a full-vehicle suspension featuring magnetorheological dampers, Vehicle System D ynamics, 38(5), 341–360, 2002. 15 N. M. Wereley a nd L. P ang, N ondimensional a nalysis o f semi-active ele ctrorheological a nd ma gnetorheological dampers usin g a pproximate pa rallel p late mo dels, Smart Materials and Structures, 7, 732–743, 1998. 16 R. Stanway, J. T. Sproston, and A. K. EI-Wahed, Applications of electro-rheological  uids in vib ration control: A sur vey, Smart Materials and Structures, 5, 464–482, 1996. 17 S. R. Hong and S. B. Choi, Vibration control of a str uctural system usin g mag neto-rheological  uid mount, Journal of Intelligent M aterial S ystems a nd S tructures, 16, 931–936, 2005. 18 Y. Yu, S. M. Peelamedu, N. G. Nagathan, and R. V. Dukkipati, Automotive vehicle eng ine mo unting systems: A sur vey, ASME J ournal o f D ynamic S ystems, M easurement a nd Control, 123, 186–194, 2001 19 A. Premont, J. Dufour, and C. Malékian, Active damping by a local force feedback with piezoelectric actuators, Journal of Guidance Control and Dynamics, 15(2), 390–395, 1992. 20 S. H. Kim, S. B. Choi, and S. R. Hong, Vibration control of a exible str ucture usin g a h ybrid mo unt, International Journal of Mechanical Science, 46(1), 143–157, 2004. 21 N. M. Wereley a nd L. P ang, N ondimensional a nalysis o f semi-active ele ctrorheological a nd ma gnetorheological

43722_C025.indd 15

22

23 24 25 26

27

28

29

30

31

32 33

34

35

36

dampers usin g a pproximate pa rallel p late mo dels, Smart Materials and Structures, 7, 732–743, 1997. Y. T. Choi and N. M. Wereley, Comparative analysis of the time response of ele ctrorheological and mag netorheological da mpers usin g nondimensional pa rameters, Journal o f Intelligent Material Systems and Structures, 13(7/8), 443–451, 2002. F. Amirouche, L. P alkovics, a nd J . Woodrooffe, Optimal driver seat suspension design for heavy tucks, Transportation Systems ASME, 277–291, 1994. R. Andersson, Ā e low back pain of bus drivers in an urban area of california, Spine, 1481–1488, 1982. M. J . G riffin, Handbook o f H uman Vibration, A cademic Press, London, 1990. S. B. Choi, Y. T. Choi, and D. W. Park, A sliding mode control of a full-car electrorheological suspension system via hardwarein-the-loop simulation, ASME Journal of Dynamic Systems, Measurement, and Control, 122(1), 114–121, 2000. P. E. Boileau, S. Rakheja, X. Yang, and I. Stiharu, Comparison of bi odynamic re sponse ch aracteristics of v arious h uman body mo dels as a pplied to s eated v ehicle dr ivers, Noise & Vibration Worldwide, 7–15, 1997. P. E. B oileau a nd S. R akheja, Whole-body v ertical b iodynamic response characteristics of the seated vehicle driver measurement and model development, International Journal of Industrial Ergonomics, 449–472, 1998. S. B Choi, Y. T. Choi, and D. W. Park, A sliding mode control of a full-car electrorheological suspension system via hardwarein-the-loop simulation, ASME Journal of Dynamic Systems, Measurement and Control, 122(1), 114–121, 2000. Y. M. Han, J. Y. Jung, S. B. Choi, and N. M. Wereley, Sliding mode co ntrol o f ER s eat susp ension co nsidering h uman vibration model, International Journal of Modern Physics B, 19(7–9), 1689–1695, 2005. J. D . C arlson, D . M. C antanzarite, a nd S T. K. A. Cla ir, Commercial magneto-rheological uid devices, Proceedings of the 5th I nternational C onference o n ER Fl uids, MR Suspensions and Associated Technology, 1995, p. 20.. B. F. S pencer J r., S. J. D yke, M. K. Sa in, a nd J. D. C arlson, Phenomenological model for a magnetorheological damper, Journal of Engineering Mechanics, ASCE, 123, 230, 1997. G. M. Kamath, N. M.Wereley, and M. R. Jolly, Characterization of semi-active magnetorheological uid lag mode dampers, SPIE C onference o n S mart S tructures a nd I ntegrated Systems, 3329, 1998. S. B. Choi, H. S. Lee, and Y. P. Park, H-innity control performance of a full-vehicle suspension featuring magnetorheological dampers, Vehicle System D ynamics, 38(5), 341–360, 2002. D. C. M cFarlane a nd K. Glo ver, Rob ust co ntroller desig n using normalized coprime factor plant description, Lecture Notes in C ontrol I nformation S ciences, S pringer-Verlag, 1989. S. H. Chang, H. S. Kim, J. K. Choi, and D. G. Lee, A study on the desig n o f vib ration da mper f or hig h sp eed CD-R OM

10/3/2008 10:32:51 PM

25-16

37

38

39

40

41

Smart Materials

drive, Journal of Korean Society Mechanical Engineering (A), 22, 939–952, 1998. J. Lee and W. K. Van Moorhen, Analytical and experimental analysis of a self-compensating dynamic balancer in a rotating me chanism, ASME J ournal o f D ynamic S ystem, Measurements and Control, 118, 468–475, 1996. S. C. Lim, J. S. Park, S. B. Choi, and Y. P. Park, Vibration control of a CD-ROM feeding system using electro-rheological mounts, Journal of Intelligent Material System and Structures, 12(9), 629–638, 2002. J. S. P ark, S. C. Lim, S. B . Cho i, J . H. K im, a nd Y. P. P ark, Vibration reduction of a CD-ROM drive base using a piezoelectric sh unt cir cuit, Journal o f S ound a nd Vibration, 269(3–5), 1111–1118, 2004. N. W. H agood a nd A. v on Flo tow, D amping o f str uctural vibrations with piezoelectric materials and passive electrical networks, Journal o f S ound a nd Vibration, 146, 243–268, 1991. J. Kim, Y. H. Ryu, and S. B. Choi, New shunting parameter tuning metho d f or p iezoelectric da mping bas ed o n measured electrical impedance, Smart Materials and Structures, 9, 868–877, 2000.

25.2 Smart Materials for Sound and Vibration Control Cai Chao, Lu Chun, Tan Xiaoming, and Zheng Hui 25.2.1 Introduction Some materials have the ability to change shape or size simply by applying an electric potential, adding a little bit of heat, or to change their state from a liquid to a solid almost instantly when near a m agnet; t hese m aterials a re c alled sm art m aterials. Ā e structures with smart materials are also called smart structures. Generally, sm art m aterials c an b e g rouped i nto t he f ollowing categories: 1.

2.

3.

43722_C025.indd 16

Piezoelectric m aterials: Ā ey can generate an electric potential in response to an applied mechanical stress (direct piezoelectric effect) and can produce the mechanical stress when subjected to an electric eld (converse piezoelectric effect). Electrostrictive materials: Electrostriction is a property of all dielectric materials. Ā e mechanical strain is proportional to the square of the electric eld. Unlike piezoelectric m aterials, e lectrostriction c annot b e r eversed; mechanical d eformation w ill not i nduce a n e lectric  eld and t he re versal of t he ele ctric  eld do es not re verse t he direction o f t he de formation. Ā e reason is that electrostrictive materials are not poled. Magnetostrictive m aterials: M agnetostriction i s t he p roperty that causes a material to change its length when subjected to

4. 5.

6.

an electromagnetic  eld. Ma gnetostriction p roperties a lso cause materials to generate electromagnetic elds when they are deformed by an external force. Shape mem ory alloy s (S MA): Ā ey a re a lloys t hat c an “remember” t heir ge ometries. Ā e m aterials p roduce shape changes when subjected to a thermal eld. Optical  bers: Ā ey u se i ntensity, p hase, f requency, o r polarization o f mo dulation to me asure s train, tem perature, electrical/magnetic  elds, pressure, a nd ot her measurable qu antities. O ptical  bers c an be u sed a s va rious excellent sensors. Electrorheostatic (ER) and Magnetorheostatic (MR) materials: Ā ey a re smart  uids, which can experience a d ramatic c hange i n t heir v iscosity i n a m illisecond w hen exposed to a magnetic (MR) or electric eld ( ER). Ā e effect can be completely reversed just as quickly when the eld is removed.

Although t here e xist m any sm art m aterials a s men tioned above and they have been installed and are being installed into various applications from the control of aerospace structures to the de velopment o f b iomechanical de vices, t his c hapter o nly covers p iezoelectric m aterials a nd t heir app lications i n s tructural sound and vibration suppression. Ā e Cu rie b rothers d iscovered t he p iezoelectric e ffects in 1880, but it was not f ully used until the 1940s when high input impedance ampliers were de veloped. I n t he 1950s, t he piezoelectric e ffect bec ame commercialized w ith t he advent of electrometer tubes. Now, piezoelectric materials have become very popular i n pa ssive a nd ac tive s tructural c ontrol app lications, both as sensors and actuators. Lead zirconate titanate (PZT) and polyvinylidene  uoride ( PVDF) a re t wo c ommonly a vailable examples of piezoelectric materials in engineering applications. Generally, t here a re t hree m ajor me ans to app ly p iezoelectric materials f or s tructural s ound a nd v ibration c ontrol: pa ssive, active, and hybrid. 25.2.1.1

Passive Control with Piezoelectric Materials

Piezoelectric materials are used for passive sound and vibration control in various shunt circuit techniques. Ā e main feature of piezoelectric materials is energy transduction. Ā is occurs when mechanical work is done on piezoelectric elements and some portion i s c onverted to a nd s tored a s d ielectric energ y. I n a vibrating s tructure, a s hunting ne twork c an b e c ongured to accomplish vibration control by modifying the dynamics of the electrical system. Lesieutre [1] described the four types of shunt circuits: re sistive, i nductive, c apacitive, a nd s witched. Ha good and vo n F lotow [ 2] i nvestigated t he p ossibility o f d issipating mechanical energ y w ith p iezoelectric m aterials s hunted w ith passive electrical circuits. A ldrich et a l. [3] conducted research in t he de sign of a pa ssive control s ystem for a s pace structure. Wu [4] presented the results of a p iezoelectric element shunted with a parallel resistor–inductor circuit. Ā e research interest in this area involves the shunt circuit design and the shunt circuit tuning method, etc.

10/3/2008 10:32:51 PM

25-17

Vibration Control for Smart Structures

25.2.1.2

Active Control with Piezoelectric Materials

Piezoelectric-enabled s tructures ar e b eing u sed in creasingly in the control or cancellation of sound and vibration in the low- to mid-frequency range. Ā ey serve to bridge the limitation of the conventional me thod o f pa ssive d amping te chniques, w hich i s only effective when used in high-frequency applications. Ā er e are many applications of piezoelectric materials for active sound and vibration control, such as the helicopter blades [5] and gun barrels [6] to reduce structural vibrations [7], an active control of sound emanating f rom a s tructure [ 8,9], a nd s ound r adiation c ontrol from a h armonically e xcited pa nel u sing m ultiple p iezoelectric actuators [10]. Ā ere are two important issues that need to be considered in using piezoelectric materials as actuators for active control s ystems: (1) t hey u sually re quire l arge a mount of p ower for operation a nd (2) t he complexity of t he hardware i nvolved w ith active control (added hardware and control law design, etc.) [11]. Ā e research i nterest i n t his a rea i nvolves how to mo del t he dynamic properties of piezoelectric-enabled structures, optimal placement o f p iezoelectric ac tuators, opt imal s election o f PZ T actuator numbers, controller design, controller tuning method, power management, etc. 25.2.1.3

Hybrid Control with Piezoelectric Materials

Ā ere are also some applications when active–passive hybrid systems are used to ac hieve the best solution for vibration suppression [12]. In a hybrid vibration control system, an active control s ystem w ith a vol tage s ource i s i ntegrated w ith a pa ssive shunt circuit. Ā e advantage of such a system lies in the increase in the overall stability of the active control system. For whatever means above, it is necessary to have accurate and reliable numerical models for design and analysis of the piezoelectric-enabled structures. Some of t he earliest models [13–15] used the i nduced s train b y t he p iezoelectric ac tuators a s a n app lied strain that contributed to t he total strain of the nonactive structures similar to a t hermal strain contribution. To  nd analytical solutions, e lectrical an d m echanical e quilibrium o r g overning equations have to be solved for a set of boundary conditions [16]. It i s gener ally re cognized t hat t heoretical mo dels a re app licable only to a very limited range of well-dened geometries and boundary c onditions. F or p ractical app lications, n umerical me thods such as nite element method (FEM) are often used for modeling, simulation, and analysis of different smart structure designs. Ā ere a re laminated-type a nd d iscrete-type smart structures for structural sound and vibration control in practice. Āe former means that piezoelectric materials cover or are embedded in the entire structure. Ā e latter means that the piezoelectric materials occupy a rel atively sm all a rea o f t he s tructure. Ā e modeling approaches and analysis techniques diff er considerably between the laminated-type and the discrete-type smart structures. Ā e purpose of this chapter is to provide an overview of various numerical modeling methods of the dynamic properties of piezoelectric-enabled sm art s tructures a nd rel ated i ssues. Ā e review i s b rief a nd m akes no at tempt to b e e xhaustive. I t i s organized into the following sections: (1) linear theory of elec-

43722_C025.indd 17

trodynamics t hat o utlines t he f undamental go verning e quations o f sm art s tructure mo deling; ( 2) me thods f or laminated-type sm art s tructures t hat i nclude t he el astic a nd plate theories in addition to coupled layerwise theory; (3) methods for discrete type smart structures that include the equivalent l ine moment approach, the energy method together with assumed mo de me thod a nd  nite elemen t app roach; a nd ( 4) new research areas in smart structures, which include nonlinear c haracteristics o f t he p iezoelectric m aterials, ac curacy description of electric potentials in the piezoelectric layers, etc.

25.2.2 Linear Theory of Electrodynamics Ā e l inear t heory of electrodynamics t reats a ll t he elastic, piezoelectric, and dielectric coefficients as constants independent of the magnitude and frequency of applied mechanical stresses and electric elds. Ā e governing equation in the absence of the body forces and body charges consists of stress equations of motion [17,18]:

− LT σ = 0 (2 ρU d

5.9)

Ā e c harge e quation o f e lectrostatics a nd t he e lectric  eldelectric potential relations are ∇ ⋅ D = 0 (2

5.10)

E = −∇φ (2

5.11)

and

where the symbolic dot denotes differentiation with respect to time t the superscript T the transposition r is the material density (kg/m3) U the particle displacement vector (m) s is the stress vector (N/m2) Ld the operator [17,18] D the electric displacement vector (C/m2) E the electric eld vector (V/m) f the scalar electric potential (V) Ā e constitutive relations for a piezoelectric layer are σ = cε − e TE (2

5.12)

D = eε + gE (2

5.13)

and

where c is t he st iff ness m atrix ( N/m2) at constant electric eld strength g the dielectric matrix (F/m or C/Vm) at constant mechanical strain e the piezoelectric stress matrix (N/mV or C/m2)

10/3/2008 10:32:52 PM

25-18

Smart Materials

Combination of the equations above yields the three-dimensional differential e quations o f t he l inear p iezoelectric c ontinuum i n terms o f t he f our de pendent v ariables ( three d isplacement components and one electric potential) as

(2 LTd cL dU + LTd e T∇ϕ = ρ U

5.14)

∇ ⋅ (eL dU − g∇ϕ )= 0 (2

5.15)

In the presence of boundaries, the appropriate boundary conditions must be adjoined to Equations 25.14 and 25.15. If there is a material surface of discontinuity, across the surface there are the continuity co nditions o f st resses, d isplacements, el ectric d isplacements, and electric potentials.

25.2.3 Methods for Laminated-Type Smart Structures Ā e developments of piezoelectric laminate models satisfy the need for more accurate mechanic models to account for the characteristics of composite laminates and the need for more accurate electroelastic mo dels to ac count f or t he c oupling e ffects bet ween t he electrical an d m echanical  elds in side t he p iezoelectric c omponents. Ā e f undamental works by e arlier re searchers [17,18] provide much of t he necessary t heoretical ba sis for t he modeling of static and dynamic behavior of laminated-type smart structures. 25.2.3.1 Elastic Theory Ā e elastic theory utilizes the elastic wave propagation method to establish the state variable relationship between upper and lower surfaces of a l aminate. Elastic theory can provide an exact solution of a problem, which satises the governing equations at every point of the domain and the boundary conditions of the problem. Ā ere are several methods to obtain a solution of governing equations within a piezoelectric layer in the frequency domain. 25.2.3.1.1

Transfer Matrix Method

Ā e continuity of displacements, stresses, and electric parameters at the interfaces of different l ayers i s i mposed. O n t he top a nd bottom surfaces of the laminate, normal displacements, stresses, and electric variables have to be continuous as well. Ā e boundary c ondition approach c onstitutes t he ba sis o f t ransfer m atrix method [20,21] and the surface impedance tensor method. 25.2.3.1.2 Surface Impedance Tensor Approach A signicant formulation to a matrix of piezoelectric composite materials app ears to b e t he eig ht-dimensional s tate vector formalism proposed originally by K raut [19]. At t he heart of t his formulation is the well-known fact that the ordinary differential equations can be transformed into a set of rst-order equations. As an extension, Honein et al. [22] proposed a systematic methodology about the surface impedance tensor. Ā e surface impedance tensor relates the components of particle displacement and the normal component of the electric displacement on a surface

43722_C025.indd 18

to the electric potential and the component of traction acting on the same surface. Once the surface impedance tensor for a single layer is obtained, a single recursive algorithm allows the evaluation of t he surface impedance tensor for any number of layers. Ā e surface impedance tensor approach is proposed to overcome a numerical difficulty, which exists when getting the solution for many l ayers a s t he p roduct o f t he s olutions o f e ach l ayer. Ā e method has been successfully applied to solve the acoustic reection a nd t ransmission p roblem w hen a p iezoelectric-enabled structure is subjected to an acoustic wave incidence [23,24]. Ā e advantages of the approach lie in that it can be used not only for wave propagation (the high-order theory) in the piezoelectric l ayer b ut a lso f or v ibration a nalysis ( the lo wer-order theory) of general piezoelectric laminates. 25.2.3.2

Classical Laminated Plate Theory

Classical l aminated p late t heory (CLPT) i s a n e xtension o f t he classical plate theory to composite laminates [25]. Lee and Moon [26], a nd L ee [ 27] u sed t he a ssumptions o f C LPT to der ive a simple theory for a piezoelectric laminate. Reddy [28] presented the theoretical formulation of laminated plates with piezoelectric layers a s s ensors or a ctuators. M any i nvestigators [ 29,30] u sed CLPT models and its variations to design piezoelectric laminates for different applications. Ā ese models used simplifying approximations in characterizing t he induced strain  eld a nd ele ctric elds generated due to an applied voltage or external load. CLPT assumes that the Kirchhoff hypothesis holds for the laminate. CLPT i mplies t hat for a c ase of both plane strain a nd plane stress. However, from practical considerations, it uses the restriction of each layer, which is in a state of plane stress, because most laminates a re t ypically t hin c ompared to t he i n-plane d imensions. It is one pertinent assumption in establishing the constitutive relationships for the laminae of a laminated structure. 25.2.3.3 First-Order Shear Deformation Theory To en hance t he CP LT m odel, a n a lternative m odel,  rst-order shear de formation t heory ( FSDT) w as de veloped [ 28], w hich belongs to t he h ierarchy o f e quivalent si ngle-layer ( ESL) l aminate theories as well. In FSDT, the third Kirchhoff hypothesis is released. It me ans t hat t ransverse nor mals do not re main p erpendicular to the mid-surface after deformation. It is also called as Rei ssner-Mindlin p late t heory o r a c onstant s hear a ngle theory. I t i s note d t hat s hear c orrection f actors a re no rmally adopted i n c omputing t he t ransverse s hear f orce re sultants to correct t he discrepancy between t he actual stress state a nd t he constant stress state predicted by the FSDT. 25.2.3.4 Third-Order Shear Deformation Theory In the FSDT model, the shear correction factors are not easy to determine. To ach ieve b etter approximation, t he h igher-order expansions o f t he d isplacement  eld h ave b een d eveloped. A third-order s hear de formation t heory (T SDT) mo del ac commodates the vanishing of transverse shear strain on the top and bottom surfaces of a laminate [28,31]. Reddy [28] presented the Navier solutions and  nite element models of CLPT and TSDT.

10/3/2008 10:32:54 PM

Vibration Control for Smart Structures

25-19

25.2.3.5 Coupled Layerwise Theory

observation was t hat t he effectiveness moments resulting f rom the piezoactuators can be seen as concentrated on the two ends of t he ac tuators w hen t he b onding l ayer i s a ssumed i nnitely thin. A s a n e xtension o f t he o ne-dimensional t heory [ 14], Dimitriadis [ 44] de veloped t he  nite t wo-dimensional p iezoelectric elements perfectly bonded to t he upper a nd lower su rfaces o f el astic p late t hrough s tatic a nd dy namic a nalysis. Ā e loads i nduced b y t he p iezoelectric ac tuator to t he supp orting thin elastic structure are estimated in their research. Ā en, the equivalent m agnitude o f t he e dge mo ments a re app lied to t he plate to replace the actuator patch such that the bending stress at the surface of the plate is equal to the plate’s interface stress when the patch is activated. As mentioned before, the approach used the induced strain by the piezoelectric actuators as an applied strain that contributes to the total strain of the nonactive structures, similar to a t hermal strain contribution. Strictly speaking, it is not a f ully coupled a nalysis between mechanical a nd piezoelectric structures. It only c onsiders t he c onverse piezoelectric e ffect. Ā e electromechanical coefficient d31 is the only source of coupling between the electric and the mechanical displacement elds. Many investigators have used this analytical model for various applications where p iezoelectric patc hes a re u sed f or c ontrolling b eams, plates, and shells. However, in many device geometries, all elements of the piezoelectric tensors couple to t he strain and electric  elds a nd t he a ssumption t hat o nly d31 p lays a role i s no longer true.

Discrete layer theories [31] and layerwise theories [32] have been developed f or t he s tatic a nd dy namic a nalysis o f p iezoelectric laminates. Unlike the simplied electric and displacement elds through t he t hickness of t he laminate in t he equivalent singlelayer piezoelectric laminate theories above, the mechanical displacements and the electric potential are assumed to be piecewise continuous across the thickness of the laminate in the layerwise theories. Ā ese theories provide a much more kinematically correct r epresentation o f c ross-sectional w arping a nd ca pturing nonlinear v ariation of ele ctric p otential t hrough t he t hickness associated with thick laminates. Ā e developments of layerwise laminate theory for a laminate with embedded piezoelectric sensors a nd ac tuators a re p resented i n t he l iterature [ 33,34]. Comparisons of the predicted free vibration results from the layerwise t heory w ith t he e xact s olutions f or a si mply supp orted piezoelectric l aminate b ring o ut t he i mproved ac curacy a nd robustness of the layerwise theory over CLPT and FSDT [35] for piezoelectric laminates. Tani and Liu [36,37] originally proposed a numerical method to investigate the surface waves in functionally gradient piezoelectric plates. L iu a nd Achenbach [38] na med t he method t he strip element method (SEM). Ā e technique has a c lear advantage o ver t he  nite elemen t me thod f or l ayered s tructures i n terms o f s torage re quirements [ 39]. Re cently, Ha n e t a l. [ 40] extended SEM to analysis of surface wave properties in a hybrid multilayered piezoelectric circular cylinder.

25.2.4.2

25.2.4

Methods for Discrete Type Smart Structures

Discrete monolithic pieces of piezoelectric material can also be used f or s ensing a nd c ontrol p urposes i nstead o f c overing t he entire surface in the form of layers due to weight and difficulty in fabrication c onsiderations. Ā e d iscrete patc hes c an ei ther b e attached to t he structure as in the literature [41–44] or embedded within the substrate as in the work [14]. 25.2.4.1

Equivalent Line Moment Approach

In this approach, a static study of the behavior of twin, symmetric p iezoelectric ac tuators b onded o n a s tructure w as c arried out. It w as demonstrated t hat a p erfectly b onded ac tuator w as equivalent to external line moments acting along the boundaries of the piezoelectric elements. Ā e representative papers are from Crawley and deLuis [14] and Dimitriadis et al. [44]. In all these previous models, the usual assumption was that unless an electric  eld was applied, the presence of the piezoelectric material on or in the substrate does not alter the overall structural properties signicantly. It was also assumed that the thickness of the bonding ad hesive l ayer w as ne gligible a nd c aused ne gligible property changes. Crawley a nd deLu is [ 14] p resented a r igorous s tudy o f t he stress–strain–voltage behavior of piezoelectric elements bonded to a nd e mbedded i n o ne-dimensional bea ms. A n i mportant

43722_C025.indd 19

Hamilton’s Principle with a Rayleigh-Ritz Formulation

Hagood et al. [45] studied the damping of structural vibrations with piezoelectric materials and passive electric networks. Ā ey derived a n a nalytical m odel f or a n el ectroelastic s ystem w ith piezoelectric m aterials u sing Ha milton’s p rinciple w ith a Rayleigh-Ritz f ormulation. Ā e re quirement t hat a ll c oupling between piezoelectric devices and substrate has to be included in the formulation is effectively facilitated by the Hamilton variational principle or energy methods, which is an effective method for complex structures because mechanical and electrical equilibrium equations do not need to be solved explicitly. Āe physics of the entire structure has been fully accounted for in the energy integrals and there is no need to derive equations based on forces and mome nts. Ā e resultant equation of motion is solved by using a Rayleigh-Ritz method. Hagood et al. [45] found that the piezoelectric energ y transformation properties highly coupled to the dynamics of the electric c ircuit a nd el astic s ystem. Ā ey st udied t his co upling f or symmetry piezoelectric actuators to add d amping to t he structural modes of a cantilevered beam. Gibbs and Fuller [46] developed an analytical static model to describe the response of an innite beam subjected to a n asymmetric actuation induced by a p erfectly b onded p iezoelectric elemen t. P lantier e t a l. [ 47] derived a dynamical model for a beam driven by a single asymmetric piezoelectric actuator. Ā eir model included the effect of the adhesive bonding layer, i.e., the actuator is not assumed to be

10/3/2008 10:32:55 PM

25-20

Smart Materials

perfectly bonded to t he ba se structure. L am e t a l. [48] der ived the energy equations via virtual work rst. Ā e assumed modes method w as u sed to d iscretize t he energ y e quations a nd Lagrange’s equation was used to obtain the equations of motion for different treatments [7]. 25.2.4.3

Finite Element Approach and Other Numerical Methods

For discrete piezoelectric patches embedded or bonded on beam and plate structures, i nvestigators have mostly u sed numerical methods such as the nite element method since obtaining exact solutions is difficult. A potential advantage of the use of the nite element model for piezoelectric devices is t hat t he  elds in t he material m ay b e p recisely de scribed w ithout si mplifying assumptions. Finite element me thods were put i nto application for piezoelectric structures since Allik et al. [49], Ha e t al. [43], and Rao and Sunar [50] developed a  nite element method to s tudy t he dynamic as well as the static response of plates containing distributed piezoelectric ceramics based on the variational principles. Hwang [30] used a f our-node quadrilateral element based on classic laminate theory with the induced strain actuation and Hamilton’s principle. Ā ey assume that no stress eld is applied to t he ac tuator l ayer a nd ac cordingly t he e quivalent ac tuator moments per unit length are found as external excitation loads. Lam et a l. [51] a lso de veloped a  nite element model ba sed on the classic laminated plate theory for the active vibration control of a composite plate containing distributed piezoelectric sensors and actuators. Zhou [31] used the same concept to establish nite element mo dels o f p iezoelectric c omposite l aminates ba sed o n Hamilton’s principle as well as CLPT and FSDT and TSDT theories re spectively. Sh en [52] de veloped a o ne-dimensional  nite element formulation for the  exural motion of a beam containing d istributed p iezoelectric de vices. I n h is w ork, t he gener alized variational principle was u sed to i nclude t he v irtual work done by the inertial and electric forces to get the functional principle. Chen et al. [53] presented the general nite element formulations f or p iezoelectric s ensors a nd ac tuators b y u sing t he virtual work principle. Kim et al. [54] developed a transition element to connect the three-dimensional solid elements in t he piezoelectric region to the at-shell elements used for the plate. Ā ey adopted some special techniques to overcome the disadvantages and inaccuracy of modeling a plate with three-dimensional elements. Āis approach has mer its i n ter ms o f ac curacy i n mo deling t he p iezoelectric patches and computational economy for the plate structure. In addition, Liu [55,56] originally proposed a point interpolation method (PIM) that belongs to the mesh-free or meshless methods. It uses the nodal values in the local support domain to interpolate t he s hape f unctions. To t he b est k nowledge o f t he authors, L iu e t a l. [57] i nitially de alt w ith piezoelectric de vice analysis with mechanical and electrical coupled problems using the p oint i nterpolation me thod. I n t his me thod, t he p roblem domain was represented by a set of arbitrarily distributed nodes. A polynomial basis was used to construct the shape functions,

43722_C025.indd 20

which possess delta function properties. Piezoelectric structure with arbitrary shape was formulated using polynomial PIM in combination w ith variational principle a nd l inear constitutive piezoelectric equations. Numerical examples [57] provide solid validations of the proposed method. It was found that the present method is easy to implement, and very exible and stable for static a nd dy namic a nalysis o f p iezoelectric s tructures w ith arbitrary shape and different mechanical or electrical boundary conditions.

25.2.5

Potential Research Areas in Smart Structures

25.2.5.1 Nonlinear Characteristics of the Piezoelectric Materials Among the currently available sensors and actuators, the smallest ones are of the order of few millimeters. However, progress toward intelligent st ructures r equires u s to de velop sma rt ma terial systems that are of the order of a few microns. Ā is reduction in size has tremendous technological benet; however clear understanding of reliability and system integrity is vital to the efficient and optimum use of these material systems. As dimensions get smaller, i nduced ele ctromechanical  elds ge t l arger. Ā er efore, the convenience of linearity in modeling should be abandoned, and material and geometric nonlinearities should be accounted for. In addition, piezoelectric materials only exhibit a linear relationship between the electric eld and strain for low eld values (up to 1 00 V/mm). Ā e relationship behaves nonlinear for large elds, the material exhibits hysteresis, furthermore, piezoelectric materials s how d ielectric a ging a nd h ence l ack re producibility of strains. Using piezoelectric ac tuators, Sh i a nd Atluri [58] conducted studies in order to control the nonlinear dynamic response of a space frame. By i ncorporating the effect of large deection and in-plane loads due to the piezoactuators, Pai et al. [59] developed a fully nonlinear theory for the dynamics and active control of a laminated c omposite plate w ith i ntegrated piezoelectric ac tuators a nd s ensors u ndergoing l arge-rotation a nd sm all-strain vibration. Ro yston a nd H ouston [ 60] p resented t he no nlinear vibratory b ehavior o f a 1 –3 p iezoceramic c omposite t heoretically a nd e xperimentally. I n t heir w ork, t he t heoretical mo del for t he electroelastic behavior of t he 1–3 composite follows t he conventional a ssumption m ade by pr ior i nvestigators but includes nonlinear terms to account for hysteresis in the embedded PZT phase. 25.2.5.2 Accurate Description of Electric Potentials in the Piezoelectric Layers It is noted that in much of the literature published on the mechanics model for the analysis of the coupled structure, a constant ele ctric p otential d istribution o f t he w hole p iezoelectric actuator in its length dimension and a linear distribution i n i ts t hickness d imension a re a ssumed. H owever, to satisfy t he M axwell e quation, a c onstant ele ctric p otential

10/3/2008 10:32:55 PM

25-21

Vibration Control for Smart Structures

distribution in the longitudinal direction may not be assumed, but r ather ob tained b y s olving t he c oupled go verning e quations [61]. Ā e eld i nside l ayers of a p iezoelectric l aminate w as previously examined by Ray et a l. [62], Roh e t a l. [63], and Heyliger [16] using full elasticity theories, without any approximations on the mechanical and electrical elds. Ā e exact solution obtained using t he exact elasticity t heory i ndicated t hat t he electric a nd elastic  eld distributions were often poorly modeled using simplied theories [16]. Ā ey showed that the electric eld inside the sensor layers was not zero and both the electrical and mechanical  eld d istributions w ere e vidently a ffected b y t he rel ative values o f t he d ielectric c onstants o f t he l ayers i n a t hree-layer cross-ply PVDF laminate. To increase the order of variation of the electrical and mechanical  elds i nside t he l ayers, h igher o rder t heories s hould b e employed i n t he l aminate mo dels. Bi segna a nd M aceri [ 64] showed that when the ratio of thickness to w idth of the plate is larger than or equal to 1/5, FSDT provides results, which differ from exact solution by 20% for displacements, electric potential, and t he i n-plane s tress c omponents. M ore re cently, Yang [ 65] included h igher o rder ( quadratic) ele ctric p otential v ariation through the thickness of the actuators and obtained two-dimensional e quations f or t he b ending mot ion o f el astic p lates w ith partially ele ctroded p iezoelectric ac tuators at tached to t he top and b ottom su rfaces o f a t hick p late. A lthough ne gligible f or thin a ctuators, th is e ffect of h igher ord er e lectrical b ehavior through the thickness needs to be considered for thick actuators. Tiersten [66] derived the approximate equations for extensional and  exural mot ion o f a t hin p iezoelectric p late sub jected to large electric elds. Up to cubic order terms are included in the expansion o f ele ctric p otential ac ross t he p late t hickness to describe t he h igher o rder ele ctrical b ehavior, a nd s howed t hat for a very thin plate, the quadratic and cubic terms in the expansion can be ignored. Wang and Quek [67,68] assumed the distribution in the transverse direction to be sinusoidal for the short-circuited electrodes case when considering a lo ng and t hin beam-embedded piezoelectric ac tuators. Ā ey c oncluded t hat t he longitudinal d istribution of the electric potential in the piezoelectric layer was not constant. 25.2.5.3 Nanostructured Ceramic Nanostructured c eramic [69] has b een put i nto t he re search stage b ecause t he na noscale m icrostructures m ay re sult i n changes in properties when the feature size is less than a particular level. Among several possible smart nanoscale materials, c arbon na notubes (CNT) h ave a roused g reat i nterest i n the research community because of their remarkable mechanical, electrochemical, piezoresistive, and other physical properties. The potential changes in properties (improved strength and toughness etc.) are what has driven research in nanoceramics o ver t he l ast de cade. K ang e t a l. u sed a na nosmart material t o d evelop a nove l s ensor for s tructural he alth monitoring [70].

43722_C025.indd 21

25.2.6 Closure Five si mulation me thods o n l aminated-type sm art s tructures have been reviewed in this chapter. Ā ese are (1) elastic theory; (2) classic laminated plate theory; (3) rst-order shear deformation theory; (4) third-order shear deformation theory; and (5) coupled l ayerwise t heory. Ā ree si mulation me thods o n d iscretetype smart structures are (1) equivalent line moment approach, (2) Ha milton’s principle w ith a R ayleigh-Ritz formulation, a nd (3) nite element approach and other numerical methods. Ā er e are s everal a reas of si mulation me thods t hat ne ed at tention i n the ye ars to c ome. Ā ese i nclude (1) nonlinearity of t he piezoelectric materials, (2) accurate description of electric elds in the piezoelectric materials, and (3) new piezoelectric materials using nanotechnologies.

References 1. Lesieutre G. A. 1998. The Shock and Vibration Digest, 30(3), 187–195. 2. Hagood N. W. and von Flotow A. 1991. Journal of Sound and Vibration, 146(2), 243–268. 3. Aldrich J. B., Hagood N. W., von Flotow A., and Vos D. W. 1993. Proceedings of the SPIE—The International Society for Optical Engineering, 1917(2), 692–705. 4. Wu S. Y. 1996. Proceedings o f t he S PIE—The International Society for Optical Engineering, 2720, 259–269. 5. Straub F. K., Ealey J. A., and Schetley L. M. 1997. Proceedings of the SPIE—The International Society for Optical Engineering, 3044, 99–113. S ociety o f P hoto-Optical I nstrumentation Engineers, Bellingham, WA. 6. Mattice M. S. a nd L aVigna C. 1997. Proceedings of the SPIE—The I nternational S ociety f or O ptical E ngineering, 3039, 630–641, S ociety o f P hoto-Optical I nstrumentation Engineers, Bellingham, WA. 7. Cai C., Zheng H., Hung K. C., and Zhang Z. J. 2006. Smart Material and Structures, 15, 147–156. 8. Khan M. S., C ai C., Hung K. C., a nd Varadan V. K. 2002. Smart Material and Structures, 11, 346–354. 9. Linder D. K., Zuonar G. A., Kirley III G. C., and Emery G. M. 1996. Proceedings of the SPIE—The International Society for Optical Engineering, 2717, 543–554. 10. Fuller C. R., Dimitriadis E. K., and Bor-Tsuen W. 1991. AIAA Journal, 29(11), 1802–1809. 11. McGowan A. M. R . 1999. Proceedings o f t he S PIE—The International Society for Optical Engineering, 3674, 178–195. 12. Kahn S. P. and Wang K. W. 1994. Active Control of Vibration and N oise—ASME, D esign E ngineering Di vision, 75, 187– 194 ASME, New York. 13. Bailey T. a nd Hubbard J.E. Jr. 1985. Journal of G uidance, Control, and Dynamics, 8, 605–611. 14. Crawley E. F . a nd deL uis J . 1987. AIAA J ournal, 25, 1373–1385. 15. Robbins D . H. a nd Re ddy J . N. 1991. Computers a nd Structures, 41, 265–279.

10/3/2008 10:32:55 PM

25-22

16. Heyliger P. 1997. J. Appl. Mech. Trans. ASME, 64, 299–306. 17. Tiersten H. F. 1969. Linea r P iezoelectric p late vib rations; elements o f the line ar the ory of p iezoelectricity a nd the vibration of piezoelectric plates, Plenum, New York. 18. Nayfeh A H. 1995. N orth-Holland S eries in Applied Mathematics and Mechanics, Elsevier, Amsterdam. 19. Kraut E. A. 1969. Physical Review, 188(3), 1450–1455. 20. Cai C., Liu G. R., and Lam K. Y. 2001a. Journal of Sound and Vibration, 248(1), 71–89. 21. Cai C., Liu G. R., and Lam K. Y. 2001b. Smart Materials and Structures, 10, 689–694. 22. Honein B ., B raga A. M. B ., B arbone P., a nd H errmann G. 1991. Journal of Intelligent Material Systems and Structures, 2, 542–557. 23. Barbone P. E. a nd B raga A. M. B . 1992. Fir st E uropean Conference on Smart Structures and Materials, Forte Crest Hotel, Glasgow, pp. 325–328. 24. Braga A. M. B., Honein B., Barbone P. E., and Herrmann G. 1992. Journal of Intelligent Material Systems and Structures, 3, 209–223. 25. Reddy, J. N. 1997. Mechanics of L aminated C omposite Plates and Shells: Ā eory and Analysis, CRC Press, Boca Raton. 26. Lee C. K. and Moon F. C. 1989. JASA, 85, 2432–2439. 27. Lee C. K. 1990. JASA, 87, 1144–1158. 28. Reddy J. N. 1999. Engineering Structures, 21(7), 568–593. 29. Lam K. Y. and Ng T. Y. 1999. Smart Materials and Structures, 8(2), 223–237. 30. Hwang W. -S., P ark, H. C. 1993. AIAA J ournal, 31(5), 930–937. 31. Zhou Y. L. 1999. M aster’s thesis. N ational U niversity o f Singapore. 32. Mitchell J. A. and Reddy J. N. 1995. International Journal of Solids and Structures, 32, 2345–2367. 33. Saravanos D . A., H eyliger P . R ., H opkins D . A. 1997. International J ournal o f S olids a nd S tructures, 34(3), 359–378. 34. Saravanos D. A. and Heyliger P. R. 1995. Journal of Intelligent Material Systems and Structures, 6, 350–363. 35. Gopinathan S. V., Varadan V. V., and Varadan V. K. 2000. Smart Materials and Structures, 9(1), 24–48. 36. Tani J. and Liu G. R. 1993. JSME International Journal, 36, 152–155. 37. Liu G. R. and Tani J. 1994. Transactions of the ASME Journal of Vibration and Acoustics, 116, 440–448. 38. Liu G. R. and Achenbach J. D. 1994. ASME Journal of Applied Mechanics, 61, 270–277. 39. Liu G. R . a nd X i Z. C. 2001. Elastic Waves in Anisotropic Laminates, CRC Press, pp. 235–253. 40. Han X., Li u G. R ., a nd Oh yoshi T . 2004. Computational Mechanics, 33(5), 334–344. 41. Crawley E. F. a nd L azarus K. B . 1991. AIAA J ournal, 29, 944–951. 42. Ā omson S. P. and Loughlan J. 1995. Composite Structures, 32, 59–67.

43722_C025.indd 22

Smart Materials

43. Ha S. K., Keilers C., and Chang Fu-Kuo 1992. AIAA Journal, 30(3), 772–780. 44. Dimitriadis E. K., F uller C. R ., a nd Rog ers C. A. 1991. Transactions of the ASME, 113, 100–107. 45. Hagood N. W., Chung W. H., and Von Flotow A. 1990. AIAA Pap, AIAA-90–1087-CP 2242–2256. 46. Gibbs G. P. and Fuller C. R. 1992. JASA, 92, 3221–3227. 47. Plantier G., Guigou C., Nicolas J., Piaud J. B., and Charette F. 1995. Acta Acustica, 3, 135–151. 48. Lam M. J ., I nman D . J ., a nd Sa unders W. R . 2000. Smart Materials and Structures, 9, 362–371. 49. Allik H. and Hughes T. J. R. 1970. International Journal for Numerical Methods in Engineering, 2, 151–157. 50. Rao S. S. a nd S unar M. 1993. AIAA J ournal, 31(7), 1280–1286. 51. Lam K. Y., Peng X. Q., Liu G. R., and Reddy J. N. 1997. Smart Materials and Structures, 6, 583–591. 52. Shen M. –H. H. 1994. Smart M aterials a nd S tructures, 3, 439–447. 53. Chen C. Q., Wang X. M., and Shen Y. P. 1996. Computers & Structures, 60(3), 505–512. 54. Kim J., Varadan V. V., and Varadan V. K. 1997. International Journal f or N umerical M ethods i n E ngineering, 40(5), 817–832. 55. Liu G. R. and Gu Y. T. 2001. International Journal of Nuner. Methods and Engineering, 50, 937–951. 56. Liu G. R . 2002. Meshfree Me thods. M oving bey ond  nite element Method, CRC Press, Boca Raton, FL, in press. 57. Liu G. R ., D ai K. Y., Lim K. M., a nd G u Y. T . 2002. Computational Mechanics, 29(6), 510–519. 58. Shi G. a nd Atluri S. N. 1990. Computer & S tructures, 34, 549–564. 59. Pai P . F ., N ayfey A. H., Oh K., a nd M ook D . T . 1993. International Journal of Solids and Structures, 30, 1603–1630. 60. Royston T . J . a nd H ouston B . H. 1998. JASA, 104(5), 2814–2827. 61. Gopinathan S. V. 2001. PhD Ā esis, Ā e Pennsylvania State University. 62. Ray M. Ch., Rao K. M., and Samanta B. 1993. Computers & Structures, 47, 1031–1042. 63. Roh Y. R., Varadan V. V., and Varadan V. K. 1996. Smart Materials and Structures, 5, 369–378. 64. Bisegna P. a nd M aceri F. 1996. ASME J ournal o f Applied Mechanics, 63, 628–638. 65. Yang J. S. 1999. Smart Materials and Structures, 8, 73–82. 66. Tiersten H. F. 1993. ASME AMD, 161, 21–34. 67. Wang Q. a nd Quek S. T. 2000. Smart Materials and Str uctures, 9, 103–109. 68. Quek S. T. and Wang Q. 2000. Smart Materials and Str uctures, 9, 859–867. 69. Cain M. and Morrell R. 2001. Chemistry, 15, 321–330. 70. Kang I., Schulz M. J., Lee J. W., Choi, G. R., Jung J. Y., Choi J.-B., a nd H wang S.H. 2007. Solid S tate Ph enomena, 120, 289–296.

10/3/2008 10:32:56 PM

26 Active Truss Structures 26.1 I ntroduction .......................................................................................................................... 26-1 26.2 P iezoelectric Stack Actuator ................................................................................................ 26-1 26.3

Structure with a Stack Transducer

Integral Force Feedback ....................................................................................................... 26-3

Extension to Cable Structures

26.4 Passive Shunt Damping........................................................................................................ 26-4 Resistive Shunting • Inductive Shunting • Generalized Electromechanical Coupling Factor

B. de Marneffe Université Libre de Bruxelles

André Preumont Université Libre de Bruxelles

26.1

26.5 Active Shunt Damping ......................................................................................................... 26-6 26.6 E xperimental Results............................................................................................................ 26-7 Frequency Response Function (FRF) • Root Loci

Acknowledgments ........................................................................................................................... 26-9 References ......................................................................................................................................... 26-9

Introduction

Ā e concept of active trusses is quite natural: it consists of replacing one or several passive bars by active struts. Piezoelectric transducers are ideally suited for this purpose, because of their high stiffness and virtually unlimited resolution; other types of transducer, based, e.g., on electrostrictive materials can also be used but they are not investigated here. Ā e active truss concept was rst demonstrated in the late 1980s [1–3]; very effective active control strategies were proposed at that time. One o f them, known as in tegral force feedback (IFF), is based on a collocated force sensor and has guaranteed stability [4] as long as perfect amplier and actuator dynamics are assumed. Although active contol can be efficient, it implies power electronics (for the actuator), a sensitive signal amplier (for the sensor) and an analog or digital ter (for the controller). Ā is might be impractical in m any applications a nd h as mot ivated t he u se of t he s o-called shunt circuits, in which an electrical circuit is directly connected to a transducer embedded in the structure. Ā e tranducer acts as an energy converter: it transforms mechanical (vibrational) energy into electrical energy, which is in turn dissipated in the shunt circuit [5-8]. No seperate sensor is required, and only one, generally simple electronic circuit is used. It is generally passive, i.e. made of resistors and inductors, but it can also be active as proposed in [9]. Ā is chapter compares all these methods with each other. Ā e benchmark truss structure used for this purpose is described in Figure 26.1. It consists of 12 bays of 140 mm each, made of steel bars o f 4 mm d iameter c onnected w ith p lastic jo ints a nd i t i s clamped at t he b ottom. I t i s e quipped w ith t wo ac tive s truts (piezo transducer + force sensor) as indicated in the gure.

26.2

Piezoelectric Stack Actuator

Consider the piezoelectric linear transducer of Figure 26.2: it is made of n id entical s lices of pie zoceramic m aterial stacked together, each of them polarized through the thickness. I f o ne a ssumes t hat a ll t he me chanical a nd ele ctrical quantities a re o ne-dimensional a nd pa rallel to t he p oling direction, t he c onstitutive e quations f or t he p iezoelectric material are: ⎧D ⎫ ⎡⎢ e T d33 ⎤⎥ ⎧ E ⎫ ⎥ ⎨ ⎬ = ⎢⎢ ⎨ ⎬ (2 E ⎥ ⎩S ⎭ ⎢⎣d33 s ⎥⎦ ⎩T ⎭

6.1)

where t he s tandard I EEE not ations h ave b een u sed [ 10]. Integrating E quation 2 6.1 o ver t he vol ume o f t he t ransducer gives, with the notations of Figure 26.2: ⎧Q ⎫ ⎡⎢ C ⎨ ⎬ = ⎢⎢ ⎩ ∆ ⎭ ⎢⎣nd33

nd 33 ⎤⎥ ⎧V ⎫ ⎥⎨ ⎬ 1/K a ⎥⎥⎦ ⎩ f ⎭ (2

6.2)

where Q = nAD i s t he tot al ele ctric c harge on t he ele ctrodes of the transducer ∆ = Sl i s t he tot al e xtension ( l = nt i s th e l ength o f th e transducer) f = AT is the total force V t he vol tage app lied b etween t he ele ctrodes o f t he transducer, r esulting i n a n el ectric  eld E = V/t = nV/l 26-1

43722_C026.indd 1

10/3/2008 10:33:21 PM

26-2

Smart Materials

Force sensor Piezoelectric transducer (b)

Strut 1 Strut 1: Shunt + Measurements Strut 2

Z

Strut 2: Excitation

(c)

(a)

FIGURE 26.1

(a) Truss structure used in the experiment; (b) detail of an active strut; and (c) disposition of the active struts (zoom).

f

Cross-section: A Thickness: t # of disks in the stack: n l = nt

E = V/t Q

V

Q

t

Electric charge: Q = nAD Capacitance: C = n 2e A/l Free expansion: d = nd33V

f

FIGURE 26.2

inverting Equation 26.2 that the stiff ness of the transducer with open electrodes (Q = 0) reads:

Electrode

Piezoelectric linear transducer.

Ā e capacitance of the transducer with no external load (f = 0) is C = eT An2/l, a nd K a = A/sEl is the s tiff ness w ith s hortcircuited electrodes (V = 0 ). Ā e e lectromechanical coupling factor o f t he m aterial ( in t he d irection o f p olarization) i s another material constant; it is de ned by k2 =

2 2 d 33 n 2d 33 Ka = (2 E T s e C

6.3)

It measures the efficiency of the conversion of mechanical energy into electrical energy and vice versa. Finally, one nds upon

43722_C026.indd 2

∆ Ka = (2 f Q = 0 (1 − k 2 )

6.4)

which has important consequences in the following. Various r ealizations o f c ommercially a vailable p iezoelectric uniaxial t ransducers a re p resented i n F igure 2 6.3. Ā e rst one (Figure 26.3a) is the one used in this work. Realizations (a) and (b) follow exactly the principle of Figure 26.2. Ā e latter is delivered with an in-built external prestressing mechanism while the former requires the introduction of an internal wire to this end. In the (c) and (d) designs, the piezoelectric stack is mounted along the main axis of an elliptical structure, which acts as a motion amplier. Ā e expansion of the actuator is in the direction perpendicular to that o f th e p iezostack. Ā is design leads to more compact (but softer) actuators. Ā e gure also shows force sensors and exible tips that prevent torques to be applied on the stack.

26.2.1

Structure with a Stack Transducer

Suppose t hat t he p iezoelectric t ransducer de scribed i n t he previous section is embedded in a l inear, undamped structure. A voltage V is applied across the electrodes and an electric charge Q ows into it (there is a relation between Q and V but it has no impact o n t he d iscussion). T he dy namic e quations o f t he structure alone, i.e., without the piezo, read classically:

10/3/2008 10:33:22 PM

26-3

Active Truss Structures PI 840.30

Pl P010-30H

(a)

Piezo stack

Prestressing wire

(b)

APA 100 M

APA 50 S

(d)

(c)

Piezo stack

FIGURE 2 6.3 Realizations of u niaxial t ransducers ( a) a nd ( b) c lassical d esign (P hysik I nstrumente G mbh), ( c) a nd ( d) a mplied design (CEDRAT). In (a), (b), and (c), a force sensor is collocated with the actuator. (c) and (d) are equipped with exible joints.

Mx + Kx = F (2

6.5)

where K and M are the stiffness and mass matrices of the structure (without t he piezo), obtained, for e xample, by me ans of a nite element model, and F is the external force vector. Here we consider that the only forces exerted on the structure come from the transducer, i.e. F = bf (2

6.6)

where b is the projection vector relating the end displacements of the strut to t he global coordinate system, and f is the force exerted by the piezo (Equation 26.2). Taking into account that ∆ = bTx, Equations 26.5 and 26.6 can be combined with Equation 26.2; one gets 2

T

Ms x + (K + K abb )x = bK and 33V (2 C (1 − k 2 )V + nd 33 K ab T x = Q (2

6.7) 6.8)

where K + Ka bbT is the global stiff ness matrix (including the transducer) and s is the Laplace variable. 26.2.1.1

Various Natural Frequencies

Analyzing further Equations 26.7 and 26.8, three different eigenvalue problems can be dened, corresponding to t he boundary conditions f = 0, V = 0, and Q = 0, respectively. 1. From E quation 2 6.7, t he eigen value p roblem w hen t he axial stiff ness of the actuator is cancelled, i.e., Ka = 0, i s given by: ( Ms 2 + kx ) = 0 (2 2.

43722_C026.indd 3

6.9)

If V = 0, i .e., if t he piezo is short-circuited, t he structure obeys the following relation:

⎛ ⎜⎝

M s2 + K + K abb T ⎞⎟⎠ x = 0 (2

6.10)

3. Finally, if the structure is charge-driven instead of voltagedriven, V can be eliminated from Equations 26.7 and 26.8 and the new equation is ⎛

Ms 2 x + ⎜⎜ K + ⎜⎝

⎞ Ka K Q bb T ⎟⎟ x = b a 2 nd 33 (2 2 ⎟ C 1− k 1− k ⎠

6.11)

and if the structure is open-circuited, i.e., Q = 0, it obeys the following relation: ⎛ ⎜ ⎜ ⎜⎝

Ms 2 + K +

⎞ Ka bb T ⎟⎟ x = 0 (2 2 1− k ⎠⎟

6.12)

which is the same as Equation 26.10 but with the short-circuit stiffness Ka replaced by the open-circuit one Ka/(1 − k2 ). Ā e solutions of these eigenvalue problems are three different sets of natural frequencies. In this work, the natural frequencies when Ka = 0 a re called zi (i = 1 , …, n), t hose when t he piezo is short-circuited are called ωi, and those when the piezo is opencircuited a re c alled Ωi. B ecause t he s tiff ness o f t he t ransducer increases when it is left open (Equation 26.4), one has Ωi > w i.

26.3 Integral Force Feedback Ā is section considers the active damping of the truss based on a force s ensor c olinear w ith t he p iezoelectric t ransducer. U pon inverting Equation 26.2, the output force f is given by f = K a (b T x − nd 33V ) (2

6.13)

where once a gain bTx = ∆ i s t he a xial elongation o f t he t ransducer. Ā e IFF consists of V (s ) =

g y (s ) nd 33 K a s

(26.14)

10/3/2008 10:33:24 PM

26-4

Smart Materials

where g is the gain of the controller and the coefficient 1/nd33 Ka has b een i ntroduced f or t he p urpose o f no rmalization. Combining E quations 2 6.7, 2 6.13 a nd 2 6.14, t he c losed-loop poles are given by ⎡ 2 ⎤ g T (K abb T )⎥ x = 0 (2 ⎢ Ms + (K + K abb ) − s g + ⎣ ⎦

6.15)

As g → 0, E quation 26.15 becomes identical to E quation 26.10, which denes the short-circuited natural frequencies w i. On the other h and, t he a symptotic ro ots a s g → ∞ a re t he op en-loop zeros zi (Equation 26.9), which correspond to the situation where the a xial s tiffness o f t he ac tive s trut h as b een remo ved. I f t he natural f requencies o f t he s tructure a re w ell s eparated, e ach mode is independent of the others. In these conditions, Equation 26.15 can be reduced to a set of uncoupled Equation 26.4: 1+ g

s 2 + z i2 =0 s ⎛⎜⎝ s 2 + w i2 ⎞⎟⎠

(26.16)

Such a root locus (starting from w i and ending at zi) is presented in Figure 26.4a when zi and w i are not too far apart. In this case, the maximum modal damping is given by [4]: x max =

wi − z i ⎛ wi ⎞ ⎜ z i ≥ ⎟⎠ 2z i ⎝ 3

(26.17)

When the distance between zi and w i increases, the shape of the root locus changes as described in Figure 26.4b. For zi < w i/3, the root locus touches the real axis, which means that critical damping c an b e at tained. F igure 2 6.4 i ndicates t hat t he c losed-loop poles are in the left half plane (and thus stable) as long as g > 0: this g uaranteed s tability i s a c onsequence o f t he c ollocation between t he ac tuator a nd t he s ensor. It rem ains v alid e ven for multimode s tructures, a s lo ng a s p erfect ac tuator a nd s ensor dynamics are assumed.

26.3.1

Extension to Cable Structures

Ā e u se of c ables to ac hieve l ightweight spacecrafts i s not ne w: it can, f or ex ample, be f ound i n H erman O berth’s ea rly boo ks o n astronautics. In terms of weight, the use of guy cables is probably the most efficient way to s tiffen a structure; cables can also be used to prestress a deployable structure and eliminate the geometric uncertainties due to the gaps. Ā e dynamic modeling of such a structure and t he de sign of a rob ust c ontroller a re however c omplex t asks, because of the nonlinear behavior of the cables and because of the complex interactions between the cables and the structure. In this context, the use of a collocated controller with guaranteed stability such as the IFF is very appreciated. Indeed, it is easy to show that this control law is passive, i.e., it always extracts energy from the system, even in presence of large nonlinearities. Tendon control with IFF has been investigated, for example, in Refs. [11–13]. It has been found that, as long as the dy namics of t he c ables c an b e ne glected, t he c losed-loop poles follow exactly the same root locus as that of the previous section (Equation 26.16). Ā is root-locus is once again shown in Figure 26.5. Here w i is identified as the structural natural frequencies when the cable is connected, and zi as those when the cable is removed. Ā e theory is readily extended to the use of s everal de centralized ac tive tendons w ith t he s ame ga in. In this case, w i corresponds to the structure when all the cables are c onnected, a nd t he zi to that when all the cables are removed. E xperiments h ave s hown t hat E quation 2 6.16 p rovides an accurate estimate of the closed-loop poles.

26.4

Passive Shunt Damping

When a p iezoelectric s tructure i s c onnected to a n ele ctrical circuit, the voltage V across the electrodes of the transducer and the current I that ows into them are related by: I (s ) = −Yshunt (s )V (s )

(26.18)

Im (s) ximax

ximax =

wi

w

zi

z1 z2

wi – zi 2zi

z3

IFF

z4

(a)

Re (s)

(b)

FIGURE 26.4 (a) Root locus of the IFF for a single-mode structure and (b) evolution of the locus as the zero zi moves toward the origin. Only one half of the locus is shown; the dotted lines correspond to the maximum damping.

43722_C026.indd 4

10/3/2008 10:33:29 PM

26-5

Active Truss Structures

Natural frequency with the active cables 1+g

(s 2 + zi2 ) s (s 2 + wi2 )

=0

wi

Im (s) ximax

wi zi

ximax =

wi – zi 2zi

zi

Re (s)

Active cables removed

FIGURE 26.5

Open-loop poles and zeros and root locus of an active cable structure with IFF.

where Yshunt is the admittance of the shunt circuit and the minus sign has been introduced to be consistent with the directions of V and I dened in Figure 26.2 (I is the derivative of Q, which in Laplace va riable beco mes I = sQ). I ntroducing E quation 2 6.18 into Equation 26.7 and 26.8 the following eigenvalue problem is obtained: ⎛ 2 ⎞ k2 K abb T T ⎜ Ms + K + K abb + (1 − k 2 ) 1 + Y /SC (1 − k 2 ) ⎟ x i = 0 ⎝ ⎠ shunt

With Yshunt = 1/R, Equation 26.19 becomes: ⎡ 2 ⎤ k K abb T ⎢ Ms + (K + K abb ) + ⎥x = 0 2 (1 − k ) + 1/sRC ⎥⎦ ⎢⎣ T

43722_C026.indd 5

1 s 2 + w i2 = 0 (2 2 RC (1 − k ) s(s 2 + Ωi2 )

6.22)

Despite t he fact t hat t he IFF a nd t he resistive shunting share a similar root-locus, t hey have very d ifferent performances, as is experimentally demonstrated below.

26.4.2

Inductive Shunting

Yshunt (s ) = ( 26.20)

Going into modal coordinates, all the modes uncouple into a set of independent characteristic equations: 1+

Ωi − w i (2 2w i

When the shunt consists of a resistance in series with an inductance, its admittance is

Resistive Shunting

2

ximax ,R =

(26.19)

It becomes identical to Equation 26.10 when Yshunt → ∞, leading to t he f requencies w i ( short-circuited), a nd i t i s iden tical to Equation 26.12 when Yshunt → 0, leading to Ωi (open electrodes).

26.4.1

According to Equation 26.21, the closed-loop poles follow a root locus similar to that of the IFF (Figure 26.4) when R varies from +∞ to 0. Ā e open-loop poles are in this case at ±jΩi (open electrodes) and the open-loop zeros are at ±jΩi (short-circuited). Just as with the IFF, the maximum achievable damping is given by

6.21)

1 (2 Ls + R

6.23)

Ā e c onnection o f t he s hunt i nductance w ith t he i nherent capacitance of the piezo creates an electric resonance which, if properly tuned, can increase the damping over that of a R shunt. With the same methodology as in the previous section, Equation 26.19 can once more be rearranged in a root locus form: 1 + 2xew e

s(s 2 + Ωi2 ) = 0 (2 s + (Ωi2 + w e2 )s 2 + w i2w e2 4

6.24)

10/3/2008 10:33:32 PM

26-6

Smart Materials

p1 we >

Ω2i wi

p1 we =

jΩi

Ω2i wi

jΩi

Q

jwi p2

jwi p2 Optimal damping

Resistive shunting

– Ωi

(a)

(b)

– Ωi

FIGURE 26.6 Root locus of the structure with a series RL shunt. (a) Ā e electrical frequency w e is not optimally tuned and (b) Optimal tuning of w e.

where we have dened the electrical frequency and damping: w e2 =

R 1 2xew e = (2 LC (1 − k 2 ) L

6.25)

Two pairs of poles p1 and p2 are now present; their evolution as xe changes is shown in Figure 26.6 for two values of w e. Ā e locus consists of two loops, starting respectively from p1 and p2; one of them goes to jΩi and the other goes to the real axis, near − Ωi. If w e > Ωi2/w i (Figure 26.6a), the upper pole is always more heavily damped t han t he lower o ne, w hile t he opp osite o ccurs i f w e < Ωi2/w i. I f w e = Ωi2/w i ( Figure 2 6.6b), t he t wo p oles a re a lways equally damped u ntil t he t wo branches touch each ot her i n Q. Ā is double root can be regarded as the optimum tuning of the inductive shunting; the corresponding eigenvalues satisfy 1/2

⎛ Ω2 ⎞ s 2 + Ωi2 + Ωi ⎜ 2i − 1⎟ s = 0 (2 ⎝ ωi ⎠

6.26)

and the corresponding damping ratio can be obtained easily by identifying the previous equation with the classical form of the damped oscillator, s2 + 2xi Ωis + Ωi2 = 0, leading to xi = ξimax ,RL =

1 Ωi2 − w i2 (2 w i2 2

6.27)

Ā is value is signicantly higher than that achieved with purely resistive s hunting. I t i s app roximatively t he s quare-root o f Equation 26.22. Note, however, that it is much more sensitive to the tuning of the electrical parameters on the targeted modes.

26.4.3

Generalized Electromechanical Coupling Factor

Ā e quantity K i2 =

43722_C026.indd 6

Ωi2 − w i2 w i2

(26.28)

is k nown a s t he gener alized ele ctromechanical coupling factor for mo de i. I t c ontrols t he c onversion me chanism b etween mechanical and electrical energy. It can be shown [14] that K i2 ≈

k 2 ni (2 1 − k2

6.29)

where k is the material electromechanical coupling factor dened in Equation 26.3 ni is the fraction of modal strain energy in the piezoelectric transducer for mode i ni qu anties t he c apability to c oncentrate t he v ibration s train energy of mode i in the transducer, and k quanties the conversion mechanism in the material. Ā e performances of the R and RL s hunts a re d irectly rel ated to Ki; i ndeed, E quations 2 6.22, 26.27, and 26.28 give ximax ,R ≈

26.5

K i2 Ki (2 and ximax ,RL = 4 2

6.30)

Active Shunt Damping

Equation 2 6.29 h ighlights t he i mportance o f a h igh k in t he shunt d amping p erformances. Ā e p iezoelectric m aterial le ad zirconate titanate (PZT) typically has k ≈ 0.7; it is currently difcult t o  nd m aterials w ith h igher v alues, e ven i f p romising materials suc h a s P MN-PT w ith k as high as 0.9 have been recently reported. k can also be articially increased by the use of an active electronic circuit that simulates the behavior of a negative capacitance. Ā e concept was rst introduced by Forward in 1979 [9]. It remained however l argely f orgotten f or 2 0 ye ars, u ntil s ome re searchers reintroduced the idea [15–17]. Since then, the idea has been used in various practical applications (e.g., [18–20]). Ā e “negative capacitance” by itself does not dissipate energy, but it does enhances the dissipation of energy in the passive (R or RL) electric circuit.

10/3/2008 10:33:35 PM

26-7

Active Truss Structures

Q

–C⬘

Vt

Q

Vt

Vp

(a)

(b)

Shunted piezo

Equivalent piezo (k ∗ > k, K *a < K a)

FIGURE 26.7 Piezoelectric transducer shunted on a series negative capacitance −C′ and equivalent transducer.

Consider a piezoelectric linear actuator shunted in series by a negative c apacitance c ircuit o f v alue − C′ ( Figure 2 6.7a): o ne immediately nds Vt = V p − Q /C ′

(26.31)

where Vp is the voltage across the transducer Q is the electrical charge owing into it Introducing Equation 26.31 into Equation 26.32, it is found that the s hunted t ransducer i s e quivalent to a “ simple” piezo transducer with a higher electromechanical coupling coefficient k∗ and smaller stiffness Ka∗ (Figure 26.7b). More precisely, one has: ∗

Ka =

K a (C ¢ − C ) C ¢ − C (1 − k 2 )

∗

k 2=

k 2C ¢ (2 C ¢ − C (1 − k 2 )

6.32)

Ā e op en-circuit s tiffness (dened i n E quation 2 6.4), on t he other hand, does not change: ka∗/(1 − k∗2) = ka/(1 − k2). A series negative capacitance thus decreases the short-circuit natural frequencies ωi∗ and has no impact on the Ωi∗, which in turn i ncreases t he p erformances o f t he R an d RL sh unts (Equations 26.22 and 26.27). Ā e negative capacitance is an active shunt that can become unstable if improperly tuned. A stability analysis showed that a sufficient (but not necessary) stability condition reads: C ′ > C (2

6.33)

with C the constant-force capacitance of the transducer (this value of C′ leads to k∗ = 1 a nd Ka∗ = 0). In practice, reasonable stability margins must be taken because of t he large nonlinearities of t he piezoelectric materials. Ā roughout this work and for the purpose of illustration, the negative capacitances were tuned as close as 90% of the stability limit. Ā is was made possible by the accurate knowledge of the structural impedance. In realistic situations where the uncertainties on the impedance are much larger, negative circuits at 60–70% of the limit would be more advisable. Unfortunately, the performances rapidly drop when C′ moves away from the stability limit. A negative capacitive circuit can be implemented as described

43722_C026.indd 7

in [ 21]. I t i s a si mple c ircuit i nvolving a n op erational a mplier. Note that the required voltage level can be quite high.

26.6 Experimental Results In this section, the different d amping techniques a re quantitatively evaluated and compared with each other. To this end, they are implemented on the benchmark truss structure represented in Figure 26.1. We focus on t he damping of t he  rst structural mode, a nd t he c ontroller or s hunt c ircuit i s applied on s trut 1 only (a mo dal a nalysis s howed t hat t his s trut h as a re asonable authority o n t he t argeted mo de). I n t his w ork, s trut 2 i s u sed only as a mean to excite the structure.

26.6.1

Frequency Response Function (FRF)

Ā e FRF between t he voltage at s trut 2 a nd t he force in strut 1 (see Figure 26.1a) is measured for different shunt circuits. Ā ese FRFs a re s hown i n F igure 2 6.8,  rst w hen t he s hunt i s p urely resistive, then when a negative capacitance is added. Ā e R shunt alone can raise x1 from 0.18% to 0.43%. Ā e negative capacitance shunt can raise it further to approximately 3%. Ā e same measurements are presented in Figure 26.9, but with a RL shunt. Ā e damping ratio of the two pairs of poles p1 and p2 is measured to be 2.7% and 2.5%, respectively, while the addition of a ne gative c apacitance b rings t hese d amping r atios to 9 .3% a nd 5.4%, re spectively. I t i s w orth not icing t hat a pa ssive RL sh unt alone performs as well as an active “negative capacitance + R” shunt. Finally, t he IFF controller is i mplemented. Ā e corresponding FRF is a lso shown i n Figure 26.9. T his control law was found extremely easy to implement as a result of its guaranteed stability.

26.6.2 Root Loci Ā e structural poles have been recorded for various values of the gain g of the IFF or of the resistance R of the shunt; these poles are p resented i n F igure 2 6.10 (each dot re presents a me asurement). Ā e solid lines represent theoretical predictions based on Equation 26.16 and on the identied w1, Ω1, and z1; the agreement between theory and experiment is very good.

10/3/2008 10:33:38 PM

26-8

Smart Materials

20

Open circuit

Amplitude (dB)

10

R only

0

R + negative capacitance

−10 −20 −30 −40 14

FIGURE 26.8

16

18

20 22 Frequency (Hz)

24

26

30

35

28

Structural response with an active and a passive resistive shunt.

20 10 Open circuit

Amplitude (dB)

0 RL only

−10

RL + negative capacitance

−20 −30 −40

Integral force feedback

−50 −60 5

FIGURE 26.9

10

15

20 25 Frequency (Hz)

40

Structural response with an active and a passive RL shunt and with the active IFF controller.

20 IFF 15

5 z1 −20

FIGURE 2 6.10 measurement.

43722_C026.indd 8

−15

−10 Real (Hz)

−5

Im (Hz)

10

R shunt IFF

Negative capacitance

Ω1 w1

w 1∗

0

Identied ro ot lo cus of t he I FF a nd of t he R s hunts ( with a nd w ithout t he ne gative c apacitance). E ach dot re presents a

10/3/2008 10:33:39 PM

Active Truss Structures

It is found t hat t he root locus of t he I FF is extremely large: critical d amping c an b e at tained. Ā is is d ue t o t he p hysical nature of the zero z1, which corresponds to the natural frequency of the truss when the axial stiff ness of the transducer is removed, i.e., almost a mechanism. On the other hand, the loci obtained with R shunts (even when a negative capacitance is added) are much smaller, in spite of the relatively large values of ni and k2. Figure 2 6.10 h ighlights th e f act th at th e a ttainable p erformances with a separate collocated transducer/force sensor pair are o ne o rder o f m agnitude l arger t han t hose t hat c an b e attained with a shunt circuit, even active. Ā e boundary conditions leading to the different eigenfrequencies ω1, Ω1, and z1 are determinant w hen e valuating t he p otential p erformances o f such a control system.

Acknowledgments Ā is study was supported by the Inter University Attraction Pole IUAP 5 o n A dvanced M echatronic S ystems a nd b y EU -FW6 within t he p rojects I NMAR: I ntelligent M aterials f or A ctive Noise re duction a nd C ASSEM: C omposite a nd A daptative Structures: Simulations, Experimentation and Modeling.

References 1. J.L. Fanson, G.H. Blackwood, and C.C. Chu. Active-member control of pre cision s tructures. I n SDM C onference, pp . 1480–1494. AIAA paper 89-1329 CP, 1989. 2. G.S. Chen, B.J. Lurie, and B.K. Wada. Experimental studies of a daptive s tructures for pre cision p erformance. I n SDM Conference, pp. 1462–1472. AIAA paper 89-1327 CP, 1989. 3. A. Preumont, J.P. Dufour, and C. Malekian. Active damping by a local force feedback with piezoelectric actuators. AIAA Journal of Guidance, 15(2):390–395, 1992. 4. A. Preumont. Vibration Co ntrol o f Active S tructures: a n Introduction, 2nd ed., Kluwer, Dordrecht, 2002. 5. R.L. F orward. Ele ctronic da mping o f vib rations in o ptical structures. Applied Optics, 18(5):690–697, 1979. 6. N.W. H agood a nd A. v on Flo tow. D amping o f str uctural vibrations wi th p iezoelectric m aterials a nd pas sive e lectrical networks. Journal of Sound and Vibration, 146(2):243–268, 1991. 7. J.J. H ollkamp. M ultimodal p assive vib ration su ppression with piezoelectric materials and resonant shunts. Journal of I ntelligent M aterial S ystems and Str uctures, 5:49–57, 1994.

43722_C026.indd 9

26-9

8. S.Y. Wu. Method f or multiple mo de piezoelectric shunting with single PZT transducer for vibration control. Journal of Intelligent Material Systems and Structures, 9:991–998, 1998. 9. R.L. Forward. Electromechanical Transducer-Coupled Mechanical S tructure wi th N egative C apacitance C ompensation Circuit. U.S. patent 4,158,787, 1979. 10. IEEE S td. IEEE S tandard o n P iezoelectricity, 1988. ANSI/ IEEE Std 176-1987. 11. Y. Achkire. Active C ontrol o f C able-Stayed B ridges. P hD thesis, U niversité Lib re de B ruxelles (ULB), B elgium, 1997. 12. A. Preumont, Y. Achkire, and F. Bossens. Active tendon control of large trusses. AIAA Journal, 38(3):493–498, 2000. 13. A. Preumont and F. Bossens. Active tendon control of vibration of t russ s tructures: Ā eory and e xperiment. Journal o f Intelligent M aterial S ystems a nd S tructures, 11(2):91–99, 2000. 14. A. Preumont. Mechatronics: Dynamics of Electromechanical and Piezoelectric Systems. Springer, 2006. 15. S.-Y. Wu. B roadband p iezoelectric sh unt f or str uctural vibration control. U.S. Patent 6,075,303, 2000. 16. M. Date, M. Kutani, and S. Sakai. Electrically controlled elasticity u tilizing p iezoelectric co upling. J. A pplied Physics, 87(2):863–868, 2000. 17. J. Tang and K.W. Wang. Active-passive hybrid piezoelectric netw orks f or vib ration co ntrol: co mparisons a nd improvement. Smart M aterials a nd S tructures, 10:794– 806, 2001. 18. J.S. Kim, K.W. Wang, and E.C. Smith. High-authority piezoelectric ac tuation system syn thesis t hrough me chanical resonance a nd ele ctrical t ailoring. Journal of I ntelligent Material Systems and Structures, 16:21–31, 2005. 19. M. N eubauer, R . Oleskie wicz, K. Popp, a nd T. K rzyzynski. Optimization o f da mping a nd a bsorbing p erformance o f shunted p iezo elemen ts u tilizing nega tive ca pacitance. Journal of Sound and Vibration, 298:84–107, 2006. 20. H. Yua, K.W. Wang, a nd J. Zha ng. P iezoelectric netw orking with e nhanced e lectromechanical co upling f or vi bration delocalization o f mist uned p eriodic st ructures-theory a nd experiment. Journal o f S ound a nd Vibration, 295:246–265, 2006. 21. Philbrick Researches Inc. Application manual for computing a mpliers f or mo delling, me asuring, ma nipulating & much els e. Technical r eport, N imrod Pr ess, B oston, MA, 1965.

10/3/2008 10:33:40 PM

43722_C026.indd 10

10/3/2008 10:33:40 PM

27 Application of Smart Materials and Smart Structures to the Study of Aquatic Animals 27.1

Introduction and Chapter Overview................................................................................... 27-1

De ning Smart Materials and Smart Structures • Applications for Smart Materials and Smart Structures in the Study of Animal Behavior • Chapter Overview

27.2 Biotelemetry or Biologging: Introduction .......................................................................... 27-2 27.3 Biotelemetry or Biologging: Equipment .............................................................................27-4 Commercially Available Transmitters and Receivers

27.4 Biotelemetry and Biologging: Applications and Findings ................................................ 27-7

Jesse E. Purdy Southwestern University

Alison Roberts Cohan Pacific Whale Foundation

27.5

Locating and Tracking Marine Animals: Fishes, Turtles, and Cephalopods • Locating and Tracking Marine Animals: Marine Mammals • Assessing Energy Expenditure with Acoustic Telemetry: Fishes, Cephalopods, and Turtles • Assessing Energy Expenditure with Acoustic Telemetry: Marine Mammals • Assessing Energy Expenditure Using Video Telemetry: Fishes, Cephalopods, and Turtles • Assessing Energy Expenditure Using Video Telemetry: Marine Mammals

Predicting the Future........................................................................................................... 27-16

Life History Transmitter • Biotelemetry and Smart Dust • Methods of Testing Behavioral Hypotheses Using Biotelemetry

27.6 C onclusions .......................................................................................................................... 27-18 References ......................................................................................................................................... 27-18

27.1

Introduction and Chapter Overview

27.1.1

Defi ning Smart Materials and Smart Structures

Smart materials sense aspects of t he environment and respond appropriately. Ā ey also reverse that response when conditions change. For example, ferroelectric-based piezoelectrics generate electric polarity when subjected to mechanical stress. Conversely, if o ne app lies a vol tage to p iezoelectric c rystals, t he c rystals respond by changing shape in a linear direction. Electrostrictive and magnetostrictive a re a lso smart materials. Ā e se materials change size in response to electric or magnetic stimuli, respectively, and produce a voltage when stretched. Rheological smart

materials a re  uids that change state instantly in response to electrical (electrorheological) or magnetic (magentorheological) stimuli. E lectrochromic m aterials h ave t he a bility to c hange their optical properties when an electric current is applied. And smart ge ls s hrink or s well by f actors of 1000 or more a f actor useful to absorb or release uids. Similarly, smart fabrics are fabrics with a special coating that causes them to contract when strain surpasses a specied limit. Ā e applications for smart materials are numerous. In snow skis, p iezoelectric m aterials de tect v ibrations a nd c ause ot her piezoelectrics to cancel that vibration (Ashley, 1996). Ā e resulting re duction i n v ibration ke eps mo re o f t he s ki o n t he sno w and can be used to keep the tires of a car on the road or reduce vibrations in buildings and bridges (Gibbs, 1996; Perkins, 1997). In addition, piezoelectrics are used in sensors that cause airbags

27-1

43722_C027.indd 1

10/16/2008 7:15:37 PM

27-2

Smart Materials

to deploy. Electrostrictive and magnetostrictive smart materials are found i n pumps, w ind tunnels, a nd landing gear a nd t hey also  nd application i n t he health sciences. R heological smart materials a re  uids also used to reduce vibration and electrochromic m aterials u nderlie l iquid c rystal d isplays ( LCD). Because sm art gel s b ecome i ncreasingly s tiff in response to mechanical stress, they have been used in ski outts to reduce the c hance o f i njury. Si milarly, t he sm art b ra i s m ade f rom a fabric with a special coating that causes it to contract when strain surpasses a specied limit. Ā e smart bra is ideal for exercise in that it stiffens and provides better support when needed (Collins, 2003). Smart m aterials c an b e c ombined to p roduce sm art s tructures. Ā ese structures perform sensing, controlling, and actuation activities. Or, as Cao et al. (1999) state, “a critical element of smart structures is the development of an optimized control algorithm that guides the actuators, following detection of certain s ensory s timuli.” U sing f eedback c ontrol me chanisms, smart structures can provide precise active damping functions and a re u sed f or t hat p urpose i n t he s pace s huttle a nd s pace station. Interestingly, Cao et al. predict that future smart structures will be used to develop supersensitive noses, ears, and eyes t hat a re smart enough to d irectly communicate w ith t he human brain. Such structures could enable humans to expand their s ensory i nput e xponentially. W e c ould h ear t he i nfrasounds produced by ele phants a nd t he u ltrasounds produced by bats and many marine mammals. We could see infrared like snakes, de tect u ltraviolet r adiation l ike a gold sh, a nd smel l the world like dogs. Ā ese abilities would have a profound effect on humans and additionally could enlighten us as to what it is like to be an elephant, a bat, a goldsh, or a dog. By experiencing t he s ensory w orlds o f no nhuman a nimals t hrough sm art materials a nd sm art s tructures, w e c ould le arn mo re a bout nonhuman animal species. And that brings us to the main topic o f t his c hapter. W hat c an sm art m aterials a nd sm art structures tel l u s a bout t he b ehavior a nd ph ysiology o f a nimals, specically, aquatic animals?

27.1.2

Applications for Smart Materials and Smart Structures in the Study of Animal Behavior

Smart m aterials a nd sm art s tructures h ave f ound w idespread application in the study of animal behavior. Often, these applications involve apparatuses that convert responses in the form of mechanical en ergy i nto el ectrical en ergy u sing f erroelectricbased piezoelectrics. Piezoelectric ceramics convert energy from one f orm to a nother a nd a s suc h a re t ransducers. E xamples include m icrophones, h ydrophones, lou dspeakers, t hermometers, light emitting diodes (LEDs), photocells, position and pressure s ensors, a nd a ntennas. T ransducers c an a lso b e u sed to transmit ultrasonic signals in air or water. Scientists studying animal behavior have used smart materials and sm art s tructures to s tudy a gonistic b ehavior, f oraging, homing, learning, migration, and territoriality, in the  eld and

43722_C027.indd 2

in the laboratory. For example, Wolcott and his colleagues have developed smart structures to s tudy t he behavior of blue crabs (Shirley a nd Wolcott, 1991; Wolcott, 1996; Wolcott a nd H ines, 1996; Clark, Wolcott, T., Wolcott, D., a nd Hines, 1999). In one application, Clark et al. (1999) used microcontroller-based smart transmitters and receivers to study the feeding and threat behaviors o f bl ue c rabs i n t heir nat ural en vironment. S ensors t hat detected and transmitted a sig nal indicating contraction of the mandibular ( chewing) m uscle w ere at tached to t he c rab. Contractions indicated that the animal was feeding. In addition, sensors were attached to the both claws. Ā e s pread o f b oth chelae a sig nal i ndicated t hat a mer al t hreat d isplay h ad b een emitted. W hen p opulation den sity w as h igh, c rabs e xpended more effort defending their territory and behaved more aggressively. Ā is increased aggression reduced the amount of time the crabs spent feeding. Ā us, high population density reduced feeding because there was a marked increase in antagonistic behavior a nd not b ecause o f l imited a vailability o f f ood. A dditional studies examined molting behavior (Shirley and Wolcott, 1991). Blue crabs were tracked using ultrasonic transmitters and reed switches and magnets were used to de tect molting. Under high predation p ressure, c rabs s ought s hallow w ater a nd mo ved to areas of less predation. Low predation pressure allowed molting to occur over a broader range.

27.1.3 Chapter Overview Ā ough the application of smart materials and smart structures is widespread in the study of animal behavior, the focus of this chapter is on telemetry studies with aquatic animals, specically within the m arine environment. Aquatic a nimal re searchers h ave u sed telemetry technology to t rack animals to de termine home range and movement pat terns, qu antify muscle movement, a nd determine energ y u se a nd re quirements t hrough me asures o f h eart rate, ph ysical ac tivity, a nd i nternal tem perature. I n add ition, telemetry studies inform how animals budget their time between foraging,  ghting, resting, mating, etc. Ā e c hapter b egins w ith specic de tails a bout t he b iotelemetry e quipment t hat aqu atic animal researchers use and representative companies that provide such e quipment c ommercially. Ā is s ection will b e followed by examples o f t he me thods a nd  ndings f rom re searchers t hat incorporated t his t echnology. Ā e literature in this eld is growing rapidly a nd t his chapter is not me ant to b e exhaustive. Our goal is to provide representative examples of equipment and applications that the reader can then explore in greater detail.

27.2

Biotelemetry or Biologging: Introduction

Aquatic a nimals b y v irtue o f c amouf lage a nd l ifestyle a re difficult to ob serve f or lo ng p eriods o f t ime. H owever, suc h observations a re required to u nderstand t heir basic physiology and behavior. For example, does the animal expend a great deal of energ y to s wim lo ng d istances o n a d aily ba sis l ike s ome sharks or do es it l ive a s edentary l ifestyle l ike c uttlesh? Does

10/16/2008 7:15:38 PM

Application of Smart Materials and Smart Structures to the Study of Aquatic Animals

the a nimal spend its entire life in one relatively small area like rocksh or does it migrate great distances like some whales? In addition, it i s i mportant to k now how t he environment a ffects the a nimal. F or e xample, do es t he tem perature r ange o f t he water animal restrict the animal’s movement and is the animal conned to a narrow or wide range of depth? Ā ese questions are addressed b y tel emetry s ystems. I n t hese s ystems, t ransmitters determine location and sense aspects of the animal and its environment. D epending on t he t ransmitter, t he d ata a re s tored or transmitted to a re ceiver o r d ata s canner. C ertain t ransmitters even provide preliminary analyses of the data. Receivers and data scanners co nnect t o a co mputer t hat decodes , p rocesses, a nd stores that information to produce a useful result. Ā us, telemetric s ystems a re sm art s tructures a nd m any o f t he t ransmitters and certain of the scanners/receivers utilize smart materials. Remote ob servation o f a nimals t hrough teleme try e volved through four stages (Ropert-Coudert and Wilson, 2005). First, it was realized that animals could carry telemetric devices without compromising u nduly t heir nat ural b ehavior. S econd, de vices were developed that sensed basic behavioral, environmental, or physiological data and enabled scientists to locate and track animals. Ā is seco nd st age, co mbined w ith aco ustic a nd s atellite telemetry, allowed scientists to s tudy migratory movement and habitat u se. Ā e t hird s tage s aw t he de velopment o f teleme try devices t hat placed t he a nimal’s ac tivities on a t imeline. Ā ese devices recorded telemetry data for subsequent analysis and led to studies where time budgets could be determined and energy use could be assessed. Ā us, scientists could keep a record of the animal’s physiology or behavior over time. Naito (2004) termed this activity “biologging.” Ā e fourth stage saw the development of devices that not only monitored physiological and behavioral data, but also monitored and stored or transmitted information about th e an imal’s e xternal e nvironment. W ith th is d evelopment, it became possible to p lace t he a nimal’s behavior w ithin the c ontext o f i ts en vironment a nd t hereby ac hieve a n u nderstanding of how the environment affects the animal’s behavior and in turn, how the animal affects its environment. Ā ese f our s tages a re appa rent i n a teleme try s ystem de veloped b y D avis e t a l. (2004). D avis e t a l. de veloped a n a nimalborne v ideo a nd d ata re corder t hat ut ilized t ransducers f or pressure, water speed, and compass bearing. Ā e se transducers were sampled once per second and data stored on a ash memory card. Ā e video component recorded up to 6h of video data from the animal’s perspective on Hi 8 cassettes. Ā e two audio channels o n t he v ideo re corder w ere a lso em ployed. O ne c hannel recorded ambient sound received by a hydrophone affixe d to the animal and the second channel recorded the analog signal from an accelerometer. Ā e accelerometer was sampled 16 t imes per second and determined  ipper stroke frequency. A g lobal positioning system (GPS) was attached to the head-mounted camera and could determine the seal’s precise location in less than 30 s. With this apparatus, Davis et al. were able to observe the types of prey W eddell s eals p ursued, t he s trategies t hey u sed to h unt, obtain e stimates of t he energ y u sed during d ives, a nd re create three-dimensional tracks of their dives. Davis et al. also reported

43722_C027.indd 3

27-3

plans to d evelop t he ne xt ge neration v ideo d ata a cquisition platform t hat w ill mo nitor s everal t imes p er s econd, a mbient dissolved oxygen, ambient temperature, bioluminance, compass bearing, conductivity, light level, pressure, swim speed, and tilt, pitch, and roll. In addition, a d igital video recorder will record for 80 h and a GPS will acquire precise location in a few seconds. Figure 27.1 depicts a female Weddell seal with video data acquisition platform attached. Ā e de velopment o f s ophisticated sm art teleme try de vices have allowed scientists to proceed from manual tracking used to study g ross mo vement to t he s ystematic i nvestigation o f c omplex behaviors including exploration, foraging, habitat selection, migration, navigation, predator–prey interactions, reproductive, and social behavior. Ā ese devices provide data from free-ranging aquatic animals of such high quality and quantity that it is becoming possible to assess cause and effect relations in the eld as opposed to just the laboratory. Ā e outcome is that scientists can now maximize internal and external validity. Ā e d iscussion b egins w ith c onsideration o f t he e quipment required to lo cate a nd t rack aquatic a nimals i n a m arine environment, me asure ph ysiological a spects o f t he o rganism, a nd record data about t he external environment. Ā is discussion is organized a round re presentative c ompanies t hat p roduce t his equipment and equipment that is being developed by individual researchers for specic applications. Ā e following section considers re presentative s tudies i nvolving m arine  shes, turtles, cephalopods, a nd m arine m ammals. Ā e c hapter en ds w ith a brief discussion of the future of biotelemetry technology and its applications to the study of nonhuman animals.

FIGURE 2 7.1 Weddel s eal o uttted wi th an imal-borne v ideo an d data recorder. Note camera mou nted on he ad a nd v ideo data acquisition platform attached to back of animal. Satellite and radio signals are generated by respective transmitters and the animal also carries speed sensors (mounted on shoulder) and an accelerometer mounted on tail. In addition, the seal carries an electronic compass located just behind the video acquisition platform. Photo by Don Calkins, Alaska Sea Life Center, Seward, AK. With permission.

10/16/2008 7:15:38 PM

27-4

27.3

Smart Materials

Biotelemetry or Biologging: Equipment

27.3.1.1

Smart m aterials a nd s tructures a re u sed to lo cate a nd t rack aquatic animals in marine environments. As radio waves quickly attenuate in salt water, researchers use ultrasonic telemetry. In a typical system, aquatic animals are  tted with transponders or pingers that emit an ultrasonic signal. One or more hydrophones sense the acoustic signal and the animal’s location is determined either f rom th e s ignal s trength o r th rough t riangulation. I n addition to lo cating a nd t racking a n aqu atic a nimal, t hese devices c an a lso en code ph ysiological a nd b ehavioral d ata. Receivers a re t hen u sed to de code t he sig nal a nd t ransmit o r store the data. At present, it is becoming more common to use transmitters t hat p eriodically s end d ata to a s atellite a nd t hen forwarded to the user (see Figure 27.2).

27.3.1

Commercially Available Transmitters and Receivers

Ā ere are a number of companies that specialize in the construction of biotelemetry apparatus. Some of these companies provide such equipment for use in marine a nd freshwater systems whereas ot her c ompanies s pecialize i n j ust m arine s ystems. Given the focus on marine applications, we consider companies that supply biotelemetry equipment for ma rine environments. A r epresentative l ist o f s uch co mpanies i ncludes L otek, S onotronics, Sony, Star-Oddi, Vemco, and Wildlife Computers.

Satellite uplink Fixed station Lab

Wirless hydrophone (WHS)

Transmission to WHS

FIGURE 2 7.2 Typical s etup for biot elemetry. A quatic a nimals a re outtted with acoustic pingers that can be detected by hydrophone. In this de piction, ma natees a re o uttted w ith c ombined a coustic/radio transmitters that send both radio (detected by a receiver on shore) and acoustic signals (detected by wireless hydrophone) and sent by radio to receiver on short. Data are then sent to satellite and then to the laboratory. Figure modeled after Bridger, C. J., Booth, R. K., McKinley, R. S., Scruton, D. A ., a nd L indstrom, R . D., J. Ap pl. Ichthyol., 17, 126, 2 001. With permission.

43722_C027.indd 4

Lotek Wireless Fish and Wildlife Systems

According to their Web site (www.lotek.com), Lotek offers a full range o f b iotelemetry s ystems f or l arge a nd sm all m arine a nimals. Ā ese systems typically involve the use of transmitters or archival tags to encode and produce the signal, hydrophones to receive a nd t ransduce t he sig nal, a nd re ceivers to de tect a nd store information. Ā rough their MAP Series, Lotek offers coded acoustic transmitters for locating and tracking aquatic animals. Ā ese t ransmitters r ange i n si ze f rom 8 .5 mm i n d iameter a nd 32 mm long to 32 mm in diameter and 101 mm long. Ā ese transmitters, set at a burst rate of every 5 s, last anywhere from 16 days to g reater t han 5 ye ars de pending o n c hoice. Sig nals f rom t he transmitters a re de tected b y h ydrophones t hat c an b e w ired directly to a receiver or can be wireless where in they send transmissions through radio waves to a land-based or mobile receiver. Lotek’s M AP 6 00 acoustic teleme try s ystem c an support up to eight diff erent h ydrophones a nd i s a ble to t rack h undreds o f transmitters or monitor thousands of transmitters on the same frequency. Ā e s ystem i s c apable o f ac curate m easurement o f position in both a 2D and 3D format. Lotek a lso offers c oded ac oustic t ransmitters t hrough t heir CAFT series. Because of t heir unique codes, systems designed around t hese t ransmitters c an d istinguish up to 2 12 d ifferent signals o n a si ngle ac oustic f requency w hile s till m aintaining individual identication. Ā ese transmitters work with the SRX 600 a nd SR X 4 00 d ata log ging re ceivers a nd c an b e u sed to assess ha bitat u tilization, m igration, a nd pa ssage d ata a mong other app lications. F or sm aller a nimals, L otek off ers the NanoTag Series. Ā ese tags are claimed to be the smallest digitally en coded t ransmitter a vailable. Ā ese t ags ut ilize L otek’s proprietary c oding s ystem (CAFT s eries) b ut do not h ave t he same duration. However, some nanotags can continue to transmit on 10 s i ntervals for 678 days. Nanotags a re used to lo cate and track aquatic animals. Lotek s ensors c an a lso de tect, re cord, a nd t ransmit or s tore basic ph ysiological a nd b ehavioral d ata. S ensor t ypes i nclude depth, motion, and temperature. Ā ese sensors can be combined into a si ngle coded transmitter. In addition, Lotek can provide electromyogram data through their EMG Series. EMG tags are 16 mm by 53 mm in size and can be used to assess metabolic costs of t ail st rokes, oxygen consumption, energetic cost s, a nd transmit tem perature d ata, i f e quipped w ith t he tem perature sensor option. Lotek also offers archival tags (LTD Archival Tags) that record and s tore v ast qu antities o f d ata rel ated to t ime, tem perature, depth, and light. Ā rough variations in sun light levels and temperature, L TD t ags c an p rovide ge olocation d ata t hrough onboard geolocation algorithms that give daily estimates of latitude and longitude. One can re ne the geolocation data by crosschecking t hese d ata w ith s ea tem peratures a nd l ight le vels available f rom o ceanographic su rveys. Ā e re al-time c lock o n the t ransmitter is temperature compensated a nd is accurate to within 3 0 s p er ye ar. A mbient l ight i s me asured w ith b uilt i n light s ensors o r a “ sensor s talk.” L ight le vel d ata i s ba nd pa ss

10/16/2008 7:15:38 PM

Application of Smart Materials and Smart Structures to the Study of Aquatic Animals

 ltered to maximize the depth at which the sensor can operate. Ā e LTD is capable of operating at depths of 2000 m with a maximum depth of 3000 m. Depending upon conditions, battery life can exceed 5 years. Archival tags are becoming the tags of choice in m arine en vironments d ue to t he d ifficulty of maintaining contact w ith t he t ransmitter i n t he op en o cean. F inally, L otek offers i ts C ART S eries (combined ac oustic r adio t ransmitters). Ā ese t ags use a c onductivity sensor to de termine whether t he animal is in fresh water or salt water. If the animal is in saltwater, acoustic sig nals a re t ransmitted a nd w hen op erating i n f resh water, t he t ransmitter s witches to r adio sig nals. L ocating a nd tracking are the primary use of these tags. Scanning re ceivers (Suretrack S TR-1000-W1) a nd c ombination receiver/data loggers (SRX 600, DRX-600, and SRX 400) are also available. Depending on the model, researchers can manually track aquatic animals on foot, boat, or aircraft; or both scan and log i nformation relevant to lo cation and tracking. In addition, t he re searcher c an log ot her d ata i ncluding tem perature, physiological, mortality, geolocation, and historical data as well as conduct 3D mapping. Ā e SRX 400 can monitor multiple frequencies, track up to 212 individuals on a single frequency, identify transmitters as sensors, beepers or coded devices, download data to a c omputer, and can monitor both ultrasonic and radio frequency. Ā e SR X 600 improves the performance of SR X 400 by o ffering i ntegrated GPS c apability, e xpanded memory, USB capability, an enhanced detection algorithm, and a dual processor. Appl ications i nclude  sh pa ssage d ata, su rvival, p resence/ absence, site delity, and habitat utilization. 27.3.1.2

Sonotronics

Sonotronics ( www.sonotronics.com) u ses sm art m aterials a nd smart structures to acoustically locate and track aquatic animals in a ma rine en vironment. Ā eir p roducts i nclude a v ariety o f ultrasonic t ransmitters w ith v arious c ombinations o f l ifetime, size, range, and telemetry options. Manual tracking systems for real-time determination of animal behavior are available as well as automated stations for log ging a nimal detections, i ncluding cellular up link a nd w eb te chnology. To de termine  sh movement,  sh passage, and  sh location standard coded ultrasonic transmitters are available. Arranged from smallest to largest, Pico T ags ( PT s eries) c an b e de tected f rom 3 00 to 7 50 m, Miniature Tags (I BT-96 s eries) f rom 5 00 m o r mo re, Tracking Tags (CT series) have a r ange of 1000 m, a nd High Power Tags (CHP-87 series) can be detected from 3000 m. Sonotronics also offers c ombined ac oustic a nd r adio t ransmitters m ade i n c onjunction with Telonics Company. Ā ese transmitters are used in detecting aqu atic a nimals t hat mo ve f rom f resh w ater to s alt water environments or from areas of low salinity to high salinity. Ultrasonic teleme try t ags a re a lso a vailable. Ā e Temperature Tag (CTT-83 s eries) me asures tem perature to w ithin . 5°C a nd can b e de tected f rom 1 000 m. Ā e D epth T ag ( DT-97 s eries) measures depth to 5000 PSI and can be detected from 3000 m. A smaller t ransmitter (Mini Depth Tag, I BDT-97 s eries) i s available with a range of 500 m and a depth range of 100 PSI. Finally, an acoustic st orage t ransmitter (AST-05) collects temperature,

43722_C027.indd 5

27-5

pressure, a nd c onductivity d ata at p reprogrammed i ntervals. Ā ese d ata a re a rchived a nd t he c ode t ransmitted o nce e very 3 min to allow for retrieval and data download. Detection of the acoustic signal is via Sonotronics’ directional (mod # D H-4) a nd o mnidirectional ( mod # s D H-3 a nd SH -1) hydrophones. Ā e D H-4 d irectional h ydrophone p rovides t he greatest range in locating tags and has noise reduction capability, which improves precision. Ā e SH-1 omnidirectional hydrophone can detect tags up to 1 km. Sonotronics also offers a towed hydrophone pac kage (T H-1) to re duce t he t ime ne cessary to transect an area. An audio amplier (HPR-95) can be used with the piezoelectric hydrophones to allow for listening and recording of underwater sounds. Data logging receivers are available. Ā e USR-96 narrow band receiver can scan 10 preset frequencies and a t wo-line L CD d isplays b oth f requency a nd i nterval. Ā e USR-96 underlies the MANTRAK kit that has all the equipment necessary to m anually t rack aqu atic a nimals. S onotronics a lso offers a sub mersible ultrasonic receiver (SUR-1) that can detect and log d ata to  ash memo ry. Ā e SU R-1 c onsists o f a n i ntegrated hydrophone, ash memory, and a transponder that allows the user to interrogate the unit from a distance and determine if information is available. Ā e CUB-1 cellular uplink buoy can be used automate the system and send text messages to a user’s cell phone or personal computer. 27.3.1.3

Star-Oddi Marine Device Manufacturing

Star-Oddi Ltd ( www.star-oddi.com) a c ompany ba sed i n Reykjavik, Iceland, produces data storage tags or archival tags, GPS  sh p ositioning s ystems, tem perature r ecorders, a nd t agging equipment for marine animals. Ā e DST-CTD and DST-CT salinity loggers detect and store for later retrieval measurements on salinity, temperature, and depth. Ā e tags measure 15 mm × 46 mm and weigh only 12 g in water. Memory capacity is 43,582 measurements per sensor a nd for t he salinity a nd temperature tag, 65,314 measurements per sensor can be obtained. Tags have a battery life of 4 years and offer 12 bit resolution. Ā e logger can be downloaded and reprogrammed as often as needed until the battery is depleted. A communication box and soft ware is available for the tags allowing wireless transfer of data to a personal computer. St ar-Oddi a lso off ers t he DS T P itch & Rol l log ger. Ā ese tags measure pitch and roll, temperature, and depth. Ā ey offer a bat tery l ife o f 5 ye ars a nd c an de tect a nd s tore de pth information to 2000 m. Size and weight of these tags is the same as the DST-CTD tag. Ā e Pitch & Rol l logger has 12 bit resolution a nd me mory c apacity of 1 74,600 me asurements i n tot al. Ā ese tags nd applications for analyzing movement of marine animals and other underwater gear. Ā e DST compass logger is a sm all sub mersible c ompass log ger w ith t he opt ion o f a lso recording tem perature a nd de pth. Ā e ele ctronic c ompass h as ±15° ac curacy a nd s hows eig ht p oint c ompass d irection o r i n degrees. Ā e compass logger is more accurate in the horizontal plane. Size and weight are the same and the memory capacity is 128,000 measurements in total with 12 bit resolution. Finally, t he Star-Oddi GPS Fish Positioning system consists of a d ata s torage  sh t ag ( DST-GPS) a nd a G PS t ransmitting

10/16/2008 7:15:40 PM

27-6

Smart Materials

sonar/transmitter. Ā e t ag i s at tached e xternally o r i mplanted and receives information about geographical location from satellites v ia Si mrad s onars o nboard t he v essel. Ā e v essel’s G PS position i s c oded a nd t ransmitted u nderwater v ia t he Si mrad sonar sig nal. Tagged  sh, s wimming w ithin 4 km o f t he G PS sonar, receive and store the geographical position as well as data on tem perature, de pth, a nd d ate/time ( see F igure 2 7.1). Ā e system is designed for medium-sized round sh (e.g., cod, saithe, salmon) a nd works well i n m arine environments. For applications in rivers, lakes, and smaller ocean areas, Star-Oddi offers a portable  sh p ositioning s ounder ( FPS). B oth s ystems c an b e used to determine migration patterns, habitat use, and foraging strategies among other possibilities. In collaboration with the Icelandic Marine Research Institute, Star-Oddi h as de veloped t he u nderwater t agging e quipment (UTE) t hat c an b e u sed to rob otically t ag de epwater  sh that cannot survive large changes in temperature and pressure. Ā e UTE is placed in the cod end of a trawl where the sh is enclosed by a g rid d iverting it to t he t agging place. Ā e sh is observed through a v ideo c amera a nd t he re searcher moves t he t agging gun into position. A robotic arm is used to make a small incision and press the DST tag into the body cavity. In just a few seconds, the  sh is released through a channel in the device. Ā i s system allows for tagging on site in the sh’s natural habitat (see Figure 27.2). In October 2003, June 2004, and June 2005, successful tagging cruises were made during which nearly 2000 redsh were tagged at depths from 500 to 800 m, with tags that measured depth a nd tem perature. A s c ommercial a nd s port  sheries recapture t agged  sh, activity data can be downloaded to provide valuable information for the  shing industry in the waters off Iceland a nd el sewhere. Ā e U TE s ystem i s s howing sig nicant p romise a s a me ans to s tudy de epwater  shes. Technical specications of the UTE and a detailed discussion of how deepwater sh are tagged can be found in Sigurdsson et al. (2006). 27.3.1.4

Vemco

Vemco (www.vemco.com) specializes i n t he de sign a nd manufacture of acoustic telemetry systems. Ā e company offers a wide variety o f t ransmitters, h ydrophones, re ceivers, a nd d ata log gers. Coded and continuous expendable pingers and data telemetry transmitters are offered to scientists who want to track and monitor aquatic animals in a marine environment. Six standard diameters ranging in size from 7 mm (V7) to 32 mm can be constructed w ith u ser-determined bat tery l ife, p ower o utput, a nd life expectancy options. Ā ese u ltrasonic a coustic t ransmitters can a lso b e e quipped w ith s ensors to me asure tem perature o r depth. As a n e xample, t he V16-coded t ransmitter me asures 16 mm in diameter, with length and weight varying from 48 to 90 mm and 9 to 1 4 g, re spectively. V 16 t ransmitters c an b e e quipped with silver oxide or lithium batteries, and, depending on minimum a nd m aximum del ays f or t ransmission, re quired p ower output, a nd s ensor t ype, p rojected bat tery l ife v aries f rom 2 6 days to 7.73 years. Users have choice of sensor type (depth or temperature), bat tery t ype, p ower le vel, f requency c ase s tyle,

43722_C027.indd 6

temperature range, maximum depth, coded type, code ID information, and random delay range for transmission. Ā e V X32TP-CHAT t ag p rovides t wo-way c ommunication between the tag and a receiver. Users can locate and track tagged marine animals and they can receive archived data from a transponder all without recapturing the animal. Data can be recorded or transmitted in real time and partial downloads are possible in the event that the animal moves out of range of the receiver. Ā e CHAT tags have a long life span and high memory capacity; up to 5.6 years with a sample interval of 1 min and an average interval of 1 day, and multiple memory downloads are possible. Tracking receivers use reply time to estimate distance and sample rate and average interval of response is reprogrammable. Ā ese tags a re equipped with depth and temperature sensors and a swim speed sensor is available on request. Ā e tag is best suited for large and slow moving marine animals and measures 32 mm in diameter and 150 mm i n leng th. It weighs 190 g i n a ir a nd 75 g i n w ater. Vemco will also manufacture experimental transmitters for specic applications. Examples include heart rate transmitters, differential pressure transmitters that can be used to determine swimming speed i n  sh or jetting i n cephalopods, accelerometers, integrating accelerometers, and transmitters that detect tilt. Vemco m anufactures si x d ifferent re ceivers, r anging f rom single c hannel re ceivers ( VR2) t hat t rack c oded t ransmitters automatically to re ceivers t hat me asure re al-time a nd p rovide detailed position information including true latitude and longitude by using three detection buoys and a two-way radio link to a base station (VRAP). Further examples include Vemco’s VR3 monitoring receiver that provides remote communication to satellite, u nderwater m odem, c ell p hone, o r u nderwater obs ervatory and can be left unattended for as long as 12 months. Satellite transmission is achieved through the ARGOS Satellite and GPS positioning system. Ā e VR28 tracking system is connected to a four-element hydrophone a rray t hat prov ides 3 60° of c overage and all functions including frequency, gain, and echo rejection blanking interval are user determined. Ā e VR28 is optimal for use with CHAT tags. Vemco also offers a variety of hydrophones to be used with their receivers to de tect acoustic tags. With the exception of the VHLF hydrophone that is used to monitor sounds in the audible frequency range and requires a separate system, Vemco hydrophones connect to the receivers via cable. Ā e hydrophones are designed to be t rolled t hrough t he w ater a nd a re t uned to s pecic acoustic transmitters. Ā e V H4x series hydrophones a re designed for u se with the VR28 tracking system. Finally, Vemco offers two different data loggers (Minilog-8 bit and the Minilog-12 bit). Ā ese loggers record temperature or temperature and depth information. Ā ey are waterproof, r ugged, a nd a re de signed to do wnload u sing t he Minilog-PC interface that requires no external electrical connections. Ā ese loggers have a battery life expectancy of 5 years. 27.3.1.5

Wildlife Computers

Wildlife Computers (www.wildlifecomputers.com) offers satellite tags a nd a rchival t ags. Ā e A rgos s atellite s ystem de tects t he transmissions from satellite tags. Ā ese tags can detect wet and

10/16/2008 7:15:40 PM

Application of Smart Materials and Smart Structures to the Study of Aquatic Animals

dry c onditions, t hus p roviding d ata o n h aul-out b ehavior b y seals o r s ea l ions, a nd t hey a re c apable o f t ransmitting de pth, temperature, and light level data. Data are received in near real time. Ā ree t ypes o f s atellite t ags a re c urrently a vailable. Ā e smart position and temperature transmitting (SPOT) tag is the smallest and lightest and can be used in a marine environment. Ā e SPOT5 t ag i s t heir l atest version. Ā is t ag i s available i n a variety of shapes a nd can be deployed on cetaceans, penguins, seals, and turtles among other species. Ā e SPOT5 tag transmits temperature data in the form of histograms to allow researchers to easily assess where the animal spends its time within specied temperature ra nges a nd h aul-out sta tistics a re r eported a s t he percentage of time the tag is dry for each hour of the day. Using the Argos satellite system and Cricket, the Argos-certied transmitter de veloped b y Wi ldlife C omputers, t he S POT5 t ag p rovides locations with an accuracy of ±350 m. Ā e Argos Satellite SPLASH tag is a new satellite tag that combines the sampling and data storage functions of an archival tag with a n A rgos t ransmitter. S PLASH s ensors me asure de pth, temperature, light level, and wet and dry periods. Temperature and depth data are collected, analyzed, summarized, and compressed for transmission. Approximately 14 Mbytes of memory are available w ith re covery of t he t ag. Available me ssage t ypes for t ransmission i nclude h istograms f or d ive d uration, m aximum d ive de pth, t ime-at-depth, t ime-at-temperature, p ercentage t imelines, a nd 2 0 min t imelines. L ocation ac curacy i s t he same as the SPOT5 tag. Ā e third type of satellite tag is the Pop-up Archival Transmitting Tag (Mk 10-PAT). Ā is tag also combines the sampling and data storage p roperties o f a n a rchival t ag w ith s atellite te chnology. The t ag i s de signed to t rack l arge-scale mo vements a nd behaviors of pelagic  sh and other animals that do not come to the surface often enough to use real-time satellite tags. Wildlife Computers contend that this tag is by far their most intelligent and c omplex t ag. Ā e P AT t ag c an p rovide d ata w ithout t he animal being recaptured, but the tag is recovered a full archival record is maintained in nonvolatile. Designed for large pelagic aquatic animals, PAT tags have been deployed on large marlin, sharks, sea turtles, swordsh, and tuna. Ā e tag is attached with a tether that is designed to detach at a user-determined date and time. Once detached, the tag  oats to the surface and transmits summarized d ata v ia Cr icket to t he A rgos s ystem. Ā e tag archives depth, temperature, and light level data. By analyzing either premature releases or when the PAT has been at a c onstant depth (surface or bottom) for too long, t he tag c an p rovide re searchers w ith i nformation o n at tachment failures an d an imal m ortality. A m echanical guill otine a utomatically c uts t he te ther i f t he t ag i s d ragged to de pths t hat would crush it, t hus ensuring transmission of important data. Ā e l ithium bat tery a llows t he PAT to re cord d ata for up to 1 year and over the course of 7 days, it can transmit 10,000 32-byte transmissions. Users can scale the transmitted data to meet the requirements of their deployment length. Sensors include 12 bit analog to d igital converters for temperature a nd de pth, 10 bit analog to d igital c onverter f or l ight le vel a nd bat tery vol tage,

43722_C027.indd 7

27-7

and a wet and dry sensor that can detect the ideal time to transmit and provide haul out data. As with the other satellite tags, geographical positions are accurate to within 350 m. Researchers have u sed t his t ag to s tudy m igration pat hs, de pth, a nd tem perature p references, a nd o ceanographic d ata i n t he f orm o f depth–temperature pro les. In addition to smart satellite tags, Wildlife Computers offers a range of archival tags. Using these tags, researchers can collect and record data on the swimming behavior of marine animals. Depending o n t he t ag, de pth, en vironmental tem perature, heart-rate, stomach temperature, and light level can be recorded. Light level data can be used to determine approximate location and time is encoded with the stored data. Ā e Mk9 is the company’s s mallest a rchival t ag a nd i s ty pically c ongured with depth, temperature, light level, and wet and dry sensors. Ā e tag is 67 mm long by 17 mm by 17 mm and weighs less than 30 g in air. Ā e tag can withstand pressure to 1000 m and the 16 Mbytes of  ash memo ry c an s tore 8 to 1 6 m illion s ensor re adings. A supplied interface provides download capability to the user’s PC. As mentioned a h eart rate-temperature recorder (HTR), a heart r ate t ransmitter (H RX), a nd a s tomach temperature pill (STP) can provide data on heart rate, internal temperature, and stomach tem perature. Ā e S TP t ags c an b e u sed to i ndicate when the animal ingested cold prey, thus providing important data o n f oraging h abits. Ā e c ompany i ndicates t hat t hey a re currently de veloping t he M k10-AF t ag t hat w ill u se F astloc technology to acquire position data in a very short time that are accurate to 10 m.

27.4 Biotelemetry and Biologging: Applications and Findings In this section, we examine studies that have used smart materials a nd sm art s tructures b iotelemetry to s tudy m arine a nimals. W e w ill e xamine re presentative s tudies a nd d iscuss briey their  ndings.

27.4.1

Locating and Tracking Marine Animals: Fishes, Turtles, and Cephalopods

27.4.1.1 Fishes Biotelemetry ha s b een u sed e xtensively to l ocate a nd t rack marine  shes (N elson e t a l., 1 997; L owery a nd Sut hers, 1 998; Candy a nd Q uinn, 1 999; P epperell a nd D avis, 1 999). M ore recently, Bridger et al. (2001) monitored the behavior of domestic steelhead t rout (Oncorhynchus mykiss) i n a n u nconstrained environment. Ā e a uthors w ere in terested in d etermining whether do mestic  sh t hat h ad e scaped f rom a n aqu acultural facility would show site delity or would d isperse. I f domestic fish showed site fidelity then recapture could reduce the economic lo ss of l arge numbers of  sh e scaping a nd t he environmental i mpact t hat dome sticated  sh might have on wild stock. Given that the salinity changed dramatically according to

10/16/2008 7:15:40 PM

27-8

depth o f wa ter, t he a uthors de termined t hat co mbination acoustic an d r adiosmart tr ansmitters ( Lotek C ART tr ansmitters) offered the most efficient way to monitor  sh location over large areas. In total, 240 trouts were implanted with a CART transmitter. One hundred and fty sh were released during the summer and 90  sh were released during the winter months. Ā e authors used remote data-logging stations that were  xed and self-sustaining. Ā e system included a Lotek SRX 400 data logger, ASP_8 (antenna switching peripheral), an ultrasonic upconverter, photocontroller, bat tery supp ly, a nd s atellite d ata t ransceiver. I n addition, a w ireless hydrophone system (Lotek WHS-1000) was used to de tect a sig nal f rom t he t ransmitter. Ā e W HS s ystem was found to be a better solution than tethered hydrophones due to deployment and maintenance difficulties, and reliability over time. A h ybrid s ystem o f w ind t urbines a nd ph otovoltaic c ells kept the system’s 400 amp hour battery charged and operational for 10 days in the absence of wind or sun light. Given the remoteness of the area (Bay d’Espoir, Newfoundland, Canada), a remote data l ink to t he data-logging stations was required. A t wo way satellite-based data link was developed to allow Bridger et al. to download date and time of transmitter detection, channel, code, antenna, p ower le vel, n umber o f e vents a nd h ourly bat tery status. Ā e re sults showed t hat s teelhead t rout rele ased during the summer months showed strong site delity w hereas  sh released during the winter months showed decreased site delity. Bridger et al. concluded that attempts to recapture domestic sh that had escaped their holding pens during the summer could be successful and reduce economic loss and ecological interactions between domestic and wild stock. To assess site delity o n a rticial re efs i n t he no rtheastern Gulf of Mexico, Szedlmayer and Schroepfer (2005) tagged and tracked 5 4 re d snapp er ( Lutjanus c ampechanus) du ring a ll seasons of the year. Fish were implanted with ultrasonic transmitters and monitored with Vemco remote receivers and Sonotronics surface receivers. Manual tracking from the surface was used to follow four red snappers over night. Positions were recorded e very h our f or 9 o r 16 h. Sz edlmayer a nd S chroepfer found t hat t he m ajority o f re d snapp ers s howed s trong si te delity, with 67% of the sh staying on articial reefs from 117 to 595 days. High residency rates were independent of season and the authors concluded that articial reefs provided suitable habitat for red snapper residing in the northeastern Gulf of Mexico. Estimates o f  sh d istribution a nd a bundance a re often in error to due to acoustic dead zone loss and avoidance reactions to v essel a nd su rvey e quipment ( Aglen, 1 994; H jellvik e t a l., 2002). Ā e ac oustic de ad z one re fers to t he a rea w here aqu atic animals c an s wim b elow t he de tection z one o f t he re cording equipment and are not c ounted. Stensholt and Stensholt (2004) compared t he e stimates o f a bundance o f A rctic c od ( Gadus morhua L.) based on data from acoustic surveys, trawls, and data storage tags (DST) in an attempt to de velop a mo del that could improve such estimates. Ā e authors worked on the assumption that samples from trawls along the bottom were biased because sh higher in the water column were not c ounted and acoustic

43722_C027.indd 8

Smart Materials

surveys were biased because it was possible to miss sh that were swimming on the bottom and out of the detection range of the acoustic equipment. Ā e argument that both methods together would provide a b etter e stimate w as m itigated by t he f act t hat water depth and temperature played a role in how much overlap there was between the two methods. In addition, prey dispersal also entered into the equation. To de velop a b etter mo del f or e stimating a bundance, t he authors b egan b y no rmalizing t he de pth d ata. Ā en, ba sed o n estimates f rom ac oustic su rveys a nd t rawls a nd d ata f rom c od tagged with data storage transmitters that provided depth and temperature information, the authors developed their model. In essence, t heir work resulted in modication of t he  sh vertical distribution due to spatial and temporal variation. Ā is improvement i n vertical d istribution a llowed t he authors to a ssess  sh abundance more accurately. Holland et a l. (2001) t agged t iger sharks (Galeocerdo cuvier) with  ve t ypes o f t ags to e valuate t he e xtent to w hich a ll t ags provided comparable data and to assess the horizontal and vertical movement of t iger sharks off the coastline of Oahu, Hawaii. Fine s cale mo vement o f t he s harks w as re corded u sing s onic tracking methods for periods up to 50 h, whereas acoustic pingers and automated, anchored data loggers provided long-term periodicity a nd re turn f requency d ata. Ā e ve t ag t ypes i ncluded standard identication t ags (Casey a nd Kohler, 1992), ac oustic tags for active tracking, long-life acoustic tags detected by submerged data loggers, a rchiving tags, a nd a p op-up a rchival tag that t ransmitted s tored d ata to a s atellite. T iger s harks w ere caught using long-line  shing techniques, tagged, a nd released less than 5 km from Waikiki beach, Oahu, Hawaii. Over a 5-year period (1994–1999), 133 t iger sharks were captured a nd tagged with i dentication t ags. S eventeen o f t hese s harks w ere l ater recaptured at the same site after intervals ranging from 14 to 336 days p ostcapture. A tot al of eig ht s harks were t racked ac tively with either internal or external tags. Ā e tracks of the eight animals were remarkably consistent with extended periods of straight-line swimming and movement over large distances with seven of the tracks ending at Penguin Banks located 32 km from the starting site. Depth recorders revealed that the sharks stayed near the bottom when in shallow water, but moved up to between 40 and 100 m depth when in open water. Twenty s harks w ere i mplanted w ith lo ng-term p ingers t hat were detected by data loggers in the original shing area. Ten of these tags were later detected. Sharks returned to t he area at a ll times of the day with some sharks showing more daytime detections than nighttime detections. Ā ere was no discernable periodicity to the sharks’ movements. Four sharks received archival acoustic modem tags. Vemco CHAT tags measuring 28.8 × 6.6 cm with a no minal l ife s pan o f 2 4 mo nths s tored de pth a nd tem perature data. A 5 min sample period with 6 h averages was used. Data were sonically downloaded via acoustic modem technology to a submerged data logger anchored to the bottom. When a tag was detected within 500 m of t he data logger, downloading commenced. At the time of writing, one shark had been detected and i ts d ata do wnloaded. H orizontal a nd v ertical mo vements

10/16/2008 7:15:40 PM

Application of Smart Materials and Smart Structures to the Study of Aquatic Animals

were consistent with the active tracking data. One shark received an archival satellite tag developed by Wi ldlife Computers, Ltd. Ā is tag recorded depth and temperature data condensing them into h istogram f ormats. Ā e t ag w as e xternally at tached a nd equipped with a corrosible link that caused the tag to release and oat to the surface where it transmits data to a s atellite. Ā e tag surfaced a nd t ransmitted d ata, b ut i t app eared t hat rele ase occurred early yielding only 13 days of data and 15 days of surface drifting. At the time of release, the tag was located 140 miles southwest of Oahu. Ā e authors concluded that the ve different tags revealed consistent data showing that tiger sharks tend to show site delity to a home range, which is apparently very large, with several weeks between returns to a specic site within that range. Ā e authors also commented that more data could be collected from the acoustic and CHAT tags than from the identication and recapture method, resulting in greater accuracy. Pacic bluen t una (Ā unnus thynnus orientalis) constitute an important commercial shery and a clear understanding of their migratory and foraging habitats is critical to ensuring the sustainability of this important resource. Domeier et al. (2005) sought to de velop an automated algorithm that could provide more accurate e stimates of latitude by u sing sea su rface temperatures from satellites in combination with temperature and light le vel i nformation p rovided b y a rchival ac oustic t ags. Researchers h ave u sed l ight le vels f rom suc h t ags to p rovide accurate estimates of longitude, but estimates of latitude were less reliable, particularly as the subject approached the equator (Hill and Braun, 2001). To develop their algorithm, the authors tagged blue n tuna with pop-up satellite archival tags (PSAT) supplied by Wildlife Computers, satellite-based archival tags (PTT-100 P SAT) f rom M icrowave Telemetry, I nc. (Columbia, MD), or L otek w ireless non transmitting a rchival t ags (LTD2310). A ll t ags r ecorded de pth, wa ter tem perature, a nd light level. Ā e Lotek archival tags recorded these variables and also provided data on internal sh tem perature. Da ta f rom these t ransmitters w ere re corded ei ther e very 2 min ( Lotek), every 2 min, and summarized hourly (Wildlife Computers), or hourly (Microwave Telemetry). Using d ata f rom t hese t ransmitters a nd f rom s atellite acquired s ea su rface tem peratures, t he a uthors de veloped a n automated s ystem f or t racking c alled t he PS AT T racker Information System (PTIS). Ā e key pa rameters of t he s ystem include time and position of tag deployment, time and position of t ag re covery, l ight-based e stimates o f lo ngitude, m aximum swimming s peed, a nd a b racketed r ange o f l atitude e stimates within which the program looks for temporal matches from the sea surface temperature data. Ā is algorithm is fully described in the paper and produced tracks that were less variable than those produced by light levels alone. Ā e authors concluded that their algorithm showed increased accuracy in tracking particularly i n t hose a reas where t here is a l arge north to s outh temperature gradient. PTIS results correlated well with estimates calculated from light-based algorithms and appeared to provide better data in those situations where light-based estimates failed (near t he e quator). Wi th re spect to t he m igratory h abits o f

43722_C027.indd 9

27-9

blue n t una, t he a uthors f ound t hat t una r anged f rom t he California-Oregon border to southern Baja California, Mexico, but s pent t he m ajority o f t heir t ime o ff t he c oast o f c entral Baja, Mexico. Finally, t he authors found evidence for seasonal migration with tuna preferring the central Baja location during the w inter a nd s pring a nd mo ving no rth to ward O regon a nd then back south during the summer and fall. 27.4.1.2

Cephalopods

Yatsu et a l. (1999) tracked jumbo  ying squid (Dosidicus gigas) in the Eastern Pacic Ocean using ultrasonic telemetry. Jumbo ying s quid p rovide a n i mportant  shery i n J apan a nd K orea that increased each year up to 1994. Perhaps due to dispersion of the species as a result of environmental changes, the catch went from a high of 165,000 metric tons in 1994 to an astonishing 80,000 metric tons in 1995. As a re sult, it became important to understand the behavior and movement patterns of this species. To t hat en d, t he a uthors de ployed t hree s quid r anging i n si ze from 35 to 43 cm total mantle length. Ultrasonic transmitters supplied b y V emco ( models V 16P-1H a nd V 16P-4H) w ere attached just forward of the n with a rubber band. Transmissions were detected w ith a n onboard hydrophone a nd GPS locations of the ship were determined by the NNSS-GPS navigation system. B oth t ag t ransmissions a nd G PS lo cations were t ransmitted at 1 s intervals. Squid were captured and released several hours a fter su nset a nd t racked b etween 8 ( one s quid) a nd 14 h (two squid). All squid were lost when they dived below 1020 m, a depth that exceeded the limit of the transmitters. Tracking data allowed the researchers to a ssess horizontal and vertical movement. Typically, s quid remained above 2 00 m during t he n ight and dived below 1000 m either at t wilight or 5 h before sunrise. Diving speed varied between 2 and 28 m per minute. During the time of tracking, squid moved between 3 a nd 5 m iles horizontally. Ā e results provided important data for further work and showed that tracking the jumbo ying squid is possible through ultrasonic tagging techniques. More recently, Jackson et al. (2005) conducted initial tests on the adv antages o f u sing hybrid ac oustic/archival t ags to lo cate and track squid and cuttlesh. Australian giant cuttlesh (Sepia apama) and tropical squid (Sepioteuthis lessoniana) were tagged with a hybrid tag developed by Vemco. Ā e V8 acoustic tag was potted w ith a V emco m inilog tem perature/depth t ag a nd attached to four animals. Ā e a nimals w ere mo nitored w ithin radioacoustic positioning telemetry (RAPT) buoy system arrays. Ā ese arrays included bottom-mounted sensors that transmitted independent tem perature re cords a nd re ference s tandards f or sound c onductivity a nd p osition. D uring t he c ourse o f t he deployment, all four tags moved out of RAPT range, but two of the tags were recovered using a b oat-mounted hydrophone and the VR60 Vemco receiver. A VUR96 diver operated receiver also aided i n t he re covery. A lthough h alf o f t he t ags w ere lo st, t he authors concluded t hat t heir system provided a n a lternative to more expensive satellite pop-up tags and had the added advantage that it could be deployed on small species that live in or return to near-shore environments.

10/16/2008 7:15:40 PM

27-10

Smart Materials

In a follow up paper, Aitken et al. (2005) describe the results acquired from Sepia apama that were tagged with acoustic transmitters and monitored with the RAPT system. Ā e authors used position only and jet pressure transmitters. Results showed that the h ome r ange f or S. a pama w as b etween 5 300 a nd 23 ,700 square meters. S. Apama was shown to be more active during the day, s pending 3 .7% o f t heir t ime f oraging d uring t he d ay a nd 2.1% of their time foraging at night. Over all, S. apama was active during t he d ay f or 3 2 d ays, b ut w as ac tive f or o nly 18 n ights. Interestingly, t he g iant c uttlesh spent 95% of its t ime re sting, which places its behavior more in line with the octopus than the squid. Ā e authors concluded that the cuttlesh’s predisposition to qu iescence a llowed it to c hannel its energ y i nto fast g rowth and later into mating. 27.4.1.3

Turtles

As i ndicated o n t he Re dlist ( www.redlist.org) s ea t urtles a re endangered in all parts of the world. Greens (Chelonia mydas), olive r idelys ( Lepidochelys o livacea), a nd l oggerhead ( Caretta caretta) a re l isted a s en dangered. M ore p ressing, h awksbill (Eretmochelys imbricata), Atlantic ridley (Lepidochelys kempii), and leatherback (Dermochelys coriacea) turtles are listed as critically en dangered. Bioteleme try s tudies c an h elp s cientists understand their behavior including foraging, migratory habits, and predator defense, and the effects of sheries. A clear understanding o f t hese v ariables a nd t he rel ated ph ysiology c an inform the development of conservation policies and hopefully stave off extinction. Recent e xamples o f b iotelemetry s tudies w ith s ea t urtles include investigations by James et al. (2005) and Polovina et al. (2004). James e t a l. outtted 38 le atherback t urtles, a c ritically endangered species that may be facing extinction in the Pacic (Spotila et al., 2000), with one of ve different s atellite t ags to assess their migratory habits. Ā e authors collected data for 8288 tracking days over an average interval of 218 days. Eleven turtles were tracked for more t han 1 ye ar. Tracking data revealed t hat turtles spent up to 4 months off the coastlines of eastern Canada and the northeastern United States. Southward migrations typically began in October, with some turtles moving south as early as August 12, and others as late as December 15. Southern destinations were determined for 25 t urtles. Eleven turtles returned to waters adjacent to nesting beaches along the north-east coast of South America, the Antilles, Panama, and Costa Rica. Other turtles m igrated to p elagic w aters b etween 5° a nd 23 ° N o r to shelf wa ters off t he c oast o f s outh-eastern U nited St ates. I n February and March, turtles returned to the northwest Atlantic, north of 38 ° N. T ypically t hese t urtles re turned w ithin s everal hundred k ilometers o f w here t hey o ccurred t he p revious ye ar. Of pa rticular i nterest to t he a uthors w as t he  nding th at th e turtles did not use the same paths when heading south or north, rather they spread out over much of the ocean. Also of interest was t he  nding t hat le atherback t urtles i n e astern C anada weighed 33% more than turtles found in the southern site of St. Croix, U.S. Virgin Islands and that turtles showed foraging site delity to shelf and slope waters.

43722_C027.indd 10

Polovina et al. (2004) examined the forage and migratory habitat of 26 loggerhead and 10 olive ridley sea turtles in the central northern Pa cic. Argos-linked Telonics ST-10 or ST-18 position transmitters were attached to 32 of these turtles. Four turtles were outtted with Wildlife Computer Argos-linked satellite transmitters with dive recorder capabilities (SDR-T10). Transmissions per turtle ranged from 30 to 485 days and resulted in 2470 transmission days. In addition to position, depth, and temperature recordings, s ea su rface tem perature, su rface c hlorophyll a lpha concentration, a nd ge ostrophic c urrent w ere ob tained f rom t he University of M iami a nd v ia s atellite a ltimitry. Sub surface temperature a nd c hlorophyll i n t he re gion w ere ob tained t hrough shipboard oceanographic t ransects. G enetic a nalyses of t he t urtles showed that all loggerhead turtles came from Japanese nesting beaches, si x ol ive r idleys c ame f rom e astern Pac ic nesting beaches, and three came from the western Pacic. Ā e results showed little overlap in the migratory paths of the two species of turtles. All loggerheads but two traveled west from their point of capture whereas olive ridleys traveled both eastward and westward. O ver t he co urse o f t ransmissions, l oggerheads m oved north o r s outh ( through t he re gion 2 8°–40° N) , de pending o n season. F rom J anuary to J une, log gerheads f avored t he s outhern section of their range and from July to D ecember, they were typically found i n t he northern s ection. With re spect to tem perature preference, loggerheads kept in waters where the surface temperatures ranged from 15°C to 25°C. Loggerheads tended to stay near the surface, spending 40% of their time at t he surface and 90% of their time overall, within 40 m o f the surface. In addition, for the loggerheads, it appeared that the Transition Zone Chlorophyll Front and the Kuroshio Extension Current provided important foraging and m igratory habitats. Interestingly, t he t ransmissions f rom one loggerhead whose recordings lasted 485 days, showed that it traveled over 9000 km in 458 days. Ā is t urtle’s westward movement was assisted by the westward  ow along the north side of the counterclockwise rot ating e ddies gener ated b y t he e astward  owing Kuroshio Extension Current. Also of interest was the nding that these areas marked regions of enhanced surface chlorophyll. Olive r idleys, o n t he ot her h and, w ere f ound s outh o f t he loggerhead habitat occupying the region bounded by 8°–31° N in the central Pacic. Ā ere was evidence that olive ridleys also displayed a north/south preference depending on season. Olive ridleys preferred warmer waters (23–28°C) and spent more time in deeper water compared to loggerheads. On average, ridleys spent 20% of t heir t ime at t he su rface a nd 6 0% of t heir t ime w ithin 40 m o f t he su rface. It a lso app eared t hat t he w estern s tock o f olive ridleys occupied different oceanic habitats from the eastern stock a f act t hat has implications for species formation a nd for conservation eff orts. Ā e lo ngest t ransmission f rom a n ol ive ridley showed that it traveled 7282 km in 193 days. Ā is turtle’s movements were also aided by equatorial currents, particularly the E quatorial C ounter Cu rrent a nd t he N orth E quatorial Current. Ā e authors concluded that their method of using satellite remote-sensed a nd s hip-based ph ysical a nd b iological d ata i n combination with the tracking data led to a more accurate picture of the habitat needs of loggerhead and olive ridley turtles.

10/16/2008 7:15:40 PM

Application of Smart Materials and Smart Structures to the Study of Aquatic Animals

27.4.2 Locating and Tracking Marine Animals: Marine Mammals Biotelemetry has been used to locate and track marine mammals for over 40 years, beginning with simple VHF radio transmitters and time-depth recorders (TDR) in t he mid si xties (Kooyman, 1965). M ost e arly b iotelemetry s tudies f ocused o n p innipeds because they came ashore at predictable locations and intervals. Ā is characteristic a llowed researchers to g lue t he tag onto t he dorsal surface of the animal and retrieve the tag later when the animal h auled o ut f or b reeding o r mol ting ( Butler a nd J ones, 1997; K ooyman, 2 004). Ā e i ncorporation o f A rgos s atellite receivers a llowed s cientists to ob tain i nformation o n a nimal location and other data from the comfort of their office, removing the need to follow animals utilizing a VHF signal. Satellitetracking studies proliferated in the 1990s as the use of platform transmitting ter minals ( PTT) b ecame mo re re adily a vailable and were used on an expanding range of taxa (Kooyman, 2004; see Hooker and Baird (2001) for a ranging studies in whales and dolphins). However, the amount of data transferable via satellite is limited, particularly for cetaceans (whales, dolphins, and porpoises) t hat o nly b reak t he w ater su rface s poradically a nd f or seconds at a time. Ā us cetacean researchers often chose archival tags that send only limited information via satellite, such as GPS location, and store the rest (Hooker and Baird, 2001). Studies of marine m ammals i ncorporating b iotechnology c over a w ide array of research topics including migration routes and destinations ( Mate, L agerquist a nd C alambokidis, 1 999); lo cal mo vements related to f oraging, socializing, a nd breeding (Chilvers et al., 2 005; M ate e t al ., 2 005; S mall e t al ., 2 005); p ostrelease monitoring of rehabilitated animals (Gulland, 2003); and diving behavior (Davis et al., 2003; Baird et al., 2005). North Atlantic right whales are one of t he most endangered whale p opulations w ith le ss t han 3 00 i ndividuals rem aining (Moore et al., 2004). Baumgartner and Mate (2005) used satellitemonitored radio tags to track right whale movements in relation to environmental variables including depth, depth gradient, surface and bottom hydrographic properties, and remotely s ensed surface tem peratures a nd c hlorophyll c oncentrations. Ā e authors collected data over a number of years from 1989 to 2000. Time-depth recorders were used to record diving behavior while separate s atellite-linked U HF r adio t ags (Telonics S T-3, S T-6, ST-15) were used to track movements. Satellite transmission was only possible when the tag was at the surface and this was determined by conductivity sensors to determine if the tag was “wet” or “ dry.” Ā is methodology saved bat tery life when t he a nimal was below the surface. Tags could transmit from once every 10 to 40 s. Ā e authors e xpected t hat r ight w hales would move f rom the B ay o f Fu ndy, a f avorite f eeding h abitat, off shore t o d eep basins where copepod density is high. Instead whales chose areas with low temperatures, h igh su rface s alinity, a nd h igh su rface stratication. Ā e authors speculated that the low temperatures and high vertical stratication in shallow basins caused higher densities of their primary copepod prey, Calanus nmarchicus. Ā e stratication increased the shoal-layering effect that promotes

43722_C027.indd 11

27-11

Calanus  nmarchicus su rvival a nd a bundance. R ight w hales were able to feed longer at these prey depths (approximately 150 m) compared to whales expending energy to dive in deep oceanic basins (Baumgartner et al., 2003).

27.4.3

Assessing Energy Expenditure with Acoustic Telemetry: Fishes, Cephalopods, and Turtles

Ā e m ajority o f teleme try s tudies h ave i nvolved lo cating a nd tracking animals. Such studies provide insight into the size of an animal’s home range, its migratory habits of short duration (diel movement) or lon g duration ( lunar c ycles, s easonal movement or annual patterns of movement) and how it utilizes its habitat. However, t his form of t elemetry cannot prov ide  ne scale data on how much energy the animal expends in a day, when and on what the animal is feeding, when the animal is mating, or how predators are avoided. Finally tracking studies do not tell us how the environment affects the animal’s behavior. To address these questions more sophisticated telemetry studies are required and often this entails a greater reliance on smart materials and smart structures. As evident f rom t he d iscussion above, biotelemetry companies a nd re searchers a re w orking toge ther to p roduce “smart t ags” t hat re cord a spects o f t he a nimal’s en vironment (e.g., a mbient d issolved o xygen, b ioluminance, de pth, l ight intensity, s alinity, s ounds, temperature, a nd t ime) a nd provide information from the animal itself (e.g., compass bearing, heart beat f requency, lo cation, m uscle mo vement, re spiration r ates, swim speed, tilt, pitch, and roll, tail stroke frequency, and vocalizations). Ā ese de vices a re mo re c ommonly u sed to s tudy o f marine mammals, but their application to understanding energy expenditure in shes, cephalopods, and turtles are growing. We begin with examples of researchers using biotelemetry to detect muscle mo vement (used to a ssess h eart r ate, s troke f requency, and metabolic rate) and internal temperature (to assess foraging bouts and energy intake and expenditure) in shes. Most attempts to telemeter muscle activity have been conducted in the laboratory and in freshwater. Ā ese tags estimated energetic costs or activity levels, but they were not able to replicate the complex structure of the EMG waveform that would indicate the precise timing and extent of muscle recruitment. In addition, the older EMG tags could not telemeter in marine systems due to their reliance on radio wave technology. Dewar et al. (1999) developed an acoustic transmitter that could be used in a marine environment. Ā is t ransmitter wa s te sted u nder lab oratory co nditions with diff erent s pecies o f  sh, including toadsh (Opsanus t au), spiny d ogsh (Squalis aca nthias), y ellow n t una ( Ā unnu s albacares), and eastern Pacic bonito (Sarda chiliensis). Ā e E MG t ag ac quired m uscle ac tivity t hrough  ve steps. First, bipolar electrodes sensed the EMG signal from the fast- or slow-twitch muscles of the species in question. Second, the signal was amplied 2400 times with a differential amplier. Ā ir d, the signal was ltered with a bandpass  lter that set the low pass and high pass poles at 400 and 10 Hz, respectively. Fourth, the signal was sent to a voltage-controlled oscillator that produced an FM

10/16/2008 7:15:40 PM

27-12

signal w hose mo dulations e xactly m imicked t he w aveform o f the signal. Finally, an acoustic transducer (PZT-4 ceramic tube resonance f requency o f 1 22 kHz) s ent t he w aveform to t he recording equipment. Ā e authors developed three different receivers connected to one or more hydrophones to detect the FM signal and convert it back to produce the EMG waveform. Receiver 1 was abandoned because i t c ould not p rovide re al t ime d isplay o f t he de coded EMG a nd t here w ere p roblems at tributed to m ultipath, a common problem in water acoustic systems that results from the fact that sound travels in all directions, simultaneously allowing wave reections to arrive at the hydrophone at different times. A second receiver improved the design, but the authors developed a third receiver to further reduce the problem of multipath and to p rovide  ner re solution o f t he sig nal. Ā e transmitted data were acquired from a single hydrophone, passed through an 80–260 kHz ba ndpass  lter ( AP280-5-SR, A P Ci rcuit C orp.), and then demodulated with a Kenwood TS440 FM receiver. Ā e signal was then low-pass  ltered and sent to a speaker to provide an a udible sig nal c orresponding w ith t he t ail b eat o f t he  sh. With Receiver 3 and their transmitter tag, the authors calculated tail-beat f requency a nd i n t he a bsence o f m ultipath p roblems replicated t he E MG w aveform, a llowing t hem to me asure t he onset–offset duration, number of zero crossings, area under the rectied waveform, and the product of the peak and mean peak intensity. Ā us, this system could lead to  ne scale examination of m uscle re cruitment a nd b etter e stimates o f energ y re quirements for free-swimming sh. In conclusion, the authors argued that t he t ag’s de sign c ould b e c ombined w ith remote s ensing devices t hat p roduce c omplex a nalog sig nals ( strain-gauges, ow-probes) to a nswer mo re c omplex que stions re garding animal physiology and behavior. Recently, Campbell et al. (2005) used a data logger to record the h eart r ate o f bl ack c od ( Paranotothenia an gustata) o ver a long period of time. Heart rate is sensitive to stress (Cooke et al., 2002) a nd c orrelates w ith me tabolic r ates a ssociated w ith changes in activity and feeding (Armstrong, 1986; Lucas, 1994). Consequently, heart rate can be used to estimate energy requirements a nd e xpenditure. C ampbell e t a l. de termined w hether surgically implanted or externally attached electrodes provided the b est me asure o f h eart r ate. Ā e a uthors u sed a m iniature electronic m icroprocessor c ontrolled d ata lo gger d eveloped at the University of Wales Bangor. Ā is data logger provided highresolution ECG recordings (512 Hz) from a free-swimming sh. Ā e m icroprocessor a nalyzed t he w aveform a nd de termined interbeat intervals and instantaneous heart rate. Ā e system was programmed to sample the rst 20 min of every hour a nd a lso recorded t wo c omplete EC G w aves e very 4 000 b eats to c heck signal quality. In addition, the data logger housed a temperature sensor that measured ambient temperature each minute. Data were stored in a nonvolatile ash memory card to protect against power failure. Once c od h ad re covered f rom t he i mplantation o r e xternal attachment of the electrodes and data logger, they were released into a s altwater p ond w here t hey w ere sub ject to ob servation.

43722_C027.indd 12

Smart Materials

Ā e a uthors s howed t hat bl ack c od o uttted wi th e xternally attached electrodes showed an initial heart rate of 46 beats per min for the  rst 24 h, but by day 20, the average heart rate had decreased to 34 beats per min. In contrast, black cod with surgically i mplanted ele ctrodes h ad i nitial h eart r ates o f 4 4 h eart beats per min, a rate that did not decrease over the 20 day period. Campbell et al. concluded that the method of electrode attachment i n E MG s tudies w ith  sh co uld be spec ies-specic, with some sh showing normal heart rates with surgically implanted electrodes and some responding better to external attachment techniques. Sharks are often depicted as cold-blooded killers. Indeed the name seems to make good sense as most of us believe that all sh are cold-blooded. Yet, not all that long ago, it was discovered that a small group of lamnoid sharks (Order Lamniformes) possess vascular c ountercurrent h eat e xchangers lo cated i n t he b rain, musculature, a nd v iscera. Ā ese h eat e xchangers a llow t hese sharks t o m aintain a r elatively c onstant in ternal t emperature, making th em s omething l ess th an c old-blooded kill ers. G reat white sharks (Carcharodon carcharias) a re lamnoid sharks a nd McCosker (1987) hypothesized that white sharks could regulate their stomach temperature. Goldman et al. (1996) used telemetry to further support this ability and Goldman (1997) conrmed this ability. Goldman fed a transmitter, hidden inside a piece of elephant seal blubber, to a great white. With the use of thermistors a nd p ressure s ensors p otted i nside V emco t ransmitters, Goldman re corded tem perature a nd de pth i nformation f rom three different sharks. Whereas the ambient temperature of the water varied from 12.3°C to 16.4°C, the stomach temperature of the sharks averaged 26.0°C–27.1°C. Ā us white sharks were able to maintain a relatively high and narrow range of stomach temperature even in water temperatures that were 10°C–15°C colder and more variable. Another of t he l amnoid s harks i s t he s almon s hark (Lamna ditropis). L ike t he g reat w hite s hark, t hese s harks a lso p ossess vascular countercurrent heat exchangers that allow for retention of me tabolically gener ated h eat. G oldman e t a l. ( 2004) te sted this sha rk’s ability to r egulate i nternal temperature. Four f reeranging sharks i n t he Prince Wi lliam Sound a rea of Southeast Alaska were i nduced to s wallow Vemco t ransmitters ( V22-TP) or Sonotronics transmitters (CHP-87-LT) that were hidden inside herring ba it. Ā e Vemco t ransmitters re corded s tomach temperature and pressure. Ā ey recorded data for 9–10 days and had a r ange o f 1 km. S onotronics t ransmitters re corded o nly stomach temperature, transmitted for 60 days, and were detectable at half the distance of the Vemco transmitters. A directional hydrophone ( Dukane C orporation, St . C harles, I L, mo del N30A5A) detected the signal from the transmitters and Vemco receivers (VM-10, or V R-60) or U ltrasonic Telemetry Systems, Brea, C A c onverted t he a nalog sig nal to d igital. Re cords w ere obtained over time periods ranging from 3.8 to 20.7 h. Similar to the great white shark, internal temperature ranged from 25°C to 25.7°C w hereas t he tem perature o f t he w ater r anged f rom 5°C to 16°C. Goldman et al. argue that salmon are able to maintain a relatively stable internal temperature and are essentially warm

10/16/2008 7:15:40 PM

Application of Smart Materials and Smart Structures to the Study of Aquatic Animals

27-13

blooded. Ā is novel approach to homeothermy, according to the authors, p robably u nderlies t his s hark’s a bility to p ursue a nd catch highly active prey like salmon in relatively cold water.

27.4.4

Assessing Energy Expenditure with Acoustic Telemetry: Marine Mammals

Energy expenditure parameters of marine mammals were rst measured u nintentionally t hrough t he u se o f a n ac oustic d ata logger on Stel ler s ea l ions (Fletcher e t a l., 1996). F letcher e t a l. investigated the effects of noise on Steller sea lion diving behavior and communication using digital audio tape recorder (DAT) tags. Ā e authors were surprised to discover that the tagged animal’s heartbeat could be detected when the seal was at t he surface a nd qu iet, a nd t hat s wim s trokes c ould b e de tected acoustically. Swim speed was measured by counting and storing the number of revolutions of a paddle wheel, and comparing the animal’s rate of change in depth with the number of revolutions. Burgess et al. (1998) followed this tag design and combined the DAT with a compact acoustic probe (CAP), which together measure dive patterns, ambient and animal noise exposure, possible animal vo calizations, pl us a coustic si gnatures of s wim s troke, surface re spiration, a nd h eart r ate. Bu rgess e t a l. w ere a ble to record cardiac activity at depth as well when ow noise did not obscure animal sounds. Biotelemetry research took a large step forward upon the release of t he D TAG b y re searchers f rom Woods H ole O ceanographic Institute (see Figure 27.3). Ā e DTAG is a digital acoustic recording t ag t hat c ontinuously re cords t he o rientation, h eading, a nd depth of the tagged animal in synchronicity with sounds sensed by a h ydrophone ( Johnson a nd T yack, 2 003; F igure 2 7.4). Orientation is calculated from pitch and roll as determined by three-axis ac celerometers (Analog D evices A DXL202). A t hreeaxis m agnetometer co mprising l ow-power m agnetoresisitve bridge sensors (Honeywell HMC1021) is used to measure heading, via t he direction of t he earth’s magnetic  eld relative to t he tag. Ā e audio c omponents c onsist of a p iezoceramic hydrophone, a preamplier, antialias  lter, and analog-to-digital converter. Ā e DTAG ut ilizes a d igital sig nal p rocessor to c ombine s ensor measurements (consisting o f a cceleration, m agnetic  eld, and pressure sensors) with audio data to store into a memory array. Orientation can be sampled at rates up to 50 Hz. A ash memory card o f up to 3 GB c an b e u sed; i n t he i nitial t ag de sign, t he DTAG h ad o nly 4 00 Mb o f memo ry. Ā e t ag ele ctronics a re combined with a bat tery, a V HF radio beacon for tag recovery, and t wo suc tion c ups f or at tachment to t he m arine m ammal. Ā is method is noninvasive and allows up to tens of hours of data c ollection. Ā e suc tion c ups u sually rele ase o n t heir o wn due to natural leakage or social and surface-active behaviors on the tagged individual’s part; an active release is also included via a n ickel-chromium w ire t hat c orrodes a nd rele ases a n a ir l ine into e ach suc tion c up ( Johnson a nd Tyack, 2 003). To d ate, t he longest at tachment o f a D TAG w as 2 1 h o n a no rthern r ight whale and 10 h on a sperm whale.

43722_C027.indd 13

Battery container Electronics/hydraulics container Tagging chamber

FIGURE 27.3 Underwater Tagging E quipment ( UTE). D eveloped by Star-Oddi. Picture taken from www.star-oddi.com. With permission.

Sperm w hales a re c hampion d ivers t hat c an d ive at le ast 1200 m and hold their breath for over 1 h (Watwood et al., 2006). Ā e DTAG system described by Johnson and Tyack (2003) was used to determine the effects of drag and buoyancy on the swimming ga its a nd d iving pat terns o f s perm w hales ( Miller e t a l., 2004). Ā e whale’s length, mass, and surface area are important to estimate drag and buoyancy, and were measured via allometry

FIGURE 2 7.4 DTAG d eveloped by Joh nson a nd Tyack ( 2003). Top portion si mplied s chematic of t he c ircuit. B ottom pic ture s hows t he internal s tructure of t he D TAG a nd it s si ze. P icture re printed f rom Johnson, M .P. a nd Ty ack, P. L., J. Oceanic E ng., 2 8, 3, 2 003. W ith permission.

10/16/2008 7:15:40 PM

27-14

while t he re search v essel w as a longside t he w hale at t he w ater surface. Ā e DTAG weighed 500 g in air and consisted of a digital re corder, de pth s ensor, a udioband ac oustic re corder, a nd three-axis a ccelerometers a nd m agnetometers. Sa mpling ra te averaged 5. 88 Hz, a llowing qu antication of animal pitch and depth, which allows uking, descent, and ascent rates and speed to be calculated. Sperm whales most commonly used steady uking on descent and stroke-and-glide uking on ascent. Mean descent s peeds were slower t han a scent s peeds, d espite wh ales stroking steadily during descents. Ā e results indicate that sperm whales m ake b ehavioral de cisions ba sed up on b uoyancy, a nd decisions may change as buoyancy parameters change by depth, air volume, and tissue density (Miller et al., 2004). Drag an d b uoyancy aff ect signicantly t he d aily energ y expenditure of marine mammals. Buoyancy is dependent on the amount of air in the animal’s tissues and the changes in air density as an animal dives. Drag affects both horizontal and vertical movement w hereas buo yancy on ly a ffects ve rtical move ments. One way that marine mammals compensate for drag and buoyancy is to moderate their swimming “gait” or style (e.g., gliding versus steady uking, Williams et al., 2000). Nowacek et al. (2001) used the DTAG setup to study the effect of buoyancy on right whales. Ā ey set the sensor sampling rate at 23 Hz to track subtle and quick movements and the audio bandwidth was set at 8 kHz, within frequency range of right whales. Ā e authors determined t hat volume of body c avities, t he proportion of positively buoyant tissue like blubber, and the proportion o f n egatively b uoyant ti ssue lik e b one aff ected tissue buoyancy. Right whales were found to be positively buoyant even when hydrostatic pressure at depth can cause negative buoyancy in s ome s pecies, suc h a s Weddell s eals ( Williams e t a l., 2 000). Ā is poses strong implications for right whale management, as positively buoyant whales on an ascent likely have reduced maneuverability. Ā erefore r ight w hales m ay h ave l ittle to no control to escape oncoming ships that pose a threat of collision. Indeed, ship strikes are a primary factor in North Atlantic right whale mortality, accounting for more than one-third of known right whale deaths (Laist et al., 2001). Sato et al. (2003) investigated the effects of seal obesity on buoyancy and energy expenditure in Weddell seals in the wild. Swim s peed, de pth, t wo-dimensional ac celerations (stroke f requency and body angle), and temperature were recorded with a data logger (Little Leonardo Corporation, Tokyo, Japan). Ā inner f emales e xhibited p rolonged g liding d uring de scent, while f atter f emales e xhibited s wim-and-glide s wimming throughout their descent and ascent. Results suggested that the added positive buoyancy of fatter individuals shortened their ascent time with less energy expenditure but increased the effort required to descend. Ā is supports the conclusions of Williams et al. (2000) that prolonged gliding is a more efficient method of swimming. Foraging e vents a re o ften inferred from two-dimensional time–depth pro les and temporal distribution of diving bouts. However, because one cannot actually observe the animal, these inferences are not always correct (R. Davis, quoted in Senkowsky, 2003). To obtain more accurate data on foraging bouts, researchers

43722_C027.indd 14

Smart Materials

use stomach temperature transmitters (STT) to quantify the foraging f requency i n l arge m arine p redators (Carey e t a l., 1984; Wilson et al., 1992). STTs provide better data because the stomach temperature of a warm-blooded marine mammal decreases markedly when it ingests relatively cold prey items. Andrews (1998) tested remotely releasable stomach temperature transmitters on captive Stel ler sea lions to de termine t he efficacy of the method for u se i n studying f ree-ranging pinniped feeding f requency. A small transmitter with a temperature sensor is introduced into the animal’s stomach and transmits data to a log ger mounted to the animal’s skin or fur. Ā e d ata log ger pac kage (Tattletale Fast Lite, Onset Computer Corporation, MA) records dive de pth, s wim s peed, a mbient tem perature, a nd s tomach temperature. A s atellite t ransmitter (Telonics S T-6, Mesa, A Z) and V HF t ransmitter a re a lso i ncluded i n t he pac kage. W hen sampling dive depth, swim speed, and stomach temperature every 10 s a nd water temperature every 60 s, t he logger’s 512 kb memory could last almost 12 days (Andrews, 1998). Eventually the stomach transmitter is regurgitated or passed. Ā e tag package atop o f t he s eal w as remotely rele asable, ut ilizing a sm all heating element to mel t a n a nchor l ine t hat at taches t he tag to the an imal’s fu r. Ā e re searchers c ould t hen re trieve t he t ag package via satellite transmission and VHF signal. Austin et al. (2006) has expanded on the use of this technology, and provided the rst quantitative analysis of feeding frequency in a free-ranging marine mammal (grey seals) based on stomach temperature t elemetry. Austin e t a l. w ere interested in h ow f eeding frequency would differ between males and females, and how time of day and between-meal duration affected foraging timing. A STT (Wildlife Computers) was inserted into an anesthetized gray seal’s stomach via an equine intubation tube. Ā e STT was designed to be large enough to delay passage but small enough so that over time the foam in the apparatus would break down and allow passage. A satellite-relay data logger (SRDL, Wildlife Computers, WA or ST-18-Telonics, AZ) was attached to t he dorsal surface of the seal and included a time–depth recorder; the data package was removed upon the seals’ return to their rookery. Ā ere were signicant feeding diff erences b etween m ale a nd f emale s eals. M ales f ed mo re often than females and spent more time feeding per day. Ā is nding was not surprising in light of the fact that gray seal males in this study were between 30% a nd 45% heavier t han female seals. For both sexes, the largest meal size was at dawn and the smallest meal at dusk. In addition, seals waited longer to eat their next meal after gorging t hemselves t he p revious me al. S ome v ariability i n i ndividual feeding behavior was documented. Ā e authors speculated that prey patchiness and individual foraging tactics between individuals likely caused some variance (Austin et al., 2006).

27.4.5

Assessing Energy Expenditure Using Video Telemetry: Fishes, Cephalopods, and Turtles

Greg Marshall (www.nationalgeographic.com/speakers/prole_ marshall.html) i s a p ioneer i n t he a rea o f v ideo teleme try a nd may h ave b een t he  rst to at tach a c amera to a f ree-ranging

10/16/2008 7:15:42 PM

Application of Smart Materials and Smart Structures to the Study of Aquatic Animals

marine animal and record its behavior. Ā e idea occurred to him while he was watching a shark swimming with a host of attached remora sh. Marshall realized that it might be possible to attach a v ideo camera to t he shark to ob tain a s hark’s eye v iew of t he world. Such a view might allow scientists to better understand its world. With this camera, one could observe the strategies used to hunt prey, to  nd mates or to a void predators under natural conditions. With those thoughts, Marshall developed and tested “crittercam.” Crittercam te chnology ( www.nationalgeographic.com/crittercam/) i ncorporates s ophisticated v ideo, a udio, a nd t ime– depth re cording s ensors t hat, f or t he  rst t ime, a llow de tailed examination of the factors that inuence tness and survival of marine mammals (Marshall, 1998). In addition, transmitters are also attached that record stroke frequency through accelerometers, c ompass b earing, de pth, tem perature, l ight le vels, a nd velocity. Ā is combined technology allows direct observation of swimming periods, stroke frequency, glide sequences, and allows researchers to actually observe, rather than postulate, why marine mammals are diving—whether to forage, rest, or socialize. Ā e c amera s ystem c an ei ther f ace f orward, i n o rder to observe head movements, foraging and socializing, or it can face backward, i n order to ob serve propulsive movements of  ukes and  ippers a nd p erhaps s ocial i nteractions. Ā e rst public viewing of images from crittercam occurred in 1993 on Geographic E XPLORER. To date, crittercam has been used on more t han 3 0 m arine s pecies i ncluding g reat w hite s harks, emperor penguins, killer whales, sea turtles, seals, sperm whales, and w alruses. Cu rrently, t his te chnology i s b eing ad apted f or land a nimals w ith c ameras b eing at tached to b ears, l ions, a nd tigers among others. Video telemetry is relatively new a nd examples of published studies with shes, turtles, or c ephalopods are rare or none xistent. Most of t he publications u sing v ideo telemetry have been conducted with marine mammals, particularly seals. In the next section, we will consider representative studies from these animals. H ere, w e w ill t alk a bout a d ifferent app roach to v ideo telemetry t hat i s b eing de veloped b y D avid S cheel at A laska Pacic U niversity a nd h is c olleagues a nd s tudents f rom t he Colorado S chool o f M ines a nd t he U niversity o f A rizona. D r. Scheel, w ith f unding f rom t he National Science Foundation, is developing a rob otic c amera t hat c ould t rack a Gi ant Pac ic octopus w herever i t v entured. Ā e rob otic c amera h as b een named Shad ow ( http://marine.alaskapacic.edu/octopus/ Shadow.php). Ā e camera will be housed in an autonomous underwater v ehicle t hat c an f ollow t he o utput o f a n ac oustic pinger attached to the octopus. Ā is underwater vehicle is a very sophisticated sm art s tructure re quiring a n a bility to mo ve through the water, navigate around obstacles, and keep the octopus in the camera’s viewnder. Ā e project is divided into three areas with Dr. Scheel’s team working on collecting and analyzing b iological t hat c an p rovide i nsight i nto t he pa rameters needed to f ollow a mo ving octopus, developing a v ideo system that provides enough scale to assess what the octopus is actually doing, a nd de termining h ow t he v ideo d ata w ill b e a nalyzed.

43722_C027.indd 15

27-15

Faculty advisor to the project, Dr. Ā omas Vincent, University of Arizona, w ill le ad t he te am t hat de velops t he p latform a nd designs the thrusters, buoyancy controls, and stabilizer allowing Shadow to move through the water. Faculty advisor, Dr. Tyrone Vincent, Colorado School of Mines, will work with students to design t he t racking s ystem. Ā is system will need to sense the acoustic sig nal a nd de termine t he r ange a nd d irection to t he octopus. B oth t he N avigation a nd P ower C ontrol te ams w ill work together to develop the controls necessary to keep the octopus within the viewnder of the camera and switch it on or off, depending on the physical activity of the animal. With Shadow, Scheel hopes to shed light on the life history of the Giant Pacic Octopus, w hich i s e xperiencing a p opulation de cline i n t he Pacic Northwest. Dr. Scheel is hoping to learn more about the migration habits, foraging strategies, and predator defense strategies of the Giant Pacic Octopus (Enteroctopus doeini) found in Prince William Sound among other places.

27.4.6

Assessing Energy Expenditure Using Video Telemetry: Marine Mammals

Davis et al. (1999) attached a small video imaging system to the head and a d ata logger to t he dorsal surface of Weddell seals in the Antarctic. Ā e data logger recorded time, depth, water speed, and compass bearing once per second;  ipper stroke frequency and ambient sounds were recorded continuously with a h ydrophone. Ā e video system recorded images of the seals’ heads and the en vironment i n f ront o f t he a nimal. Re sults s hed l ight o n how W eddell s eals s talk l arge p rey i n t hree-dimensional environments, and also pointed to the possibility that pinnipeds may use passive detection of prey as determined by their lack of emitted sounds. Williams e t a l. (2000) ut ilized si milar me thods to i nvestigate diving strategies in marine mammals. Weddell seals, an elephant seal, a blue whale, and a bottlenose dolphin were outtted with a video i maging s ystem a nd a t ime–depth re corder (T DR). Swimming mode, relative stroke amplitude, stroke frequency, and gliding periods were determined using a mot ion analysis system (Peak Performance). Data were matched to duration and depth of the associated dive. To assess effect of changes in locomotion on energetics, p ostdive o xygen c onsumption o f W eddell s eals w as measured b y h aving t he s eals b reathe i nto a me tabolic h ood. Results i ndicated t hat W eddell s eals a nd l ikely ot her m arine mammals t ake adv antage o f l ung c ompression a nd sub sequent negative buoyancy by engaging in prolonged gliding on descent; this behavior c onserves energ y, a llowing m arine m ammals to increase aerobic dive duration and achieve deeper dives. Ā e most recent application of this video-data logging system is the estimation of energy expended during foraging dives by Weddell seals in the Antarctic. Williams et al. (2004) measured the associated energetic costs of foraging including metabolism, locomotion, a nd prey d igestion by me asuring p ostdive oxygen consumption. Costs of locomotion increased with the number of ipper strokes. Also, recovery time due to oxygen consumption

10/16/2008 7:15:42 PM

27-16

Smart Materials

was i ncreased b y 4 5% w hen s eals to ok f oraging d ives t hat t he authors at tribute to t he ne cessary energ y f or w arming a nd digestion of prey (Williams et al., 2004).

27.5 Predicting the Future Predicting t he near f uture of biot elemetry technology is pretty simple. Tags will get smaller and smarter with a w ider range of sensors that will operate for longer periods of time and possess more memory. It is also likely that biotelemetry systems in the future w ill p rocess mo re d ata a nd p rovide mo re s ophisticated analyses, making easier the problem of guring out what 1 million or more data points are telling us about the animal. In addition, we expect that behavioral and physiological ecologists and others interested in nonhuman animals will use biotelemetry to study a n i ncreasing v ariety o f s pecies w ith a g reat v ariety o f applications. Ā at is in the near term. It is more difficult to attempt to p redict w hat t he b iotelemetry e quipment w ill lo ok like 50 or even 20 years from now or how scientists interested in aquatic animals will use such equipment. Researchers who were instrumental in the early development and promotion of telemetric systems and who today are providing important examples of th is m ethodology h ave a ddressed th is q uestion r ecently. Block ( 2005), C ooke e t a l. ( 2004), a nd Rop ert-Coudert a nd Wilson ( 2005) p rovide e xcellent re views o f t he h istory o f t his technology, current applications, and they offer insight into what the future holds and does not hold for its use. In this section we will not paraphrase their insight, but instead we examine a new smart tag that is currently being developed and tested that may provide better data on mortality and survival of marine mammal populations. I n a ddition, we br iey c onsider t he p ossibility o f telemetry systems based on nanotechnology and termed “Smart Dust” by developers. Finally we speculate on the possibility that biotelemetry could be used to directly test various theories proposed by behavioral ecologists a nd ex perimental psychologists to account for the behavior of animals.

27.5.1

Life History Transmitter

Horning and Hill (2005) describe the development of “Ā e Life History Transmitter” that they predict will become an important to ol i n t he s tudy o f aqu atic a nimal p opulation dy namics. Ā e L ife H istory T ransmitter i s de signed to c ollect v ital d ata from marine vertebrates for up to 10 years, which could amount to the animal’s entire lifetime. Ā e tag would record life history data from the time it is implanted until the animal dies. At this point, t he t ransmitter i s rele ased t hrough de cay a nd up on reaching the surface of the ocean or exposed to air on a beach it transmits its data to a satellite. If successful, the authors contend that long-term survival rates and population dynamics of marine v ertebrates c ould b e i nvestigated mo re ac curately. I n addition, it should be possible to obtain better estimates of the energy required to survive and reproduce. Horning and Hill (2005) began by modeling their transmitter after Wildlife Computer’s pop-up archival transmitter (PAT)

43722_C027.indd 16

tags because these tags record information until they are exposed to the air at which time they transmit their data to a satellite via the ARGOS system. Ā e researchers then detailed the problems the tag would have to solve. For example, the tag had to discriminate b etween t wo s tates, mo rtality a nd e xtrusion. W hen t he animal dies, a mortality event has occurred and when the tag is exposed to a ir either via the surface of the ocean or on a b each following t he a nimal’s death, a n extrusion event has occurred. Ā e authors decided t hat an internal temperature sensor could be used to determine mortality. To determine the point at which the tag was exposed to air following extrusion, the authors used a c ombination l ight a nd c onductivity/immersion s ensor. Horning and Hill then set out to build two prototypes of the Life History Tag. Ā e rst tag (Standard LHX) would house temperature, light-level, and immersion sensors only. As such, it would nd application in population dynamic studies where researchers w ant b etter d ata o n su rvival. Ā e mo re s ophisticated t ag (Enhanced LHX) would record pressure and movement through motion sensors in addition to internal temperature, light-level, and conductivity/immersion. Ā e enhanced version would provide b ehavioral d ata ( total d ives, t otal v ertical d isplacement, stroke frequency) over the lifetime of the animal and contribute greatly to s tudies on energetics. To focus their efforts, Horning and Hill ranked the necessary components in terms of the potential to fail. In decreasing order of probability to fail, battery, sensors, A to D converter, housing, and soft ware comprised the list of problems to solve. Ā e a uthors p ower t heir t ransmitter o ver a 1 0-year p eriod with t wo p rimary bat tery s ystems. A B CX85PC1 l ithium/ bromine-chloride c omplex bat tery ( Wilson Gre atbatch T echnologies, M A) p owers t he c ontroller s oft ware a nd a L SH26180 lithium/thionyl-chloride battery powers the transmitter’s amplier. Both batteries lose 3% of their power annually resulting in 70% c apacity a fter 10 years. Ā e s tandard L HX t ag ut ilizes a n YSI44017 t hermistor ( YSI I nc., B eavercreek, O H) to me asure internal temperature, a TAOS TSL257 light-to-voltage converter (TAOS Inc., Plano, TX) to measure light level, and a contact-free immersion sensor developed by Wildlife Computers, Redmond, WA. H ousing re quirements i ncluded p ositive b uoyancy, p ressure rated to 1000 m, a helical antenna with a cover that allowed for p ropagation o f U HF sig nals, a nd mo isture p roong for residing i nside t he a nimal for a de cade or more. Ā e L HX t ag measured 4 2 mm i n d iameter a nd 1 22 mm i n leng th. Ā e tag weighed 115 g in air, or about .1% of the target species, stellar sea lions. Ā e tag’s buoyancy rating in salt water was −38 g. To ensure the tag remains operational for 10 years, the tag is interrogated every 30 min for sensor-data acquisition, error-check, tag-state, mortality, and extrusion events. As mentioned briey, mortality is measured by internal temperature. For example, if the recorded temperature falls outside the user-dened range of appropriate temperature for 96 interrogations, the animal is presumed to be dead. Ā is positive test for TST1 triggers a heuristic algorithm to determine the highest likelihood o f m aking a c orrect de cision w ith re spect to e xtrusion. If all sensors are operating correctly, the light level sensor

10/16/2008 7:15:42 PM

Application of Smart Materials and Smart Structures to the Study of Aquatic Animals

indicates t hat t he t ag i s i n a mbient l ight a nd t he i mmersion sensor detects air the tag assigns a YES to TST2 and assumes an extrusion e vent h as o ccurred. At t his p oint, d ata t ransmission begins. Ā e speed with which data is transmitted depends upon the condence level of the heuristic algorithm. If the algorithm is highly condent, data are transmitted over a 10-day period. If the algorithm has detected possible problems with the sensors, it produces a m oderately co ndent v alue a nd t he t ransmitter reduces t ransmission s peed a nd s ends t he d ata o ver a 2 0 d ay period. At the lowest level of condence, transmission is reduced still further and could take up to 1 00 days to t ransmit all data. Ā e longer periods of time to transmit are provided to allow for full e xtrusion i n t he e vent t he t ag h as b een o nly pa rtially exposed. To date, the Life History Transmitter has been tested on a few Steller sea lions a nd subjected to f ull simulation studies w ith a computer. Ā e authors continue to de velop the Enhanced LHX and p lan a mo re e xtensive de ployment o f t he St andard t ag. I f possible, H orning a nd H ill p lan to i mplant b oth t he s tandard and the enhanced version in the same animal to assess reliability. Ā e authors contend that this tag offers a strong advance over existing methods for assessing population dynamics and could lead to sub stantially mo re a nd mo re ac curate mo rtality a nd behavioral data involving fewer animals.

27.5.2 Biotelemetry and Smart Dust In the past few years, scientists and engineers have been working on a p roject t hat has i mportant i mplications for biotelemetry studies. Smart Dust (http://robotics.eecs.berkeley.edu/~pister/ SmartDust/ a nd w ww.computerworld.com/mobiletopics/ mobile/story/0,10801,79572,00.html) re fers to de vices t hat are very sm all w ireless s ( MEMS) t hat s ense s timuli, p rocess data, a nd p rovide b idirectional c ommunication. Ā e go al o f the Smart Dust project at the University of California, Berkeley was to design sensors that could detect temperature, humidity, barometric p ressure, l ight i ntensity, t ilt a nd v ibration, a nd magnetic  elds. I n add ition, t he de vice c ould op erate f or 2 years or more with a 1% duty cycle, and provide bidirectional radio c ommunication. Fi nally, a n onb oard m icroprocessor could provide for preliminary analysis of the data. Amazingly, the sensors, power supply, analog circuitry, bidirectional optical communication, and programmable microprocessor would all  t w ithin 1 c ubic millimeter a nd data could be communicated at distances up to 1000 m. Ā e investigators, Kris Pister, Joe Kahn, Bernhard Boser, and Steve Morris, suggest that Smart Dust has obvious applications to t he m ilitary w here suc h de vices c ould p rovide bat tleeld surveillance, treaty monitoring, and equipment movement. Such devices c ould a lso p rovide p roduct qu ality mo nitoring, sm art offices with individually tailored environments for workers, and interfaces f or d isabled p ersons. W e c ontend t hat sm art d ust could be of signicant value to the study of nonhuman animals. Ā ese de vices c ould p rovide c ontinuous d ata o n en vironmental variables like temperature, humidity, salinity, light levels, magnetic

43722_C027.indd 17

27-17

elds, and movement. Ā ese data could be recorded over long or short periods of t ime depending upon t he degree of resolution desired. Smart Dust could be used to lo cate and track animals, assess na vigational a bility a nd en vironmental c onstraints o n daily or annual migratory behavior, determine movement, and provide e stimates o f energ y re quirements a nd u se. O f i mportance, t his could a ll b e done w ith e quipment s o small a s to b e completely unobtrusive to the animal.

27.5.3

Methods of Testing Behavioral Hypotheses Using Biotelemetry

In 2001 and 2002, one of us (J.E.P.) spent a tot al of 5 mo nths in Antarctica working with Randy Davis, Texas A&M University at Galveston, L ee Fu iman, U niversity o f T exas M arine S cience Institute, Port Aransas, Terrie Williams, University of California at Santa Cruz, and Markus Horning, Hateld Marine Mammal Center, O regon St ate U niversity s tudying f oraging a nd d iving characteristics of Weddell s eals u nder t he f ast ic e i n McMurdo Sound. In analyzing certain of the data, an application for the use of biotelemetry became apparent that has not been utilized to any great extent. In their rst seasons in McMurdo Sound, Davis and coworkers used an isolated hole protocol that was rst developed by Kooyman (1965). In this paradigm, researchers go out on the ice and nd a location where there are no cracks or holes in the ice within a circle whose diameter is 2 km or greater. Ā en, a hole is drilled in the center of this circle and seals are caught and transported to this location where they are outtted with video telemetry e quipment (see a bove de scription f rom D avis e t a l., 2 004). Seals are released from this location where they conduct series of exploratory a nd f oraging d ives o ver a p eriod o f s everal d ays. Because t he s eal c annot ge t to a nother h ole to b reathe, i t h as return to the isolated hole. When the seal returns or when it hauls out at a location close to camp, the equipment and all physiological, environmental, and video data are retrieved. In analyzing second by second compass bearing data from the diving seals, it was observed that seals would typically head out on a specic heading, say 90°, and return along the same heading, now 270°. Ā is dive would typically last 16–20 min and was followed b y re sting i n t he h ole f or 3 –5 min a nd d iving a gain. Interestingly, t he s eal w ould t ypically h ead off in a diff erent direction from the rst, say 135° and return along a heading of 315°. W hen one lo oked at a ll of t he d ives toge ther, it app eared that the seal left the hole on a heading that appeared to be chosen at r andom, b ut o ne t hat d id not re peat t he p revious h eadings. Ā is behavior has implications for the foraging strategies of the Weddell s eal. I f t here a re l imited s chools o f p rey t hat mo ve around haphazardly, it is likely that the probability of nding food in the last known place is less than that of  nding food in a new place. If true, the strategy of heading off in a different direction each t ime a nd not re peating d irections might be t he most efficient. Ā is s trategy re quires t hat W eddell s eals a re a ble to maintain a specic heading and that they remember where they have previously been. In essence, Weddell seals must possess a good sense of direction and good spatial memory.

10/16/2008 7:15:42 PM

27-18

What makes this even more interesting is that this strategy appears to b e very similar to t hat used by a r at in a r adial arm maze. I n a n eig ht-arm r adial a rm m aze, r ats app ear to c hoose one of eight arms at random and explore to t he end looking for food. If food is found, because it is not replaced, it is to the rat’s advantage to choose a different arm on the next trial. In solving the maze, rats pick one arm, go down that arm, nd t he food, and t hen re turn to t he c enter. F rom h ere, t he r at w ill p ick, at random, another arm and run down it. Because food at the end of the arm is not replaced, the rat must remember which arms it has visited and which it has not. Several studies have shown that the rat is very good at this task and has shown remarkable spatial memory ( Olton a nd S amuelson, 1 976; Rob erts, 1 984) I n f act, researchers have shown that the rat’s spatial memory is such that it can remember which arms it has visited in a maze that has as many as 24 arms (Roberts, 1979). One of the difficulties of these studies on spatial memory is that they were conducted under laboratory conditions. Ā us, we know the animal is capable of the task, but we do not know if this ability is ecologically relevant. Biotelemetry may offer a w ay for behavioral e cologists a nd e xperimental ps ychologists to s tudy t hese abilities in a f ree-ranging animal. By k nowing where the animal is, the tracks that it used to navigate and when it is foraging it should be possible to judge whether the behavior observed in the laboratory reects that observed in the eld. In the end, biotelemetry m ight p rovide a me ans f or te sting i n t he a nimal’s nat ural environment, the theories that behavioral ecologists and experimental psychologists posit to account for an animal’s behavior.

27.6 Conclusions Smart materials and smart structures serve as a foundation for a wide variety of the equipment used to study aquatic animals. Ā is is particularly true in the area of biotelemetry where smart tags and re ceivers a re p laying a n i ncreasing role i n u nderstanding foraging, habitat selection, mating behavior, migration, navigation, population dy namics, a nd t he w ays i n w hich t he en vironment constrains behavior. Ā e application of smart materials and smart materials to t he s tudy of aqu atic a nimals h as i ncreased sig nicantly but scientists have probably realized only a fraction of the applications that are possible. In the next 10 years, we should see biotelemetry equipment becoming increasingly smaller a nd more s ophisticated a nd a s a re sult t he n umber o f s pecies t hat could be studied with this methodology will increase as well as the t ypes o f que stions t hat c an b e a nswered. Suc h s tudies w ill contribute greatly to the knowledge base and will go a long way toward sustaining a great resource, aquatic animals.

References Aitken, J. P., O’Dor, R. K., and Jackson, G. D. 2005. Ā e secret life of the giant Australian cuttlesh Sepia apama (Cephalopoda): Behaviour and energetics in nature revealed through radio acoustic p ositioning a nd t elemetry (RAPT). Journal o f Experimental Marine Biology and Ecology, 320, 77–91.

43722_C027.indd 18

Smart Materials

Aglen, A. 1994. S ource o f err or in aco ustic estima tion o f  sh abundance. I n: Marine  sh B ehaviour in C apture an d Abundance Estimation. Eds. A. Ferno and S. Olsen. Fishing News Books, Oxford, U.K., pp. 107–133. Andrews, R. D. 1998. Remotely releasable instruments for monitoring the foraging behaviour of pinnipeds. Mar. Ecol. Prog. Ser. 175, 289–294. Armstrong, J. D. 1986. Heart rate as an indicator of activity, metabolic ra te, f ood in take a nd dig estion in p ike, Esox l ucius. Journal of Fish Biology, 29, 207–221. Ashley, S. 1995. S mart skis a nd o ther ada ptive str uctures. Mechanical Engineering, 117, 76081. Austin, D., B owen, W. D., McMillan, J. I., and B oness, D. J. 2006. Stomach t emperature t elemetry r eveals t emporal pa tterns of f oraging success in a f ree-ranging ma rine ma mmal. Journal of Animal Ecology, 75, 408–420. Baird, R. W., Hanson, M. B., and Dill, L. M. 2005. Factors inuencing the diving behaviour of sh-eating killer whales: Sex differences and diel a nd interannual variation in di ving rates. Canadian Journal of Zoology, 83(2), 257–267. Baumgartner, M.F. and Mate, B. R. 2005. Summer and fall habitat of North Atlantic right whales (Eubalaena glacialis) inferred from s atellite telemet ry. Canadian Journal of Fisheries and Aquatic Science, 62, 527–543. Baumgartner, M. F., C ole, T. V. N., Cla pham, P., a nd M ate, B . R . 2003. North Atlantic right whale habitat in the lower Bay of Fundy a nd o n the SW S cotian S helf d uring 1999–2001. Marine Ecology Progress Series, 264, 137–154. Block, B . A. 2005. P hysiological e cology in the 21st cen tury: Advancements in b iologging s cience. Integrative a nd Comparative Biology, 45, 305–320. Bridger, C. J., B ooth, R . K., M cKinley, R . S., S cruton, D. A., a nd Lindstrom, R . D . 2001. M onitoring  sh b ehaviour wi th a remote, co mbined aco ustic/radio b iotelemetry syst em. Journal of Applied Ichthyology, 17, 126–129. Burgess, W., Tyack, P., L eBoeuf, B . J., a nd C osta, D . P. 1998. An intelligent aco usatic r ecording t ag  rst r esults f rom f reeranging northern elep hant s eals. Deep-Sea Re search II , 45, 1327–1351. Butler, P. J. and Jones, D. R. 1997. Physiology of diving of birds and mammals. Physiol. Rev. 77, 837–899. Campbell, H. A., Bishop, C. M., Davies, D. A., and 7 E gginton, S. 2005. Re cording lo ng-term hea rt ra te in Paranotothenia angustata u sing a n e lectronic da talogger. Journal of Fi sh Biology, 67, 1150–1156. Cao, W., Harley, C. H., and Waser, R . 1999. Smart materials and structures. Proceedings of the National Academy of Science, USA, 96, 8330–8331. Candy, J . R . a nd Quinn, T. P. 1999. B ehavior o f ad ult c hinook salmon ( Oncorhynchus ts hawytscha) in B ritish C olumbia coastal wa ters determine d f rom u ltrasonic telemet ry. Canadian Journal of Zoology, 77, 1161–1169. Casey, J. G. and Kohler, N. E. 1992. Tagging studies on the shortn mako sha rk ( Insures o xrinchus) in the Western N orth

10/16/2008 7:15:42 PM

Application of Smart Materials and Smart Structures to the Study of Aquatic Animals

Atlantic. Australian J ournal o f M arine a nd F reshwater Research, 43, 45–60. Chilvers, B. L., Rob ertson, B. C., Wilkinson, I. S., Duignan, P. J., and G emmell, N. J . 2005. M ale ha rassment o f f emale New Z ealand s ea lio ns, P hocarctos ho okeri: mo rtality, injury, a nd ha rassment a voidance. Canadian J ournal of Zoology, 83 (5), 642–648. Clark, M. E., Wolcott, T. G., Wolcott, D. L., and Hines, A. H. 1999. Foraging a nd a gonistic ac tivity co-o ccur in f ree-ranging blue crabs (Callinectes sapidus): observation of animals by ultrasonic telemetry. Journal of Experimental marine Biology and Ecology, 233, 143–160. Collins, S. 2003. Brainy bra gives lift when you need it, 11/2/2006. Cooke, S. J., Hinch, S. G., Wikelski, M., Andres, R. D., Kuchel, L. J., Wolcott, T. G., and Butler, P. J. 2004. Biotelemetry: a mechanistic approach to ecology. Trends in Ecology and Evolution, 19, 334–343. Davis, R . W., F uman, L., Williams, T. M., C ollier, S., H agey, W., Kanatous, S. B., Kohin, S., and Horning, M. 1999. Hunting behavior of marine mammal beneath the Antarctic fast-ice. Science, 283, 993–996. Davis, R . W., F uiman, L. A., Williams, T. M., H orning, M., a nd Hagey, W. P. 2003. Classication of Weddell seals dives based on thr ee-dimensional mo vements a nd vide o r ecorded observations. Mar. Ecol., Prog, Ser., 264, 109–122. Davis, R . W., Hagey, W., a nd Horning, M. 2004. Monitoring the behavior a nd m ulti-dimensional mo vements o f Weddell seals using an animal-borne video and data recorder. Mem. Natl. Inst. Polar. Res., Spec., Issue, 58, 150–156. Dewar, H ., De ffenbaugh, M., Ā urmond, G., L ashkari, K., a nd Block, B. 1999. Development of an acoustic telemetry tag for monitoring electromygrams in free-swimming sh. Journal of Experimental Biology, 202, 2693–2699. Domeier, M. L., K iefer, D ., N asby-Lucas, N., Wagschal, A., a nd O’Brien, F. 2005. T racking P acic bluen t una ( Ā u nnus thynnus orientalis) in the northeastern Pacic with an automated alg orithm tha t estima tes la titude b y ma tching s easurface-temperature da ta f rom s atellites wi th t emperature data from tags on sh. Fisheries Bulletin, 103, 292–306. Fletcher, S., LeBoeuf, B. J., Costa, D. P., Tyack, P. L., and Blackwell, S. B. 1996. Onboard acoustic recording from diving elephant seals. Journal of the Acoustical S ociety of America, 100, 2531–2539. Gibbs, W. W. Scientic American Explorations (May, 1996). Goldman, K. J. 1997. Regulation of body temperature in the white shark, Carcharodon c archarias. J ournal o f Co mparative Physiology B, 167, 423–429. Goldman, K. J., Anderson, S. D., McCosker, J. E., and Klimley, A. P. 1996. Temperature, swimming depth, and movements of a white shark at the South Farallon Islands, California. In Great White Sharks: Ecology And Behavior, Eds. A. P. Klimley and D. G. Ainley, San Diego, CA: Academic Press, pp. 111–120. Goldman, K. J., Anderson, S. D., L atour, R . J., a nd M usick, J. A. 2004. Homeothermy in ad ult salmon sharks, Lamna ditropis. Environmental Biology of Fishes, 71, 403–411.

43722_C027.indd 19

27-19

Hjellvik,V., Michalsen, K.,Aglen, A., and Nakken, O. 2002. Examining the effective height of the bottom trawl using acoustic survey recordings. I CES S ymposium o n Acoustics in Fisheries a nd Aquatic Ecology, June 10–14, 2002. Paper no. 118. 20pp. Hill, R. D. and Braun, M. J. 2001. Geolocation by light-level. Ā e next st ep: la titude. I n Electronic T agging a nd T racking i n Marine Fisheries, Eds. J. R. Sibert and J. Nielsen, Dordrecht: Kluwer Academic, pp. 237–248. Holland, K. N., Gush, A., Meyer, C. G., Kajiura, S., Wetherbee, B. M., and Lowe, C. G. 2001. Five tags applied to a single species in a sin gle lo cation: Ā e tig er sha rk exp erience. I n Electronic Tagging an d Tracking in M arine Fi sheries, Eds. J. R. Sibert and J. Nielsen, Dordrecht: Kluwer Academic, pp. 237–248. Hooker, S. K. and R. W. Baird. 2001. Diving and ranging behaviour of o dontocetes: A metho dological r eview a nd cri tique. Mammal Review 31, 81–105. Horning, M. and Hill, R. D. 2005. Designing an archival satellite transmitter f or lif e-long dep loyments o n o ceanic v ertebrates: Ā e life history transmitter. IEEE Journal of Oceanic Engineering, 30, 807–817. Jackson, G. D., O’Dor, R . K., and Andrade, Y. 2005. First tests of hybrid acoustic/archival tags on squid and cuttlesh. Marine and Freshwater Research, 56, 425–430. James, M. C., Ot tensmeyer, C. A., a nd M yers, R . A. 2005. Indentication o f hig h-use ha bitat a nd thr eats t o lea therback sea turtles in northern waters: New directions for conservation. Ecology Letters, 8, 195–201. Johnson, M. P. and Tyack, P. L. 2003. A digital acoustic recording tag for measuring the response of wild marine mammals to sound. Journal of Oceanic Engineering, 28, 3–12. Kooyman, G. L. 1965. T echniques us ed in me asuring de ving capacities of Weddell seals. Polar Records, 12, 391–394. Kooyman, G. L. 2004. Genesis and evolution of bio-logging devices: 1963–2002. Mem. Natl. Inst. Polar Res. Spec. Issue, 58, 15–22. Laist, D. W., Knowlton, A. R., Mead, J. G., Collet, A. S., and Podesta, M. 2001. Collisions between ships and whales. Marine Mammal Science, 17, 1632–1644. Lander, M. E. and Gulland, F. M. D. 2003. Rehabilitation and postrelease monitoring of Steller sea lion pups raised in captivity. Wildlife Society Bulletin, 31(4), 1047–1053. Lowry, M. B. a nd S uthers, I. M. 1998. Home ra nge, ac tivity a nd distribution pa tterns o f a t emperate r ocky-reef  sh, Cheilodactylus fuscus. Marine Biology, 132, 569–578. Lucas, M. C. 1994. Heart rate as an indicator of metabolic rate and activity in adult Atlantic salmon, Salmo salar. Journal of Fish Biology, 44, 889–903. Marshall, G. J. 1998. Crittercam: An animal-borne imaging and data logging system. Marine Technology Society Journal, 32, 11–17. Mate, B . R ., L agerquist, B . A., a nd C alambokidis, J . 1999. Movements of North Pacic blue whales during the feeding season off southern California and their southern fall migration. Marine Mammal Science, 15(4):1246–1257. Mate, B . R ., D uley, P., L agerquist, B . A., Wenzel, F., S timpert, A., and Clapham, P . 2005. Obs ervations o f a f emale N orth Atlantic right wh ale (Eubalaena g lacialis) in sim ultaneous

10/16/2008 7:15:42 PM

27-20

copulation with two males: supporting evidence for sperm competition. Aquatic Mammals, (31)(2): 157–160. McCosker, J. E. 1987. Ā e white shark, Carcharodon carcharias, has a warm stomach. Copeia, 1, 195–197. Miller, P. J. O., Johnson, M. P., Tyack, P. L., and others. 2004. Swimming gaits, passi ve dra g a nd b uoyancy o f di ving sp erm w hales Physeter macr ocephalus. J ournal of E xperimental B iology, 207, 1953–1967. Moore, M . J., K nowlton, A. R ., K raus, S. D., McLellan, W. A., and Bonde, R . K. 2004. M orphometry, gr oss mo rphology a nd available hist opathology in N orth Atlantic rig ht w hale (Eubalaena g lacialis) mo rtalities (1970–2002). Journal o f Cetacean Research and Management, 6(3), 199–214. Naito, Y. 2004. Bio-logging science. Mem Natl Inst Polar Res, Spec, Issue, 58, 118–132. Nelson, D . R ., M cKibben, J . N ., S trong, J r., W. R ., L owe, C . G ., Sisneros, J. A., Schroeder, D. M., and Lavenberg, R. J. 1997. An acoustic tracking of a megamouth shark, (Megachasma pelagios): a cr espuscular v ertical migra tor. Envrionmental Biology of Fishes, 49, 389–399. Nowacek, D. P. et al. 2001. Buoyant balaenids: the ups and downs of buoyancy in right whales. Proceedings of the Royal Society of London, Series B, 268, 1811–1816. Olton, D . S. a nd Sa muelson, R . J . 1976. Remem brance o f paces passed: S patial memo ry in ra ts. Journal of E xperimental Psychology, Animal Behavior Processes, 2, 97–116. Pepperell, J. G. and Davis, T. L. O. 1999. Post-release behavior of black marlin, Makaira in dica, c aught o ff the Great Barrier Reef with sportshing gear. Marine Biology, 135, 369–380. Perkins, S. Science News Online (November 22, 1997). Polovina, J. J., Balazs, G. H., Howell, E. A., Parker, D. M., Seki, M. P., and Du tton, P. H. 2004. F orage a nd migra tion ha bitat o f loggerhead (Caretta caretta) and olive ridley (Lepidochelys olivacea) s ea t urtles in the cen tral N orth P acic Ocean. Fisheries Oceanography, 13, 36–51. Roberts, W. A. 1979. Spatial memory in the ra t on a hiera rchical maze. Learning and Motivation, 10, 117–140. Roberts, W. A. 1984. S ome issues in a nimal spa tial memo ry. In Animal C ognition, E ds. H. L. Ro itblat, T. G. B ever, a nd H. S. Terace, Hillsdale, NJ: Erlbaum, pp. 425–443. Ropert-Coudert, Y. and Wilson, R . P. 2005. Trends and p erspectives in animal-attached remote sensing. Frontiers in Ecology and the Environment, 8, 437–444. Sato, K., Mitani, Y., Cameron, M. F. Siniff, D. B., and Naito, Y. 2003. Factors affecting stroking patterns and body angle in diving Weddell s eals under na tural co nditions. Ā e J ournal o f Experimental Biology, 206, 1461–1470. Senkowsky, S. 2003. High tech surveillance: Adapting new technology for marine studies. BioScience, 53, 1150–1157.

43722_C027.indd 20

Smart Materials

Shirley, M. A. a nd Wolcott, T . G. 1991. A t elemetric st udy o f microhabitat selection by premolt and molting blue crabs, Callinectes s apidus ( Rathbun) wi thin a sub estuary o f the Pamlico Ri ver. North Ca rolina M arine Be havior a nd Physiology, 19, 133–148. Sigurdsson, T., Ā orsteinsson, V., a nd Gustafsson, L. 2006. In si tu tagging o f de ep-sea r edsh: Application o f a n under water, sh-tagging system. ICES J ournal o f M arine S cience, 63, 523–531. Small, R . J., L owry, L . F., Hoef, Jay, M . V., Frost, K . J., D eLong, R. A., and Rehberg, M. J. 2005. Diff erential movements by harbor s eal p ups in co ntrasting Alaska en vironments. Marine Mammal Science, 21(4), 671–694. Spotila, J. R ., Reina, R . D., S teyermark, A. C., P lotkin, P. Tl , a nd Paladino, F. V. 2000. Pacic leatherback turtles face ext inction. Nature, 405, 529–530. Stensholt, B. K. and Stensholt, E. 2004. Geographical variation in the v ertical distrib ution o f co d (Gad us mo rhua L.) a nd availability to survey gears. GIS/Spatial Analyses in Fishery and Aquatic Sciences, (109–126). Szedlmayer, S. T. and Schroepfer, R. L. 2005. Long-term residence of red snapper on articial reefs in the northeastern Gulf of Mexico. Transactions of the American Fisheries Society, 134, 315–325. Watwood, S. L., Miller, P. J. O., Johnson, M., and others. 2006. Deep-diving f oraging b ehaviour o f sp erm w hales (Physeter macrocephalus). Journal of Animal Ecology 75 , 814–825. Williams, T. M., Davis, R. W., Fuiman, L. A., and others. 2000. Sink or swim: strategies for cost-efficient diving by marine mammals. Science, 288, 133–136. Williams, T. M., Fuiman, L. A., Horning, M., and Davis, R. W. 2004. Ā e cost of foraging by a marine predator, the Weddell seal (Leptonychotes weddellii): Pricing by the stroke. J. Exp. Biol, 207, 973–982. Wolcott, T. G. 1996. N ew o ptions in p hysiological a nd b ehavioural e cology t hrough multichannel telemet ry. Journal of Experimental Marine Biology and Ecology, 193, 257–275. Wolcott, T. G. and Hines, A. H. 1996. Advances in ultrasonic biotelemetry for animal movement and behavior: Ā e blue crab case s tudy. I n: Methods a nd T echniques o f U nderwater Research, Pr oceedings o f t he American Academy o f Underwater S ciences S cientic D iving Symposium, October 12–13, 1996, Eds. M.A. Lang and C. C. Baldwin, Washington, DC: Smithsonian Institution, pp. 229–236. Yatsu, A., Yamanaka, K., a nd Yamashiro, C. 1999. Tracking exp eriments of the jumbo ying squid, Dosidicus gigas, with an ultrasonic telemetry system in the eastern Pacic Ocean. Bulletin of the Natural Resources Institute, Far Seas Fishery, 36, 55–60.

10/16/2008 7:15:42 PM

28 Molecular Imprinting Technology

David A. Spivak Louisiana State University

28.1

28.1 Introduction and Applications of Molecularly Imprinted Polymers ..............................28-1 28.2 Fabrication Methods of Molecular Imprinting Technology ............................................28-2 28.3 Molecular Imprinting Formats ............................................................................................28-3 28.4 Molecular Imprinting Templates ........................................................................................ 28-4 References .......................................................................................................................................... 28-4

Introduction and Applications of Molecularly Imprinted Polymers

Ā ere a re m any re views o n mole cular i mprinting te chnology that have chronicled the development of this eld over the last two de cades [ 1–4]. M olecular i mprinting te chnology c reates “smart m aterials” b y i ntegrating a tem plate i nto t he p olymer formulation during the fabrication of these polymeric materials (Scheme 2 8.1). I n t his w ay, mole cular le vel i nformation i s transferred from the template to the polymers in the form of a three-dimensional “memory” of the template’s shape and other features. Because molecularly imprinted polymers (MIPs) can be tailored to specically bind a targeted molecule or catalyze a de sired re action, t hese sm art m aterials h old t remendous promise for applications such as • • • • • •

Chemical and biological separations [5,6] Immunoassays [7,8] Catalysts [9] Sensors [10,11] Nanotechnology Biomedical, e.g., drug delivery [12–14]

Compounds c urrently ut ilized i n t hese a reas a re no nspecic polymers, b iological mole cules, a nd d esigned sm all mole cule receptors. MIPs are benecial when the applications listed above require specic binding for a target molecule that is not available

from nonspecic polymers such as silica gel, carbohydrates, or other p olymer m aterials. W hile b iological mole cules e xhibit good specicity toward a large range of molecules, MIP materials have improved properties such as • stability to o rganic s olvents, aque ous s olvents, a nd h igh temperatures [15] • ease, s peed, a nd lo w c ost o f p reparation i n a ny volume • tailorable b inding a ffi nities em ploying a ny f unctional group desired (including those not available to biological molecules) whereas antibodies and enzymes dena ture under va rious conditions suc h as hea t, o rganic s olvents, acids a nd bas es, et c. Furthermore, development of biological molecules is generally expensive and often co me from animal sources and cannot be stored. A third advantage is that MIPs can be stored for several years without loss of performance unlike biological molecules. Small mo lecule receptors can b e the r ugged alternative, however, the syn thesis o f thes e co mpounds is o ften l ong and arduous, w hereas mo lecular im printing ca n b e ca rried o ut in one day. MIPs also have potential to create enzyme-like catalysts, specically t ailored to c arry o ut a de sired re action w ith sub strate-selective control of stereochemistry a nd regiochemistry [9]. One of the main ways to accomplish this takes advantage of transition state t heory, outlined in Figure 28.1, 28-1

43722_C028.indd 1

10/3/2008 10:34:07 PM

28-2

Smart Materials

Solution phase prepolymer complex with +

Functional monomers

Remove template

Copolymerization with cross-linker

−

+ Rebind template

SCHEME 28.1 Outline of the molecular imprinting strategy.

Transition state analog

Transition state R O

O OH

R P O

O OH

R O O

−OH

R O HO

Starting materials

+

OH

Products Reaction coordinate

FIGURE 28.1 Energy diagram for the MIP catalyzed hydrolysis of a carboxylate ester.

the reaction coordinate diagram for hydrolysis of esters. If a MIP can be elicited to b ind t he transition state of a re action (bold f igure s hown u nder t he b ottom re action c oordinate, and around t he transition state analog), it lowers t he overall energy ba rrier a nd c atalyzes t he re action. H owever, si nce the t ransition s tate i s to o u nstable to u se a s a tem plate, a stable transition state analog is used for imprinting in place of the actual transition state. Ā e MIP formed to the transition state analog can be added to the solution of starting material, which w ill c atalyze t he o verall re action a s s hown i n F igure 28.1. MIPs that catalyze a reaction and control regiochemistry or stereochemistry, but do not have turnover (i.e., the reaction occurs once in the imprinted site) are referred to as microreactors [16].

43722_C028.indd 2

28.2 Fabrication Methods of Molecular Imprinting Technology Ā e modern strategy employed for molecular imprinting (shown in S chemes 2 8.1 a nd 2 8.2) i s much t he s ame a s i ntroduced b y Wulff i n 1 972, em ploying t wo t ypes o f mo nomers, f unctional monomers a nd c ross-linkers [ 17]. Ā e f unctional monome rs incorporate a moiety that interacts with a template molecule; the template c an b e a sm all mole cule o f i nterest (e.g., d rugs, h ormones, sugars, etc.) or larger molecules such as proteins or other macromolecules. Ā e functional monomer typically has a single polymerizable gr oup t hat d oes n ot c ross-link t he m aterial. However, cross-linking monomers that incorporate a functional group for i nteractions w ith t he template have a lso been h ighly successful. The primary role of the functional monomer is to provide tem plate–polymer i nteractions, re ferred to a s t he prepolymer complex, which provides rebinding interactions for the template molecule. The prepolymer complex can either be c ovalent or nonc ovalent i n n ature; ho wever, nonc ovalent interactions are the most convenient to use and are generally preferred. An example of this method is shown in the rst step of Scheme 28.2, where the functional monomer (methacrylic acid) forms a complex in solution with the template molecule (an adenine derivative). Ā e s econd t ype o f mo nomers a re c ross-linkers, w hich h ave two or more polymerizable groups, which form a r igid network polymer t hat s erves a s a s caffold t hat h olds t he tem plate-organized functional monomers in their positions. Some examples of cross-linkers used for molecular imprinting are shown in Figure 28.2. To form the imprinted polymer, the prepolymer complex is copolymerized w ith a n excess of c ross-linking monomer in t he presence of a n equal volume of inert solvent a nd a p olymerization-initiator. Ā ermal- or photochemical-initiated p olymerization results in a h ighly cross-linked polymer material, as shown in the second step of Scheme 28.2, with the template still inside.

10/3/2008 10:34:08 PM

28-3

Molecular Imprinting Technology

NH2 N

N

N

N

Prepolymer

+

complex

O O O OH NH2 H N N

HO N H O

O

O H N

Cross-linker O

N

O

H

O NH2 O O H N

N N

O H

Extract Rebind

O O O H

O

O

H

O

SCHEME 28.2 Molecular imprinting of a derivative of the DNA base adenine. (Adapted from Spivak, D., Gilmore, M.A., and Shea, K.J., J. Am. Chem. Soc., 119, 4388, 1997. With permission.)

as “batch-rebinding.” MIPs are used most extensively utilized in batch re binding mo de b y i ndustry f or s ample p reparation v ia O O solid-phase e xtraction o f f ood, b iological, o r en vironmental O O samples. Ā e h ighly c ross-linked nat ure o f t he M IP m aterials p-DVB EGDMA NOBE makes t hem p redisposed f or c hromatographic supp orts. Ā us, FIGURE 28.2 Examples of crosslinkers used for molecular imprinting. analytical separation of molecules by MIPs is achieved by packp-DVB, para-divinylbenzene; E GDMA, et hyleneglycol d imethacrylate; ing the MIP powder into a tube, which can be attached to liquid NOBE, N,O-bismethacryloyl ethanolamine. (From Sibr ian-Vazquez, M. (or gas) chromatography apparatus. However, chromatographic and Spivak, D.A., J. Am. Chem. Soc., 126, 7827, 2004; Sibrian-Vazquez, M. applications are generally more efficient if the MIP powder is in and Spivak, D.A., Macromolecules, 36, 5105, 2003. With permission.) a s pherical b ead f ormat, w ith a na rrow si ze r ange. To ac hieve this type of material, polymer beads can be formed from suspension, em ulsion, o r d ispersion p olymerization me thods [ 21]. Subsequent removal of t he template by extraction or hydrolysis Another method of forming polymer beads is the “seed-swelling” leaves t he  nal b inding si tes i n t he p olymer (step 3 o f S cheme technique, which uses a linear polystyrene latex particle as a seed 28.2). Ā ese sites are complementary in size, shape, and function- particle [ 22]. Ā e s eed pa rticles c an a bsorb t he M IP s olution, ality to the template, resembling the “lock and key” paradigm of which u pon p olymerization w ill f orm c ross-linked MI P enzymes. Rebinding of the template molecule to the molecularly beads w ith a na rrow si ze r ange. O ther t ypes o f c omposite imprinted sites, as shown in step 3 of Scheme 28.2, is favored for MIP beads can be formed, often starting with a core particle that the template molecule versus any other molecule in a mixture. serves as a support for another MIP coating [23]. An alternative method to grinding MI Ps a nd p acking in to c hromatographic columns is to p olymerize directly within the column; however, 28.3 Molecular Imprinting Formats the solvents needed to form t he needed porous monolith l imit MIPs are most frequently fabricated as polymer monoliths that the types of templates that can be used [24]. are ground into a powder made up of irregularly shaped particles MIPs can also be formed in  lm f ormats [25], h owever, t he (Scheme 28.3). Ā e powder is generally fractionated into a speci- traditional formulations for MIPs a re often too brittle to m ake ed si ze r ange, e .g., 25 –45 µm, f or v arious app lications. A s a them p ractical f or m aking  lms. Ā ere a re s ome re ports o f powder, MIPs can be added to a s olution mixture to absorb the composite polymer membranes that integrate the MIP polymer targeted molecule selectively from the mixture; this is referred to within the pores or as a c oating on a s table membrane support [26,27]. In some instances, nontraditional MIP formulations have been used to directly create the membrane material [28]. Surface-imprinted p olymer ph ases a re o f i nterest, pa rticularly for large molecules, such as proteins, t hat cannot be imprinted using traditional bulk polymerization formats [29]. Ā is is due to th e f act th at m olecules wi th m olecular w eights l arger th an 1000 g/mol will remain trapped in the MIP matrix and cannot be removed. Ā us, MIP binding sites formed on or near a surface can allow access to large molecules, thereby improving the mass transfer k inetics a nd u ltimate re cognition o f t he tem plate molecule. Both thin- lm formatted MIPs and surface-imprinted SCHEME 28.3 Initial polymer monolith in a t est tube and t he  nal materials a re often em ployed f or s ensor de velopment to ward powdered MIP product. large biomolecules [30]. O

O

43722_C028.indd 3

H N

O

10/3/2008 10:34:08 PM

28-4

28.4

Smart Materials

Molecular Imprinting Templates

Many diff erent t ypes of templates have been used for molecular imprinting; some examples of these are shown in Table 28.1. MIP recognition o f tem plates c an b e u sed to de tect o r s eparate t he target compound f rom ot her compounds, or even stereoisomers of the template. To create these selective sites, there must be at least one s trong b inding i nteraction b etween t he tem plate a nd t he monomers i n t he pre polymerization m ixture. For nonc ovalent imprinting, t hese in teractions ar e o ften i onic o r h ydrogen bonding interactions. Several other structural features have been evaluated for optimization of MIP performance such as number of i nteractive g roups o n t he tem plate, t he s patial d isposition o f groups, a nd s hape s electivity [31,32]. I t s hould b e note d t hat i n general, not a ll of the template can be removed from MIPs. Ā is leads to p roblems i n t race a nalyses by t hese m aterials, w hich i s solved by using a template that is not the same as (but similar to) the molecule targeted for analysis. Imprinting of large molecules, such as proteins, can be accomplished by directly imprinting the molecule u sing s pecial f ormulation s trategies (e.g., lo w le vels o f cross-linker) or novel formats (e.g., surface imprinting). Another method of imprinting large molecules is known as the “epitope approach,” where only a f ractional piece of the larger template is actually imprinted and recognition of the whole molecule is mediated t hrough t hat p iece [33]. A l ast rem ark o n M IPs i s t hat t he selective recognition of molecules is an average effect of a distribution o f b inding si tes t hat i s d ue to a n a rray o f v ariances i n t he binding site structure that arise during polymerization. Āer efore, the actual binding site generally behaves as an average ensemble of the different structures, rather than as a discrete structure that is often illustrated for exemplary purposes in the literature.

TABLE 28.1 Selected Examples of Templates Used in Various Molecular Imprinting Processes Compound Class Carbohydrates

Bioactive compounds

Amino acids and peptides

Proteins Coenzymes Nucleotide bases Metal ions

43722_C028.indd 4

Example Mannose Galactose, fucose Cellobiose, maltose Morphine Caffeine Nicotine Ā e ophylline Benzodiazepines Phenylalanine Tryptophan Oligopeptides Cytochrome c, BSA, lysozyme, ribonuclease Pyridoxal Adenine DNA Transition metals, lanthanides, group I and II metals

References [34] [35] [36] [37] [38] [39,40] [41] [42] [43] [44] [37,45–47] [48,49] [50,51] [18,52,53] [54] [55,56]

References 1 Alexander, C., Andersson, H.S., Andersson, L.I., Ansell, R.J., Kirsch, N., N icholls, I.A., O ’Mahony, J ., a nd Whitcombe, M.J., M olecular im printing s cience a nd t echnology: A survey o f the li terature f or the y ears u p t o a nd inc luding 2003. J. Molec. Recog., 19, 106, 2006. 2 Yan, M. a nd R amstrom, O . (E ds.), Molecularly I mprinted Materials: Science and Technology, Marcel Dekker, New York, 2005, 734pp. 3 K omiyama, M., T akeuchi, T ., a nd M ukawa, H. (E ds.), Molecular I mprinting: F rom F undamentals to Applications, Wiley-VCH, New York, 2003, 147pp. 4 Sellergren, B . (E d.), Molecularly I mprinted P olymers: Man Made M imics o f Antibodies a nd Ā ei r Application in Analytical Chemistry, Elsevier, Amsterdam, 2001, 550pp. 5 Sellergren, B. Imprinted p olymers: Stable, reusable a ntibodymimics for highly selective separations, Am. Lab., 14–20, 1997. 6 Pichon, V. a nd Haupt, K., Affinity s eparations o n mo lecularly im printed p olymers wi th sp ecial em phasis o n s olidphase extraction, J. Liquid Chromatogr. Related Technol., 29, 989, 2006. 7 Haupt, K., M ayes, A.G., a nd M osbach, K. H erbicide ass ay using an imprinted polymer based system analogous to competitive uoroimmunoassays, Anal. Chem., 70, 3936, 1998. 8 Vlatakis, G., Andersson, L.I., M uller, R ., a nd M osbach, K. Drug ass ay usin g a ntibody mimics made b y mo lecular imprinting, Nature, 361, 645, 1993. 9 Wulff, G. Enzyme-lik e ca talysis b y mo lecularly im printed polymers, Chem. Rev., 102, 1, 2002. 10 Piletsky, S.A., Turner, and Anthony, P. F. New materials based on imprinted polymers and their application in optical sensors, Opt. Biosensors, 397, 2002. 11 Kriz, D., Ramstrom, O., and Mosbach, K., Molecular imprinting: N ew p ossibilities f or s ensor t echnology, Anal. C hem., 345A, 1997. 12 Norell, M.C., Andersson, H.S., a nd N icholls, I.A. Towards molecularly im printed p olymer bas ed co ntrolled r elease materials, J. Mol. Recogn., 11, 98, 1998. 13 Allender, C.J ., B rain, K.R ., H eard, C.M., a nd P ellett, M.A. Imprinted polymers, novel excipients in controlled delivery systems, Proc. I nt. S ymp. C ontrolled Re lease B ioact. M ater. 24, 585, 1997. 14 Spivak, D.A., Optimization, evaluation, and characterization of molecularly imprinted polymers, Adv. Drug Delivery Rev., 57, 1779, 2005. 15 Sellergren, B. and Shea, K.J., Inuence of polymer morphology on the ability of imprinted network polymers to resolve enantiomers, J. Chromatogr., 635, 31, 1993. 16 Shea, K.J . M olecular im printing o f syn thetic netw ork polymers: the de novo synthesis of macromolecular binding and catalytic sites, Trends Polym. Sci., 2, 166, 1994. 17 Wulff, G. a nd Sa rhan, A. U se o f p olymers wi th enzymeanalogous structures for the r esolution of racemates., Ang. Chem., Int. Ed., 11, 341, 1972.

10/3/2008 10:34:09 PM

28-5

Molecular Imprinting Technology

18 Spivak, D ., G ilmore, M.A., a nd S hea, K.J ., E valuation o f binding a nd o rigins o f sp ecicity o f 9-eth yladenine imprinted polymers, J. Am. Chem. Soc., 119, 4388, 1997. 19 Sibrian-Vazquez, M. and Spivak, D.A., Molecular imprinting made easy, J. Am. Chem. Soc., 126, 7827, 2004. 20 Sibrian-Vazquez, M. and Spivak, D.A., Enhanced enantioselectivity of imprinted polymers formulated with novel crosslinking monomers, Macromolecules, 36, 5105, 2003. 21 Fairhurst, R.E., Chassaing, C., Venn, R.F., and Mayes, A.G., A direct comparison of the performance of ground, beaded and silica-grafted MIPs in HPLC and turbulent ow chromatography applications, Biosensors Bioelectronics, 20, 1098, 2004. 22 Zhang, L., Cheng, G., Fu, C., and Liu, X., Tyrosine imprinted polymer beads with different functional monomers via seed swelling and suspension polymerization, Polymer Eng. Sci., 43, 965, 2003. 23 Sulitzky, C., Rueckert, B., Hall, A.J., Lanza, F., Unger, K., and Sellergren, B ., G rafting o f mo lecularly im printed p olymer lms on silica supports containing surface-bound free radical initiators, Macromolecules, 35, 79, 2002. 24 Schweitz, L., Andersson, L.I., a nd N ilsson, S., R apid ele ctrochromatographic enantiomer separations on short molecularly imprinted polymer monoliths, Anal. Chim. Acta, 435, 43, 2001. 25 Titirici, M.-M. and Sellergren, B., Ā in molecularly imprinted polymer  lms via r eversible addition-fragmentation chain transfer polymerization, Chem. Mater., 18, 1773, 2006. 26 Piletsky, S.A., Panasyuk, T.L., Piletskaya, E.V., Nicholls, I.A., and Ul bricht, M ., R eceptor an d t ransport prop erties of imprinted p olymer mem branes—a r eview, J. Me mbr. S ci., 157, 263, 1999. 27 Takeda, K. and Kobayashi, T., Hybrid molecularly imprinted membranes f or t argeted b isphenol deri vatives, J. Me mbr. Sci., 275, 61, 2006. 28 Wang, H.Y ., X ia, S.L., S un, H., Li u, Y.K., C ao, S.K., a nd Kobayashi, T., Molecularly imprinted copolymer membranes functionalized b y phas e in version im printing f or uraci l recognition a nd p ermselective b inding, J. Ch romatogr., B , 804, 127, 2004. 29 Husson, S.M. and Gopireddy, D. Molecular engineering of surfaces f or s elective s eparations, Separation Sc i. Technol., 38, 2851, 2003. 30 Turner, N.W., Jeans, C.W., Brain, K.R., Allender, C.J., Hlady, V., and Britt, D.W., From 3D t o 2D: a review of the mo lecular imprinting of proteins, Biotechnol. Progr. 31 Spivak, David, A., Selectivity in molecularly imprinted polymers, Molecularly I mprinted M aterials, S cience and Technology, Yan, M. and Ramstrom, O. (Eds.), Marcel Dekker, New York, 2005, p. 395. 32 Spivak, D.A., Simon, R., and Campbell, J. Evidence for shape selective ca vity f ormation in mo lecularly im printed p olymers, Anal. Chim. Acta, 504, 23, 2004. 33 Minouraa, N., R achkov, A., H iguchi, M., a nd S himizu, T., Study of the fac tors in uencing peak asymmetr y on chromatography usin g a mo lecularly im printed p olymer p repared by the epitope approach, Bioseparation, 10, 399, 2002.

43722_C028.indd 5

34

35

36 37

38

39 40

41 42

43

44

45 46

47

48

Wulff, G ., Vesper, W., G robe-Einsler, R ., a nd S arhan, A., Enzyme-analog built polymers, 4. Ā e synthesis of polymers containing chiral cavities and their use for the resolution of racemates., Makromol. Chem., 178, 2799, 1977. Mayes A.G., Andersson L.I., and Mosbach K., Sugar binding polymers showing high anomeric and epimeric discrimination obtained by noncovalent mole cular imprinting, Anal. Biochem., 222, 483,1994. Striegler, S., I nvestigation o f dis accharide r ecognition b y molecularly imprinted polymers, Bioseparation, 10, 307, 2002. Andersson, L.I., Mueller, R., Vlatakis, G., and Mosbach, K., Mimics of the binding sites of opioid receptors obtained by molecular im printing o f en kephalin a nd mo rphine. Proc. Natl. Acad. Sci. USA, 92, 4788, 1995. Farrington, K., Magner, E., and Regan, F., Predicting the performance o f m olecularly i mprinted po lymers: S elective extraction of caffeine by molecularly imprinted solid phase extraction, Anal. Chim. Acta, 566, 60, 2006. Spivak, D.A. and Sibrian-Vazquez, M., D evelopment of an aspartic acid-bas ed cr oss-linking mo nomer f or im proved bioseparations, Bioseparation, 10, 331, 2002. Andersson, H.S., Karlsson, J.G., Piletsky, S.A., Koch-Schmidt, A.-C., Mosbach, K. Nicholls, and Ian, A. Study of the nature of r ecognition in mo lecularly im printed p olymers II. Inuence of monomer-template ratio and s ample load o n retention and selectivity, J. Chromatogr., A, 848, 39, 1999. Vlatakis, G., Andersson, L.I., Mueller, R ., and Mosbach, K. Drug ass ay usin g a ntibody mimics made b y mo lecular imprinting, Nature, 361, 645, 1993. Hart, B.R., Rush, D.J., and Shea, K.J., Discrimination between enantiomers of structurally related molecules: Separation of benzodiazepines by molecularly imprinted polymers, J. Am. Chem. Soc., 122, 460, 2000. Sellergren, B ., Ek berg, B ., a nd M osbach, K., M olecular imprinting of amino acid derivatives in macroporous polymers. D emonstration o f substra te- a nd ena ntio-selectivity by chr omatographic r esolution o f racemic mixt ures o f amino acid derivatives. J., Chromatogr., 347, 1, 1985. Kim, H. a nd G uiochon, G., Adsorption o n mo lecularly imprinted p olymers o f st ructural a nalogues o f a tem plate. Single-component adsorption isotherm data, Anal. Chem., 77, 6415, 2005. Titirici, M.M. and Sellergren, B., Peptide recognition via hierarchical imprinting, Anal. Bioanal. Chem., 378, 1913, 2004. Hart, B.R. and Shea, K.J., Synthetic peptide receptors: molecularly imprinted p olymers f or t he recognition of p eptides using p eptide–metal in teractions, J. Am. C hem. S oc., 123, 2072, 2001. Nicholls, I.A., R amstroem, O ., a nd M osbach, K., I nsights into the role of the hydrogen bond and hydrophobic effect on recognition in molecularly imprinted polymer synthetic peptide receptor mimics, J. Chromatogr., A, 691, 349, 1995. Nishino, H., Huang, C.-S., and Shea, K.J., S elective protein capture b y ep itope im printing, Angew. Chem ., Int. E d., 45, 2392, 2006.

10/3/2008 10:34:09 PM

28-6

49 Shi, H., Tsai, W.-B., Garrison, M.D., Ferrari, S., and Ratner, B.D., Template-imprinted nanostructured surfaces for protein recognition, Nature, 398, 593, 1999. 50 Andersson, L.I. and Mosbach, K. Molecular imprinting of the coenzyme–substrate analog N-pyridoxy-l-p henylalaninanilide, Makromol. Chem. Rapid Commun., 10(9), 491–495, 1989. 51 Svenson, J., Zheng, N., Nicholls, and Ian, A. A., Molecularly imprinted p olymer-based syn thetic t ransaminase, J. Am . Chem. Soci., 126(27), 8554–8560, 2004. 52 Umpleby, R.J., II, Rushton, G.T., Shah, R.N., Rampey, A.M., Bradshaw, J.C., Berch, J.K., Jr., and Shimizu, K.D., Recognition d irected s ite-selective c hemical m odication of molecularly im printed p olymers, Macromolecules, 34, 8446, 2001.

43722_C028.indd 6

Smart Materials

53 M atsui, J ., H igashi, M., a nd T akeuchi, T ., M olecularly imprinted p olymer as 9-eth yladenine r eceptor ha ving a porphyrin-based recognition center, J. Am. Chem. Soc., 122, 5218, 2000. 54 Spivak, D.A. and Shea, K. J., Investigation into the scope and limitations of mole cular imprinting with DNA mole cules, Anal. Chim. Acta, 435, 65, 2001. 55 Rao, T., Prasada, K.R., and Daniel, S., Metal ion-imprinted polymers—novel materials for selective recognition of inorganics, Anal. Chim. Acta, 578, 105, 2006. 56 Murray, G.M. a nd S outhard, E., M etal io n s elective molecularly imprinted materials. In Molecularly Imprinted Materials, Yan, M. a nd R amstrom, O . (E ds.), M arcel Dekker, New York, 2005, p. 579.

10/3/2008 10:34:09 PM

29 Biomedical Sensing

Dora Klara Makai Budapest University of Technology and Economics

Gabor Harsanyi Budapest University of Technology and Economics

29.1

29.1 I ntroduction ...........................................................................................................................29-1 29.2 Operation of Biosensors .......................................................................................................29-2 29.3 Smart Polymers in Biomedical Sensing ..............................................................................29-2 29.4 Enzyma tic Biosensors ...........................................................................................................29-3 29.5 D NA Sensors ..........................................................................................................................29-3 29.6 I mmunosensors .....................................................................................................................29-3 29.7 Alternatives for Natural Receptors ..................................................................................... 29-4 29.8 Dr ug Delivery........................................................................................................................ 29-4 29.9 F uture Prospects ....................................................................................................................29-5 References ...........................................................................................................................................29-5

Introduction

Biomedical sensing is a typical example of interdisciplinary elds that poses great challenges and desires the cooperation of experts of n umerous s cientic a reas suc h a s me dicine, b ioengineering, chemistry, e lectrical e ngineering, m aterial s ciences, n anotechnology i n developing biosensors a nd ot her sensitive biomedical constructions such as drug delivery systems. Ā e development in the  eld o f m icro- a nd na notechnology a nd m aterial s cience enabled t he p reparation o f m iniaturized b iosensors t hat a re applicable in biomedical diagnostics and monitoring. Biosensors are special sensors using a biologically active material f or me asurement o r i ndication. Ā ey c onsist o f t wo m ain parts, the receptor and the transducer (this is not a physical separation). Ā e receptor interacts with the analyte selectively; the transducer produces a signal (electrical or optical) as a result of the former interaction. Ā is signal carries information about the concentration of the analyte if certain conditions are ensured. Receptor pa rts c ontain b iologically ac tive c omponents t hat are c apable f or s pecic c hemical re actions w ith t he a nalyte. Upon t he nature of receptor t ypes, biosensors can be classied into the following four main groups: • Enzymatic biosensors e xploit t he specic re action b etween enzyme a nd i ts sub strate. E nzymes a re b iocatalysts. Ā ey enable c ertain b iochemical re actions to p roceed t hrough their catalytic activity. Immobilizing the enzyme as receptor, its substrate as analyte can be detected.

• Ā e operation of immunosensors is based on the formation of a specic binding between antibody and antigen. • DNA sensors exploit the interactions between the complementary DNA sequences. • Living biosensors use microorganisms or living tissues as receptors for selective sensing. Biosensors c an e mploy diff erent t ransducers l ike c alorimetric types, el ectrochemical t ypes ( potentiometric, a mperometric, o r conductometric e lectrochemical c ells, as w ell as m odied Ion– sensitive eld effect transistor (ISFETs) ber optic transducers, or gravimetric type resonator transducers. Latest de velopments i n t he a rea o f b iomedical s ensing a re stimulated by t he re cent progresses i n smart materials de signing. Smart materials are mostly polymers that respond to external s timuli w ith c onsiderable p roperty c hanges. E ven sm all variations in the environmental pH, temperature, ionic strength, electric or m agnetic eld c an alter their structure, solubility, or can lead a sol–gel transition. Ā ese phenomena can be exploited in different internal or external biomedical applications [1]. Biomedical appl ications c over not on ly re search l aboratory applications or ho spital i nstruments but home d iagnostic k its commercially a vailable f or c ommon p eople a nd app licable i n everyday l ife suc h a s g lucose s ensors for d iabetic pat ients. Ā e widespread availability of biosensors is preceded by long optimization processes. Mostly complicated problems have to be solved during a s ensor development such a s choosing t he appropriate receptors, transducer types, immobilization techniques, preventing

29-1

43722_C029.indd 1

10/3/2008 9:07:40 PM

29-2

Smart Materials

receptors f rom p oisoning d uring o peration, as suring a ppropriate stability for t he receptors, compensating interferences a nd disturbing e ffects, a nd de veloping f abrication, s torage, p reoperation treatment, calibration, and data-processing techniques that enable t he w ide r ange m ass app lication o f low-cost, s hort-lifetime sensors in practice [2]. It has to b e mentioned that the global biosensor market is presently dominated by medical applications but besides t he glucose sensors no b iosensor has ye t achieved t he t rue mass market sales [3]. After the glucose sensors, DNA chips represent the l argest s ector o f to day’s b iosensor m arket [3]. A lthough promising, DNA sensor technologies still need to br idge t he gap b etween e xperimental s tatus a nd t he re ality o f c linical and diagnostic applications [4].

enzymatic reaction of glucose and glucose oxidase. Ā e substrate specicity of enzymes or the specic interactions between complementary D NA s equences o r b etween a ntigen a nd a ntibody alone do not ensure the specicity of the analytical method as an analytical method is considered specic i f it i s p erfectly s elective. I n t he c ase o f ele ctrochemical b iosensors, i nterference i s usually hindered by multilayered membrane structures. Considering th is am ount o f th e aff ecting pa rameters, i t i s clear t hat opt imizing a b iosensing s ystem i s not a si mple t ask. Simulation o f t he i mpacts o f t hese d ifferent pa rameters b y computer is helpful and effective method for the optimization.

29.2 Operation of Biosensors

Since biomolecules are highly sensitive structures, their immobilization requires special circumstances i n order to m aintain their functionality and stability. Ā e development of the material s ciences h as o ffered t he p ossibility o f c hoosing a mong a wide range of ideal matrix materials for biosensor fabrication. Besides providing a biocompatible environment to the receptor molecules, the matrix should ensure the contact with the transducer i t s hould p rovide to t he t arget mole cules to ge t to t he receptors. Ā e use of advanced materials can improve the performance o f b iosensors v ia i mproving t he s tability a nd t he activity of biomolecules or ensuring the operation of biosensors in e xtreme c onditions. Suc h sm art m aterials a re h ydrogels, sol–gel der ived m aterials, s ol–gel/hydrogel h ybrids, a nd l ipid membranes. Hydrogels a re t hree-dimensional p olymer ne tworks t hat swollen w ith w ater c an p rovide a b iocompatible en vironment for biomolecules. Fernandez a nd e t a l. entrapped g lucose oxidase in polyacrylamide hydrogels by aqueous cross-linking copolymerization of acrylamide a nd N,N¢-methylene bi sacrylamide, which can be incorporated into an amperometric glucose biosensor [9]. Changes in structure of smart materials in response to different external stimuli can cause changes in functional properties of t he material. Designing such smart polymeric systems has a great importance and a promising future from the perspective of biomedical application. Zhang et al. introduced a d -glucose-sensitive hydrogel based on cross-linking carboximethyl-dextrane (CMD) w ith t he g lucose-binding lectin concanavalinA (ConA) [10]. Between ConA and ter minal g lucose mo ieties o f de xtrane, b esides c ovalent bindings, a ffinity i nteractions t ake p lace a s w ell. I ncreasing external g lucose c oncentration p rovokes c ompetitive d isplacement of a ffinity interactions inducing this way changes in morphology a nd p ermeability. Ā e p ermeability o f t he h ydrogel increases with increasing external d-glucose concentration. Kim et al. studied the behavior of poly(acrylic acid)/poly(vinyl sulfonic acid) (PAAc/PVSA) copolymer hydrogel i n a n electric eld [11]. Ā e polymer showed reversible volume change in constant electric  eld indicating that this material is a good candidate for biosensor preparation.

In the background of the apparently simple principles of biosensor op eration, a n i mmense c omplexity i s h idden, n umerous parameters should be considered in course of preparation and operation of the biosensor that are summarized below by introducing en zymatic biosensors prepared by embedding en zymes into conducting polymers. In order to prepare a biosensor, biologically active molecules should be immobilized onto a substrate in a way that they could keep their stability and stay operable. Ā ere are different immobilization methods ranging from a “simple” adsorption or covalent b inding to emb edding i nto p olymer m atrices, a ll o f t hem have their advantages and drawbacks. Electrochemical polymerization is a simple, reproducible, low cost, r apid me thod to de posit c onducting p olymer m atrices on electrode surfaces [5,6]. Ā e embedding of the enzymes in a polymer layer can be achieved in one-step procedure conducting the electrodeposition from a solution containing the monomers and the enzymes as well. Ā e thickness of the deposited  lm is controlled by the charge passed during the electropolymerization. Ā e s ensor re sponse de pends o n t he p roperties o f t he ele ctrodeposited po lymer, wh ich a re a ffected b y t he c onditions o f the deposition such a s t he me thod of deposition (cyclovoltammetric, ga lvanostatic, p otentiostatic d eposition me thods), t he enzyme and monomer concentration in the solution, the pH of the s olution, ot her c omponents i n t he s olution, s peed o f t he deposition, a nd c urrent den sity. Ā e amount of enzyme in the solution has a g reat inuence on t he forming process, permeability, and morphology of the  lm. High enzyme concentration inhibits the nucleation process in the early stage of the deposition, results in a dense, less permeable, less porous lm [7]. High monomer concentration encumbers the enzyme incorporation, which leads to lower sensor response [8]. It h as to b e c onsidered t hat t he op eration o f a b iosensor i s inuenced by the sample itself as well. Other components in the sample to be analyzed should not be ignored. For example, blood contains some components that during the amperometric determination o f blo od g lucose u ndergo ele ctrode re action at the applied potential, thus the signal derives not only from the

43722_C029.indd 2

29.3

Smart Polymers in Biomedical Sensing

10/3/2008 9:07:41 PM

29-3

Biomedical Sensing

Besides h ydrogels, s ol–gel-derived m aterials a nd s ol–gel/ hydrogel, hybrid materials provide a biocompatible environment for biomolecules. Tan et al. described an amperometric cholesterol b iosensor u sing s ol–gel c hitosan/silica h ybrid c omposite  lm [12]. A ccording to t he a uthors, t he h igh s ensitivity, go od repeatability, reproducibility and selectivity, the rapid response, and lo ng-term s tability c an b e pa rtly at tributed to t he h ybrid composite lm.

29.4 Enzymatic Biosensors Enzymatic b iosensors a re w idely app lied i n t he me dical  eld. Different analytes in different biological samples (blood, urine, saliva e tc.) c an b e de tected a nd me asured v ia en zymatic re actions in vitro and in vivo as well. Enzymatic biosensors can also be used by production of pharmaceutical agents in bioreactors for controlling reactions. As men tioned a bove, a mong en zymatic b iosensors, g lucose sensors a re t he mo st w idely u sed. A t le ast 1 71 m illion p eople suffer from diabetes worldwide and this  gure may increase in the following decades [13]. One of the most important approaches of d iabetes t reatment i s t he s elf-monitoring o f blo od g lucose (SMBG), which means regularly checking blood g lucose levels. In this way, hypo- and hyperglycemia can be controlled in order to reduce or prevent the diabetes complications. Ā e S MBG i s recommended f or a ll d iabetic p eople e specially f or t hose w ho need to take insulin regularly [14]. Retama e t a l. p resented a n a mperometric g lucose s ensor using polypyrrole–microgel composites [15]. For immobilizing glucose oxidase, a new immobilization matrix, water-dispersed complex o f p olypyrrole–polystyrensulfonate ( Ppy) emb edded in polyacrylamide (PA) microgel has been prepared. The polyacrylamide–polypyrrole (PAPPy) microparticles combine the c onductivity o f p olypyrrole a nd t he p ore si ze c ontrol o f polyacrylamide. Besides g lucose, t here a re ot her me tabolites t he h igher o r lower le vels o f w hich i ndicate i llnesses. l- Lactate i s t he  nal product of anerobic glycolysis. Elevated l- lactate level in blood can b e a m arker of d ifferent d iseases suc h a s hypoxia or l actic acidosis [16]. A s lactate le vel i ndicates t he oxygenation state of tissues continuous monitoring of lactate level is critical during operations, intensive therapy, in critical care and sport medicine as well. For the amperometric detection of lactate, mainly NADdependent lactate dehydrogenase or lactate oxidase is used but l-l actate–cycochrome c oxidoreductase based lactate sensor has also been reported [16]. Schuvailo et al. introduced highly selective microbiosensors for glucose, lactate, and glutamate sensing that can be used in in vitro measurements [17]. Ā ere a re c linically i mportant mole cules t hat c annot b e detected amperometrically using only one type enzyme but with consecutive enzymatic reactions. Ā e more enzyme type immobilized the more difficult the sensor preparation. Cholesterol and its esters are precursors of different biomolecules, e .g., h ormones. M onitoring o f c holesterol le vel ga ins importance i n d iagnosis a nd p revention o f h eart d iseases a nd

43722_C029.indd 3

arteriosclerosis. S alinas e t a l. i ntroduced a m ultienzymatic biosensor ba sed o n t he c onsecutive en zymatic re actions c atalyzed by cholesterol esterase, cholesterol oxidase, and peroxidase for total cholesterol determination [18]. In order to eliminate the interference of l -ascorbic acid a micropacked column containing ascorbate oxidase is integrated as prereactor. Multianalyte biosensors are able to detect and measure more analyte si multaneously. Multienzymatic a mperometric s ensors have more working electrodes, each functionalized with different enzymes for detection of different substrates. By fabrication of multianalyte biosensors, t he c ross-talk phenomenon should be considered. Ā e molecules generated by the enzymatic reaction undergo electrode reaction in course of the sensing process. Ā ese product molecules gain the electrode surface by diffusion. Cross-talk derived from the diff usion can lead to false results. In order to p revent ge tting f alse re sults b ecause of t he c ross-talk, the s ensor c onstruction ( interelectrode d istance, ele ctrode geometry, etc.) should be rightly chosen. Simulation of processes by means of computer can help to  nd t he mo st ade quate construction.

29.5 DNA Sensors Ā e op eration o f D NA s ensors i s ba sed o n t he i nteraction between c omplementary D NA s equences. Ā e h ybridization between the immobilized ssDNA strands and the DNA strands in the sample to be analyzed can be detected by various methods. Similar to enzymatic sensors, electrochemical transduction is an attractive s olution due to i ts h igh s ensitivity a nd t he rel atively inexpensive instruments. By the immobilization of DNA sequences, the strong interaction between streptavidin and biotin is often exploited. Riccardi et al. presented an amperometric DNA biosensor for hepatitis C virus de tection [ 19]. Ā e biot inylated ol igonucleotide prob es have been attached to the streptavidin entrapped into siloxane– poly(propylene o xide) s ol–gel  lms de posited o nto a g raphite electrode surface. Okumura et al. also used the streptavidin–biotin interaction in the immobilization process [20]. Ā ey used hydrogel nanospheres a s en hancers i n su rface p lasmon re sonance ( SPR) imaging s tudies f or h ighly s ensitive de tection o f K-ras p oint mutation.

29.6 Immunosensors Ā e specic reaction between antigens and antibodies is exploited in t he op eration of i mmunosensors. A ntigens a re foreign sub stances t hat i nvade t he b ody. E ntering a s u nwanted i nvaders, foreign bioagents i n l iving t hings ac tivate t he protecting f unctions of the immune system. Antibody production starts against the a ntigens i n o rder to s pecically b ind t hem as  rst s tep o f their elimination. Each antibody bears a s pecial section, which ensures a selective tting with a specic antigen. Antigens ca n dir ectly b e d etected b y imm obilizing s pecic antibodies on the sensor surface. Ā is st rategy s uffers f rom a

10/3/2008 9:07:41 PM

29-4

Smart Materials

serious drawback namely the usually low concentration of antigens, which can lead to false-negative output. Sensitivity can be improved v ia i ndirect de tection, i .e., t he de tection of t he a ntibodies (instead of antigens) that are produced in huge amounts. Pohanka e t a l. presented a no vel me thod for i ndirect de tection o f Francisella t ularensis (causative agent of tularemia, a highly infectious pathogen bacterium, potential biological warfare agent) [21]. Ā e piezoelectric immunosensor is based on the measurement of antitularemic antibodies with covalently immobilized antigens in infected mouse serum. Petrosova et al. reported an optical immunosensor for the detection o f a ntibodies to E bola v irus s trains Sud an a nd Zaire in animal and human sera [22]. The inactivated Ebola virus receptors have been photoimmobilized to the photoactivable e lectrogenerated p olymer f ilm mo dified f iber opt ic surface. Synthetic peptides that mimic the special immunoreactive sections provide a promising alternative to immobilized antigens. Vaisocherova e t a l. de veloped a su rface p lasmon re sonance biosensor f or l abel-free de tection o f a ntibody a gainst Eps tein– Barr virus using synthetic peptide as a receptor [23]. Antibodies can b e p roduced not o nly a gainst bac teria, v iruses, o r ot her microorganisms b ut a gainst “ simple” mole cules suc h a s h istamin. Li et al. introduced a regenerable surface plasmon resonance immunosensor for detection of antihistamine antibody [24].

29.7

Alternatives for Natural Receptors

Since biomolecules are highly sensitive materials, it is of crucial importance to handle them carefully, which practically makes difficult to work with them in practice. Replacing these molecules by alternative receptors can make biosensor preparation easier. Molecularly i mprinted p olymers ( MIPs) a re p olymers w ith articial re cognition si tes ( receptors) c omplementary to t he target mole cules, w hich c an b e p roteins, D NA s equences, enzymes, a ntibodies b ut sm all mole cules a s w ell. D uring t he process of mole cular i mprinting, f unctional monome rs a nd cross-linkers form a polymer network around the templates that are el iminated at t he en d o f t he p rocedure. M IPs re present attractive alternatives to expensive, relatively unstable bioreceptores. Ā ey a re e asy to p repare w ith h igh re producibility a nd their ma ss p roduction is cost-effective. Ā ey c an w ithstand to more e xtreme c onditions t han t he nat ural re ceptors a nd s how high a ffinity and specicity. Fu rther adv antages o f M IPs o ver the natural receptors are that they are more compatible with the micromachining techniques and make possible creating receptors for important analytes for which no natural receptors exits. Besides these advantages MIPs usually suffer from some disadvantages such as template leakage, not e asily reachable binding sites or nonspecic bindings. Ā ere is still no general procedure for MIP preparation a nd t here a re still d ifficulties to de al w ith such as the integration of transducers or the transformation of binding event to signal [25].

43722_C029.indd 4

Numerous MIPs have been created for different clinically important analytes but mainly for small molecules. For example, Cheng et a l. introduced a M IP for glucose, which can be f unctionalized as a part of a capacitive glucose sensor [26]. Unlike small molecules, protein imprinting is still proved to be a c hallenging task mainly since t heir 3D structure is highly sensitive to t he en vironment, tem perature, io nic s trength, pH, etc. Ā ere exist more methods for imprinting proteins. Āe grafting method involves the formation of a t hin polymer  lm on a solid sub strate a round t he p rotein. A nother me thod, c alled epitope approach, is based on that only a little part of the protein, a s hort p eptide s equence i s u sed a s tem plate i n t he M IP preparation. Template molecules can be immobilized on a solid sacricial substrate as well. Once the polymerization is over, the sacricial support can be eliminated by different chemical treatments. Ā is method enables preparing MIPs with templates that are insoluble in the polymerization mix. Peptide n ucleic ac ids ( PNAs) a re ol igonucleotide a nalogs having u ncharged ps eudeopeptide bac kbone i nstead o f su gar– phosphate backbone [27]. PNAs can hybridize with complementary DNA, RNA, or PNA strands. Ā e PNA–DNA duplexes show higher thermal stability than the respective DNA–DNA duplexes [28]. Ā e stability of PNA–DNA complexes, unlike that of DNA double helices, is barely affected b y t he io nic s trength o f t he medium. PNAs show great resistance against enzymatic degradation [28]. Brandt et al. introduced PNA microarrays combined with time-of-ight secondary mass spectrometry (ToF-SIMS) as a sensitive method for label-free DNA diagnostic on a chip [29].

29.8 Drug Delivery Drug delivery by smart materials can also be considered as biomedical sensing process although the way they are working is markedly diff erent f rom t he a forementioned b iosensors. Ā e use of smart materials (such as nanosystems, microgels, MIPs) in the eld of drug delivery has a signicant impact on the efficiency of treatment improving the clinical and the commercial value of the products. Ā e increased efficiency of smart materials in drug delivery is mainly due to the targeted drug delivery and the i ntelligent d rug rele ase. T argeting a nd i ntelligent, c ontrolled rele ase o f d rugs en sures t hat t he d rug c an re ach t he desired site of the body and the desired way (sustained or pulsative). Drug targeting reduces the side-effects of medications as the drug is carried to the desired site of action, minimizing this way the exposure of other tissues and organs to t he drug. Active t argeting c an b e ac hieved b y u sing c arriers b earing a motif that can bind only with receptors, which are specic for the target cells. Intelligent drug release can be realized by drug carriers, which respond in a predictable way to specic stimuli (e.g., changes of pH). Molecularly imprinted polymers can function as drug carriers as they can be prepared to able to selectively bind to the target motifs. After binding to the target, they can release their payload. Nanosystems are systems containing submicron-sized dispersed

10/3/2008 9:07:41 PM

Biomedical Sensing

particles a s na noparticles, na nocapsules, p olymeric s elf-assemblies, or dendrimers (nanometer-sized synthetic polymer macromolecules) [30]. Ā e major advantages of na noparticles i n d rug delivery derive from the size and their nature. Ā eir size makes possible to penetrate small capillaries and to be taken up by cells while their biodegradable nature ensures a sustained drug release over a longer period after administration [31]. By the aid of nanosystems, water solubility and stability of drugs can be improved [30]. Drug del ivery by na noparticles plays a n i mportant role i n cancer therapy, gene therapy, AIDS therapy, delivery of protein, antibiotics, and vaccines. Ā e surface of nanoparticles can be modied in order to further improve the efficiency of drug delivery. Surface modication gain importance, for example, in intravenous administered drugs as it can reduce the opsonization (interaction with special blood proteins) of nanoparticles and increase the circulation tim e (stealth n anoparticles). Ā ese effect w ere o bserved with poly(ethylene-glycol)-modied nanoparticles [31]. Delivering of insulin is a highly investigated eld in the drug delivery searching. In case of insulin as most of the other drugs, the o ral ro ute o f ad ministration i s t he mo st de sirable w ay o f administration due to its high patient compliance. As insulin is a p rotein, i ts o ral ad ministration p oses g reat c hallenges t hat have to be overcome by developing intelligent drug carrier systems. Oral administration of proteins is not efficient as the majority of the molecules suffer degradation from the enzymes of gastrointestinal tract before reaching the blood stream. Peppas et a l. de veloped a h ydrogen-bonding c opolymer ne twork o f poly(methacrylic aci d) g rafted wi th p oly(ethylene-glycol) f or oral delivery of peptides and proteins [32]. In the acidic environment of the stomach, the proteins remain entrapped in the delivery system unreachable for the enzymes, safe from degradation. Ā e proteins can reach this way the intestine where they should be absorbed and can be released from this delivery “vehicle.” Ā e mechanism behind this protection and release phenomenon can be explained by the changes in mesh size of the hydrogel.

29.9

Future Prospects

Ā e development of biomedical sensing systems is still on t he rise. Advances in material sciences and in nanotechnology can be e xploited i n b iosensor p reparation i n o rder to f abricating more a nd more sophisticated biosensing s ystems u sing smart materials. E xperts still face challenge of combining t he existing and the newly developed materials and methods and optimizing t he b iosensing s ystems to ge t b iosensors t hat c an b e widely used.

References 1. Jeong, B. and Gutowska, A., L essons f rom nature: Stimuliresponsive p olymers and t heir b iomedical a pplications, Trends Biotechnol., 20, 305, 2002. 2. Harsanyi, G., Sensors in Biomedical Applications, Technomic Publishing Company, Lancaster, Basel, 2000, 226.

43722_C029.indd 5

29-5

3. Bogue, R ., D evelopments in b iosensors—where a re tomorrow’s markets? Sens. Rev., 25, 180, 2005. 4. Hahn, S. et al., Nucleic acid based biosensors: Ā e desires of the user, Bioelectrochemistry, 67, 151, 2005. 5. Cosnier, S., B iomolecule immob ilization o n ele ctrode surfaces by entrapment or attachment to electrochemically polymerized lms, Biosens. Bioelectron., 14, 443, 1999. 6. Ā anachasai, S., Yoshida, S., and Watanabe, T., Effects of fabrication pa rameters o n the enzyme loadin g a nd s ensor response of enzyme-carrying conductive polymer electrodes, Anal. Sci., 19, 665, 2003. 7. Shin, M.C. and Kim, H.S., Electrochemical characterization of polypyrrole/glucose oxidase biosensor: Part I. Inuence of enzyme c oncentration on t he g rowth and prop erties of the lm, Biosens. Bioelectron., 11, 161, 1996. 8. Shin, M.C. and Kim, H.S., Electrochemical characterization of polypyrrole/glucose oxidase biosensor: Part II. Optimal preparation co nditions f or the b iosensor, Biosens. Bioelectron., 11, 171, 1996. 9. Fernandez, E. et al ., Viscoelastic and swelling properties of glucose o xidase lo aded p olyacrylamide h ydrogels a nd t he evaluation of t heir prop erties a s g lucose s ensors, Polymer, 46, 2211, 2005. 10. Zhang, R. et al., Synthesis and characterization of a d-glucose sensitive hydrogel based on CM-dextran and concanavalin A, React. Funct. Polym., 66, 757, 2006. 11. Kim, H. I. et al., Electroactive polymer hydrogels composed of polyacrylic acid and poly(vinyl sulfonic acid) copolymer for application of biomaterial, Synth. Met., 155, 674, 2005. 12. Tan, X. et al., An amperometric cholesterol biosensor based on multiwalled carbon nanotubes and organically modied sol-gel/chitosan hybrid composite lm, Anal. Biochem., 337, 111, 2005. 13. http://www.who.int/diabetes 14. http://www.fda.gov/diabetes/glucose.html#2 15. Retama, J.R. et al., Design of an amperometric biosensor using polypyrrole–microgel composites containing glucose oxidase, Biosens. Bioelectron., 20, 1111, 2004. 16. Smutok, O. et al., A novel l-lac tate-selective biosensor based on  avocytochrome b2 f rom meth ylotrophic y east Hansenula po lymorpha, Biosens. B ioelectron., 20, 1285, 2005. 17. Schuvailo, O.M. et al., Highly selective microbiosensors for in vivo measurement of glucose, lactate and glutamate, Anal. Chim. Acta, 573, 110, 2006. 18. Salinas, E. et al., Multienzymatic-rotating biosensor for total cholesterol determination in an FIA system, Talanta, 70, 244, 2006. 19. Dos Santos Riccardi, C. et al ., Immobilization of streptavidin in sol–gel lms: Application on the diagnosis of hepatitis C virus, Talanta, 70, 637, 2006. 20. Okumura, A. et al., Point mutation detection with the sandwich method employing hydrogel nanospheres by the sur face plasmon resonance imaging technique, Anal. Biochem., 339, 328, 2005.

10/3/2008 9:07:41 PM

29-6

21. Pohanka, M., P avlis, O ., a nd S kladal, P., Dia gnosis o f tularemia usin g p iezoelectric b iosensor t echnology, Talanta, 71, 981, 2007. 22. Petrosova, A. et al., Development of a highly sensitive, eld operable biosensor for serological studies of Ebola virus in central Africa, Sens. Actuators, B, 122, 578, 2007. 23. Vaisocherova, H. et al., Surface plasmon resonance biosensor f or dir ect det ection o f a ntibody a gainst E pstein-Barr virus, Biosens. Bioelectron, 22, 1020, 2007. 24. Li, Y. et al ., Surface plasmon resonance immunosensor for histamine based on an indirect competitive immunoreaction, Anal. Chim. Acta, 576, 77, 2006. 25. Piletsky, S.A., T urner, N.W ., a nd L aitenberger, P ., Molecularly imprinted polymers in clinical diag nostics— Future potential and existing problems, Med. Eng. Phys, 28, 971, 2006. 26. Cheng, Z., Wang, E., and Yang, X., C apacitive detection of glucose usin g mole cularly im printed p olymers, Biosens. Bioelectron., 16, 179, 2001.

43722_C029.indd 6

Smart Materials

27. Nielsen, P.E. et al ., Sequence-selective recognition of DNA by st rand displ acement w ith a t hymine-substituted polyamide, Science, 254, 1497, 1991. 28. Egholm, M. et al ., PN A h ybridizes t o co mplementary oligonucleotides ob eying the Watson–Crick h ydrogenbonding rules, Nature, 365, 566, 1993. 29. Brandt, O. et al., Development towards label- and amplication-free genotyping of genomic DNA, Appl. S urf. S ci., 252, 6935, 2006. 30. Kidane, A. a nd B hatt, P .P., Re cent ad vances in small molecule dr ug delivery, Curr. O pin. C hem. B iol., 9, 347, 2005. 31. Xu, T., et al ., Modication of nanostructured materials for biomedical applications, Mater. Sci. Eng., C, 27, 579, 2007. 32. Peppas, N.A. and Kavimandan, N.J., Nanoscale analysis of protein an d p eptide ab sorption: I nsulin ab sorption u sing complexation a nd pH-s ensitive h ydrogels as deli very vehicles, Eur. J. Pharm. Sci. (in press); corrected proof available online May 10, 2006.

10/3/2008 9:07:41 PM

30 Intelligent Chemical Indicators 30.1 I ntroduction ...........................................................................................................................30-1 30.2 Chemical Indicating Devices Are Smart.............................................................................30-1 Classication • G eneral Operating Principles • C hoice of Indicators • p H Indicators • T emperature and Time–Temperature Indicators • A nticounterfeiting and Tamper Indicator Devices • Indicating Device Issues and Limitations

Christopher O. Oriakhi Hewlett-Packard Company

30.3 C onclusion ..............................................................................................................................30-9 Acknowledgments .............................................................................................................................30-9 References ...........................................................................................................................................30-9

30.1 Introduction Most p eople rememb er a c hemical i ndicator f rom t heir h igh school ch emistry. Ā is k ind o f i ndicator i s a m aterial w hich changes color to signify the endpoint of a titration reaction or to provide a rel ative i ndication o f t he ac idity o r a lkalinity o f a chemical subst ance. Ā e u se o f i ndicators e xtends f ar b eyond this. F or e xample, f ood, c osmetic, ph armaceutical, a nd ot her chemical f ormulations u ndergo c omplex c hemical, en zymatic, and microbial interactions when they are exposed to UV light or temperature uctuations over time. Consequently, product quality may b e de graded, a nd m ay le ad to add itional s afety c oncerns. Ā e c hallenges f acing t he p roduce i ndustry i nclude suc cessful implementation of good manufacturing practices (GMP), hazard analysis and critical control point (HACCP), total quality management (TQ M) p rograms, a nd ot her re gulations dem anding compliance [1]. To a ddress t hese iss ues a nd i ncrease consumers’ condence on product quality, safety, and authenticity, many manufacturers i ncorporate i nexpensive i ndicator mo nitoring de vices i nto their products during production, packaging, or storage. A large number of consumer-readable indicators are available commercially. Some examples of these types of indicators are tags, labels, seals, a nd t hermometers. S ome a re de signed to g ive a v isual color change in response to degradation of product quality, tampering, or to de tect a co unterfeit. Ā ey a re u sed e xtensively i n the c hemical, f ood, a nd ph armaceutical i ndustries w here c onsumers need assurance of product integrity, quality, and safety during postmanufacture handling. Generally, ch emical i ndicators ma ybe de ned a s s timulus responsive materials that can provide useful information about changes in their environment. Organic dyes, hydrogel or “smart polymers,” shape memory alloys, thermochromic or photochromic

inks, and liquid crystals are some examples. Ā ey may function by f orming s tructurally a ltered io nic o r mole cular c omplexes with species in their environment through chemical or physical interactions in volving p roton e xchange, c helation, h ydrogen bonding, dipole–dipole interactions or van der Waal forces [1]. Ā e resulting characteristic biochemical, chemical, optical, magnetic, thermal, or mechanical changes can be tailored to provide the desired indication response. Ā is chapter will focus on inexpensive disposal chemical indicating d evices s uch as p H in dicators, t emperature in dicators, time–temperature i ndicators (T TI), t ampering a nd c ounterfeit indicators. Ā e tem perature a nd T TIs a re w idely u sed i n t he food and pharmaceutical products where date coding on a package may sometimes be inadequate.

30.2

Chemical Indicating Devices Are Smart

Smart m aterials o r de vices m aybe de  ned a s m aterials t hat produce strong visually perceptible changes in physical or chemical property in response to small physical or chemical stimuli in the medium. Ā e m aterial p roperties me asured m ay i nclude p H, concentration, composition, solubility, humidity, pressure, temperature, l ight i ntensity, ele ctric a nd m agnetic  eld, sha pe, a ir velocity, h eat c apacity, t hermal c onductivity, mel ting p oint, o r reaction r ates [ 2–6]. C hemical-indicating de vices c an re spond reversibly o r i rreversibly to sma ll c hanges i n t he p hysical o r chemical properties in their environment in a predictable manner. Ā ey may be regarded as smart materials because of the range of material properties they encompass. Typical materials include shape memo ry a lloys, p iezoelectric m aterials, m agnetostrictive substances, e lectrorheological a nd m agnetorheological  uids, hydrogel polymers, photo and thermoresponsive dyes [2–6]. 30-1

43722_C030.indd 1

10/3/2008 9:08:13 PM

30-2

30.2.1

Smart Materials

Classification

Indicators may be classied on the basis of the response mechanism, o r op erating p rinciples, a nd app lication. Ā us t here a re chemical, b iological, b iochemical, ele ctrical, m agnetic, o r mechanical i ndicators ac cording to t he re sponse me chanism. Based on the intended application, indicators may be classied as temperature i ndicator, t ime–temperature i ndicator, pH i ndicator, c ounterfeit i ndicator, t amper i ndicator, f reeze a nd t haw indicator, or freshness indicator.

30.2.2

General Operating Principles

Ā e re sponse me chanism o f mo st i ndicators i ncludes o ne o r more of t he fol lowing: ph ysical, c hemical, ph ysicochemical, electrochemical, a nd b iochemical. Ā e ph ysical me chanisms are ba sed o n ph otophysical p rocesses, ph ase t ransition, o r other c ritical m aterial properties suc h a s melting, g lass t ransition, c rystallization, b oiling, s welling, o r c hanges i n s pecic volume. In most cases, these transitions are driven by changes in the interaction forces (e.g., hydrophilic–hydrophobic forces) within o r a round t he i ndicator m aterial. I ndicator re sponse mechanism may a lso be based on chemical, biochemical, a nd electrochemical reactions. Examples include acid–base, oxidation– reduction, photo chemical, p olymerization, e nzymatic, a nd microbial reactions. Many of these changes are irreversible and the o nset o r ter mination c an b e ob served v isually a s c olor change, color movement, or mechanical distortion [1–6].

30.2.3 Choice of Indicators Some factors governing the selection of a given indicator device include the following: Cost: It must be relatively inexpensive. Ā e i ndicator should not be more expensive than the product it is protecting. Application: Easy to attach to a variety of containers or packages. Once installed, the device must remain intact and readable during the service life of the packages. Response: Ā e response mechanism must have fast kinetics on the order of seconds to a f ew hours and be reproducible. Ā er e should be no time delay in response between reactions involving solid, l iquid, o r ga s. M ost app lications re quire t he re sponse i n the indicator to be irreversible to preserve the needed indication record. Sensitivity: Ā e i ndicator must b e h ighly s ensitive, ac curate, and easily activated. A u ser-friendly indicator that can provide useful information when needed will make both the product manufacturer and the consumer happy. Shelf-life: Ā e i ndicator m ust h ave a s helf-life e qual to o r longer than that of the product it is monitoring. It should be technically difficult to duplicate or counterfeit the indicator’s re sponse. I n t his c ase, t he i ndicator i s ac ting a s a “smart” locking mechanism.

43722_C030.indd 2

30.2.4

pH Indicators

Ā e p H i ndicator i s p robably t he olde st a nd si mplest sm art chemical indicating device k nown. Ā e chemistry of acid–base indicators is well documented [7] and involves proton or electron exchange reactions. It will be mentioned only briey here. By Brönsted and Ostwald’s denition, indicators are weak acids or bases. Upon dissociation, they exhibit a s tructural and color change that is different from the undissociated form. Indicators commonly used in acid–base and redox titrations are conjugated organic dye s c ontaining o ne o r mo re l ight-absorbing g roups called chromophores. Ā e electronic structure of these dyes can be changed by redox or proton exchange reactions. Ā is also changes the absorption energ y in the visible region of the electromagnetic s pectrum, a nd c onsequently, t he c olor. Ā e color change therefore provides a visual indication of an endpoint of a titration [1,7]. 30.2.4.1

Indicator Materials

Indicators for pH measurement may exist as a s olution, emulsion, c olloidal gel, pap er, o r ele ctrode. M ethods f or p reparing them a re de scribed i n t he l iterature [ 1,7,8]. T able 3 0.1 l ists examples of common indicators used in solution form or as indicator papers. Ā ere are also nondye pH-sensitive smart polymers, which exhibit critical material changes in response to changes in the concentration of hydrogen or ot her ions. Typically, t here a re polymer electrolytes derived from homo- or copolymerization of acidic or basic functionalized monomers. Examples include poly(ethylene o xide), p oly(dimethylsiloxane); c opolymers o f N,N-dimethylacrylamide, N-t-butylacrylamide, a nd ac rylic acid; m ixtures of methacrylic ac id a nd methyl methacrylate; mixtures of diethylaminoethyl methacrylate and butyl methacrylate; po lystyrenesulfonate; po lyacrylamide, f unctionalized cellulose copolymers, etc. [8,9]. In the presence of specic ions, these polymer gels can swell or shrink over a range of pH at ro om tem perature, re sulting i n a v isible c olor c hange. Indicators b ased on t hem a re h ighly se lective, se nsitive, a nd inexpensive [8].

30.2.5 Temperature and Time–Temperature Indicators Recently, there has been a growing interest in the use of critical temperature indicators and TTIs in intelligent packaging technology. Ā e p roduction, p rocessing, d istribution, s torage, a nd point-of-use of temperature-sensitive products pose a great challenge to i ndustries. T emperature p lays a le ading role i n t he growth of microorganisms in foods and thus in the incidence of food poisoning. As a result, many perishable food and nonfood products a re p rone to tem perature-induced qu ality lo ss a nd degradation. Pharmaceutical products, medical devices, chemical reagents, industrial f ormulations s uch as inks , p aints a nd c oatings,

10/3/2008 9:08:14 PM

30-3

Intelligent Chemical Indicators TABLE 30.1 Selected Indicators for pH Measurement Common Name (nm)

Formula

pKaa

pH Range

Bromothymol blue

C27H28O5Br2S

7.3

6.0–7.6

Yellow

Blue

617

Bromocresol green Bromocresol purple o-Cresol red Chlorophenol red Methyl red Phenol red Phenolphthalein Tetrabromophenol blue a-Naphtholphthalein Ā y mol blue

C21H14O5Br4S C21H16O5Br2S C21H15O5S C19H12O5Cl2S C15H15N3O2 C19H14O5S C20H14O4 C19H6O5Br8S C28H18O4 C27H30O5S

4.9 6.3 8.2 6.25 5.0 8.0 9.5 3.56 8.4 9.2

3.8–5.4 5.2–6.8 0.2–1.8 4.8–6.4 4.4–6.3 6.4–8.2 8.0–9.8 3.0–4.6 7.4–8.8 1.2–2.8

Yellow Yellow Red Yellow Red Yellow Colorless Yellow Colorless Red

Blue Purple Yellow Red Yellow Red Red–Violet Blue Green–Blue Yellow

618 588 572 572 526 559 552 605 660 598

a

Color in Acid

Color in Alkaline

λmaxa

Measurements made in solution at zero ionic strength and for a temperature ranging from 15°C to 30°C.

photographic m aterials, a nd ot her rel ated i tems a re h ighly sensitive to temperature abuse. Ā erefore, it is useful to monitor the temperature history from production to point of consumption for such products. Ā ere has b een a g reater ne ed to mo nitor a nd prevent overheating-related failure of machine and electronic parts. Routine quality control a nd preventive maintenance of automobile a nd aircraft engines system against overheating may help avoid catastrophic and costly failures. To e nsure p roduct sa fety, f reshness, m inimize l osses a nd retain customer’s condence, dedicated temperature indicators or T TI a re at tached to m any c ommercial products. Ā ey show the tem perature o r re cord tem perature h istory o f t he p roduct and indicate if abuse has occurred, thus allowing the consumer to make judgments about product quality or safety. Ā e indicator may incorporate a color-changing dye, polymeric substance, enzyme, or time–temperature integrating materials, which must be thermally activated to function [10].

diagnosis, and troubleshooting. For example, a crayon would be useful w here a k nowledge o f t he e xact tem perature o f eng ine parts is desirable [13]. Such crayons are made of materials with known melting points calibrated to have a response time of a few seconds with an accuracy of ±1%. Two commercial examples include t he Telatemp tem perature-indicating c rayon (made b y Telatemp C orporation) a nd Tempilstik tem perature i ndicators (made b y T EMPIL D ivision, A ir L iquide A merica, S outh Plaineld, NJ). Ā ey are available for use in temperature ratings between 1 25°C a nd 1 200°C. I n a n app lication, t he i ndicating crayon i s u sed to m ark a de sired su rface o f k nown op erating temperatures suc h a s t he e xhaust m anifold, c ylinder heads, or fuel injection pump. An instantaneous color change upon marking i ndicates t he r ated tem perature h as b een e xceeded. Alternatively, if a color change is observed after 1 or 2 s, the rated temperature of the indicator has been reached. When the material is operating below the rated temperature, a color change may not be observed or take much longer time to occur.

30.2.5.1 Critical Temperature Indicators

30.2.5.3

Critial temperature indicators (CTIs) visually indicate that a material or product has been exposed to a n undesirable temperature below or above a reference critical temperature for a time long enough to alter product quality or cause safety concerns [11]. Several temperature indicators have been described in the patent literature although only a f ew have been commercialized. Ā eir principle of operation depends on temperature changes resulting from freezing, melting t ransition, l iquid c rystal f ormation, p olymerization, enzyme re actions, o r ele ctrochemical c orrosion [ 12]. S ome formulations are derived from a dye and a chemical blend or a thermoresponsive hydrogel polymer with a specic freezing or melting transition. Some commercial indicators are described below.

Temperature labels are designed to permanently and irreversibly respond to a nd re cord o verheating o f su rfaces. Ā e indicator labels a re m ade b y s ealing o ne o r mo re tem perature-sensitive materials i n a h ighly s table s elf-adhesive p olymeric o r pap er strip [14]. Temperature ratings are printed on, below, and above each indicator window in degree centigrade (°C) and Fahrenheit (°F). Several models in custom sizes and shapes and customerselected t emperature ra nges a re a vailable f rom T elatemp Corporation. To u se i ndicator l abels, t hey a re i nstalled o nto a desired su rface. W hen t he c alibrated a nd r ated temperature i s exceeded, the indicator window of the label undergoes a noticeable color change (from si lver to bl ack i n most Telatemp products). Response time is usually less than 1 s with a guaranteed tolerance of ± 1%. S ome u ses in clude m onitoring o verheating in e ngine components, e lectronics, t ransformers, c hemicals, f oods, a nd pharmaceutical items.

30.2.5.2

High-Temperature Indicating Crayons

Ā ese a re u sed m ainly f or p lant m aintenance, qu ality a ssurance, engine mo nitoring, h eat t reatment o f me tals a nd a lloys,

43722_C030.indd 3

Irreversible Temperature Labels

10/3/2008 9:08:14 PM

30-4

30.2.5.4 Defrost Temperature Indicators Ā ese are color-change indicators that are mostly used to monitor chill and frozen temperatures of chemical and food products during s torage o r d istribution [15]. Ā ey a re go /no go w ithout delay or go/no go w ith s ome del ay t ime de vices. Ā e go /no go without delay indicators provide information that the temperature has risen above a t hreshold value over a short time period. Ā e go/no go w ith delay indicators show the indicator has been exposed to a predetermined temperature for a specied length of time. Some examples are as follows: 1. ColdMark indicators (IntroTech, Inc.) are used to show if a p roduct h as b een e xposed to c ritical c old tem perature conditions with an accuracy of ±1°C. Several designs have been de scribed i n t he patent l iterature [16]. Ā ey consist essentially of a glass bulb lled with a colorless uid and a tube pa rtially  lled w ith a c olorless a nd a v iolet c olored uid separated by a green barrier uid. Ā e tube is enveloped in a transparent plastic casing provided with a pressure-sensitive a dhesive b acking for at tachment on to t he exterior or interior surface of a desired package [17]. For a freeze-monitor i ndicator, t he re sponse t emperature i s 32°F/0°C. Upon exposure to temperatures at or below the response temperature for about 30 min, the colorless uid in the bulb freeze and contract, causing the colored  uid in the tube to ow do wnward, c oloring t he b ulb. W hen warmed above the response temperature, the bulb remains permanently c olored. W hen not i n u se, t he i ndicator i s stored above the response temperature. 2. TwinMark i ndicator (I ntroTech, I nc.) i s a d ual p urpose indicator that can respond to cold and hot temperatures. It works l ike t he C oldMark a nd s hows i rreversible c olor change in the bulb when the cold side or hot side of the indicator is exposed to temperatures below or above predetermined values [17]. 3. ColdSNAP indicator (Telatemp Corporation) is a temperature recorder that contains a b imetallic sensing element [14]. When attached to a product, a safety tab is pulled to activate the sensor. A clear window in the indicator means safe p roduct s torage tem perature. W hen e xposed to a damaging c ritical tem perature, t he b imetallic s ensor snaps into the indicator window permanently changing it from clear to red. Ā e ac curacy o f t he i ndicator i s ± 2°C and c ustomers c an s elect snapp ing p oint tem peratures ranging from −20°C to +40°C. 30.2.5.4.1 Liquid Crystal Display Labels Liquid c rystal d isplays ( LCDs) a re m iniature re versible t hermometers engineered on a label. Ā ey are provided with an indicator w indow a nd a s elf-adhesive bac king. A l iquid c rystalline material o r p olymer i s p laced b ehind t he i ndicator w indow. When installed, the material changes from opaque to transparent at rated temperatures [12,14]. It is possible to select the indicator’s color change of interest. LCDs can be designed to c ontinuously

43722_C030.indd 4

Smart Materials

monitor su rface tem peratures c overing v arious tem perature ranges or indicate selected temperatures. For example, Telatemp markets LCD reversible temperature decals covering a r ange of −30°C to 120°C in 2–5°C increments and with visual tan/green/ blue color changes permitting readings to 1°C. 30.2.5.5 Time–Temperature Integrator or Indicators A t ime–temperature i ntegrator/indicator is a co ntinuous monitoring device or tag that measures both the cumulative exposure time and temperature of perishable products from production, distribution, s torage, a nd e ven p oint of u se. W hen a T TI i s at tached to a product and activated, t he time–temperature history to w hich t he product has been exposed is recorded and integrated into a single visual result (such as color change). In theory, TTI can be used as an informational, monitoring, and decision-making protocol at any point in the distribution chain of product quality and safety [11,18]. Ā e response mechanism of a TTI is based on the irreversible physical, c hemical, o r b iochemical c hanges suc h a s o xidation and re duction re actions, p rotein denat uration, en zymatic browning, subl imation o f ic e, a nd re crystallization o f ic e t hat occurs fr om t he tim e o f a ctivation. Ā e r ate o f t hese c hanges increases w ith tem perature. Ā e app lication a nd rel iability o f TTIs have been studied extensively by several workers [19–22]. Applications of T TI include food produce, pharmaceutical and blood products, cosmetics, adhesives, paints and coating, photographic and lm products, avionics, plastics, as well as shipment of live plants and animals. 30.2.5.6

Chemical Kinetic Basis for TTI Application

Major factors responsible for the deterioration of food and related chemical products include microbial growth, physical changes and biochemical or chemical reactions occurring within the perishable products. Some examples of spoilage reactions include acid–base catalysis, e nzymatic r eactions, f ree ra dical p rocesses, h ydrolytic reactions, lipid oxidation, meat pigments oxidation, nonenzymatic browning reactions, and polymerization or cross-linking processes [23]. Like any chemical reaction, the rates of degradation or quality loss of perishables are inuenced by environmental factors such as temperature, concentration (e.g., composition of internal gaseous phases), and water activity. Increasing the temperature, for instance, increases the rate of these reactions. Ā e re sponse of T TIs a nd c hanges i n food qu ality (or ot her perishables) can both be modeled by chemical kinetic equations. It is critical to gather accurate and reliable kinetic data for TTI to be suc cessful a nd rel iable. S everal s tudies a re a vailable o n t he subject [23–25]. Br iey, s tudies o n s everal f rozen a nd re frigerated f oods i ndicated t hat t he re sponse o f T TIs c orrelates w ith storage-related qu ality changes [26]. In a nother study, t he sensory c hanges i n f rozen h amburger w ere re ported to c orrelate with the response of a commercially available TTI [27]. Several i nvestigators h ave s tudied t he i nuence o f tem perature on the rate of quality loss. It is generally accepted that the rate o f qu ality lo ss b ehaves a s a n e xponential f unction o f t he reciprocal o f t he abs olute tem perature. Ā is is t he A rrhenius equation shown below:

10/3/2008 9:08:14 PM

30-5

Intelligent Chemical Indicators TABLE 30.2 Activation Energies of Some Selected Reactions Resulting in Food Quality Losses

TTI response

?

Shelf life of food

Activation Energy Range Type of Reaction

(kJ/mol)

(kcal/mol)

Diffusion controlled Acid/base catalysis Enzyme reaction Hydrolysis Lipid oxidation Nutrient losses Maillard reaction Protein denaturation Spore destruction Vegetable cell destruction

0–60 80–120 40–60 50–60 80–100 20–120 100–180 300–500 250–350 200–600

0–14.3 19.1–28.6 9.5–14.3 11.9–14.3 19.1–23.9 4.8–28.6 23.9–43 71.6–119.3 59.7–83.5 47.7–143.2

60–200

14.3–47.7

Microbial growth

Source: From Taoukis, P. S. and Labuza, T. P., J. Food Sci., 1989, 54(4), 783; Fu, B., Taoukis, P. S., and Labuza, T. P., Drug D ev. Indu. Pharm., 1992, 18(8), 829. With permission.

− EA ⎞ k = Z exp ⎛ ⎝ RT ⎠ where k is the reaction rate constant Z is the temperature independent preexponential factor EA is the activation energy, which describes the temperature sensitivity of the quality loss reaction R is the universal gas constant T is the absolute temperature in Kelvin (K) To measure loss of quality or shelf-life in perishables, one or more characteristic measurable quality factor, denoted X are selected. Āe quality f actor X c an b e a c hemical, m icrobiological, or ph ysical parameter. Ā e rate of change of the quality factor, with time under isothermal storage conditions is given by the following: d[ X ] = − k[ X ]n dt where t is the reaction time n is the reaction order For most food quality losses, reaction order of zero or one is typical for a simple rate constant. By considering the kinetics of changes in food quality at v arious temperatures the activation energ y EA for the quality loss reaction can be obtained [19,24,25]. Table 30.2 lists typical values of EA(Food) for quality losses. From the TTI kinetics, the activation energy of the indicator (EA(TTI)) can be obtained. 30.2.5.7

Correlation of TTI Response with Food Shelf Life

Taoukis a nd L abuza h ave de veloped a c orrelation s cheme f or predicting the shelf life of a product based on the TTI response [19,20]. Ā is s cheme i s ba sed o n a s eries o f k inetic e quations describing quality loss and TTI response (Figure 30.1).

43722_C030.indd 5

TTI kinetics

Teff(TTI)

Food kinetics

Assumption Teff(TTI) = Teff(food)

Teff(food)

FIGURE 30.1 Application of TTI as food quality monitors. (From Fu, B. a nd L abuza, T . P ., J. Food Distr. Res., 1992, 23(1), 9. With permission.)

Ā e effective tem perature f or a v ariable t ime–temperature distribution is determined from the TTI response kinetics. Ā is value is used in conjunction with the food kinetics to estimate the rem aining s helf l ife o r qu ality lo ss. Ā e s cheme s hown i n Figure 30.1 assumes that the effective temperature and the activation energies of both the TTI device and the food are the same. However, this is not always the case. For practical applications, a difference between EA(food) and EA(TTI) of ±8.4 kJ/mol (2 kcal/mol) generally gives a good prediction [19,20]. 30.2.5.8

Commercially Available TTIs

Ā ere are several types of TTI devices that have been described in the patent literature, but only three have continued to receive signicant at tention f rom t he industrial a nd scientic community for the past 15 years [11,18]. Ā ese include the 3M Monitormark, Lifeline, and VITSAB TTI. Ā ey are available as labels, tags, pallet sticks, or pallet with sheathes. Each of these is described below. 30.2.5.9

3M Monitormark TTI

Ā e principles, characteristics, and operation of the 3M Monitormark TTI (3M Identication a nd C onverter Systems Division, St . Paul, MN) is well documented [29,30]. It is designed to give a visual response on ly a fter a c ritical t hreshold tem perature h as b een exceeded. Ā e indicator consists of an assemblage of porous paper wick, a nd a re ctangular pap er pad s aturated w ith a bl ue c olored fatty acid ester and phthalate mix that has a specic melting point, a polyester barrier layer, which separates the wick from the reservoir pad (paper pad) and adhesives. Ā e response critical temperature is usually t he melting point of t he specialty chemical. Ā e prepared TTI tag remains inactive until it is activated. Ā e i ndicator i s ac tivated w hen a p olyester p ull-strip i s removed allowing contact between the porous paper and the reservoir. When exposed to temperatures higher than the threshold temperature, t he c hemical i n t he re servoir mel ts a nd m igrates along the porous wick. Ā e indicator has ve windows that progressively change color as the migrating chemical runs through.

10/3/2008 9:08:14 PM

30-6

Smart Materials

Ā e ob servation o f c olor i n t he  rst w indow i mplies t hat t he indicator h as b een e xposed to a tem perature h igher t han t he preset t hreshold tem perature. Ā e e xtent o f c olor mo vement through these windows indicates the cumulative exposure time spent above the threshold temperature. Response cards provided by the manufacturer are used to correlate the time–temperature relationship of each indicator. When all ve windows turn blue, the product is interpreted to be unacceptable. If stored according to manufacturer’s guidelines, the tag shelf life is 2 year from the date of manufacture. A p lot o f t he s quare o f t he r un-out d istance a gainst t ime i s linear for a Monitormark TTI and obeys the Arrhenius equation [20]. Ā e design and operation of this indicator relies on a diffusion mechanism so the activation energy of this TTI is limited to 0–60 kJ/mol (0–14.3 kcal/mol). Ā erefore, it is recommended for use i n ma ny en zyme a nd d iffusion-controlled spoilage reactions. 30.2.5.10

Lifeline Fresh-Scan and Fresh-Check TTI

Lifeline Technologies (L ifeline Technologies, Mor ristown, N J) offers two “full history” TTIs that monitor the freshness of perishables independent of temperature threshold. Ā e se products are the Fresh-Scan and Fresh-Check indicators and are based on the color change resulting f rom solid state polymerization of a diacetylenic monomer. Ā e construction, operation, and characteristics o f t he de vice i s de scribed i n de tail b y Patel a nd h is coworkers [31,32] as well as by Field and Prusik [33]. Ā e Fresh-Scan indicator consists of the following: a standard bar c ode c ontaining p roduct i nformation a nd a d iacetylenic monomer de posited o n a p ressure-sensitive ba nd; a p ortable microcomputer with a laser scanner, which measures changes in reectance of the indicator; and a data analysis work station. Ā e indicator ba nd i nitially s hows a bout 1 00% re ectance. As t he diacetylenic m onomer u ndergoes a t ime–temperature de pendent s olid s tate p olymerization d uring s torage, t he i ndicator band d arkens, c ausing a de crease i n t he me asured re ectance. Ā e r ate of c olor c hange follows A rrhenius b ehavior s o h igher temperatures en hance c olor de velopment r ate. Ā e microcomputer co mbines t his t ime–temperature cha racteristic w ith t he product information on the bar code to p redict the shelf life of the product. Ā e F resh-Check i ndicator i s a c onsumer-readable v isual label, which is also based on the color change of an incorporated polymerizable monome r, but w ith a d ifferent de sign a nd c onguration [15]. Ā e circular device consists of an inner polymer shell, c ontaining t he monomer a nd a n outer nonpolymer shell painted w ith a re ference c olor. W hen e xposed to s ome t ime– temperature storage conditions, the monomer at the center part converts to polymer, causing a progressive color development at a r ate t hat i ncreases w ith tem perature. I f t he p olymer c enter become darker than the outer reference, the consumer is advised to discard the product, regardless of the printed expiration [15]. Ā ese t wo L ifeline i ndicators a re ac tive a s s oon a s t hey a re manufactured and must be refrigerated at very low temperatures (< −24°C) to preserve high initial reectance. Wells a nd Si ngh

43722_C030.indd 6

have investigated the use of these indicators as a quality change monitor f or p erishable a nd s emiperishable p roducts suc h a s tomatos, lettuce, canned fruitcake, and UHT-sterilized milk [22]. It was shown t hat t he response k inetics of t hese i ndicators correlates with quality loss in the products studied. The indicators studied have activation energies of 84–105 kJ/mol (20–25 kcal/mol). Major limitations in these indicators are that the p olymerization re action i s ph otosensitive a nd a lso t heir inability to respond to short time–temperature history. 30.2.5.11

VITSAB TTI

Numerous studies have been carried out on the time–temperature-dependent enzyme reaction on which the Vitsab TTI (formally k nown a s I -Point) i s ba sed [34,35]. I t i s a “ full h istory” indicator that records the temperature conditions and time history of a p roduct i ndependently of t hreshold temperature. Ā e device uses two color-coded compartments to store the chemicals. Ā e green part houses a mixture of enzyme solution and a pH-indicating dye. Ā e gray part contains a lipase substrate suspension. Ā e line separating the two chambers represents a pressure-sensitive ba rrier. Ā e de vice i s ac tivated b y b reaking t he barrier between the two chambers. Ā is causes the enzyme solution and the substrate to mix to form the indicating solution. As the re action p rogresses, t he l ipase sub strates ( triglycerides) hydrolyze i nto t heir c omponent f atty ac ids, c ausing t he pH to drop. A change in the pH causes the dye to change color from an initial green color to a  nal yellow color at a rate that is temperature-dependent. A v ariety o f V ITSAB i ndicators c an b e c ustom-designed f or speci c t emperature-sensitive c ommodities by j udicious s election o f en zyme c oncentration a nd en zyme– substrate combination. At present, there are four different types of indicators, which come in one and three dots standard congurations depending on t he le vel o f p roduct s afety r isk to b e c ommunicated. Ā e single dot indicator will change color from green to yellow above a predetermined temperature. Ā e t hree-dot i ndicator c onveys the degree of temperature exposure with high reliability. Ā e ac tivation energ ies of t he m ajor T TI products a re su mmarized in Table 30.3. By matching the EA of these TTIs to the EA values f or s ome c ommon de terioration re actions i n f ood a nd pharmaceutical pro ducts ( Table 3 0.2), c urrent T TI t echnology should p rovide a r ange o f s elections t hat w ill me et m anufacturer’s growing interest and demand for these products.

30.2.6 Anticounterfeiting and Tamper Indicator Devices Counterfeiting, forgery, tampering, and piracy of valuable documents a nd pro ducts h ave e xisted f rom t ime i mmemorial. Counterfeiting has become a multibillion dollar business that is prospering more than many of the victim companies. Ā is highly organized t rade re sults i n loss re venue to b oth c ompanies a nd governments, as well as loss of jobs in private and public sectors. Ā e mot ivation for t he professional counterfeiters is t he ease in making huge prot w ith very low r isk of b eing c aught a nd

10/3/2008 9:08:18 PM

30-7

Intelligent Chemical Indicators TABLE 30.3 Activation Energy Values of Some Common Commercial TTIs TTI Activation Energy Manufacturer Visual Indicator Tag Systems, AB-Malmo, Sweden

3M Packaging System Div., St. Paul, MN Lifeline Technology Inc., Morris, Plain

a

TTI Model

(kJ/mol)

(kcal/mol)

Reference

VITSAB Type 1 VITSAB Type 2 VITSAB Type 3 VITSAB Type 4 I-Point 3014a I-Point 4007a I-Point 4014a I-Point 4021a All types Fresh-Check A7 Fresh-Check A20 Fresh-Check A40 Fresh-Check B21 Fresh-Scan-18 Fresh-Scan-41 Fresh-Scan-68

57.4 94.7 126.1 194.8 47.8 137.0 101.8 141.2 41.1 155.0 81.3 81.7 88.0 113.1 85.8 82.5

13.7 22.6 30.1 46.5 11.4 32.7 24.3 33.7 9.8 37.0 19.4 19.5 21.0 27.0 20.5 19.7

[36] [36] [36] [36] [19] [19] [19] [19,20] [19,36] [36] [19] [19] [19] [19] [19,20] [19,20]

Vitsab was formally known as I-Point.

penalized. It is impossible to enumerate products or documents that are counterfeited or quantify the impact of these goods on vulnerable consumers. Activities of counterfeiters have increased in developed countries. Ā e incidence of counterfeiting is even greater in developing countries, which have become a dumping ground for counterfeit products. A pa rtial list of goods that are commonly f aked i nclude a ntiquities, c urrency, c redit le tters, credit c ards, je welry, c hemicals, p rocessed f ood, ph armaceuticals, a nd c osmetic p roducts, p roduct iden tication marks, videos, c ompact d isks (CDs), c lothing a nd appa rel ac cessories, electronics, a uto a nd a irplane p arts, c hildren’s t oys, w atches, sporting goods, travel documents such as passports, identity cards and driving permits, and textile materials [37–39]. Advances i n t echnology s uch a s c omputers, l asers, s canners, charge-coupled devices (CCD), color printers, optical holograms, and a nalytical tools have en hanced t he state of t he a rt for a nticounterfeiting security and encryption. Ironically, these advances are also being exploited by the adversaries to promote high-tech forgery. To combat these activities, research efforts have continued to develop tamper-resistant security mechanisms that will make co unterfeiting to o e xpensive to b e l ucrative. Ā is subject has been reviewed recently [37,38]. A summary of some of the available technologies relevant to this chapter will be presented.

Ā e rel atively t hicker D enisyuk h ologram re quires a mo re stringent m anufacturing p rocess. I n add ition to b eing mo re expensive to produce, it is not as mechanically robust as the embossed hologram. Ā is hologram is the more easily counterfeited optical hologram. Holographic Dimensions, Inc. offers the Verigram security system that consists of two elements, the display hologram and the machine readable Verigram™ hologram, which afford a completely secured device. Although the display component may be counterfeited, the Verigram™ element is claimed to be impossible to forge [40]. One of the leading manufacturers of anticounterfeiting holograms is American Bank Note Holographics (ABNH) Inc. Ā ei r products are used for various commercial and security applications. Ā ey i nclude h olographic s tripe, h olomagnetic s tripe, holographic thread, transparent holographic laminate, hot stamp foil, h ologuard, f rangible v inyl a nd p ressure-sensitive a nd tamper-evident labels. ABNH can manufacture holograms with magnetic stripe, visible, and invisible bar codes, or microprinting for added security, product tracking, audits, and authentication. Ā e l ist o f p roducers a nd v endors o f c ommercial a nd s ecurity holograms is endless. Only two are mentioned here as examples.

30.2.6.1

Taggants a re m icroscopic c olor-coded pa rticles der ived f rom several layers of a highly crossed-linked melamine polymer and are u sed e xtensively a s m arkers [ 41]. L ayers o f m agnetic a nd uorescent materials are added to the taggants’ particles during formulation to simplify detection and decoding. Ā e technology, originally developed for explosives, nds applications in antiterrorism, a uthentication, p iracy, c ounterfeiting, a nd p roduct

Holographic Technology

An opt ical hologram is a t hree-dimensional photo created on a photoresist lm or plate by the reection and refraction pattern of laser l ight i ncident o n a n ob ject. Two o f t he h olograms w idely used in security applications are the Embossed and the Denisyuk hologram. Ā e embossed hologram can be readily mass-produced as a mechanically tough and durable thin lm at a very low cost.

43722_C030.indd 7

™

30.2.6.2 Microtaggants Technology

10/3/2008 9:08:19 PM

30-8

Smart Materials

identication [41]. Taggant particles can be coded with information suc h a s t he p roduct m anufacturer, p roduction d ate a nd batch number as well as the distributor. When used in antiterrorism, t aggants a re t horoughly m ixed w ith t he e xplosives. When a b omb c ontaining t agged e xplosive i s de tonated, rem nant taggants maybe collected from debris with the aid of a UV light and a m agnet. Ā e coded color sequence information can be re trieved w ith a n opt ical m icroscope. Ā is reliably helps i n tracing the origin of the explosive. Tagged particles can also be added to printing ink for security printing. Ā is provides a way to combat counterfeit, forgery, and product d iversion a nd a ssist p roduct/property iden tication. Microtaggants c an b e adde d to m any p roducts ei ther d irectly during f ormulation o r b y t hermal t ransfer,  lms, laminates, spraying, a nd s o o n. S ome o f t he c ompanies t hat m arket Microtaggant i dentication pa rticles a re M icrotrace I nc., Minneapolis, M N; M ICOT C orporation ( St. Pa ul, M N); S W Blasting ( BÜLACH, S witzerland); a nd P last L abor ( Bulle, Switzerland). Pat ronage c omes f rom government a gencies, l aw enforcement, and industry. 30.2.6.3

Smart Ink Technologies

Smart inks or coatings change color predictably in response to alteration of their environment. Color change may or may not be reversible and smart inks or coatings are  nding applications in areas such as security printing, lenses, and other optical devices. Ā e two main classes of smart inks include those derived from photochromic and thermochromic materials. Photochromic inks are formulated from light-sensitive dyes, pigments, p olymers, o r ot her c olorless c hemical c ompounds. Ā ese i nks c hange c olor w hen e xposed to u ltraviolet l ight. Examples i nclude e xtrusion l amination i nks, i nvisible i nk security m arkers, t agged i nks, a nd i nks ba sed o n c opper-free metallics. Ā ermochromic inks incorporate temperature-sensitive polymers, dyes, or pigments. When exposed to a predened temperature, t he i nk u ndergoes a c olor c hange. Typical c olor-sensitive components i nclude d iacetylenic co mpounds, h ydrogel p olymers, and thermochromic liquid crystals. Security inks in corporates  uorescent d yes, photo chromic dyes, thermochromic dyes, or organic pigments in their formulation. Ā ese dyes or pigments reversibly change color when exposed to u ltraviolet l ight o r h eat o f p redened i ntensity o r temperature. Ā is forms a basis for their use in security printing, counterfeit o r f orgery i ndicators, t amper-evident, a nd ot her applications. Several companies around the world manufacture and distribute counterfeit detector pens or specialty uorescent light bulbs for quick authentication of currency bills, gift certicates, and other documents. 30.2.6.4

Tamper-Indicating Devices

Tamper indicators are security seals or labels designed to detect product or document tampering. Indicator devices available are based on  ber opt ic ac tive s eal te chnologies, opt ical t hin  lm, holograms, a nd c hemical re actions. A m ajor re quirement f or

43722_C030.indd 8

these m aterials i s t hat t hey p rovide u nambiguous e vidence o f tampering. Ā ey must be resistant to environmental conditions such as heat or solvents and relatively cheap. Several low-cost tamper i ndicator p roducts a re c ommercially a vailable to me et various customers’ needs. Telatemp (T elatemp C orporation) s ells T elaSEAL s ecurity labels constructed of polyester  lm, high strength instant-bond adhesive, and other proprietary chemicals [14]. A repeating geometrical pattern is printed onto the TelaSEAL during construction for additional security. Ā e seal can be applied to a v ariety of su rfaces i ncluding p lastics, me tals, w ood, a nd pa inted su rfaces. A d isrupted re peating pat tern w ith a c ustom me ssage “OPENED” is transferred to the applied surface when the device is peeled from the substrate. Ā is p rovides p ermanent an d unambiguous evidence of tampering. Meyercord (Sentinel Division, Carol Stream, IL) supplies several tamper-evident adhesive-base materials that leave a “VOID” message on surfaces, thus providing indication of unauthorized opening or tampering of products. Ā e “VOID” message can be customized to i nclude c ompany’s na me, t rade m ark, logo s, o r serial number. 3M h as de veloped s everal a uthenticating de vices ba sed o n optical thin lm technology for counterfeit detection and tamper indication. A proprietary metallic surface coating is formulated and de posited o n a p ressure-sensitive m aterial. Sp ecialized imaging technologies are used to generate unique covert graphics and print product information or serial number on the device. When the device is peeled from the substrate, the covert image is revealed fol lowed by a n i rreversible c olor c hange, w hich pr ovides a su re indication of tampering. Ā ese devices  nd uses in asset control, package sealing a nd identication, parts identication, serial numbering, and consumer information.

30.2.7 Indicating Device Issues and Limitations Interest in indicating devices continues to g row and numerous potential app lications h ave b een p roposed. D espite sig nicant advances made in chemical indicator technology during the last 25 ye ars, o nly a f ew de vices h ave b een c ommercialized w hen compared to the number of patents granted each year. Ā ere are several re asons w hy m any m anufacturers, go vernments, a nd consumers a re rel uctant to adopt c hemical i ndicating de vices. Major c onsiderations o n t he pa rt o f m anufacturers i nclude device rel iability, c ost, a nd e ducation. M anufacturers f ear t he consumers w ill b e rel uctant to emb race p roducts c ontaining these indicators for reasons of additional cost. Ā e addition of a leak indicator to the headspace of packaged food or pharmaceutical products is likely to be greeted with stiff resistance on the part of the consumer for safety reasons. To gain acceptance, commercial indicators must be relatively inexpensive compared to t he p roducts to w hich t hey a re app lied. A lso, t hey m ust b e technically competent to perform their intended functions. Response mechanisms of many indicators are poorly understood [42]. For example, to produce a reliable time–temperature

10/3/2008 9:08:19 PM

30-9

Intelligent Chemical Indicators

indicator, the quality index, kinetics, and mechanism of food or chemical p roduct de gradation u nder v arious en vironmental conditions m ust b e u nderstood. S ome de gradation re actions may not always follow simple Arrhenius kinetics. Ā er efore, predictions ba sed o n t hem m ay b e f aulty a nd t hus a ffect product reliability. Consumer surveys [12,30] on the use of TTI on food products o n a go od note i ndicate t hat t he p ublic re ception o f these indicators is growing. It also raised awareness on the need to e ducate t he c onsumers o n t he u sefulness, rel iability, a nd application of indicator devices in order to i ncrease the current level of acceptance. Even t hough most of t he counterfeit a nd t amper i ndicator technologies c urrently a vailable a re rel iable, t hey a re qu ite expensive. F or t his re ason, m any i ndustries a re rel uctant to implement t hem. S ome t he i ndicators a re l imited i n t heir scope of application a nd a re e asily forged or c opied a s i s t he case f or m any opt ical h olograms [ 37–40]. Ā is h as bee n a major p roblem w ith m any c redit a nd v arious iden tication cards. Ā e challenge is to produce indicators that are reliable, technically difficult to copy or fake, and are available at a relatively low price. Recently, s everal i ndicating de vices h ave m ade i nroads i nto the market place with unsubstantiated claims of what the product can off er. In some cases, instructions for consumers on the application a nd i nterpretation o f i ndicator re sponse o r c olor change are inadequate. Such consumers appear to b e confused. Information about the safety of the indicator should be provided too. Efforts should be made to design indicators that are childproof to avoid ingestion and other accidents. A general specication for the various classes of indicators needs to be established. Such a s pecication s hould p rovide p erformance c riteria t hat each indicator design must meet for a given application. It should also set standards for acceptable qualication testing. Ā is will ensure t hat c ommercial i ndicating de vices me et t he ne eds o f industry.

30.3 Conclusion Ā e application of chemical indicating devices has come to stay and now available. In many cases, they are living up to expectations in reducing the incidence of counterfeiting, forgery, tampering or piracy, and losses of perishable products. Temperature and TTIs are now being used in routine packaging of some critical food and chemical products. For TTI, only limited information o n t he k inetics a nd re sponse me chanism i s c urrently available in the literature. New indicators covering a large range of activation energies for many spoilage reactions need to be designed. Ā is will allow the scope of application of TTIs to b e extended. M ore re search a nd d evelopment e ffort c ould a nd should be directed toward addressing the present technical and manufacturing c hallenges h indering c ommercialization o f many patented inventions. Public education will accelerate interest and ultimate acceptance of these products. Such acceptance may create sufficient volumes of use, which would lead to lower costs for manufacturing these smart devices.

43722_C030.indd 9

Acknowledgments Ā e a uthor w ish to t hank D r. Ted L abuza f or p roviding ke y references on TTI, Dr. Annapoorna Akella for help with patent literature search, Dr. John Evans, Dr. Jim Harvey and, Dr. David Landman, f or h elpful c omments; a nd to c ompanies w ho g raciously provided relevant information.

References 1. Boside, G. a nd A. H armer., Chemical a nd B iochemical Sensing With O ptical F ibers a nd Waveguides, B oston, M A: Artech House, Inc. 1996. 2. Nagasaki, Y. a nd K. K ataoka, P olysilamines as in telligent materials, CHEMTECH, 1997, 3. 3. Hoffman, A. S., Intelligent p olymers in me dicine and biotechnology, Artf. Organs, 1995, 19, 458–467. 4. Galaev, I. Y., S mart p olymers in b iotechnology a nd me dicine, Rus. Chem. Rev., 1995, 64, 471. 5. Mucklich, F. a nd H. J anocha,Smart ma terials—Ā e IQ of materials in systems, Z. Metallkd., 1997, 87, 357. 6. Snowden, M. J., M. J. Murry, and B. Z. Chowdry, Chem. Ind., 1996, 531. 7. Kolthoff, I. M. a nd C. Ros enblum, Acid-Base I ndicators, London: Macmillan, 1937. 8. B romberg, L. a nd G. L evin, Macromol. R apid C ommun., 1996, 17, 169. 9. Dagani, R., Intelligent gels, Chem. Eng. News, 1997, 26. 10. Bryne, C. H., Temperature indicators—the state of the a rt, Food Technol., 1976, 6, 66. 11. Taoukis, P. S., B . F u, a nd T. P. L abuza, T ime–temperature indicators, Food Technol., 1991, 45(10), 70 a nd r eferences therein. 12. Woolfe, M. L., Temperature monitoring and measurement. In: Chilled Foods, A Comprehensive Guide, Dennis, C. and M. Stringer (eds.), New York: Ellis Horwood Ltd, 1992, Ch. 5. 13. Anon. T emperature indica tors  nd m any ap plications i n engine e quipment ma intenance, Diesel P rog. E ng. D rives, 1996, 8, 36. 14. Telatemp C orporation, F ullerton, CA P roduct ca talog, 1997. 15. Selman, J. D., Time/temp’ indicators: how the y work, Food Manufacture, 1990, 65(8), 30 and references therein. 16. Manske, W. J ., Cri tical Temperature I ndicator, U.S. P atent No. 4,457,252, 1984. 17. Introtech, Inc., St. Paul, MN Product catalog, 1992. 18. Schoen, H. M. a nd C. H. B yrne, D efrost indica tors: ma ny designs have been patented yet there is no ide al indicator, Food Technol., 1972, 26(10), 46. 19. Fu, B. and T. P. Labuza, Considerations for the application of time–temperature integrators in f ood dist ribution, J. Fo od Distr. Res., 1992, 23(1), 9. 20. Taoukis, P. S. and T. P. Labuza, Applicability of time–temperature indica tors as f ood q uality mo nitors under no nisothermal conditions, J. Food Sci., 1989, 54(4), 783.

10/3/2008 9:08:19 PM

30-10

21. Taoukis, P. S. and T. P. Labuza, Reliability of time–temperature indicators as f ood quality monitors under is othermal conditions, J. Food Sci., 1989, 54(4), 789. 22. Wells, J. H. and R. P. Singh, Application of time–temperature indicators in mo nitoring cha nges in q uality a ttributes o f perishables a nd s emi-perishable f oods, J. Fo od S ci., 1988, 53(1), 148. 23. Labuza, T. P., Application of chemical kinetics to deterioration of foods, J. Chem. Edu., 1984, 61(4), 348. 24. Hendrickx, M., G. M aesmans, S. D e C ordt, J. Noronha, A. Van Loey, and P. Tobback, Evaluation of the integrated timetemperature effect in thermal processing of foods, Crit. Rev. Food Sci. Nutrir, 1995, 35(3), 231 and references therein. 25. Villota, R . and J. G. Hawkes, Reaction kinetics in f ood systems. In: Hand Book of Food Engineering, Heldman, D. R. and D. B. Lund, (Eds.), New York: Marcel Dekker, 1992, p. 39. 26. Wells, J. H. and R. P. Singh, A kinetic approach to food quality p rediction usin g f ull-history time-t emperature indicators, J. Food Sci., 1988, 53(6), 1866. 27. Grisius, R., J. H. Well, E. L. Barrett, and R. P. Singh, Correlation of time–temperature indicator response with microbial growth in pasturized milk, J. Food Proc. Preserv., 1987, 11, 309. 28. Fu, B., P. S. Taoukis, and T. P. Labuza, Ā eoretical design of a variable activation energy time–temperature integrator for prediction of food or drug shelf life, Drug Dev. Indu. Pharm., 1992, 18(8), 829. 29. Manske, W. J ., S elective T ime I nterval I ndicating D evice, U.S. Patent No. 3,954,011, 1976. 30. Selman, J. D., Time–Temperature Indicators, In: Active Food Packaging, Rooney, M. L., (Ed), New York: Blackie Academic & Professional, 1995, Ch. 10 and references therein.

43722_C030.indd 10

Smart Materials

31. Patel, G. N. and K. C. Yee, Diacetylene Time–Temperature Indicator, U.S. Patent No. 4,228,126, 1980. 32. Patel, G. N., A. F. P reziosi, a nd R . H. B aughman, T ime– Temperature History Indicators, U.S. Patent No. 3,999,946, 1977. 33. Field, S. C. and T. Prusik, Shelf life estimation of beverage and food products using bar coded time–temperature indicator labels. In: The Shelf Life of Foods and Beverages, Charalambous, G., (E d.), Amsterdam: Els evier, 1986, p. 23. 34. Blixt, K., S. I. A. Tornmarck, R. Juhlin, K. R. Salenstedt, and M. Tiru, Enzymatic Substrate Composition Adsorbed on a Carrier. U.S. Patent No. 4,043,871, 1976. 35. Blixt, K. and M. Tiru, An enzymatic time/temperature device for mo nitoring t he ha ndling o f p erishable co mmodities, Dev. Biol. Std., 1977, 36, 237. 36. Vitsab TTI Product catalog, 1997, @ www.vitsab.com. 37. van Renesse, R. L., (ed.) Optical Document Security, Boston, MA: Artech House, 1994. 38. Lancaster, I., Progress in counterfeit deterrence; the contribution of information exchange, Proc. SPIE, 1996, 2659, 2. 39. McGrew, S. P., Countermeasures against hologram counterfeiting, Proc. Optical Security Systems Symposium, Zurich, Oct. 14–16, 1987. 40. Brown, K. G., J. Weil, and M. O. Woontner, Counterfeit Proof and Machine Readable Holographic Technology Verigram Systems, @ http://www.hgrm.com/counterf.htm. 41. Rouhi, A. M., G overnment, ind ustry eff orts yield a rray o f tools to combat terrorism, Chem. Eng. News, 1995, July 24. 42. Selman, J. D., Time–temperature indicators: do the y work? Food Manufacture, 1988, 63(12), 36 and references therein.

10/3/2008 9:08:19 PM

31 Piezoelectric Polymer PVDF Microactuators

Yao Fu Silverbrook Research Pty. Ltd.

31.1 I ntroduction ........................................................................................................................... 31-1 31.2 Piezoelectric Polymer Cantilever......................................................................................... 31-2 31.3 Modeling and Simulation ..................................................................................................... 31-2 Fabrication of Piezoelectric Composite Cantilevers ......................................................... 31-3

31.5

Testing and Evaluation of Cantilevers ................................................................................31-4

Erol C. Harvey Swinburne University of Technology

Muralidhar K. Ghantasala Western Michigan University

Parametric Analysis for the Cantilever Deection

31.4

Excimer Laser Ablation · Punching of the PVDF Cantilever · E lectroplating Permalloy on the Metallized PVDF Determination of Piezocoefficient of PVDF Using a Bimorph Cantilever · Te sting of Unimorph Cantilever

31.6 S ummary .................................................................................................................................31-6 References ...........................................................................................................................................31-6

31.1 Introduction Microactuators a re o ne o f t he mo st i mportant c omponents i n microelectromechanical systems (MEMS). Ā e simplest and earliest actuating elements are microcantilevers. In the late 1980s, microactuators received increasing attention when electrostatically d riven m icromotors started to app ear [1,2]. I n t he last 15 years, m any t ypes o f m icroactuators ut ilizing v arious d riving forces (e.g., e lectrostatic, e lectromagnetic, p iezoelectric, s hape memory a lloy ( SMA) e tc.) h ave b een de veloped. P iezoelectric actuation i s one of t he most p opular ac tuation principles u sed for microactuators. Piezoelectric m icroactuators gener ate l arge f orces b ut o nly small d isplacements b y app lying a vol tage to t he p iezoelectric materials. Typically u sed c onstructions a re t he b imorph [ 3,4] and multilayer structures. In addition, more and more unimorph composite b eams [ 5–8] a re em ployed to me et t he i ncreasing requirements of special applications. Ā e piezoelectric actuation is employed in many applications, such a s electric fans [4], hydrophones [9], m icrophones [10,11], inkjet p rinters, c ontrol v alves [ 12], m icropumps [ 13], t actile sensor [14], acoustic control [15], and micromotors [16], etc. Ā e main advantages of this actuation are its high precision, speed, and mechanical power. In these applications, piezoelectric ceramic materials such as zinc oxide [17] and PZT [7,8] are most commonly used, as they exhibit large piezoelectric coefficients. However, the major difficulty

associated with their use in many applications is the requirement of advanced deposition technologies and facilities to prepare stoichiometric thin  lms. Further, these are usually brittle, and have a relatively large Young’s modulus, limiting the achievable strain. S ome composite ac tive polymers [18] comprised of piezoelectric ceramics a nd epoxies were fabricated to c ompensate f or t hese d isadvantages. P iezoelectric p olymers suc h a s polyvinylidene  uoride ( PVDF) a nd i ts c opolymers c an o vercome some of these difficulties even though they have a relatively low pie zoelectric c oefficient. L ow n umerical Young’s mo dulus values of these polymers [19,20] have the potential for enabling relatively large strain piezoelectric actuators. Ā e s pecial c haracteristics o f p iezoelectric p olymers h ave attracted the attention of many researchers from many different disciplines. A lot o f devices were designed from these materials since the 1980s. A long list of papers and patents can be found in the p iezo lm s ensors te chnical m anual f rom M easurement Specialist [21]. In some of these devices, PVDF has been used mostly i n t he form of a t hin foil, either s tretched or mounted. Ā e a bility to pat tern PV DF i nto re quired s hapes a nd si zes i s highly de sirable i n order to i ncrease t he r ange of applications, particularly in the millimeter to micrometer scale. Several different methods have been employed to pattern PVDF, including the u se o f a n e xcimer l aser [ 22], U V l ight s ource [ 23], x -rays [24,25], a nd h ot emb ossing [ 26]. E ach me thod h as i ts o wn strengths and problems. Excimer laser ablation will change the UV-VIS photoabsorption spectrum of PVDF polymer [22]. Heat 31-1

43722_C031.indd 1

10/3/2008 9:08:37 PM

31-2

Smart Materials

diff usion can easily damage the piezoelectric effect of PVDF and can also burn the edge of the pattern [27]. A direct pattern transfer i nto PV DF i s p ossible t hrough x-ray-induced e tching [24]. However, t he x -ray f abrication i s m uch mo re e xpensive t han laser ablation. Hot embossing PVDF will have the risk of damaging t he p iezoelectric e ffect i n PV DF p olymer a s i t t ypically requires molding temperatures in t he range of 175°C to 185°C. One group in Japan [28] used an electrospray deposition method to deposit thin-layer PVDF on the substrate biased with a voltage of 8–15 kV. A piezoelectric polymer composite cantilever was designed by using  nite element method si mulation. Ā e a im of t he design was to  nd the optimal parameters and dimensions of a p iezoelectric cantilever to ac hieve large tip deection and force. Ā e inuence o f t he pa rameters o f t he c antilever o n t ip de ection was analyzed based on the simulations. Ā e designed piezoelectric unimorph actuator was fabricated using punching and electroplating te chniques. Ā e m icrostructures o n t he top o f t he piezoelectric polymer PVDF were realized by a low-temperature hot embossing process, thus avoiding the depoling of the PVDF polymer. Ā e m aterial p roperty p iezocoefficient o f PV DF w as characterized. Finally, the experimental test results of the composite cantilever are explained.

Z

31.3

Modeling and Simulation

Simulation a nd modeling using numerical methods employing a vailable f inite elemen t s oftware i s a n e fficient w ay to design and optimize the microstructures. We performed the simulation using ANSYS and CoventorWare. The 3-D model

PVDF

Permalloy

FIGURE 31.1

43722_C031.indd 2

Cross-section of the piezoelectric polymer cantilever.

X −20.00

10.00

Y

Displacement magnitude

40.00

70.00

100.00

(a)

Z

31.2 Piezoelectric Polymer Cantilever A unimorph cantilever beam is chosen as the basic element. Ā e cantilever i s m ade o ut o f a c ommercially a vailable p iezoelectric polymer PV DF s heet, w ith a t hin p ermalloy l ayer ele ctroplated onto one side in our laboratory. Ā e schematic cross-section of the cantilever is shown in Figure 31.1. Ā e t hickness of t he original PVDF sheet is 28 µm a nd i s c overed w ith a 4 0 nm t hick n ickelcopper alloy metal electrode layers on both sides. Ā e PVDF polymer i s p olarized i n a d irection pa rallel to i ts t hickness. Ā e nonpiezoelectric layer of the unimorph cantilever is fabricated by electroplating p ermalloy to a t hickness o f app roximately 5 µm. Ā e length of the cantilever beam is 6 mm and the width is 1 mm.

Potential (reol)

Y

X 0.00

17.96

35.93

53.89

71.86

(b)

FIGURE 31.2 Simulation of a c antilever beam using CoventorWare. (a) D istribution of t he e lectrical p otential a cross t he PV DF p olymer. (b) Displacement of the cantilever caused by piezoelectric effect.

of a c antilever beam simulated using CoventorWare package is shown in Figure 31.2. It is showing two layers, one for PVDF and t he ot her f or p ermalloy. T he b eam d imensions w ith a length of 5 mm, a width of 1 mm, and a thickness of 28 µm for the P VDF l ayer wi th th e th ickness o f n onpiezo p ermalloy layer o f 5 µm. T he Y oung’s mo duli o f PV DF a nd t hat o f permalloy w ere i nput a s 2 .8 [ 21] a nd 1 50 GPa [ 29], re spectively. T he t ip de flection at a n app lied p otential o f 100 V i s found to be 71.9 µm. Figure 31.2a shows t he geometric d istribution o f t he ele ctrical p otential o ver t he t hickness o f piezoelectric e lement a nd F igure 31.2b t he b eam de flection under this potential.

31.3.1 Parametric Analysis for the Cantilever Deflection Ā e t ip de ection of a unimorph cantilever is determined by its geometrical d imensions, m aterial p roperties, a s w ell a s b y t he external electrical eld. In order to design an effective microactuator with maximized deection, i t i s ne cessary to opt imize a ll t hese parameters. A detailed analysis of the effect of individual parameters on the tip deection of the cantilevers was performed.

10/3/2008 9:08:38 PM

31-3

Deflection of beam (PVDF + 5 µm permalloy)

800

300 200 100 0 0

50 100 Thickness of PVDF (µm)

150

Beam deflection (µm)

Deflection of beam (µm)

Piezoelectric Polymer PVDF Microactuators

9 µm 28 µm

600 400 200 0 0

FIGURE 31.3

Tip deection versus PVDF thickness.

As expected, the tip deection increases proportionally to the square o f t he b eam leng th a nd i s p roportional to t he app lied voltage. Ā e width of the cantilever has little effect on the beam deection. Ā erefore in the simulation, the width of the cantilever is xed at 1 mm and applied voltage on the cantilever was set at 100 V. Ā e t ip de ection o f a c omposite c antilever b eam de pends on t he t hickness of both of its piezo and nonpiezo layers. Ā e effect o f t he p iezo l ayer t hickness i s  rst st udied. Figure 31.3 shows the variation of the beam deection, as the thickness of the PV DF l ayer i s v aried i n t he r ange 9 –110 µm. Ā is graph clearly s hows t hat t he t hinner t he PV DF l ayer, t he l arger t he deection of the beam. Ā is kind of variation in beam deection with the piezolayer thickness can be explained by the fact that t he i nduced strain on t he t hinner PV DF u nder t he same electrical p otential i s l arger d ue to t he l arge ele ctrical  eld (voltage/thickness o f t he b eam) a nd a lso t he t hinner c rosssection o f t he c antilever b eam h as le ss mo ment o f i nertia, therefore th e d eection o f th e c antilever i ncreases wi th th e decrease of the PVDF thickness. Even though the 9 µm t hick PVDF l ayer p rovides t he l argest de ection, it is not practical for use in most applications; as such, a structure is very exible and provides little force. Ā erefore, a 2 8 µm thick PVDF layer was is used in this design. It is equally important to determine the effect of the nonpiezo layer. F igure 3 1.4 s hows t he v ariation o f t he b eam de ection with th e th ickness o f th e n onpiezo p ermalloy l ayer, w hen all other pa rameters a re ke pt c onstant. I t c an b e s een f rom t he graph t hat f or a 2 8 µm t hick PV DF l ayer t he de ection ha s a maximum of around 75 µm with a permalloy layer thickness of about 2 µm, but decreases very little up to a t hickness of 5 µm. When t he t hickness i s i ncreased to 3 0 µm, t he de ection decreases to 9 µm. For a 9 µm thick piezo layer, the beam deection can reach up to 700 µm with a nonpiezo layer thickness of 1 µm, but decreases to 50 µm with a nonpiezo layer thickness of 15 µm. Ā en the two curves converge after the thickness of the nonpiezo layer increases to values above 20 µm. Ā is means that when t he no npiezo l ayer t hickness e xceeds 2 0 µm, t he b eam deection is mainly affected by the nonpiezo layer only. In this design, a thickness of 5 µm was chosen as an optimal value for a 28 µm piezoelectric polymer, as a thinner permalloy layer might get d amaged du ring a ctuation. Ā is a nalysis s hows t hat w ith

43722_C031.indd 3

FIGURE 31.4 layer.

10 20 30 Thickness of nonpiezo layer (µm)

40

Tip deection versus the thickness of nonpiezoelectric

certain material the thickness of the nonpiezoelectric layer can be optimized to achieve the maximum deection. Young’s mo dulus o f t he no npiezoelectric l ayer h as b een identied a s a nother i mportant pa rameter w ith re spect to t he beam deection. In order to achieve the maximum tip deection of t he c antilever, a n opt imal m aterial o f t he no npiezoelectric layer can be worked out from this simulation. It is observed from the simulation that the useful range of Young’s modulus from a practical material point of view is in the range of 10–1000 GPa. Ā e numerical results obtained from the nite element analysis (FEA) si mulations w ere c ompared a nd c on rmed w ith t he theoretical a nalysis ba sed o n b imetal t heory. Ā e optimal parameters f or t he p iezoelectric c omposite c antilever a re determined b y t he t hickness a nd t he Young’s mo dulus o f t he nonpiezoelectric layer. For a given material of the nonpiezo layer, an optimum thickness can be calculated for obtaining the maximum tip deection. Similarly, a suitable material can be selected from this kind of simulation depending on the chosen actuator conguration and the thickness of the layer. Ā e potential materials for the nonpiezoelectric layer can be metals (e.g., aluminum) or polymers, such as polyimide and Mylar. Ā e use of a magnetic material a s t he nonpiezoelectric layer facilitated t he for mation of a h ybrid m icroactuator. Si nce t he si mulation s tudies indicated a required nonpiezo layer modulus in the range of 10 to 1000 GPa, we selected a soft m agnetic p ermalloy (Ni 80Fe20) with a modulus of 150 GPa for this purpose. Ā is simulation and modeling work helped us in arriving at some of the optimal parameters of the beam length of 5 mm (depending on the required deection), width of 1 mm, and thicknesses of the PVDF and permalloy layers of 28 and 5 µm respectively.

31.4

Fabrication of Piezoelectric Composite Cantilevers

Even though PVDF can potentially be used in many microactuator applications, i t i s s till a c hallenge f or M EMS re searchers to pattern and form the polymer foils into the required shapes and sizes with high accuracy. Ā e present work demonstrates a novel route for this purpose using laser micromachining, electroplating, and punching (microembossing) techniques.

10/3/2008 9:08:38 PM

31-4

Smart Materials

31.4.1 Excimer Laser Ablation Excimer l aser m icromachining h as b een u sed a s a to ol f or machining and/or patterning many polymers or their layers on different sub strates. L aser m icromachining h as a lready b een demonstrated as an excellent tool for patterning different photoresists, p olyimides, p olycarbonate, PET, a nd si milar m aterials. Ā e a vailability o f a n e xcimer l aser ba sed m icromachining system (Exitech Ltd., United Kingdom) has prompted us to explore t he p ossibility o f app lying t his me thod f or m achining the cantilevers using PVDF material. A typical machined pattern in these experiments is shown in Figure 3 1.5. Ā e s quare h ole i s app roximately 3 00 × 3 00 µm, machined u sing 6 4 s hots at a l aser  uence o f 1 135 mJ/cm2. Unfortunately, the edges turned out to be quite rough and could not be improved signicantly, at a range of laser machining conditions tried. As the number of shots increased, the edges became worse. Ā is m ay m ainly b e at tributed to t he l aser a bsorption properties ( @ 2 48 nm) a nd t he t hermal c haracteristics o f t he PVDF polymer. Ā is led to t he conclusion t hat t he d irect excimer laser micromachining (@ 248 nm) is not suitable for machining or patterning PVDF polymer foils.

31.4.2

Punching of the PVDF Cantilever

Punching is one of the most commonly used processing methods for thermoplastics [30,31]. In conventional punching, a pair of p reheated m ale a nd f emale p unch to ols i s u sed to c ut t he polymer by shear forces. A polarized piezoelectric polymer can be exposed to no more than its Curie temperature (70–90°C for PVDF), beyond which the material loses its piezoelectric characteristics. Ā erefore, c onventional pu nching pr ocess c annot b e directly used in this case. Instead of using male and female plug tools, a nickel shim with the required punch pro le was used to cut t he PV DF p olymer c antilever i n a h ome-made emb ossing system. Ā is fac ilitated a l ow-temperature p unching p rocess, forming the designed pro le of the piezoelectric polymer cantilever w ell b elow i ts Cu rie tem perature. Ā e n ickel s him w as

450 µm

FIGURE 3 1.5 SEM images of metallized PVDF patterned with an excimer laser at energy density of 1135 mJ/cm 2.

43722_C031.indd 4

Pressure

Al plate Insert PVDF Shim Al Plate

FIGURE 31.6

Hot embossing system used for the punching process.

fabricated u sing l aser m icromachining o f a mold i n a g iven polymer substrate and subsequent electroforming. Ā e punching of t he c antilever f rom a PV DF polymer sheet was p erformed u sing a m icroembossing s ystem at ro om tem perature. Ā e procedure used for t his purpose is similar to t he embossing technique described by Becker et a l. A n ickel shim, PVDF  lm, a nd t he i nsert w ere  xed i n a s andwich s tructure between t he t wo a luminum p lates s hown i n F igure 31.6. Ā is arrangement i s en closed i n a v acuum c hamber i nside t he h ot embossing system. A hydraulic pump was used to apply the force on the chamber. A force of 0.27 bar/mm2 was applied onto the sample for one minute. PVDF cantilevers with the designed shape were punched out along the PVDF stretch direction.

31.4.3

Electroplating Permalloy on the Metallized PVDF

As t he c omposite c antilever re quires a no npiezo l ayer o n t he PVDF cantilever structure, the next logical step would be to metallize, the PVDF cantilever to form the required bilayer. Permalloy is chosen for this purpose to b uild a hybrid actuator combining piezoelectric a nd ele ctromagnetic p rinciples. I n t his c ase, t he permalloy serves the purpose of a soft magnetic layer, which can assist in providing additional force. It is deposited by employing electroplating te chniques. Ā e bat h c omposition u sed f or t his purpose is presented by Ref. [32]. A basic sulfate bath with a pH of 3.5 was used. Ā e electroplating was carried out at ro om temperature (25°C) at a current density of 5 mA/cm2. With a plating rate of about 6 µm/h, it takes less than an hour to plate 5 µm.

31.5

Testing and Evaluation of Cantilevers

Ā e testing and evaluation of the fabricated cantilevers consisted of two parts. Āe rst part of the testing used a bimorph piezocantilever to determine the piezocoefficient, which ensured that the value used for t he PVDF is realistic. Ā e second pa rt of t he evaluation mainly compared t he de ection obtained by t he a nalytical a nd numerical (FEA simulation) methods with those obtained experimentally using a fabricated unimorph composite cantilever.

10/3/2008 9:08:39 PM

31-5

Piezoelectric Polymer PVDF Microactuators

31.5.1 Determination of Piezocoefficient of PVDF Using a Bimorph Cantilever Ā e piezocoefficient of t he PV DF polymer was veried using a bimorph c antilever h aving a leng th o f 6 mm a nd a w idth o f 1 mm. Ā e bimorph constitutes two PVDF punched cantilevers from a 28 µm thick PVDF sheet and attached together. When an AC vol tage w as app lied to t his te st s tructure, t he b imorph vibrated like a fan. Ā e measured resonance frequency was found to b e a bout 6 00 Hz. Ā e t ip de ections o f t he c antilever w ere measured u nder a n opt ical m icroscope. Ā e c antilever w as excited by an oscillator. Ā e variation of tip deection with the applied vol tage w as me asured i n t his e xperiment i s s hown i n Figure 31.7. Ā e bottom curve gives the tip deection of the cantilever a s a f unction o f t he app lied vol tage at 5 0 Hz, w hile t he upper p ink c urve i ndicates t he de ection w hen e xciting a t i ts resonance frequency. It maybe noted that the deection at 50 Hz excitation is similar to the values obtained with pure DC excitation. Due to a Q-value larger than 1, the amplitude at t he resonance frequency is about ve times higher for small voltages. At high voltages, the amplitude does not show a linear dependence any more, which may be due to a shift of the resonance frequency for h igh de ections. W hen u sing t he l inear de pendence of t he deection o n t he vol tage i n t he lo w-frequency c ase, t he p iezo strain coefficient d31 of the PVDF can be calculated using the following equation: 3 l2 d = d 31 2 V 4 t Where l is the cantilever length t is the thickness of a single PVDF sheet V is the applied voltage By substituting t he deection ob served i n t he a bove e quation, d31 was obtained by equation as 25 × 10−12 m/V, which is exactly the same value as the one used in the simulation. Voltage vs. tip deflection of bimorph PVDF 80

Deflection (µm)

70 60 50

Testing of Unimorph Cantilever

Ā e dimensions of the unimorph cantilever used in this experiment were 6 × 1 mm w ith a PV DF s heet t hickness of 2 8 µm having a 5 µm thick electroplated permalloy layer on one side. Ā e cantilever was actuated by a DC voltage in the range from 60 to 180 V. Ā e tip deection of the cantilever was observed in each c ase u nder a n opt ical m icroscope mo unted w ith a micrometer. Ā ree s amples w ith t he s ame no minal d imensions were measured to ensure the repeatability of the results. Ā e deections observed with each of these beams are shown in Table 31.1. To ensure the reproducibility of the readings, deections were measured i n e ach c ase at le ast t wice (switching t he voltage off and on) and the range of the values obtained is provided in Table 31.1. It may be noted that the observed deection increases four times when t he applied voltage is tripled. A lso t he values f rom beam to beam were very consistent. Ā e variation of the average tip deection as a function of the applied voltage is plotted i n Figure 31.8, a long w ith t he deection v alues c alculated a nalytically a nd u sing  nite element method ( FEM)s a nalysis f or a b eam leng th o f 6 mm, w idth o f 1 mm, a nd t hicknesses o f PV DF a nd p ermalloy 2 8 a nd 5 µm, respectively. In general, the deection of t he cantilever is linearly proportional to the applied voltage. Ā ese results showed that the values of the deection obtained through analytical calculations are in excellent agreement with the numerical values obtained through CoventorWare si mulation u sing  nite el ement m ethods. However, the deections observed experimentally were 20% less than the values obtained through analytical numerical computations. Ā i s difference between the experimental and computed values maybe attributed to a combination of uncertainties in the thickness me asurement, no nuniformity o f t he ele ctroplated permalloy layer, residual stresses w ithin t he electroplated permalloy  lm, a nd a d ifference i n t he Young’s mo dulus o f t he electroplated p ermalloy c ompared to t he b ulk v alue, w hich i s used in the simulation. Ā e ele ctroplated l ayers a re, i n gener al, k nown to c ontain residual stress essentially due to the electrochemical kinematics occurring during the plating process [33]. Ā e plating parameters suc h a s c urrent den sity, bat h c oncentration, p H, a nd temperature a re k nown to c ontribute to t he re sidual s tress

40 30 20

TABLE 31.1 Measurement of Tip Deections of a Composite Cantilever (µm)

10 0 0

10

20 30 Voltage (V)

Tip deflection at 50 Hz sin wave

40

50

Resonance vibration amplitude

FIGURE 31.7 Bimorph cantilever deection as a f unction of applied voltage for different frequencies.

43722_C031.indd 5

31.5.2

Potential (V)

Beam 1

Beam 2

Beam 3

60 100 120 150 180

35–40 90–100 105–115 130–135 145–155

40–45 85–95 100–110 130–145 150–160

40–45 90–95 105–115 135–140 150–160

10/3/2008 9:08:40 PM

31-6

Smart Materials

200

Beam deflection (µm)

180 160 140 120 100 80

Testing

60

Simulation

40

Theory

20 0 0

FIGURE 31.8

50

100 Applied voltage (V)

150

200

Comparison of experiment and simulation.

developed during the deposition process. In addition, residual stress i n ele ctroplated permalloy  lms is a lso sensitive to t heir iron content, beyond certain values of the nickel to iron ratio. Different te chniques h ave b een u sed to me asure t he i nternal stress during electroplating [33,34]. However, it is hard to nd a process variable t hat does not i nuence internal stresses of t he deposited layers. In this work, we found that current density is a critical parameter with respect to the reduction of residual stress. As the current density increased the plating rate also increased at the same time increasing the surface internal stress. Ā er efore, in order to minimize the stresses a lower current density (5 mA/ cm2) was used instead of 10 mA/cm2 during the electroplating of the permalloy layer. However, it is extremely difficult to deposit completely st ress-free  lms. Ā ese re sidual s tresses w ithin t he electroformed permalloy layer might have contributed to s ome extent t o t he bea m d eection o bserved i n t he ex periment. It m aybe note d t hat b oth t he si mulation a nd t he a nalytical calculations assumed a stress-free metallic layer. Ā e Young’s modulus of the nonpiezo layer used in the simulation m ight b e a nother p ossible c ause f or t he d ifference in deection observed in the experiment. Ā e Young’s modulus of the nonpiezo layer used in the simulation is mainly a bulk value for permalloy w ith a no minal composition of 81% Ni a nd 19% iron. Ā e Young’s modulus of the electroplated could be different from the bulk value reported in the literature. All of these factors might have together contributed to t he deviation of the experimental cantilever deection from the simulation. Ā i s certainly requires f urther i nvestigations to de termine a nd iden tify t he specic reasons for such a large difference.

31.6 Summary A p iezoelectric p olymer c omposite u nimorph c antilever w as designed a nd optimized by  nite element method modeling. Ā e u nimorph c antilever c onsists o f a p iezoelectric p olymer PVDF a nd a no npiezo el astic l ayer. Ā e effect o f va rious parameters o n t he t ip de ection o f t he c antilever w as a nalyzed. Ā ese studies showed that an increase in the thickness

43722_C031.indd 6

of t he p iezoelectric l ayer re duced t he t ip de ections o f th e composite cantilever. Ā ickness of the nonpiezo layer affected the t ip de ection d ramatically. I n o rder to app ly t his u nimorph c antilever a s a c omponent o f a l ater h ybrid ac tuator, which c ombines ele ctromagnetic ac tuation a long w ith t he piezoelectric cantilever, permalloy was chosen as the material for the nonpiezo layer of the unimorph cantilever. Considering the a bove re sults a nd t he d ifferent a spects o f f abrication, a 28 µm t hick PV DF l ayer w ith a 5 µm p ermalloy h as b een selected f or t his w ork. Ā e t ip de ection of the c antilever i s found to be proportional to the square of its length. Ā e details of the fabrication routes followed in making the PVDF cantilevers are described. Two types of cantilevers, unimorph a nd bimorph, were f abricated a nd te sted. Ā e cantilevers we re pu nched u sing e lectroformed n ickel s hims, w hich was l aser m icromachined. Ā e b imorph c antilevers w ere formed b y at taching t wo PV DF c antilever toge ther, w hereas the unimorph cantilevers constituted a single PVDF cantilever with o ne side h aving a 5 µm t hick ele ctroformed p ermalloy layer. Ā e effect of microstructures on the deection of cantilever b eams w as a lso i nvestigated b y f abricating t hem u sing microembossing techniques. Ā e unimorph composite cantilever has generated a 70 µm tip deection at an excitation voltage of 1 00 V. Ā e me asured re sults o f t he c antilever h ave s hown reasonable agreement with the values obtained by analytical and FEA simulation. Punching is a promising way to shape and fabricate polymer lms, which are exible and sensitive to temperature. Furthermore, it can be used favorably to s hape a su rface microstructure on a PVDF polymer lm.

References 1. Fan, L., Tai, Y., and Muller, R .S., IC-processed ele ctrostatic micromotors, Sensors and Actuators 20: 41–47, 1989. 2. Fujita, H., Microactuators and micromachines, Proceedings of the IEEE 86: 1721–1732, 1998. 3. Robertson, C., B imorph dri vers-an ele ctric s olution, Pennwalt (Europe) 79: 1–8, 1979. 4. Todals, M. et al ., A ne w ele ctromotional de vice, RCA Engineer 25: 24–27, 1979. 5. DeVoe, D.L., Piezoelectric thin lm micromechanical beam resonators, Sensors and Actuators A 88: 263–272, 2001. 6. Elvin, N.G., El vin, A.A., a nd S pector, M., A s elf-powered mechanical st rain ener gy s ensor, Smart M aterials a nd Structures 10: 293–299, 2001. 7. Kueppers, H. and Leuerer, T., PZT thin lms for piezoelectric microactuator applications, Sensors and Actuators A 97–98: 680–684, 2002. 8. Zurn, S. et al., Fabrication and structural characterization of a resonant frequency PZT microcantilevers, Smart Materials and Structures 10: 252–263, 2001. 9. Lau, S.T. et al ., P iezoelectric co mposite h ydrophone a rray, Sensors and Actuators A 3145: 1–7, 2001.

10/3/2008 9:08:41 PM

Piezoelectric Polymer PVDF Microactuators

10. Lee, G. a nd Chen, S.H., Microfabricated p lastic c hips by hot embossing methods and their applications for DNA separation and detection, Sensors and Actuators B 75: 142–148, 2001. 11. Schellin, R. and Hess, G., A silicon subminiature microphone based o n p iezoresistive p olysilicon st rain ga uges, Sensors and Actuators A 32: 555–559, 1992. 12. Roberts, D .C. a nd Li , H., A p iezoelectric micr ovalve f or compact high-frequency, high-differential pressure hydraulic micropumping syst ems, Journal o f M icroelectromechanical Systems 12: 81–92, 2003. 13. Yoseph, B .C. a nd Cha ng, Z., P iezoelectrically ac tuated miniature peristaltic pump, Proceedings of SPIE’s 7th Annual I nternational S ymposium o n S mart S tructures a nd Materials, Newport, March 1–5, 2000. 14. Dargahi, J., Parameswaran, M., and Payandeh, S., A micromachined p iezoelectric ta ctile se nsor f or a n e ndoscopic grasper—theory, fa brication a nd exp eriments, Journal o f Microelectromechanical Systems 9: 329–335, 2000. 15. Li, L.X. and Shen, Y.P., Ā e optimal design of piezoelectric actuators f or aco ustic co ntrol, Smart M aterials and Structures 10: 421–426, 2001. 16. Ruprecht, R. et al ., Molding of LIGA microstructures from uorinated polymers, Microsystem Technologies 2: 182–185, 1996. 17. Jenkins, D.F.L. et al., Ā e use of sputtered ZnO piezoelectric thin  lms as b road-band micr oactuators, Sensors a nd Actuators A 63: 135–139, 1997. 18. Friese, C., Goldschmidtboing, F., and Woias, P., Piezoelectric microactuators in p olymer-composite te chnology, Pr oceedings of Transducers’03, Ā e 12th I nternational Conference on Solid State Sensors, Actuators and Microsystems, Boston, MA, June 8–12, pp. 1007–1010, 2003. 19. Sun, D. and Mills, J.K., Control of a rotating cantilever beam using a torque actuator and a distributed piezoelectric polymer actuator, Applied Acoustics 63: 885–899, 2002. 20. Xu, T.B., Cheng, Z.Y., and Zhang, Q.M., High performance micromachined unimorph actuators based on electrostrictive poly (vin ylidene  uoride-triuoroethylene) copolymer, Applied Physics Letters 80, 2002. 21. www.msiusa.com 22. Izumi, Y. et al ., Irradiation eff ects of excimer laser light on poly(vinylidene  uoride) (PVD F)  lm, Bulletin o f t he Chemical Society of Japan 71: 2721–2725, 1998.

43722_C031.indd 7

31-7

23. Katan, E., Narkis, M., and Siegmann, A., Ā e effect of some uoropolymers’ structures on their response to UV irradiation, Journal o f Applied P olymer S cience 70: 1471–1481, 1998. 24. M anohara, H.M. a nd M orikawa, E., P attern tra nsfer b y direct pho to etchin g o f p oly(vinylidene  uoride) using x-rays, Journal of Microelectromechanical Systems 8: 417– 422, 1999. 25. Duca, M.D., Plosceanu, C. L., and Pop, T., Effect of x-rays on poly(vinylidene  uoride) in x-ra y p hotoelectron sp ectroscopy, Journal o f Applied P olymer S cience 67: 2125–2129, 1998. 26. Ruprecht, R. et al ., Molding of LIGA microstructures from uorinated polymers, Microsystem T echnologies 2: 182– 185, 1996. 27. Fu, Y., H arvey, E.C., Gha ntasala, M.K., a nd S pinks, G.M., Design, fa brication a nd t esting o f p iezoelectric p olymer PVDF micr oactuators, Smart Materials a nd S tructures 15: S141–S146, 2006. 28. Fujitsuka, N., et al., Monolithic pyroelectric infrared image sensor using PVDF thin  lm, Sensors and Actuators A 66: 237–243, 1998. 29. Khoo, M. a nd Li u, C., Micr o ma gnetic silico ne elast omer membrane ac tuator, Sensors and Actuators A 89: 259–266, 2001. 30. Becker, H. a nd Ga rtner, C., P olymer micr ofabrication methods f or micr oudic a nalytical a pplications, Electrophoresis 21: 12–26, 2000. 31. Simdekova, I. et al., A study of hot embossed microchannels using confocal microscopy, Proceedings of SPIE Nano- and Microtechnology: M aterials, Pr ocesses, P ackaging, and Systems, Melbourne, Australia, Dec. 2002, pp. 82–92. 32. Ghantasala, M.K. et al ., Design and fabrication of a micr o magnetic b earing, Smart Materials a nd S tructures 9: 235– 240, 2000. 33. Stein, B ., A p ractical guide to under standing, me asuring and c ontrolling st ress in ele ctroformed met als, AESF Electroforming S ymposium, L as Vegas, M arch 27–29, 1996. 34. Hyoung, J.C. and Ahn, H.C., Magnetically-driven bi-directional opt ical m icroscanner, Journal of M icromechanics and Microengineering 13: 383–389, 2001.

10/3/2008 9:08:41 PM

43722_C031.indd 8

10/3/2008 9:08:41 PM

32 Ultrasonic Nondestructive Testing and Materials Characterization

John A. Brunk National Nuclear Security Agency’s Kansas City Plant

32.1 Scope and Purpose ..............................................................................................................32-1 32.2 Sound Frequency Range for Materials Characterization and Testing ..........................32-1 32.3 Flaw Detection Practices ....................................................................................................32-2 32.4 Flaw Detection References .................................................................................................32-2 32.5 Basic Equations ....................................................................................................................32-2 32.6 Ultrasonic Velocity and Attenuation Measurements ......................................................32-3 32.7 Sources of Ultrasound ........................................................................................................32-3 32.8 Couplants..............................................................................................................................32-3 32.9 Complications ......................................................................................................................32-4 32.10 S implications ......................................................................................................................32-4 32.11 Example 1: Real Defects Used to Establish and NDT Procedure ..................................32-4 32.12 Example 2: Material Property Variation Shown during “Flaw Detection” ...................32-6 32.13 Example 3: Filtering to Separate Discontinuities from Material Variations ................32-7 32.14 C onclusion............................................................................................................................32-9 References .........................................................................................................................................32-10

32.1 Scope and Purpose Sound a nd u ltrasound have a t remendous variety of i ndustrial and s cientic appl ications i n m aterials c haracterization,  aw detection, and measurement. Although the same basic principles always apply, there are huge differences in the practical details. There is continuous improvement in equipment, and new techniques are always being developed to the point of being practical for routine use. In 2006, it is impossible to g ive a broad view of the te chnology t hat w ill b e up -to-date f or e ven a f ew mo nths. There are many good sources of basic information, such as Ā e NDT H andbook [ 1] R ather t han d uplicate a vailable t raining material, t his c hapter w ill de scribe a f ew t hings t hat c an b e accomplished w ith u ltrasound, w ith s ome e xplanation o f w hy they need to be worked with some caution. Some differences between  aw de tection a nd material cha racterization methods are described. Ultrasonic evaluation is only one of many nondestructive methods for these applications. It is sometimes necessary to utilize two or more to obtain all of the desired information. An e xample o f u sing x -ray r adiography a s a c omplementary technique is included.

32.2 Sound Frequency Range for Materials Characterization and Testing One denition of sound is mechanical vibrations transmitted by an elastic medium. We generally think of sound as what we can hear, from 15 or 20 Hz or cycles/second up to as high as 20,000 Hz (20 k Hz) f or no rmal yo ung p eople. The upp er l imit ten ds to decrease with age, more for men than for women. A middle-aged person may have an upper limit of 12 to 14 kHz. “Ultrasound,” above the frequency of human hearing, is generally considered to be anything above about 20 kHz. Applications for the lower frequencies are limited or indicated by the fact that some animals hear (or transmit and detect) higher frequencies t han h umans. These i nclude dol phins, p orpoises, whales, bats, and some species of frogs. Ultrasonic burglar alarms that op erate i n t he 2 0–30 kHz r ange a re not su itable for homes with dogs . E lephants c ommunicate u sing i nfrasound (frequencies too low for humans to hear) down to as low as about 14 Hz. Frequencies from as low as 3 kHz (sonic) up to at le ast 250 MHz are used for aw detection, material property evaluation, measuring, 32-1

43722_C032.indd 1

10/3/2008 9:08:58 PM

32-2

and process monitoring. The most common industrial application maybe testing structural welds using frequencies from 1 to 5 MHz.

32.3 Flaw Detection Practices Ultrasonic  aw detection is required by many industrial codes and specications, such as the ASME Boiler and Pressure Vessel Code [ 2] a nd t he A merican W elding S ociety’s D 1.1/D1.1M Structural W elding Co de-Steel [ 3]. These do cuments p rovide inspection a nd a cceptance cr iteria f or s pecic p roduct f orms. They also include requirements for training, experience, examination, a nd c ertication o f perso nnel wh o per form u ltrasonic and other nondestructive tests. Personnel requirements are made by reference to other publications. In the United States, these are usually SNT-TC-1A Personnel Qualication and Certication in Nondestructive T esting [ 4] o r N AS 4 10, N AS C ertication & Qualication of Nondestructive Test Personnel [5]. Inspector qualication is beyond the scope of this discussion. More i nformation c an b e ob tained f rom t he re ferenced do cuments and their publishers. Appropriate and sufficient training is essential for successful evaluation of product forms and materials to t heir specications. However, training directed toward specic types of test objects does not necessarily cover information important for other applications.

32.4 Flaw Detection References Flaw detection to determine whether a product or material meets acceptance criteria requires some denitions before any testing is initiated: 1. “Flaw” must be dened. “ Discontinuity” is a t erm o ften used for an inhomogeneity that has been detected and must be evaluated in accordance with requirements. Denitions m aybe gener al suc h a s “ linear i ndications” and “rounded i ndications” or s pecic suc h a s (in w elds) “slag inclusions” and “lack of fusion.” 2. Smallest discontinuity or test signal to be evaluated must be stated. This maybe an actual dimension, but in ultrasonic testing, it is usually an indication that gives a certain signal amplitude compared to a reference reector. 3. Portion of the test object to be evaluated should be stated. This maybe the entire volume, or only part, such as what will remain after subsequent machining. Most references are not real  aws. One reason for this is a need for standardization. No two real  aws would be alike. The most common r eferences u sed to si mulate def ects a re si de-drilled holes (with the sidewall of the hole as the reector) and at-bottom h oles ( with t he  at hole b ottom a s t he re ector). Others include v-notches a nd u-notches in t he material surface. Many specications dene reference requirements [6]. Some applications require special references to provide a reasonably re alistic app roximation o f de fects t hat a re o f c oncern. There a re s pecications c overing f abrication o f re ferences i n solid composite laminates [7] and in composite honeycombs [8].

43722_C032.indd 2

Smart Materials

An important limitation to real aws is that they often can be fully characterized only by cross-sectioning or some other destructive means, leaving nothing to b e used as a re ference in the future. There are exceptions, such as in Example 1 below.

32.5 Basic Equations Ultrasound can be used to determine variations in bulk properties or localized changes within a material sample. Wave propagation i s rel ated to t he el astic p roperties o f t he p ropagating medium. Longitudinal waves are those where particle displacement is in the direction of propagation. They may exist in solids, liquids, and gases. Shear o r t ransverse w aves a re t hose w here pa rticle mot ion i s transverse to the direction of propagation. They may exist in solids, and in a few very viscous liquids, but not in gases. For bulk materials (when t he d imensions o f t he m aterial a re m any t imes l arger than the wavelength of the ultrasound) the relationships are Young’s Modulus of Elasticity, E = [rvs2 (3v l2 − 4vs2)]/(v l2 −vs2) where r = density, kg/m3 [or lb/in.3] v l = longitudinal velocity, m/s [or in./s] vs = transverse velocity, m/s [or in./s] E = Young’s modulus of elasticity, N/m2 [or lb/in.2] Poisson’s ratio, s = [1 − 2(vs/v l)2/2[1 − (vs/v l)2] where s = Poisson’s ratio vs = ultrasonic transverse velocity, m/s [or in./s] v l = ultrasonic longitudinal velocity, m/s [or in./s] Shear modulus, G = rvs2 Bulk modulus, K = r[v l2 − (4/3)vs2] Surface w aves m ay t ravel o n t he su rfaces o f b ulk m aterials, and also may exist in thin layers. There are several types of surfaces waves, but the type most often used in ultrasonic testing is a Rayleigh wave. These have an elliptical particle motion in the vertical plane. Their velocity is typically about 90% of the longitudinal wave velocity in the material. From these relationships, it can be seen that measurements of ultrasonic velocity can be used to indicate variations in material properties. A very important parameter affecting the transmission and reection of ultrasound at boundaries where material properties change is “acoustic impedance.” This is the product of density and wave velocity: Acoustic impedance, z = rv l where z is the acoustic impedance (kg/m2 s [or lb/in.2 s]). The percentage of incident power that is reected f rom a nd transmitted across a boundary is given by

10/3/2008 9:08:58 PM

32-3

Ultrasonic Nondestructive Testing and Materials Characterization

Reection coefficient for energy (R): R = (Z2 − Z1)2/(Z2 + Z1)2

32.8 Couplants

where Z1 is the acoustic impedance in medium 1 Z2 is the acoustic impedance in medium 2

Some sort of material must be used between the transducer face and the entry surface of the test object. Most contact transducers have hard surface wear plates. The purpose of a couplant is to reduce the acoustic attenuation mismatch that would exist with an air gap (although air-coupled testing is de nitely possible f or s ome app lications). Su itable c ouplants f or mo st c ontact tests include light-grade oil a nd g lycerine. For t ransverse wave te sting, t here a re t wo t ypes o f me thods. D irect c ontact shear w ave t ransducers p roduce a t ransverse w ave t raveling perpendicular to the transducer face. Th is is the same path followed b y a lo ngitudinal w ave f rom t he s ame s ample su rface location. The best practice is to have both beams follow the same pat h, especially for a ny materials t hat may be a nisotropic. Transverse waves may be coupled with high-viscosity liquids such as honey, or with slightly compressible solids. For either type of wave, liquid intrusion into porous materials must be avoided. Dry couplings may be an integral part of the transducer (instead of a hard face) [15], and air-coupling is possible for some materials [16]. Transverse waves may also be produced by refraction. When the angle of incidence at the test surface is not zero (the center of the incoming beam is not perpendicular to t he test surface) the direction o f u ltrasound en tering t he te st m aterial i s g iven b y Snell’s Law:

Transmission coefficient for energy (T): T = (4Z2Z1)/(Z2 + Z1)2 The change in acoustic impedance at a boundary allows detection of discrete aws suc h a s ga s p orosity o r c racks i n me tals, measurement o f l ayers suc h a s s tainless s teel w eld c ladding applied to lo w-alloy s teel f or c orrosion re sistance, a nd ot her types of discrete material changes within test objects.

32.6 Ultrasonic Velocity and Attenuation Measurements There is an ASTM document covering ultrasonic velocity measurements w ith c onventional pu lse-echo u ltrasonic  aw detectors [9]. It also includes methods that require more specialized equipment and can produce more accurate measurements. Ther e are other documents covering specialized methods; for example, for measuring velocity and measuring attenuation in advanced ceramics [10,11].

32.7

Sources of Ultrasound

Most u ltrasonic t ransducers u se piezoelectric elements to c onvert e lectrical i mpulses i nto me chanical v ibrations a nd v iceversa. Re ference [ 1] i ncludes a d iscussion o f p iezoelectric transducers a nd a lso o f a lternatives suc h a s ele ctromagnetic acoustic transducers and laser-generated ultrasound. In general, piezoelectric transducers used with immersion techniques or in contact with some sort of liquid coupling are used unless there is some special reason to do otherwise. For example, when it is desired to test metal while it is very hot or when a material is so porous, its properties would be altered by liquid intrusion [12]. Piezoelectric transducers are made from several types of materials a nd constructed in various ways for different applications. A good transducer catalog is full of useful information, and some manufacturers can provide a great deal of assistance in selecting the right transducer for a specic application. This is too complex a subject to try to cover here. In Example 1, a set of self-coupling contact t ransducers w as e specially m ade for t he application. I n Example 2, it can be seen that two different immersion transducers produced very different signals from a steel sample. Some re cent de velopments i nclude h igh-efficiency air/gas coupled t ransducers t hat c an el iminate t he ne ed for a ny ot her couplant for s ome appl ications [ 13] a nd ph ased a rrays w here selective s equential pulsing of individual elements in a n arr ay can be used to “steer” the resulting ultrasound beam [14]. NDT. net is a n excellent resource for following new de velopments i n ultrasonic a nd ot her N DT me thods. I t g ives ac cess to pap ers from many technical conferences.

43722_C032.indd 3

sin f sin f

=

VLW VM

q Water or plastic wedge

Test material f where q f VLM VSM

= angle of incident beam = angle of refracted beam = velocity of longitudinal waves in water or wedge = velocity of shear or longitudinal waves in material under examination

The shear wave angle is always smaller than the longitudinal wave angle. When the calculated angle increases to 90°, the longitudinal wave becomes a surface wave. By varying the incident angle, d ata c an b e t aken i n m ultiple d irections. Th is is ea sier with w ater c oupling c ompared to u sing a n umber o f p lastic wedges giving different angles. Water coupling is also more convenient for scanning large test objects and collecting data during the scan. Scanners are available for contact testing. W hen using longitudinal waves w ith a thin l ayer o f l iquid c ouplant, ac tual s canning i s p ossible. F or shear w ave t ransducers re quiring a v ery v iscous c ouplant, a nd with se lf-coupling t ransducers, t he t ransducer must be picked up and put down for each data point.

10/3/2008 9:08:58 PM

32-4

Smart Materials

32.9 Complications Characterization o f m aterials m ust b e a pproached w ith ca ution because the most readily measured quantities, velocity, and attenuation, are functions of several variables. In addition, the relationships to variables change under different conditions. For example, “all else being equal” the attenuation of ultrasound passing through a particular material increases with increasing frequency. That is, higher f requencies a re at tenuated more t han lower. But t he rel ationship between attenuation and frequency is not a simple one. In a polycrystalline material, where the wavelength of ultrasound is much greater than the average grain diameter, attenuation is proportional to t he fourth power of t he f requency. W hen t he wavelength is approximately equal to the grain diameter attenuation is proportional to the square of the frequency. In between these two, there is a continuous change. To further complicate matters, “test frequency” is not a single value but a frequency range. Transducers may p roduce b roader o r na rrow ba ndwidths ac cording to t heir construction and how they are excited, and receivers may be broadbanded or tuned. In many cases, the received signal has lost relatively mo re o f i ts h igher f requency c omponents. B eam s pread i s also a pa rt of total attenuation, and this is dependent upon transducer si ze a nd f requency a s w ell a s m aterial p roperties. A v ery good discussion of the complexities of ultrasonic materials characterization is given by Papadakis [17]. Ultrasound can still be a very valuable tool for characterizing or comparing materials. Attenuation as a function of frequency can be compared for s amples of t he s ame a lloy w ith d ifferent g rain si zes even t hough t here i s not o ne si mple c orrelation f unction f or t he entire range of sizes studied. Ultrasonic measurements are effectively used for a very wide assortment of applications. To name just a few: • • • •

Characterization of processes for freezing foods [18] Evaluation of cheese as it matures [19] Characterization of graphite particles in cast iron [20] Evaluation o f den sity v ariations i n g reen a nd si ntered ceramics [21] • Monitoring g rain si ze w hile p roducing s eamless me tal tubing [22] It is important to note that material characterizations are comparative, just as are aw detection methods. That is, quantitative re sults a re p roduced o nly w hen s amples a re c ompared with materials with known properties determined by some other method.

32.10

Simplifications

Useful app lications re quire l imiting t he n umbers o f v ariables and/or unknowns in the test objects in addition to proper selection of equipment. Because mathematical models can become extremely c omplicated a nd m ust t ake s ome v ariables a s c onstants, they can at b est give the investigator a go od start in the right direction. It is also important to avoid adding unexpected variables with the equipment used. It is well understood that ultrasonic attenuation and grain scattering in metals is in uenced b y g rain si ze a nd s hape, a nd w ill

43722_C032.indd 4

vary w ith t he p ropagation d irection i n s amples t hat h ave b een worked in a m anner that produces a p referred grain orientation. Anisotropy m ay a lso o ccur i n m aterials w here i t m ight not b e expected, such as in blocks of polystyrene. Whenever possible, it is a good idea to prepare test references from material that is as close as possible to the samples to be evaluated, and to consider that ultrasonic properties may be directional. Howard and Enzukewich reported an excellent practical example of dealing with the effects of microstructure on ultrasonic attenuation of alloy steels such as 4330 and 4340 [23]. Normalizing a hardened alloy steel reference was found to effectively reduce ultrasonic attenuation and “noise” to about that from fully annealed material to be tested. AWS D1.1/D1.1M covers ultrasonic examination of we lds in a limited range of structural steels. Material characterization is not an issue, so several test simplications are possible. Flaw detectors are required to have certain operating characteristics. Dimensions and nominal frequencies of transducers are prescribed. Standard test blocks are dened for d istance a nd b eam a ngle c alibration. The s ame blocks have holes to p rovide reference sig nals for d iscontinuity e valuation. W ith s pecic m aterial a nd t ransducer characteristics, attenuation of ultrasound by the material is taken to b e a c onstant 2 dB/in. a fter t he  rst i nch. O nly one re ference hole is used, and this attenuation factor is used to compensate for actual m etal pa ths. The intent is that everyone using approved equipment and procedures will be able to accept welds that meet code requirements and reject any indications that do not. The ASME Boiler and Pressure Vessel Code covers a much wider range of materials a nd t he requirements for u ltrasonic examinations apply to lo w-alloy steels, various stainless steels, a nd a lloys such as Monel and Inconel. Calibration blocks are required to be of the s ame material to b e tested a nd where applicable i n t he s ame heat treat condition. Instead of calculating equivalent hole signal amplitudes for d ifferent d istances, s everal side -drilled h oles a re used to de termine t he d istance–amplitude r elationship f or t he same size hole in different locations. There is more latitude in the choice of transducers, but the distance–amplitude relationship must be d etermined f or ea ch t ransducer u sed. The appropriate setup for material characterization can be quite different from one optimized for aw detection When straightforward aw detection procedures a re p erformed a s s pecied, t here c an b e t imes w hen material p roperty c hanges ot her t han t he “ defects” o f i nterest create unusual results that must be explained.

32.11 Example 1: Real Defects Used to Establish and NDT Procedure This app lication w as u nusual b ecause i t em ployed t hroughtransmission over a path with two 90° corners, and special drycoupling 4 MHz t ransducers. I t w as de veloped to te st sm all U-shaped ferrite cores to de termine i f t hey had c racked i n t he course of being potted and then incorporated into an assembly with a s tainless steel housing. They were ground ush with the steel surface of the housing. After potting, it was still possible to radiograph t he c ores f or c racks, b ut x-ray w as i neffective for the assemblies. Figure 32.1 shows parts before and after potting.

10/3/2008 9:08:59 PM

Ultrasonic Nondestructive Testing and Materials Characterization

32-5

FIGURE 32.3 Less severe crack shown in radiograph.

FIGURE 32.1 Parts before and after potting.

The circular ends are 0.200 in. (5.1 mm) in diameter with a center-to-center spacing of 0.350 in. (8.9 mm). A number of potted cores were radiographed until some with cracks were identied. F igure 32.2 i s a r adiograph t hat s hows a completely fractured leg. Figure 32.3 shows a less severe leg crack

FIGURE 32.2 Radiograph shows fractured leg.

43722_C032.indd 5

near its junction with the bridge between the legs. Cracks in these locations should be detectable by a pulse-echo ultrasonic test with a transducer coupled to the round end of each leg in turn. Figure 32.4 shows a c rack i n t he bridge. This wa s of particular i nterest because t his crack could not b e detected by a si mple pulse-echo test. It was used to develop a through-transmission test to detect a crack in the bridge. A xture was made to hold a pair of dry-coupling contact transducers for a through-transmission test. Transmission

FIGURE 32.4 Cracks in the bridge shown by through-transmission test.

10/3/2008 9:08:59 PM

32-6

Smart Materials Supplier A

Supplier B

5

10

µs

5

10

µs

15

5

10

µs

15

0

5

10

µs

15

5

10

µs

15

5

10

µs

15

Volts

−0.2 0.0 0.2

15

Volts

−0.2 0.0 0.2

0

Volts

−0.2 0.0 0.2

0

FIGURE 32.5 New cores not potted.

through an undamaged core was easy. The reections and redirections between the transmitter to t he receiver would be extremely difficult if not impossible to model. It was sufficient to demonstrate that it could be done with repeatable results in the absence of a defect. To establish repeatability, a number of similar parts from two suppliers w ere e xamined. F igure 3 2.5 s hows s ome t ypical through-transmission time–amplitude displays of cores that had not been potted. Cores from supplier B gave consistent responses. Those f rom supp lier A w ere sub stantially mo re v ariable. They were also found to vary more in density and in other properties of interest. A number of parts were tested after potting and then put into assemblies. Being in assemblies did not signicantly change the ultrasonic response, indicating the cores had not been damaged. Some assemblies were subjected to mechanical shocks that might crack t he c ores a nd te sted a gain. A f ew w ere f ound to h ave cracked. Figure 32.6 shows the difference between cores considered good and the sample shown in Figure 32.4.

32.12

Example 2: Material Property Variation Shown during “Flaw Detection”

A 3 04L stainless steel round ba r stock was te sted w ith a w ater immersion p ulse-echo te chnique w ith t he u ltrasound b eam directed radially into the bar. Reference bars with at-bottom holes (drilled r adially to v arious de pths f rom t he b eam en try

43722_C032.indd 6

surface with at bottoms parallel to a plane tangent to the entry point) a re m ade for e ach lot o f m aterial. D iscontinuity i ndications are compared with a calibration curve of amplitude versus depth for a specic at-bottom hole diameter. The amplitude of the bac k su rface re ection is also monitored. Unfavorably oriented  aws could reect energy so that it is not received by the transducer, b ut s till re duce bac k su rface re ection. References for larger bars are often made with holes drilled no deeper than the radius so the “back half” of the sound path through the bar is tested. The upper part of Figure 32.7 shows a time–amplitude signal typical of most material in a particular lot of 4 in. (102 m) diameter bar. The lower part shows a h igher “noise” signal level and lower bac k re ection t hat a roused s uspicion. The transducer used had been found to b e suitable based on detection of reference holes i n a ba r f rom t he s ame lot . It had a no minal c enter frequency of 5 MHz, 0.75 in. (19 mm) d iameter, spherical beam focus with a focal length in water of 6 in. (152 mm). The suspicious region was reexamined using a transducer with a nominal center frequency of 10 MHz. Compared to the 5 MHz unit, i t h ad a m uch lo wer energ y o utput a nd a m uch b roader bandwidth. It had an element diameter of 0.4 in. (10 mm), spherical focus with a focal length in water of 1.4 in. (35 mm). The resulting display is shown in Figure 32.8. This clearly shows a reector at a de pth where t he lower pa rt of t he previous  gure has what appears to be part of the front surface “noise.” Th is is an example of a large difference in response with a test frequency change. A slab with a leng th of about 0.2 in. (5 mm) was cut from the suspicious part of the bar and scanned with ultrasound directed

10/3/2008 9:09:00 PM

Ultrasonic Nondestructive Testing and Materials Characterization

FIGURE 32.6

Top three are good cores in assemblies. Bottom is the core in Figure 32.4.

in the axial direction. Figure 32.9 shows a plan view plot of the back reection a mplitude. This s can w as m ade w ith a 2 0 MHz transducer to increase amplitude variations. There is a very denite r ing o f lo wer a mplitude d ividing t he s ample i nto w hat appears to be an inner and an outer section. Figure 32.10 is a photograph of the beam entry surface after it was etched to b ring out the grain structure. The original ultrasonic te st w as re sponding to t he w ell-dened c hange i n g rain structure. The original setup for detecting “real” defects, as represented by at- bottom holes, was not capable of indicating this change except by increased “noise” and loss of back reection. If the initial test had utilized a lower test frequency, the condition might not h ave been de tected at a ll. I f a “ cored” structure had been suspected a nd testing for it had been pa rt of t he requirements, the material would probably have been tested at two different frequencies with different transducers. In fact, a uniform grain structure was required for this material.

32.13

Example 3: Filtering to Separate Discontinuities from Material Variations

Figure 32.11 shows responses from a 7.5 in. (190.5 mm) diameter 304 s tainless s teel ba r w ith a c ored g rain s tructure si milar to

43722_C032.indd 7

32-7

the one in the previous example. Longitudinal wave contact tests were do ne w ith a 2 .25 MHz c eramic c omposite t ransducer, 0.375 in. (9.5 mm) in diameter. The upper part of the gure shows the re sponse u sing a s pike p ulser a nd a b roadband ( 0.1 to 10 MHz) amplier. The f ront su rface re ection no ise ob scures anything else within the rst third of the sound path. The lower part of the gure was produced with the same transducer in the same location. The same pulser- receiver was used, but a tunable preamplier/bandpass  lter was added between t he transducer and the receiver input. The  lter was tuned to 2.25 MHz and the preamplier s et to 3 0 dB. The m ain re ceiver ga in w as re duced until the back reection amplitude was the same as before. This reduced t he f ront su rface no ise g reatly, a nd a d iscontinuity within Gate 1 was clearly resolved. Frequency spectra of the back surface re ection indicate that removing frequencies below about 1.5 MHz eliminated most of the front surface noise. Tuning to a f requency m uch h igher t han 4 MHz re sulted i n e xcessive loss of b ack re ection. The bac k su rface re ection s hows t hat sufficient ultrasonic energy did make the round trip through the material. P roper t uning made i t p ossible to de tect si de-drilled 0.020 in. ( 0.5 mm) re ference h oles t hroughout t he m aterial at distances from 0.5 in. (12.7 mm) to 7.0 in. (178 mm). Pulse t uning, u sing a s quare w ave o r to neburst p ulser c an produce si milar re sults b y c ausing a b roadband t ransducer to produce a na rrow f requency s pectrum w ith h igher p ower

10/3/2008 9:09:01 PM

32-8

Smart Materials Ultrasonic signals from “typical” 4 in. diameter material

Ultrasonic signals from suspect material in the same lot

FIGURE 32.7 Time–amplitude signals from typical and suspect material, tested with the usual procedure using a me dium–bandwith 5 M Hz transducer.

Time–amplitude display with ultrasound beam directed radially (normal scan direction). Tested with a highly damped 10 MHz transducer, 0.4 in. (10 mm) diameter, 1.4 in. (35 mm) focal lenght.

FIGURE 32.8

43722_C032.indd 8

Display from same region of suspect material using a 10 MHz broadband transducer.

10/3/2008 9:09:01 PM

Ultrasonic Nondestructive Testing and Materials Characterization

32-9

FIGURE 32.9 Plan view plot.

Small grain region is approximately 2.2 in. diameter

Large grains

compared to w hen a s pike pulser i s u sed. A b roadband t ransducer u sed t his w ay s till h as i ts f ull no rmal ba ndwidth a s a receiver. Pulse tuning and receiver tuning can be used simultaneously. It is important to remember that techniques for removing u nwanted no ise sig nals m ay a lso remo ve i nformation o f interest when the purpose is material characterization instead of aw detection.

32.14 Conclusion

Small grains

FIGURE 32.10 Photograph of beam entry surface.

43722_C032.indd 9

Ultrasound provides a tremendous assortment of useful tools for aw detection and material property characterization. It can be easy to go astray because there is no simple universal tool kit. Fortunately, a very large amount of useful information is readily available. It is important to understand how the characteristics o f a ll t he c omponents o f a te st s ystem c an i nuence the results. These i nclude t ransducer f requency, si ze, d amping characteristics ( bandwidth), a nd f ocusing. P ulser v ariables include t ype (spike, s quare w ave, to neburst), a nd t he n umber and type of controls. For example, spike pulsers may have continuously variable pulse damping, a l imited number of damping s ettings, o r no op erator ad justment f or d amping. S ome have two or more settings for pulse voltage. Receivers may be

10/3/2008 9:09:03 PM

32-10

Smart Materials

40

60

Frequency spectrum of first back reflection

0.0

Gate 1 Gate 2

20 (149.40, 0.00)

[FFT] of str8100 time slot 1: board 1 Volts 0.2 0.4 0.6 0.8

Volts 0.00 0.05 0.10 0.15 0.20

FWR of str8100 time slot 1: board 1

80 100 120 Time (µs) (17.68, 0.00)

5

10 Frequency (MHz)

Broadband amplifier 0.1–10 MHz. Front surface noise exceeds gate 1 level for 20 µs

20 (124.58, 0.00)

Volts 0.2 0.4 0.6 0.8

Gate 1 Gate 2

40

60

[FFT] of str8100 time slot 1: board 1

0.0

Volts 0.00 0.05 0.10 0.15 0.20

FWR of str8100 time slot 1: board 1

80 100 120 Time (µs) (25.00,0.03)

Frequency spectrum of first back reflection

5

10 Frequency (MHz)

Same amplifier with added tunable bandpass filter. Discontinunity at 15 µs is resolved.

FIGURE 32.11 Response in steel bar with cored grain structure.

broadband with ranges up to perhaps 10, 50 or 100 MHz. Tuned reception frequencies may include a limited number of specic passbands, but a tuneable bandpass  lter is often m ore useful. Portable instruments designed for a limited number of applications often h ave fewer adjustments in the interest of reliability and re peatability w ith le ss c hance f or op erator er ror. Ha ving the r ight e quipment i s a s i mportant a s t he r ight k nowledge. Understanding how test object geometry and possible material property variations may interact with ultrasound is essential for understanding test data.

References 1.

2. 3. 4. 5.

43722_C032.indd 10

Ā e NDT Handbook, 2nd ed., Vol. 7 Ultrasonic Testing. 1991 The American S ociety f or N ondestructive T esting, 1711 Arlingate Lane, Columbus, OH 43228-0518. (The third edition is expected in late 2006 or early 2007.) Published b y The American S ociety o f M echanical Engineers, Three Park Avenue, New York, NY 10016-5990. Published by American Welding Society, 550 N.W. LeJeune Rd., Miami, FL 33126. Published b y The American S ociety f or N ondestructive Testing, I nc., 1711 Arlingate L ane, C olumbus, O H 432280518. Published by Aerospace Industries Association of America, Inc., 1250 Eye St. N. W., Washington, DC, 2005.

6. ASTM E 137-06. S tandard P ractice f or F abrication a nd Checking Aluminum Alloy Ultrasonic Standard Reference Blocks. ASTM International, 100 B arr Harbor Drive, West Conshohocken, PA 19428. 7. SAE ARP 5605 “Solid Composite Laminate NDI Reference Standards”, SAE I nternational, 400 C ommonwealth Dri ve, Warrendale, PA 15096. 8. SAE ARP 5606 “Composite H oneycomb ND I Ref erence Standards”, SAE International. 9. ASTM E 494-05.Standard Practice for Measuring Ultrasonic Velocity in Materials. ASTM International, 100 Barr Harbor Drive, West Conshohoken, PA 19428–2959. 10. ASTM C 1331-96. S tandard T est M ethod f or M easuring Ultrasonic Velocity in Advanced Ceramics with Broadband Pulse-Echo Cross-Correlation Method. 11. ASTM C 1332-96. S tandard T est M ethod f or M easuring Ultrasonic Attenuation Coefficients in Advanced Ceramics by Pulse-Echo Contact Techniques. 12. Brunk, J . U ltrasonic exa mination o f p orous ma terials. Materials Evaluation, April 1988. The American Society for Nondestructive Testing. 13. Hillger, W. et al . Inspection of CFRP co mponents by ultraonic ima ging wi th a ir co upling. ND T.net, O ctober 2002, Vol. 7, No. 10. 14. Poguet, J. and Ciorau, P., Reproducibility and reliability of NDT p hased a rray p robes, P aper p resented a t the 16th

10/3/2008 9:09:04 PM

Ultrasonic Nondestructive Testing and Materials Characterization

15.

16.

17.

18.

43722_C032.indd 11

World C onference o f N ondestructive T esting, M ontreal, 2004. Available at NDT.net, Oct 2004, Vol. 9, No. 10. Brunk, J . Applications a nd ad vantages o f s elf-coupling transducers f or ma terials c haracterization a nd in spection, Acousto-Ultrasonics Ā eo ry and Application. J . Dule (e d.). Plenum Press, New York, 1988. Bhardwaj, M. a nd S tead, G. “Introduction t o co ntact-free ultrasonic char acterization of c onsolidated ma terials”, Application o f N on-Destructive E valuation in P owder Metals S eminar, I owa S tate U niversity, Ames, I A, April 2000. Papadakis, E. The inverse problem in materials characterization through ultrasonic attenuation and velocity measurements. N ondestructive M ethods f or M aterial P roperty Determination. C. R uud a nd R . G reen, J r. (e ds.), P lenum Press, New York, 1984. Jivanuwong, S. et al. Design and development of an on-line system to study ultrasonic properties of model food systems

19. 20. 21. 22.

23.

32-11

during freezing. Paper presented at the 2002Annual Meeting and Food Expo, Anaheim, CA. Graham-Rowe, D. That cheese sounds great, New Scientist, August 1998, p. 20. Kruger, S.E. et al. Ultrasonic backscattering formulation applied to cast iron characterization. Proceedings of the 7th European Conference on Nondestructive Testing, Copenhagen, 1998. Bhardwaj, M. “Non-contact ul trasonic c haracterization o f ceramics a nd co mposites”, Pr oceedings o f t he American Ceramic Society, Vol. 89, 1998. Levesque, D. et al . Thickness and grain size mo nitoring in seamless t ube makin g p rocess usin g las er ul trasonics”, Proceedings of the Fifth International Workshop, Advances in Sig nal Pr ocessing f or N ondestructive E valuation o f Materials, Quebec City, August 2005. Howard, Q. and Enzukewich, S. The effects of microstructure o n th e ul trasonic t esting o f all oy s teels. M aterials Evaluation, December, 1997.

10/3/2008 9:09:04 PM

43722_C032.indd 12

10/3/2008 9:09:04 PM

33 Lipid Membranes on Highly Ordered Porous Alumina Substrates

Andreas Janshoff Johannes Gutenberg Universität

Claudia Steinem Georg-August Universität

33.1 I ntroduction ...........................................................................................................................33-1 33.2 Preparation of Highly Ordered Porous Alumina Substrates............................................33-1 33.3 Preparation of Pore-Suspending Lipid Membranes..........................................................33-2 Nano–Black Lipid Membranes · Solvent-Free Pore Suspending Membranes

33.4 Insertion of Ion Channel Forming Peptides and Proteins ...............................................33-4 33.5 C onclusions ............................................................................................................................33-5 References ...........................................................................................................................................33-5

33.1 Introduction For a living cell, the main barrier between the inside and outside world is a 4–6 nm thick layer mainly composed of lipids and proteins, t he p lasma memb rane. Ā e l ipids, a rranged i n a b ilayer, form a barrier, impermeable for water-soluble ions and molecules. Membrane spanning proteins a re embedded, which a re i n pa rt responsible for the selective transport of otherwise impermeable ions a nd mole cules. Ā e f undamental u nderstanding o f t hese membrane proteins as well as their applications in biosensors and screening assays make the development of membrane structures attached to a surface, which allow for the functional insertion of transmembrane proteins, very at tractive. Such bilayers provide, however, o nly a n app ropriate en vironment f or memb rane p roteins if several requirements are ful lled. Each leaet of the lipid bilayer s hould b e i n t he  uid s tate a nd t he en tire memb rane should be surrounded by an aqueous phase. Several different bilayer systems have been developed over the past 30 years to fulll t hese re quirements, t wo o f w hich w ill b e b riey discussed here: (1) Solid supported membranes, rst described by Brian and McConnell i n 1984 on g lass su rfaces [1], a llow to f unctionalize conducting as well as nonconducting surfaces [2–8]. Ā e resulting membranes have, however, some major drawbacks. Āe membrane i s i n d irect c ontact to t he su rface, w hich h ampers t he insertion a nd f unctionality of t ransmembrane proteins. Due to the lack of a second aqueous compartment, the transport of ions and small molecules mediated by ion channels, transporters, and pumps c annot b e f ollowed e asily. To mo nitor t he t ransport o f ions or mole cules f rom one c ompartment t o a nother, a l ipid

bilayer ne eds to b e p repared t hat s pans a sm all ap erture, t hus separating t wo aque ous c ompartments. Suc h memb ranes h ave been developed in the 1970s by Müller and Rudin [9]. However, they are very fragile, and automation and parallization as required for the design of biosensors and automated screening systems, is not possible. Hence, attempts have been made to suspend lipid bilayers across small apertures, manufactured in glass or silicon [10–14]. In recent years, a hybrid system has been established, which i s ba sed o n h ighly o rdered me soporous sub strates, o n which lipid bilayers are prepared in a way that they suspend the pores. Ā ese membranes are to be expected to combine the merits of s olid supp orted a nd f reestanding l ipid bilayers su spending a 1–100 µm sized single pore.

33.2 Preparation of Highly Ordered Porous Alumina Substrates Ā e sub strate t hat i s u sed i s h ighly o rdered p orous a lumina. Anodized porous alumina has been extensively investigated over the last  ve decades [15,16]. Ā e material is su fficiently stable in aqueous solution and hydrophilic so that the pores are lled with water. I t c an b e p repared i n v arious p ore si zes r anging f rom around 20 to 400 nm, making it an ideal mesoporous/macroporous m aterial f or t he de velopment o f a ne w b ilayer s ystem. To obtain a sie ve-like s tructure o f h exagonally o rdered p ores w ith diameters o f a bout 6 0 nm, t he f ollowing p rocedure i s p ursued (Figure 33.1). Aluminum foils (thickness 0.5 mm, purity 99.999%) are c leaned w ith e thanol, e lectropolished, a nd a nodized i n 33-1

43722_C033.indd 1

10/3/2008 9:09:37 PM

33-2

Smart Materials

First anodization

Dissolution of AI2O3

Second anodizatiion

Dissolution of AI

Pore bottom opening

FIGURE 33.1 Schematic drawing of the procedure to generate a sieve like structure of hexagonally ordered pores in alumina. Further details can be found in the text.

aqueous 0.3 M oxalic acid solution at U = 40 V and T = 2°C for 3 h. Ā e formed porous oxide layer is removed by wet chemical etching in a m ixture of phosphoric acid (6 w t%) and chromium(VI) oxide (1.8 wt%) at 60°C for at least 3 h [17–19]. Ā e remaining pattern on the aluminum substrate serves as a m ask for the second anodization process using the same parameters as in the rst step. Ā e second anodization is performed for several days to achieve a porous layer of reasonable thickness, which guarantees mechanical stability. Under the chosen conditions, the pores grow around 1–2 µm i n de pth p er h our. Ā e re sulting p orous a lumina sub strates are then incubated in a saturated HgCl2 solution to remove the underlying aluminum layer. Finally, pore bottoms are removed by ch emical etch ing a t T = 3 0°C w ith 1 0 w t% pho sphoric a cid solution. Ā e process of pore opening can be monitored online by impedance spectroscopy [20]. Ā e resulting porous substrates are analyzed by scanning electron microscopy.

g

din

rea

sp

ion rpt iso em let td rop idso lve n

le

sic

Ve

Lip

ion

Ch

rpt

iso

em

Ā e basic idea is to span lipid bilayers on the mesoporous alumina substrate. A prerequisite for such a preparation is that the membrane c overs t he upp er pa rt o f t he p orous m aterial, w hile t he inner walls are noncovered. To achieve this, the bottom surface of t he sieve l ike a lumina structure is coated w ith a t hin 25 nm gold layer, a llowing for the chemisorption of thiol compounds. As y et, t wo diff erent s trategies we re re alized to obt ain p oresuspending l ipid memb ranes o n p orous a lumina sub strates dependent on the requirements of the membrane.

Ch

33.3 Preparation of Pore-Suspending Lipid Membranes

Gold evaporation

33.3.1 Nano–Black Lipid Membranes If highly insulating lipid bilayers are required for the insertion of peptidic and proteinacous ion channels to monitor their activity down to t he si ngle c hannel le vel, memb ranes a re p repared i n analogy to t he te chnique to ob tain bl ack l ipid memb ranes (BLMs) developed by Müller and Rudin (Figure 33.2). First, the gold-coated substrate is functionalized with 1,2-dipalmitoyl-snglycero-3-phosphothioethanol (DPPTE) to render it hydrophobic

43722_C033.indd 2

FIGURE 33.2 Schematic drawing of the formation of pore suspending membranes. O n t he le ft-hand sid e, t he for mation of a n ano-BLM i s illustrated, w hile on t he r ight-hand sid e t he f usion a nd s preading of unilamellar vesicles onto functionalized porous alumina substrates is shown leading to solvent-free pore suspending lipid membranes.

10/3/2008 9:09:37 PM

33-3

Lipid Membranes on Highly Ordered Porous Alumina Substrates

and to p rovide a mo nolayer o n t he s olid pa rt o f t he supp ort, which needs to b e completed by a s econd l ipid monolayer. Ā e formation of a f ull lipid bilayer covering the solid part and suspending the pore part is achieved by applying a solution of 1,2diphytanoyl-sn-glycero-3-phosphocholine (DPhPC) (2% w/v) in n-decane to t he su rface of t he f unctionalized p orous substrate [20–22]. Ā e p rocess o f l ipid b ilayer f ormation a s w ell a s t he characteristic electrical parameters of the membrane are investigated by impedance spectroscopy. To model the impedance data, an equivalent circuit is required, which in this case is composed of a parallel connection of a capacitance Cm and a resistance R m representing t he p ore-suspending memb ranes i n s eries to a n Ohmic resistance representing the electrolyte resistance. In general, a good accordance between data and t is achieved, resulting in a mean specic membrane capacitance of Cm = (0.4 ± 0.1) µF/cm2 taking the porous area into account [20]. Ā is value is in good agreement with specic capacitance values obtained from classical BLMs generated by the Müller-Rudin technique, which are r eported t o be a round 0 .5 µF/cm2 [23]. To at tribute to t his similarity, we called the membranes nano-BLMs. Āe membrane capacitance is a n i nvaluable pa rameter to c ontrol a nd evaluate the process of na no-BLM formation. Ā e second cha racteristic parameter is the membrane resistance R m, which is determined at v ery lo w f requencies i n t he i mpedance s pectrum. A na noBLM w ith a memb rane re sistance o f >1 GΩ i s e ssential f or i ts application in single-channel recordings. In almost all preparations, t he re sistance o f t he p ore-spanning membranes e xceeds the critical value of around 1 GΩ, which is sufficient for low conductance me asurements a nd s tays at t his v alue f or t he e xperimental time period of 8–10 h. Ā e memb rane re sistance t hen slowly and continuously decreases owing to the individual rupturing of some pore suspending membranes.

33.3.2 Solvent-Free Pore Suspending Membranes It is desirable to create pore spanning lipid bilayers that are entirely free of solvent due to unwanted electrical and mechanical c ontribution fr om s olvent m olecules in teracting w ith t he membrane c onstituents. A s traightforward w ay to ob tain s olvent-free membranes is to spread and fuse unilamellar vesicles on solid supports forming planar supported l ipid bilayers f ree of de tergents a nd s olvents. Ā e s uccess o f t his p reparation method depends strongly on the adhesion between the attached vesicles and the substrate. Hence, in most cases, strong attractive electrostatic interactions are employed to overcome the threshold ad hesion energ y a nd to en force r upture o f v esicles leading to p lanar l ipid b ilayers. S olvent-free memb ranes o n porous sub strates a re pa rticularly c hallenging d ue to t he reduced c ontact a rea o f t he v esicle w ith t he p orous m atrix, which re sults i n a n i nherently re duced ad hesion energ y. O ne way to create planar pore spanning bilayers is to chemisorb thiols w ith charged end g roups, w hich a llow to de posit opp ositely c harged ve sicles (Fi gure 3 3.2). Ā e gold su rface i s  rst functionalized wi th a m onolayer of 3-mercaptopropionic acid

43722_C033.indd 3

25 nm

80 nm

FIGURE 33.3 Scanning force microscopy image of a lipid bilayer suspending t he p ores of a p orous a lumina s ubstrate. Two p ores a re not covered by a membrane.

followed by an incubation for 4 h at T = 5 5°C w ith large u nilamellar vesicles with a diameter of 100 nm composed of either positively charged N,N,-dimethyl-N,N,-dioctadecylammonium bromide (D ODAB) or 1 ,2-dioleoyl-3-trimethylammoniumpropane chloride (DOTAP). In situ atomic force microscopy (AFM) images clearly reveal the presence of a single lipid bilayer (DODAB) covering the porous alumina matrix next to u ncovered pores (Figure 33.3). Pore spanning bilayers offer t he u nique p ossibility to c hallenge t heir ele ctrical p roperties o n a lo cal s cale. Re cently, Steltenkamp et al. [24]. managed to directly measure the restoring f orces o f f ree-standing l ipid memb ranes c omposed o f DODAB a nd D OTAP c overing a na noscopic h ole o f s elected pore si ze, w hich i s d isplaced b y a n A FM t ip w ith a de ned normal force. With this technique, it is feasible to infer elastic properties of lipid bilayers, i.e., the bending modulus and lateral tension f rom o nly a f ew l ipid mole cules ( 5000–30,000 l ipids). Ā e indentation experiments were quantied by means of variational calculus de scribing t he i ndentation o f a p ore-spanning bilayer with a nite-sized tip. Typical force–distance curves obtained from uid phase DOTAP a nd gel ph ase D ODAB memb ranes a re de picted i n Figure 3 3.4. A ll c urves w ere t aken i n t he c enter o f t he p ores. Figure 3 3.4 s hows t race ( indentation) a nd re trace ( relaxation) curves t aken o n gel -phase D ODAB a nd  uid-phase DOTAP bilayers covering pores with an average radius of 90 nm (insets). Strikingly, the indentation curves exhibit a linear dependency of the re storing f orce o n t he p enetration de pth f or D ODAB a s well as for D OTAP membranes over t he f ull r ange. Ā e mean slope representing t he appa rent “spring constant” of t he membrane i s a f unction o f t he ph ysical s tate o f t he b ilayer. F luid DOTAP bilayers show an apparent spring constant of 90 nm = (0.0039 ± 0.0008) N m−1, while DODAB bilayers in t he kDOTAP gel-state are considerably stiffer, displaying a spring constant of 90 nm kDODAB = (0.015 ± 0.004) N m−1. I mportantly, b oth t race a nd retrace curve lie on top of each other except for the snap-on and snap-off positions, illustrating that under these conditions, it is possible to assess static, elastic properties. It is further possible to

10/3/2008 9:09:38 PM

33-4

Smart Materials

DODAB 3

F (nN)

DOTAP 2

1

0 −200

0 h0 (nm)

200

FIGURE 3 3.4 Force-distance c urves ( black l ine: i ndentation; g rey line: re traction) t aken i n t he c enter of ge l-phase D ODAB a nd  uidphase D OTAP bi layers c overing a lumina p ores w ith a d iameter of 180 nm. Ā e i nsets s how A FM i mages t aken at lo w a nd h igh nor mal forces illustrating the situation on the surface.

extract t he bending modulus of t he bilayer a nd t he lateral tension representing t he ad hesive force of t he membrane attached to the pore rims by solving the Euler-Lagrange equation yielding the shape of the indented membrane. Ā is method is the rst one to assess bending rigidity of membranes on a nanoscopic scale.

33.4

Insertion of Ion Channel Forming Peptides and Proteins

Pore-suspending membranes h ave b een i nvented w ith t he a im to i nsert io n c hannel p eptides a nd p roteins a nd mo nitor t heir activity. If the resistance of such a membrane is sufficiently high, namely i n t he Gigaoh m re gime, it i s p ossible to mo nitor t he ion  ow t hrough a si ngle ion channel using high ga in current

ampliers. Ā e gener al f unctionality o f na no-BLMs w as  rst demonstrated u sing g ramicidin a s a c hannel-forming p eptide [20]. Ā e p eptide i s adde d to t he aque ous ph ase. Wi th t ime, a peptide spontaneously partitions into the bilayer, which can be monitored by the resulting ion-current that is driven by a holding potential of Vh = +70 mV across the bilayer in 0.5 M symmetric KCl solution. Ā e peptide activity is characterized by stepwise current changes, which can be attributed to single conductance states a nd multiples of t hose. A si ngle op en s tate O 1 ex hibits a current  ow o f ( 4.2 ± 0. 15) p A u nder t he c hosen c onditions, which translates in a gate d conductance state of (60 ± 2) pS. In general, gramicidin is a good indicator of the formation of single lipid bilayers, since gramicidin is, with a leng th of 2.6 nm, only capable of spanning the hydrophobic part of one bilayer leading to the observed current steps. Besides peptides, also large proteins can be inserted in membranes su spending p ores w ith d iameters o f 6 0 nm. A go od example is t he reconstitution of t he outer membrane protein F (OmpF) f rom Escherichia coli. It is composed of 16 antiparallel aligned β-sheets (β-barrel), connected by amino acid sequences referred to a s lo ops a nd t urns, building up a w ater-lled pore. Ā ree o f t hese mo nomeric u nits w ith a mole cular w eight o f 37.1 kDa [25] a nd a leng th of 5 nm [26] a re a rranged a round a threefold molecular axis. Loop 3 (L3) folds into the barrel forming a c onstriction zone of (11 × 7 ) Å 2 at appro ximately half the height of the channel. Ā e constriction zone is assumed to b e a decisive factor regarding conductivity and ion selectivity as elucidated by means of computer simulations and mutant studies. To achieve the insertion of OmpF in preformed nano-BLMs, the protein is added from a detergent solution to the aqueous phase [27]. Ā e reconstitution of the trimeric protein into the membrane causes b oth a t hree-step i ncrease i n c onductance a nd a s tep decrease i n io nic c urrent d ue to t he op ening a nd c losing o f the channel subunits (Figure 33.5A). From a ll-point histogram analysis of such current traces, the three different conductance levels w ith G1 = ( 1700 ± 8 0) p S, G 2 = ( 3360 ± 8 0) p S, a nd

2s

O2 O1

Frequency

200 pA

C 0.03

O3

C (A)

0.02 0.01 0.00

(B)

O1

O2 O3

0

1000 2000 3000 4000 5000 G (pS)

FIGURE 33.5 (A) Time trace of a characteristic channel activity (change in current) of one OmpF-trimer reconstituted in nano-BLMs at a holding potential of Vh = −100 mV. Ā re e different current levels of the different opening states (O1, O2, O3) as well as the closed state “C” can be distinguished. (B) All-point histogram of the current trace.

43722_C033.indd 4

10/3/2008 9:09:38 PM

Lipid Membranes on Highly Ordered Porous Alumina Substrates

G3 = (5060 ± 50) pS (Figure 33.5B) can be determined. Ā e conductance of one subunit monomer c alculated a s t he d ifference between t he c onductance s tates averages Gm = ( 1690 ± 2 0) p S. Ā e current traces unambiguously demonstrate that even large proteins can be inserted into nano-BLMs. To give an impression about t he si zes, o ne c an c alculate t hat o ne p ore o f t he p orous alumina holds about 4100 lipid molecules, taking the molecular surface area of one DPhPC molecule of 69 Å 2 into account [28]. Ā e OmpF protein occupies an area of around 80 nm 2 [26], which is around 3% of the total membrane area on one pore.

33.5

Conclusions

Pore-suspending l ipid memb ranes o n h ighly o rdered p orous alumina substrate are developed as a versatile alternative to classical freestanding and solid-supported membranes. It has been shown t hat na no-BLMs on p orous a lumina substrates a re well suited for the functional insertion of ion channels, allowing the analysis o f t heir ac tivities do wn to t he si ngle c hannel le vel. Owing to t he c oupling o f t he memb ranes to a supp ort, t hese membranes a re i n principle su ited for t he de velopment of biosensors and screening devices.

References 1. Brian, A. A., McConnell, H. M., Allogeneic stim ulation o f cytotoxic T cells b y su pported p lanar mem branes, Proc. Natl. Acad. Sci. USA, 81, 6159, 1984. 2. Hillebrandt, H., Wiegand, G., T anaka, M., Sac kmann, E., High ele ctric r esistance p olymer/lipid co mposite  lms on indium-tin-oxide electrodes, Langmuir, 15, 8451, 1999. 3 . Janshoff, A., S teinem, C., Transport acr oss a rticial membranes—an a nalytical p erspective, Anal. B ioanal. Chem ., 385, 433, 2006. 4. Kiessling, V., Tamm, L. K., Measuring distances in supported bilayers by  uorescence in terference-contrast micr oscopy: Polymer supports and SNARE proteins, Biophys. J., 84, 408, 2003. 5. Knoll, W., Morigaki, K., Naumann, R., Sacca, B., Schiller, S., Sinner, E.-K., Functional tethered bilayer lipid membranes, in Ultrathin Elec trochemical C hemo- a nd B iosensors, V. M. Mirsky, ed., Springer, Berlin, 2004, p. 239. 6. Sackmann, E., Supported membranes: Scientic and practical applications, Science, 271, 43, 1996. 7. Steinem, C., J anshoff, A., U lrich, W.-P., S ieber, M., Galla, H.-J., Impedance analysis of supported lipid bilayer membranes: A s crutiny of d ifferent p reparation te chniques, Biochim. Biophys. Acta, 1279, 169, 1996. 8. Tanaka, M., Sackmann, E., Polymer-supported membranes as models of the cell surface, Nature, 437, 656, 2005. 9. Müller, P., Rudin, D. O., Action potentials induced in biomolecular lipid membranes, Nature, 217, 713, 1968.

43722_C033.indd 5

33-5

10. Cheng, Y., Bushby, R. J., Evans, S. D., Knowles, P. F., Miles, R. E., Ogier, S. D., Single ion channel sensitivity in suspended bilayers on micromachined supports, Langmuir, 17, 1240, 2001. 11. Fertig, N., Klau, M., George, M., Blick, R. H., Behrends, J. C., Activity o f sin gle io n cha nnel p roteins dete cted wi th a planar microstructure, Appl. Phys. Lett., 81, 4865, 2002. 12. Ogier, S. D., Bushby, R. J., Cheng, Y., Evans, S. D., Evans, S. W., Jenkins, T. A., Knowles, P. F., Miles, R. E., Suspended planar phospholipid bi layers on m icromachined s upports, Langmuir, 16, 5696, 2000. 13. Osborn, T. D ., Yager, P., F ormation o f p lanar s olvent-free phospholipid bi layers b y L angmuir-Blodgett t ransfer of monolayers to micr omachined a pertures in si licon, Langmuir, 11, 8, 1995. 14. Schmidt, C., Mayer, M., Vogel, H., A chip-based biosensor for t he f unctional ana lysis of s ingle i on chan nels, Angew. Chem. Int. Ed., 39, 3137, 2000. 15. Masuda, H., F ukuda, K., Or dered met al na nohole a rrays made by a two-step replication of honeycomb structures of anodic alumina, Science, 268, 1466, 1995. 16. Ā ompson, G. E., Wood, G. C., Anodic lms on aluminum, in Treatise o n M aterials S cience a nd T echnology, Vol. 23, J. C. Scully, ed., Academic Press, New York, 1983, p. 203. 17. Li, A. P ., M üller, F ., B irner, A., N ielsch, K ., G ösele, U ., Hexagonal pore arrays with a 50–420 nm interpore distance formed b y s elf-organization in a nodic al umina, J. Ap pl. Phys., 84, 6023, 1998. 18. Li, A. P ., M üller, F ., B irner, A., N ielsch, K ., G ösele, U ., Polycrystalline nanopore arrays with hexagonal ordering on aluminum, J. Vac. Sci. Technol. A, 17, 1428, 1999. 19. Li, A. P ., M üller, F ., B irner, A., N ielsch, K ., G ösele, U ., Fabrication a nd micr ostructuring o f hexa gonally o rdered two-dimensional nanopore arrays in a nodic alumina, Adv. Mater., 11, 483, 1999. 20. Römer, W., S teinem, C., I mpedance a nalysis a nd sin glechannel recordings on nano-black lipid membranes based on porous alumina, Biophys. J., 86, 955, 2004. 21. Horn, C., Steinem, C., Photocurrents generated by bacteriorhodopsin ads orbed o n na no-black li pid mem branes, Biophys. J., 89, 1046, 2005. 22. Römer, W., Lam, Y. H., Fischer, D., Watts, A., Fischer, W. B., Göring, P ., Wehrspohn, R . B ., G ösele, U ., S teinem, C ., Channel ac tivity o f a viral tra nsmembrane p eptide in micro-BLMs: Vpu1–32 from HIV-1, J. Am. C hem. S oc., 126, 16267, 2004. 23. Benz, R ., F röhlich, O ., L äuger, P ., M ontal, M ., E lectrical capacity of black lipid lms and of lipid bilayers made from monolayers, Biochim. Biophys. Acta, 394, 323, 1975. 24. Steltenkamp, S., Müller, M. M., Deserno, M., Hennesthal, C., Steinem, C., J anshoff, A., M echanical p roperties o f p orespanning lipid bilayers probed by atomic force microscopy, Biophys. J., 91, 217, 2006.

10/3/2008 9:09:39 PM

33-6

25. Buehler, L. K., Kusumoto, S., Zha ng, H., Ros enbusch, J. P., Plasticity of Escherichia coli porin channels. Dependence of their conductance on strain and lipid environment, J. Biol. Chem., 266, 24446, 1991. 26. Cowan, S. W., Schirmer, T., Rummel, G., Steiert, M., Ghosh, R., Pauptit, R. A., Jansonius, J. N., Rosenbusch, J. P., Crystal structures explain functional properties of two E. coli porins, Nature, 358, 727, 1992.

43722_C033.indd 6

Smart Materials

27. Schmitt, E. K., Vrouenraets, M., Steinem, C., Channel activity of OmpF monitored in nano-BLMs, Biophys. J., 91, 2163, 2006. 28. Pownall, H. J ., P ao, Q ., B rockman, H. L., M assey, J . B ., Inhibition o f le cithin-cholesterol ac yltransferase b y diphytanoyl phosphatidylcholine, J. Biol. C hem., 262, 9033, 1987.

10/3/2008 9:09:39 PM

Index A Active structures, 7-4 concepts, 7-12, 7-16 Active hybrid control, 25-17 Active truss structures, 26-1 active shunt damping, 26-6 inductive shunting, 26-5 integral force feedback, 26-3 extension to cable structures, 26-4 passive shunt damping, 26-4 piezoelectric stack actuator, 26-1 resistive shunting, 26-5 Actuators, 4-2, 5-7, 6-6, 9-12 MSMA, 20-5 piezoelectric, 9-25, 25-5 sensors, 19-13 SMA design methods, 20-15 Adaptive concepts, 7-13 materials, 5-7 systems, 5-7 Adaptive controls, 7-4 Adhesives, smart applications, 11-2 coatings, 11-3 electrically conductive, 11-12 properties, 11-2 Aerodynamic control concepts, 7-11 Aeroelasticity, 7-7 effects, dynamic, 7-9 static, 7-8 Aircraft control, 7-10 design, 7-4 structures, 7-5 Alumina substrates, 33-1 Anisotropically conductive adhesives (ACA), 11-18 properties electrical, 11-19 thermal, 11-20 reliability, 11-21 Anisotropically conductive  lms (ACF), 11-19 properties electrical, 11-19

thermal, 11-20 reliability, 11-21 Applications, smart materials active structures, 7-7 aircraft, 7-1 functionally graded polymer blends, 3-13 gels, 10-6 SMAs, 4-1–4-7 Aquatic animals, 27-1–27-18 smart materials, 27-1 smart structures, 27-2 Battery materials, smart electrochemical concepts, 8-1 types, 8-2 carbon anode, 8-6 layered oxide cathode, 8-4 lithium ion, 8-2 olivine oxide cathode, 8-6 spinel oxide cathode, 8-5

B Bioanalytical systems, 24-24 Biocatalysts, 24-25 Biologging, see Biotelemetry, 27-2 applications and ndings, marine animals, cephalopods, 27-9; sh, 27-7; turtles, 27-10 energy expenditure, 27-11 acoustic telemetry, 27-13, cephalopods, 27-11; sh, 27-11; turtles, 27-11 marine mammals, 27-13 video telemetry, 27-14, cephalopods, 27-14; sh, 27-14; turtles, 27-14 marine mammals, 27-14 future behavioral hypothesis, 27-17 biotelemetry and smart dust, 27-17 transmitter, 27-16 locating and tracking, marine mammals, 27-11 Biomacromolecule - sensitive polymers, 13-4 Biomedical sensing, 29-1 biosensors, 29-2 Bioseparation, 24-17, 24-19 Biotechnology, 24-13 Biotelemetry, 27-2

alternatives for natural receptors, 29-4 DNA sensors, 29-3 drug delivery, 29-4 enzymatic, 29-3 immunosensors, 29-3 smart materials, aquatic animals, 27-2 smart polymers, 29-2 smart structures, aquatic animals, 27-2

C Cantilever, piezoelectric polymer, 31-1 Carbon anode battery, 8-6 Carbon microtubes, 22-1 synthesis and morphology, 22-3, 22-6 CCNT, see Conical carbon nanotubes CD-ROM drive with piezoelectric shunt, 25-12 CMT, see Carbon microtubes Ceramics, smart, 9-1 electrostrictive, 9-6 applications, 9-9 ferroelectric, 9-1 materials, 9-15 nanostructured, 25-21 perovskite, 11-5 piezoelectric, 9-6 applications, 9-9 processing, 9-7 Chemical indicators, 30-1 classication, 30-2 Chitosan-based gels applications controlled release matrix, 10-6 ECM, 10-9 immobilization supports, 10-9 separation membrane, 10-8 network types, 10-1 coordination, 10-6 grafted and block, 10-4 hydrogen bonding, 10-1 polyelectrolyte (PEC), 10-3 self-assembly, 10-5 Chromates, 23-2 Closed loop control sensors, 16-6 Coatings, 16-7 Composites, 4-2, 5-1, 18-2

I-1

43722_C034.indd 1

10/16/2008 10:39:05 PM

I-2

Index

hybrid, HCM, 5-5 applications, 5-5, 5-7 bers, natural, 5-2 properties, 5-3 synthetic, 5-2 reinforcements, 5-2–5-5 types, 5-1 MMC, 5-1 CMC, 5-1 PMC, 5-1 matrix hybridized, 5-4, 5-5 piezoelectric, 9-8, 9-18 Concepts active structures and materials, 7-12 adaptive, 7-13 aerodynamic control, 7-11 deformation and effectiveness, 7-15 intelligent processing, 18-1 variable shear, 7-11 virtual, 7-10 wing tip design, 7-19 Concrete, 4-2 Conducting polymers, 24-1 applications energy conversion and strength, 24-5 sensing, 24-2 synthesis, 24-2 Conical carbon nanotubes, 22-1 applications, 22-11 drug delivery, 22-13 electrode materials, 22-11 eld emissions, 22-12 uid ow, 22-13 miscellaneous, 22-13 porous carbons, 22-12 template for nanoelectrode ensembles, 22-11 morphology, 22-2, 22-10 synthesis, 22-8 Control, aircraft, 7-10 Corrosion protective coatings, 23-1–23-15 Critical temperature indicators, 30-3 CTI, see Critical temperature indicators Cu-based alloys, 20-30 Cure monitoring optical ber sensors, 12-1 Curing process, 16-4–16-10 thermosetting resins, 18-3

pH-sensitive, 13-3, 13-12, 24-14 stimuli-responsive, 13-5 thermoresponsive, 13-2 pulsatile systems, 13-11 self-regulated systems, 13-12 stimuli-modied release systems, 13-2 Dynamic elastic effects, 7-9 Dynamic systems, seat suspension with ER damper, 25-7 vehicle suspension using MR damper, 25-9

D

FBG, see Fiber Bragg grating FDEMS, see Frequency-dependent electromagnetic sensors Fe-based alloys, 20-29 Ferroelectric ceramics electrostrictive effects, 9-1 piezoelectric effects, 9-1 Ferromagnetic shape memory alloys (FSMA), 6-2 actuator material, see Fe-Pd, 6-6 characteristics, 6-4 mechanisms, 6-6

DNA sensors, 24-3 Dopants, 23-4–23-11 Doping, 11-8 Drug delivery systems, 10-16, 13-1, 13-10, 29-4 nanotechnology, 13-6 polymers biomacromolecule-sensitive, 13-4 bioseparation, 24-19 controlled drug delivery system, 13-10

43722_C034.indd 2

E ECA, see Electrically conductive adhesive Electrically conductive adhesive (ECA), 11-12 types, 11-12 ACA, anisotropically conductive, 11-18 ICA, isotropically conductive, 11-13 NCA, nonconductive, 11-21 Electrochemical battery concepts, 8-1 Electrochromic smart windows, 14-26 control, 14-29 materials, 14-28 structure, 14-27 Electrochromic (EC) sol-gel coatings, 11-25 Electrochromism, 11-25 Electrodynamics, 25-17 Electromagnetic sensors, 16-1 Electrorheological uids, 4-2, 21-1 alternative modeling, 21-4 Bingham model, 21-5 ER damper, 25-7 ER uid-based mount, 25-1 high-yield strength materials, 21-1 lamellar particle structure, 21-3 properties, 21-2 Enzymatic biosensors, 29-2 Enzymes, 13-12 ER, see Electrorheological uids ERF, see High-yield strength electrorheological uids Equipment, aquatic animals receivers, 27-4 transmitters, 27-4 Electrostrictive materials, 9-6 applications, 9-9

F

phase transformations, 6-4, 6-5 properties, 6-6 Fiber Bragg grating, 14-8 Fiber Bragg grating sensor, 12-4, 14-8–14-10 Fiber optic systems optical ber concepts, 14-8 sensors, 14-7, 14-8 optical ber modulation mechanisms, 14-4 intensity-modulated sensors, 14-4 interferometric sensors, 14-6 polarimetric sensors, 14-5 optical ber sensors and windows, 14-1, 18-5 applications, 14-5, 14-10–14-14 properties, 14-3 Films, see Anisotropically conductive (ACF) Films, thin, 1-1–1-12, 9-18 Flexible structures, 25-1 vibration control, 25-1 ER uid-based mount, 25-1 MR uid-based mount, 25-4 passive piezoelectric actuator-based mount, 25-6 Flexures, 20-16 Flip-chip under ll lead-free alloys, 15-6 materials, 15-2 no-ow, 15-7 silica  llers, 15-9 process, 15-4 properties, 15-3 reliability, 15-4 Frequency-dependent electromagnetic sensors, 16-1 closed loop control, 16-6 coatings, 16-7 cure process, 16-4 curing by UV and EB radiation, 16-7 life monitoring, 16-10 resin infusion of preforms, 16-5 theory, 16-2 Functionally graded polymeric materials, 3-1 applications, 3-13 preparation methods, dissolution-diffusion, 3-2 polymerization-diff usion, 3-3, 3-4 properties, 3-2, 3-8 types PBMA/PEO, 3-8 PC/PS, 3-7 PEO/PBMA, 3-12 PEO/PLLA, 3-11 PVC/PHMA, 3-10 PVC/PMMA, 3-5

G Galvanic cells, 11-25 GER, see Giant electrorheological effect GER effects, 21-3 Giant magnetocaloric effect, 20-4

10/16/2008 10:39:05 PM

I-3

Index Glucose, 13-4 Glutathione, 13-4 GMCE, see Giant magnetocaloric effect GMM, see Magnetostrictive materials

H HCM, 5-5, 5-6 Health monitoring, 4-6, 12-3 sensing methods, 12-3 sensors, 12-3 ber Bragg grating (FBG), 12-4 PVDF (polyvinylidene uoride), 24-10, 31-1 heterogeneous polymers, 24-16 PZT (lead zirconate titanate), 12-5 High-temperature SMAs, 20-31 High-yield strength ERFs, 21-1 alternative modeling, 21-4 Bingham model, 21-5 lamellar particle structure, 21-3 Hybrid active control with piezoelectric materials, 25-17 Hydrogels, Chitosan-based applications, 10-16 derivatives, 10-13 types chemical, 10-15 physical, 10-15 Hydrothermal hybrid, 2-4 Hydrothermal synthesis, 2-1–2-4 ceramics, oxides, 2-4 crystal size, 2-7 morphology, 2-7 nonthermodynamic variables, 2-8

I ICP, see Inherentlt conductive polymers Immunosensors, 29-3 Ion channels peptides and proteins, 33-4 Inherently conductive polymers, 23-2, 24-3, 24-5 cathodic disbonding, 23-4 corrosion literature, 23-5–23-7 Intelligent chemical indicators, 30-1 classication, 30-2 factors in choice, 30-2 application, 30-2, 30-4, 30-5 cost, 30-2 response, 30-2 sensitivity, 30-2 shelf life, 30-2 indicator materials, 30-2 issues and limitations, 30-8 operating principles, 30-2 pH indicators, 30-2 smart devices/indicators, 30-1 temperature and time-temperature indicators, 30-2, 30-4, 30-5 CTI, 30-3

43722_C034.indd 3

defrost temperature indicators, 30-4 high-temperature indicating crayons, 30-3 holographic technology, 30-7 irreversible temperature labels, 30-3 LCD, 30-4 microtaggants technology, 30-7 smart ink technologies, 30-8 tamper-indicating devices, 30-8 Intelligent processing, 18-1 composite manufacturing sensors and sensory techniques, 18-4 RTM monitoring, 18-5 composite materials, 18-2 cure thermosetting resins, 18-3 concept, 18-1 metallic materials, 18-9 Iron-palladium (FePd), 6-3–6-7 Isotropically conductive adhesives (ICA) or polymer solder, 11-13–11-18

L Laminated - type smart structures methods, 25-18–25-20 classical laminated plate theory, 25-18 coupled layerwise theory, 25-19 elastic theory, 25-18 rst-order shear deformation theory, 25-18 surface impedance tensor approach, 25-18 third-order shear deformation theory, 25-18 transfer matrix, 25-18 Laser annealing SMAs, 20-14 Layered oxide cathodes, 8-4 LCD, see Liquid crystal displays Life monitoring sensors, 16-10 Linear theory of electrodynamics, 25-17 Liposomes - trigger release, 24-23 Lipid membranes nanoblack, 33-2 pore-suspending, 33-2 ion channel formed peptides and proteins, 33-4 porous alumina substrates, 33-1 preparation, 33-2 solvent-free, 33-3 Liquid crystal display, 30-4

M Magnetic eld-assisted superelasticity, 20-4 Magnetic eld-induced strain, 20-2 Magnetic shape memory alloys, 20-2, 20-3, 20-31 applications actuators, 20-5 damping, 20-5 sensing and power generation, 20-5 magnetic shape memory, 20-5

magnetically controlled SMAs, 20-2 properties, 20-3 production, 20-5 Magnetic shape memory effect, 20-4 Magnetorheological (MR) uids, 4-2 applications, 17-6 basic operating modes, 17-1 direct shear, 17-6 MR damper, 25-4 MR uid-based mount, 25-4 valve, 17-5 composition, 17-3 additives liquid vehicle particles, 17-3 history, 17-1 properties, 17-4 Magnets, organic/polymer, 19-1 types, 19-5 Heacyanometallate - room temperature, 19-5 M(TCNE)2 - high room temperature, 19-5 V(TCNE)x - room temperature, 19-5 uses, 19-6 Magnetostrictive materials, 19-1, 19-11 applications, 19-11 torsional/actuator/torque sensor, 19-13 valve ring indentation, 19-14, 19-15 vibration sensor/actuator, 19-13 powder metallurgy and characteristics, 19-8 properties, 19-9 chemical, 19-10 physical, 19-10, 19-12 Martensitic transformation, 20-28 Materials modeling, 20-36–20-41 Mathematical modeling, 20-17 macroscopic, 20-21 micromechanical, 20-21 thermomechanics, 20-22 transformation kinetics, 20-24 Materials characterization, 32-1 complications, 32-4 sound frequency, 32-1 Medicine, 24-13 applications, 24-1, 24-3-4, 24-7, 24-10, 24-13, 24-16, 24-17, 24-23 bioseparation, 24-17, 24-19 smart pills, 24-17 MEMS, see Microelectromechanical systems Metallic material processing, 18-9 MFAS, see Magnetic eld-assisted superelasticity MFIS, see Magnetic eld-induced strain Microactuators, PVDF, 31-1 Microstructural mechanisms of SME, 20-41 macroscopic, 20-43 microscopic, 20-42 Microtubes, carbon, 22-1 growth mechanisms, 22-3 synthesis methods, 22-3 Molded under ll, 15-11 Molecular imprinting technology, 28-1

10/16/2008 10:39:05 PM

I-4 Applications, 28-1, 28-3 biomedical -drug delivery, 28-1 catalysts, 28-1 chemical and biological separation, 28-1 immunoassays, 28-1 nanotechnology, 28-1 sensors, 28-1 fabrication methods, 28-2 properties, 28-1 ease, 28-1 low cost, 28-1 speed, 28-1 stability, 28-1 tailorable binding affi nities, 28-1 Molecularly imprinting polymers, 28-1 formats, 28-3 templates, 28-4 Mounts ER uid-based, 25-1 MR uid-based, 25-4 passive piezoelectric actuator-based, 25-6 MR, see Magnetorheological uids MSM, see Magnetic shape memory MSMA, see Magnetic shape memory alloys MSME, see Magnetic shape memory effect

N Nanoblack liquid membrane, 33-2 Nanostructured ceramics, 25-21 Nanotubes, carbon, 22-1, 22-2 Natural bers, 5-2 Natural receptors, 29-4 NCU (noncontact ultrasound), 9-28 signal processing, 9-30 system, 9-30 transducers, 9-28 NDT, see nondestructive testing Network-type gels, 10-1, 10-6 Nickel-titanium (NiTi), 6-8–10, 20-30 NIPAAM, see poly(N-isopropylacrylamide) Nonconductive adhesives (NCA), 11-21 l ms,(NCF), 11-21 properties, 11-22 Nondestructive testing, 32-1

O Olivine oxide cathodes, 8-6 Optical ber sensors, 12-1, 14-2 Optical ber windows, 14-1 Optical properties, 1-8, 1-11 Organic/polymer magnets, 19-1 Oriented growth, 1-9 Other SMAs b-Ti alloy, 20-32 Oxides, see thermoelectric materials metallic, 11-7, 11-10 mist layered, 11-7 properties, 11-9

43722_C034.indd 4

Index

P PANI, see Polyaniline Passive control-sound and vibration, 25-16 Passive piezoelectric actuator-based mount, 25-6 Passive piezoelectric-based mount, 25-6 Passive structures, 7-6 PBMA/PEO system, 3-8 PC/PS blend, 3-7 PEO/PBMA blend, 3-12 PEO/PLLA blend, 3-11 Perovskite ceramics, 11-5, 11-10 pH-sensitive polymers, 13-3, 13-13, 24-14 Piezoelectric (PZT) applications, 9-9, 9-20 actuators, 5-7, 9-25, 25-5 composites, 9-8 effect, 9-12 hybrid control, 25-17 materials, 9-6, 9-15 ceramics, 9-15 crystals, 9-15 linear transducer, 26-1 passive control, 25-16 research in smart structures, 25-20 resonator and  lter, 9-21 transformer, 9-21 polymer, 24-11, 31-1 sensors, 5-7 Piezoelectricity, 9-13, 24-10 applications, 9-19 Piezoelectric polymer (PVDF) composite cantilever, 31-1, 31-2 fabrication, 31-3 laser micromachining, 31-4 metallization, 31-4 punching, 31-4 modeling and simulation nite element method simulation, 31-2 parametric analysis, 31-2 PVDF microactuators, 31-1 testing and evaluation unimorph and bimorph cantilever, 31-5, 31-6 Polyaniline, 23-2, 23-8 Polyester synthesis, 16-10 Polymerization-diff usion methods, 3-3 Polymers, 9-18, 13-2, 19-1, 23-2, 24-1, 24-10, 24-13, 31-1 Poly (N-isopropylacrylamide), 24-20 Polypyrrole, 23-2, 24-1, 24-6 Polyvinylidene uoride, 24-10, 31-1 properties, 24-11 Pore-suspending lipid membranes, 33-2 preparation lipid membranes, 33-2 Powder metallurgy, 19-8 GMM, 19-11 magnetostrictive actuator sensor, 19-8 PPy, see Polypyrrole Properties, SMAs, 20-32 Protective  lms, 23-14

Pulsatile systems, 13-11 PVC/PHMA blend, 3-10 PVC/PMMA polymer system, 3-6 Properties, 3-8–11 smart performance, 3-9–11 tensile, 3-8–9 thermal shock resistance, 3-9 PVDF, see Polyvinylidene uoride

R Rayleigh wave, 9-22 Receivers, aquatic animals, 27-4 Residual stress, 1-1 measurement, 1-2 theory, 1-12 XRD, see XRD, 1-3, 1-4, 1-9 Reversibly cross-linked polymers, 24-16 RTM monitoring ber optic sensors, 18-5

S SAW device, see Rayleigh wave Self-regulated systems, 13-12 Sensing biomedical, 29-1 Sensors, 5-7, 12-1, 12-3, 14-2, 16-1, 19-4, 19-8, 19-13 Bragg grating types, 14-8–14-9 composite manufacturing, 18-4 frequency-dependent electromagnetic 16-1 intensity-modulated, 14-4 interferometric, 14-6 polarimetric, 14-5 Shape memory alloy, see SMAs, 6-1, 20-1, 20-28, ferromagnetic, 6-2 martensitic transformation, 20-28 mechanisms, 6-6 properties, 20-32 Shape memory alloy systems, 20-29 Cu-based alloys, 20-30 Fe-based alloys, 20-29 high-temperature SMAs, 20-31 magnetic shape memory, 20-5 magnetic SMAs, 20-3 magnetically controlled SMAs, 20-2 properties, 20-3 Ni-Ti alloys, 20-30 Other SMAs, 20-31 Ternary Ni-Ti alloy systems, 20-31 Ti-Ni-Cu, 20-31 Ti-Ni-Nb, 20-31 Shape memory effect (SME), 6-1, 20-9, 20-20, 20-41 methods extrinsic, 20-12 intrinsic, 20-10 microstructured mechanisms macroscopic, 20-43 microscopic, 20-42

10/16/2008 10:39:05 PM

I-5

Index Shape memory materials experimental observations, 20-18 mathematical modeling, 20-17 Simple harmonic motion, 5-8 wing, 5-9 SHM, see Simple harmonic motion SMAs, 20-1, 20-8 Smart battery materials, 8-1 Smart ceramic materials, 2-1 composition and morphologies, 2-4 hydrothermal synthesis, 2-1, 2-2 merits, 2-2 nonthermodynamic variables, 2-8 production, 2-4 thermodynamic modeling, 2-7 thermodynamic variables, 2-7 Smart corrosion protective coatings, 23-1 chromates, 23-2 electrochemical potential, 23-2 partially soluble pigments, 23-2 environmental responsiveness, 23-2 Smart devices/indicators CTIs, 30-3 indicator materials, 30-2 issues and limitations, 30-8 types high-temperature indicating crayons, 30-3 ph indicators, 30-2 TTIs, 30-2, 30-4, 30-5 temperature, 30-2 Smart materials, aquatic animals applications, 27-2 biologging or biotelemetry, 27-2 denition, 27-1 equipment, 27-4 receivers, 27-4 transmitters, 27-4 types Lotek wireless sh and wildlife systems, 27-4 Sonotronics, 27-5 Star-Oddi Marine Device Manufacturing, 27-5 Vemco, 27-6 Wildlife Computers, 27-6 future, 27-16 behavioral hypotheses, 27-17 biotelemetry and smart dust, 27-17 transmitter, 27-16 Smart materials, modeling, 20-36 classication of modeling techniques, 20-40 conservation laws, 20-39 constitutive equations, 20-39 equations of state, 20-39 Smart memory alloys, 4-1, 6-1, 20-1 SME, see Shape memory effect SMM, see Shape memory materials Smart polymers, 24-1, 29-2 actuating, 24-3 bioanalytical systems, 24-24 biotechnology, 24-13

43722_C034.indd 5

elastomers, 24-1 medicine, 24-13 processing and device fabrication, 24-6 reversibly soluble biocatalysts, 24-25 sensing, 24-3 synthesis, 24-2 types heterogeneous, 24-16 ph-sensitive, 24-14, 24-19 reversibly cross-linked, 24-16 smart surfaces, 24-20 thermosensitive, 24-14 Smart structures, aquatic animals applications, 27-2 animal behavior, 27-2 biologging or biotelemetry, 27-2 denition, 27-1 equipment, 27-4 receivers, 27-4 transmitters, 27-4 types Lotek wireless sh and wildlife systems, 27-4 Sonotronics, 27-5 Star-Oddi Marine Device Manufacturing, 27-5 Vemco, 27-6 Wildlife Computers, 27-6 future, 27-16 behavioral hypothesis, 27-17 biotelemetry and smart dust, 27-17 transmitter, 27-16 Smart structures, discrete type, 25-19 Smart structures, laminated-type, 25-18 potential research, 25-20 Smart structures, vibration control, 25-1 Smart surfaces cell detachment, 24-20 controlled porosity “chemical value”, 24-22 liposomes - trigger release, 24-23 poly (NIPAAM), 24-20 temperature-controlled chromatography, 24-21 Smart window systems, 14-17 applications, 4-1–4-6 actuators, 4-2, 20-10 composites, 4-2 concrete, 4-2 electrorheological uids, 4-2 exures, 20-16 magnetorheological uids, 4-2 piezoelectric, 4-2, 5-7 retrotting, 4-4 structural uses, 4-2 active control, 4-2, 4-5 health monitoring, 4-6 hybrid control, 4-4 passive control, 4-3 SMT, see Surface Mount Technology, 15-10 Sol-gel coatings, electrochromic, 11-25 Sol-gel processing, 11-27

Solvent- ll lipid membranes, 33-3 Sound control piezoelectric materials, 25-16 Spars, 5-7–5-10, 5-13 Spinel oxide cathodes, 8-5 Static elastic effects, 7-8 Stimuli-modulated release systems, 13-2 Stimuli-responsive polymers, 13-5 Strain, 1-5–1-10 Structural uses of SMAs, 4-2 Structures active, 7-4 aircraft, 7-5 passive, 7-6 Surface mount technology, see ip-chip underll Synthetic bers, 5-2

T Ternary Ni-Ti alloy systems Ti-Ni-Cu, 20-31 Ti-Ni-Nb, 20-31 Thermodynamic modeling, 2-4–2-7 Ther moelectric materials Heikes formula, 11-5 Thermoresponsive polymers, 13-2 Thermosensitive polymers, 24-14 Thermosetting resins, 18-3 Thin  lms, 1-1, 9-18 properties, 1-4–1-8 Tissue engineering, 10-16 Transducers, 9-28 Transmitters, 27-4 TTI, see Time-temperature indicators TTI applications anticounterfeiting and tamper, 30-6 commercially available, 30-5 food shelf life, 30-5 Lifeline Fresh-Scan and Fresh Check, 30-6 3M monitor mark, 30-5 VITSAB, 30-6

U Ultrasonic non destructive testing, 32-1 attenuation measurements, 32-3 basic equations, 32-2 complications, 32-4 couplants, 32-3 examples, 32- 4–32-9 aw detection, 32-3–32-6 material variation, 32-6, 32-7 real defects, 32-4 references, 32-2 practices, 32-2 simplications, 32-4 sources of ultrasound, 32-3 ultrasonic velocity, 32-3

10/16/2008 10:39:05 PM

I-6

Index

Ultrasonic transducers, 9-20 Underl l molded, 15-11 water level, 15-12 Underl l packaging lead-free alloys, 15-6 no-ow, 15-7 USM, see Piezoelectric actuator

seat suspension using ER damper, 25-7 vehicle suspension using MR damper, 25-9 exible structures, 25-1 hybrid control, 25-17 smart piezoelectric materials, 25-16, 25-17 Virtual concept, 7-10

V

W

Variable shear concept, 7-11 Vibration control, 25-1 dynamic systems, 25-7 CD-ROM drive with piezoelectric shunt, 25-12

Water level under ll, 15-12 Windows, 14-17–14-30 architectural glazing, 14-18 characteristics, 14-20

43722_C034.indd 6

physics, 14-18 types electrochromic, 14-21 gasochromic, 14-25 PLDC, 14-24 suspended particle, 14-25 thermochromic, 14-22 thermotropic, 14-22

X XRD, 1-3–1-9

Y Yield diagrams, 2-5–2-7

10/16/2008 10:39:05 PM