Sensors Based on Nanostructured Materials

Sensors Based on Nanostructured Materials

Francisco J. Arregui Editor

Sensors Based on Nanostructured Materials

13

Editor Francisco J. Arregui Universidad Pu´blica de Navarra Depto. Ingenieria Electrica y, Electronica Campus Arrosadia, s/n 31006, Pamplona, Spain

ISBN: 978-0-387-77752-8 e-ISBN: 978-0-387-77753-5 DOI: 10.1007/978-0-387-77753-5 Library of Congress Control Number: 2008928442 # Springer ScienceþBusiness Media, LLC 2009 All rights reserved. This work may not be translated or copied in whole or in part without the written permission of the publisher (Springer ScienceþBusiness Media, LLC, 233 Spring Street, New York, NY 10013, USA), except for brief excerpts in connection with reviews or scholarly analysis. Use in connection with any form of information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed is forbidden. The use in this publication of trade names, trademarks, service marks, and similar terms, even if they are not identified as such, is not to be taken as an expression of opinion as to whether or not they are subject to proprietary rights. Printed on acid-free paper springer.com

Acknowledgments

As Editor, I want to express my gratitude to all the members of Springer US for their assistance and help, to Angela DePina and Jennifer Mirski for their infinite patience. Steven M. Elliot, who truly believed in this project since the very first moment and who made it possible to join all the pieces of this challenging puzzle, deserves a special mention. My colleagues and friends Rick Claus and Ignacio Matias kept their support for this book from the start and I am very grateful to them. The availability of the rest of contributors of this book has been amazingly encouraging. These scientists are some of the most active authors in the world in the nano & sensor field. It has been really great to work with Professors Craig A. Grimes, Yoke Khin Yap, Manuel Vazquez, Michael J. McShane, John T.W. Yeow, Joan R. Morante, Richard O. Claus and Ignacio R. Matias. They were also the ones who convinced other colleagues to participate in this book. Of course, I would not like to forget any of the authors, with a special and thankful remark to Prof. Zheng Wei Pan who attended the request made by Prof. Yoke Khin Yap. For the rest, I want to acknowledge the work made by the people in Pennsylvania; Boston; Waterloo (Canada); Barcelona, Madrid and Pamplona (Spain); Texas and last, but not least, Virginia: N. Sinha, Kristen E. La Flamme, Abhishek Prasad, Samuel Mensah, Teresa Andreu, Jordi Arbiol, Andreu Cabot, Albert Cirera, Joan Daniel Prades, Francisco Hernandez-Ramirez, Albert Romano-Rodriguez, Ignacio Del Villar, A. Asenjo, M.P. Morales, K.R. Pirota, G. Badini-Confalonieri, M. HernandezVelez, Jennifer Lalli, Bradley Davis and Christelle Jullian. Special thanks also to Javier Goicoechea for his world-record response time. I also want to acknowledge all the students, colleagues and friends of the Universidad Publica de Navarra who have shared their time, dedication and enthusiasm with us. They have been inspiring. I am also grateful to Dr. Paul Urquhart for his always interesting suggestions and to Ms. Manjula Jude for the final editing of the book. Besides the authors, editorial staff, reviewers and others who participated in this book, we would also like to express our gratitude to all their families, friends and colleagues for their understanding. v

vi

Acknowledgments

If I express gratitude to the participants’ families, I cannot forget mine, just saying thanks to them is a very small thing (a nano thing!) compared to all their support over these years. Thanks to my Encarna, Modesto, Mikel, Joseba, Miriam-Larri and Isabel. Everyday they remind me that there is ‘‘plenty of room at the bottom’’.

Contents

1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Francisco J. Arregui

1

2

Carbon Nanotube and Fullerene Sensors . . . . . . . . . . . . . . . . . . . . . . John T. W. Yeow and Niraj Sinha

11

3

Non-carbon Nanotubes: Hydrogen Sensors Based on TiO2 . . . . . . . . Kristen E. LaFlamme and Craig A. Grimes

29

4

Alternative Nanostructured Sensors: Nanowires, Nanobelts, and Novel Nanostructures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Abhishek Prasad, Samuel Mensah, Zheng Wei Pan, and Yoke Khin Yap

5

Nanosensors: Controlling Transduction Mechanisms at the Nanoscale Using Metal Oxides and Semiconductors. . . . . . . . . . . . . . . . . . . . . . Teresa Andreu, Jordi Arbiol, Andreu Cabot, Albert Cirera, Joan Daniel Prades, Francisco Hernandez-Ramı´ rez, Albert Romano-Rodrı´ guez, and Joan R. Morante

59

79

6

Quantum Dots for Sensing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Javier Goicoechea, Francisco J. Arregui, and Ignacio R. Matias

131

7

Nanostructured Magnetic Sensors . . . . . . . . . . . . . . . . . . . . . . . . . . . Manuel Va´zquez, Agustina Asenjo, Maria del Puerto Morales, Kleber Roberto Pirota, Giovanni Badini-Confalonieri, and Manuel Herna´ndez-Ve´lez

183

8

Encapsulated Probes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Michael J. McShane

253

vii

viii

Contents

9

Optical Fiber Sensors Based on Nanostructured Coatings . . . . . . . . . Francisco J. Arregui, Ignacio R. Matias, Javier Goicoechea, and Ignacio Del Villar

275

10

Nanostructured Flexible Materials: Metal RubberTM Strain Sensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Christelle Jullian, Jennifer Lalli, Bradley Davis, and Richard Claus

303

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

319

Contributors

Francisco J. Arregui Electric and Electronic Engineering Department Universidad Pu´blica de Navarra, Edificio de los Tejos Campus Arrosadı´ a, 31006 Pamplona, Navarra, Spain Teresa Andreu Department of Electronics Faculty of Physics University of Barcelona Barcelona E-08028 Spain Jordi Arbiol Department of Electronics Faculty of Physics University of Barcelona Barcelona E-08028 Spain Agustina Asenjo Instituto de Ciencia de Materiales CSIC 28049 Madrid Spain Richard Claus NanoSonic Inc. 1485 South Main Street Blacksburg, VA 24060 Andreu Cabot Department of Electronics Faculty of Physics University of Barcelona Barcelona E-08028 Spain Albert Cirera Department of Electronics Faculty of Physics University of Barcelona Barcelona E-08028 Spain Giovanni Badini-Confalonieri Instituto de Ciencia de Materiales CSIC 28049 Madrid Spain Bradley Davis NanoSonic Inc. 1485 South Main Street Blacksburg, VA 24060 Ignacio Del Villar Public University of Navarre, 31006 Pamplona Spain Craig A. Grimes The Pennsylvania State University Electrical Engineering University Park, PA 16802 Javier Goicoechea Electric and Electronic Engineering Department Universidad Publica de Navarra, Edificio de los Tejos, Campus Arrosadia, 31006 Pamplona, Navarra, Spain Francisco Hernandez-Ramı´ rez Department of Electronics Faculty of Physics University of Barcelona Barcelona E-08028 Spain Manuel Herna´ndez-Ve´lez Instituto de Ciencia de Materiales CSIC 28049 Madrid Spain

ix

x

Contributors

Christelle Jullian Department of Materials Science and Engineering Virginia Tech Blacksburg, VA 24061 Jennifer Lalli NanoSonic Inc. 1485 South Main Street Blacksburg, VA 24060 Kristen E. LaFlamme Boston University Boston, MA Samuel Mensah Department of Physics Michigan Technological University Houghton, MI 49931 Joan R. Morante Department of Electronics Faculty of Physics University of Barcelona Barcelona E-08028 Spain Ignacio R. Matias Electric and Electronic Engineering Department Universidad Pu´blica de Navarra, Edificio de los Tejos Campus Arrosadı´ a, 31006, Pamplona, Navarra Spain Maria del Puerto Morales Instituto de Ciencia de Materiales CSIC 28049 Madrid Spain Michael J. McShane Biomedical Engineering Department Texas A&M University College Station, TX 77843 Abhishek Prasad Department of Physics Michigan Technological University Houghton, MI 49931 Zheng Wei Pan Faculty of Engineering & Department of Physics and Astronomy University of Georgia Athens, GA 30602 Joan Daniel Prades Department of Electronics Faculty of Physics University of Barcelona Barcelona E-08028 Spain Kleber Roberto Pirota Instituto de Ciencia de Materiales CSIC 28049 Madrid Spain Albert Romano-Rodrı´ guez Department of Electronics Faculty of Physics University of Barcelona Barcelona E-08028 Spain Niraj Sinha Department of Systems Design Engineering, University of Waterloo, 200 University Avenue (W), Waterloo, ON N2L 3G1, Canada. Manuel Va´zquez Instituto de Ciencia de Materiales CSIC 28049 Madrid Spain John T.W. Yeow Department of Systems Design Engineering, University of Waterloo, 200 University Avenue (W), Waterloo, ON N2L 3G1, Canada. Yoke Khin Yap Department of Physics Michigan Technological University Houghton, MI 49931

Chapter 1

Introduction Francisco J. Arregui

1.1 Some Data About Nanotechnology In this digital era in which we live, human beings are creating new devices, machines and systems that apparently make our lives easier, more pleasant and more comfortable. These new artifacts have to be incorporated in cheaper, faster, smaller and more complex ‘‘technological organisms’’ that organize themselves in cheaper, faster, smaller and more complex ‘‘technological organizations’’. Thanks to the effort that humanity is making now, perhaps the future of the world will be better but what is for sure is that we can achieve things today that were not possible yesterday. For instance, we can chat, talk and see our beloved ones when they are a world away from us, just by using a small computer in our living room. We can have access to the biggest encyclopedia that the man was able to create: Internet. We can have medicines, diagnosis tools and implants that can deal with health problems in a way that was unknown until today. This continuous progress is possible because all of us, in the Newtonian way, are able to see farther than others because we stand ‘‘on the shoulders of giants’’. As was stated above, it is obvious that the partial democratization of the technology, that is, the possibility to access computers that can connect each other in this world, has opened the door to the generation of a new kind of knowledge as a result of the cooperative effort of many. This revolution would not have been possible without the continuous innovation of the semiconductors. The Moore’s Law which very optimistically predicted that ‘‘the number of transistors that can be inexpensively placed on an integrated circuit will be increasing exponentially, doubling approximately every two years’’ has been proved not so optimistic and has been ruling the microlectronics market for more than 40 years now. That word ‘‘microelectronics’’ has been the key of all we are talking about now. The key to make possible that we (the authors of this book) could control and monitor new experiments, or just press the keyboard, to obtain strange symbols and graphs in the screen of a computer F.J. Arregui Electric and Electronic Engineering Department, Universidad Pu´blica de Navarra, Edificio de los Tejos, Campus Arrosadı´ a, 31006 Pamplona, Navarra, Spain

F.J. Arregui (ed.), Sensors Based on Nanostructured Materials, DOI: 10.1007/978-0-387-77753-5_1, Ó Springer ScienceþBusiness Media, LLC 2009

1

2

F.J. Arregui

and some months later, maybe years, some readers (yourself!) have the chance to see what we saw after working very hard to obtain our modest results. That part of the word ‘‘micro’’ has been with us in the last decades; ‘‘micro’’ has been part of the common travel that humanity initiated. But now when we are just beginning to see that the ‘‘micro’’ world is quite big, we are just beginning to observe even smaller things. Things that go beyond the micron, things that belong to the nanometer scale, to the nano reign. Half a century has already passed since Nobel laureate Richard P. Feynman stated in 1959 that there was ‘‘plenty of room at the bottom’’, what is considered by many as the formal initiation of nanotechnology [1]. The visionary challenge and proposal of Feynman pointed out that ‘‘At the atomic level, we have new kinds of forces and new kinds of possibilities, new kinds of effects. The problems of manufacture and reproduction of materials will be quite different’’. This aspect of nanotechnology, where a number of properties of the matter change when the nanoscopic level is reached, is what is very attractive to explore for the design of materials with new properties and applications not known yet. Today, we can already find nanotechnologic applications on diverse fields such as drug delivery, hard coatings or antimicrobial textiles. Just the appearance of the word ‘‘nano’’ in referred scientific papers has exploded as is shown in Fig. 1.1. In fact, in 2008 there are more scientific journal articles that include the word nano published in just 1 day, around 115, than in the whole year 1970. According to the databases of ISI Web of KnowledgeSM and ScopusTM, the sum of all the scientific journal articles that include this simple word, nano, has surpassed a quarter million articles (around 350,000 papers including conference papers).

45,000

Number of Journal Articles

40,000 35,000 30,000 25,000 20,000 15,000 10,000 5,000 0 07 20 06 20 05 20 04 20 03 20 02 20 01 20 00 20 99 19 98 19 97 19 96 19 95 19 94 19 93 19 92 19 91 19 90 19 89 19 88 19 87 19 86 19 85 19 84 19 83 19 82 19 81 19 80 19

Year of Publication

Fig. 1.1 Number of scientific journal articles with the word ‘‘nano’’ in the abstract, keywords or title since 1980 (data extracted from ScopusTM database)

1 Introduction

3

5%

5.18%

4.82%

3.94%

3.77%

2.99%

2.25%

1.88%

1.87%

2.08%

1.48%

1.92%

1.73%

2.45%

2.03%

0.10%

1.42%

0.12%

0.00%

0.17%

1%

0.37%

2%

2.87%

3%

4.63%

4%

0.00%

Ratio of Journal Articles (%)

6%

0% 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

Year of Publication

Fig. 1.2 Ratio of journal articles that include the word ‘‘sensor’’ from the ones which already include the word ‘‘nano’’ (data extracted from ISI Web of KnowledgeSM and ScopusTM databases)

Besides, not only the number of scientific publications with the word ‘‘nano’’ has increased exponentially, but in this specific group of articles around the nano-world the ratio of scientific publications related to sensors has also increased as is plotted in Fig. 1.2. After studying this data, it is easy to find an explanation to this trend, but if we have to name some of the reasons we could say that the advantages that nanotechnology introduces in the sensor field are many: obviously a decrease in the sensors’ dimensions (or at least the sensing films), an increase of surface area, a shorter diffusion time (therefore a shorter response time) and, the most interesting thing, the possibility of designing materials with tailored and new properties not presented in the bulky or macroscopic versions of the same compounds. The top 50 most cited journal articles that include the words ‘‘nano’’ and ‘‘sensor’’ when the introduction of this book is being written, May 2008, are indicated in the bibliography [2, 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, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51]. It is clear also that great economical efforts have been made by the governments, agencies and private foundations which have been able to share this vision. Today the tendency to dedicate funds to nanoscience and nanotechnology is increasing and numerous ‘‘nanotechnology centers’’ are appearing around the globe, usually linked to universities or research institutions. As an example of this institutional support, in 2000, the United States National Nanotechnology Initiative (NNI) was founded to coordinate federal nanotechnology research and development; the budget of this agency is plotted in Fig. 1.3. Until now, all things that have been said in this introduction with respect to nanotechnology are only advantages with respect to the consolidated technologies and with all the promising benefits of nanotechnology it is easy to fall in self-complacence. It is evident that when a new world is in front of our eyes, everything looks an adventure, a challenge and a new opportunity, and it is easy to set off the epidemic of a gold fever, the nano-fever because this is a new world

4

F.J. Arregui

1,527 *

2009*

1,491 *

2008*

1,425

2007

1,351

Year

2006 1,200

2005 989

2004 863

2003 697

2002 464

2001 0

200

400

600

800

1,000

1,200

1,400

1,600

NNI budget (US dollars in millions)

Fig. 1.3 The United States National Nanotechnology Initiative budget (source the NNI web site, www.nano.gov *2008 and 2009 show the estimated and proposed budgets, respectively)

with new rules, a new world to explore and discover and a new world that needs pioneers. Inevitably, a time for quest is also a time of confusion and it is easy to talk and promise gold but all that glitters is not gold and we have to be rigorous, hard workers and have the patience to pursue real objectives. Today, it is easy to see that many applications that are presented as ‘‘nanotechnology’’ are in fact a recasting of straightforward materials science, and some researchers already talk that there may be a danger that a ‘‘nano bubble’’ will form from the overuse of the term by scientists and entrepreneurs to collect funding, spoiling the transformative possibilities of nanotechnology. In this quest, we will have to find the equilibrium with this new nano-world, learn its new language and be respectful with its laws. On the other hand, sensors are our devices for perception of sight, hearing, taste, smell, touch and many more things beyond our five senses. Sensors can be used for detection of bacteria, viruses, explosives, biological markers, magnetic fields, infrared radiation, traces of pollutants, medical diagnostic tools and many other applications. Nanotechnology and sensors are two disciplines that have in common its high degree of interdisciplinarity and can be combined very easily. Since we are able now to synthesize new materials arranged at the nanometer scale we are also able to fabricate materials with properties different to the original bulk materials. This has opened the door to new and real applications for sensing. These nanostructured materials take advantage of the arrangement of matter on the atomic and molecular scales, generally 100 nm or smaller, and the fabrication of devices that lie within that size range, an intermediate size between molecular and microscopic (micrometer-sized) structures. There are different approaches to achieve these nanostructures; an easy classification is to

1 Introduction

5

make two big groups: top-down techniques and bottom-up techniques. The topdown techniques aim to fabricate smaller devices by using larger dimension objects; many of these technologies descended from the conventional semiconductor industry which is now capable of creating devices or arranging materials in dimensions smaller than 100 nm. The bottom-up techniques arrange structures from smaller components such as atom by atom or molecule by molecule. In current applications both approaches, top-down and bottom-up, tend to merge in order to take advantages of both techniques as will be seen in the different chapters of the book whose structure is described below.

1.2 About This Book This book reviews the works already presented in the literature with respect to the fabrication of sensors based on nanostructured materials. The book has been organized by topics of high interest. In order to offer a fast read of the state of art of each topic, every chapter in this book is independent and selfcontained. On the other hand, since nanotechnology is interdisciplinary by definition, some chapters overlap others and are in some way related between them. The nine chapters which follow the introduction of this book try to keep the same structure: first an introduction to the specific topic under study; second, the different fabrication methods of nanostructured materials for sensing in that particular field and, third, the sensing applications already reported in the literature of these devices. In Chapter 2 a review about different carbon nanotube (CNT) and fullerene sensors is presented. Chapter 3 deals with non-carbon nanotubes, and more specifically TiO2 nanotube arrays and its applications as hydrogen sensors, since this a less covered topic in the literature the chapter gives very detailed information. Nanowires, nanocombs, nanobelts, nanorods, nanoswords and nanosquids of mainly ZnO and SnO are presented in Chapter 4; these structures show some of the most beautiful arrangements in nature. Chapter 5 is focused on the utilization of metal oxides and semiconductors for controlling the transduction mechanisms of sensors at the nano scale, paying special attention to nanotemplates and nanowires. A review about the utilization of quantum dots for sensing is presented in Chapter 6 specially covering many of its applications in biosensing. Chapter 7 is a detailed review about the different types of nanostructured magnetic sensors: magnetic nanoparticles as well as magnetic nanowires and films. Encapsulated probes, term that refers to a class of multimolecular cocktails which have been physically encased within a protective package, are studied in Chapter 8. Optical fiber sensors based on nanostructured coatings are described in Chapter 9. The last chapter of the book, Chapter 10, deals with nanostructured flexible materials, more specifically presents some of the characteristics of Metal RubberTM, a commercially available product. Some examples of

6

F.J. Arregui

A

B

Fig. 1.4 Some examples of the nanostructures that are studied in this book. A TiO2 nanotubes for hydrogen sensing presented by the contributors in Chapter 2, B badmintonlike SiOx nanowires presented by the contributors in Chapter 4

these nanostructures can be seen in Fig. 1.4. The sensing applications that are covered along these nine chapters include temperature, pressure, strain, radiation, flow, magnetic fields, gas, volatile organic compounds, ions, chemicals, immunoassays, DNA detection, biological recognition, glucose, enzymatic detection, cell detection, magnetic resonance imaging (MRI) and others. Each chapter has been written by different contributors who give their personal perspective of these topics, a strategy that enriches the contents of the book because the chapters sometimes overlap and at least a few themes are dealt with from different angles. With respect to the contributors of this book, we have the great satisfaction of having convinced the most active authors in this field in the world to participate: according to ScopusTM database, 5 of the first 11 authors with more scientific journal publications with the words ‘‘nano’’ and ‘‘sensor’’ are authors of some chapters in this book. The first author of the most cited article with the words ‘‘nano’’ and ‘‘sensors’’ [2] (more than 2000 times according to Scopus) is also a contributor of this book. Most of the contributors are members in the editorial boards of journals related to the field and three of them are editors-in-chiefs of different journals. The purpose of this book is to make a humble contribution but a real one to some discipline that is emerging, thanks to the new applications of the nanotechnology: sensors based on nanostructured materials. This book does not deal ‘‘what it could be’’ but ‘‘what it is’’, already published results, a fact, tested devices and this is what, perhaps, makes this book valuable because there is not in the literature a collection dedicated to sensors based on nanostructured materials. We hope that readers enjoy it and that can be a valuable tool for those who want to have a summary of sensors fabricated with many diverse techniques, always dealing with nanostructured materials.

1 Introduction

7

Paradoxically, these nanostructures have been possible because we stood yet on the shoulders of the microdevices, and we do not have to forget that, thanks to those giants we are able to see farther; farther and deeper, in that place where there is plenty of room at the bottom.

Bibliography 1. R.P. Feynman (1992). ‘‘There’s plenty of room at the bottom’’, J Microelectromech Syst 1, 60–66. 2. Zheng Wei Pan, Zu Rong Dai, and Zhong Lin Wang (2001). ‘‘Nanobelts of semiconducting oxides’’, Science 291, 1947–1949. 3. R.H. Baughman, A.A. Zakhidov, and W.A. De Heer (2002). ‘‘Carbon nanotubes – the route toward applications’’, Science 297, 787–792. 4. J. Kong, N.R. Franklin, C. Zhou, M.G. Chapline, S. Peng, K. Cho, and H. Dai (2000). ‘‘Nanotube molecular wires as chemical sensors’’, Science 287, 622–625. 5. C.A. Mirkin, R.L. Letsinger, R.C. Mucic, and J.J. Storhoff (1996). ‘‘A DNA-based method for rationally assembling nanoparticles into macroscopic materials’’, Nature 382, 607–609. 6. Y. Cui, Q. Wei, H. Park, and C.M. Lieber (2001). ‘‘Nanowire nanosensors for highly sensitive and selective detection of biological and chemical species’’, Science 293, 1289–1292. 7. A. Zettl (2000). ‘‘Extreme oxygen sensitivity of electronic properties of carbon nanotubes’’, Science 287, 1801–1804. 8. M.S. Gudiksen, L.J. Lauhon, J. Wang, D.C. Smith, and C.M. Lieber (2002). ‘‘Growth of nanowire superlattice structures for nanoscale photonics and electronics’’, Nature 415, 617–620. 9. A.N. Shipway, E. Katz, and I. Willner (2000). ‘‘Nanoparticle arrays on surfaces for electronic, optical, and sensor applications’’, ChemPhysChem 1, 18–52, 2000. 10. M.A. Burns, B.N. Johnson, S.N. Brahmasandra, K. Handique, J.R. Webster, M. Krishnan, T.S. Sammarco, P.M. Man, D. Jones, D. Heldsinger, C.H. Mastrangelo, and D.T. Burke (1998). ‘‘An integrated nanoliter DNA analysis device’’, Science 282, 484–487. 11. M.R. Falvo, G.J. Clary, R.M. Taylor II, V. Chi, F.P. Brooks Jr., S. Washburn, and R. Superfine (1997). ‘‘Bending and buckling of carbon nanotubes under large strain’’, Nature 389, 582–584. 12. P. Alivisatos (2004). ‘‘The use of nanocrystals in biological detection’’, Nat Biotechnol 22, 47–52. 13. C.L. Haynes and R.P. Van Duyne (2001). ‘‘Nanosphere lithography: A versatile nanofabrication tool for studies of size-dependent nanoparticle optics’’, J Phys Chem B 105, 5599–5611. 14. A.C. Comely and S.E. Gibson (2001). ‘‘Self-assembling organic nanotubes’’, Angew Chem Int Ed 40, 988–1011. 15. H.G. Craighead (2000). ‘‘Nanoelectromechanical systems’’, Science 290, 1532–1535. 16. I. Willner and E. Katz (2000). ‘‘Integration of layered redox proteins and conductive supports for bioelectronic applications’’, Angew Chem Int Ed 39, 1181–1218. 17. J.W. Perry, K. Mansour, I.-S. Lee, X. Wu, P.V. Bedworth, C. Chen, D. Ng, S.R. Marder, P. Miles, T. Wada, M. Tian, and H. Sasabe (1996). ‘‘Organic optical limiter with a strong nonlinear absorptive response’’, Science 273, 1533–1536. 18. R.J. Chen, S. Bangsaruntip, K.A. Drouvalakis, N. Wong Shi Kam, M. Shim, Y. Li, W. Kim, P.J. Utz, and H. Dai (2003). ‘‘Noncovalent functionalization of carbon nanotubes for highly specific electronic biosensors’’, Proc Natl Acad Sci USA 100, 4984–4989.

8

F.J. Arregui

19. F. Favier, E.C. Walter, M.P. Zach, T. Benter, and R.M. Penner (2001). ‘‘Hydrogen sensors and switches from electrodeposited palladium mesowire arrays’’, Science 293, 2227–2231. 20. E. Comini, G. Faglia, G. Sberveglieri, Z. Pan and Z.L. Wang (2002). ‘‘Stable and highly sensitive gas sensors based on semiconducting oxide nanobelts’’, Appl Phys Lett 81, 1869. 21. H. Dai (2002). ‘‘Carbon nanotubes: Synthesis, integration, and properties’’, Acc Chem Res 35, 1035–1044. 22. J. Huang, S. Virji, B.H. Weiller, and R.B. Kaner (2003). ‘‘Polyaniline nanofibers: Facile synthesis and chemical sensors’’, J Am Chem Soc 125, 314–315. 23. R.A. Wolkow (1999). ‘‘Controlled molecular adsorption on silicon: Laying a foundation for molecular devices’’, Annu Rev Phys Chem 50, 413–441. 24. T.G. Drummond, M.G. Hill, and J.K. Barton (2003). ‘‘Electrochemical DNA sensors’’, Nat Biotechnol 21, 1192–1199. 25. X.Y. Kong and Z.L. Wang (2003). ‘‘Spontaneous polarization-induced nanohelixes, nanosprings, and nanorings of piezoelectric nanobelts’’, Nano Lett 3, 1625–1631. 26. Z.R. Dai, Z.W. Pan, and Z.L. Wang (2003). ‘‘Novel nanostructures of functional oxides synthesized by thermal evaporation’’, Adv Funct Mater 13, 9–24. 27. M.S. Arnold, P. Avouris, Z.W. Pan, and Z.L. Wang (2003). ‘‘Field-effect transistors based on single semiconducting oxide nanobelts’’, J Phys Chem B 107, 659–663. 28. I.L. Medintz, A.R. Clapp, H. Mattoussi, E.R. Goldman, B. Fisher, and J.M. Mauro (2003). ‘‘Self-assembled nanoscale biosensors based on quantum dot FRET donors’’, Nat Mater 2, 630–638. 29. J.G.G. Borst and B. Sakmann (1996). ‘‘Calcium influx and transmitter release in a fast CNS synapse’’, Nature 383, 431–434. 30. X. Wang, C.J. Summers, and Z.L. Wang (2004). ‘‘Large-scale hexagonal-patterned growth of aligned ZnO nanorods for nano-optoelectronics and nanosensor arrays’’, Nano Lett 4, 423–426. 31. K. Kalyanasundaram and M. Gra¨tzel (1998). ‘‘Applications of functionalized transition metal complexes in photonic and optoelectronic devices’’, Coord Chem Rev 177, 347–414. 32. T. Bein (1996). ‘‘Synthesis and applications of molecular sieve layers and membranes’’, Chem Mater 8, 1636–1653. 33. H. Zheng, J. Wang, S.E. Lofland, Z. Ma, L. Mohaddes-Ardabili, T. Zhao, L. SalamancaRiba, S.R. Shinde, S.B. Ogale, F. Bai, D. Viehland, Y. Jia, D.G. Schlom, M. Wuttig, A. Roytburd, and R. Ramesh (2004). ‘‘Multiferroic BaTiO3-CoFe2O4 Nanostructures’’, Science 303, 661–663. 34. J. Wang and M. Musameh (2003). ‘‘Carbon nanotube/Teflon composite electrochemical sensors and biosensors’’, Anal Chem 75, 2075–2079. 35. K. Besteman, J. Lee, F.G.M. Wiertz, H.A. Heering, and C. Dekker (2003). ‘‘Enzymecoated carbon nanotubes as single-molecule biosensors’’, Nano Lett 3, 727–730. 36. R. Cush, J.M. Cronin, W.J. Stewart, C.H. Maule, J. Molloy, and N.J. Goddard (1993). ‘‘The resonant mirror: A novel optical biosensor for direct sensing of biomolecular interactions. Part I: Principle of operation and associated instrumentation’’, Biosens Bioelectron 8, 347–353. 37. N. Nath and A. Chilkoti (2002). ‘‘A colorimetric gold nanoparticle sensor to interrogate biomolecular interactions in real time on a surface’’, Anal Chem 74, 504–509. 38. M. Law, H. Kind, B. Messer, F. Kim, and P. Yang (2002). ‘‘Photochemical sensing of NO2 with SnO2 nanoribbon nanosensors at room temperature’’, Angew Chem Int Ed 41, 2405–2408. 39. R. Skomski (2003). ‘‘Nanomagnetics’’, J Phys Condens Matter 15. 40. M. Bognitzki, W. Czado, T. Frese, A. Schaper, M. Hellwig, M. Steinhart, A. Greiner, and J.H. Wendorff (2001). ‘‘Nanostructured fibers via electrospinning’’, Adv Mater 13, 70–72.

1 Introduction

9

41. M. Trau, N. Yao, E. Kim, Y. Xia, G.M. Whitesides, and I.A. Aksay (1997). ‘‘Microscopic patterning of orientated mesoscopic silica through guided growth’’, Nature 390, 674–676. 42. A.D. McFarland and R.P. Van Duyne (2003). ‘‘Single silver nanoparticles as real-time optical sensors with zeptomole sensitivity’’, Nano Lett 3, 1057–1062. 43. B. Mayers, B. Gates, Y. Yin, and Y. Xia (2001). ‘‘Functionalized carbon nanotubes for molecular hydrogen sensors’’, Adv Mater 13, 1384–1386. 44. J. Hahm and C.M. Lieber (2004). ‘‘Direct ultrasensitive electrical detection of DNA and DNA sequence variations using nanowire nanosensors’’, Nano Lett 4, 51–54. 45. J.R. Grider, A.E. Foxx-Orenstein, and J. Jin (1998). ‘‘5-Hydroxytryptamine4 receptor agonists initiate the peristaltic reflex in human, rat, and guinea pig intestine’’, Gastroenterology 115, 370–380. 46. Z.L. Wang (2004). ‘‘Zinc oxide nanostructures: Growth, properties and applications’’, J Phys Condens Matter 16. 47. Q. Wan, Q.H. Li, Y.J. Chen, T.H. Wang, X.L. He, J.P. Li, and C.L. Lin (2004). ‘‘Fabrication and ethanol sensing characteristics of ZnO nanowire gas sensors’’, Appl Phys Lett 84, 3654–3656. 48. S.F. Bent (2002). ‘‘Organic functionalization of group IV semiconductor surfaces: Principles, examples, applications, and prospects’’, Surf Sci 500, 879–903. 49. S.Y. Chou (1997). ‘‘Patterned magnetic nanostructures and quantized magnetic disks’’, Proc IEEE 85, 652–671. 50. D. Routkevitch, A.A. Tager, J. Haruyama, D. Almawlawi, M. Moskovits, and J.M. Xu (1996). ‘‘Nonlithographic nano-wire arrays: Fabrication, physics, and device applications’’, IEEE Trans Electron Devices 43, 1646–1658. 51. P. Gomez-Romero (2001). ‘‘Hybrid organic-inorganic materials – in search of synergic activity’’, Adv Mater 13, 163–174.

Chapter 2

Carbon Nanotube and Fullerene Sensors John T.W. Yeow and Niraj Sinha

2.1 Introduction The first fullerene was discovered in 1985 by Sir Harold W. Kroto from the University of Sussex and Richard E. Smalley and Robert F. Curl Jr. from Rice University [1] inadvertently when they were studying the nucleation of carbon in the atmosphere of a cool carbon-rich red giant star. Fullerenes refer to a family of carbon allotropes. Each carbon molecule is composed of at least 60 carbon atoms such as C60. When the atoms are arranged in the form of hollow sphere, it is referred to as buckyballs. Fullerenes that take the form of a cylinder are referred to as carbon nanotubes (CNTs). By 1990, it was relatively easy to synthesize macroscopic quantities of C60. Donald Huffman of University of Arizona and Wolfgang Kratschmer of Max Planck Institute developed a technique by which C60 was produced by evaporating graphite electrodes via resistive heating in an atmosphere of 100 Torr helium [2]. Since the discovery of CNTs in 1991 by Iijima [3], the interest in practical applications of fullerenes has skyrocketed. Carbon nanotubes (CNTs), which can essentially be thought of as a layer of graphite rolled up into a cylinder, have shown great promise in the field of nano electromechanical systems (NEMS). The advantage of CNTs over other materials is due to their small size, high strength, high electrical and thermal conductivity, and high specific area. There are two types of CNTs: singlewalled nanotubes (SWNTs) and multi-walled nanotubes (MWNTs). These two types of CNTs differ in the arrangement of their graphene cylinders. While SWNTs have only one layer of graphene cylinder, MWNTs have many layers. Although SWNTs are structurally similar to a single layer of graphite (that is, a semiconductor with zero band gap), they can be either metallic or semiconducting depending on the tube diameter and the chirality (the sheet direction in which the graphite sheet is rolled to form a nanotube cylinder) [4, 5].

J.T.W. Yeow Department of Systems Design Engineering, University of Waterloo, 200 University Avenue (W), Waterloo, ON N2L 3G1, Canada

F.J. Arregui (ed.), Sensors Based on Nanostructured Materials, DOI: 10.1007/978-0-387-77753-5_2, Ó Springer ScienceþBusiness Media, LLC 2009

11

12

J.T.W. Yeow, N. Sinha

The diameter (d) and the chiral angle () can be obtained by an integer pair (n, m) using Eqs. 2.1 and 2.2 [6]: pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi a m2 þ mn þ n2 p " pffiffiffiffiffi # 3n  ¼ arctan 2m þ n

d¼

(2:1)

(2:2)

where a is the lattice constant in the graphite sheet. Depending on the relation between n and m, three categories of CNTs are defined: (i) armchair (n = m and chiral angle equal to 308); (ii) zigzag (n = 0 or m = 0 and chiral angle = 08); and (iii) chiral (other values of n and m and chiral angles between 0 and 308) [7]. All armchair nanotubes are metals, as well as those with n  m = 3j (j being a nonzero integer). All others are semiconductors, which have the band gap that is inversely related to the diameters of the nanotubes [8].

2.2 Carbon Nanotube Synthesis Techniques For synthesis of defect-free CNTs of macroscopic lengths in desired quantities, development of reliable synthesis techniques is essential. Controlling the chirality of CNTs for a specific application is very challenging. The state-of-the-art synthesis techniques produce statistical distributions of chiralities, and hence electrical properties [9]. In general, following three techniques are used for synthesizing CNTs: (i) carbon arc-discharge technique; (ii) laser-ablation technique; and (iii) chemical vapor deposition (CVD) technique. The three techniques are discussed in detail below.

2.2.1 Carbon Arc-Discharge Technique In the carbon arc-discharge technique, two carbon electrodes are kept in a vacuum chamber. The electrodes are used to generate an arc by DC current. An inert gas is supplied to the chamber to increase the speed of carbon deposition. After the stabilization of pressure, the power supply is turned on (about 20 V) and the positive electrode is gradually brought closer to the negative electrode to strike the electric arc. The electrodes become red hot and a plasma forms. The rods are kept about a millimeter apart upon stabilization of the arc. During this period, the CNT deposits on the negative electrode. The power supply is cut off and the machine is left for cooling once a specific length is reached. The two most important parameters to be taken care of in this method

2 Carbon Nanotube and Fullerene Sensors

13

are (i) the control of arcing current and (ii) the optimal selection of inert gas pressure in the chamber [10]. Using arc-discharge technique, MWNTs were first discovered by Iijima [3] in 1991 when he saw very thin and long tubes of pure carbon under electron microscope. The growth of SWNTs for the first time was demonstrated by Iijima and Ichihashi [11] and Bethune et al. [12] in 1993 using arc-discharge technique. Arc-discharge technique produces high-quality CNTs. While MWNTs do not need a catalyst for growth, SWNTs can only be grown in the presence of a catalyst. MWNTs can be obtained by controlling the pressure of inert gas in the discharge chamber and the arcing current. The by-products are polyhedron-shaped multi-layered graphitic particles in case of MWNTs. The growth of SWNTs by arc-discharge technique was optimized by Journet et al. [13] using graphite cathode (16 mm diameter, 40 mm long), graphite anode (6 mm diameter, 100 mm long), mixture of catalysts (Ni–Co, Co–Y, or Ni–Y), helium pressure of 660 mbar, arcing current of 100 A, and voltage drop of 30 V between the electrodes. The scanning electron microscopy (SEM) revealed that the deposited material consisted of high amount of entangled carbon ropes of diameters 5–20 nm. Li et al. [14] synthesized SWNTs by modifying the arcdischarge method and by using FeS as a promoter. As evaluated by scanning electron microscopy, thermogravimetric analysis, and Raman spectroscopy, the synthesized SWNT fibers were 80% pure by volume. Later, Paladugu et al. [15] demonstrated that CNTs can be synthesized by arc-discharge in open air. As their method does not require a controlled atmosphere, the cost of production may be reduced.

2.2.2 Laser-Ablation Technique Intense laser pulses are utilized in the laser-ablation technique to ablate a carbon target, which in the presence of an inert gas and catalyst forms CNTs. An analysis by X-ray diffraction (XRD) and transmission electron microscopy (TEM) revealed that the SWNTs produced by Thess et al. [16] using laser ablation were ropes (or bundles) of 5–20 nm diameter and tens to hundreds of micrometers of length. It was found by Arepalli et al. [17] that individual nanotubes of lengths tens of microns are formed in the vicinity of the target at the beginning, which subsequently coalesce into bundles. Based on spectral emission and laser-induced fluorescence measurements, Scott et al. [18] suggested that the carbon for the formation of CNTs comes from direct ablation as well as from carbon particles suspended in the reaction zone. In addition, the confinement of CNTs in the reaction zone within the laser beam leads to the purification and annealing of CNTs by laser heating. Braidy et al. [19] used pulsed KrF laser ablation of a graphite pellet to synthesize SWNTs and other nanotubular structures. It was observed by them that relatively high UV laser intensity had an adverse effect on the growth of SWNTs. By using high-vacuum laser ablation, multi-layered MWNTs were grown selectively

14

J.T.W. Yeow, N. Sinha

by Takahashi et al. [20] by dispersing graphite powder on a Si (100) substrate. During the growth of MWNTs, high substrate temperature was maintained by them. In general, the amount and type of catalysts, laser power and wavelength, temperature, pressure, type of inert gas present, and the fluid dynamics near the carbon target are some of the parameters that determine the amount of CNTs produced [21]. The by-products of SWNTs in case of arc-discharge and laserablation techniques are fullerenes, graphitic polyhedrons with enclosed metal particles, and amorphous carbon [21].

2.2.3 Chemical Vapor Deposition Technique In CVD, energy is imparted to hydrocarbons (the commonly used sources are methane, ethylene, and acetylene) to break them into reactive radical species in the temperature range of 550–7508C. These reactive species diffuse down to a heated and catalyst-coated surface where they remain bonded. As a result, CNTs are formed. The commonly used energy sources are electron beam and resistive heating. By catalytic decomposition of acetylene over iron particles at 7008C, microtubules of up to 50 mm length of CNTs were synthesized by Yacaman et al. [22]. Vardan and Xie [23] developed a CVD technique, which used microwave energy for synthesizing MWNTs. The use of acetylene as the hydrocarbon and cobalt as the catalyst at a temperature of 7008C resulted in MWNTs with 26 layers and average diameter of 20–30 nm. A sequential combination of radio frequency plasma-enhanced CVD (RF PECVD) and thermal CVD was used by Park et al. [24] to synthesize CNTs from acetylene and hydrogen gas mixture on stainless steel plates. Wei et al. [25] used CVD with gas-phase catalyst delivery to direct the assembly of CNTs in a variety of predetermined orientations, building them into one-, two-, and three-dimensional arrangements. The key parameters that affect the synthesis of CNTs by CVD include the nature of hydrocarbons, catalysts, and the growth temperature.

2.2.4 Purification In all the three synthesis techniques mentioned above, the CNTs come with a number of impurities. The type and amount of impurities depend on the synthesis technique that was used to produce CNTs. The purity of CNTs can be analyzed by spectroscopic techniques. For example, Raman spectroscopy is commonly used for qualitative evaluation of purity of CNTs, while near infrared spectroscopy can be used for the quantitative assessment. Carbonaceous materials are the most common impurities observed. The other types of impurities include metals. As carbonaceous impurities have high oxidation rates, the

2 Carbon Nanotube and Fullerene Sensors

15

impurities in the carbon arc-discharge technique can be purified by oxidation. Generally, two approaches are followed for purification by oxidation: (i) gasphase oxidation and (ii) liquid phase oxidation. Ebbesen et al. [26] used gas-phase oxidation for purification of CNTs. As low yield of purification was observed by them, liquid phase oxidation for better homogeneity was tried by Hiura et al. [27]. Bandow et al. [28] used a cationic surfactant and trapped SWNTs on a membrane filter to purify SWNTs synthesized by pulsed laser-ablation technique. Although high purity by weight (90%) was observed after purification, this technique was not found very useful for purifying large batches. To overcome this limitation, a macroscale purification technique was proposed by Rinzler et al. [29]. Xu et al. [30] developed a process for purification of SWNTs grown by CVD of carbon monoxide that included sonication, oxidation, and acid washing steps. For MWNTs grown by CVD, Biro et al. [31] used wet and dry oxidation to remove impurities and traces of catalysts. Several other techniques have been proposed to purify CNTs. However, they may change the electrical and mechanical properties of CNTs since the structural surfaces of CNTs are modified after purification. Therefore, current research focus is on producing high-purity CNTs directly.

2.3 Carbon Nanotube and Fullerene Sensors The sensor market is expanding at a very fast rate. Some of the potential applications of sensors are in biomedical industry, environmental monitoring, agricultural industry, fishing industry, food industry, automotive industry, electronics industry, defence and homeland security. All these applications require improved sensitivity, selectivity and stability beyond the commercially available sensors currently available. In order to meet these requirements, many different approaches have been considered by CNT research community, and several manuscripts related to CNT and fullerene sensors have been published [32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 6162, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120]. This section presents a survey of sensors, which utilize the remarkable properties of CNTs. As is common with any emerging field, there are many proof-of-concept prototypes available for CNT and fullerene sensors. However, commercially available CNT sensors are not very common.

2.3.1 Force Sensors – Pressure and Strain For the first time, the potential of SWNTs as molecular and macroscopic pressure sensors was demonstrated by Wood and Wagner [32]. They applied high hydrostatic pressures to SWNTs by using a diamond anvil and recorded

16

J.T.W. Yeow, N. Sinha

the micro-Raman spectroscopy. The Raman spectrum of the CNTs was monitored under different pressures. The shifts in various Raman bands were observed simultaneously. The observed shifts in Raman peaks were found to be highly reversible, which showed potential of CNTs in pressure sensing based on mechanism of Raman band shifts. Liu and Dai [33] grew SWNTs on suspended square polysilicon membranes and demonstrated that pressure sensors can be realized by using their piezoresistive properties. A change in resistance in the SWNTs was observed when uniform air pressure was applied on the membranes. Moreover, the membrane was restored to its original condition when the gas was pumped out, which indicated that the process was reversible. In their study, Wu et al [34] demonstrated using first-principle quantum transport calculations, molecular-dynamics simulation, and continuum mechanics analysis that hydrostatic pressure can induce radial deformation. The radial deformation results in electrical transition of SWNTs. Transition from metal-to-semiconductor property in armchair SWNTs was observed when pressure was applied. This observation provides a basis for designing nanoscale tunable pressure sensors. Later, Fung et al. [35] demonstrated that piezoresistive pressure sensors can be built using MWNTs. They positioned a network of MWNTs across a PMMA membrane using dielectrophoretic (DEP) manipulation. When pressure was applied, the membrane deflected causing a bending in MWNTs. The advantage of this device is that it can be made through conventional micromachining processes. It was argued by Dharap et al. [38] that the conventional sensors have disadvantage that they are discrete point, fixed directional, and are not embedded at the material level. To overcome these limitations, a CNT film sensor for strain sensing on macroscale was presented by them. It was based on the principle that the electronic properties of CNTs change when subjected to strains. Since randomly oriented bundles of SWNTs were used, the film was isotropic in nature. The isotropic nature of CNT films helps in measuring strains in multiple locations and in different directions. A nearly linear relationship between the measured change in voltage and the strains in CNT films was observed experimentally when the films were subjected to tensile and compressive stresses. In another study, Li and Chou [39] developed SWNT-based sensors to measure strain and pressure at nanoscale. The sensors were based on the shift in resonant frequency of carbon nanotube resonator when subjected to a strain resulting from an external loading. Simulation studies by atomistic modeling revealed that the resonant frequency shifts are linearly dependent on the applied axial strain and the applied pressure. In addition, it was found that the reduction in tube length and diameter enhances the sensitivities of sensors. Berdinsky et al. have reported the use of fullerene-based sensor for temperature and pressure detection [40]. Pure fullerene was evaporated at a temperature of about 600–6808C in an atmosphere of argon gas at a flow rate of 5–15 ml/min. The argon pressure within the chamber was 670–1340 Pa. The

2 Carbon Nanotube and Fullerene Sensors

17

thickness of the fullerene thin film deposited on the substrates was 2–3 mm. The response of the sensor to changes in temperature, humidity, and pressure is shown in Fig. 2.1a–c. It was demonstrated that the sensitivities of the sensors were increased by subjecting the fullerene thin film to an oxidation process.

(a)

(b)

Fig. 2.1 (a) Temperaturedependent resistance change of fullerene thin films that are treated at various treatment temperatures: (1) initial sample; (2) 3708C; (3) 3908C. (b) Humidity-dependent resistance change of fullerene thin films: (1) initial sample; (2) treated at 3008C. (c) Pressure-dependent resistance change of fullerene thin films that are treated at 3908C at various testing conditions: (1) 208C; (2) 308C; (3) 708C [40]

(c)

18

J.T.W. Yeow, N. Sinha

2.3.2 Flow Sensors Ghosh et al. [41] have proposed a flow sensor based on SWNTs. The sensor is based on the generation of a current/voltage in a bundle of SWNTs, when the bundle is kept in contact with a flowing fluid. Generation of such current/ voltage was theoretically predicted by Kral and Shapiro [42] earlier. In general, an electric current is generated when the flow of free charge carrier is induced in any material. According to Kral and Shapiro, the electric current generation in CNTs is due to the transfer of momentum from the flowing liquid molecules so as to have a dragging effect on the free charge carriers in the nanotubes. The predicted relationship between the electric current and the fluid flow velocity is linear, which is in sharp contrast to the experimental findings of Ghosh et al. [41]. It was found experimentally that the induced voltage fits logarithmic velocity dependence over nearly six decades. It was observed experimentally that the ionic strength of the flowing liquid significantly affected the induced voltage. For flow velocities of the order of 10–5 m/s, the induced voltage was found to be saturated. The experimental data points fitted empirically to a logarithmic equation. The CNT-based flow sensors presented by Ghosh et al. [41] has great potential to be used in micromachines that work in fluidic environment such as heart pacemakers, which need neither heavy battery packs nor recharging.

2.3.3 Temperature Sensors As temperature changes, CNTs embedded in polymer matrices exhibit a shift in the Raman D* band. It was observed by Wood and Wagner [32] that the wavenumber of D* band increases with decreasing temperatures as CNTs experience compression. Although the proof-of-concept prototype demonstrates the potential of CNT composites for temperature sensing, an indepth study is warranted to understand the dispersion and interconnectivity of the CNT network. In another study, Wong and Li [43] used electrical properties of CNTs and developed MWNT-based thermal sensors. They manipulated bulk MWNTs by AC electrophoresis to form resistive elements between gold microelectrodes and demonstrated that MWNTs can potentially serve as temperature and anemometry sensors. The MWNT sensor chip was put inside an oven after packaging on a PCB for data acquisition. The oven temperature was monitored by a Fluke type-k thermocouple, which was attached on the surface of the PCB. As the temperature inside the oven was varied, change in resistance of MWNT sensors was measured. The representative data set indicated a linear relationship between the resistance and the temperature in the range of 20–608C. The I–V measurements of the devices revealed power consumption in mW range when MWNTs were used in

2 Carbon Nanotube and Fullerene Sensors

19

constant current configuration, which indicated that CNTs could be a promising material to fabricate ultra low-power-consumption thermal sensors.

2.3.4 Chemical Sensors Kong et al. [44] reported that the electrical resistance of semiconducting SWNTs dramatically changes when exposed to gaseous molecules such as nitrogen dioxide (NO2), ammonia (NH3), and oxygen (O2). They found that the nanotube sensors are at least an order of magnitude faster than those based on solid-state sensors. In addition to their small size, semiconducting SWNTs operate at room temperature with sensitivity as high as 103. This forms a basis for building semiconducting SWNT-based chemical sensors. However, it was pointed by Modi et al. [45] that the carbon nanotube gas sensors based on electrical conductance changes have certain limitations. The limitations include poor diffusion kinetics, inability to identify gases with low adsorption energies, and low capability to distinguish between gases or gas mixtures. It was also argued that the conductance of CNTs is highly sensitive to changes in moisture, temperature, and gas-flow velocity. Gas ionization sensors, featuring the electrical breakdown of a range of gases and gas mixtures at the tips of CNTs, were proposed by them to overcome these limitations. The cathode used for the purpose was aluminum and the anode was vertically aligned MWNT film (25–30 nm in diameter, 30 mm in length, and 50 nm separation between nanotubes) grown on SiO2 substrate. A glass insulator was used to separate the electrodes. The proposed sensors demonstrated by them were found to have good selectivity and sensitivity and were unaffected by various environmental conditions (moisture, temperature, and gas flow). Chopra et al. [51] developed microwave resonant sensors coated with either SWNTs or MWNTs for detection of ammonia. Comparative experiments revealed that SWNT sensors were more sensitive than the MWNT sensors. This sensor system can be used for applications that prohibit the use of physical connections or require non-destructive testing. In another study, Chopra et al. [65] demonstrated that SWNTs can be used as chemical sensors for detection of other gases in addition to NH3. The sensors developed by them showed sensitivity to CO, N2, He, O2, and Ar. Vapor sensors based on SWNT field effect transistors (FETs) have been developed by Someya et al. [66] for the detection of alcohols. The structure of the FET sensor and the corresponding experimental geometry are schematically shown in Fig. 2.2. The drain current measurements as a function of time are shown in Fig. 2.3. A sharp spike is observed few seconds after the saturated ethanol vapor is delivered to the surface and then the current decreases and reaches a steady value. Staii et al. [69] have proposed chemical sensors based on single-stranded DNA (ss-DNA) as the chemical recognition site and SWNT field effect transistors as the electronic read-out component. These sensors have rapid response and fast recovery times

20

J.T.W. Yeow, N. Sinha

Fig. 2.2 Cross-sectional structure of the FET-based sensor and the experimental geometry. Reprinted with kind permission from [66], Someya et al., Nano Lett., 3, 877 (2003). Copyright @ 2003 American Chemical Society

Fig. 2.3 Drain current measurements as a function of time with a source–drain bias of –100 mV and a gate bias of –10 V. Reprinted with kind permission from [66], Someya et al., Nano Lett., 3, 877 (2003). Copyright @ 2003 American Chemical Society

on the scale of seconds and are able to detect a variety of gases. The schematic of the experimental setup and the gases used in the experiment is shown in Fig. 2.4. The change in sensor current upon exposure to different gases is shown in Fig. 2.5. These sensors are self-regenerating: samples maintain a constant response with no need for sensor refreshing for approximately 50 gas exposure cycles. These features can be used for applications ranging from homeland security to disease diagnosis. It was observed by Snow et al. [71] that the capacitance of SWNTs is highly sensitive to a wide range of vapors. Therefore, this property can be utilized to develop fast, low-power sorption-based chemical sensors. In another study, Jang et al. [72] proposed a chemical sensor employing laterally grown MWNTs as the active sensing element and found that the electrical resistance of MWNTs changes upon exposure to air or NH3. Also, as the measurement temperature and gas concentration were increased, fast response time and higher sensitivity were observed. Penza et al. [73] developed surface acoustic wave (SAW) sensors that were coated by CNTs for chemical detection of volatile organic compounds (such as ethanol, ethyl acetate, and toluene in

2 Carbon Nanotube and Fullerene Sensors

21

Fig. 2.4 (a) Schematic of experimental setup. (b) Gases used in the experiment. Reprinted with kind permission from [69], Staii et al., Nano Lett., 5, 1774 (2005). Copyright @ 2005 American Chemical Society

Fig. 2.5 Change in sensor current upon exposure to different gases. Reprinted with kind permission from [69], Staii et al., Nano Lett., 5, 1774 (2005). Copyright @ 2005 American Chemical Society

nitrogen). These sensors were found to be highly sensitive during experiments. Ong et al. [74] built a gas sensor comprising MWNT-silicon dioxide (SiO2) composite on the principle that the conductivity and permittivity of the composite change with the absorption of different gases in the MWNT-SiO2 layer.

22

J.T.W. Yeow, N. Sinha

The sensor developed by them has the advantage of allowing remote monitoring conditions inside the opaque, sealed containers. Szymanska et al. [75] synthesized a supported lipid bilayer membrane (S-BLM) that was modified using C60 for the detection of neutral odorant molecules. C60 displays good electron transfer properties and behaves as moderate electron acceptor molecules. Bilayer lipid membranes were formed on cleaved Teflon-coated stainless steel wire. The structure was immersed in a lipid solution with 2% l--phosphatidylcholine, 0.5% cholesterol in n-decane saturated with C60. The C60-enhanced electrochemical sensor showed a general larger response to neutral smell compounds. Fig. 2.6 shows the schematic of the C60 electrochemical sensor.

2.3.5 Biosensors It has been found that the electrical properties of CNTs are highly sensitive to external charges. Even small amount of electron transfer between CNTs and adsorbates may result in significant change in the conductance of the device. Since biomolecules such as DNA, RNA, and proteins are heavily charged molecules, their adsorption onto the CNT surface can change the electronic properties of CNTs in a similar way as gas molecules. Therefore, CNTs can be potentially used to develop biosensors. An amperometric biosensor was developed by Sotiropoulou and Chaniotakis [102] using CNTs as immobilization matrix. Aligned MWNTs were grown on platinum (Pt) substrate, which served as the transduction platform for signal monitoring. The schematic diagram of the CNT array biosensor is shown in Fig. 2.7. The CNT arrays were purified by

Fig. 2.6 Schematic of electrochemical sensor based on C60-modified s-BLM for detection of neutral odorant compounds[75]

2 Carbon Nanotube and Fullerene Sensors

23

Pt O H O O H O O H O O H O O H O

O H

e–

Glucose

O O H O O HH O O H O O H O

Gluconic Acid

O H O O H O O H O O H O O H O

Fig. 2.7 Schematic diagram of the CNT array biosensor. Reprinted with kind permission from [102], Sotiropoulou and Chaniotakis, Anal. Bioanal. Chem., 375, 103 (2003). Copyright @ Springer Science and Business Media (2002)

acid and air treatments. The SEM images of the Pt-aligned CNT arrays are shown in Fig. 2.8. After the enzyme was immobilized, the response and sensitivity of the sensor treated by acid were found to be very high when compared to the air-treated sensor. In another study, a single-strand DNA chain was chemically attached onto the surface of a CNT suspended on gold electrodes [106]. This helped in detection of complementary DNA and/or target DNA chains of specific sequences. To overcome the challenges of tissue penetration and the natural autofluorescent media, Barone et al. [109] developed near-infrared optical sensors based

Fig. 2.8 SEM images of Pt-aligned CNT arrays (a) In original state, (b) after acid treatment, and (c) after air oxidation. Reprinted with kind permission from [102], Sotiropoulou and Chaniotakis, Anal. Bioanal. Chem., 375, 103 (2003). Copyright @ Springer Science and Business Media (2002)

24

J.T.W. Yeow, N. Sinha

on SWNTs. The sensors were developed making use of the fact that CNTs fluoresce in a region of the near infrared where human tissue and biological fluids are particularly transparent to their emission. Zhang and Gorski [112] have developed an electrochemical sensing platform based on the integration of redox mediators and CNTs in a polymeric matrix. It was found that the incorporation of CNTs decreased the overpotential for the mediated process by an extra 0.30 V and reduced the response time from 70 to 5 s. This concept can find useful applications in electrochemical devices such as sensors, biosensors, and biological fuel cells and reactors.

2.3.6 Radiation Sensors Ionization chambers are currently the most important dosimeters due to their sensitivity and relatively flat energy response. However, their applications for in vivo dose measurements are limited because of their large size and high bias voltage requirements for achievement of acceptable ionization collection efficiency. To this end, radiation sensors utilizing CNTs have been proposed [119, 120]. The radiation sensor developed by Ma et al. [120] is based on the principle that upon interaction between ionization radiation and the gas in the active volume of the detector chamber, energy is transferred by high-energy particles to the gas molecules. As a result, the gas molecules split into electron and hole pairs. When an external electric field is applied, holes and electrons move to the cathode and anode, respectively, which can be used to quantify the radiation dosage. The sensor was characterized by signal saturation characteristic, signal response linearity, electrode separation distance, and oblique incident beam measurements. Excellent linear responses to exposure and high sensitivity to oblique incident beams were found. Overall, these results indicate that the ionization collection efficiency of CNTs can be utilized to miniaturize the state-of-the-art ionization chambers and lower the bias voltages.

2.4 Conclusions The discovery of fullerene as the third form of pure carbon besides diamond and graphite has started a new field of fullerene science. This important discovery in the laboratory will be felt by the general public as many practical applications of CNTs and fullerenes are developed. In particular, improved sensors with increased sensitivity and selectivity have already been demonstrated by incorporating these nanomaterials. It is anticipated that the intense research on CNT and fullerene-based sensors will have a significant impact on a wide range of industries. Hospitals will be able to provide timely and accurate diagnosis by having biosensors that respond to minute quantity of targeted analytes. Industries will benefit by having miniature sensors that detect pressure, temperature,

2 Carbon Nanotube and Fullerene Sensors

25

and other physical parameters. Multiple sensors could be deployed simultaneously as companies take advantage of the small size and mass of the sensors. Environmental monitoring such as chemical and gas sensing will be increasingly relevant. CNT and fullerene sensors will play a major role in our society that has an insatiable need for more information. Not only will we have sensors that deliver better performance, but they will be smaller, faster, and more power efficient.

References 1. H. W. Kroto, J. R. Heath, S. C. O’Brien, R. F. Curl, and R. E. Smalley (1985). Nature, 318, 162. 2. W. Kratschmer, L. D. Lamb, K. Fostirapoulos, and D. R. Huffman (1990). Nature, 347, 354. 3. S. Ijima (1991). Nature, 354, 56. 4. R. Saito, M. Fujita, G. Dresselhaus, and M. S. Dresselhaus (1992). Appl. Phys. Lett., 60, 2204. 5. R. H. Baughman, A. A. Zakhidov, and W. A. de Heer (2002). Science, 297, 787. 6. M. Terrones (2003). Ann. Rev. Mat. Res., 33, 419. 7. J. W. G. Wildoer, L. C. Venema, A. G. Rinzler, R. E. Smalley, and C. Dekker (1998). Nature, 391, 59. 8. S. G. Louie (2001). Topics in Applied Physics, In: M. S. Dresselhaus, G. Dresselhaus, and P. Avouris (Ed.) Springer, New York, 80, 113. 9. T. W. Odom, J. L. Huang, P. Kim, and C. M. Lieber (1998). Nature, 391, 59. 10. T. W. Ebbesen (1997). Carbon Nanotubes: Preparation and Properties, In: T. W. Ebbesen (Ed.) CRC Press, Boca Raton, 139. 11. S. Ijima and T. Ichihashi (1993). Nature, 363, 603. 12. D. S. Bethune, C. H. Kiang, M. S. de Vries, G. Gorman, R. Savoy, J. Vazquez, and R. Beyers (1993). Nature, 363, 305. 13. C. Journet, W. K. Maser, P. Bernier, A. Loiseau, M. L. de la Chapelle, S. Lefrant, P. Deniard, R. Lee, and J. E. Fischer (1997). Nature, 388, 756. 14. H. Li, L. Guan, Z. Shi, and Z. Gu (2004). J. Phys. Chem. B, 108, 4573. 15. M. C. Paladugu, K. Maneesh, P. K. Nair, and P. Haridoss (2005). J. Nanosci. Nanotechnol., 5, 747. 16. A. Thess, R. Lee, P. Nikolaev, H. J. Dai, P. Petit, J. Robert, C. H. Xu, Y. H. Lee, S. G. Kim, A. G. Rinzler, D. T. Colbert, G. E. Scuseria, D. Tomanek, J. E. Fischer, and R. E. Smalley (1996).Science, 273, 483. 17. S. Arepalli, P. Nikolaev, W. Holmes, and B. S. Files (2001). Appl. Phys. Lett., 78, 1610. 18. C. D. Scott, S. Arepalli, P. Nikolaev, and R. E. Smalley (2001). Appl. Phys. A, 72, 573. 19. N. Braidy, M. A. El Khakani, and G. A. Botton (2002). J. Mat. Res., 17, 2189. 20. S. Takahashi, T. Ikuno, T. Oyama, S. I. Honda, M. Katayama, T. Hirao, and K. Oura (2002). J. Vac. Soc. Jpn, 45, 609. 21. H. Dai (2001). Topics in Applied Physics, In: M. S. Dresselhaus, G. Dresselhaus, and P. Avouris (Ed.) Springer, New York, 80, 29. 22. M. J. Yacaman, M. M. Yoshida, L. Rendon, and J. G. Santiesteban (1993). Appl. Phys. Lett., 62, 202. 23. V. K. Varadan and J. Xie (2002). Smart Mat. Struc., 11, 610. 24. D. Park, Y. H. Kim, and J. K. Lee (2003). Carbon, 41, 1025. 25. B. Q. Wei, R. Vajtai, Y. Jung, J. Ward, R. Zhang, G. Ramanath, and P. M. Ajayan (2002). Nature, 416, 495.

26

J.T.W. Yeow, N. Sinha

26. T. W. Ebbesen, P. M. Ajayan, H. Hiura, and K. Tanigaki (1994). Nature, 367, 519. 27. H. Hiura, T. W. Ebbesen, and K. Tanigaki (1995). Adv. Mat., 7, 275. 28. S. Bandow, A. M. Rao, K. A. Williams, A. Thess, R. E. Smalley, and P. C. Eklund (1997). J. Phys. Chem. B, 101, 8839. 29. A. G. Rinzler, J. Liu, H. Dai, P. Nikolaev, C. B. Huffman, F. J. Rodriguez-Macais, P. J. Boul, A. H. Lu, D. Heymann, D. T. Colbert, R. S. Lee, J. E. Fischer, A. M. Rao, P. C. Eklund, and R. E. Smalley (1998). Appl. Phys. A, 67, 29. 30. C. Xu, E. Flahaut, S. R. Bailey, G. Brown, J. Sloan, K. S. Coleman, V. C. Williams, and M. L. H. Green (2002). Chem. Res. Chinese Univ., 18, 130. 31. L. P. Biro, N. Q. Khanh, Z. Vertesy, Z. E. Horvath, Z. Osvath, A. Koos, J. Gyulai, A. Kocsonya, Z. Konya, X. B. Zhang, G. V. Tendeloo, A. Fonseca, and J. B. Nagy (2002). Mat. Sci. Eng. C., C19, 9. 32. J. R. Wood and H. D. Wagner (2000). Appl. Phys. Lett., 76, 2883. 33. J. Liu and H. Dai (2002). [Online]. Available: http://www.nnf.cornell.edu/2002re u/ Liu.pdf. 34. J. Wu, J. Zang, B. Larade, H. Guo, X. G. Xong, and F. Liu (2004). Phys. Rev. B, 69, 153406. 35. C. K. M. Fung, M. Q. H. Zhang, R. H. M. Chan, and W. J. Li (2005). Proc. 18th IEEE Conf. MEMS, 251. 36. I.-M. Choi and S. -Y. Woo (2006). Metrologia, 43, 84. 37. V. L. Pushparaj, L. Ci, S. Sreekala, A. Kumar, S. Kesapragada, D. Gall, O. Nalamasu, J. Suhr, and P.M. Ajayan (2007). Appl. Phys. Lett., 91, 153116. 38. P. Dharap, Z. Li, S. Nagarajaiah, and E. V. Barrera (2004). Nanotechnology, 15, 379. 39. C. Y. Li and T. W. Chou (2004). Nanotechnology, 15, 1493. 40. A. S. Berdinsky, Y. V. Shevtsov, A. V. Okotrub, S. V. Trubin, L. T. Chadderton, D. Fink, and J. H. Lee (2000). Chem Sustain. Dev., 8, 141. 41. S. Ghosh, A. K. Sood, and N. Kumar (2003). Science, 299, 1042. 42. P. Kral and M. Shapiro (2001). Phys. Rev. Lett., 86, 131. 43. V. T. S. Wong and W. J. Li (2003). Proc. IEEE Int. Symp. Circuits Sys., 4, IV844. 44. J. Kong, N. R. Franklin, C. Zhou, M. G. Chapline, S. Peng, K. Cho, and H. Dai (2000). Science, 287, 622. 45. A. Modi, N. Koratkar, E. Lass, B. Wei, and P. M. Ajayan (2003). Nature, 424, 171. 46. Z. Hou, H. Liu, X. Wei, J. Wu, W. Zhou, Y. Zhang, D. Xu, and B. Cai (2007). Sens. Actuators B, 127, 637. 47. Y. M. Wong, W. P. Kang, J. L. Davidson, A. Wisitsora-at, and K. L. Soh (2003). Sens. Actuators B, 93, 327. 48. I. Sayago, E. Terrado, E. Lafuente, M. C. Horrillo, W. K. Maser, A. M. Benito, R. Navarro, E. P. Urriolabeitia, M. T. Martinez, and J. Gutierrez (2005). Synth. Met., 148, 15. 49. Y. Hayakawa, Y. Suda, T. Hashizume, H. Sugawara, and Y. Sakai (2007). Jpn. J. Appl. Phys., 46, L362. 50. J. Suehiro, S.-I. Hidaka, S. Yamane, and K. Imasaka (2007). Sens. Actuators B, 127, 505. 51. S. Chopra, A. Pham, J. Gaillard, A. Parker, and A. M. Rao (2002). Appl. Phys. Lett., 80, 4632. 52. N. D. Hoa, N. V. Quy, Y. Cho, and D. Kim (2007). Sens. Actuators B, 127, 447. 53. F. Picaud, R. Langlet, M. Arab, M. Devel, C. Girardet, S. Natarajan, S. Chopra, and A. M. Rao (2005). J. Appl. Phys., 97, 114316. 54. Q. Zhao, M. B. Nardelli, W. Lu, and J. Bernholc (2005). Nano Lett., 5, 847. 55. P. Qi, O. Vermesh, M. Grecu, A. Javey, Q. Wang, H. Dai, S. Peng, and K. J. Cho (2003). Nano Lett., 3, 347. 56. E. Bekyarova, M. Davis, T. Burch, M. E. Itkis, B. Zhao, S. Sunshine, and R. C. Haddon (2004). J. Phys. Chem. B, 108, 19717. 57. M. Lucci, P. Regoliosi, A. Reale, A. D. Carlo, S. Orlanducci, E. Tamburri, M. L. Terranova, P. Lugli, C. D. Natale, A. D’Amico, and R. Paolesse (2005). Sens. Actuators B, 111–112, 181.

2 Carbon Nanotube and Fullerene Sensors

27

58. K. S. V. Santhanam, R. Sangoi, and L. Fuller (2005). Sens. Actuators B, 106, 766. 59. Y. X. Liang, Y. J. Chen, and J. H. Wang (2004). Appl. Phys. Lett., 85, 666. 60. L. Valentini, C. Cantalini, I. Armentano, J. M. Kenny, L. Lozzi, and S. Santucci (2004). Diam. Relat. Mat., 13, 1301. 61. J. Li, Y. Lu, Q. Ye, M. Cinke, J. Han, and M. Meyyappan (2003). Nano Lett., 3, 929. 62. F. Qu, M. Yang, J. Jiang, G. Shen, and R. Yu (2005). Anal. Biochem., 344, 108. 63. S. Hrapovic, Y. Liu, K. B. Male, and J. H. T. Luong (2004). Anal. Chem., 76, 1083. 64. R. P. Deo, J. Wang, I. Block, A. Mulchandani, K. A. Joshi, M. Trojanowicz, F. Scholz, W. Chen, and Y. Liu (2005). Anal. Chim. Acta, 530, 185. 65. S. Chopra, K. McGuire, N. Gothard, A. M. Rao, and A. Pham (2003). Appl. Phys. Lett., 83, 2280. 66. T. Someya, J. Small, P. Kim, C. Nuckolls, and J. T. Yardley (2003). Nano Lett., 3, 877. 67. C. Wei, L. Dai, A. Roy, and T. B. Tolle (2006). J. Am. Chem. Soc., 128, 1412. 68. K. Cattanach, R. D. Kulkarni, M. Kozlov, and S. K. Manohar (2006). Nanotechnology, 17, 4123. 69. C. Staii, A. T. Johnson, M. Chen, and A. Gelperin (2005). Nano Lett., 5, 1774. 70. J. T. W. Yeow and J. P. M. She (2006). Nanotechnology, 17, 5441. 71. E. S. Snow, F. K. Perkins, E. J. Houser, S. C. Badescu, and T. L. Reinecke (2005). Science, 307, 1942. 72. Y. T. Jang, S. I. Moon, J. H. Ahn, Y. H. Lee, and B. K. Ju (2004). Sens. Actuators B, 99, 118. 73. M. Penza, F. Antolini, and M. A. Vittori (2004). Sens. Actuators B, 100, 47. 74. K. G. Ong, K. Zeng, and C. A. Grimes (2002). IEEE Sens. J., 2, 82. 75. I. Szymanska, H. Radecka, J. Radecki, D. Kikut-Ligaj (2001). Biosens. Bioelect., 16, 911. 76. M. L. Y. Sin, G. C. T. Chow, G. M. K. Wong, W. J. Li, P. H. W. Leong, and K. W. Wong (2007). IEEE Trans. Nanotech., 6, 571. 77. M. Penza, P. Aversa, G. Cassano, W. Wlodarski, and K. Kalantar-Zadeh (2007). Sens. Actuators B, 127, 168. 78. R. K. Roy, M. P. Chowdhury, and A. K. Pal (2005). Vacuum 77, 223. 79. Y. Li, M. J. Yang, and Y. Chen (2005). J. Mat. Sc., 40, 245. 80. M. Penza, F. Antolini, and M. A. Vittori (2005). Thin Solid Films, 472, 246. 81. M. Penza, M. A. Tagliente, P. Aversa, and G. Cassano (2005). Chem. Phys. Lett., 409, 349. 82. S. G. Wang, Q. Zhang, R. Wang, and S. F. Yoon (2003). Biochem. Biophys. Res. Comm., 311, 572. 83. P. Young, Y. Lu, R. Terrill, and J. Li (2005). J. Nanosci. Nanotechnol., 5, 1509. 84. B. Perez, M. Pumera, M. del Valle, A. Merkoci, and S. Alegret (2005). J. Nanosci. Nanotechnol., 5, 1694. 85. Y. Lin, F. Lu, Y. Tu, and Z. Ren (2004). Nano Lett., 4, 191. 86. J. Wang and M. Musameh (2003). Anal. Chem., 75, 2075. 87. L. B. da Silva, S. B. Fagan, and R. Mota (2004). Nano Lett., 4, 65. 88. J. Wang, M. Musameh, and Y. Lin (2003). J. Am. Chem. Soc., 125, 2408. 89. M. Penza, G. Cassano, P. Aversa, F. Antolini, A. Cusano, M. Consales, M. Giordano, and L. Nicolais (2005). Sens. Actuators B, 111–112, 171. 90. Y. Zhang, K. Yu, R. Xu, D. Jiang, L. Luo, and Z. Zhu (2005). Sens. Actuators A, 120, 142. 91. J. Suehiro, G. Zhou, and M. Hara (2005). Sens. Actuators B, 105, 164. 92. J. Suehiro, G. Zhou, H. Imakiire, W. Ding, and M. Hara (2005). Sens. Actuators B, 108, 398. 93. M. Zhang, A. Smith, and W. Gorski (2004). Anal. Chem., 76, 5045. 94. J. Wang, G. Liu, and M. R. Jan (2004). J. Am. Chem. Soc., 126, 3010. 95. B. Philip, J. K. Abraham, A. Chandrasedhar, and V. K. Varadan (2003). Smart Mater. Struct., 12, 935. 96. O. K. Varghese, P. D. Kichambre, D. Gong, K. G. Ong, E. C. Dickey, and C. A. Grimes (2001). Sens. Actuators B, 81, 32.

28

J.T.W. Yeow, N. Sinha

97. I. Sayago, E. Terrado, M. C. Horrillo, M. Aleixandr, M. J. Fernandez, H. Santos, W. K. Maser, A. M. Benito, M. T. Martinez, J. Gutierrez, and E. Munoz (2007). Proc. 2007 Spanish Conf. Elect. Dev., 189. 98. J. Kombakkaran, C. Clewett, and T. Pietra (2007). Chem. Phys. Lett., 441, 282. 99. E. H. Espinosa, R. Ionescu, B. Chambon, G. Bedis, E. Sotter, C. Bittencourt, A. Felten, J.-J. Pireaux, X. Correig, and E. Llobet (2007). Sens. Actuators B, 127, 137. 100. Y. Sun and H. H. Wang (2007). Adv. Mater., 19, 2818. 101. A. Yang, X. Tao, R. Wang, S. Lee, and C. Surya (2007). Appl. Phys. Lett., 91, 133110. 102. S. Sotiropoulou and N.A. Chaniotakis (2003). Anal. Bioanal. Chem., 375, 103. 103. M. Gao, L. Dai, and G. G. Wallace (2003). Electroanalysis, 15, 1089. 104. Y. Lin, F. Lu, Y. Tu, and Z. Ren (2003). Nano Lett., 4, 191. 105. Y.-L. Yao and K.-K. Shiu (2007). Electrochim. Acta, 53, 278. 106. P. He and L. Dai (2004). Chem. Commun., 3, 348. 107. J. Wang, G. Liu, and M. R. Jan (2004). J. Am. Chem. Soc., 126, 3010. 108. H. Boo, R.-A. Jeong, S. Park, K. S. Kim, K. H. An, Y. H. Lee, J. H. Han, H. C. Kim, and T. D. Chung (2006). Anal. Chem., 78, 617. 109. P. W. Barone, S. Baik, D. A. Heller, and M. S. Strano (2005). Nature, 4, 86. 110. D. A. Heller, E. S. Jeng, T.-K. Yeung, B. M. Martinez, A. E. Moll, J. B. Gastala, and M. S. Strano (2006). Science, 311, 508. 111. M. Consales, A. Crescitelli, S. Campopiano, A. Cutolo, M. Penza, P. Aversa, M. Giordano, and A. Cusano (2007). IEEE Sens. J., 7, 1004. 112. M. Zhang and W. Gorski (2005). Anal. Chem., 77, 3960. 113. Y.-C. Tsai and C.-C. Chiu (2007). Sens. Actuators B, 125, 10. 114. Q. Zhou, Q. Xie, Y. Fu, Z. Su, X. Jia, and S. Yao (2007). J. Phys. Chem. B, 111, 11276. 115. L. Zhu, R. Yang, J. Zhai, and C. Tian (2007). Biosens. Bioelect., 23, 528. 116. S. Timur, U. Anik, D. Odaci, and L. Gorton (2007). Electrochem. Commun., 9, 1810. 117. T. Hirata, S. Amiya, M. Akiya, O. Takei, T. Sakai, and R. Hatakeyama (2007). Appl. Phys. Lett., 90, 233106. 118. J. Wei, L.-L. Sun, J.-L. Zhu, K. Wang, Z. Wang, J. Luo, D. Wu, and A. Cao (2006). Small, 2, 988. 119. T. Kotani, N. Kawai, S. Chiba, and S. Kitamoto (2005). Physica E., 29, 505. 120. J. Ma, J. T. W. Yeow, J. C. L. Chow, and R. B. Barnett (2007). Int. J. Robot. Autom., 22, 49.

Chapter 3

Non-carbon Nanotubes: Hydrogen Sensors Based on TiO2 Kristen E. LaFlamme and Craig A. Grimes

3.1 Introduction Sensors for the detection of gases such as oxygen, water vapor, and hydrogen are becoming increasingly important for a number of areas such as manufacturing, environmental monitoring, medicine, and defense/security [19]. Hydrogen sensing in particular is needed for industrial process control, combustion control, and in medical applications where the presence of hydrogen is indicative of certain types of health conditions [18]. Considerable effort has historically been spent in both broad and specific development of sensor technologies. In order to achieve a useful sensing device, it is necessary to simultaneously solve a number of design parameters including but not limited to size, cost, and durability, which are often contradictory in nature. Perhaps a starting point in considering a sensing platform is the transduction mechanism: do we seek, for example, to detect changes in electrical impedance, electrical phase, magnetic properties, frequency, elasticity, or mass? Is the material being used as a sensor selective for the target species or does it require an e-nose approach? Once a target sensor material is identified, operational issues that must be determined include sensitivity, dynamic range, resolution, hysteresis, fatigue, and drift. Materials investigated for sensing utility may possess one or more favorable attributes but be unusable because of a severe limitation in another [19]. Metal oxide materials are widely recognized for their outstanding gas-sensing properties. Capable of operating at elevated temperatures and in harsh environments, they are mechanically robust, relatively inexpensive, and offer exquisite sensing capabilities. Furthermore, it has become evident that introducing nanoscale architectural features onto the surfaces of metal oxides results in superior and unexpected gas-dependent electrical behavior [13]. Titanium oxide (TiO2), or titania, has previously earned much attention for its oxygen-sensing capabilities. With proper manipulation of the microstructure, crystalline phase, and/or addition of impurities or surface functionalization, K.E. LaFlamme Boston University, Boston, MA

F.J. Arregui (ed.), Sensors Based on Nanostructured Materials, DOI: 10.1007/978-0-387-77753-5_3, Ó Springer ScienceþBusiness Media, LLC 2009

29

30

K.E. LaFlamme, C.A. Grimes

this material can also be used for the detection of reducing gases. In this work, the authors describe the development of highly ordered TiO2 nanoarrays for hydrogen sensing. These devices have exhibited unprecedented changes in electrical conductivity in response to hydrogen, representing the largest known change in electrical properties of any material in response to any gas at any temperature. The fabrication of the TiO2 nanotube arrays will be outlined first, followed by a discussion on their operating characteristics, and finally a presentation of the performance of these sensors in specific environmental and medical applications.

3.2 Fabrication of TiO2 Nanotube Arrays Titania nanotube arrays are fabricated using a simple anodization approach. Briefly, titanium foils (typical starting thickness 250 mm) are subjected to potentiostatic anodization in a two-electrode electrochemical cell connected to a dc power supply using a platinum foil counter electrode. A diagram of a typical setup is shown in Fig. 3.1. In combination with HF, KF, or NaF to provide fluoride ions, it is possible to obtain nanotube arrays up to 222 mm in length using a variety of organic electrolytes, including dimethyl sulfoxide (DMSO), formamide (FA), ethylene glycol, and N-methylformamide (NMF). The formation of nanotube arrays in a fluoride-containing electrolyte is the result of several simultaneously occurring processes. During anodization, the nanotube formation occurs as a result of the combined effect of oxide growth at the oxide–Ti interface, the electric field-assisted dissolution at the oxide–electrolyte interface, and the chemical

Fig. 3.1 Anodization setup comprising a glass beaker, an electrolyte, and a dc power supply [21]

3 Non-carbon Nanotubes: Hydrogen Sensors Based on TiO2

31

dissolution of the titanium oxide by fluoride ions. Shallow pores are formed at the initial stages of anodization, leaving unanodized metallic portions between the pores. As the pores become deeper, the electric field at these protruded metallic regions increases, enhancing the field-assisted oxide growth and dissolution there. This results in the formation of well-defined inter-pore voids simultaneously with the growth of the pores. Both these processes finally yield a tubular structure. The mechanism of titania nanotube formation by anodization is discussed in detail elsewhere [11]. The chemistry appears quite flexible, with substitution between the fluoride-containing acids and organics generally resulting in nanotube arrays several tens of microns in length. If longer tubes are desired, it is necessary to minimize water content in the anodization bath to less than 5%. Prior to anodization, the foil is cleaned with ethanol. The anodization procedure is carried out at approximately 208C, with the anodization current being monitored using a digital multimeter interfaced with a computer. The resulting structure consists of adjacent, parallel-oriented nanotubes separated from the unanodized foil by a thin oxide layer (barrier layer), the initial thickness of which is approximately equal to the nanotube radius [21]. Physical features of the nanotubes can be controlled by changing the anodization conditions [13, 21]. This ability to tailor nanotube array dimensions has enabled the ability to meaningfully test the gas-sensing properties of the nanotube arrays. The length of the nanotube array is dependent upon the pH, with higher pH values that remain acidic resulting in longer nanotubes. Tube length is also somewhat dependent on anodization voltage. The wall thickness and pore diameter are controlled by the anodization bath temperature and anodization voltage. The scanning electron micrographs in Fig. 3.2 show some examples of nanotubes that were fabricated under different conditions. The as-fabricated, amorphous nanotube arrays demonstrate no hydrogensensing capabilities. Consequently, the nanotube arrays must be annealed at high temperatures ranging from 480 to 5808C for several hours. The annealing process crystallizes the nanotubes, primarily in the anatase phase, but with some rutile crystals on the barrier layer, enhancing the sensitivity of the hydrogen response [10]. Anatase, the polymorph of titania, has been reported to have high sensitivity for reducing gases like hydrogen and carbon monoxide. Thus, as the diffusing hydrogen atoms go to the interstitial sites and as the c/a ratio of anatase is almost four times that of rutile, it appears that the anatase phase accommodates hydrogen easily and hence makes a higher contribution to hydrogen sensitivity. Studies indicated that no significant differences in gas-sensing properties were found as a function of annealing temperature; however, above 5808C, protrusions start to come out through the nanotubes, an effect which spreads with increasing temperature. These protrusions, which are due to oxidation of the titanium substrate, collapse the nanotubes [18].

32

K.E. LaFlamme, C.A. Grimes

Fig. 3.2 Cross-sectional and top view SEM images of TiO2 nanotubes fabricated under different conditions. A–C Formamide-based electrolyte at 35 V for 48 h, tube length 70 mm; DE Dimethyl sulfoxide electrolyte containing 2 vol% HF at 60 V for 70 h, tube length 93 mm; F Ethylene glycol electrolyte containing 0.25 wt% NH4F at 60 V for 48 h, tube length 222 mm;. Magnification: A, 650; B, 55,000; C, 120,000; D, 750; E, 50,000; F, 45,000 [22]

3 Non-carbon Nanotubes: Hydrogen Sensors Based on TiO2

33

3.3 Sensor Development and Operating Characteristics 3.3.1 Sensor Design To transform the nanotube arrays into gas sensors, it is necessary to achieve electrical contact with the material so that the resistance or impedance of the material can be measured as a function of gas ambient. This is accomplished by depositing metal contacts onto the array. Typically, platinum electrodes have been used with these types of hydrogen sensors [7, 9, 10, 11, 13, 17, 18, 19, 20, 21, 22]. Upon deposition of the electrodes, metal wires are bonded onto the pads. Using an electronic interface, the changes in electrical resistance in response to variations in hydrogen concentration can be quantified. A diagram of a typical sensor setup can be found in Fig. 3.3. When the sensor comes into contact with hydrogen, the resistance is reduced. For a sensor measurement, a constant current I is applied to the sensor and the resulting dc voltage V is recorded. The sensor resistance R is then calculated by R = V/I. The electrical resistance of the nanotubes is sensitive to hydrogen concentrations ranging from 10 ppm to 4% [17]. The dominant mechanism behind the observed hydrogen-sensitive electrical behavior of the nanotubes appears to be chemisorption of hydrogen. During chemisorption, hydrogen acts as a surface state and a partial charge transfer from hydrogen to the titania conduction band takes place. This creates an electron accumulation layer on the nanotube surface, enhancing its electrical conductance. Upon removal of the hydrogen ambient, electron transfer back to the hydrogen molecule takes place, which subsequently desorbs, restoring the original electrical resistance of the material.

Fig. 3.3 Schematic diagram of a typical setup used for gas-sensing experiments [20]

34

K.E. LaFlamme, C.A. Grimes

Another factor that may play a role in the hydrogen sensitivity (and selectivity) is the platinum electrodes. At elevated temperatures hydrogen dissociation can occur on platinum surfaces. These dissociated hydrogen atoms may spill onto the nanotube surface in accordance with the well-known spillover mechanism of hydrogen by platinum. Once on the nanotube surface, the hydrogen atoms diffuse into the material, affecting its electrical properties.

3.3.2 Operating Features The nanotube arrays are extremely sensitive in the presence of hydrogen and their response behavior is highly reproducible. Figure 3.4 demonstrates the change in resistance of a 76 nm diameter titania nanotube sensor as the hydrogen concentration was repeatedly cycled at 2908C between a pure nitrogen ambient and a 500 ppm hydrogen in discrete steps of 100 ppm [7, 17, 19]. A variation in resistance of 3 orders of magnitude is observed at only 100 ppm hydrogen. Moreover, the electrical behavior of the sensor is consistent, recovering its original resistance after repeated exposure to hydrogen at different concentrations without hysteresis.

3.3.3 Tunability In these titania nanotube array-based sensors, hydrogen can produce changes in electrical resistance of about 3–8.7 orders of magnitude from baseline dependent on the operating conditions. Figure 3.5 shows the normalized change in electrical conductance of a nanotube array (tube diameter = 76 nm; tube

Fig. 3.4 Measured electrical resistance of a 76 nm titania nanotube hydrogen sensor when exposed to different concentrations of hydrogen at 2908C [19]

3 Non-carbon Nanotubes: Hydrogen Sensors Based on TiO2

35

Fig. 3.5 Normalized change in electrical conductance of a 76 nm diameter TiO2 nanotube array to 1,000 ppm hydrogen at different temperatures. Gg is the conductance in the presence of hydrogen and G0 the base resistance associated with a nitrogen atmosphere [17]

length = 400 nm) as a function of temperature as it is switched from a nitrogen ambient to nitrogen plus 1,000 ppm hydrogen and then back to nitrogen. With reference to this figure, Gg is the conductance in the presence of hydrogen and G0 is the base conductance associated with a nitrogen atmosphere. The magnitude of the conductance variation increases with temperature, starting from a marginal variation at 1808C to a variation of 3 orders of magnitude at approximately 4008C [17]. Physical properties of the nanotubes can also affect their performance. For example, Fig. 3.6 demonstrates the variation in electrical conductance for sensors with nanotube diameters of 76, 53, and 22 nm and wall thicknesses of 27, 17, and 13 nm, respectively [17]. It is clear that the conductance variation is more prominent at smaller nanotube diameters. The sensitivity of the sensors also changes with respect to tube diameter, as represented in Fig. 3.7.

Fig. 3.6 Normalized change in electrical conductance of TiO2 nanotube arrays of 76, 53, and 22 nm diameter to 1,000 ppm hydrogen at 2908C [17]

36

K.E. LaFlamme, C.A. Grimes

Fig. 3.7 The sensitivity variation of samples with pore diameters of 22, 53, and 76 versus hydrogen concentration [17]

The sensitivity can be calculated as s¼

Ggas  G0 G0

(3:1)

where Ggas and G0 represent the conductance of the sensors in hydrogen and at baseline, respectively. Assuming the density of the nanotube is 90% that of single crystal titania, the calculated surface area of the 76 nm diameter nanotube is approximately 19 m2g1 and that of the 22 nm array 38 m2g1. Interestingly, while the reduction of the nanotube diameter from 76 to 22 nm increases the surface area by only a factor of 2, the sensitivity is increased by approximately a factor of 200! This suggests that the nanotube wall thickness and the number of contact points between the nanotubes are additional factors that play a role in determining the nanotube sensitivity. With the reduced wall thicknesses, the space charge layer created by the chemisorption of hydrogen extends throughout the wall, resulting in a significant reduction in resistance consistent with the observed differences in sensitivity between nanotubes with different diameters. Also, as can be seen in Fig. 3.2, the nanotubes are in contact with each other, creating a high resistance path for electrons to travel. Upon the creation of the space charge layer due to hydrogen, these tube-to-tube contact points become highly conducting relative to the rest of the nanotube. For a bulk conductivity constant with nanotube diameter, the greater the number of contact points the greater will be the resistance change upon exposure to hydrogen. Therefore, the smaller diameter tubes, with thinner walls and greater number of contact points will exhibit higher sensitivities than their larger diameter counterparts [17]. On the other hand, sensors with smaller diameter tubes also tend to be more brittle and harder to handle without breakage. Therefore, a balance must be struck between sensitivity and mechanical durability [18]. Tube length is another physical parameter that affects the functionality of these sensors. Figure 3.8A demonstrates the gas-sensing properties of a sensor composed of a nanotube array with tube length of 1 mm. A change in electrical

3 Non-carbon Nanotubes: Hydrogen Sensors Based on TiO2

A

37

B

Fig. 3.8 A Hydrogen gas-sensing properties of a 1 mm long nanotube array; B hydrogen gassensing properties of a 20 mm long nanotube array [22]

resistance of 8.7 orders of magnitude is seen in response to 1,000 ppm hydrogen, while the response time is approximately 20 s. Similar sensitivities and response times are obtained from sensors made of nanotube arrays of that approximate length. On the other hand, Fig. 3.8B shows the change in electrical behavior for a sensor made from 20 mm long nanotube arrays. While the sensor still demonstrates an excellent change in resistance of 3 orders of magnitude in response to 2,000 ppm hydrogen, there is a clear reduction in sensitivity accompanied by longer response and recovery times of nearly 2 h due to the time required for hydrogen to diffuse inside the long pores [22].

3.3.4 Cross-Sensitivity Up to this point, the performance of these sensors has been discussed in the context of a nitrogen ambient. What about other gases? In the real world, sensors are exposed to a number of other gases such as oxygen, carbon dioxide, carbon monoxide, humidity, and organic vapors. Thus, it is important to assess the cross-reactivity of these sensors with other gases. Figure 3.9 shows the variation in resistance of a nanotube array hydrogen sensor (tube diameter = 76 nm; tube length = 400 nm) upon exposure to various gases including carbon dioxide, carbon monoxide, ammonia, and oxygen. As shown in Fig. 3.9A, when the sensor is switched from a pure nitrogen environment to nitrogen plus 10% CO2, there is no detectable change in resistance; this is expected because carbon dioxide is neither a strong reducing agent nor a strong oxidizing gas [21]. Figure 3.9B shows the normalized change in resistance with respect to oxygen, ammonia, and carbon monoxide [18]. The sensitivity of the nanotubes to carbon monoxide and ammonia is negligible compared to that of hydrogen. The

38

K.E. LaFlamme, C.A. Grimes

A

B

Fig. 3.9 Variation of resistance of a 76 nm diameter titania nanotube array sensor upon exposure to A carbon dioxide [21] and B carbon monoxide, ammonia, and oxygen [19]

resistance of the nanotubes did increase in the presence of oxygen; however, as shown in Fig. 3.10, while oxygen does reduce the sensitivity of the sensor, there is still a clearly detectable hydrogen signal. Furthermore, the sensitivity difference was only about 1 order of magnitude between a pure nitrogen environment and one with 20% oxygen (atmospheric condition). On the other hand, the sensors are significantly affected by moisture. Figure 3.11 plots the change in sensor sensitivity (at 1,000 ppm hydrogen) as a function of relative humidity (RH). Between RH of 3 and 56%, the sensitivity of the hydrogen sensor varies

108 H2 (1%) + 20% O2 H2 (1%)

7

10

H2 (1%) + 15% O2

H2 (1%) + 2% O2

6

Resistance (Ohms)

10

105 104 103 102 101

Air 0

200

400

600

800

1000

1200

1400 1600

Time (Seconds) Fig. 3.10 Electrical behavior of the hydrogen sensor in the presence of increasing concentrations of oxygen [21]

3 Non-carbon Nanotubes: Hydrogen Sensors Based on TiO2

39

Fig. 3.11 The sensitivity of the hydrogen sensor is reduced with increasing humidity [21]

from 2.5  108 to 7  104, a change of more than 3 orders of magnitude. For these reasons, it is important to calibrate the sensor to RH.

3.3.5 Room Temperature Sensing From Fig. 3.5 we have seen that temperature plays a role in the performance of titania nanotube-based hydrogen sensors, with the sensors generally displaying higher sensitivity at higher temperatures. However, elevated operating temperatures are not favorable for many applications, particularly those involving flammable environments and those requiring low-power operation. To realize hydrogen sensing at room temperature, a thin film of palladium (Pd) can be deposited on the surface of the tubes. A typical sensor response upon exposure to 20–1,000 ppm hydrogen gas is shown in Fig. 3.12. In this case, the sensor used had a tube length of approximately 200 nm and a diameter of about 22 nm. As seen in this plot, the initial resistance is of the order of 106 , and falls to below 102 when exposed to 1,000 ppm hydrogen. Typical 90% response times are approximately 15 s. The sensor recovers completely after H2 flow is terminated in each case with no detectable drift [20]. The catalytic properties of palladium for hydrogen dissociation and its hydrogen adsorption–desorption characteristics are well known. Similar to platinum, palladium dissociates hydrogen into atomic form. The activated hydrogen removes adsorbed oxygen and is adsorbed onto the Pd surfaces of

40

K.E. LaFlamme, C.A. Grimes 108

Resistance (Ω)

107

20

40

60

Hydrogen concentration in ppm 80 100 200 400 600

800

1000

6

10

105 104 103 102

air air air

air

1

10 0 100

3

2 10

3

4 10

air 3

6 10

air

air 3

8 10

air 4

1 10

air

1.2 104

air

1.4 104

Time (Sec) Fig. 3.12 Typical sensor response at 258C to different concentrations of hydrogen in the 20–1,000 ppm range [20]

the sensor. Hydrogen adsorption reduces the Pd work function; therefore, the height of the potential barrier at the metal–titania interface is lowered, which in turn reduces the electrical resistance of the sensor. In this manner, Pd enhances the H2 sensitivity of the nanotubes [20].

3.4 Applications of Titania Nanotube Hydrogen Sensors 3.4.1 Self-Cleaning Sensors A critical concern of any sensor platform is the potential for unwanted contamination, or poisoning, which introduces spurious measurements and generally ends the useful lifetime of a sensor. A sensor used in a non-controlled environment faces potential contamination from volatile organic vapors, carbon soot, oil vapors, dust, and pollen to cite just a few examples. While the titania nanotubes that have been described in this chapter are cheap to manufacture and could be used disposably, an important advance in sensor technology would be a sensor able to self-clean, thereby extending its useful lifetime and minimizing the potential for inaccurate measurements. The photocatalytic activity of titania can be used to mitigate the effect of contamination on these sensors by irradiating the surface of the titania surface with light with an energy equal to or greater than the titania band-gap energy (3.0 eV for the rutile phase and 3.2 eV for the anatase phase). Titania is a semiconductor characterized by a filled valence band and an empty conduction band. When a photon with energy h matches or exceeds the band-gap energy, an electron is promoted from the valence band to the conduction band, leaving a hole behind. The valence band holes are powerful oxidants, whereas the conduction band electrons are good reducers. Most organic photodegradation

3 Non-carbon Nanotubes: Hydrogen Sensors Based on TiO2

41

reactions utilize the oxidizing power of electron holes, either directly or indirectly, generally producing CO2 and H2O without the production of potentially toxic side effects [10, 20]. The physical properties of the tubes contribute to the self-cleaning ability. Nanotube wall thickness, shape, surface area, and length are important factors in potentially enhancing the photocatalytic efficiency on which the self-cleaning capabilities discussed here are based. For example, when the feature size of the titania structure is less than 25 nm, the wave function of the charge carriers spreads throughout the structure, and hence the photogenerated electrons and holes are readily accessible to the donors or acceptors on the surface, resulting in a high degree of photocatalysis. Additionally, the crystallinity of the material plays a role: the anatase phase of titania is optimal for photocatalytic activities, while the rutile phase contributes to its hydrogen-sensing capabilities [10, 20]. In one study, Mor et al. demonstrated the ability of titania nanotube-based hydrogen sensors to completely recover their hydrogen-sensing abilities with exposure to UV light after being completely extinguished by immersing the sensor in motor oil [9]. Figure 3.13 shows the real-time electrical resistance of an illustrative hydrogen sensor operated at room temperature in response to different atmospheres, contamination, and UV light exposure. These self-cleaning sensors were tested in a 60 cm3 Plexiglas test chamber, with an opening for introducing the motor oil onto the sensor surface and a quartz window for passing the UV illumination onto the sensor surface. The sensor was exposed to a hydrogen–nitrogen mixture of 1,000 ppm hydrogen. After reaching the saturation resistance, the gas was switched back to air with the sensor returning to its original state (Fig. 3.13A). Prior to sensor contamination (Fig. 3.13A) there was a change in sensor resistance with exposure to 1,000 ppm hydrogen of approximately 175,000%. While in air (Fig. 3.13B), the sensor was then contaminated with an approximately 0.06 mm layer of 10 W–30 (Penzoil) motor oil after which the sensor demonstrated virtually no change in electrical resistance with hydrogen exposure over a period of approximately 10 min (Fig. 3.13B at time 1,600 s). The oil-contaminated region on the sensor (Fig. 3.13B) was uniformly illuminated with UV light in the presence of air. The recovery of the sensor’s hydrogen-sensing capability after sequential UV exposures with durations of approximately 1.3 h can be seen in Fig. 3.13B and C; 3.8 h in Fig. 3.13C; and 4.4 h in Fig. 3.13C and D. As seen from Fig. 3.13D, at the end of the UV-driven sensor-cleaning period, the measured relative change in electrical resistance with exposure to 1,000 ppm hydrogen is 100,000%. The clean and recovered air sensor has similar resistance values with exposure to air; the recovered sensor has a 1,000 ppm hydrogen resistance of approximately 100

compared with the resistance of 60 of the sensor prior to contamination. The photocatalytic properties of the nanotubes can be qualitatively understood by viewing FE-SEM images of a sensor contaminated with stearic acid before and after cleaning. Figure 3.14A shows a 22 nm sensor coated with stearic acid, applied by dripping melted stearic acid onto the sample. As seen in Fig. 3.14B, after 1 h of UV exposure, the sample is essentially clean.

42

K.E. LaFlamme, C.A. Grimes

A

B Fig. 3.13 A plot of real-time variation of resistance change before, during, and after cleaning the contaminant, motor oil 10 W–30, with UV exposure. The plot, broken into four parts for clarity, shows A The original sensor behavior from time 10 to 1,000 s. B Behavior of the sensor over time 100–6,000 s during which the sensor is contaminated with oil losing its hydrogensensing capabilities and is initially exposed to UV light C The behavior of the sensor from time 5,000 to 45,000 s. At time 7,000 s, the UV is turned off, with the sensor regaining its nominal starting resistance of approximately 100,000 W, at which point it is exposed to 1,000 ppm hydrogen and shows relative change in resistance of approximately 50. The sensor is then again exposed to UV, from roughly time 15,000 to 29,000 s. After this second UV exposure, the sensor is again exposed to 1,000 ppm hydrogen, showing an approximate factor of 500 change in electrical resistance. The sensor is once again exposed to UV from time 36,000 s. D Sensor behavior from time 45,000 to 70,000 s continues with UV exposure of the sensor to time 52,000 s, after which the sensor is repeatedly cycled between air and 1,000 ppm hydrogen showing a relative change in impedance of approximately 1,000. Compared to the hydrogen sensitivity of a non-contaminated sensor, the relative response of the ‘recovered’ sensor is within a factor of 2 [9]

3 Non-carbon Nanotubes: Hydrogen Sensors Based on TiO2

43

C

D Fig. 3.13 (continued)

To more closely observe the recovery of sensor properties with UV exposure time, Mor et al. measured the changes in relative resistance of a sensor in response to incremental UV exposure following contamination with both stearic acid and a 1 mm thick layer of motor oil [10]. Figure 3.15 shows the relative resistance change, Rair/RH, of the sensor with each incremental 1 h UV exposure step; Rair is the resistance of the clean sensor in air, generally equivalent to that of the contaminated sensor, and RH is the steady-state resistance of the sensor exposed to 1,000 ppm hydrogen. The value of Rair/RH for the clean sensor is shown at an arbitrary time, whereas that of the contaminated sensor is shown at

44

K.E. LaFlamme, C.A. Grimes

Fig. 3.14 FE-SEM top views of the stearic acid-contaminated titania nanotubes A before and B after cleaning using a UV exposure of 1 h [10]

time 0. This plot clearly demonstrates the successive recovery of the Rair/RH value toward that of the clean sensor. Notably, the electrical resistance of the sensor rapidly drops with UV illumination due to photogeneration of charge carriers; yet, after the UV light source is turned off, a relatively long time is required for resistance to be regained. This phenomenon can be easily seen in Fig. 3.13C when the UV illumination is turned off at approximately 7,000 s. Since the UV light source used in these experiments has a shutter that closes when the light source is

Fig. 3.15 A semi-log plot illustrates the stepwise improvement of contaminated sensor. Here the relative change in resistance of a sensor measured with respect to 1,000 ppm of hydrogen (in nitrogen) is shown as a function of duration of UV exposure. The fresh sensor is shown at arbitrary time, while the dirty sensor is at time 0 [10]

3 Non-carbon Nanotubes: Hydrogen Sensors Based on TiO2

45

turned off, and charge carrier lifetimes do not run into the minutes, it is clear that another mechanism must be responsible for the relatively slow recover times. This behavior is explained by the role of oxygen in manipulating the sensor conductivity. On UV illumination, the chemisorbed oxygen must be desorbed, increasing the conductivity. Hence, the conductivity increase upon UV illumination can be attributed to both the photogenerated current and the electrons that are donated by the oxygen. Once the UV illumination is removed, the oxygen is readsorbed and the electrons are extracted from the sensor; however, the process of oxygen readsorption is slow, on the order of several minutes, thus resulting in the increased time for the sensor to recover its original resistance [9, 10]. It is well established that the presence of oxygen and water plays a crucial role in the photocatalytic cleaning of titania, hence the sensor was exposed to air during UV exposure to facilitate removal of the oil from the sensor surface. If the relevant redox potential of the contaminant does not lie within the band gap of titania, the organic contaminants cannot be oxidized by photogenerated electron–hole pairs on the surface of titania. However, because the potential of water and oxygen exists within the band gap of titania, the photogenerated holes in the valence band can oxidize water to produce a highly reactive hydroxyl radical (.OH) and the photogenerated electrons in the conduction band can reduce oxygen to form highly reactive superoxide (O2.) ions (see Equations 3.2–3.4), which then assist in oxidizing the organic species. TiO2 þh ! Hþ þe

(3:2)

H2 O þ Hþ ! OH þ Hþ

(3:3)

O2 þe ! O2 :

(3:4)

Because the utility of a given chemical sensor technology is determined, in part, by how rapidly its properties degrade with non-specific contamination, development of a self-cleaning sensor technology of excellent sensitivity is an exciting prospect. In this section, a room temperature hydrogen sensor that demonstrates over a 170,000% change in electrical resistance upon exposure to 1,000 ppm hydrogen is described. The sensor is able to self-clean photocatalytically with UV exposure, fully recovering its initial properties lost due to contamination.

3.4.2 Hydrogen Sensing for Biomedical Applications: Transcutaneous Sensors In addition to their utility in environmental sensing, industrial process control, and manufacturing, hydrogen sensors also have a place in the biomedical arena. For example, breath hydrogen is a clinically important parameter used as an

46

K.E. LaFlamme, C.A. Grimes

indicator for such disorders as lactose intolerance [3], fructose maladsorption [12], bacterial growth [15], fibromyalgia [14], and neonatal necrotizing enterocolitis [6]. Transcutaneous gas monitoring has also been used as a means for diagnosing disease as well as monitoring treatment effectiveness. For example, transcutaneous carbon dioxide is routinely monitored for bronchopulmonary dysplasia, apnea, upper airway obstruction, mechanical ventilation, and respiratory problems associated with hypercapnia [1]. Transcutaneous carbon dioxide is also routinely used in neonatal intensive care units [2, 16] or in the preclinical detection of pneumothorax [8]. In a similar vein, Varghese and coworkers developed a sensor to measure hydrogen transcutaneously for the diagnosis of lactose intolerance [21]. Lactose intolerance is caused by the inability to digest significant amounts of lactose, a major sugar in milk. Failure to digest lactose results in fermentation of the lactose in the colon, which subsequently leads to the formation of various gases including hydrogen. The hydrogen is then absorbed from the intestines and carried into the bloodstream. A portion of the hydrogen is released in the skin, thus allowing hydrogen detection with a non-invasive hydrogen sensor [21]. A picture of the transcutaneous sensor is shown in Fig. 3.16. As previously discussed, metal oxide sensors such as the one described in this chapter can respond to some extent to other gases, resulting in unwanted cross-reactivity. Thus, it is necessary to use other sensors and cross-correlate the responses of these different sensors to obtain an absolute measurement of the target gas from within a complex environment. Therefore, a humidity–temperature sensor is incorporated with the transcutaneous hydrogen sensor. In this device, the hydrogen sensor and humidity–temperature sensor are secured in a white

A

B

Fig. 3.16 A Packaging of sensors comprising a hydrogen sensor and a humidity–temperature sensor. B The connector twists onto the adhesive-backed ring. Vent holes (not visible) on the side prevent the accumulation of water vapor and gases in the housing [21]

3 Non-carbon Nanotubes: Hydrogen Sensors Based on TiO2

47

Teflon housing, which screws onto the adhesive-backed ring that is applied to the skin of the patient undergoing the test. The sensors were tested on patients who were referred by their doctors for a breath hydrogen test due to lactose deficiency. The transcutaneous sensor was placed on the forearm of children and adult volunteers, and the sensor documented the transcutaneous excretion of hydrogen. The results demonstrated a good correlation between the transcutaneous sensor readings and the exhaled breath hydrogen levels as measured by a commercial gas chromatography instrument. Furthermore, some of the patients tested had no detectable transcutaneous hydrogen excretion or detectable levels of exhaled breath hydrogen, thus indicating that the sensor does not respond to other volatile molecules that might be excreted through the skin. Figure 3.17 shows an illustrative measurement of transcutaneous hydrogen levels using the described transcutaneous sensor for a lactose-intolerant adult volunteer who drank a small sample of milk; note the sensor measurement is continuous in operation, and the hydrogen levels (ppm) in exhaled breath are measured using a gas chromatograph. Figure 3.18 shows the decrease in the sensor measurement, which indicates an increase in transcutaneous hydrogen, coinciding with an increase in exhaled breath hydrogen concentration. In addition to lactose intolerance, the transcutaneous hydrogen sensor can also be applied to other diseases such as necrotizing enterocolitis (NEC), a devastating disease of uncertain etiology and pathogenesis that causes high levels of mortality and morbidity in about 10% of preterm infants in neonatal

Fig. 3.17 Measurement of transcutaneous hydrogen concentrations using the described sensor (y-axis on left-hand side) and exhaled breath hydrogen levels (y-axis on right-hand side) measured by a gas chromatograph from a lactose-intolerant adult volunteer [21]

48

K.E. LaFlamme, C.A. Grimes

Fig. 3.18 Correlation between the response of the transcutaneous hydrogen sensor and exhaled hydrogen concentration [21]

intensive care units throughout the world. Almost 20 years ago, several investigators demonstrated that elevated levels of exhaled breath hydrogen gas precede the onset of the clinical and radiographic signs of NEC [4, 6]. From a practical standpoint, it is very difficult to measure gas levels in exhaled breath of preterm infants because of their small tidal volumes and rapid respiratory rates. In neonatal intensive care units, the use of endotracheal tubes, nasal CPAP systems, and nasal cannula flow systems are all barriers to the successful collection of undiluted end expiratory gas samples for gas analysis. Therefore, transcutaneous sensing of hydrogen gas in preterm infants has the potential to facilitate the early identification and treatment of cases of NEC. This would permit the early withdrawal of enteral feeds and the early initiation of systemic antibiotics and supportive therapy with fluids, pressors, and blood products. In addition to saving lives (NEC mortality rate is about 35%), the early detection of NEC may also significantly reduce hospital costs, which was about $95 million per year in the United States in 2002 [5].

3.4.3 Sensor Networks In many instances, practical application of hydrogen sensors would be significantly enhanced by the development of an inexpensive, wide-area sensor network technology incorporating hydrogen sensors capable of real-time in situ detection of 0.1–1,000 ppm H2 gas. As a step toward that ultimate goal, Grimes and coworkers incorporated titania nanotube hydrogen sensors into a wireless

3 Non-carbon Nanotubes: Hydrogen Sensors Based on TiO2

49

sensor network to measure hydrogen concentrations from multiple sources in real time [7]. The hydrogen sensor itself is composed of a highly ordered thin film of titania nanotubes to which platinum electrodes are connected, as described earlier in this chapter. Briefly, the nanotubes are highly sensitive to hydrogen due to the chemisorption of molecular hydrogen, displaying a change in electrical resistance of approximately 3–8.7 orders of magnitude in response to hydrogen gas, depending on the operating conditions such as temperature, physical parameters of the tubes, or environmental conditions. The hydrogen sensors are completely reversible, with a response time of about 5 min and a recovery time of about 30 min. The sensor network consists of an array of nodes distributed throughout the area of interest, with a ‘host’ or ‘parent’ node connected directly to a computer that graphically presents the collected data. The main component of a sensor node is a microcontroller that oversees node operation, including transmission of node sensor information as well as relaying of information from distant nodes toward nodes closer to base. Each node has a RF transceiver integrated with the microcontroller for wireless communication between nodes. Interface circuitry is used to convert the response of the hydrogen sensor, a change in electrical resistance, to a voltage variation so that it can be digitized and processed by the microcontroller. The host node also contains a serial interface to communicate with a computer via RS232 protocol.

3.4.3.1 Design of the Sensor Network Before a sensor node can transmit data it needs to know the identity of its ‘parent,’ or the node it should relay its data to. To achieve this, the computer first instructs the parent node to send a broadcast signal that contains its identity. When a nearby node receives the signal, it remembers the parent node as its parent node and in turn sends its own broadcast signal to other nodes. This process is repeated until all nodes in the network know their parent node identity. The broadcasting operation is important for the nodes in the network to initialize communication links or re-establish communications after some nodes are damaged or removed. When relaying sensor information, all nodes transmit data to their parent nodes. This ensures that the data of all nodes are eventually sent to the host node and downloaded to the host computer. The circuit schematic of the sensor node is shown in Fig. 3.19. The microcontroller (AT90S8535) is connected to a RF transceiver chip (TR1000) via its serial bus. The microcontroller oversees all operations of the node, including acquiring data from the sensor, transmitting sensor data, and relaying the data of other nodes. A temperature sensor is included in each sensor node (AD7418). The resistance of the hydrogen sensor is converted to a voltage by a sensor interface circuit (see Fig. 3.20). The converted voltage is sent to the analog-to-digital converter port (ADC0, pin 40) of the microcontroller to be digitized.

50

K.E. LaFlamme, C.A. Grimes

Fig. 3.19 Circuit schematic of the sensor node. The ADC0 port (pin 40) of the microcontroller AT90S8535 is connected to the output of the sensor interface circuit [7]

Figure 3.20 shows the schematic diagram of the hydrogen sensor interface circuit. The voltage reference IC6 (ADR380) provides a constant voltage of 2 V across pins 2 and 3. When R5 is set to 10 k , IC6 provides a constant current of 200 mA through the hydrogen sensor. By Ohm’s law, the output voltage of the unity-gain op amp IC7 (AD820) V is proportional to the sensor resistance RS: V¼RS 200A

(3:5)

3 Non-carbon Nanotubes: Hydrogen Sensors Based on TiO2

51

Fig. 3.20 Circuit schematic of the sensor interface. The output (pin 6 of the op amp) is connected to the ADC0 port (pin 40) of the microcontroller AT90S8535 (see Fig. 3.19) [7]

The microcontroller has a 10-bit analog-to-digital converter with a maximum sampling voltage of 3 V. In other words, the sampling voltage from 0 to 3 V will be linearly converted to a digital value of 1 to 1,023, with a voltage resolution of about 3 mV. From Equation 3.5, a 3 mV voltage resolution leads to a resistance resolution of 15 , which cannot accurately represent the resistance variation of the hydrogen sensor when the hydrogen concentration is larger than 100 ppm because the electrical resistance of the sensor becomes less than 60 at 100 ppm. To increase the resolution, another op amp, IC9, with a gain of 15 is used to amplify the output voltage from IC7. The outputs from both IC9 (with a gain of 15) and IC7 (with a gain of 1) are sent to a switch (IC10), which is controlled by a comparator (IC8). When the output of IC7 is between 0 and 200 mV (corresponding to an RS of 0–1 k ), the comparator sends out an output of 1 to pin 1 of the switch so that the switch can choose IC9 as its input. Conversely, when the voltage output of IC7 is larger than 200 mV, the switch will select IC7 as its input. With this design, the sensor interface circuit will generate a voltage of 200 mV–3 V when the sensor resistance is from 1 to 15 k and 0 to 3 V when the sensor resistance is 0–1 . As a result, the resistance resolution increases to 1

when the sensor resistance is less than 1 k . The output of the comparator is also sent to the I/O port of the microcontroller (PA1) and is included as part of the data. When the host computer receives a data package from the sensor node, it will divide measurements by 15 if the value of PA1 is 1. Doing so will ensure that the measured voltage is linearly proportional to the resistance from 0 to 15 k . The operational characteristics of this sensor network were tested using an experimental setup illustrated in Figs. 3.21 and 3.22. The hydrogen sensor itself

52

K.E. LaFlamme, C.A. Grimes

Fig. 3.21 The experimental setup for testing the sensor response. For elevated temperature measurements only the hydrogen sensor is placed within the test chamber, with an electrical connection passing to the interface circuitry. Using other sensor nodes as data relays the hydrogen measurements are sent to the host node and downloaded to the host PC for graphical presentation [7]

was placed in a test chamber, consisting of an 8 cm diameter and 60 cm long quartz tube, in which the temperature and atmosphere could be controlled (Fig. 3.21). Node 2, which was connected to the hydrogen sensor, transmitted sensor information to Node 3 since it was the only node within transmission range of Node 2. Similarly, Node 4 was the parent node of Node 3 due to its proximity and Node 1 was the parent node of Node 4 (Fig. 3.22). The test

Fig. 3.22 Nominal physical layout of hydrogen sensor network nodes. Node 2 is connected with the hydrogen sensor placed within the test chamber, Nodes 3 and 4 are relays, and Node 1 is the host node; the nodes can of course be re-located within the general area. The described configuration allows Node 3 to relay sensor information for Node 2 and Node 4 for Node 3; for this configuration the communication to Node 2 will be lost if either Node 3 or Node 4 is damaged [7]

3 Non-carbon Nanotubes: Hydrogen Sensors Based on TiO2

53

chamber temperature was held constant at 2908C, and the environment was repeatedly cycled between a nitrogen ambient and nitrogen + 1,000 ppm H2. Every 2 s, Node 2 converted the electrical resistance of the hydrogen sensor to voltage and transmitted the information to Node 1 using Nodes 3 and 4 as relays. The host PC downloaded the measurement information from Node 1 every 6 s. The host computer also instructed the host node to send out a broadcast signal every 5 min to re-establish communication links between nodes. As the sensor is switched from a nitrogen to a nitrogen + 100 ppm hydrogen atmosphere, there is a factor of 200 change in measured voltage, as shown in Fig. 3.23. Figure 3.24A shows the voltage recorded by Node 2 as the concentration of hydrogen is varied from 50 to 104 ppm, and Fig. 3.24B shows the results as the concentration of hydrogen is varied from 50 to 1,000 ppm. The voltage ranges from 0.049 to 2.046 V, linearly corresponding to sensor resitivities of 245–10.23 k . The behavior of the sensor is consistent, recovering its original resistance after repeated exposure to varying hydrogen concentrations. The sensor responds linearly to hydrogen concentrations of approximately 500 ppm, with an exponential response above this value. Figure 3.23 shows a factor of 200 change in measured electrical impedance upon exposure to 100 ppm hydrogen. With respect to the ultimate sensitivity we can expect from titania nanotube hydrogen sensors using resistive bridge

Fig. 3.23 The voltages recorded by the sensor network when the hydrogen concentration is cycled between 0 and 100 ppm [7]

54

K.E. LaFlamme, C.A. Grimes

Fig. 3.24 A The sensitivity of the sensor toward hydrogen. The response is best fitted with a function V = 11.98  h0.8375, where V is the measured voltage in volts and h is the hydrogen concentration in ppm. Note: 10,000 ppm is equal to 1% hydrogen atmosphere. B The curve fit on the figure is log V = C1h3 + C2h2 + C3h + C4 where V is the measured voltage in Volts, h is hydrogen concentration in ppm and C1 = 1.22  109, C2 = 4.48  106, C3 = 5.04  103, and C4 = 0.271 [7]

3 Non-carbon Nanotubes: Hydrogen Sensors Based on TiO2

55

circuits, changes in electrical resistivity of 1% can be readily measured. However, if we consider a 10% change in electrical resistance, and linear scaling between hydrogen concentration and electrical resistance as seen in Fig. 3.24B for low hydrogen concentrations, a hydrogen resolution of 0.05 ppm appears readily achievable. The sensor network is a robust system, which is able to function if some of the nodes are damaged or removed. This was demonstrated by both relocating nodes within the network and removing nodes from the network. For example, consider rearranging Nodes 3 and 4 so that both are within the communication range of Node 2, as shown in Fig. 3.25. At first, Node 2 recognizes Node 3 as its parent since it receives the broadcast signal from Node 3 before Node 4. Node 2 occasionally switches its parent node to Node 4 when the sensor nodes reestablish communications every 5 min. When Node 3 is turned off, the host node stops receiving data from Node 2 since the link is cut off. However, when the host node sends out another broadcast signal, Node 2 re-connects to the network again through Node 4 and the measurement continues. The power of a sensor network is that widespread communication can continue even if one or more of the nodes is removed from the network. In this design, the sensor node can operate for over 100 h on three AA batteries. In the future, the sensor node can become even more energy efficient by self-engaging low-power mode when idle and turning on only when it needs to collect and transmit data. This can extend the lifetime of the sensor node to a few months or even years. Beyond hydrogen monitoring, the sentinel sensor network can also be applied to other applications by integrating the nodes with different types of sensors. For example, with proper integration of the nodes with desired sensors,

Fig. 3.25 This configuration allows both Nodes 3 and 4 to relay sensor data for Node 2. Although shorter in communication range, the network can still function even if Node 3 or Node 4 is damaged [7]

56

K.E. LaFlamme, C.A. Grimes

the network can be used to monitor for chemical or biological events in airports, theaters, and other public venues. The sensor network can also be used for monitoring of pollution levels, be they industrial, medical, or at home.

3.5 Summary This chapter summarizes the fabrication of TiO2 nanotube arrays for hydrogensensing applications via anodization. The nanotube sensors contain both anatase and rutile phases of titania and show appreciable sensitivity to hydrogen at temperatures as low as 258C; furthermore, their response behavior is highly reproducible. The sensitivity of the sensors increases drastically with temperature, showing an increase of nearly 3 orders of magnitude from a starting temperature of 180–4008C. Physical properties of the nanotubes can also affect their performance, with higher hydrogen sensitivity associated with decreased tube diameter and increased tube length. The sensors are highly selective toward hydrogen compared with carbon monoxide, ammonia, and carbon dioxide. While oxygen does reduce the sensitivity of sensor, there is still a clearly detectable hydrogen signal. On the other hand, the sensors are significantly affected by moisture, thus it is important when designing a sensor to calibrate it to RH. Hydrogen sensors such as the ones described here are exceedingly versatile and can be used in many different applications. The sensors are able to selfclean in a photocatalytic manner, which is valuable in a number of environmental and industrial applications where the potential for contamination by vapors, soot, and the like is high. The sensors can also be used in the medical field to detect a variety of ailments such as lactose intolerance or necrotizing enterocolitis. The sensors can also be incorporated into networks for widespread hydrogen monitoring. In conclusion, work so far on the interaction of titania nanotube arrays with hydrogen has revealed an unprecedented gasdependent shift in electrical resistance. Extension of the technology presented here to other metal oxides should enable dramatically improved gas-sensing materials.

References 1. Capovilla J, VanCouwenberghe C, Miller WA (2000) Noninvasive blood gas monitoring. Crit Care Nurs Q 23:79–86 2. Carter BG, Wiwczaruk D, Hochmann M, Osborne A, Henning R (2001) Performance of transcutaneous PCO2 and pulse oximetry monitors in newborns and infants after cardiac surgery. Anaesth Intens Care 29:260–265 3. Chong SKF, Ramadan AB, Livesey E, Wood G (2002) The use of a portable breath hydrogen analyser in screening for lactose intolerance in paediatric patients with chronic abdominal pain or chronic diarrhoea. Gastroenterology 122:M1827 Suppl. 1821 APR 4. Engel RR, Virnig NL (1973) Origin of mural gas in necrotizing enterocolitis. Pediatric Res 7:292A

3 Non-carbon Nanotubes: Hydrogen Sensors Based on TiO2

57

5. Garstin WIH, Boston VE (1987) Sequential assay of expired breath hydrogen as a means of predicting necrotizing enterocolitis in susceptible infants. J Pediatr Surg 22:208 6. Godoy G, Truss C, Philips J (1986) Breath hydrogen excretion in infants with necrotizing enterocolitis. J Pediatr Res 20:348A 7. Grimes CA, Ong KG, Varghese OK, Yang X, Mor G, Paulose M, Dickey EC, Ruan C, Pishko MV, Kendig JW, Mason AJ (2003) A Sentinel sensor network for hydrogen sensing. Sensors 3:69–82 8. McIntosh N, Becher JC, Cunningham S, Stenson B, Laing IA, Lyon AJ, Badger P (2000) Clinical diagnosis of pneumothorax is late: use of trend data and decision support might allow preclinical detection. Pediatr Res 48:408–415 9. Mor GK, Carvalho MA, Varghese OK, Pishko MC, Grimes CA (2004) A roomtemperature TiO2-nanotube hydrogen sensor able to self-clean photoactively from environmental contamination. J Mater Res 19:628–634 10. Mor GK, Varghese OK, Paulose M, Grimes CA (2003) A self-cleaning, room-temperature titania-nanotube hydrogen gas sensor. Sens Lett 1:42–46 11. Mor GK, Varghese OK, Paulose M, Mukherjee N, Grimes CA (2003) Fabrication of tapered, conical-shaped titania nanotubes. J Mater Res 18:2588 12. Moukarzel AA, Lesicka H, Ament ME (2002) Irritable bowel syndrome and nonspecific diarrhea in infancy and childhood relationship with juice carbohydrate malabsorption. Clin Pediatr (Phila) 41:145–150 13. Paulose M, Varghese OK, Mor GK, Grimes CA, Ong KG (2006) Unprecedented ultrahigh hydrogen gas sensitivity in undoped titania nanotubes. Nanotechnology 17:398–402 14. Pimentel M, Chow EJ, Lin HC (2000) Comparison of peak breath hydrogen production in patients with irritable bowel syndrome, chronic fatigue syndrome and fibromyalgia. Gastroenterology 118:2141 Part 2141 Suppl 2142 15. Riordan SM, McIver CJ, Duncombe VM, Thomas MC, Bolin TD (2000) Evaluation of the rice breath hydrogen test for small intestinal bacterial overgrowth. Am J Gastroenterol 95:2858–2864 16. Tobias JD, Wilson WR, Jr, Meyer DJ (1999) Transcutaneous monitoring of carbon dioxide tension after cardiothoracic surgery in infants and children. Anesth Analg 88:531–534 17. Varghese OK, Gong D, Paulose M, Ong KG, Dickey EC, Grimes CA (2003) Extreme changes in the electrical resistance of titania nanotubes with hydrogen exposure. Adv Mater 15:624–627 18. Varghese OK, Gong D, Paulouse M, Ong KG, Grimes CA (2003) Hydrogen sensing using titania nanotubes. Sens Actuat B 93:338–344 19. Varghese OK, Grimes CA (2003) Metal oxide nanoarchitectures for environmental sensing. J Nanosci Nanotechnol 3:277–293 20. Varghese OK, Mor GK, Grimes CA, Paulose M, Mukherjee N (2004) A titania nanotube-array room-temperature sensor for selective detection of hydrogen at low concentrations. J Nanosci Nanotechnol 4:733–737 21. Varghese OK, Yang X, Kendig J, Paulose M, Zeng K, Palmer C, Ong KG, Grimes CA (2006) A transcutaneous hydrogen sensor: From design to application. Sens Lett 4:120–128 22. Yoriya S, Prakasam HE, Varghese OK, Shankar K, Paulose M, Mor GK, Latempa TA, Grimes CA (2006) Initial studies on the hydrogen gas sensing properties of highly-ordered high aspect ratio TiO2 nanotube-arrays 20 mm to 222 mm in length. Sens Lett 4:334–339

Chapter 4

Alternative Nanostructured Sensors: Nanowires, Nanobelts, and Novel Nanostructures Abhishek Prasad, Samuel Mensah, Zheng Wei Pan, and Yoke Khin Yap

4.1 Introduction State-of-the-art microfabrication techniques have led to smaller and faster computers and electronic devices. The emergence of nanoscale science and engineering (NSE or the so-called nanotechnology) has led toward the fabrication of much smaller and faster devices based on nanostructured materials. One-dimensional (1-D) nanostructures such as carbon nanotubes (CNTs) [1, 2, 3, 4, 5, 6, 7, 8] and boron nitride nanotubes (BNNTs) [9, 10, 11, 12, 13, 14,] have attracted tremendous research interest in the past decade. While the growth of these nanotubes with desired electronic properties is still challenging, scientists and engineers have extended their interest into alternative nanostructures such as nanowires that may offer better chances to achieve controllable electronic properties. These alternative nanostructures did not have the unique closed-shell structures like CNTs and BNNTs. Instead, they are often based on the known semiconductors such as Si, InP, GaP, ZnO, and GaN. These nanostructures usually maintain the crystal structures of their bulk precursors and thus have well-established physical and chemical properties that can be employed for applications such as biological and chemical sensors. In this chapter, we will review the current research status of these alternative nanostructures in various aspects including their synthesis processes, and their applications for sensors. This chapter will cover nanostructures such as nanowires, nanocombs, nanobelts, nanorods, nanoswords, and nanosquids. The overview and appearance of these nanostructures will first be introduced in Section 4.2, followed by their synthesis and fabrication techniques in Section 4.3. Examples of these nanostructures for sensing will be given in Section 4.4.

A. Prasad Department of Physics, Michigan Technological University, Houghton, MI 49931, USA

F.J. Arregui (ed.), Sensors Based on Nanostructured Materials, DOI: 10.1007/978-0-387-77753-5_4, Ó Springer ScienceþBusiness Media, LLC 2009

59

60

A. Prasad et al.

4.2 Novel Nanostructures Nanoscale materials offer larger surface area (A) to volume (V) ratios (A/V) than the bulks. Thus nanostructures are believed to have better performances than materials in the micro- or larger scales for sensing applications. The earlier investigations of these nanostructures have been demonstrated on the growth of carbide and superconducting nanorods [15, 16]. Later, efforts have been devoted on the study of semiconducting nanowires such as Si [17], GaN [18], InP [19], and GaP [20]. Interesting ribbon- or beltshaped nanostructures were then demonstrated for a series of metal oxide nanostructures including SnO2 [21], ZnO, In2O3, and CdO [22], which are now generally referred as nanobelts. Among these metal oxides, ZnO appeared to be structurally interesting. Nanostructures of ZnO can appear as nanowires, nanobelts, nanopropellers, nanocombs, nanotubes, nanoswords, nanotripods, nanotetrapods, nanosquids, nanorods, nanotips, etc., as shown in Fig. 4.1 [23, 24, 25, 26].

4.3 Methods of Synthesis and Fabrication In this section we will introduce various growth techniques employed for the synthesis of some alternate nanostructures, including nanowires [17, 18, 19, 20, 27, 28, 29], nanobelts [21, 22, 23, 24, 25, 26, 30, 31,32], nanocombs [23, 24, 25, 26, 33, 34], nanotubes, and nanosquids [23–26]. We have classified these growth techniques into physical vapor deposition (PVD, Section 4.3.1), chemical vapor deposition (CVD, Section 4.3.2), solution-based chemistry (Section 4.3.3), and other synthesis techniques (Section 4.3.4).

4.3.1 Physical Vapor Deposition Physical vapor deposition (PVD) are vapor–solid (VS) deposition techniques that involve the generation of reactant vapors by physical processes such as heats, plasmas, and lasers.

4.3.1.1 Laser-Assisted Catalytic Growth Laser-assisted catalytic growth (LCG) was among the earlier approaches for the growth of semiconducting nanowires [17, 18, 19, 20]. This technique involves the use of a pulsed laser to ablate a target that containing the element(s) desired in the nanowires and the metal catalyst component. In this technique,

4 Alternative Nanostructured Sensors

61

Fig. 4.1 In addition to nanowires, ZnO nanostructures can appear in various morphologies including (left to right, top to bottom) nanobelts, nanopropellers, nanocombs, nanotubes, nanoswords, nanotripods, nanotetrapods, nanosquids, nanorods, and nanotips. Scale bar = 1 mm

the target was kept inside the growth chamber of a tube furnace. The vaporized compounds were then grown into nanowires of Si [17], GaN [18], InP [19], and GaP [20], through the vapor–liquid–solid (VLS) mechanism [35]. This approach was used to prepare bulk quantities of single-crystalline nanowires

62

A. Prasad et al.

with diameters of as small as 3 nm and lengths as long as 30 mm [17]. This technique was similar to the growth of single-wall carbon nanotubes by laser vaporization [36].

4.3.1.2 Thermal Evaporation Thermal evaporation was one of the most widely used techniques for the growth of alternative nanostructures. In most cases, horizontal furnaces are used with a tubular reaction chamber constructed by an alumina tube or a quartz tube. This was the technique used for the growth of nanobelts of SnO2 [21], ZnO, In2O3, and CdO [22]. For example, the growth of SnO2 nanobelts was conducted in a horizontal tube furnace with an alumina tube chamber. SnO2 powders were used as the source material that was evaporated at 1,3508C under a pressure of 200–300 Torr and Ar gas flow rate of 50 sccm. These vapors were then condensed as belt-like nanostructures in a narrow region downstream where the temperature was 900–9508C. This vapor–solid condensation process did not involve the use of other metal catalysts, although Sn itself could have mediated the formation of SnO2 nanobelts. This approach was used for the growth of interesting diskettes of SnO2 [31] in a horizontal tube furnace as shown in Fig. 4.2. In this case SnO2 powders were vaporized at 1,0508C under a pressure of 500–600 Torr with the flow of Ar gas. SnO2 diskettes were deposited at the low temperature (LT) region (200–4008C). By using the same approach, Ga2O3 nanoribbons and nanosheets were produced by thermal evaporation of GaN powders in a horizontal alumina tube furnace at 1,1008C [30]. These Ga2O3 nanostructures were condensed at 800–8508C under a pressure of 300 Torr created by the flow of Ar gas at a rate of 50 sccm. Various ZnO nanostructures can be produced by thermal evaporation of Zn powders. ZnO nanotetrapods were synthesized in a horizontal tube furnace with a fused-quartz tube chamber at 9008C. A quartz plate held with a few spherical Zn pellets (3 mm in diameter) was inserted into the tube and vaporized in air ambient. After heating for about 2 min, ZnO nanotetrapods were found on the surface of the quartz plate [37]. These ZnO nanotetrapods were used for humidity sensors. On the other hand, ZnO nanobelts can be grown by evaporation of Zn powder at 6008C. Results indicate that the control of gas flow rates and partial pressures of Ar, O2, and Zn vapors are important for the growth of ZnO nanobelts [38]. Finally, the growth of ZnO nanowires in quantity of several grams was demonstrated by heating the mixture of Zn powders and NaCl to 600–7008C in the flow of Ar (25 sccm) and O2 (20 sccm) gases that create a pressure of 2 Torr [39]. About 70–80% of the Zn powders were converted to ZnO nanowires when NaCl was used, otherwise 5–10% without NaCl.

4 Alternative Nanostructured Sensors

63

Fig. 4.2 (Top) Various appearances of SnO2 diskettes (a to e). (Bottom) Schematic diagram for a horizontal tube furnace for the growth of various oxide nanostructures by thermal vaporization. Nanostructures can be condensed at high (H.T.), medium (M.T.), and low temperature (L.T.) regions downstream

64

A. Prasad et al.

4.3.1.3 Radiofrequency Magnetron Sputtering Vapor–solid growth of nanostructured materials can also be obtained by other physical vapor deposition technique such as radiofrequency (rf) magnetron sputtering. For example, ZnO nanobelts were grown by using rf magnetron sputtering of a ZnO target [40]. These nanobelts were deposited on sapphire substrates without the use of catalyst at relatively low pressure of 40 mTorr and an rf power of 300 W for 60 min. The sputtering gas used in this technique is not indicated in this report. The deposition of these nanobelts was carried out without external heating to the substrates. This approach usually results in the formation of ZnO films but nanobelts were deposited when sapphire substrates were used.

4.3.2 Chemical Vapor Deposition Chemical vapor deposition (CVD) techniques involved chemical decomposition or chemical reaction in the formation of nanostructured materials. These chemical reactions required to either create the growth species or form nanostructures by the use of catalyst through the vapor–liquid–solid (VLS) mechanism. We will discuss two types of CVD techniques, namely thermal CVD and metal–organic chemical vapor deposition (MOCVD).

4.3.2.1 Thermal Chemical Vapor Deposition Various ZnO nanostructures can be grown by carbothermal reduction of ZnO powders either with or without the use of catalyst such as Au [23–26]. In this case, the precursor ZnO and graphite powders were mixed and loaded into a ceramic boat, which was placed in a small quartz reaction tube. This reaction tube was then inserted into a larger quartz tube of a horizontal tube furnace. This approach was sometimes referred as double-tube thermal CVD. Oxidized Si substrates (or other substrates) were usually placed downstream to the source materials at the lower temperature zones. In most cases, these precursor powders were combusted at 1,1008C so that ZnO powders will be reduced into Zn and ZnOx vapors in the presence of graphite powders. These Zn and ZnOx vapors were then condensed as ZnO nanostructures in O2 gas at a pressure of several Torrs. ZnO appeared as nanowires, nanobelts, nanopropellers, nanocombs, nanotubes, nanoswords, nanotripods, nanotetrapods, nanosquids, nanorods, nanotips, etc., as shown in Fig. 4.1 [23, 24, 25,26]. We found that nanorods can be grown without Au catalyst. These ZnO nanorods can be transformed into nanotubes and nanosquids with appropriate cooling during the growth [23, 24, 26]. These nanorods can also be transformed into long nanowires by placing a Au-coated substrate beside the sampling substrates during the growth, a so-called ‘‘side-catalyst’’ approach [25].

4 Alternative Nanostructured Sensors

65

This technique was applied for the growth of ZnO nanocombs by heating the tube furnace at 9008C for 30 min. Argon (100 sccm) and oxygen (1 sccm) were used in this case without Au catalyst [41]. Under this oxygen-deficient ambient, Zn vapors/droplets might be formed that functions as a catalyst for the growth of ZnO nanocombs. These nanocombs were tested as the biosensors for glucose detection. ZnO nanowires can also be grown by using ZnO and graphite powders with germanium (Ge) as the catalyst [29]. In this case, Ge catalysts were supplied in two ways. In one method, GeO2 powders were reduced by the graphite powders together with the ZnO powders. In the other method Ge dots were patterned on the SiO2-coated silicon wafer by photolithography. ZnO nanowires were condensed at the temperature zone of 500–6508C with diameters and lengths of 50–400 nm and 50–200 mm, respectively. It is interesting to note that the Ge catalysts remain at the tips of these nanowires with diameters much larger than those of the nanowires. In addition to ZnO, other nanostructured metal oxides were grown by this vapor–liquid–solid approach. For example, SnO2 nanowires were grown directly on the electrodes of sensors and tested for NO2 gas sensing [42]. The synthesis of these SnO2 nanowires was carried out by thermal evaporation of Sn metal powders at low temperatures of 600–7008C. This was performed in the flows of O2 gas at a flow rate of 10 sccm and growth pressure of 1–10 Torr. Gold films of 0.2–0.3 nm were used as the catalyst for the growth of these nanowires. The diameters of the resulted SnO2 nanowires were ranged from 50 to 100 nm, which was determined by the growth temperatures. The growth of silicon oxide (SiOx) nanowires was demonstrated by a simple thermal CVD approach [28]. In this case, GaN powders were used as the Ga source, i.e., the metal catalyst. The Si sources can be the Si wafers, SiO powders, or silane (SiH4) gas. The GaN powders were placed at the location with temperature around 1,1508C, no matter what Si source was used. This technique was based on the VLS process. In fact, bundles of these SiOx nanowires can be grown from a micrometer-sized Ga droplet. They can assemble into gourdlike, spindle-like, badminton-like, and octopus-like morphologies at different temperature zones [43]. For example, the badminton-like structures can be formed at 980–1,0108C (Fig. 4.3a and b). These SiOx nanowires exhibit branching growth features with a height of 10–15 mm (Fig. 4.3c). Thermal CVD was used for the growth of doped nanostructures. For example, Sb-doped SnO2 nanowires were grown for gas sensor application [44]. In this case, metal Sn and Sb powders mixed in the weight ratio of 10:1 were used as the evaporation sources in a thermal CVD system. Si substrates coated with 5 nm gold were placed downstream in the furnace. The temperature at the center of quartz tube was 9008C, and a constant flow of 1% oxygen and 99% nitrogen was maintained at a flow rate of 5 l/min. The as-grown nanowires have a diameter of 40–100 nm and lengths up to several tens of micrometers. As discussed, CVD has been a very useful and versatile technique. Nanostructures that were initially grown by other techniques have later been

66

A. Prasad et al.

Fig. 4.3 Badminton-like SiOx nanowire at different magnification (a, b). (c) TEM image of a badminton-like structure

successfully grown by CVD. For example, instead of laser-assisted catalytic growth, silicon nanowires are now commonly grown by thermal CVD [27, 45]. This was carried out using Au nanoparticles as the catalyst and Silane (SiH4, 10% in He) gas as the Si source.

4.3.2.2 Metal–Organic Chemical Vapor Deposition (MOCVD) Thermal CVD techniques discussed so far are not always the suitable approach for the growth of certain nanostructures. For example, GaN nanowires have been grown by laser ablation [18] and thermal CVD [46, 47]. High temperatures were needed to generate Ga vapor source in the thermal CVD technique, and Ni, Fe, and Au were used as the catalysts. The use of a solid Ga is technically

4 Alternative Nanostructured Sensors

67

simple but often leads to nonconstant vapor pressures for continuous growth of GaN nanowires. Metal–organic chemical vapor deposition (MOCVD) was found to be a more versatile technique for the growth of GaN nanowires. The MOCVD technique seems to overcome the vapor pressure problem and yield high-quality GaN nanowires [48]. In the MOCVD approach, trimethylgallium (TMG) and ammonia gas were used as Ga and N precursors. Silicon substrates or c-plane and a-plane sapphire substrates coated with 2–10 nm thin film of Ni, Fe, or Au were used for the growth at 800–1,0008C. The reaction was carried out in an oxygen-free environment at atmospheric pressure. TMG was kept cool in a – 108C temperature bath. Nitrogen was used as a carrier gas and percolated through the TMG precursor and coupled with a second nitrogen line to give a total nitrogen flow rate of 250 sccm. Hydrogen and ammonia gases were supplied at a total flow rate of 155 sccm. The as-grown GaN nanowires were having diameters of 15–100 nm and lengths of 1–5 mm, oriented predominantly along the [210] or [110] direction. MOCVD technique was also used for growing nanostructured IrO2 crystals [49]. The nanostructured crystals were grown on a gold-coated quartz substrate and their gas-sensing properties were studied by quartz crystal microbalance (QCM) technique. The growth of these nanostructures was performed by using (methylcyclopentadienyl) (1,5-cyclooctadiene) iridium, (MeCp)Ir(COD), as the precursor. The quartz substrate temperature (Ts) was kept between 350 and 5008C. High-purity oxygen was used as carrier gas at a flow rate of 100 sccm, which leads to a growth pressure of 7.2 Torr. The temperature of the precursor reservoir (Tp) was varied between 95 and 1058C. Different morphologies, such as nanoblades, layered columns, incomplete nanotubes, and square nanorods, were observed at various combinations of Ts and Tp.

4.3.3 Solution-Based Chemistry In addition to the vapor phase growth techniques discussed so far, several solution phase deposition techniques have been used for the growth of various alternative nanostructures.

4.3.3.1 Hydrothermal Synthesis Tungsten oxide nanowires (WO2.72) were grown by hydrothermal technique and were tested for hydrocarbon sensing [50]. Sensors based on WO2.72 nanowires show high sensitivity for 50–2,000 ppm of LPG (propane–butane mixture) at 2008C as well as relatively short recovery and response times. These WO2.72 nanowires were prepared by solvothermal synthesis. Tungsten hexachloride (1 g) was loaded into a 25 ml autoclave filled with ethanol up to 90% of its volume. Hydrothermal synthesis was carried out at 2008C for 24 h. The product

68

A. Prasad et al.

obtained by centrifugation was washed with ethanol. The as-grown nanowires were having diameters of 5–15 nm and lengths of 100–200 nm. ZnO nanorods can also be grown by hydrothermal techniques [51]. Zinc and cetyltrimethylammonium bromide (CTAB) were used as the precursors. CTAB (1.5 g) was dissolved in deionized water (35 ml) to form a transparent solution. Then, zinc powders (1.8 g) were added to the above solution under continuous stirring. The resulting suspension was transferred into a teflonlined stainless steel autoclave (volume 40 ml) and sealed tightly. Hydrothermal treatments were carried out at 1828C for 24 h. The autoclave was then cooled down. Next precipitates were collected, washed with deionized water for several times, and dried in air. The lengths of the ZnO nanorods are usually shorter than 1 mm and their diameters ranging from 40 to 80 nm. These nanorods were used for the sensing of various vapors including alcohol, LPG, gasoline, ammonia, etc. Vanadium oxide (V2O5) nanobelts coated with Fe2O3, TiO2, and SnO2 nanoparticles have been prepared by mild hydrothermal reaction [52]. For the growth of the V2O5 nanobelts, nitric acid was added dropwise to a 0.1 M ammonium metavanadate solution until the final pH value of the solution reached about 2–3 under stirring. Solution obtained was transferred to a teflon-lined autoclave and filled with deionized water up to 80% of the total volume. Then the autoclave was kept at 1808C for 24 h. The final product was washed with deionized water and pure alcohol several times to remove any possible remnants. The as-grown nanobelts were tens of micrometers long with smooth surfaces, typically 60–100 nm wide and 10–20 nm thick. These nanobelts were then coated with the oxide nanoparticles for the sensing of alcohol, benzene, cyclohexane, gasoline, ammonia, etc. NiFe2O4 nanospheres, nanocubes, and nanorods were prepared by a hydrothermal method [53]. For synthesizing these NiFe2O4 nanostructures, Ni(NO3)2 6H2O and Fe(NO3)3 9H2O were dissolved in deionized water to form a mixed solution with [Na2+] ¼ 0.10 mol/l and [Fe3+] ¼ 0.20 mol/l. NaOH solution (6.0 mol/l) was added dropwise under stirring into 20.0 ml of the mixed solution until the desired pH value was attained to form an admixture. In the next step, the admixture was transferred to a teflon autoclave (50 ml volume) with a stainless steel shell up to 80% of the total volume. The autoclave was kept at 120–2008C for 24–96 h and then cooled down. The final product was washed with deionized water and alcohol for several times and then dried. The length and diameter of the nanorods were about 1 mm and 30 nm, respectively; the side length of the nanocubes was about 60–100 nm. It was found that sensors based on NiFe2O4 nanorods were relatively sensitive and selective to triethylamine. 4.3.3.2 Hydrolysis The growth of Fe2O3/ZnO core/shell nanorods was reported by a solution phase-controlled hydrolysis process of Zn2+ ions in the presence of Fe2O3

4 Alternative Nanostructured Sensors

69

nanorods [54]. One hundred milligram of Fe2O3 nanorods (100 nm long, 25 nm in diameter) were dispersed in 200 ml deionized water by ultrasonication. Then 20 mg of zinc acetate (Zn(Ac)2 6H2O)) was introduced to the solution, and the suspension was heated in an oil bath at 408C under vigorous stirring. Twenty milliliters of 5% ammonia was then added into the suspension for 30 min, and the reaction was maintained at the temperature for 1 h. The colloidal suspension was then centrifuged and sintered at 4008C for 2 h to obtain Fe2O3/ZnO nanorods. These core/shell structures have diameters of 30 nm. 4.3.3.3 Aqueous Chemical Growth ZnO nanostructures were deposited on glass substrates by the aqueous chemical growth technique at 958C [55]. The growth process involves the use of an equimolar (0.01 M) aqueous solution of Zn (NO3)2 6H2O and C6H12N4 as the precursors. The solution and the substrates were then placed in glass bottles and heated at 958C for 1, 5, 10, and 20 h. After each induction time, the substrates were thoroughly washed with deionized water and dried in air. Flower-like aggregations of ZnO nanorods were deposited in all cases. The diameters of these nanorods are relatively big (500–1,000 nm) and increased with the deposition duration. Their lengths are usually <10 mm.

4.3.4 Other Synthesis Techniques 4.3.4.1 Electrospinning The electrospinning technique was employed to produce nanofibers of MoO3 and WO3 [56]. In general, the electrospinning process is based on the principle that material is extracted by an electric field from a polymer solution reservoir. When the electric field has reached a critical intensity, the polymer solution will be ejected toward the opposite electrode (substrate). During this ejection process, a series of hydrodynamic instabilities and solvent evaporation processes will lead to the deposition of polymer nanofibers. For the growth of MoO3 and WO3 nanofibers, molybdenum isopropoxide and tungsten isopropoxide were used as precursors, respectively. These precursors were mixed with n-butanol to form 0.1 M solutions. The agitated and aged metal oxides were then combined with polyvinylpyrrolidone (PVP) solution (0.05–0.1 mM in ethanol). Each mixture is magnetically stirred for 45 min before filling into a syringe. The syringe pump was programmed to dispose 0.5 ml of solution at various flow rates. Substrates were mounted onto aluminum foil and placed on the grounded collecting screen. A high voltage (18 kV) was then applied to the needle of the syringe to create electric field between the syringe and the substrate to extract the sol-gels into nanofibers. All samples were annealed in air to obtain the end products. These nanofibers have diameters ranging from 80 to 200 nm.

70

A. Prasad et al.

4.4 Sensing Applications: Biological Sensing and Chemical Sensing In this section we attempt to review the applications of various alternative nanostructures in sensing of biological and chemical species. Examples are given for nanocombs, nanowires, nanotubes, and nanobelts.

4.4.1 Biological Sensing ZnO nanocombs were investigated for glucose detection [41]. The zinc oxide nanocomb glucose biosensor showed a high sensitivity (15.33 mA/cm2 mM) for glucose detection and high affinity of glucose oxidase (GOx) to glucose. ZnO nanocombs were transferred from silicon substrate to a standard gold electrode. Then ZnO/gold electrode was wetted by 0.01 M phosphate buffer solution (PBS) and dried in nitrogen gas. Five microliter of GOx solution was dropped onto the surface of the ZnO/gold electrode. After the evaporation of water, a 5 ml 0.5% weight Nafion solution was added dropwise onto the GOx/ ZnO/gold electrode. Next the electrode was dried overnight at 48C to form a film, which is essential for the attachment of GOx /ZnO nanocombs to the gold electrode. Amperometric performance of the GOx/ZnO biosensor was measured. Figure 4.4a shows amperometric response of the ZnO biosensor after addition of successive aliquots of 20 mM glucose in PBS under stirring. Biosensor shows a rapid and sensitive response to the change of glucose concentration, indicating a good electrocatalytic property of Nafion/GOx/ZnO/gold electrode. The response time for the electrode is less than 10 s. The corresponding calibration curve (solid square) of the glucose biosensor is shown in Fig. 4.4b. With the increase of the glucose concentration, the response current increases and saturates at a high glucose concentration of 13 mM. The linear range of the calibration curve is from 0.02 to 4.5 mM with a sensitivity of 15.33 mA/cm2 mM. The detection limit of glucose is 0.02 mM. Si nanowires were used for the detection of DNA sequence [57]. These biosensors were based on two-terminal silicon nanowire electronic devices. The p-type Si nanowires were synthesized by thermal CVD as described earlier [27, 45] and functionalized with peptide nucleic acid (PNA) receptors. These receptors were linked to the surface of these nanowires using intervening avidin protein layer. These devices have a microfluidic delivery structure as shown in Fig. 4.5a. This design enabled the flow of test fluid to flow across the functionalized Si nanowires for sensing. PNA with complementary sequence was chosen since it will recognize the sequence of DNA of interest. The flow of solution without the DNA will result in no substantial change in conductance of the device (as in Fig. 4.5b). The flow of complementary DNA solution will initiate hybridization and lead to a change of transient conductance (as in Fig. 4.5c).

4 Alternative Nanostructured Sensors

71

Fig. 4.4 (a) Amperometric responses of Nafion/GOx/ZnO/gold electrodes with the successive addition of 20 mM glucose to the 0.01 M, pH 7.4 PBS buffer under stirring. (b) The calibration curve (solid square) of ZnO/GOx biosensor and the Lineweaver–Burk plot (open circle). The straight line is the linear fit to the calibration curve

The specific conductance changes due to PNA–DNA hybridization were obtained from the time-dependent conductance recorded with the same SiNW device. Direct comparison of these data shown in Fig. 4.6a highlights the net conductance change associated with hybridization of the DNA complementary to the PNA receptor. Figure 4.6b shows a summary of concentration-dependent data. This shows that the wild-type DNA can effectively and selectively detect

72

A. Prasad et al.

Fig. 4.5 (a) Schematic of a sensor device consisting of a Si nanowire and a microfluidic channel, where the arrows indicate the direction of sample flow. (b) The Si nanowire surface with PNA receptor. (c) PNA–DNA duplex formation

concentrations as low as 10 fM. Figure 4.6c shows a systematic increase in net conductance with increasing DNA concentrations. Nanofluidic devices using silica nanotubes were created for single DNA sensing [58]. Figure 4.7a shows the schematic of nanofluidic device which contains a single silica nanotube bridging across two fluidic channels/reservoirs. Figure 4.7b shows transmission electron microscope image of a silica nanotube synthesized by translating silicon nanowire by a series of oxidation/etching process. Note that single-crystalline ZnO nanotubes can now be grown in one process and may enhance sensing capability of similar devices in the future [24]. Figure 4.7c is another view of the nanofluidic device with channels and ports. Scanning electron microscope image of a single nanotube with channel is shown in Fig. 4.7d. The inset image shows the opening of the nanotube embedded between two silicon dioxide layers. In a typical sensing experiment, both channels are filled with 2 M potassium chloride (KCl) buffer solution and change in ionic current is recorded with change in applied voltage. No transient current change was observed with pure buffer solution. When DNA molecules were introduced into the 2 M KCl buffer solution the ionic current exhibited frequent drops in current corresponding to the passage of DNAs through the nanotube. Current drop may be due to the

4 Alternative Nanostructured Sensors

73

Fig. 4.6 (a) Conductance graphs obtained from 100 fM of wild-type (solid line) and mutant (dashed line) fragments on the same nanowire. (b) Net conductance changes versus time for 100 fM (1), 30 fM (2), 10 fM (3) and 1 fM (4) DNA samples. (c) Conductance versus DNA concentration where data points shown as squares and circles were obtained from two independent SiNW devices

geometrical exclusion effect of conducting ions since DNA molecule has finite size which acts as the blockages in transient ionic current. Also the duration and current decay characteristics measured at different bias and ionic concentrations provide useful information.

4.4.2 Chemical Sensing Gas sensors were demonstrated by using SnO2 nanobelts [59]. Nanobelts were synthesized using thermal evaporation technique. Gas-sensing characterization showed that the nanobelts were sensitive to CO, NO2, and ethanol. Electrical characterization was carried out by a volt-amperometric technique at constant bias of 1 V, and changes in electrical current were measured by a picoammeter. Figure4.8a, b and c showsthe sensorresponse to ethanol, CO, and NO2,respectively. Sensor response is defined as the relative variation in conductance due to the introduction of the gas at 4008C in the case of ethanol and CO, or 2008C in the

74

A. Prasad et al.

Fig. 4.7 (a) Schematic of nanofluidic device. (b) Transmission electron microscopy image of a single silica nanotube (scale bar represents 100 nm). (c) View of nanofluidic device showing ports. (d) Scanning electron microscopy image of the nanofluidic device (scale bar represents 10 mm). Inset is the cross-section view of the silica nanotube (scale bar represents 100 nm)

case of NO2 and 30% relative humidity (RH). Sensor response is 4,160% for ethanol, 90% for CO, and 1,550% for NO2. The results clearly demonstrate the potential of nanobelt sensors with sensitivity at the level of a few ppb. Field-effect transistors using ZnO nanowires were investigated as chemical sensors for detection of NO2 and NH3 [60]. ZnO nanowires were synthesized by catalytic chemical vapor deposition technique. For these sensors, the detection sensitivity can be tuned by the back-gate potential. In addition, it was found that negative gate voltages can be used to increase the gas molecule desorption process at room temperature making these devices refreshable for further sensing. NO2 and NH3 gases were used for sensing characterization. The gate voltages required to refresh 1% NH3 is 30 V. A gate voltage of 60 V is required to refresh devices after exposure to 10 ppm of NO2 gas. The minimum gate voltages required to refresh the devices depend on the tested gas concentration and the type of tested gas as shown in Fig. 4.9a. The time-dependent response for the NO2 gas is shown in Fig. 4.9b. The conductance of these ZnO nanowires decreases when a negative gate voltage pulse of –20 V was applied at the back gate. The inset of Fig. 4.9b shows the

4 Alternative Nanostructured Sensors

75

(a)

(b)

(c)

Fig. 4.8 (a) Sensor response to ethanol at 4008C and 30% relative humidity. (b) Sensor response to CO at 4008C and 30% relative humidity. (c) Sensor response to NO2 at 4008C and 30% relative humidity

slopes (k) of the linear recovery regions for NO2 gases at various concentrations. The corresponding recovery slopes for NH3 gas are shown in Fig. 4.9c at different concentrations. As shown, these slopes are concentration independent and are different for NO2 and NH3. The larger recovery slopes for NH3 indicate that NH3 is much easier to be electrodesorbed than NO2.

76

A. Prasad et al.

Fig. 4.9 (a) Gate refresh voltage dependence on the concentration of NO2 and NH3 gases. (b) The recovery of the nanowire conductance under Vg = 20 V pulse in 10 ppm NO2. Inset shows the recovery curves at different concentrations of NO2 gas. (c) Recovery of curves for different concentrations of NH3 gas under Vg = 20 V pulse. (d) Temperature dependence of the nanowire conductance for Ar, NH3, and NO2

The difference of these recovery rates for the two gases is due to the difference of binding strength of these molecules on the ZnO nanowires. From these results, it is believed that NH3 gas molecules are having weaker surface-binding strength on ZnO nanowires as compared to NO2. According to the Volkenstein

4 Alternative Nanostructured Sensors

77

model, the authors suggested that acceptors like chemisorption states (Ea) are formed within the band gap of these ZnO nanowires. These chemisorbed gas molecules tend to capture electrons. Also the depth of these Ea states in the forbidden band is related to the conductance activation energy EA. As shown in Fig. 4.9d, EA (NH3) < EA (NO2). Acknowledgments Yoke Khin Yap acknowledges supports from the U.S. Department of Army (Grant number W911NF-04-1-0029 through the City College of New York), National Science Foundation CAREER Award (Award number 0447555, Division of Materials Research), the Defense Advanced Research Projects Agency (Contract number DAAD1703-C-0115 through the U.S. Army Research Laboratory), and the Department of Energy, the Office of Basic Energy Sciences (Grant No. DE-FG02-06ER46294, the Division of Materials Sciences and Engineering). Zheng Wei Pan is from the Faculty of Engineering, the University of Georgia.

References 1. S. Iijima, Nature 354, 56 (1991). 2. S. Iijima and T. Ichihashi, Nature 363, 603 (1993). 3. D. S. Bethune, C. H. Kiang, M. S. deVries, G. Gorman, R. Savoy, J. Vazquez, and R. Beyers, Nature 363, 605 (1993). 4. M. S. Dresselhaus and G. Dresselhaus, Eds., Carbon Nanotubes: Synthesis, Structure, Properties and Applications (Springer-Verlag, Berlin, 2001). 5. K. Hata, D. Futaba, K. Mizuno, T. Namai, M. Yumura, and S. Iijima, Science 306, 1362 (2004). 6. V. K. Kayastha, Y. K. Yap, Z. Pan, I. N. Ivanov, A. A. Puretzky, and D. B. Geohegan, Appl. Phys. Lett. 86, 253105 (2005). 7. J. Menda, B. Ulmen, L. K. Vanga, V. K. Kayastha, Y. K. Yap, Z. Pan, I. N. Ivanov, A. A. Puretzky, and D. B. Geohegan, Appl. Phys. Lett. 87, 173106 (2005). 8. V. K. Kayastha, S. Wu, J. Moscatello, and Y. K. Yap, J. Phys. Chem. C (Lett.) 111, 10158 (2007). 9. N. G. Chopra, R. J. Luyken, K. Cherry, V. H. Crespi, M. L. Cohen, S. G. Louie, and A. Zettl, Science 269, 966 (1995). 10. A. Loiseau, F. Willaime, N. Demoncy, G. Hug, and H. Pascard, Phys. Rev. Lett. 76, 4737 (1996). 11. W. Han, Y. Bando, L. Bourgeois, K. Kurashima, and T. Sato, Appl. Phys. Lett. 73, 3085 (1998). 12. C. Tang, Y. Bando, and T. Sato, Chem. Phys. Lett. 362, 185 (2002). 13. J. Wang, V. K. Kayastha, Y. K. Yap, Z. Y. Fan, J. G. Lu, Z. W. Pan, I. N. Ivanov, A. A. Puretzky, and D. B. Geohegan, Nano Lett. 5, 2528 (2005). 14. C. Zhi, Y. Bando, C. C. Tang, R. Xie, T. Sekiguchi, and D. Golberg, J. Am. Chem. Soc. 127, 15996 (2005). 15. H. Dai, E. W. Wong, Y. Z. Lu, S. Fan, and C. M. Lieber, Nature 375, 769 (1995). 16. P. Yang and C. M. Lieber, Science 273, 1836 (1996). 17. A. M. Morales and C. M. Lieber, Science 279, 208 (1998). 18. X. Duan and C. M. Lieber, J. Am. Chem. Soc. 122, 188 (2000). 19. X. Duan, Y. Huang, Y. Cui, J. Wang, and C. M. Lieber, Nature 409, 66 (2001). 20. M. S. Gudiksen and C. M. Lieber, J. Am. Chem. Soc. 122, 8801 (2000). 21. Z. R. Dai, Z. W. Pan, and Z. L. Wang, Solid State Comm. 118, 351 (2001). 22. Z. W. Pan, Z. R. Dai, and Z. L. Wang, Science 291, 1947 (2001).

78

A. Prasad et al.

23. S. L. Mensah, Master of Science Dissertation 2006, Department of Physics, Michigan Technological University. 24. S. L. Mensah, V. K. Kayastha, I. N. Ivanov, D. B. Geohegan, and Y. K. Yap, Appl. Phys. Lett. 90, 113108 (2007). 25. S. L. Mensah, V. K. Kayastha, and Y. K. Yap, J. Phys. Chem. C (Lett.) 111, 16092 (2007). 26. S. L. Mensah, A. Prasad, J. Wang, and Y. K. Yap, J. Nanosci. Nanotech. (Special Issue on Nanowires) 8, 233 (2008). 27. F. Patolsky, G. Zheng, and C. M. Lieber, Nat. Protoc. 1, 1711 (2006). 28. Z. W. Pan, S. Dai, and D. H. Lowndes, Tsinghua Sci.Technol. 10, 718 (2005). 29. Z. W. Pan, S. Dai, C. M. Rouleau, and D. H. Lowndes, Angew. Chem. Int. Ed. 44, 274 (2005). 30. Z. R. Dai, Z. W. Pan, and Z. L. Wang, J. Phys. Chem. B 106, 902 (2002). 31. Z. R. Dai, Z. W. Pan, and Z. L. Wang, Adv. Funct. Mater. 13, 9 (2003). 32. Z. L. Wang, Adv. Mater. 15, 432 (2003). 33. Z. W. Pan, S. M. Mahurin, S. Dai, and D. H. Lowndes, Nano Lett. 5, 723 (2005). 34. H. Q. Yan, R. R. He, J. Johnson, M. Law, R. J. Saykally, and P. D. Yang, J. Am. Chem. Soc.125, 4728 (2003). 35. G. W. Wagner and W. C. Ellis, Appl. Phys. Lett. 4, 89 (1964). 36. T. Guo, P. Nikolaev, A. Thess, D. T. Colbert, and R. E. Smalley, Chem. Phys. Lett. 243, 49 (1995). 37. X. Wang, Y. Ding, J. Zhang, Z. Zhu, S. You, S. Chen, and J. Zhu, Sens. Actuators B 115, 421 (2006). 38. Y.Huang,J. He, Y.Zhang, Y.Dai,Y. Gu, S. Wang, andC.Zhou, J.Mater. Sci. 41,3057 (2006). 39. J. Yang, W. Wang, Y. Ma, D. Z. Wang, D. Steeves, B. Kimball, and Z. F. Ren, J. Nanosci. Nanotech. 6, 2196 (2006). 40. A. Z. Sadek, W. Wlodarski, K. Kalantar-zadeh, and S. Choopun, Sensors 2005 IEEE, 1326. 41. J. X. Wang, X. W. Sun, A. Wei,Y. Lei, X. P. Cai, C. M. Li, and Z. L. Dong, Appl. Phys. Lett. 88, 233106 (2006). 42. I. Hwang, Y. Choi, J. Park, J. Park, K. Kim, and J. Lee, J. Korean Phys. Soc. 49, 1229 (2006). 43. Z. W. Pan, S. Dai, D. B. Beach, and D. H. Lowndes, Nano Lett. 3, 279 (2003). 44. Q. Wan and T. H. Wang, Chem. Commun. 30, 3841 (2005). 45. Y. Cui, L. J. Lauhon, M. S. Gudiksen, J. Wang, and C. M. Lieber, Appl. Phys. Lett. 78, 2214 (2001). 46. C. C. Chen, C. C. Yeh, C. H. Chen, M. Y. Yu, H. L. Liu, J. J. Wu, K. H. Chen Chen, L. C. Chen, J. Y. Peng, and Y. F. Chen, J. Am. Chem. Soc. 120, 2791 (2001). 47. J. C. Johnson, H. Choi, K. P. Knutsen, R. D. Schaller, P. Yang, and R. J. Saykally, Nat. Mater. 1, 106 (2002). 48. T. Kuykendall, P. Pauzauskie, S. Lee, Y. Zhang, J. Goldberger, and P. Yang, Nano Lett. 3, 1063 (2003). 49. T. W. Chao, C. J. Liu, A. H. Hsieh, H. M. Chang, Y. S. Huang, and D. S. Tsai, Sens. Actuators B 122, 95 (2007). 50. C. S. Rout, A. Govindaraj, and C. N. R. Rao, J. Mater. Chem. 16, 3936 (2006). 51. J. Xu, Y. Chen, D. Chen, and J. Shen, Sens. Actuators B 113, 526 (2006). 52. J. Liu, X. Wang, Q. Peng, and Y. Li, Sens. Actuators B 115, 481 (2006). 53. X. Chu, D. Jiang, and C. Zheng, Sens. Actuators B 123, 793 (2007). 54. S. Si, C. Li, X. Wang, Q. Peng, and Y. Li, Sens. Actuators B 119, 52 (2006). 55. G. Kenanakis, D. Vernardou, E. Koudoumasa, G. Kiriakidis, and N. Katsarakis, Sens. Actuators B 124, 187 (2007). 56. K. M. Sawicka, A. K. Prasad, and P. I. Gouma, Sens. Lett. 3, 1 (2005). 57. J. Hahm and C. M. Lieber, Nano Lett. 4, 51 (2004). 58. R. Fan, R. Karnik, M.Yue, D. Li, A. Majumdar, and P.Yang, Nano Lett. 5, 1633 (2005). 59. E. Comini, G. Faglia, G. Sberveglieri, Z. Pan, and Z. L. Wang, Appl. Phys. Lett. 81, 1869 (2002). 60. Z. Fan and J. G. Lu, Appl. Phys. Lett. 86, 123510 (2005).

Chapter 5

Nanosensors: Controlling Transduction Mechanisms at the Nanoscale Using Metal Oxides and Semiconductors Teresa Andreu, Jordi Arbiol, Andreu Cabot, Albert Cirera, Joan Daniel Prades, Francisco Hernandez-Ramı´ rez, Albert Romano-Rodrı´ guez, and Joan R. Morante

5.1 Introduction Nanotechnology is defined as the design and engineering of functional materials and devices through control of matter in dimensions of roughly 1–100 nm, where unique phenomena enable novel applications [1]. While nanotechnology allows us to take advantage of these exclusive phenomena and related properties, it offers us new possibilities and relationships among the different multidisciplinary effects. Nanotechnology not only occupies the fields of material science and engineering but also applies to fundamental physics, chemistry and biology. Figures 5.1–5.3 show examples of functional semiconductor nanostructures. The design and engineering of materials at the nanometer scale can be achieved from two opposed directions: (a) the reduction of bulk dimensions of the material to the nanometer scale, which is known as the top-down approach and (b) the assembly of molecules and atoms into structures up to the nanometer scale, known as the bottom-up approach. The scaling down performed in the microelectronics area during the last decades of the past century presents a clear example of the ‘‘top-down approach’’. Moore’s law describes the evolution followed by the computing power of the silicon chips. This evolution is doubling every 18–24 months due to the increase in the number of transistors integrated by area unit. Nowadays, transistor sizes have already reached dimensions down to 50 nm. However, in the near future, integration would not be able to follow this law since it will no longer be possible to shrink dimensions due to the intrinsic material characteristics. The bottom-up approach is a promising nanotechnology alternative to the extrapolation of the top–bottom methods inherited from the macroworld. During the last years, both top-down and bottom-up strategies have overlapped and coexisted, facilitating and accelerating the discovery of new functional materials. Organic and molecular-based transistors coming from the T. Andreu Department of Electronics, Faculty of Physics, University of Barcelona, Barcelona, E-08028, Spain

F.J. Arregui (ed.), Sensors Based on Nanostructured Materials, DOI: 10.1007/978-0-387-77753-5_5, Ó Springer ScienceþBusiness Media, LLC 2009

79

80

T. Andreu et al.

Fig. 5.1 Faceted TiO2 anatase nanocrystals

bottom-up approach are overlapping with the options obtained from the top-down processes. It is clear that all of these possibilities, developments or implementation of ideas dealing with the nanoworld require adequate tools and processes that allow us to measure, fabricate, characterize and, importantly, manipulate

Magnified Selections

FFTs (Digital Diffractograms)

Filtered Selections

3.26 ?

c 3.39 ?

a d b 3.32 ?

3.22 ?

Fig. 5.2 HRTEM analysis of tiny distorted and non-faceted SnO2 nanoparticles

5 Nanosensors

81

Fig. 5.3 Gold clusters decorating TiO2 anatase nanoparticles

nanostructures. In fact, new instruments and techniques with nanoscale resolution are accelerating scientific knowledge and technological implementations that promote new challenges in miniaturization. In spite of the growing activities and efforts in this field – that are revealed by a high number of published activities, specialized conferences, workshops and patents – it must be kept in mind that the way for converting basic discoveries into marketable products is long and hard. However, the new options are so flashing, exciting and attractive that scientists, technologists, policy makers, entrepreneurs and high-tech enterprises are converging in their activities for launching a broad range of novel products reaping social benefits for a sustainable and intelligent ambient world where nanosensor devices will play an outstanding role. Despite having such high expectations, it should be noted that, due to the lack of standardizations, nanotechnology is accompanied by many possible dangers. Many difficulties, hurdles and challenges have yet to be overcome. Among them, interfacing between the nanoworld and the macroworld or the effects of the nanomaterials on our health and on our environment are still relatively unclear.

5.2 Nanosensors The motivation for having sensors is given by our needs of monitoring – receiving information – the environment around us and having the capability of using the obtained data, after processing it, for interacting again with our environment. So, it is through sensors that we connect with the world, in the same way humans use

82

T. Andreu et al.

their senses. The concept sensor is derived from the Latin word ‘‘sentire’’ which means perceive (or sense, which is not clear and hence we need to go to Latin). The word sensor denotes an entity that responds to external stimulus or ‘‘energy variation’’ by generating a functionally related output [2] as a measure, either directly or indirectly, of this information. Sensors transduce information in the form of variations of different forms of energy, such as thermal, mechanical, optical, electrical, magnetic, radioactive, chemical or biochemical, into another form of energy. For example, the information about the kinetic energy in a mobile machine could be revealed by the increase of temperature due to the heat released by applying a disc brake. Nevertheless, as one of the main objectives is the processing of this information, in the same way human senses send the collected information to our brain, we typically are looking for a direct or indirect transduction from any energy form to a final electrical one, for example, if we use the resistance variation with the temperature for monitoring the kinetic energy information in the abovedescribed case. Although there is no totally accepted difference between transducers and sensors, and even both concepts are used as equivalent, the word sensor is mainly used when there is an electrical output signal. On the other hand, detector is reserved for those cases for which non-quantitative measurements are required but only sensitivity to an input signal is obtained. Our present society has become characterized by the significant incorporation of the information technologies. These define the present century as that of the intelligent ambient thanks to the continuous development of sensors. As the needs for physical, chemical and biological recognition systems and transduction platforms grow very fast, sensor technology is continuously pushing up. Sensors are used in applications ranging from environmental monitoring, intelligent buildings and smart ambient; health care and medical diagnostic; industrial manufacturing, transport elements and automotive; defence and security; food control and agriculture and everything that can contribute to make our lives better, easier and safer as well as more sustainable to our world. In such a scenario of increasing demand of sensors and of requested new applications, novel approaches for sensor technology are pursued and for it, nanotechnology is offering one of the most outstanding impacts on the basis of small device sizes – integration, miniaturization and low power consumption; inexpensive – capability for high production volume at low cost; high efficiency – controlled transduction mechanisms at the nanoscale tightening up homogeneity and uniformity; long-term stability – material properties improved and designed at the nanoscale and improved sensor dynamic – minimized time required for sensor response. In fact, the enhancement of the interactions taking place at the nanoscale enables the efficiency of the transduction mechanisms and, hence, the implementation of more advantageous sensors than the conventional ones. Thus, bottom-up nanotechnology approach enables to design and synthesize novel materials with tailored properties adequate for the sensing processes that

5 Nanosensors

83

cannot be imagined from top-down approaches or conventional microtechnologies. Due to these tailored properties, detection limits can be lowered – smaller quantities of samples – increasing the sensitivity. The small size, lightweight and high surface-to-volume ratios, Figs. 5.1–5.3, are the best candidates for improving the capability for transducing chemical and biological species or to have different electrical, magnetic, optical or phonon properties and even quantum effects. Likewise, the better controlled transduction mechanisms may favour the sensor selectivity or, alternatively, improve its sensor performances. It is expected that nanostructured material with modified or functionalized surface may also assist in having greater selectivities. So, for example, biosensing applications are enhanced by specific surface functionalizations, Fig. 5.3. Nevertheless, one of the advantages is to use multiple nanosensor elements as active part of the macrosensors that may compensate for the loss of performance of the individual nanosensor element. Furthermore, the nanotechnological synthesis and processing open the option for tuning material properties and making them more stable by controlling its crystalline defects and surface orientation. Some of the micro/nanofabrication technologies are very mature and widely used, especially those used for the top-down approach, whilst others, more related with the bottom-up approach, are still in their early stages or waiting for tools for interfacing nano with macroworld that are needed for nanomanipulation, nanopatterning, nanolithographic machinery and other procedures driving knowledge towards application of nanodevices and their integration when needed. In this framework, the nanostructured material characteristics become the cornerstone of the possible nanosensor building and, hence, their synthesis methods are the clue for a successful development of nanotechnology.

5.3 Nanomaterial Synthesis for Sensing Materials can be confined at the nanoscale level in 0D, 1D, 2D or 3D defining quantum dots, thin films, nanowires and/or nanoparticles. At the same time, they can adopt multitude of shapes, such as nanorods, nanospheres and nanocubes. These nanomaterials exhibit optical, thermal, mechanical, electrical and surface properties that are strongly dependent on their dimensions. The characteristic wavelength for phonons (101 nm) and visible photons (102 nm) is located at the nanometer scale. At the same time quantum effects are already observed at the 100 nm range and functional semiconductors have Debye lengths at or below the 101 nm scale. All these particularities of the nanoscale offer multiple new sensing processes. Furthermore, nanomaterials are also characterized by huge ratios of surface-to-bulk atoms. Then they could present phonon, photon or electron confinement effects that can be used for sensing processes but they also present interesting surface properties that

84

T. Andreu et al.

involve all the interactions of this unit with the ambient. All of these effects can be enhanced by engineering and acting on their dimensions, shapes and the own nature of these materials that can be metals, metal oxides, semiconductors or magnetic materials [3]. At the present, there is no generally available technology for using only one of these units as sensing elements because there is no easy way for connecting individually one of them or for getting the information facilitated for one of them. Therefore, for their application in sensing devices, these particles suspended in liquid or gas phase require, in general, their organization on substrates or their deposition as thin films. Furthermore, the possibility to add these nanomaterials into more complex assemblies including even organic materials gives rise to alternatives for novel applications, for example, in solar cells [4] or in biosensing [5, 6, 7]. There are numerous techniques and methods for synthesizing nanoparticles, although selection of one of them depends on the particle nature – metal, metal oxide, semiconductor, magnetic, etc.– its functionality and the surface to which they should be attached. The capability of these methods is directly related to its possibility to be applied with tight control of the conditions and parameters during synthesis [8]. The main difficulty is the general trend to easily aggregate or precipitate although stabilizing additives can change the growth, solubility and surface charge in such a way that particles are kept separated and suspended in the liquid. The use of reverse micelles as nanounit reaction is one of the most common methods in colloidal synthesis. Then nanomaterials are obtained via chemical reduction of metal ions or via coprecipitation reactions [9]. All of these techniques and processes offer a broad variety adapted to the final desired properties in the nanomaterials. Thus, it is well known that the magnetic softness/hardness is straightforwardly related to its magnetic exchange length that highlights the significant importance of the nanometric dimensions. As a consequence, there are many reported methods for the magnetic nanoparticles synthesis [10, 11]. Likewise, the synthesis of metallic nanoparticles has received much attention due to their inert nature and catalytic properties [12]. Moreover, their size and quantum confinement modify their light scattering and absorption several orders of magnitude more than other materials. Therefore, there are also a broad range of techniques for their synthesis [13, 14]. In this nanoparticle area, one of the most extended techniques is based on the named sol-gel technique [15] that involves the transition of a system from a liquid ‘‘sol’’ (usually colloidal) into a solid or ‘‘gel’’ phase through hydrolysis, polymerization and condensation steps. Generally, metal or metalloid elements surrounded by reactive ligands are used as precursors and among them metal alkoxides, aluminates, titanates or borates are commonly used in these processes. It is quite compatible with well-known methods for obtaining thin films on substrate such as spin coating, dip coating, spray coating or drop coating [16]. In fact, sol-gel is one of the most employed technologies for sensing applications using thin films

5 Nanosensors

85

based on nanostructured materials [17] as it is advantageous from an economical point of view and at the same time offers easy options to investigate new compounds and nanomaterials [18]. Maybe one of the major inconveniences is the control of the postdeposition processes. Annealing parameters are critical for adequately removing organic additives used to adjust the viscosity or to control the stability of the suspension. Likewise, aside of the layer morphology – nonuniformities due to the cracks or holes formation – it can alter the final crystallization and grain growth as well as the adhesion to the substrate. Just to avoid many of these drawbacks, there are many efforts in other directions for obtaining thin films. So, Langmuir–Blodgett technique [19] gives to the nanomaterial synthesis community an interesting alternative for assembling 1D nanoscale building blocks, allowing final superstructures that depend on the fine-tuning of the compression process [20]. During these last years, its use has been increased together with other innovative techniques based on different variations of the chemical solution deposition [21, 22, 23] processes or just using nanotemplates [24] or even electrodeposition techniques that can or cannot be also combined with nanotemplates. In fact, the use of a replica obtained from the solution filling of a previously synthesized silica nanotemplate allows to easily have 3D distribution of nanostructures after performing the thermal annealing inside the nanotemplate and removing it by means of chemical etching processes [25, 26]. Their characteristics are according to the employed nanotemplate typically based on anodic aluminium oxides, Figs. 5.4 and 5.5, or silica [27], Figs. 5.6 and 5.7. One of them that deserves the attention of the sensing community is the surface faceting of these tiny or very small particles. In many cases, the minimization of the constitution energy involves a quasi-spherical shape without well-faceted surfaces. Then particles’ surfaces present more reactive centres and many of their properties become affected by these boundary conditions [28]: (i) phonon confinement and phonon propagation affecting the thermal conductivity; (ii) free charge scattering affecting electrical mobility; (iii) enhanced photon absorption by the high surface state density; (iv) surface built-in potential

Fig. 5.4 SEM pictures of anodic aluminium oxides used as nanotemplates

86

T. Andreu et al.

Fig. 5.5 SnO2 nanowires obtained by impregnation and calcinations of anodic aluminium oxide nanotemplates

modifying the inner conduction channel at the nanostructure and (v) adsorption states affecting the interaction of the nanostructure with the surrounded ambient. To overcome these possible drawbacks as well as to favour the possibility for individual electrical access, one of the options is the synthesis of the 1D nanostructures. These nanomaterials keep the surface-to-volume ratio quite large and besides they can also be deposited or used applying standard or lowcost nano and microtechnologies maintaining well-defined surface from a crystallographic point of view. They are almost free of crystallographic defect and they are almost perfect monocrystals. In the specialized literature, several names can be found concerning these 1D nanostructures that, in a general

Fig. 5.6 Examples of silica-based nanotemplates showing the regular porous distribution. Reproduced with permission from E. Rossinyol et al., Sensors and Actuators B, 109, 1, 57 (2005). Copyright Elsevier

5 Nanosensors

87

Fig. 5.7 Negative replica of different metal oxides obtained from silica nanotemplates. Reproduced with permission from [24]

way, can be determined by its length L, width W and thickness T. If L is comparable to W or T they receive the name of nanostick [29] or nanorod [30]. If L is much larger than W or T, they receive the name of nanowire [31, 32] or nanobelt [33] according to the ratio W/T be less than or not greater than 5. A very particular case of these 1D nanostructures is that defined by the concept of nanotube [34]. It corresponds to one, SWNT, or several atomic planes, MWNT, that fold on themselves. The most well known are the carbon nanotubes but also many other materials like TiO2, for example [35], have been reported as nanotubes. Varied techniques have been proposed to grow these nanostructures on several substrates, Fig. 5.8. On one side there is the method based on the evaporation–condensation processes. In the pioneering work of Z.L. Wang and coworkers [36], these authors used commercially available nanopowders of metal oxides. They evaporated them using low-vacuum deposition conditions at temperatures much lower than their actual bulk evaporation and condensed them at lower temperature sites along the horizontal oven. Several

88

T. Andreu et al.

Fig. 5.8 Chemical vapour deposition reactor adapted to grow nanowires

initial theories have been proposed for explaining the growth mechanisms. But, up to now, they fail to describe a general model which can design the growth in different situations [37]. A variation of this method is based on the use of a second phase that helps in the condensation process on the surface. In fact, this second phase acts as a catalytic nucleation site [38]. Catalyst species are deposited on the substrate as a space distribution of small nanodroplets that at the growing temperature become liquids, Fig. 5.9. Each one acts as a trap for the growth species. The catalyst must be inert and the capture species are amalgamate or diluted, enhancing its deposition and growing in one direction limited by the wettability of the catalyst elements. One of the most currently used catalysts is gold even though other elements such as Cu, Ga, Fe, etc. have also been applied mainly for metal oxides and typical semiconductors as silicon or gallium arsenide for example, Figs. 5.10 and 5.11. Around these two principles, evaporation–condensation or vapour–liquid–solid, there is a plethora of variations using the well-known physical vapour deposition methods that based on conditions of solid–vapour transformation obtain nanostructured thin films. In this sense, molecular beam epitaxy-based techniques are one of the most promising options for growing nanowires. MBE grows at very-low deposition rate, few Angstrom by second, in ultra-high vacuum, allowing the control of the deposited material as well as the ideal growth conditions of Frank-van der Merwe [39]. Under these conditions, it is effective to force the growth of nanowires using adequate substrate with a catalytic seed distribution. Precisely, the ways used to obtain this distribution of nanocatalytic clusters, the selected substrate and the applied catalytic species define the broad variety of procedures that can be found in the specialized literature. Gold clusters have

5 Nanosensors

89

Fig. 5.9 SEM and HRTEM images of ZnO nanowires grown on SiO2-coated silicon substrates

widely been used for growing many nanowires [40] but many other species have also been proposed [41]. For other materials, like GaAs, self-catalyst of Ga droplets has also been proposed [42]. Likewise, different patterning processes have been proved in spite of the fact that the control of the growth sites [43, 44] and latter nanomanipulation [45, 46] still need much more efforts. Moreover, all of these procedures are also pushing for controlling doping [47] and the formation of lateral or coaxial junctions [48] or heterostructures [49, 50] or even containing radial multiquantum wells [51], Fig. 5.12.

90

T. Andreu et al.

Fig. 5.10 High-resolution transmission microscopy images of a SnO2 nanowire showing the presence of some dislocations. Reproduced with permission from J. Arbiol et al., J. Crystal Growth, 310, 1, 253 (2008). Copyright Elsevier

A totally different approach for obtaining nanowires from some material that can be had as foils or wafers is based on nanoelectrochemistry [52]. In this case, the doping is that previously existing in the starting substrate, which is an interesting advantage28. Essentially, for the case of silicon nanowires, the process is based on the electroless metal deposition [53] based on a FH solution that gives rise simultaneously to anodic and cathodic processes on the substrate surface [54].

Fig. 5.11 GaAs nanowires obtained on SiO2-coated (111)B GaAs wafers. Reproduced with permission from [42]

5 Nanosensors

91

BFSTEM

BF STEM

Fig. 5.12 (a) BFTEM image and (b) magnified detail of a GaAs NW with coaxial QWs [51]. (c) HRTEM image of an axial modulated nanowire showing a region with heterostructures [41]

5.4 Nanosensing Mechanism Features After all of these plethora of techniques, methods and procedures, finally, it is the control of the nanocrystals properties and performances that becomes more outstanding. Crystallographic structure and defects, impurities concentrations and distributions and surface characteristics determine (i) the phonon propagations and confinement, (ii) the charge transport and charge scattering mechanisms and (iii) the absorption and recombination processes as well as the optical confinement, Stokes shift and optical resonance. On the other hand, the surface properties and its functionalization reveal the solid–ambient interactions that define the chemical and biochemical sensing processes. As a consequence, the different sensing processes at nanoscale and, hence, the types of nanosensors can be classified using many different criteria. Thus, on a very general basis, it is standard to classify the nanosensors as (i) chemical, (ii) biochemical and (iii) physical sensors. These last ones include the optical, thermal, electrical, magnetic and mechanical sensors. However, here, we will focus on the more significant features taking place at the nanoscale mechanisms. Here, among many other possibilities and casuistry, we are going to consider only four nanosensor groups that concern mainly the more outstanding transduction mechanisms that nowadays have already developed significant nanosensor applications: a) Surface characteristics. Chemical and biochemical nanosensors are mainly based on this type. Special attention must be paid to the functionalized surface-based sensors. In this group the charge transfer processes from the outer molecules towards the inner part of the solid nanomaterial through the surface is the most outstanding feature of these transduction mechanisms. It gives rise to nanosensors based on electrical interaction through the surface. b) Photons. The photon interaction with the nanomaterial determines the optical-based nanosensors (photon absorptions, Stokes shifts, resonances,

92

T. Andreu et al.

luminescence pathways and so on). It gives rise to the development of nanosensors based on photon capture that are available for the photodetection at the nanoscale. c) Plasmon resonances. Small variation of the dielectric constant defined at the nanoscale domain can easily modify the surface plasmon resonance. It gives rise to the development of nanosensors based on plasmon resonance that reveal the influence of the dielectric constant variations at the nanoscale. d) Mechanical properties. The progressive scaling down of the mechanical structures allows having new significant relation among the different physical forces. It gives rise to the development of nanosensors based on mechanical resonances which present an important influence on its resonance values of the small mass changes taking place on these mechanical nanostructures. Due to the small dimensions of the nanostructures, the models deduced for bulk cases are not more applicable, whilst the analyses of the sensing mechanism features require the application of new simulations and models of these nanostructures. So, a broad variety of methods starting from the classical calculations up to the most advanced quantum mechanical calculations have been developed during the last years [55, 56]. Properties such as energy of the structure, work function changes, binding energies as well as chemisorption and physisorption properties [57] of the surface can be calculated on the basis of the DFT (density functional theory) and wave methods. In spite of the huge effort performed in this area, there is still an important task to be done. Without this theoretical support, many of the experimental activities carried out are lacking complete scientific analysis. The knowledge advancement is fortunately very positive but it is not going fast enough.

5.5 Nanosensors Based on Electrical Interaction Through the Surface It is well known that metal oxide materials due to their surface electronic structure present very interesting properties that give rise to interactions with the gas molecules, tin oxides being one of the most representative of these materials [58] although at the beginning ZnO was the metal oxide used for building the first solid state-based gas sensor. Many years ago, it was already pointed out that attention on tiny crystals of metal oxide semiconductors shows outstanding changes of their electric resistance upon exposure to reducing or oxidizing gases [59, 60, 61], Fig. 5.13. As found with pure SnO2 devices, electric resistances in air as well as those under exposure to H2 diluted in air began to increase sharply when the size of component SnO2 grains decreased beyond a critical point (about 6 nm in diameter). In addition, sensor response to H2 increased remarkably with decreasing size. A similar size effect was also observed with the response of other metal oxide device sensors to oxidizing gases [62, 63, 64]. It was suggested that extension of

5 Nanosensors

93

Fig. 5.13 Detail of the intergrain connectivity between nanocrystals and its electrical equivalent model considering grain, intergrain and electrode resistances

space charge layer to a whole region of component crystals could be responsible for such effects, though the precise mechanisms involved were left open. They also appear to be related to the surface characteristics [65]. Typically, porously sintered assemblies of these tiny crystals as thick or thin film layer [66] are used, Fig. 5.14. It is working as a 3D network of interconnected tiny crystals presenting the above-described electrical properties and hence contributing to the total 3D network equivalent resistance changes when the ambient composition is varied. This property has provided a base to develop semiconductor gas sensors

Fig. 5.14 Cross-section of a sensitive metal oxide layer based on nanocrystals

94

T. Andreu et al.

for use in various fields. Despite marvellous development, this type of gas sensors still has some basic problems on gas response left unsolved or not well understood like porosity influence, crystal grain boundary interconnections, surface crystal orientations and shapes and sizes of the component crystals. What will happen when the size minimized drastically up to the geometrical sizes are comparable with the depletion zone dimension, the free scattering-less length or the screening length? What is representative of a solid state gas-sensing material is its capability for absorbing gas molecules according to their oxidizing or reducing character onto surface sites identified by their acidic or basic properties. So, for example, gases like O2 and NO2 are very well known to be adsorbed on oxide semiconductors to form anionic adsorbates. The most standard one is O2 that is adsorbed as follows, see Fig. 5.15: O2 þ 2e ¼ 2O :

(5:1)

The equilibrium of this chemical process is expressed by ðKO2 PO2 Þ1=2 ½e ¼ ½O :

(5:2)

Here KO2 and PO2 are, respectively, the adsorption constant and the partial pressure of oxygen, and [e] and [O–] are surface densities of free electrons and O– adsorbates, respectively. One must observe that the former quantity determines resistance of the sensor device so that solving how it is related with PO2 is a central subject in semiconductor gas sensors. Although [e] as well as [O–] are surface quantities, these are deeply associated with the bulk of the semiconductor. In fact, this equation points out the adsorption equilibrium on a semiconductor and, as it is shown, it cannot be solved without considering the electronic equilibrium of the semiconductor. Therefore, the electronic equilibrium

Fig. 5.15 Schematic of the surface interactions taking place on metal oxides

5 Nanosensors

95

depends not only on gas adsorption strength but also on shape and size of component crystals that determine the charge neutrality into the crystal. Therefore, the general aspects of the electronic equilibrium become straightforwardly connected with the chemical equilibrium and its description and analyses are harder as the crystal dimensions scale down to the nanoworld, and the application that was useful for larger crystals becomes no more acceptable. It is known that, if there are traps of electrons on the surface of an n-type semiconductor, conduction electrons are transferred from the subsurface to the traps, leaving an electron-depleted layer behind. This transfer continues until an electronic equilibrium is reached throughout all regions from surface to bulk. The simpler scheme of electron transfer for a plane surface is drawn in Fig. 5.16. Conventionally it is assumed that donors are ionized completely and that all of the conduction electrons up to depth W are completely transferred to the surface (abrupt model). Under the latter assumption, density of surface charges, Qsc, is nominally equal to –qNdW, where q is elementary charge of electron and Nd is the density of donors in semiconductor. Under these conditions, electric potential, V, in the depletion region should satisfy the following Poisson’s equation, where x is the depth from the surface and " is the permittivity of semiconductor: d2 V=dx2 ¼ qNd =":

(5:3)

By introducing the boundary conditions that dV/dx as well as V are zero at x = W, the above equation is solved to give the following depth profile of electric potential: VðxÞ ¼ ðqNd =2"Þðx  WÞ2 :

Fig. 5.16 Charge space zone scheme at the surface due to the interaction with the ambient

(5:4)

96

T. Andreu et al.

Potential energy of electron, qV(x), as well as surface potential barrier height, qVs, are directly obtained from the above equations: qV=kT ¼ ðq2 Nd =2"kTÞðx  WÞ2 ¼ ð1=2Þfðx  WÞ=LD g2 ;

(5:5)

qVs=kT ¼ ð1=2ÞðW=LD Þ2 :

(5:6)

Here we have introduced an important quantity called Debye length, LD, defined by LD = (q2Nd/"kT)–1/2 for simplifying the expressions. The surface chemical equilibrium can be solved now because the quantities involved have been made explicit with the properties of the underlying semiconductor. ½e ¼ Nd expðqVs =kTÞ;

(5:7)

½O  ¼ Qsc =q ¼ Nd W:

(5:8)

Using these relations, chemical equilibrium can be rewritten as ðKO2 PO2 Þ1=2 =LD ¼ ðW=LD Þ expððW=LD Þ2 =2Þ;

(5:9)

and then resistance for the sensor device, R, can be estimated from the [e] values R  expðqVs =kTÞ

or

R  expððW=LD Þ2 =2Þ:

(5:10)

In the presence of a reducing gas such as H2 under fixed PO2, [O–] is modulated by its surface reactions. Even in this case, once [O–] is properly estimated as a function of PH2, then one can estimate the correlation between R and PH2. It should be remarked that the performed hypothesis for this calculation is only a coarse approximation. Actually, the used abrupt model is not necessarily rationalized. It ignores that conduction electrons are never perfectly absent in the depletion region but actually present rather abundantly in the vicinity of the border to the inner intact region. The phenomenon is called a distribution tail. This tailing eventually reduces Qsc from that estimated above and also it overestimates W and the bending potential qVs. It becomes important as we are dealing with very tiny crystals for which the accurate dimensions are very significant. In this case, for example, there is no possibility for a non-limited PO2 increase because there is no room to extend W. When the absorption of gas molecules starts, electrons are depleted from surfaces – nanocrystal shapes are essential for a right description; it is not equivalent spherical, columnar or plate like shapes – leaving the inner region intact, but subsequently the whole area including the central part is deprived of electrons thereafter achieving the condition of deep depletion shrinking the inner neutral part.

5 Nanosensors

97

In spite of the existing classic models [67, 68], an accurate and completed theory which describes resistance and response of a semiconductor gas sensor using very tiny component crystals has still to be constructed. It must involve the combination of the chemical equilibrium on the surface of semiconductor with the electronic equilibrium inside. In fact, nanomaterials are one of the most promising alternatives for gas sensing. They represent the natural bridge between single molecules and macroscopic bulk materials. Their finite size and limited number of electronic states create a lot of restrictions for the transport mechanisms, specially for the charge conduction as the number of surface states can be comparable with that of the bulk. Then, their electrical behaviour is expected to present novel electrical and optical properties which can be used to fabricate nanodevices with improved capabilities [69]. This is the case of small nanowires where the area–volume ratio is easily maximized as the diameter size is decreased [70]. In this scenario, for small crystals, there are three paramount important findings that must be underlined: (i) how will the catalytic additive affect at the nanoscale level [71, 72, 73]? (ii) how much probable is it to achieve deep depletion conditions in the nanocrystal as partial pressure of adsorbing gas increases? and (iii) how much nanocrystals are needed to have a reliable sensing platform based on the surface interactions? It is this last question that has arisen many requests for using nanotechnologies. In fact, in spite of the complexity of the mathematics models to find the equivalent impedance between two terminals to a 3D distribution of impedances, it is clear that the final value is proportional to that of a single element [74]. It tells us that we will have similar information just using only one of these nanostructures instead of an entangled bunch of them. In this case, no grain boundaries exist and, hence, the electrical transduction effects induced by the adsorbed gas molecules onto the surface of these nanowires can straightforwardly be revealed by the electrical magnitudes of these single nanocrystals. Several attempts for using only one of a few grains of metal oxide nanocrystals were already done by J. Tamaki et al. [75, 76, 77] using different gaps open in a gold electrode. Nevertheless, the lack of repetitiveness is still a challenge and the quality of contacts of gold/particle and among the particle seems not to be enough. On the contrary, this situation has moved the efforts from sphericallike nanocrystals towards the use of nanowires, nanotubes or structures based on them. Unlike the small quasi-spherical nanoparticles with non-faceted surfaces, these nanostructured materials present well-defined crystallographic directions as crystal surfaces as well as looking like a perfect monocrystal. No grain boundaries are present. Therefore, the electrical transduction effects induced by the adsorbed gas molecules onto the surface of these nanostructures can straightforwardly be revealed by the electrical performances of these single nanostructures [78]. Moreover, as the gas molecules adsorption takes place at the surface, one of the most important issues for tailoring the sensor material response is the control of its active surface area. Many authors [59, 60, 61, 79]

98

T. Andreu et al.

have shown that there is a direct relationship between the active surface areas presented by the sensing material and the response given by the macroscopic sensor fabricated with these nanomaterials. According to these results, an increase of the surface area should involve an increase of the sensor response. This is the case of nanowires and 3D nanocrystal networks where the area–volume ratio is easily maximized as the diameter size is decreased, although as is discussed below it is needed to reduce a lot of the nanowire diameter to have high surface/volume ratio like it happens for the nanograins. The majority of the experimental works show gas sensor based on multiple nanostructures or entangled bunch of nanowires, Fig. 5.17. A large number of these nanostructures are contacted and their electrical parameters estimated. However, only some equivalent mean values of these parameters are determined due to the dispersion existing among the contacted nanostructures and the grain boundaries among them. In this scenario it is not straightforward to study the sensing mechanisms onto a single metal oxide nanocrystal, Fig. 5.18. To study gas-sensing mechanism onto one individual nanocrystal is still a challenge but it is required to reach a better knowledge of the electrical transport mechanisms which take place in these nanostructures. One of the reasons to explain this lack of experimental studies is that there are many difficulties in performing reliable electrical contacts on one individual nanostructure in a controlled fabrication process at the nanoscale level. The most common techniques are optical and electron beam lithographies. Nevertheless, both techniques are multistep and time-consuming processes [80, 81], and they are difficult to tailor for each specific sample. For all these reasons, a complementary method based on FIB technique to the electron-based lithography process has been proposed as one option which could, experimentally, help to solve problems related to more conventional contacting processes.

Fig. 5.17 Electrical scheme for entangled bunch of nanowires compared with that corresponding to a nanocrystal network

5 Nanosensors

99

Fig. 5.18 Simplified electrical equivalent circuit to a resistance distribution corresponding to a nanostructure network

Focused ion beam (FIB) is a powerful technique developed during the late 1970s and the early 1980s for the patterning and, later, for the deposition of materials, with resolution in the tens of nanometer range, and commonly used in circuit edit, mask repair, microsystem technology processes and material characterization [82, 83, 84]. The basic principle of this technique is a focused ion beam of highly energetic particles that scans the sample’s surface and sputters the material of the exposed area. The scanning can be performed, similarly to a SEM, using electrostatic lenses and, thus, the milling occurs without the need of masks. At the present time, gallium (Ga+) ions are the most used particles in FIB technique due to the fact that gallium is a metallic element with a low melting temperature which allows the fabrication of long lifetime, high brightness and reliable liquid metal ion source (LMIS) required in FIB technique. Moreover, the element gallium is positioned in the centre of the periodic table (element number 31), so its momentum transfer capability is optimal for a wide variety of materials. On the contrary, lighter elements would be less efficient in milling heavier elements [82]. On the other hand, if a metal organic compound is introduced in the beam path with the help of a so-called gas delivery system by using a microneedle, decomposition occurs due to interaction of the compound with both secondary electrons and ions originated during the Ga+ ion bombardment. Part of the compound can be deposited on the sample’s surface (ion-assisted deposition) or can reactively assist the milling process (gas-assisted etching), while the rest is removed by the vacuum system. In this way, conductive and isolating materials can easily be deposited with FIB with nanometer precision. Although the purity of the deposition is generally lower as compared to conventional deposition techniques like CVD or PVD owing to contamination originated during the metal organic decomposition, the main advantage of this technique is its

100

T. Andreu et al.

flexibility and masks are not required. However, damage introduced in these materials by ion bombardment (Gaþ ions accelerated to 30 kV, in our case) during fabrication of contacts has limited the use of this technique. Alternatively, it is possible to fabricate electrical contacts by using low ion currents in the proximity of the nanomaterials in order to reduce the damage. But, ion exposure is not thoroughly eliminated, and contacted nanomaterials are still modified [85]. At the present, the appearance of the so-called dual beam systems, a traditional FIB which incorporates a scanning electron microscope (SEM), has facilitated the use of FIB thanks to the possibility of acquiring electron images in situ, avoiding the damage caused by ions when imaging. Moreover, metal organic compounds can be also dissociated with the help of secondary electrons (SE), giving rise to electron beam-induced deposition (EBID) [86], electrons accelerated to 5 keV and methylcyclopentadienyl Pt (IV) trimethyl (PtC6H16) as gas precursor. Nevertheless, the number of SE produced by the primary electrons is smaller than that produced by Gaþ ions, so the deposition rate is much lower in an electron beam-induced deposition [82]. Due to the fact that interaction between electrons and the sample is less destructive, performing an electron-assisted deposition on the nanostructure to be contacted and finishing the rest of the contact with the help of ions can avoid undesired surface damage and structure modification of the nanostructure. According to this procedure based on the combination of both electron- and ion-assisted deposition, metal oxide nanowires (NWs) and 3D nanocrystal networks can be contacted either in 2- or 4-probe configuration. The attempts of contacting nanomaterials using this method are still scarce and further research must be done in order to have a complete knowledge of the overall characteristics and performances of the electrical contacts [87, 88, 89]. The focused ion beam machine can be used to contact single nanostructures deposited previously and nano/micro manipulated onto a silicon wafer already processed to have an adequate distribution of microelectrodes. The manipulation defines a better placement of the nanostructure in front of the microelectrodes before proceeding with the Pt deposition. Then, four or two electrondeposited Pt contacts are performed, two or one, respectively, in each extreme. In this way, damage is strongly diminished and Gaþ ions effects on the nanostructure are avoided. In fact, a couple of contacts are fabricated near each extreme of the nanostructure requiring much less than a half micron. The remaining nanostructure, of a significant length (more than several couples of microns), is released of any feature relative to the contact fabrication such as is shown in Figs. 5.19–5.21. Once the influence area due to the deposition is far away from the nanostructure, ion deposition methodology is used to extend the stripes to the pre-patterned contacts, Fig. 5.20, because the deposition time is much shorter than using the electron one. The electron and the Gaþ ion beams are accelerated to 5 and 30 kV, respectively.

5 Nanosensors

101

Fig. 5.19 Details of the four electrodes array patterned on a silicon wafer. Reproduced with permission from [93]

Using this procedure, the previously presented nanostructures have been contacted in 2- and 4-probe configurations. This contact fabrication process for each nanometer-sized material takes more or less 2 h, limiting the application of this lithography technique in large-scale processes. On the other hand, its flexibility guarantees that it can be used for rapid prototyping, helping to solve some of the multistep-related problems of other nanolithography techniques. Two- and four-probe V–I electrical measurements have been performed using a Keithley Source Measure Unit (SMU) 2400. These measurements at

Fig. 5.20 (a) Schematic of the focused ion beam process for obtaining four contacts using electron (beside the nanowire) and ion (rest of the strips) beams. (b) Image of a nanowire with four electrical connections. Reproduced with permission from [93] and F. HernandezRamirez et al., Nanotechnology 17, 1134, 2006. Copyright IOP Institute of Physics (2006)

102

T. Andreu et al.

Fig. 5.21 Magnification from microscale to the nanoscale level for a connected individual nanowire. Reproduced with permission from F. Hernandez-Ramirez et al., Nanotechnology 17, 1134, 2006. Copyright IOP Institute of Physics (2006)

room temperature reveal that contact resistance contribution in 2-probe measurements is much more important than the nanowire resistance, Fig. 5.22. In some case, such contribution is even higher than 90% of total measured resistance, which range between some hundreds of kiloohms and some tens of megaohms. This huge difference between 2- and 4-probe measurements cannot be explained only by the sum of FIB-fabricated platinum stripes resistance and the contact resistance between them and pre-patterned gold microelectrodes since according to FIB platinum resistivity, lower resistance values would be expected (a few tens of kiloohms). So, the main part of the measured contact resistance is believed to be originated in the Pt–SnO2 NW junction. No degradation in this kind of contacts has been noticed after a long time of their fabrication, more than 4 weeks, and neither after applying constant current for several minutes. For example, good stability of the contacts performed on a nanowire has been observed after 1 month and no degradation after a current of 700 nA has been injected during 20 min. All measured nanowires are destroyed at current densities below 1010 A/m2, which is the critical current density value obtained for both ion- and electronassisted deposition. As has been said previously, due to the influence of the contact resistance to the final result, 4-probe electrical measurements are required if electrical parameters of nanostructure must be extracted. In order to determine the importance of this contact resistance contribution and as an example, the SnO2 nanowire shown in Fig. 5.21 has been measured in both 2- and 4-probe electrical measurements. A resistance of R = 12.53 M has been obtained under the first setup, while a resistance of R = 1.93 M , see Fig. 5.22, has been obtained in the second case. These results point out clearly how the contribution of the contact impedance distorts the measurement of the nanostructure and even, sometimes, can be higher than the resistance of the nanostructure itself. The extraction of the electrical parameters of the contacted nanowires is performed applying a simple model to estimate the material resistivity.

5 Nanosensors

103

Voltage (V)

4 3

2-probe

2 1

4-probe

0 100

200

300

Current (nA) Fig. 5.22 Electrical measurements of a single nanowire using 2 or 4 probes on a silicon substrate die mounted on a TO packaging. Reproduced with permission from F. Hernandez-Ramirez et al., Nanotechnology 17, 1134, 2006. Copyright IOP Institute of Physics (2006)

Considering that these nanowires have a cylindrical cross-section, the resistivity can be calculated with the help of Eq. (5.11). R ¼ L=A ¼ L=pr2 ;

(5:11)

where R is the measured resistance,  is the unknown resistivity, L is the length of the nanowire between the fabricated contacts and A is its crosssection, which can be determined if the radius, r, of the nanowire is measured. Once a nanowire has been contacted, the dimensions L and r are measured in the SEM using XP-Annotate# software, which is available in the FIB workstation. This software compensates for the error introduced in the measurements when the sample is tilted. In order to test the accuracy of this software, morphological characterization has been performed with an AFM working in tapping mode. It has been checked how platinum stripes have a roughness between 2 and 5 nm and their dimensions differ by less than 10% of the ones obtained with XP-Annotate#. So, the use of this software has been considered a reliable and fast method to determine dimensions (L and r) of the nanowires. Once the dimensions and the resistance of the nanowires are known, the resistivity can be easily determined by applying Eq. (5.11). In the case of the SnO2 nanowire of Fig. 5.21, a resistivity of  = 50 m cm has been found. The found value agrees with previous values of the literature related to NW studies of single SnO2 crystals [90]. These contacts show excellent electrical stability as a function of time and applied current density. Therefore, they fulfil the requirements to fabricate nanodevices able to work for a long time. However, there is still an important feature to be discussed. It is the linearity of the electrical contact. These ones are

104

T. Andreu et al.

Fig. 5.23 Intensity versus applied voltage in a nanowire measured using 2-probe contact. Reproduced with permission from F. Hernandez-Ramirez et al., Nanotechnology 17, 1134, 2006. Copyright IOP Institute of Physics (2006)

Current (nA)

50 25 0 –25 –50 –2

–1

0

1

2

Voltage (V)

symmetric but reveal a non-ohmic behaviour such as is shown in Fig. 5.23. This rectifying behaviour originates at the Pt–SnO2 NW junctions. This junction forms a Schottky barrier of height B = 0.75  0.10 eV in the absence of interface states owing to the differences between the work function of Pt (Pt = 5.65 eV) and the electron affinity of SnO2 (SnO2 = 4.9  0.1 eV). This configuration can be described as a back-to-back Schottky-like circuit. According to this assumption, the applied bias V must distribute as V ¼ Va þ Vd þ VNW þ Vi ;

(5:12)

where Va is the voltage drop in the cables, gold microelectrodes and FIBfabricated platinum stripes, Vd is the voltage drop in the direct-biased Pt–SnO2 junction, VNW is the voltage drop along the SnO2 NW and Vi is the voltage drop in the reverse-biased junction. In this back-to-back Schottky-like configuration, the total current is limited by this reverse-biased junction, insert in figure A. The contact barrier height is reduced and the current increases with increasing bias according to   qBE I ¼ AA T2 exp ; kB T

(5:13)

rffiffiffiffiffiffiffiffiffiffiffiffi rffiffiffiffiffiffiffiffiffi qE q NS  d ¼ B0   ; 4p"s "S 4p

(5:14)

where BE and sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi  ffi 2qND kB T E¼ V þ bi  ; "s q

(5:15)

5 Nanosensors

105

where A is the contact area, A** is the effective Richardson constant, BE is the effective barrier height, Bo is the ideal barrier height in the absence of image force, E is the maximum electric field at the junction, "s and ND are the dielectric constant and doping concentration of SnO2, Ns is the surface state density and bi is the built-in potential. If both the voltage drop in the direct-biased junction Vd and the voltage drop in the nanowire VNW are supposed to be small compared with the voltage drop in the reverse-biased junction Vi, then, it can be considered, as first approximation, that all the applied voltage falls in this latter junction, Vd þ VNW  Vi ) V Vi :

(5:16)

According to this assumption, ln(I) versus V1/4 should be linear at all temperatures for which the model is valid. Rectifying I–V curve values confirm that the back-to-back Schottky-like assumption can be very useful to describe the electrical response of these SnO2 NWs at room temperature, Fig. 5.24. The contact resistance presents an activation energy at room temperature, which is found depending on the nanowire, but over 200–2508C its value becomes, generally, already non-predominant. It is important if we keep in mind the gas sensor working temperatures. This behaviour can be explained assuming that some FIB-induced disorder is produced underneath the Pt contacts during the contact fabrication process, creating surface states used by electrons to overcome the barrier, even at low bias, considering multiphononassisted tunnel mechanisms. Another consequence of this contact resistance that must be taken into account as relevant in many experimental cases is the heat dissipation. So, to avoid also the DC thermal effects due to the power dissipation, as well as to avoid any problem with the species migration through the structures, AC measurements can also be applied. AC impedance spectroscopy is a well-known 100 Negative voltages Positive voltages

Current (nA)

Model line

Fig. 5.24 Logarithmic plot of the intensity versus V1/4. Reproduced with permission from F. Hernandez-Ramirez et al., Nanotechnology 17, 1134, 2006. Copyright IOP Institute of Physics (2006)

10

1 0.8

0.9

1.0 Voltage (V1/4)

1.1

1.2

106

T. Andreu et al.

and widely used technique in material science studies that can help us to overcome these limitations. Moreover, AC impedance spectroscopy is known to be useful to characterize Schottky barriers originated in Pt–SnO2 junctions. For the AC measurements, an impedance analyser Solartron SI-1260 with a maximum working frequency of f = 5 MHz and an impedance analyser Gamry EIS300 with a maximum working frequency of f = 300 kHz can be used inside of a Faraday cage. As far as we know, these were the first attempts of performing 4-probe electrical measurement using AC impedance spectroscopy on a single nanowire [91]. For it, gas sensor nanodevices were fabricated following a bottom-up process based on the use of an individual SnO2 nanowire based on DC [92] or on AC [93] measurements. These nanowires were grown by variations of the methods based on vapour–liquid–solid (VLS) mechanisms in evaporation/condensation or chemical vapour deposition processes that have been described elsewhere [94, 95, 96, 97]. In all the cases, under different growth conditions, it has been possible to have very excellent single nanocrystals as nanowires which present different crystallographic orientation and faceting, Fig. 5.25. Electrical values reveal directly the consequences of the electrical transduction of the gas–solid interaction taking place at the nanowire surface. It can be seen in Fig. 5.26 for a nanowire of 50 nm diameter measured at 2658C in two atmospheres, nitrogen and dry synthetic air (80% N2, 20% O2). On the basis of these procedures, systematic measurements of the nanowire resistance in the bottom-up gas sensor nanodevices at different temperatures and atmospheres have been performed. One of the most interesting results using metal oxide nanowires can be attained in the analysis of its resistance against the synthetic air and nitrogen atmospheres. It reveals a clear dependence on the diameter size according to (1/ D) dependence, Fig. 5.27. It gives support to the model that assumes the existence of a space charge zone created by the adsorbed oxygen which is limiting the central area of the nanowire for electrical transportation. So, adsorbed oxygen as well as the hydroxyl group modify the surface conditions for other gases and they must be considered before studying any gas-sensing process. Among the typical test gases used to evaluate sensor performances, carbon monoxide is the more studied and it presents, at the same time, simple transduction mechanisms. It is accepted that in the temperature range between 200 and 3708C, there is CO2 formation from the adsorbed oxygen [98]. Ideally, it follows a well-reported mechanism. Therefore, CO is one of the most reported gases to evaluate the performances of the nanosensors based on single nanowires. Again, the main variation in the performed measures has been found related to the values of the diameter of the nanowire. It constitutes one of the most critical points, such as it is well accepted for nanograins. The possibility of getting nanosensors with high performances depends on it such as is shown by Fig. 5.28. It corroborates the above-shown trends for the atmosphere change from nitrogen to synthetic air. So, a small amount of adsorbed molecules can easily

5 Nanosensors

107

Fig. 5.25 SnO2 nanowires

bend the bands in the radial or coaxial direction. It limits the central zone of the nanowire available for lateral transport and, as a consequence, the sample resistance is then a measure of the amount and type of gas. Obviously, it becomes significant as the transversal radius of this central zone is comparable to the dimensions of the surface space zone. Many references from the literature have reported about CO measures using individual nanowires. However, the reported sensor responses are, as it was discussed above, not important due to the large radius of the electrically contacted nanowires used in these references. In spite of the difficulties in their manipulation and contacts fabrication, values around 25 nm and below are needed to achieve higher sensor response than those reported for macroscopic gas sensor based on interconnected particles of only a few nanometers of diameter [99]. Nevertheless, it should be noted that

108

AS N2 AS Z img (Mohm)

Fig. 5.26 AC electrical measurements of a single SnO2 nanowire under different gas atmospheres. Reproduced with permission from [93]

T. Andreu et al.

1.2

0.6

0.0 0

1

2

3

Z real (Mohm)

using an individual nanowire as a single nanocrystal has more advantages from the point of view of stability, repetitivity and reliability as there are no grain boundaries. Moreover, there are well-defined crystallographic directions for the surfaces of the nanowire unlike the case of very small nanoparticles that usually present very reactive surface because they are not yet faceted surfaces [100]. Therefore, a single nanowire presents the characteristics of a monocrystal that fit much better with the stability requirements and long-term wearout claims. So, maintaining small radius in the selected nanowires, the surface variation has a stronger influence on the nanowire resistance. Nevertheless, in spite of these interesting results, more detailed studies are still needed to achieve a complete mechanism views and to have a complete modelling of it, especially if it is possible to module the working temperature. Just this interest has focused the efforts to achieve the combination of individual nanowires and micromembranes with integrated heaters.

Fig. 5.27 Nanowire section size dependence of the electrical response from nitrogen to synthetic air atmosphere

5 Nanosensors

109

Fig. 5.28 Sensor response for different sensing materials against the CO concentration (Commercial = nanocrystal of 10 nm, NW literature 200 nm)

It is one of the most promising approaches for obtaining portable devices for gas sensing since their working temperature can effectively and easily be modulated. To the best of our knowledge, only a few attempts to use this fabrication methodology have been reported in the literature [101, 102]. On the other hand, some challenging issues like the electrical stability of the final devices based on single nanowires should also be solved prior to their large-scale application. Recently, portable gas sensor based on an individual SnO2 nanowire has been reported [92], Fig. 5.29. In this example, individual SnO2 nanowires synthesized by chemical vapour deposition of a molecular precursor [Sn(OtBu)4] were transferred onto suspended micromembranes with one integrated SnO2:Sb heater and platinum interdigitated microelectrodes, Fig. 5.29a. Membranes were fabricated using silicon-on-insulator wafers. High-resolution TEM images showed single crystalline nanowires with dislocation-free bodies and a main growth direction of [100] and interplanar spacing in correspondence with the rutile structure of SnO2 [103]. Nanowires were electrically contacted to platinum microelectrodes using a FEI Dual Beam Strata 235 FIB instrument equipped with a trimethyl methylcyclopentadienylplatinum ((CH3)3CH3C5H4Pt) injector to deposit platinum, following a nanolithography process, Fig. 5.29b. Two-probe DC measurements were performed with the help of an electronic circuit, Fig. 5.29c, designed to guarantee low current levels and avoid any undesired fluctuation. This option reports the possibility of modulating the device temperature using the integrated heater. Although the temperature of these membranes was calibrated as a function of the power dissipated by the heater, accurate estimation of the effective temperature of nanowires is quite difficult since uncontrolled self-heating effects due to the existence of Schottky barriers and their conduction of charge carriers should always be considered [104]. The dynamic response of these devices was fast enough to reach complete thermal stabilization after a few seconds of changing the power applied to the heater,

110

T. Andreu et al.

Fig. 5.29 Details of a gas-integrated nanosensor based on a single SnO2 nanowire. Reproduced with permission from [92]

demonstrating that the optimal working conditions can be easily modulated in a fast and controlled way. It facilitates its use as devices for monitoring this gas such as is shown by Fig. 5.30. This last result demonstrates that these devices could be used in the future as portable and reliable gas microsensors. The sensing performances of these nanostructures can be enhanced by means of the adequate modification of the surface. It is well known that the presence of catalytic nanoclusters of different noble metals, such as gold and platinum, or inorganic compounds, such as CuO and CaO, has been widely reported to improve reactivity with some molecules. However, the modification of the surface wins a maximum significance when we are looking for the interaction with biomolecules. In this case, the interactions can cover a wide range from very strong covalent bonds to weak van der Waals interactions, Fig. 5.31.

5 Nanosensors

111

Fig. 5.30 Sensor response against CO of the integrated nanosensor shown in Fig. 29. Reproduced with permission from [92]

Many organic compounds possess different functional groups such as amines (H2N-R’), thiols (HS-R’), alcohol (OH-R’), phenol (OH-cycle-R’) and carboxylic acids which are organic molecules that contain a carboxyl group (-C=O)- whose carbon atom is bonded to a hydroxyl group (-OH) as it happens, for example, in the methanoic acid, ethanoic acid or propanoic acid (R-COOH). So, these functional groups can be used to form covalent bonds on the target organic molecules where atoms forming this bond share pairs of electrons. Sometimes, this covalent coupling can be done directly on the surface of the sensing material, whereas in many other cases, other molecules named cross-links are needed to link one molecule to another. Likewise,

Metal oxides or Metal thin

Fig. 5.31 Schematic of a typical functionalized surface

Substrate

112

T. Andreu et al.

very often, the functional groups need to be energetically activated avoiding the deterioration of these biomolecules. In general, there are many functional groups but carboxylic acids (R-COOH) are frequently used. They can be covalently coupled with other functional groups to give rise to different bonds: (i) amides, R-CO-NH-R’, (ii) thio ester R-CO-S-R’, (iii) acyl ester R-CO-O-R’ and (iv) aryl ester R-CO-O-cycle-R’. All of them allow fixing functional groups but often there is the problem about how to be fixed on the substrate that is typically based on silicon wafer (silicon dioxide films) or alternatively on metal layer or metal oxide materials. For it, self-assembling monolayers formation has become one of the most applied methodologies. Molecules with hydrophobic–hydrophilic nature, also named amphiphilic molecules, are used. These molecules consist of two different head and tail functional groups where one end sticks to the sensor surface and the other end interacts with the analyte biomolecules establishing a strong link. The alternative layer by layer adsorption of different materials with charged or functional groups [105, 106] allows the formation of ultra-thin layers (several nanometers) with a high control of their composition and structure [107]. Gold is a suitable substrate for forming SAMs since freshly deposited gold presents a high hydrophilic degree which facilitates the interaction with a thiol group. Alkane thiols, HS-CnH2n+1, are among the most used thiols for sensing applications due to the easy formation of a thiolate bond with the gold surface from an alcohol solution and the ordered alkane tail packaging because of the existing van der Waals among them. Other materials such as silver, platinum, copper can replace gold and alternatively silicon dioxide and many metal oxides can also be applied due to the oxygen bridged and hydroxyl groups on their surface that gives a strong hydrophilic character. Moreover, it also gives the possibility for using other compounds, silanol for example, as intermediate links between the substrate surface and the target analytes. On the other hand, other materials such as carbon present a strong hydrophobic character. These surface modification protocols are usually applied to functionalize the nanoparticles or substrate surfaces in order to develop sensitive and highly stable biosensors. Furthermore, during the last years, this capability has been enhanced due to the use of the surface plasmon resonance effects, SPR, which supplies a powerful tool for the sensing procedure [108], mass sensors based on quartz crystal microbalance (QCM) or surface acoustic wave (SAW) devices. Likewise, these surface modification procedures have been enriched by incorporating antibodies, enzymes, nucleotides or DNA sequences [109]. Nevertheless, although the versatility of the biosensors as analytical tools has increased and the biosensors are used for continuous monitoring of vital biochemical parameters [110, 111], some applications remain limited by functional deterioration due to surface fouling by proteins and other biological components [112]. Nowadays, nanotechnology is bringing continuously new possibilities for biosensors construction and for developing novel electrochemical bioassays. So, new developments are combining the use of the

5 Nanosensors

113

functionalized surface of carbon nanotubes or metal oxide (MO) nanowires as a gate electrode of the FET-like sensing structures [113]. Their advantages as conductive transducers are high surface area and accelerated electron transfer (ET) [114]. In spite of the very promising results, there are still many challenges due to the relatively inefficient interaction with biomolecules, low selectivity and interfacing activity with respect to specific analytes. The improvements can be directed at self-assembling a layer of specifically designed active molecules that make electrical communication between the analyte and the sensing surface (gate electrode in the case of FET scheme) more efficient, improve sensitivity and stability and even make the interfacing layer an active player in sensing mechanism [115]. The improvement of the functionalization steps and procedures becomes the right clue to have a reliable electrochemical communication between the analyte and the electrode surface with enhanced and controllable electron transfer mechanisms that still require much more effort for their understanding.

5.6 Nanosensors Based on Photon Capture: Photodetection at the Nanoscale In this section, like in the previous one where the advantages of playing in the nanoscale for the electrical interaction through the surface has been shown, we want to discuss about the advantages of using nanosized materials in lightsensing applications. Therefore, our aim is not to present a detailed description of all the photodetection strategies and achievements but to emphasize the photodetection capability at the nanoscale. Detailed information on this topic can be found elsewhere [116, 117, 118]. The simplest strategy to electrically ‘‘transduce’’ the presence of light is the use of photoconductors. In such materials, the energy of the absorbed light is used to promote an electron from the valence band to the conduction band of the material (usually, a semiconductor) leaving behind a hole in a process known as photogeneration of charge pairs. Consequently, the conductance of the material decreases after exposure to light and, in first approach, the minimum detectable wavelength corresponds to the bandgap energy of the material. Since this essentially is a bulk process, what is the advantage of using nanosized materials (with high surface-to-volume ratio) instead of other conventional technologies such as thin films [119]? Photoconductors, Fig. 5.32, based on individual nanowires, modelled as an arbitrary volume of length L, width W and thickness T can be studied using the fundamental principles ruling light carrier generation on semiconductors [120, 121, 122]. For simplicity, let us consider the case of extrinsic n-type semiconductors. In low-injection regime (i.e. under low illumination levels), the current

114

T. Andreu et al.

Fig. 5.32 Diagram of a metal oxide structure with an arbitrary volume of length L, width W and thickness T under a photon flux ph. Photocarrier generation is produced in the outer layer of the nanowire until depth –1

W

α–1

φph

L T

jph

density jph corresponding to the photogenerated electrons is given by the following equation: jph ¼ q

1     V ph ; L 1

(5:17)

where q is the elemental charge, L is the length of the nanowire,  is the absorption coefficient of the material at the wavelength of the incident light,  is the fraction of photons not reflected by the surface,  is the quantum efficiency of carrier generation by one photon, is the carrier lifetime, m* is the effective carrier mobility, V is the (external) applied voltage along the nanowire and ph is flux of photons impinging on the nanowire. Here we assumed a constant carrier generation profile until the depth –1 and, according to Matthiessen’s rule, m* can be broken down into the following factors [122]: 1 1 1 ¼ þ ;  B S

(5:18)

where mB and mS are the bulk and surface contribution. To evaluate the total photogenerated current Iph, which is the experimental response of real devices, one can assume that the nanowire is thick enough to absorb all the incident photons. That is to say, T 1 :

(5:19)

Therefore, it can be deduced that thinner nanowires (T < –1) will lead to lower photoresponses. On the contrary, the use of thicker nanowires (T >> –1) will not imply a further enhancement of the response. To exemplify this, the penetration depth –1 of near-UV photons (wavelength from 400 to 250 nm) in ZnO is almost constant at 50 nm [123]. Thus, ZnO nanowires with radii slightly above r 25 nm should be used to maximize photoresponse to UV light in this

5 Nanosensors

115

wavelength range. If the constant absorption profile approximation is maintained (light penetration depth –1 and nanowire width W), photocurrent Iph in nanowires which verify Eq. (5.19) is given by   W Iph ¼ jph 1 W ¼ q     V ph ; L

(5:20)

where three different contributions are clearly identified. The first one is related to geometric parameters of the device ðW=WLLÞ, the second one to the intrinsic properties of nanowires (b m*) and the third one depends only on the working conditions (Vph). The performance of these devices can be also analysed in terms of their photoconductive gainGph, which is defined as [121, 122,] Gph 

Iph 1 2   V; qF L

(5:21)

where the same type of contributions are identified. Concerning the geometry of photodetectors, it is clear from Eq. (5.4) that Iph is enhanced by increasing the width of photoactive area (W). Besides, Iph and Gph are also improved by decreasing the distance between the electrical contacts L (see Eqs. (5.20) and (5.21)). On the one hand, a convenient way to increase the area is using multinanowire-based configurations. These devices can be fabricated by means of self-assembly techniques, i.e. dielectrophoresis, to electrically contact large amounts of them in parallel [124, 125]. This fabrication methodology prevents parasitic effects arising from uncontrolled nanowire-to-nanowire contacts. It is noteworthy that according to Eq. (5.21), the photoconductive gain Gph obtained with these multinanowire configurations is equivalent to that provided by one single nanowire, if all of them are identical. On the other hand, both Iph and Gph increase inversely with the distance between contacts L. Thus, this experimental parameter is critical to optimize the performance of final photodetectors. The lower limit for L will strictly depend on the precision of the nanolithography technique and other size-associated phenomena like diffraction, if L approaches the wavelength of photons, or uncontrolled degradation effects produced when the electric field of rupture of the metal oxide is overcome. To exemplify this last point, it can be roughly estimated that nanowires contacted between two electrodes separated by only 50 nm [126] and polarized at 5 V [127, 128] will be subjected to electric fields as high as 1 MV/cm. As far as the intrinsic properties of nanowires are concerned, the dependence of Iph with, , m* and b must be also considered (see Eq. (5.20)). The quantum efficiency  determines the spectral response of photodetectors and was observed to increase orders of magnitude when photons with energies above the bandgap interact with these devices compared to typical responses obtained with sub-bandgap photons [129]. It is noteworthy that the bandgap edge of nanowires depends not only on the material but also on their dimensions [122,

116

T. Andreu et al.

130]. Thus, controlling the radii of nanowires is critical to tune the spectral sensitivity of the final devices. Photogenerated carrier lifetime is the second parameter directly related to the intrinsic properties of nanowires, and it is known to be higher in nanomaterials than bulk because of their large surfaceto-volume ratio and the formation of deep-level surface states [121, 131]. In the case of semiconductor nanowires, it is generally accepted that photocarrier relaxation dynamics consists of an initial decay process in the nanosecond range, explained by the fast carrier thermalization and hole trapping by surface states, followed by a slow decay dependent on the surrounding atmosphere and the nanowire surface coating [128, 132, 133, 134, 135]. This second process, with characteristic times of seconds, dominates the final response of nanowire-based photodetectors. For this reason, the carrier lifetime contribution to the photoresponse Iph (see Eq. (5.20)) can be enhanced by controlling the surface interactions of this type of nanowires. The third parameter related to the intrinsic properties of nanowires is the electrical mobility *, which is known to be dependent on their radii. To exemplify this with ZnO nanowires, mobility values ranging from 2 to 30 cm2/Vs were reported for nanowires with radii below r 100 nm [136, 137, 138, 139]. This diminished mobility tends to the bulk value (200 cm2/Vs) with thicker nanowires [128]. This behaviour is attributed to scattering and trapping of the electrons by surface defect states and becomes more important with thin nanowires, whose surface-to-volume ratio increases. In these circumstances, surface contribution mS to m* is significantly minimized (see Eq. (5.18)). Thus, nanowires as thick as possible are convenient to obtain optimal devices from a point of view of * optimization. The limitation introduced by the dependence of * with radius can be also overcome by coating the nanowires’ surface with a passivation layer. Using the previous example, it was demonstrated that the mobility of ZnO nanowires dramatically increased using this fabrication strategy (up to 4,000 cm2/Vs) [139]. The last intrinsic parameter of nanowires to be considered in this work is the fraction of photons not reflected by the surface of the metal oxide b, which was recently demonstrated to be lower in a wide spectrum range with photodetectors based on aligned nanowires instead of thin films [140]. As far as the experimental conditions are concerned, it can be demonstrated from Eqs. (5.20) and (5.21) that photoresponse rises with applied voltage V and flux of photons ph, making it difficult to compare most of the reported results, since different experimental conditions were used in all these experiments [127, 128]. For this reason the photoconductive gain (Eq. (5.21)) can be expressed in a more convenient way gph 

L2 Gph    ; V

(5:22)

which is normalized and independent of the device geometry and the experimental conditions.

5 Nanosensors

117

Besides the gain, the dynamic response of a photodetector is also important. The low-pass bandwidth is approximately BW ð2p Þ1 and the normalized gain per bandwidth becomes gph BW

  : 2p

(5:23)

Therefore, mobility and quantum efficiency are the key parameters to determine the overall performance of a photodetector. Equations (5.22) and (5.23) show the fact that high photoconductive gains are achieved by increasing the carrier lifetime but this worsens the dynamic response. Finally, it is illustrative to compare the performance already achieved using single nanowire photoconductors with thin film-based detectors [119]. We focused again in ZnO. In high-quality thin films mobility tends to the bulk value (200 cm2/Vs) [141]. The highest photoconductive gain reported with thin film photodetectors is GTF ph lit ¼ 1360 [142]. Using the polarization and geoTF metric conditions reported by the authors ð VTF ph lit ¼ 5V; Lph lit ¼ 10 Þ, the nor8 2 TF malized photoconductive gain is gph lit ¼ 3  10 m =V. This gain is clearly below the ones reachable with single ZnO nanowires in their present stage 7 of development ð GNW ph lit ¼ 5  10 Þ [128]. However, if we estimate the overall TF performance with the normalized gain per bandwidth ðlit  1:5msÞ product 3 2 TF we obtain gph lit ¼ 3  10 m Hz=V to be compared with the best figure for a 8 2 single nanowire photodetector ð gNW ph lit  BW ¼ 5  10 m Hz=VÞ [128]. Therefore, the high photogain currently achieved with individual nanowires is mainly associated with the longer lifetime of the photocarriers. In conclusion, nanowires can be used as photoconductors as an alternative to conventional technologies. However, we can only take advantage of their superior properties (high crystalline quality and extremely high mobility) if appropriate layout and operating conditions are considered. Some of these promising strategies are diminishing the distance between the electrical contacts, increasing the photoactive area or improving the electrical mobility of nanomaterials using surface coatings. Besides photoconductors, there exist a number of other devices based on the photogeneration of charge pairs that are conventionally used in photodetection [119, 120, 121, 122]. Most of them are based on the charge separation of the photogenerated pairs (electron and hole) by a built-in electric field. In these cases the junction of different materials (metal–semiconductor, heterojunction and homojunction of semiconductors) is used to generate these built-in fields. In Fig. 5.33 we summarize some of them [119]. Recently, many of these configurations have been successfully obtained using nanowires (diodes in individual nanowires [143, 144], core–shell nanowires [145, 146], nanowire–wafer unions [147]) and open the door to further improvements in photodectection with nanowires. Finally, similar nanostructures and devices can be also used in solar energy harvesting applications in which the same principle of separation of

118

T. Andreu et al.

Fig. 5.33 Schematic structure of different semiconductor photodetectors suggested by [119]

photogenerated pairs can be used to provide power (voltage and current) to an external electrical load [122, 148. 149]. Particularly, the use of nanosized materials in cost-effective dye-sensitized solar cells is necessary to efficiently collect the charge generated by light in the organic dye layer [150, 151, 152]. Other promising achievements are the integration of a fully functional solar cell in one single nanowire [143, 144, 153].

5.7 Nanosensors Based on Plasmon Resonance: Influence of the Dielectric Constant Variations at the Nanoscale Plasmons, collective oscillations of the free electron gas density, can be exploited for sensor purposes in two different ways: (i) as transducers of the measured magnitude into an optical signal or (ii) as enhancers of the spectroscopic signatures concerning the measured magnitude. As a transducer, we take advantage of the sensitivity of the plasmon resonance to the very close dielectric environment that can be easily achieved by controlling it at the nanoscale domains that present a high surface/volume ratio enhancing the role played by the surface plasmon. This dependence to a highly localized dielectric environment, coupled with the use of metal nanoparticles as sensing probes, provides with a considerable spatial resolution to probe-based sensors taking advantage of this transduction mechanism. In this direction, silver and gold nanoparticles, having resonance frequencies in the visible region, have been used as single particle optical biochemical and chemical sensors [154, 155, 156]. Usually, dark field optical microscopy is used to analyse the resonance spectra of these single nanoparticles. Its resonance frequency is very sensitive to the adsorbants on its surface (Fig. 5.34).

5 Nanosensors

119

Fig. 5.34 Cartoons of the (a) plasmon generation in metal nanoparticles by an electromagnetic field, (b) modification of the plasmon resonance frequency of single nanoparticles due to the presence of the target analyte and enhancement of the optical signature of the species nearby the metal surface, (c) waveguide surface plasmon resonance sensor and (d) tipenhanced Raman spectroscopy

Plasmon resonances are sensitive not only to the surrounding dielectric media but also to the shape, size and interaction between particles [157, 158]. While such dependences deteriorate the characteristics of nanoparticle arrays as plasmon resonance sensors, they can be used to measure geometric parameters at the nanometer scale [159]. As an example, interparticle distances, and thus molecular length, can be optically measured using metal nanoparticles linked by the entity for which distance needs to be measured. This plasmon molecular ruler can be used to further analyse the geometric parameters of suitable biological processes [160]. Probe-based sensors require somewhat demanding systems for signal discrimination and imaging. The analysis of the plasmon resonance spectra of single nanoparticles usually requires a dark field optical microscopy setup. Further integrated plasmon resonance sensors can be obtained using metal thin films in place of independent nanoparticles. In this direction, a complete sensor can be fabricated by coupling a thin layer of gold or silver with an emitter and a receptor. Due to the different momentums of surface plasmon waves and light, the coupling of the electromagnetic waves with the surface plasmons on a metal thin film requires a special configuration. There are three coupling designs predominantly used: (a) the coupling using a prism (ATR method);

120

T. Andreu et al.

(b) coupling using an optical waveguide and (c) coupling using a grating coupler [161, 162, 163, 164]. In these devices, surface plasmon propagation constants strongly depend on the wavelength, showing a propagation maximum, that is a light absorption maximum, at a given resonance wavelength. The field of the surface plasmon waves generated at the metal/dielectric interface strongly decays into both media. Therefore, the resonance wavelengths, like on surface plasmons generated in nanoparticles, are highly sensitive to the very localized dielectric medium neighbouring the metal surface. Thus, changes on this dielectric constant, such as those associated with the presence of an analyte, can be precisely detected by measuring the intensity, the coupling angle or the wavelength of resonance coupling of the surface plasmon wave and light. Resonance peak positions not only are used to sense the presence of analytes but can be used to determine distances and even analyse biological processes thanks to the sensitivity of the plasmon resonance to interparticle distances [165, 166]. The very intense fields created at the resonance frequencies in the close vicinity of the metal nanostructures [167, 168, 169] can also be used as enhancers of the spectroscopic signatures of analytes attached or lying close to the metal surface [170, 171, 172]. Surface-enhanced Raman spectroscopy uses this extreme field enrichment to improve Raman sensitivity up to 1011 times [173, 174, 175]. This technique, which can be integrated using optic fibres [176] and potential in vivo usage [177, 178], finds its most intriguing feature on its potentially huge spatial resolution. The very localized enhancement obtained at the metal surfaces allows for the simultaneous topographic and spectroscopic analysis of the sample with a resolution at the nanometer scale when coupling the surface-enhanced Raman spectroscopy with an AFM provided with goldor silver-coated tips [179]. So, the most interesting feature is its potential spatial resolution. In this direction, the coupling of the surface-enhanced Raman spectroscopy with an AFM using metal tips allows the simultaneous topographic and spectroscopic analysis of the sample [180].

5.8 Nanosensors Based on Mechanical Resonances: Influence of Small Mass Changes onto Nanostructures At the nanoscale, small mechanical forces are involved which is a challenge for using them as transduction elements in a device [181]. But, nowadays, they are used in promising mechanical devices as they are quite sensitive to very small mass changes. For this reason mechanical resonators offer a very promising standard alternative although it requires aligned and free-standing active elements at the nanoscale that are, at the present technology level, not straightforward or easy to obtain or to process, mainly due to the residual stress controls. Anyway, among many possible nanostructures, small cantilevers, which monitor the resonant frequency of the cantilever, are one of the most used sensitive mass detectors.

5 Nanosensors

121

The theoretical mass sensitivity (m/f) of a rectangular cantilever can be calculated as
m=fðkg=HzÞ 0:208 ðk=f 30 Þ;

(5:24)

assuming that the spring constant k remains constant after the mass change and that the mass is evenly distributed on the cantilever. f0 is the cantilever’s resonant frequency and f is the frequency shift of the cantilever. f0 depends on the viscous medium in which it is vibrating due to viscous damping of the cantilever, and in vacuum f0 will be equal to p fvac ¼ 0:162 E=ðw=l2 Þ;

(5:25)

where  is the mass density of the cantilever material, E the Young’s modulus and l and w the length and width of the cantilever. Therefore, mass sensitivity increases as the resonant frequency increases too [182]. It depends inversely on the size. For a cantilever with a width, height and length of 1, 2 and 50 mm, respectively, m/f is approximately 10–15 g/Hz. This equation also demonstrates that by decreasing the dimensions even further the mass sensor could theoretically detect single molecules and, of course, large proteins [183]. Normally, for a micron range size cantilever, the movement is detected by optical techniques where a laser is focused on and reflected from the cantilever surface. However, as the cantilever dimensions are decreased to nanometer scale, this method becomes difficult to apply and new readout techniques have been developed. Typically, the readout from nanometer-sized cantilevers has been rather complicated and the operation of the mass sensor has been limited to vacuum [184]. It is well known that due to the damping mechanism introduced by the air molecules at atmospheric pressure even more than two orders of magnitude in the Q-factor can exist. New alternatives have presented a nanometer-sized cantilever with integrated capacitive readout, which offers a simple readout scheme that can be operated at atmospheric pressure and vacuum [185, 186]. In this option, by monolithic integration of the nanocantilever and CMOS circuitry, it has been shown that it is possible to increase both the functionality and the resolution of the sensor, allowing operation in more viscous media or functionalities like resonant frequency tracking and Q-factor enhancement. In this technological frame, these mass nanosensors are open for new applications that require the measurement of particle fluxes in high vacuum – i.e. an atom lithography system – or a high-sensitive gas sensor. Likewise, applications that need a high spatial resolution can also be obtained, i.e. the localized measurement of particle fluxes, something that is impossible with quartz microbalances [187, 188, 189].

122

T. Andreu et al.

As alternatives to these readout methods to overcome the difficulties inherent in the optical detection for very small cantilevers, AFM-based methods have been proposed on a cantilever actuated electrostatically by means of a driver electrode placed closely parallel to the cantilever. For it, the AFM is operated in a dynamic non-contact mode using oscillation amplitudes corresponding to a low-force regime. The dependence of the static cantilever deflection on DC voltage and of the oscillation amplitude on the frequency of the AC voltage is measured and the results are fitted by a simple non-linear electromechanical model [190]. Even though this method works well and is very precise, the resulting system complexity makes it difficult to use in portable low-cost devices. For these reasons, from a practical point of view, the straightforward measurement of the cantilever deflection by means of an electrical transducer integrated on the cantilever becomes more plausible and realistic but it needs the integration compatibility with on-chip electronics fabrication. It can be obtained if the cantilever deflection is measured by means of a piezoresistor integrated in the cantilever [191]. In fact, the use of cantilevers with submicrometer thickness and width in the micrometer range gives the possibility to have small spring constant and high force sensitivity that are the requirements for detecting small forces like those involved in molecular recognition experiments [192, 193, 194, 195]. A clear example of these options is given by the detection of antibody–antigen forces. For example, cantilevers with 2 mm thickness have been reported to have a spring constant k 4 N/m. However, the forces to be measured for molecular detection are in the range of tens to hundreds of piconewton. Then, to obtain a deflection of 1 nm under a 10 pN load, the cantilever must have a small spring constant of k 0.01 N/m. This spring constant can be achieved with a silicon microcantilever with small spring constants and high force sensitivity, which as we shall see requires transverse cantilever dimensions in the submicrometer range, which should be submicron in thickness and about 1 mm in width [196, 197]. Thus, for improving the system performances, it has been proposed by several groups to use a U-shaped structure for the cantilevers [198, 199] in which the piezoresistor uses the complete surface of the cantilever. This shape allows the minimization of the cantilever dimensions, especially its width. If l >> w, where l is the length of the cantilever and w the width of each leg, the U-shaped cantilever will behave as two identical separated rectangular cantilevers corresponding to the two legs. The classical beam theory for small deformations can be applied to calculate the stress in the rectangular cantilever for a given perpendicular force F applied to the tip. The stress can be used to calculate the output signal
V when the piezoresistor is connected in a half Wheatstone bridge with an identical reference cantilever resistor. Using these procedures, it has been probed that it is possible to fabricate cantilevers with lengths between 60 and 200 mm and leg widths between 2 and 10 mm with spring constants ranging from 1.02 to 0.006 N/m and the force sensitivities from 2.4 to 11.8 mV/nN that are given a resolution, within the parameter ranges used, of about 10 pN.

5 Nanosensors

123

5.9 Summary The improved knowledge and control of the transduction mechanisms at the nanoscale level has enabled and enhanced the ideas, designs and developments of nanosensors. There are many examples that have been reported during the last decades that cover all the fields of physical, chemical and biochemical areas. Nevertheless, advancements have been more relevant in some sensor types. Among them, some examples have been highlighted as representative of the nanoscience and nanotechnology contributions to the sensor field. So, controlled performances of surface/volume-based sensors using a single nanocrystal as conductometric or transistor channel device have already been achieved and these sensors are applied in the chemical and biochemical fields using inorganic and organic materials. Surface properties control and their functionalizations become one of the most outstanding and challenging features for future developments. On the other hand, the interaction of nanomaterials with photons of different energies at the nanoscale level is being used as one of the best scenarios for charge separation, charge collection and electrical or optical confinements. Likewise, the control of in some properties such as plasmon resonances at the nanoscale level has given rise to many options for new nanosensor applications based on the enhanced Raman mechanism of fundamental importance for sensing analytes. Furthermore, mechanical resonances-based nanosensors have also been reported to bring a paramount importance for new highly mass-sensitive sensor. There are many other options to take advantage of the mechanisms and properties at the nanoscale level, and in the immediate future their associated feasibility challenge will be solved on the basis of the nanotechnology advancements and nanoscience understandings.

References 1. http://www.nano.gov 2. W. Gopel, J. Hesse, and J.N. Zemel. Sensor: A comprehensive survey. VCH, Weinheim, ¨ Germany, 1991. 3. A.P. Alivisatos, Science, 271, 933–937 (1996). 4. K. Kalyanasundaram and M. Gra¨tzel, Curr Sci, 66, 706–715 (1994). 5. W.J. Aston, Biosens Bioelectron, 7, 161–163 (1992). 6. S.P. Mohanty and E. Kougianos, IEEE Potentials, 25 (2), 35–40 (2006). 7. D. Erickson, S. Mandal, A.H.J. Yang, and B. Cordovez, Microfluidics and Nanofluidics, 4, 33–52 (2008). 8. M. Epifani, R. Diaz, J. Arbiol, E. Comini, N. Sergent, T. Pagnier, P. Siciliano, G. Faglia, and J.R. Morante, Advanced Funct. Mater. 16, 1488–1498 (2006). 9. M. Epifani, J. Arbiol, E. Pellicer, and J.R. Morante, Chem. Mater., 19 (20), 4919–4924 (2007). 10. A.-H. Lu, E.L. Salabas, and F. Schu¨th, Angew. Chem. Int. Ed., 46, 1222–1244 (2007).

124

T. Andreu et al.

11. X.-M. Lin and A.C.S. Samia, J. Magn. Magn. Mater., 305, 100–109 (2006). 12. A. Cabot, A. Die´guez, A. Romano-Rodrı´ guez, J.R. Morante, and N. Barsan, Sensors Actuat. B, 79, 98–106 (2001). 13. M.C. Daniel and D. Astruc, Chem. Rev., 104, 293–346 (2004). 14. G. Schmid, Chem. Rev., 92, 1709–1727 (1992). 15. A. Dieguez, A. Vila, A. Cabot, A. Romano-Rodriguez, J.R. Morante, J. Kappler, N. Barsan, U. Weimar, and W.Gopel, Sensors Actuat. B-Chem., 68, 94–99, (2000). 16. J. Puigcorbe, A. Vila, J. Cerda, A. Cirera, I. Gracia, C. Cane, and J.R. Morante, Sensors Actuat. A, 97–98, 379–385 (2002). 17. J. Arbiol, Metal additive distribution in TiO2and SnO2 semiconductor gas sensor nanostructured materials. PhD Thesis Dissertation. University of Barcelona (2001) ISBN84475-2636-4. 18. M. Epifani, E. Comini, J. Arbiol, E. Pellicer, P. Siciliano, G. Faglia, and J.R. Morante, J. Phys. Chemc, 111 (37), 13967–13971 (2007). 19. I.R. Peterson, J. Phys. D, 23, 379–395 (1990). 20. P. Yang and F. Kim, Chemphyschem, 3, 503–506 (2002). 21. L. Vayssieres, Adv. Mater., 15 (5), 464–466 (2003). 22. L. Vayssieres and M. Graetzel, Angew. Chem. Int. Ed., 43, 3666–3670 (2004). 23. L. Vayssieres, Appl. Phys. A, 89 (1), 1–8 (2007). ´ A. Cornet, J.R. Morante, 24. E. Rossinyol, A. Prim, E. Pellicer, J. Rodrı´ guez, F. Peiro, B.Z. Tian, T. Bo, and D. Zhao, Sens Actuat. B, 126, 18–23 (2007). ´ A. Cornet, and J.R. Morante, Adv. Funct. 25. A. Prim, E. Pellicer, E. Rossinyol, F. Peiro, Mater., 17 (15), 2957–2963 (2007). ´ A. Cornet, 26. E. Rossinyol, A. Prim, E. Pellicer, J. Arbiol, F. Herna´ndez- Ramı´ rez, F. Peiro, J.R. Morante, L.A. Solovyov, T. Bozhi, T. Bo, and D. Zhao, Adv. Funct. Mater., 17, 1801–1806 (2007). ´ A. Cornet, and 27. E. Rossinyol, E. Pellicer, A. Prim, S. Estrade, J. Arbiol, F. Peir o, J.R. Morante, J. Nanoparticle Res., 10, 369–375 (2008). 28. A.I. Hochbaum, R. Chen, R. Dı´ az, W. Liang, E.. Garnett, M.G. Najarian, A. Majumdar, and P. Yang, Nature, 451, 163–167 (2008). 29. J. Arbiol, A. Cirera, F. Peiro, A. Cornet, J.R. Morante, J.J Delgado, and J.J Calvino, Appl. Phys. Lett., 80, 329–31 (2002). 30. Y.S. Kim, S.C. Ha, K. Kim, H. Yang, S.Y Choi, Y.T Kim, J.T. Park, C.H. Lee, J. Choi, J. Paek, and K. Lee,. Appl. Phys. Lett., 86, 213105 (2005). 31. Y. Cui, Q. Wei, H. Park, and C.M. Lieber, Science, 293, 1289–1292 (2001). 32. F. Herna´ndez-Ramirez, J.D. Prades, A. Tarancon, S. Barth, O. Casals, R. Jime´nez-Diaz, E. Pellicer, J. Rodriguez, M.A. Juli, A. Romano-Rodrı´ guez, J.R. Morante, S. Mathur, A. Helwig, J. Spannhake, and G. Muller, Nanotechnology, 18, 495501–06 (2007). 33. E. Comini, G. Faglia, G. Sberveglieri, Z. Pan, and Z.L. Wang,. Appl. Phys. Lett., 81, 1869 (2002). 34. S. Chopra, A. Pham, J. Gaillard, A. Parker, and A.M. Rao, Appl. Phys. Lett., 80, 4632 (2002). 35. G.K. Mor, O.K. Varghese, C.A. Grimes, M.A. Carvalho, and M.V. Pishko, J. Mater Res. , 19, 628–634 (2004). 36. Z.W. Pan, Z.R. Dai, and Z.L. Wang, Science, 291, 1947–1949 (2001). 37. M. Borgstrom, G. Immink, B. Ketelaars, R. Algra, and E. Bakkers, Nature Nanotechnol., 2, 541–544 (2007). 38. R.S. Wagner and W.C. Ellis, Appl. Phys. Lett., 4, 89–90 (1964). 39. E. Bauer and J. H. van der Merwe, Phys. Rev. B, 33, 3657–3671 (1986). 40. Y. Cui, Q. Wei, H. Park, and C. Lieber, Science, 293, 1289 (2001). 41. J. Arbiol, B. Kalache, P. Rocai Cabarrocas, J. R. Morante, and A Fontcuberta i Morral, Nanotechnology, 18, 305606–14 (2007). 42. A. Fontcuberta i Morral, C. Colombo, G. Abstreiteer, J. Arbiol, and J.R. Morante, Appl. Phys. Lett., 92, 063112 (2008).

5 Nanosensors

125

43. A.L. Giermann and C.V. Thompson, Appl. Phys. Lett., 86, 121903 (2005). 44. W.K. Choi, T.H. Liew, H.G. Chew, F. Zheng, C.V. Thompson, Y. Wang, M.H. Hong, X. D. Wang, L. Li, and J. Yun, Small, 4, 330–333 (2008). 45. H.A. Pohl, Dielectrophoresis: the behaviour of neutral matter in nonuniform electric field, Cambridge University Press, Cambridge, UK (1978) 46. P.J. Burke, Encyclopedia of Nanoscience and Nanotechnology, In: H.S. Nalwa (Ed.) Vol 6 American Scientific Publishers, Los Angeles, CA 2004, pp 623–641. 47. Q. Wan, E. Dattoli, and W. Lu, Small, 4 (2008). 48. B. Tian, X. Zheng, T.J. Kempa, Y. Fang, N. Yu, G. Yu, J. Huang, and C.M. Lieber, Nature, 449, 885–890 (2007). 49. Z. Zhong, C. Yang, and C.M. Lieber, Nanosilicon, 176–216. (Ed. V. Kumar, Elsevier (2008) ). 50. A. Vomiero, M. Ferroni, E. Comini, G. Faglia, and G. Sberveglieri, Nano Lett., 7, 3553–3558 (2007). 51. A. Fontcuberta i Morral, D. Spirkoska, J. Arbiol, M. Heigoldt, J.R. Morante, and G. Abstreiter, Small, 4, 899–903 (2008). 52. K.Q. Peng, Y.J. Yan, S.P. Gao, and J. Zhu, Adv. Mater., 14, 1164–1167 (2002). 53. P. Gorostiza, M.A. Kulandainathan, R. Diaz, F. Sanz, P. Allongue, and J.R. Morante, J. Electrochem. Soc., Society, 147, 1026 (2000). 54. P. Gorostiza, P. Allongue, R. Dı´ az, J.R. Morante, F. Sanz, J.Phys.Chem. B, 107, 6454–6461 (2003). 55. A.R. Leach, Molecular modelling principles & applications. Addison Wesley Publishing Company, USA (1997). 56. J.D. Prades, A. Cirera, and J.R. Morante, Quantum chemical calculations of surfaces and interfaces of materials, In: V.A. Basluk and P. Ugllengo (Eds.) American Scientific Publisher California, USA (2008). 57. J.D. Prades, A. Cirera, and J.R. Morante, J. Electrochem. Soc., 154 (8), H675–H680 (2007). 58. M. Batzill and U. Diebold, Prog. Surf. Sci., 79, 47 (2005). 59. C.N. Xu, J. Tamaki, N. Miura, and N. Yamazoe, Chem. Lett., 1990(3), 441 (1990). 60. C.N. Xu, J. Tamaki, N. Miura, and N. Yamazoe, Tech. Digest of 9th Sensor Symp., Tokyo, Japan, p.95 (1990). 61. C.N. Xu, J. Tamaki, N. Miura, and N. Yamazoe, Sens. Actuat. B, 3, 147 (1991). 62. J. Tamaki, Z. Zhang, K. Fujimori, M. Akiyama, T. Harada, N. Miura, and N. Yamazoe, J. Electrochem. Soc., 141, 2208 (1994). 63. Y.- G. Choi, G. Sakai, K. Shimanoe, N. Miura, and N. Yamazoe, Sens. Actuat. B, 95, 258, (2003). 64. A. Cabot, J. Arbiol, E. Rossinyol, J.R. Morante, F. Chen, and M. Liu, Electrochem. Solid State Lett., 7 (5), G93–G97 (2004). 65. M.N. Rumyantseva, A. Gaskov, N. Rosman, T. Pagnier, and J.R. Morante, Chem. Mater., 17, 893–901 (2005). 66. A. Cabot, J. Arbiol, A. Cornet, J.R. Morante, F. Chen, and M. Liu, Thin Solid Film, 436, 64–69 (2003). 67. W. Gopel and K.D. Schierbaum, Sens. Actuat. B, 26, 1–12 (1995). 68. N. Barsan, D. Koziej, and U. Weimar, Sens. Actuat. B, 121, 18–35 (2007). 69. J. Jortner and C.N.R. Rao, Pure Appl. Chem., 74, 1489–1783 (2002). 70. M. Law, J. Goldberger, and P. Yang. Annu. Rev. Mater. Res., 34, 83–122 (2004). 71. A. Cabot, A. Die´guez, A. Romano-Rodrı´ guez, J.R. Morante, and N. Barsan, Sens. Actuat. B, 79, 98–106, (2001). 72. A. Dieguez, A. Vila, A. Cabot, A. Romano-Rodriguez, J.R. Morante, J. Kappler, N. Barsan, U. Weimar, and W. Gopel, Sens. Actuat. B-Chem., 68, 94–9, (2000). 73. J. Kappler, N. Barsan, U. Weimar, A. Dieguez, J.L. Alay, A. Romano-Rodrı´ guez, J.R. Morante, W. Gopel, and Fresenius, J. Anal. Chem., 361, 110–114 (1998). ¨

126

T. Andreu et al.

74. http://mathforum.org/library/drmath/view/65234.html 75. J. Tamaki, J. Niimi, S. Ogura, and S. Konishi, Sens. Actuat. B, 117, 353–358 (2006). 76. J. Tamaki, A. Miyaji, J. Makinodan, S. Ogura, and S. Konishi, Sens. Actuat. B, 108, 202–206 (2005). 77. J. Tamaki, Y. Okochi, and S. Konishi, Electrochemistry, 74, 159–162 (2006). 78. F. Hernandez-Ramirez, Fabrication strategies and characterization of nanodevices fabricated with focused ion beam techniques. PhD Thesis Dissertation, University of Barcelona (2007). 79. G.J. Li and S. Kawi, Matter. Lett., 34, 99–102 (1998). 80. P.A. Smith, C.D. Nordquist, T.N. Jackson, T.S. Mayer, B.R. Martin, J. Mbindnyo, and T.E. Mallouk, Chem.-Eur. J., 8, 4355 (2002). 81. M.E. Toimil-Morales, E.M. Hohberger, C.H. Schaeflein, R.H. Blick, R. Neumann, and ¨ C. Trautmann, Appl. Phys. Lett., 82, 2139 (2003). 82. ‘‘Focused ion beam technology, capabilities and applications’’. FEI Company (2004). Brochure available at (www.feicompany.com) 83. S. Reyntjens and R. Puers, J. Micromech. Microeng., 11, 287–300 (2001). 84. S. Matsui and Y. Ochiai, Nanotechnology, 7, 247–258 (1996). 85. G. de Marzi, D. Iacopino, A.J. Quinn, and G. Redmond, J. Appl. Phys., 96, 6 (2004). 86. L. Rotkina, J-F. Lin, and J. P. Bird, Appl. Phys. Lett., 83, 4426 (2003). 87. F. Hernandez-Ramirez, O. Casals, J. Rodrı´ guez, A. Vila, A. Romano-Rodriguez, J.R. Morante, M. Abid, S. Valizadeh, Mat. Res. Soc., Symp. Proc., J 5.2 (2005). 88. V. Gopal, V.R. Radmilovic, C. Daraio, S. Jin, P. Yang, and E. Stach, Nano. Lett., 4, 2059 (2004). 89. A. Vila, F. Hernandez-Ramirez, J. Rodriguez, O. Casals, A. Romano-Rodriguez, J.R. Morante, and M. Abid, Mater. Sci. Eng. C., 26, 1063–1066 (2006). 90. M. Law, H. Kind, B. Messe, F. Kim, and Y. Peidong, Angew. Chem., 114, 2511 (2002). 91. F. Herna´ndez-Ramı´ rez, J. Rodrı´ guez, O. Casals, E. Russinyol, A. Vila`, A. RomanoRodrı´ guez, J.R. Morante, and M. Abid, Sens. Actuat. B, 118, 98–203 (2006). 92. F. Hernandez-Ramirez, J.D. Prades, A. Tarancon, S. Barth, O. Casals, R. Jime´nez– Diaz, E. Pellicer, J. Rodriguez, M.A. Juli, A. Romano-Rodriguez, J.R. Morante, S. Mathur, A. Helwig, J. Spannhake, and G. Mueller, Nanotechnology, 18, 495501–06 (2007). ´ O. Casals, J. Arbiol, A. Romano-Rodrı´ guez, and 93. F. Herna´ndez-Ramı´ rez, A. Tarancon, J.R. Morante, Sens. Actuat. B, 121, 3–17 (2007). 94. Z.R. Dai, Z.W. Pan, and Z.L. Wang, Adv. Funct. Mater., 13, 9–24 (2003). 95. R.S. Wagner and W.C. Ellis, Appl. Phys. Lett., 4, 89–90, (1964). 96. Y. Ding, P.X. Gao, and Z.L. Wang, J.Am. Chem. Soc., 126, 2066 (2004). 97. S. Mathur, S. Barth, H. Shen, J.-C. Pyun, and U. Werner, Small, 1, 7, 713–717 (2005). 98. S. Lenaerts, M. Honore, G. Huyberechts, J. Roggen, and G. Maes, Sensor and Actuators B, 18/19, 478 (1994). 99. A. Cabot, J. Arbiol, E. Rossinyol, J.R. Morante, F.L. Chen, M.L. Liu, Electrochem. Solid State Lett., 7, G93 (2004). 100. M.N. Rumyantseva, A.M. Gaskov, N. Rossman, T. Pagnier, J.R. Morante, Chem. Mater., 17, 893 (2005). 101. D. Li, Y. Wu, L. Shi, P. Yang, and A. Majumdar, Appl. Phys.Lett., 83,14 (2003). 102. C. Yu, Nanomaterials characterization and bio-chemical sensing using microfabricated devices. Thesis Dissertation, University of Texas (2004). 103. S. Mathur, S. Barth, H. Shen, J.C. Pyun, and U. Werner, Small, 1, 713 (2005). 104. F. Hernandez-Ramrez, A. Tarancon, O. Casals, E. Pellicer, J. Rodrguez, A. RomanoRodrguez, J.R. Morante, S. Barth, and S. Mathur, Phys. Rev. B, 76, 085429 (2007). 105. G. Decher, Science, 277, 1232–1237 (1997). 106. A. Ulman, An introduction to ultra thin organic films from Langmuir Blodgett to selfassembly. Academic Press, New York, USA (1991).

5 Nanosensors

127

107. J. Zhang, Z.L. Wang, J. Liu, S. Chen, and G.Y. Liu, Self assembled nanostructures. Kluwer Academic/Plenum Publishing, New York, USA (2003). 108. M. La Barbera, Science, 289, 1882 (2000). 109. C.M. Niemeyer and C.A. Mirkin, Nanobiotechnology: Concepts, applications and perspectives. Wiley-VCH, Dortmund, Germany, (2004). 110. N. Wickramasinghe, S. Choudhary, and E. Geisler, Int. J. Biomed. Eng. Technol., 1, 41–58 (2007). 111. S. Vikas and C. Pundir, Sens. Trans., 82, 1405–1417 (2007). 112. D. Fraser, An introduction to in vivo biosensing: progress and problems, In: Fraser, D. (Ed.), Biosensors in the body, continuous in vivo monitoring. Wiley, London, 10–56 (1997). 113. M. Pumera, S. Sanchez, I. Ichinose, and J. Tang, Sens. Actuat. B, 123, 1195–1205 (2007). 114. X.-J. Huanga and Y.K. Choi, Sens. Actuat. B, 122, 659–671 (2007). 115. V. Chukharev, T. Vuorinen, A. Efimov, N.V. Tkachenko, M. Kimura, S. Fukuzumi, H. Imahori, and H. Lemmetyinen, Langmuir, 21, 6385–6391 (2005). 116. P. Avouris and J. Chen, Mater. Today, 9, 10, 46–54 (2006). 117. D.J. Sirbuly, M. Law, H.Q. Yan, et al., J. Phys. Chem. B, 109, 32, 15190–15213, (2005). 118. H.E. Maes, C. Claeys, R. Mertens, et al., Adv. Eng. Mater., 3, 10781–787 (2001). 119. E. Monroy, F. Omnes, and F. Calle, Semicond. Sci. Technol., 18, R33–R51 (2003). 120. A. Rose, Concepts in photoconductivity and allied problems. Interscience Publishers: New York (1963). 121. R.H. Bube, Photoelectronic properties of semiconductors. Cambridge University Press: Cambridge (1992). 122. S.M. Sze, ‘‘Physics of semiconductor devices’’. John Wiley & Sons, Inc: New York (1981). 123. H. Yoshikawa and S. Adachi, Jpn, J. Appl. Phys., 36, 10, 6237–6243, (1997). 124. S. Kumar, S. Rajaraman, R.A. Gerhardt, Z.L. Wang, and P.J. Hesketh, Electrochim. Acta, 51, 943–951, (2005). 125. J. Suehiro, N. Nakagawa, S. Hidaka, M. Ueda, K. Imasaka, M. Higashihata, T. Okada, and M. Hara, Nanotechnology, 17, 2567–2573 (2006). 126. This resolution value was taken as a rough approximation to the minimal feature of current lithography techniques. 127. H. Kind, H. Yan, B. Messer, M. Law, and P. Yang, Adv. Mater. (Weinheim, Germany) 14 (2), 158–160 (2002). 128. C. Soci, A. Zhang, B. Xiang, S.A. Dayeh, D.P.R. Aplin, J. Park, X.Y. Bao, Y.H. Lo, and D. Wang, Nano Lett., 7 (4), 1003–1009, (2007). 129. X.G. Zheng, Q. Sh. Li, J.P. Zhao, D. Chen, B. Zhao, Y.J. Yang, L. Ch. Zhang, Appl. Surf. Sci., 253, 2264–2267 (2006). 130. S. Mathur, S. Barth, H. Shen, J.-C. Pyun, and U. Werner, Small, 1, 77, 713–717 (2005). 131. J.S. Jie, W.J. Zhang, Y. Jiang, X.M. Meng, Y.Q. Li, and S.T. Lee, Nano Lett., 6, 9, 1887–1892 (2006). 132. Z.Y. Fan, P.C. Chang, J.G. Lu, E.C. Walter, R.M. Penner, C.H. Lin, and H.P. Lee, Appl. Phys. Lett., 85, 25, 6128–6130 (2004). 133. Y.W. Heo, B.S. Kang, L.C. Tien, D.P. Norton, F. Ren, J.R. La Roche, and S.J. Pearton, Appl. Phys. A: Mater. Sci. Process., 80, 3, 497–499 (2005). 134. S.A. Studenikin, N. Golego, and M. Cocivera, J. Appl. Phys., 87, 5, 2413–2421 (2000). 135. C.S. Lao, M.-C. Park, Q. Kuang, Y. Deng, A.K. Sood, D.L. Polla, Z.L. Wang, J. Am. Chem. Soc., 129 (40), 12096–12097 (2007). 136. P.-C. Chang, C.-J. Chien, D. Stichtenoth, C. Ronning, and J.G. Lu, Appl. Phys. Lett., 90, 113101 (3 pages) (2007). 137. Z.Y. Fan, D.W. Wang, P.C. Chang, W.Y. Tseng, J.G. Lu, Appl. Phys. Lett., 85, 5923 (2004).

128

T. Andreu et al.

138. P. -C. Chang, Z. Fan, D. Wang, W.. - Y. Tseng, W. - A. Chiou, J. Hong, J.G. Lu, Chem. Mater., 16, 5133–5137 (2004). 139. P. Chang, Z. Fan, C. Chien, D. Stichtenoth, C. Ronning, J.G. Lu, Appl. Phys. Lett., 89, 133113 (2006). 140. L. Hu and G. Chen, Nano Lett. 7, 11, 3249–3252 (2007). 141. A. Ohtomo and A. Tsukazaki, Semicond. Sci. Technol., 20, S1–S12 (2005). 142. Y. Liu, C.R. Gorla, S. Liang, N. Emanetoglu, Y. Lu, H. Shen, and M. Wraback, J. Electron. Mater., 29, 69–74 (2000). 143. K. Q. Peng, Z.P. Huang, and J. Zhu, Adv. Mater., 16, 73–76 (2004). 144. Chen Yang, Zhaohui Zhong, and Charles M. Lieber, 310, 1304–1307 (2005). 145. L.J. Lauhon, M.S. Gudiksen, D.Wang, and C.M. Lieber, Nature, 420, 57–61 (2002). 146. Y.-L. Chueh, L.-J. Chou, and Z.L. Wang, Angew. Chem. Int. Ed., 45, 7773–7778 (2006). 147. B. Pradhan, S.K. Batabyal, and A.J. Pal, Appl. Phys. Lett., 89, 233109,(2006). 148. M. Law, L.E. Greene, J.C. Johnson, R. Saykally, and P. Yang, Nat. Mat., 4, 455 (2005). 149. X.D. Wang, J.H. Song, J. Liu, and Z.L. Wang, Science, 316, 102 (2007). 150. M. Gratzel, Nature, 414, 338 (2001), and B. O’Regan, M. Gra¨tzel, Nature, 353, 737 (1991). 151. W.U. Huynh, J.J. Dittmer, and A.P. Alivisatos, Science, 295, 242 (2002). 152. L.E. Greene, M. Law, B.D. Yuhas, and P.D. Yang (2007). J. Phys. Chem. Lett., 111,18451 (2007). 153. B. Tian, X. Zheng, T.J. Kempa, Y. Fang, N. Yu, G. Yu, J. Huang, and C.M. Liebre, Nature, 449, 885–889 (2007). 154. Chanda Ranjit Yonzon, Douglas A. Stuart, Xiaoyu Zhang, Adam D. McFarland, Christy L. Haynes, and Richard P. Van Duyne, Talanta, 67, 438–448 (2005). 155. Adam D. McFarland and Richard P. Van Duyne, Nano Lett., 3, 1057 (2003). 156. J.J. Mock, David R. Smith, and Sheldon Schultz, Nano Lett., 3, 485 (2003). 157. T.R. Jensen, M.D. Malinsky, C.L. Haynes, and R.P. Van Duyne, J. Phys. Chem. B, 104, 10549–10556 (2000). 158. J.J. Mock, M. Barbic, D.R. Smith, D.A. Schultz, and S. Schultz, J. Chem. Phys., 116, 6755 (2002). 159. C. Sonnichsen, B.M. Reinhard, J. Liphardt, and A.P. Alivisatos, Nature Biotechnol., 23, ¨ 741–745 (2005). 160. B.M. Reinhard, S. Sheikholeslami, A. Mastroianni, A.P. Alivisatos, and J. Liphardt, Proc. Natl. Acad. Sci., USA 104, 2667–2672 (2007). 161. J. Dosta´lek, J. Tyroky´, J. Homola, E. Brynda, M. Skalsky´, P. Nekvindova´, J. Pirkova´c, J. Kvord, and J. Schrofele, Sens. Actuat. B: 76, 8–12 (2001). ¨ 162. J. Homola, Ivo Koudela b, and Sinclair S. Yee, Sens. Actuat. B, 54, 16–24 (1999). 163. J. Homola, S.S. Yee, and G. Gauglitz, Sens. Actuat. B, 54, 3–15 (1999). 164. J. Homola, Anal. Bioanal. Chem., 377, 528–539 (2003) . 165. Carsten Sonnichsen, Bjorn M Reinhard, Jan Liphardt, and A Paul Alivisatos, Nat. ¨ ¨ Biotechnol., 23, 741–745 (2005). 166. Bjorn M. Reinhard, Sassan Sheikholeslami, Alexander Mastroianni, A.P. Alivisatos, ¨ and Jan Liphardt, Use of Plasmon Rulers to Reveal the Dynamics of DNA Bending and Cleavage by Single EcoRV Restriction Enzymes, Proc. Natl. Acad. Sci. USA 104, 2667–2672 (2007). 167. E. Hao and G.C. Schatza, J. Chem. Phys., 120, 1 (2004). 168. H. Xu, J. Aizpurua, M. Kall, and P. Apell, Phys. Rev. E, 62, 4318 (2000). 169. D.A. Genov, A.K. Sarychev, V.M. Shalaev, and Alexander Wei Resonant, Nano Lett., 4, 1 (2004). 170. C.L. Haynes, C.R. Yonzon, X. Zhang, and R.P. Van Duyne, J. Raman Spectrosc., 36, 471–484 (2005). 171. Z. Zhul, T. Zhu, and Z. Liu, Nanotechnology, 15, 357–364 (2004). 172. Zhong-Qun Tian, Bin Ren, and De-Yin Wu, J. Phys. Chem. B,, 106, 37 (2002).

5 Nanosensors

129

173. Martin Moskovits, J. Raman Spectrosc., 36, 485–496,(2005). 174. Katrin Kneipp, Yang Wang, Harald Kneipp, Lev T. Perelman, Irving Itzkan, Ramachandra R. Dasari, and Michael S. Feld, Phys. Rev. Lett.,, 78, 1667 (1997). 175. Shuming Nie and Steven R. Emory, Science,, 275, 1102 (1997). 176. David L. Stokes and Tuan Vo-Dinh, Sens. Actuat. B: Chem., 69, 28–36 (2000). 177. Chanda Ranjit Yonzon, Christy L. Haynes, Xiaoyu Zhang, Joseph T. Walsh, Jr., and Richard P. Van Duyne, Anal. Chem., 76, 78–85 (2004). 178. A. Stuart, Jonathan M. Yuen, Nilam Shah, Olga Lyandres, Chanda R. Yonzon, Matthew R. Glucksberg, Joseph T. Walsh, and Richard P. Van Duyne, Anal. Chem., 78, 7211–7215 (2006). 179. Raoul M. Stockle, Yung Doug Suh, Volker Deckert1, and Renato Zenobi, Chem. Phys. ¨ Lett., 318, 131–136 (2000). 180. Raoul M. Stockle, Yung Doug Suh, Volker Deckert1, and Renato Zenobi, Chem. Phys. ¨ Lett., 318, 131–136 (2000). 181. V. Sazonova, Y. Yaish, H. Ustunel, D. Roundy, T.A. Arias, P.L. McEuen, Nature, 431, 284–287 (2004). 182. G. Abadal, Z.J. Davis, B. Helbo, X. Borrise´, R. Ruiz, A. Boisen, F. Campabadal, J. Esteve, E. Figueras, F. Pe´rez-Muranso, and N. Barniol, Nanotechnology, 12, 1–5 (2001). 183. J. Yang, T. Ono, and M. Esashi, Sens. Actuat. B, 82, 102–107,(2000). 184. A.N. Cleland and M.L. Roukes, Appl. Phys. Lett., 69, 2653–2655 (1996). 185. Z.J. Davis, G. Abadal, B. Helbo, O. Hansen, F. Campabadal, F. Pe´rez-Murano, J. Esteve, E. Figueras, J. Verd, N. Barniol, and A. Boisen, Sens. Actuat. A, 105, 311–319 (2003). 186. A. Boisen, K. Birkelund, O. Hansen, and F. Grey, J. Vac. Sci. Technol. B, 16, 2977–2981 (1998). 187. P. Engels, S. Salewski, H. Levsen, K. Sengstock and W. Ertmer, Appl. Phys. B, 69, 407–412 (1999). 188. A.S. Bell, B. Brezger, U. Drodofsky, S. Nowak, T. Pfau, J. Stuhler,T.H. Schulze, and J. Mlynek, Surf. Sci., 433–435, 40–47 (1999). 189. Z.J. Davis, G. Abadal, O. Kuhn, O. Hansen, F. Grey, and A. Boisen, J. Vac. Sci. Technol. B, 18, 612–616 (2000). 190. G. Abadala, Z.J. Davis, X. Borris!e, O. Hansen, A. Boisen, N. Barniol, F. P!erezMurano, and F. Serra, Ultramicroscopy, 97,127–133 (2003). 191. C. Hagleitner, A. Hierlemann, D. Lange, A. Kummer, N. Kerness, O. Brand, and H. Baltes, Nature, 414, 293–296 (2001). 192. G.U. Lee, L.A. Chrisey, and R.J. Colton, Science, 266, 771–773 (1994). 193. M. Rief, M. Gautel, F. Oesterhelt, J.M. Fernandez, and H.E. Gaub, Science, 276, 1109–112 (1997). 194. J. Fritz, M.K. Baller, H.P. Lang, H. Rothuizen, P.Vettiger, E. Meyer, H.-J. Guntherodt, C.H. Gerber, and J.K. Gimzewski, Science, 288, 316–318 (2000). 195. F. Oesterhelt, D. Oesterhelt, M. Pfeiffer, A. Engel, H.E. Gaub, and D.J. Muller, Science, 288,143–146 (2000). 196. J.A. Harley, and T.W. Kenny, Appl. Phys. Lett., 75, 289–291 (1999). 197. E.L. Florin, V.T. Moy, and H.E. Gaub, Science, 264, 415–417 (1994). 198. G. Villanueva, J. Montserrat, F.P_erez-Murano, G. Rius, and J. Bausells, Microelectr. Eng., 73–74, 480–486 (2004). 199. P.A. Rasmussen, J. Thaysen, O. Hansen,, S.C. Eriksen, and A. Boisen, Ultramicroscopy, 97, 371–376 (2003).

Chapter 6

Quantum Dots for Sensing Javier Goicoechea, Francisco J. Arregui, and Ignacio R. Matias

6.1 Introduction. Quantum Technology and Properties: Quantum Wells, Wires and Dots Quantum confinement has become a powerful tool for creating new materials with extraordinary properties. Since 1980s, the quantum effects on materials have become relevant as far as the scientific community has focused its attention on smaller devices. When certain particle scale is trespassed, quantum confinement effects start to play a relevant role in the macroscopic properties of the matter. Since their beginning, quantum-confined structures have been widely used in optoelectronic device technology rather than in sensor applications. Nevertheless, sensor applications based on quantum dots experiment a real boost thanks to the semiconductor nanocrystals. The possibility of having high-quality, industrially scaled-up, biocompatible quantum dot nanocrystals has supposed a real breakthrough in the biological and medical fields. Quantum dots significantly improve the sensing tools in applications such as cellular assays, cancer detection, or DNA sequencing. This chapter summarizes the state of the art of the use of quantum dots in the sensor field. First of all, it is necessary to give a definition of what quantum confinement is and why it is so important in order to develop new materials with interesting properties. Historically, the scientific community has classified the existing materials based on their composition, and more recently also based on their internal structure. In other words, the external properties of a material were attributed to its basic components (iron, aluminum, etc.), and how these basic units were geometrically and spatially arranged inside the material. This latter consideration reaches its highest importance with the composite materials, which make possible to engineer their final properties simply by controlling

J. Goicoechea Electric and Electronic Engineering Department, Universidad Publica de Navarra, Edificio de los Tejos, Campus Arrosadia, 31006 Pamplona, Navarra, Spain

F.J. Arregui (ed.), Sensors Based on Nanostructured Materials, DOI: 10.1007/978-0-387-77753-5_6, Ó Springer ScienceþBusiness Media, LLC 2009

131

132

J. Goicoechea et al.

the design and fabrication process of the overall material. These composite materials meant a real revolution in the material science some decades ago, especially in fields like aeronautics. In such materials the size did not affect to the macroscopic properties of the overall material, namely, a big piece of iron has the same thermodynamic and photophysical properties than a smaller one. This large-scale word can be described with a high degree of exactness using the Newtonian physics. Nevertheless, the classical description of the world is no longer valid if the speed is increased to very high values, or when the size is decreased until near-atomic scale; this is the nano-scale. Therefore in nano-sized materials it is possible to observe some physical effects derived from the quantum mechanics which are not visible in normal-scaled materials. The prevalence of the quantum confinement effects in small-size materials is one of the motivations of nanotechnology for making things small. The properties change dramatically as their size is reduced down to the quantum scale, opening new possibilities of fabricating a whole new family of artificial materials. The nano-structuration of materials means also a high integration capability and the possibility of self-assembly complex structures starting from simple components. In the following lines there is a basic discussion about the causes and the effects of quantum confinement that can help the reader to understand why quantum dots provide enhanced optical performance. Nevertheless, this section is meant to be accessible for a variety of readers, including non-experts, so only the very basic aspects are discussed here. For further and more detailed explanations the reader can consult any textbook on quantum mechanics for condensed matter physics. It is widely known since Planck stated the very first steps of the quantum theory that the energy is not available in a continuous form. The quantum physics caused a revolution in the conception about the intimate structure of the matter, leading to new atomic models. The quantum atomic model states that the available energy values for electrons are not continuous, but they are quantized. There is a discrete distribution of the possible energy values and positions of an electron inside an atom. Nevertheless, this discrete energy distribution is modified when a high number of atoms are together close enough to interact. This is the case of bulk semiconductors materials. In such cases, the ionic-bounded atom lattice produces a massive overlapping of the atomic energy states, concentrating into a range of energies. In such energy ranges the atomic energy states are so close that they can consider continuous energy bands, as the traditional semiconductor physics have traditionally done. Therefore, it can be assumed that metals, semiconductors, and isolators present continuous bands where the electrons can get any energy value. In the case of bulk semiconductor materials, the band energies are separated by forbidden energy gaps which provide them all their amazing electronic (and optoelectronic) properties.

6 Quantum Dots for Sensing

133

The optical characteristics of these bulk semiconductor materials are determined by the absorption and emission of photons. In semiconductor materials it is very common to speak about electrons and holes when the electrical currents are described. Thanks to their energy band distribution, it is widely known that in those semiconductor materials, a certain amount of external energy (heat, radiation, etc.) can be absorbed by the material (equal to the energy bandgap), promoting one electron from the valence band to the conduction band, and leaving a hole in the valence band. Both electron and holes can be described as conducting corpuscles with different properties which strongly depend on the material. Consequently, the electrons and holes can recombine, leading to the formation of excited states within the semiconducting material, which causes the emission (heat, radiation, etc.) of a certain amount of energy as a consequence of the excited state relaxation. This excitation and relaxation phenomena are schematically displayed in Fig. 6.1a. When an electron and a hole recombine, they experience a strong attractive force. This excited electron–hole pair (also called ‘‘exciton’’) is formed by two different particles with opposite electrostatic charge. It is important to note that the exciton since its creation and until its relaxation can be treated as a corpuscle that physically fills a space within the material, just as electrons and holes do. The exciton size can be calculated using a modified Bohr model yielding the expression shown in Eq. 6.1 [1].

Fig. 6.1 a Schematic plot of the excitation and relaxation processes inside a typical semiconductor material. b The ‘‘particle in a box’’ is the simplest theoretical problem where the quantization effect can be observed. c Representation of an exciton inside a piece of material. When the size of the material is smaller than its Borh radius, the density of states (DOS) of the material is strongly modified due to quantum confinement effects

134

J. Goicoechea et al.

r ¼ "  h2 =p  mr  e2

(6:1)

where r is the radius of the sphere defined by the 3D separation of the electron–hole pair, e is the dielectric constant of the semiconductor, mr is the reduced mass of the electron–hole pair, h is Planck’s constant, and e is the charge on the electron [1]. Therefore, it is clear that the exciton size strongly depends on the physical properties of the particular semiconducting material. Nevertheless, this electron–hole pair separation is around 1–10 nm in most common semiconductors [2, 3]. Regarding the photophysical properties of the semiconductor materials, there is another useful magnitude that gives information about the distribution of the energy transitions between the valence and the conducting bands; this is the density of states (DOS). It represents the number of possible electron–hole transitions that can take place at a given photon energy. Bulk semiconductors have a continuous distribution of the DOS, with a parabolic shape (see Fig. 6.2 left). This means that the probability to find electron–hole transitions increases with the square of the energy of the transition. All the things said previously are valid for bulk semiconductors. However, some unexpected effects appear when the excitons are confined by potential barriers. The simplest case of the 1D confinement is analyzed in a classical quantum physics problem, the ‘‘particle in a box’’. This problem consists in a particle is confined between two infinite potential barriers (see Fig. 6.1b). The Schrodinger equation is used to study the energy of the particle, and it is found ¨ that the possible energy values of the particle are quantized and are also the possible spatial positions of the particle. The quantization in the energy and the existence of forbidden places inside the potential well (spatial nodes) arise naturally from the wave-like nature of the system. The quantized values of the energy of the particle are described by the Eq. 6.2.

Fig. 6.2 Distribution of the density of states (DOS) depending on the confinement degree. As well as the excited states are spatially confined, the quantification of the DOS appears.

6 Quantum Dots for Sensing

135

En ¼

 2 p2 2 h n 2mL2

(6:2)

where n is a natural integer (n = 1, 2, 3, etc.), m is the mass of the particle, h is the reduced Planck constant, L is the quantum well length, and E is the energy. As it can be seen, the energy is quantized (only have discrete possible values). It is also important to note that the length of the potential well may affect the distribution of the energy states of the particle. Consequently, as the size of the quantum well decreases, the allowed energy values are increased. This 1D problem is the simplest example of how the quantum confinement has an influence on the distribution of the energy states and therefore on the electrooptic properties of the materials. A similar effect is observed when the electron–hole pairs are spatially confined in any direction. When this confinement happens, the energy state distribution is spited into a series of subbands, just like in the ‘‘particle in a box’’ problem. As far as the available energies suffer from quantization due to the confinement, the distribution of the available electron–hole transitions is also modified (see Fig. 6.2). Depending on the number of dimensions in which the excitons are confined, it is possible to speak about bulk semiconductors (3D, or no confinement), quantum wells (2D, confinement in one spatial direction), quantum wires (1D, confinement in two spatial directions), and quantum dots (‘‘0D’’, three spatial directions confined semiconductors), as it is displayed in Fig. 6.2. As it can be seen, the shape of the distributions of the available state for electrons and holes changes drastically with the confinement degree. One of the most interesting characteristics of the quantum-confined structures is that there is a pileup of the DOS at the band edge (see Fig. 6.2). This means that more electron–hole transitions are likely to happen in the band edge than in bulk semiconductors. Particularly, in the case of QDs, the DOS distribution remains in the discrete energy distribution of single atoms; that is why sometimes some workers refer them as ‘‘artificial atoms’’. The pileup effect in the distribution of the available transitions leads to narrower and more intense emission peaks in optoelectronic devices. There are also other benefits of the confinement in the case of QDs, ideal for optical performance, as it is summarized in the next lines:

 Narrower and more intense emission peaks.  DOS pile up near the band edge. Consequently, more transitions can contribute to the optical response at the same energy.

 Displacement of the allowed energies due to the confinement. Controlling the confinement conditions it is possible to tune the bandgap properties of the QDs.  As a consequence of level splitting, relaxation processes are slowed, leading to an increasing of the fluorescence lifetime.  Controlling some confinement parameters it is possible to achieve some control over the polarization of the emitted light.

136

J. Goicoechea et al.

 The exciton stability is increased due to the confinement effects that contribute with the coulombian attraction in keeping the electron–hole pair stable. All these properties make QDs ideal for optoelectronic applications, where they show an outstanding performance. Initially the quantum confinement technology has been more used in the design and fabrication of more efficient and better optoelectronic devices, having no relevant importance in sensing applications, Nevertheless, the appearance of QD nanocrystal colloidal suspensions, due to their surprisingly good photophysical properties and biocompatibility, have supposed a real breakthrough in the biosensing applications, boosting the power of most of the traditional biology measurement tools, and even reaching to medical diagnosis applications.

6.2 Synthesis of Quantum Dots As it has been introduced in previous paragraphs, the fabrication of QD structures consists in the creation of a low-bandgap spatial region smaller than the Bohr radius of the exciton in this particular medium, confined by higher potential barriers in the confinement directions. There are different techniques for fabricating quantum dot structures depending on the material and also on the confinement strategy. It is possible to classify those approaches mainly in two different categories; the QDs made by semiconductor growth techniques onto wafers and the synthesis of QD nanocrystal colloidal suspensions. In the first case, the QDs are created during the fabrication growth process of semiconductor devices onto silicon wafers. In the semiconductor device industry there are several well-established techniques such as Molecular Beam Epytaxy (MBE) and Metal-Organic Chemical Vapor Deposition (MOCVD) that have been used for decades in the fabrication of microelectronic devices. These techniques make possible to fabricate multilayer semiconductor devices, with a sub-nanometric control in the thickness of the different layers. This made possible to create quantum well (QW) structures since 1970s, and consequently to apply them at industrial scale in the optoelectronic device fabrication. However the impact of such structures in the sensor field has been very limited, having more repercussion in the electronic and optoelectronic device fabrication (Lasers, photodiodes, etc.). The second technology for creating QD-structured semiconductors is the nanocrystal synthesis. In this approach the semiconductor QDs are not embedded into multilayer structures grown onto silicon wafers but they consist in stand-alone particles of nanometric size, dispersed into a colloidal suspension. In this case the quantum confinement is achieved due to the abrupt difference between the bandgap structure of the semiconductor core and the isolating character of the surrounding dissolving medium, which acts as almostinfinite potential barriers. The main advantage of this approach is that the last synthesis routes do not require such complicated and expensive equipment (like

6 Quantum Dots for Sensing

137

MBE or MOCVD); it is possible to fabricate precisely controlled nanocrystals with industrially scaled processes at low price. Those QD nanocrystals are very versatile from the application point of view due to their photophysical properties and biochemical functionalization potential. In the last decades researchers have reported a huge number of QD nanocrystal-based sensing and imaging applications in the biomedicine field, for instance immunoassays, DNA tracking, cancer detection, etc. In the next subsections there is a detailed description of the most relevant techniques used for QD synthesis: the semiconductor QDs growth onto silicon wafers and the colloidal QD nanocrystal approaches.

6.2.1 QD Growth onto Semiconductor Wafers In the late 1970s and early 1980s the rapid development of planar growth techniques like Molecular Beam Epitaxy (MBE) and Metal-Organic Chemical Vapor Deposition (MOCVD) led to high-quality-layered semiconductor structures with abrupt interfaces. Both techniques give the designer the ability to control in a very precise way the thickness of each layer. At heterojunctions electrons are confined to move into thin 2D depletion layers. This opened the way for creating high-quality quantum well structures (QW). In such semiconducting structures, the electron energies and thresholds for optical excitation can be tailored, via confinement, simply by varying the thickness of the QW layer. As a consequence, QW structures have been successfully used in optoelectronic semiconductor devices like lasers or photodiodes, as far as they allow tuning the band-profile of the device, just by adjusting the thickness of several semiconductor layers. However, the fabrication of quantum wires and dots is not so easy since it implies an additional difficulty as far as it requires a very high resolution also in the plane directions. In order to achieve this coplanar patterning, lithographic techniques are needed, therefore a tremendous effort was devoted in the late 1980s and the beginning of 1990s in the development of quantum dot structures using this approach [4]. Unfortunately, the ability of making such small figures with lithographic techniques has a limit, and most of the times this limit is too high. The last advances in nanolithography techniques, like extreme ultraviolet lithography (EUVL) or X-ray lithography (XRL), light coupling nanolithography (LCN), nanoimprint lithography (NPL), or dual-beam FIB-SEM systems, may have enough resolution to fabricate QD-sized figures. However, even these ultra high definition techniques would be at their limit, as far as the figure size of the desired QDs should be below the Bohr radius of the exciton in the material, what means around 10 nm for GaAs. A simpler approach for creating QDs than the direct lithography consists in first growing a larger structure and afterwards reducing it in size. However this top-down approach met with limited success as far as it was almost impossible to make useable structures that were small enough. However, there are other

138

J. Goicoechea et al.

fabrication strategies different from the top-down approach and from the lithographic techniques. Those bottom-up techniques start from the well-established QW structure, and they introduce a modification on the geometry of the multilayer structure. This geometric modification is designed in order to achieve tiny local lower bandgap regions embedded into higher bandgap semiconductor material, and therefore create the quantum confinement conditions. Most of the bottom-up techniques derive from the well-known QW fabrication techniques, introducing some modifications in order to achieve QD confinement regions. The most important techniques are the self-assembly QD growing (Stranski–Krastanow technique), V-groove growing techniques, cleaved edge overgrowth, interfacial QDs, etc. The Stranski–Krastanow (SK) technique uses the well-known epitaxial growing techniques commonly used in QW fabrication, mainly Metal-Organic Vapor Phase Epitaxy (MOVPE) and Chemical Beam Epitaxy (CBE). This approach takes advantage of the formation of 3D islands onto a semiconductor substrate driven by the existence of a lattice misfit between the epilayer and the substrate (see Fig. 6.3a and 6.3b). This lattice mismatch gives the thermodynamic force for the transition from 2D planar epitaxial growth of the epilayer to growth of the

Fig. 6.3 Self-assembled QDs grown onto semiconductor wafers. a The self-assembled QDs (SK) derive form the QW fabrication techniques. b Cross-sectional TEM image of two InGaAs and Ga(SbAs) QDs layers. c A detail from b where it is possible to see a single QD using HRTEM Reprinted from [5], with permission from Elsevier

6 Quantum Dots for Sensing

139

epilayer as 3D islands. In Fig. 6.3c it is displayed a HRTEM micrograph of a self-assembled QD created using the SK technique [5]. One of the main drawbacks of this technique is that the in-plane position of the QDs cannot be controlled, as they grow spontaneously. The fact of using conventional semiconductor growing techniques makes this approach easier and cost effective for mass production. Figure. 6.4 shows an example of patterned growth QDs [6]. In this approach the difference in the radii of curvature of two alternate layers of GaN and In0.02Ga0.98N (higher bandgap materials) is used to create a small quantum confinement cavity (QD) in the peak of a pyramid-shaped pattern. Inside these higher bandgap barriers it is a sandwiched InGaN layer (a lower bandgap material), where the excitons are confined. In the past years QDs have kept the interest of lots of research groups worldwide. In the traditional semiconductor industry, QD structures, with their atom-like density of states distribution, are especially interesting in order to improve the efficiency of optoelectronic devices. There are currently several commertial applications, for example in LEDs, VCSELs (vertical cavity surface emitting LASER), etc.

Fig. 6.4 Patterned growth of QDs. a Schematic plot where it is showed that this technique is an evolution of the QW fabrication techniques. b Scheme of the layer sequence of the QD fabrication process. c SEM image of one of the patterned pyramids. The QD is inside the peak of the pyramid c and d are reprinted from [6], with permission from Elsevier

140

J. Goicoechea et al.

6.2.2 Nanocrystal Synthesis Semiconductor nanocrystals consist of almost spherical pieces of semiconductor crystal which may have a diameter around 2–10 nm. They grow spontaneously around a seed point and form a crystalline structure in three dimensions. When the crystal growing reaction is stopped, almost spherical nano-sized crystals are achieved as it is shown in Fig. 6.5. The semiconductor nanocrystals can be synthesized using a wide variety of materials. Nevertheless, since the very first synthesis studies one of the most used materials choices for QD synthesis was group II-VI semiconductors. More precisely the Cd-chalcogenide semiconductors have been traditionally the most popular approaches to fabricating QD nanocrystals due to the availability of the precursors and the easiness of high-quality crystallization routes. However, it is possible to use other materials for the fabrication of QDs, for example the III–V group semiconductors [7, 8] like InAs or GaAs. The material election in the nanocrystals fabrication has an impact into their bandgap properties, allowing to tune very precisely their emission color, or also to overcome some Cd-chalcogenide QD problems like Cd cytotoxicity [9, 10, 11, 12]. The emissive properties of QD nanocrystals can be modified by changing their composition or their size. As a consequence a whole family of fluorescent materials whose emission peak wavelengths can sweep the entire VIS-NIR spectrum (see Fig. 6.6). The emission color tailoring together with their narrow emission spectrum opens the door to color multiplexed measuring applications. In the 1990s some groups started to develop routes for synthesizing semiconductor QDs. In the very first works dry approaches were used. Koyama et al. in 1992 [16] used Nd:YAG-pulsed laser ablation in argon gas to achieve the nanocrystal formation. A different approach was performed by Bawendi’s group in 1994, reporting a technique based on electrospray of the precursors

Fig. 6.5 a Schematic plot of the structure of a CdSe QD nanocrystal. b HRTEM image of a single CdSe nanocrystal where the semiconductor lattice is easily visible. Reprinted from [13] with permission from the ACS publishing group

6 Quantum Dots for Sensing

141

Fig. 6.6 a Ten distinguishable emission colors of ZnS-capped CdSe QDs excited with a nearUV lamp. From left to right (blue to red), the emission maxima are located at 443, 473, 481, 500, 518, 543, 565, 587, 610, and 655 nm Reprinted from [14], with permission from Nature Publishing Group b Size- and material-dependent emission spectra of several semiconductor nanocrystals in a variety of sizes. The right-hand series represent different sizes of CdSe nanocrystals (2.1, 2.4, 3.1, 3.6, and 4.6 nm from right to left). The central series is of InP nanocrystals (3.0, 3.5, and 4.6 nm). The left-hand series represent InAs nanocrystals (2.8, 3.6, 4.6, and 6.0 nm from left to right) Reprinted from [15] with permission from AAAS

[17]. Those first attempts were eclipsed by new wet synthesis routes in the following years. The new wet techniques avoided the expensive and sophisticated equipment requirements (high vacuum need, etc.) and made possible the apparition of new applications for QD nanocrystals. The first wet approaches to nanocrystal synthesis consisted in routes which used methyl alcohol or water as solvents, and amines [18, 19] or polyphosphates [20] as coordinating groups. Those techniques were much simpler than the previous approaches, but they had several drawbacks, for example the high size dispersion and low emission efficiencies of the synthesized nanocrystals. A breakthrough paper from Murray et al. [21] described a new wet route for the synthesis of Cd-chalcogenide nanocrystals starting from dimethyl cadmium Cd(CH3)2, using organic-coordinating solvents at high temperatures. The most common coordinating solvents used were trioctylphosphine oxide (TOPO), trioctylphosphine (TOP), and hexadecylamine. These solvents, apart from being the supporting medium for the chemical reaction to occur, acted as surfactants, capping the nanocrystals and stabilizing their surface, consequently preventing the QD nanocrystals from aggregation. This surface stabilization

142

J. Goicoechea et al.

also leads to higher fluorescence quantum efficiencies, as it will be described in detail later. The QDs produced using this method showed a very high quality, showing very monodispersed size distribution and high quantum yields. The main drawback was the technical hazards of the production process since the main chemical precursor, the Cd(CH3)2 is toxic, unstable, explosive, and very expensive. The critical experimental conditions needed in this synthesis route made it unsuitable for large-scale production. In a very important work [22], Peng and coworkers selected CdO as organometallic precursor instead of Cd(CH3)2, also using coordinating solvents at high temperatures under an Argon inert atmosphere. The quality and size dispersion of the QD nanocrystals were even superior to the previously reported synthesis routes, with the additional advantage that less dangerous experimental conditions are required since CdO is far more stable than dimethyl cadmium. This breakthrough paper described a reproducible way to fabricate high-quality Cd-chalcogenide QD nanocrystals using mild and simple experimental conditions, opening the door for scale up the synthesis to industrial production of the QD nanocrystals. Several works have been reported in which QD nanocrystals were synthesized, and most of them are based on the previously described Cd-chalcogenide route using organic-coordinating solvents at high temperatures [23, 24, 25]. In the recent years, other alternative approaches for the QD have been investigated. Ludolph [26] and Crouch [27] used a new approach starting from stable non-air-sensitive precursor based on selenocarbamate derivatives of Zn or Cd, and an air-stable complex of Cd and imino-bis(diisopropylphosphine selenide) respectively, achieving similar QD quality as previously reported routes. It is very common that during the nanocrystal growth-quenching process some lattice defects may occur in the surface of the new nanoparticle (generally surface atoms that are missing at least one chemical bond). The effect of those defects in the photophysical properties of the QDs is the dramatic reduction of the fluorescence quantum yield due to the non-radiative annihilation of the excited electron–hole pairs (also called ‘‘excitons’’) in these ‘‘trap’’ sites or defects. However, it is possible to passivate such trap sites by simply reconstructing these loose bonds. This can be done by coating the QD nanoparticles with an outer layer of another semiconductor material. This topology is called core/shell quantum dots, as far as they consist in an inner core consisting of the active QD nanocrystal and an outer semiconductor coating that plays two simultaneous roles: passivate the surface traps of the core and protect the active core from external aggressive agents. The use of this core-shell structure results in much more emissive and more photostable QD nanocrystals.

6.2.3 Core-Shell Nanocrystal Quantum Dots As it has been previously introduced, it is very common to find lattice defects in the QD surface. Those defects generate local variations in the bandgap of the

6 Quantum Dots for Sensing

143

Fig. 6.7 Trap site nonradiative recombination mechanism. This effect contributes to a dramatic decrease in the fluorescence quantum yield of the nanocrystals

materials, acting as near-band trap sites, as it is shown in Fig. 6.7.The existence of those traps causes an increase in the quenching probability of the excited states, and therefore they contribute to a dramatic decrease in the emission efficiency. Nevertheless, it is possible to passivate those trap sites simply by coating the active QD cores with a different semiconducting material. The physical properties of the QD coating layer (also called shell) have a dramatic impact on the photophysical properties of the QDs. Therefore it is critical to choose a larger bandgap semiconducting material which can complete the surface loose bonds (passivating the traps), keeping, at the same time, the same confinement conditions into the core of the nanocrystal. It has been demonstrated since the very first studies [17] that capping the nanocrystal core with a shell of an inorganic larger bandgap semiconductor (e.g. ZnS) reduces this defect formation, and simultaneously contributes to the quantum confinement as far as it also acts as a potential barrier for the core excitons [28, 29]. As a result, the quantum yield is considerably enhanced. Some groups have reported that CdSe core-shell QDs are about 20 times brighter than single rhodamine 6G molecules [30]. This shell also plays a protective role against external aggressive agents. That is why the core-shell configuration is more robust to photobleaching, pH changes, or other potentially aggressive media. The photodegradation (or photobleaching) of the traditional organic dye fluorophores consists in a photoinduced oxidative reaction of some of the conjugated bounds of the organic molecule, leading to the destruction of the photophysical properties of the molecule. That degradation reaction requires the combination of the fluorophores and the reacting agent, leading to a chemical reaction triggered by the external light. In core-shell QDs, the capping layer acts a physical barrier between the environment and the

144

J. Goicoechea et al.

Fig. 6.8 Photostability comparison between the CdSe/ZnS QDs and the organic fluorophore Alexa 488 in 3T3 cells. a Top row: The nucleus was labeled with streptavidin-conjugated QDs (red), and the microtubules were labeled using anti-mouse IgG Alexa 488 (green). Bottom row: Microtubules were labeled with QDs (red), and nuclear antigens were stained green with Alexa 488 conjugated to anti-human IgG. Whereas labeling signals of Alexa 488 faded quickly and became undetectable within 2 min, the signals from QDs showed no obvious change within the experiment duration. b Time evolution of the fluorescence intensity of QDs and Alexa 488 in 3T3 cells cultivated in standard and in antifading media, using the same experimental conditions as in (a) From [31]. Reprinted with permission from Nature Publishing Group

photoluminescent core, significantly improving the phostability of the QDs. Therefore it has been demonstrated that core-shell QDs shows bright fluorescence activity up to hours under continuous illumination, compared to the few minutes that traditional organic dye fluorophores. In Fig. 6.8 it is shown the behavior of CdSe/ZnS core/shell QDs compared with a conventional fluorescein labeled sample using fluorescence microscopy [15]. As result of their discrete, atom-like electronic structure, QDs have typically very narrow emission spectra with FWHM of the luminescent emission of around 15–40 nm. Since the emission lines are comparatively much narrower that those of organic dyes, detection of the QDs suffers much less form crosstalk that might result from the emission of a different fluorophore [25]. There are other advantages of QD nanocrystals compared to the conventional organic dye fluorophores. The main and most interesting photophysical advantages are listed below:

6 Quantum Dots for Sensing

145

 High fluorescence quantum efficiency.  Narrow emission peak with no red tailing.  Wide absorption band that allows multiple color QDs excitation with only 



 

one light source. This is very useful to achieve high-quality color multiplexed imaging when using standard fluorescence applicationsLong fluorescent lifetimes (15–20 ns). This is one order of magnitude greater than most of fluorescent dyes. This permits the QDs to be distinguished from other intrinsically fluorescence background sources, with the appropriate apparatus, increasing the signal-to-noise ratio by several orders of magnitude). High extinction coefficient (0.5–5 106 M-1cm-1). This means that more excitation photons are absorbed. Combined with the high quantum yield, the overall emission with the same excitation power results to be considerably higher. High photostability (see Fig. 6.8). Color tunability by controlling the size of the nanocrystals. Currently there is a wide variety of QDs that covers the whole VIS–NIR spectrum (see Fig. 6.6).

All these properties make QD nanocrystals better fluorophores than the traditional fluorescent dyes, and therefore they had outperformed them in most of the biological and biomedical applications. However, the described synthesis routes for normal and core-shell QDs, which intrinsically produce highly hydrophobic nanocrystals, as far as they are coated by a trioctylphosphine oxide/trioctylphosphine (TOP/TOPO) native layer. Since most of the sensing applications of QD nanocrystals are in the bio field, water soluble QDs are needed. Therefore, an additional step has to be carried out for removing or modifying the native hydrophobic coating to turn it into hydrophilic. Several approaches have been developed in order to change the surface properties of the QDs, including a wide variety of functionalization strategies in order to provide the nanocrystals selectiveness, or smart behavior. These surface modification strategies will be commented in the following section.

6.2.4 Functionalization of Nanocrystal Quantum Dots One of the main keys of the success of QDs in a lot of scientific disciplines, together with photostability and efficiency, is precisely their versatility. The study of the surface chemistry of semiconductor nanocrystals has produced a whole family of high-performance fluorophores able to be present in diverse media via QD coating. The first strategy consists in replacing the TOPO native ligands by thiolanchored hydrophilic moieties (mercaptans) like hydroxyl, carbonyl, etc. [24]. It has been widely reported the use of mercaptoalcohols and meracptoacids as stabilizers [32]. A schematic plot of this functionalization strategy is shown in Fig. 6.9a. This approach takes advantage of the strong affinity between the thiol

146

J. Goicoechea et al.

Fig. 6.9 Methods of conjugation of biomolecules to QDs. a) shows the thiol bound formation between S atoms on the QD surface and S atoms in molecules terminated with carboxylic acid, amine, or hydroxyl groups. b) shows the formation of a complex between a Zn atom on the QD surface and polyhistidine residues by metal-affinity coordination. c) shows the utilization of a silica overcoat which makes possible the utilization of conventional silanization techniques to functionalize the QDs. d) shows the utilization of native TOPO ligands to functionalize the QDs using amphiphilic molecules. e) shows the functionalization of QDs using the streptavidin-biotin affinity. f) shows electrostatic interactions between QDs surfaces and oppositely charged biomolecules

group and the surface of the QDs. The main disadvantage of this approach is that QD functionalized in that way is only stable at basic pH values, hence reducing the range of potential applications of such functional probes. QDs conjugated with carboxylic terminated mercaptans can be attached to the amino group of a secondary biomolecule via EDC conjugation [33, 34, 35], increasing the number of possibilities for bio-functionalization. Another strong binding group to QD surface is the polyhistidines. In this case the attachment of the external molecule to the QD surface is done by metal complexation of polyhistidine with the Zn atoms of the QD (see Fig. 6.9b). In other works [36, 37, 38, 39, 40] some authors have reported the use of thiolated peptides or polyhistidine-containing proteins in order to achieve a direct attachment of the proteins to the QD surface. In a different approach, an additional coating over the core-shell QD configuration is used (see Fig. 6.9c). This coating is made by polymerized functional silica (SiO2) [41]. These resulting nanoparticles are more stable; it is possible to use them in a wider range of pH than when functionalized with mercaptans. Furthermore, as far as the silica chemistry is well established, it also allows further functionalization steps leading to stable bioconjugated QDs.

6 Quantum Dots for Sensing

147

Another alternative, instead of removing the native surface TOPO ligands, is to overcoat them with amphiphilic diblock or triblock copolymers or by phospholipids [31, 42, 43,], as it is shown in Fig. 6.9d. These QD conjugates show a higher biocompatibility and photostability, and less toxicity in biological media, as far as the amphiphilic coating acts as isolation between the QD and the environment, reducing their interaction. The main drawback of this approach is that it increases the overall particle size, making it unsuitable for several specific bioapplications (for example, extravasation processes, etc.). One of the most used approaches for bio-functionalization is based on the strong bonding between biotin and streptavidin (see Fig. 6.9e). Biotin also called vitamin H or B7 shows an extraordinary affinity for the streptavidin, which is a small protein. As a consequence, QD-streptavidin complexes can be used as bioprobes for selective binding biotinylated targets [44, 45]. The wide number of applications of this approach have been potentiated by the low cost of the QD-streptavidin conjugates (they are already commercially available). However, there are also problems with the large overall size of the particles (15–20 nm) which make them unsuitable for some specific biologic applications, and also concerning the immunogenicity of streptavidin [46], which restricts their in vivo applications. Based on this approach some groups have reported a three-layer scheme [47, 48], using a primary antibody versus a target molecule attached to a secondary biotinylated antibody, which at the same time is linked to a QD-streptavidin complex. This latter strategy is not only limited to antibodies, and it can be used with other kind of molecules. Finally Mattoussi’s group have reported the non-covalent adsorption of selfassembled engineered proteins [49, 50, 51]. The electrostatic attraction between the engineered proteins and the QDs is enough to keep the two species attached, as well as it allows the proteins to retain their normal activity [50]. Not only proteins have been conjugated with QDs but other biomolecules like oligonucleotides, short DNA sequences, or antibodies. This approach is schematically shown in Fig. 6.9 f. As it has been previously commented that all the functional capping layers have an additional effect as far as they act as physical protective barrier of the QD surface. On the other hand these coatings prevent the environment from negative effects coming from the QDs, such as cytotoxicity. Some works have demonstrated that PEG functionalization of the QD surface derives in a higher biocompatibility, photostability, and also it reduces their cytotoxicity and their accumulation in organs like the liver, etc.

6.3 Sensing Applications Semiconductor QDs have contributed not only in a quantitative but also in a qualitative way in the improvement of the techniques used in a wide range of scientific fields. This QD technology has supposed a real improvement in the design and fabrication of optoelectronic devices such as lasers or photodetectors,

148

J. Goicoechea et al.

leading to more efficient devices. However, the real impact of QDs in the sensing applications field came with the semiconductor nanocrystals synthesis. These QD nanocrystals have a very high performance as fluorophores, mainly due to the fact that they have higher efficiencies and suffer much less from photobleaching. As the photochemistry of these new materials has advanced, a lot of unexpected applications have appeared in which the fluorescence from QD can be modulated. Thanks to phenomena like the surface chemistry modification, the energy transfer mechanisms or the conjugation of QD with biomolecules have made possible the bursting of QD in the sensing applications. It is clear that QD nanocrystals have gone beyond their initial expectations in sensing, and they have caused a revolution in applications such as immunoassays, DNA tracking, or biomedical applications such as cancer detection.

6.3.1 Ion Indicators As it was discussed in Sec. 6.2.3 the emissive properties of QDs are dramatically influenced by the status of their surface [52]. The excited electron–hole pairs (excitons) created in the core of the QD are spatially confined by the size of the semiconductor nanoparticle. The presence of any surface charge or any chemical or physical ligand in the QD may cause the creation of near-bandgap trap sites and it obviously can affect the fluorescence emission efficiency. It has been demonstrated that the presence of some metallic ions in a solution containing QDs can affect in different ways their fluorescence quantum yield, setting the basis for several ion detection optical sensing applications. The traditional techniques for ion detection in water solutions usually involved techniques such as chemiluminescence, voltammetry, furnace atomic absorption spectroscopy, inductively coupled plasma mass spectroscopy, fluorescence based on organic dyes, etc. [25]. Additionally, QDs overcome most of the issues of the organic fluorophores, as photostability, low signal intensity, narrow excitation bands, and wide emission spectra with red tailing, therefore QDs are a good alternatively for simple and easy optical detection of metallic ions. These ion sensitive devices have a great relevance in environmental applications as far as the presence of heavy metal ions in aqueous solutions is a critical parameter for biological systems. Those heavy metal ions produce a dramatic negative impact in the trophic networks as far as they are very toxic agents for algae, bacteria, etc. These metallic cations may also cause negative effects on human health even at very low concentrations, showing in some cases (Hg, Pb, etc.) a cumulative effect in mammals that may cause serious illnesses [53, 54]. It is possible to classify the mechanisms of fluorescence variation of the QDs depending on the nature of the phenomena that cause this variation, as it is shown in Fig. 6.10. First, some groups have reported the enhancement of the fluorescent signal from the QDs due to the presence of some ions. Sphanel and

6 Quantum Dots for Sensing

149

Fig. 6.10 Mechanisms of sensing properties of QDs. (1) Enhancement of luminescence: a passivation of surface trap states by adsorption of Zn2+ or Cd2+ on QDs surface; b formation a QDs network with cysteine–Zn complexes, activating the surface states. (2) Quenching of the luminescence: a electron transfer from the dopamine to the QDs; b non-radiative recombination, facilitated by Ag2Se formation on the surface of QDs, displacing Cd2+; c combination of electron transfer and non-radiative recombination, the reduction of Cu2+ by thioglyceriol forms CdS+Cu+ quenches by facilitating non-radiative recombination, d inner filter caused by a strong absorption of Fe3+ Reprinted from [58] with permission from Taylor & Francis

coworkers have reported a pioneering work [20] where adding Cd ions in a basic medium to CdS QDs significantly improved the emission quantum yield (see Fig. 6.10a). This phenomenon is related with the formation of Cd(OH)2 that passivated the surface traps, and therefore reducing the exciton non-radiative annihilation. A similar enhancement of fluorescence was demonstrated by Chen and Rosenzweig [55] when CdS QDs functionalized with l-cysteine was exposed to Zn2+ ions (see Fig 6.10b). The emission enhancement was justified by the activation of the QD surface states due to the formation of Zn–cysteine complexes, and additionally the optical response was found to be selective for Zn2+ ions. A similar effect was reported by other groups [56]. In 2005 Chen et al. [57] demonstrated another application based on CdS-l–cysteine conjugate for Ag+ trace detector with a detection limit as low as 5 nm. These last two works, [55] and [57], are based on the high affinity of the thiol group from the cysteine to

150

J. Goicoechea et al.

some metallic cations. The cation forms a complex with the thiol group, and at the same time it is bound to the QD surface. Then, an active surface state is formed, enhancing the fluorescence emission from the QD. Nevertheless the most common sensing mechanism is the fluorescence quenching instead of enhancement. The reduction of the fluorescent signal from the QDs can be caused by several phenomena, namely, ion-binding interactions, non-radiative recombination pathways, electron transfer processes, and inner filter effect (see Fig. 6.10c, d, e, and f ). The ion-binding of external chemical species to the QD surface causes the formation of exciton annihilation traps, and consequently a reduction of the fluorescence emission. Isarov et al. [59] described in the first works demonstrating the use of QDs as selective ion probes. The addition of Cu2+ ions to a CdS QD solution in 2-propanol caused a ionic binding of the Cu2+ to the QD surface, producing near-bandgap trap sites and therefore quenching the emission from the CdS QDs. Liang et al. [60] reported an optical sensor for detection of Ag+ based on the fluorescence quenching due to ion-binding mechanisms. It was found that Ag+ strongly bound to the CdSe QDs surface forming Ag2Se, displacing Cd2+ by Ag+ and acting as electron–hole non-radiative recombination centers. Selective Cu2+ probes were reported by Xie et al. [61], where CdSe/ZnS core-shell QDs were used. Those QDs were surface modified with bovine serum albumin (BSA), and the water-soluble nanoparticles were sensitive and selective to Cu2+ ions, reaching a sensitivity of 10 nm. The fluorescence quenching is explained in terms of chemical binding of Cu2+ ions to the CdSe surface, and by a chemical displacement of the Cd2+ ions from the crystalline structure, leading to the formation of CuSe in the surface of QDs. Those CuSe sites in the surface of the QDs facilitate the non-radiative annihilation of the exciton, leading to a chemically induced fluorescence quenching. In a more recent work, Bo et al. [62] describe a route for synthesizing CdTe QDs for Cu(II) tracing based on the fluorescent quenching of the nanocrystals. CdTe QD nanocrystals were synthesized and modified by 3-mercaptopropionic acid to make them water soluble. In this work a fluorescence quenching of the QD emission was demonstrated due to ionic binding of the Cu2+ ions to the QDs surface, showing fluorescence recovery with the addition of EDTA (ethylendiaminotetraacetic acid), a powerful chelating agent for di- and trivalent metallic ions. A similar approach was followed by Fernandez-Argu¨elles et al. [63] where a sulphonic-modified CdSe QDs synthesis for Cu2+ sensing in aqueous solutions based on fluorescence quenching was described (see Fig. 6.11). In this work, the cross-sensitivity with other metallic cations was tested, resulting in a good selectivity of the probes to Cu2+ ions. Real water samples were finally measured with the sulphonic-substituted QDs showing good accordance with differential pulse anodic stripping voltammetry (DPASV) measurements. Gattas-Asfura and Leblanc [64] synthesized a peptide-coated CdS QD for optical detection of Cu2+ and Ag+. However, the CdS QDs described in this work showed wider emission peaks, indicating the presence of a high number of surface defects.

6 Quantum Dots for Sensing

151

Fig. 6.11 Effect of the addition of trace amounts of copper on sulphonic-capped CdSe QDs: The fluorescence is reduced as the concentration of Cu2+ ions is increased. This quenching mechanism is explained in terms of non-radiative recombination in the CuSe sites created in the QD surface (see Fig. 6.10-2b) Reproduced from [63] with permission from Elsevier

In other work, Li et al. reports the sensing properties of CdTe thiol-capped QDs, which have been tested against several metallic ions in aqueous solution. The effect was different depending on the nature of the metallic ion, for example an enhancement of the fluorescence was reported for Zn2+, but a strong fluorescence quenching was found for Mn, Ni, and Cd. The underlying mechanisms are not well defined but likely there might be a selective combination of Zn activation of surface emitting states and other cationic ion binding with exciton annihilation sites generation. Larkowicz [65] reported a iodide sensor based on the fluorescence quenching of polyphosphate stabilized CdS QDs. The sensitive mechanisms were explained in terms of non-radiative recombination, by electron transfer mechanisms and by inner filter effects. Some approaches have been published for cyanide ion sensor devices. These applications have a special relevance due to the huge toxicity of the CNfunctional group. Sarkar [66] described a sensor based on variations in the absorption spectrum of the QDs rather than in their emissive properties. CdSe and CdS QDs were used stabilized with potassium nitrilotriacetate and sodium selenosulfate for the first case, and thiourea for the second QDs. The variation of the absorption spectra of the QDs was explained by a strong reversible adsorption of CN- ions on the QDs surface. The absorption band shift was caused by the increased localization of the excitons inside the QD due to the compression of the electron wave function by the adsorption of the CN- groups. A different approach for cyanide ion detection was reported by Jin [67]. In this work, the authors presented CdSe QDs modified by 2-mercaptoheptane sulfonate which provided a selective CN sensitivity in aqueous solutions, reaching detection limits as low as 1.1 uM. It was also found that the addition of some surfactants (dodecyltrimethylamonium chloride, DTAC) significantly

152

J. Goicoechea et al.

improved the QD stability. This group also reported that a previous photoactivation step was performed prior to making the measurements. This photoactivation consisted of an intense radiation with solar light or a 400 nm excitation for a controlled period of time. This resulted in an increase of the measured luminescence combined with a blue shift of the maximum emission spectrum. The authors justify this increase in the fluorescence yield by the photoinduced passivation of surface traps, eliminating some of the near-bandgap trap sites that produce the non-radiative recombination mechanisms. Chen [68] in 2006 found an interesting sensitive mechanism using two different sized CdSe/ZnS QD species functionalized with 15-crown-5 (1,4,7,10,13pentaoxacyclopentadecane). The functional ligand 15-crown-5 shows a strong ‘‘host–guest’’ potential, with a high-chelating power to alkali metals. When K+ ions are present in the solution, they form complexes with the chelating ligands, bringing closer the different size quantum dots. If the QDs are close enough, close dipole-to-dipole interactions may occur, causing the energy transference via Forster resonance energy transfer (FRET) between the excited states of the ¨ smaller QDs (donors) and the bigger ones (acceptors). The output signal is the ratiometric measurement of the fluorescence emission from the two QD species. These techniques suppose good approach for optical ion sensing applications which can be an easy, fast, and inexpensive alternative to the conventional techniques. Nevertheless, their main drawback is their cross-sensitivity to experimental conditions: Most of the times they need a very specific pH, concentration of the analytes, etc. In other words, it can be said that their weak spot is their robustness. However, some approaches have reported good measurement in real water samples, showing a good match with conventional techniques measurements [63].

6.3.2 In vitro Biological Applications When in a particular sensing application, the magnitude of interest (or target) is part of a biological system, the measuring environment conditions are very exigent for the measuring technique. This is due to the presence of a lot of interfering factors that may alter the measuring signal, introducing a significant error. This is why when the sensing mechanism involves a very delicate chemical reaction, or when a very high accuracy is required, in vitro measurements are performed. In vitro measuring means that the measuring takes place in a controlled environment outside of the living organism. This control of the measuring environment together with the possibility of using techniques unsuitable in living organisms, make in vitro measuring a great tool for the biomedical analysis and diagnostics. Applications as immunoassays, DNA tracking, or cytology are widely used by the biological and medical communities, and the application of QDs technology have supposed a new powerful tool in these areas, as it will be discussed below.

6 Quantum Dots for Sensing

153

6.3.2.1 Immunoassays The immunochemistry is the part of the biochemistry which takes advantage from the specific bonding reactions between antigens and antibodies. Therefore the immunoassays can be defined as biochemical tests, generally performed in a physiologic fluid, and based on antibody–antigen interactions in order to detect selectively the presence of a target biomolecule. There are also other antibodybased detection approaches using QDs, but because they are focused on applications such as cell labeling or cancerous tissue fluorescent labeling, they will be commented in subsequent sections. Traditionally, there are some wellestablished techniques in immunochemistry, like immunoperoxidase or immunofluorescence staining techniques, the Western blot (WB) technique, and the enzyme-linked immunosorbent assay (ELISA), among others. For instance, this latter technique has been widely used as a diagnosis tool and food quality control during the past years due to its reliability and high throughput. Roughly, it is possible to distinguish two main different stages in immunoassays; a first stage (recognition) where the target biomolecule is probed with its specific antibody, followed by a second stage (visualization), where other molecules are bound to the antibodies in order to make them detectable using different techniques. In this visualization step it is common to use colorimetric, chemiluminescence, radioactive and also fluorescence detection techniques. QDs are a powerful tool in order to probe the recognized targets as a consequence of their good fluorescence emission properties and their excellent bio-conjugation potential. On of the main advantages of using QDs in immunofluorescence assays is that lower concentrations of the probed-target biomolecules can be detected due to the high quantum yield of the QDs. This high emission efficiency lowers the required number of QDs necessary to obtain a measurable amount of light. The use of QD-avidin conjugates [69, 70] is a very good alternative to organic dye fluorophores in such immunoassays due to their good properties like photostability, high quantum yield, and versatility. But they are not used only as high-quality fluorophores. They also open the way to multiplexed protein detection due to their wide absorption band (allows only one light source excitation), their narrow emission peak (allow signal multiplexing), their size control (allows color tunability for signal multiplexing), and their surface engineering (allows multiple-target-specific reactions). This multiple-target detection has been successfully proved by several groups using modified ELISA techniques [71, 72]. Other example of immunoassay is the WB technique where gel electrophoresis is used in order to separate the biomolecules by their length (denaturing conditions) or by their quaternary structure (native or non-denaturing conditions). The biomolecules are then transferred to a substrate (typically nitrocellulose or polyvynilidene fluoride membranes), where they are probed with their specific antibodies. Regarding the recognition stage, there are several variations related to the use of only one antibody reaction, or a two-step scheme where a

154

J. Goicoechea et al.

primary (for target binding) and a marked secondary antibody are used (currently the most used procedure). Afterwards, the visualization agents are linked to the antibodies, revealing the presence or not of the target biomolecule. The WB technique is currently used in applications such as the diagnosis of HIV, bovine spongiform encephalopathy, or lime disease, just to cite a few. Makrides reported in 2005 [71] a WB assay for two different proteins using two color QD-Streptavidin-conjugated nanocrystals. This was not possible using organic dye fluorophores. One example of ultrasensitive detection using QD nanocrystals in WB assays was reported by Bakalova et al. in 2005 [73]. In this work the authors reported a single antibody WB assay where a primary biotinylated antibody for two telomere target proteins, namely Telomeric Repeat binding Factor (TRF1) and TRF1-Interacting Nuclear protein 2 (Tin2). The antibodies were previously conjugated with biotin-CdSe QD conjugates using avidin as linker between the QDs and the antibody as it is shown in Fig. 6.12. Some groups have reported ELISA-based fluorescence assays using QDs as fluorescent labels bound to the antibodies. The so-commented QD-avidin conjugation make possible the easy and direct improvement of fluorescence ELISA-kind approaches, and once more, the QD nanocrystals have shown an enormous potential to replace the organic dye fluorophores due to their good photophysical properties and also due to their multiple-target detection potential. Goldman et al. reported successfully multiplex sensing using four color QDs (CdSe/ZnS core/shell), for cholera toxin, ricin, shiga-like toxin 1, and staphylococcal enterotoxin B detection [74]. The simultaneous detection of the four toxins from a single sample probed with a mixture of all four QDantibody reagents was successfully carried out in a single well of a microtiter plate. In this work, the QDs were conjugated with amine-reactive NHS biotin, and those conjugates were used as tracer reagents against nitroavidin or nitroNeutrAvidin.

6.3.2.2 DNA Detection The so-commented photophysical properties of the QD nanocrystals make them ideal for a lot of applications where organic dye fluorophores have been used for years. One of these applications where QD nanocrystals show enormous potential is in DNA detection. There are several factors that make QDs suitable for genomic applications; their high photostability, their narrow and tunable emission peak (allows multiplexing), their great bio-functionaliztion potential, and their high fluorescence quantum yield that provides higher signal-to-noise ratios with lower analyte concentrations. This high fluorescence emission allows a better spatial definition, as far as almost single QD detection can be achieved with advanced fluorescence microscopy techniques [72, 75, 76, 77, 78]. This spatial resolution is a basic requirement of some DNA detection techniques such as fluorescence in situ hybridization (FISH).

6 Quantum Dots for Sensing

155

Fig. 6.12 Ultrasensitive QD-based Western blot analysis of tracer proteins (TRF1, Tin2) and b-actin in leukemia K-562 cells using a sandwich-type arrangement (QD-biotin-biotinylated antibody conjugates by avidin bridges). The high quantum efficiency of QDs make possible to lower the detection limit of the WB assay Reprinted from [73] with permission from the ACS publishing group

In FISH assays, fluorescently labeled DNA probes are used for gene mapping and identification of chromosomal abnormalities. The genetic material (stored in the chromosomes) is placed together with the labeled complementary DNA sequences, and an in situ hybridization of the genetic material occurs. Therefore the fluorescent probe selectively is attached to the complementary nucleotide sequence, allowing the spatial localization of determined genetic sequences inside the chromosome [79]. The main drawbacks of using organic dyes in FISH applications are related with the difficulty of multiplexing more

156

J. Goicoechea et al.

than four different optical signals, and also with the high photobleaching rates of the dyes which resulted in low signal values, and the formation of freeradicals from the photodegradation reactions which interfered in the DNA– protein interactions. Several groups have reported different DNA-QD conjugation strategies in which the DNA or oligonucleotides still retained their ability to form complementary sequences of Watson–Crick base pairs, or in other words, retained their normal functional behavior. These DNA-QD conjugation approaches are usually based on non-covalent self-assembly interactions between the biomolecule and the functionalized QDs. Recently, direct multicolor imaging using QD-DNA conjugated FISH probes was achieved in Escherichia Coli by Wu and coworkers [80]. A different approach in DNA tracking was reported by Han [14]. In this elegant work, different color emitting QDs are embedded into polymer nanobeads. By controlling the amount of the each color in each polymeric sphere type, it is possible to fabricate different colored barcodes. Conveniently conjugated with oligonucleotide sequences these nano-barcodes can be used for the detection of single nucleotide polymorphisms (SNPs). These SNPs consist in DNA sequences which only differ at a single nucleotide [14, 81]. This approach also highlights the importance of QDs in single-molecule recognition. A different approach is also reported by Gerion et al. [82] for SNPs detection. There are also DNA recognition techniques based on Forster Resonance ¨ Energy Transference (FRET). Patolsky and coworkers [83] have reported FRET-based sensing mechanisms in order to monitor the telomerization process and also the replication of a viral DNA. CdSe/ZnS core/shell QDs stabilized with mercaptopropionic acid and modified with a thiolated oligonucleotide were used. These QD conjugates were incubated together with a deoxyribonucleotide triposphate (dNTP) mixture that included the nucleotide 2’-deoxyuridine 5’-triphosphate (dUTP) labeled with Texas-Red (dUTP-TR) in the presence of telomerase. As telomerization proceeds, the fluorescence from the QDs decreases, and at the same time the fluorescence from the TR increases. This is explained as result of a FRET coupling between the QDs (donors) and the dye molecules (acceptors) incorporated into the telomeric units by telomerase. The same experimental procedure was followed in order to measure the viral DNA replication. QDs were functionalized with the complementary oligonucleotide of a virus M13 DNA. The M13 DNA was incubated together with the QD-oligonucleotide complexes, with DNA polymerase and with the mixture of dNTPs (containing the TR-dUTP). The same FRET phenomenon was registered as the DNA hybridization proceeded, as far as the QDs and the TR molecules were brought closer. This specific DNA recognition using simple fluorescence techniques shows potential applications as the fast and sensitive detection of specific cancer cells, for example. This technology may result in the fabrication of chip-based DNA sensors, where the simple fluorescence signals can be measured using a CCD detector, and the FRET readout occurs when DNA hybridization (recognition of the specific target) proceeds.

6 Quantum Dots for Sensing

157

Fig. 6.13 Working principle of DNA recognition using molecular beacons (MBs). This is a powerful way to label specific oligonucleotide sequences with a high degree of selectivity, making possible to detect even single nucleotide polymorphisms (SNPs)

Another FRET-based sensing approach was reported by Kim and coworkers [84] where QDs are used conjugated with molecular beacons (MBs) for specific segment recognition. The MB technology is used for revealing the presence of specific oligonucleotide sequences using the FRET quenching of a fluorophore, and it was firstly reported by Tyagi and Kramer in 1996 [85]. MBs consist of hairpin-shaped oligonucleotide molecules with a fluorophore and a quencher in each extreme of the polynucleotide chain (see Fig. 6.13). When the MB is not hybridized, the emitter and the quencher are very close allowing the FRET to occur, but when the conjugated DNA target is localized and the hybridization reaction is carried out, the MB unfolds and both the emitter and the quencher are placed further one respect the other. Consequently, the hybridization of the DNA sequence can be observed by monitoring the fluorescence coming from the system as far as the FRET quenching has been disabled. In this work [84] CdSe/ZnS core/shell QDs functionalized with mercaptoacetic acid were conjugated with amine-terminated MBs with 4-(4’-dimethylaminophenilazo benzoic acid) (DABCYL) as quencher molecule.

6.3.2.3 Cell detection The use of staining dyes in histology and biology has been one of the most valuable tools for the study of cells and their anatomy, function, and chemistry (also known as cytology). In particular, the organic fluorescent dyes have

158

J. Goicoechea et al.

supposed a tremendous impact in the cellular research community as far as they contribute to the selective detection and tracking of some parts or processes inside the cell. The usual research tools that the cellular biologist uses together with organic fluorescent dyes are fluorescence microscopy, fluorescence confocal microscopy, multiphoton microscopy, spectral image analysis, among others. Semiconductor quantum dots (QDs) have supposed a revolution in the cytology fluorescence applications as far as they improve the performance of the observation techniques and make possible to reach new applications where organic dyes could not. Since Alivisatos and coworkers reported the first works in 1998 [15] there has been a huge effort from the scientific community to develop new bio-applications of QD nanocrystals. The main motivation of this interest is to take advantage of the QDs photophysical properties with respect to the traditional organic dye labels. However, although QDs are far much better than dyes, their application in biology is still focused on research works and it is not a common practice in real assays. This reticence is mainly related with the difficulty in fabricating stable QD-biomolecule conjugates. Nevertheless, research works have demonstrated that QDs overcome organic dyes in most of their features, making them promising substitutes to a lot of conventional organic fluorophores in the short-term [86]. The extremely high photostability and quantum efficiency of QDs compared to organic dye fluorophores, combined with confocal fluorescence microscopy techniques, have led to a significant improvement in 3D cell reconstructed imaging [87, 88]. Image reconstruction quality has been severely limited by the photobleaching of the traditional fluorophores, so QDs have supposed a great advance in this particular application. QDs also offer the possibility of having different colors of the VIS-NIR spectrum, simply by changing the nanoparticle size. This made possible to make wavelength signal multiplexing of different probes inside a cellular environment [24, 88, 89]. Finally, QDs also offer the opportunity of conjugation with biomolecules, which can provide them with some selectivity. A lot of works focused their effort in making new QD bioconjugates for new applications, just to cite a few, nanocrystals have been functionalized with Prostate-Specific Membrane Antigen (PSMA), HER kinases, glycine receptors, serotonin transport proteins, p-glycoprotein, etc. [24, 31, 43, 44, 45, 87, 90, 91, 92, 93, 94]. Another advantage of QDs is that due to their high quantum yield, lower concentrations are required to produce a significant light emission. This made possible even single QD tracking [75, 76, 78, 95]. Figure 6.14 displays a good example of single QD tracking, where live HeLa cells were stably transfected with a plasmid expressing a chimeric avidin-CD14 receptor, incubated with biotin-QDs conjugates. However, when single quantum dot emission is observed, it is common to find an undesirable effect, termed as ‘‘blinking’’ [76, 95]. The blinking phenomenon consists in sudden variations in the emission intensity and wavelength of the single QD (see Fig. 6.14D). The last reasons for this behavior are still

6 Quantum Dots for Sensing

159

Fig. 6.14 Single QD tracking inside HeLa living cells, transfected with a plasmid expressing a chimeric avidin-CD14 receptor. Biotin-QD conjugates were used as probes and observed with fluorescence microscopy. a Bright field image. b Fluorescence image. c Single QD tracking trajectory. d Emission versus time profile of the single QD where it is clearly visible the blinking effect From [75]. Reprinted with permission from AAAS

unclear, but it is believed that an alternation in the ionic state of the QD may cause photophysical variations, and therefore fluctuations in the emission spectra and intensity of QDs. This dynamical behavior can be explained by extending the photoionization model that describes blinking in semiconductor nanocrystals as a result of a QD ionization and Auger electron-exciton (holeexciton) energy transfer. This effect is observable only in single QD emission tracking because when higher QD populations are studied the emission is statistically stabilized. However, this blinking issue can be overcome by passivating the QD surface with thiol moieties [96], or by using QDs in free suspension [97]. It has also been demonstrated that SNR ratios in fluorescence microscopy applications can be boosted due to high fluorescence times of QDs [98]. This fluorescence time can be as long as 10 ns, which is two or three orders of magnitude longer than organic dyes and usual tissue autofluorescence. It is possible to insert a short-time delay between the excitation and the emission scanning, getting rid of the main part of the background signal, therefore improving the sensitivity of the measurement. In the following paragraphs some of the most relevant advances in cell tracking using QDs will be commented. As the reader will notice, most of the QD techniques used for cell-tracking applications are also commonly used in applications such as immunoassays or DNA tracking. Therefore, DNA tracking or immunoassays share some techniques with cell tracking applications, like the FISH technique, or the QD-antibody conjugation.

160

J. Goicoechea et al.

Fig. 6.15 A dual-color image of QD 605-labeled C. parvum and QD 565-labeled G. lamblia using strategy 2. Scale bar, 10 mm From [104] with permission from the American Society for Microbiology

Several groups worldwide have reported the successful labeling and visualization of numerous cellular parts and events or proteins inside living or fixed cells, for instance, nuclei, mithochondria, actin filaments, cytokeratin, etc. [99, 100, 101, 102, 103]. The FISH technique has been widely used for cellular fluorescence probing because it has a high potential to 2D and 3D imaging of QDs selectively attached to some elements of the cell. The selectivity is provided to the QDs by the conjugation with oligonucleotide chains. Zhu and coworkers have reported a FISH approach for selectively localization of Giardia Lamblia bacteria [104] as it is shown in Fig. 6.15. There are also a number of works in which immunoassay-based approaches are used for selectively observing and tracking some specific cells. In these woks antibodies have been attached to QDs to form high-quality selective labels, as in [105] where streptavidin-coated QDs were conjugated with anti-InlA (InlA is a cell surface protein of Lysteria monocytogenes) and these conjugates were used for single cell detection. A similar approach can be found in [106] where streptavidin-coated QD was attached to biotinylated anti-E. Coli antibodies for the selective detection of Escherichia Coli bacteria. This approach has been used for the selective detection of more cells of different types, like protozoa or viruses, taking advantage of the wide potential of QD-immunoassays. Other peptideconjugated QDs were used for cell detection [50], where Lucine zipper-maltose binding protein was linked to the QDs by electrostatic attractive interactions. The binding protein showed evidences of retaining its functional activity. If this selectivity is combined with the color-tuning potential of QDs, it is possible to achieve selective multi-target detection, wavelength multiplexed, as it is described in [104] where two different size QDs (emitting in two different colors) are conjugated with different antibodies. Therefore, simultaneous detection of Cryptosporidium parvum and Giardia Lamlia bacteria was achieved using fluorescence microscopy. A similar approach was reported in [107] were two-color simultaneous detection of Escherichia Coli and Salmonella Typhimurium was achieved.

6 Quantum Dots for Sensing

161

Fig. 6.16 Scheme of the main QD probe delivery mechanisms inside living cells Reprinted from [110] with permission from Elsevier

These advances made possible to have excellent fluorescent bioprobes, which allowed to study long cellular dynamics, or some specific cellular processes with a better performance than with the traditional organic dyes. QDs can be delivered inside living cells in order to study them, using techniques such as microinjection [42], peptide-induced transport [108], electroporation [109], or phagocytosis [90] as it is shown in Fig. 6.16. Using these QD loading techniques inside the cell, its internal processes can be observed, even if a cell is loaded with some QDs, they would spread in children cells during the cytokinesis process. Therefore, QDs make possible tracking long-term cellular incidences, movement, and fate of living cells [42, 45, 90, 111]. All these studies would have not been possible if organic fluorescent dyes were used instead of QDs. One of the first works on cellular dynamics imaging was reported by Chan and Nie in 1998 [30] where transferrin-QD conjugates uptake by living HeLa cells via receptor-mediated endocytosis was observed by fluorescence microscopy. Alivisatos also reported in 1998 [15] the selective multicolor target labeling of fibroblast nucleus with green silanized QDs, and red biotinylated QDs conjugated with a streptavidin/biotin-phalloidin complex to label the F-actin filaments. These two early studies were the basis for numerous QD-based cell sensing and imaging research works published recently. Just to cite a few woks, one example of cellular process tracking is [35] where QDs are conjugated with Lectin via EDC coupling in order to label grampositive bacteria. In this work QDs were tracked as they diffused into the membrane and moved within the cytosol. Jaiswal and coworkers reported in [90] the labeling of HeLa and D. discoideum cells using QDs by endocytic uptake (see Fig.. 6.17) and by QD-antibody bioconjugate probing. QDs are also useful to target and study neural cells and their dynamics. In 2001, Winter and coworkers [112] suggested a QD-biomolecule conjugated for

162

J. Goicoechea et al.

Fig. 6.17a HeLa and b D. discoideum cells confocal microscopy images Reprinted from [90] with permission from Nature Publishing Group

specifically bind the v membrane subunit of cultured living neurons. These conjugates were synthesized by linking mercaptoacetic acid-coated QDs with polypeptides, which contain the RGD recognition group for binding vb1 and vb2 integrins. In this work [112] the authors suggested the possibility of generating a QD-electrical response when the sample is illuminated, causing photo induced responses in the nerve cells, therefore interfering with the neuronal activity, causing photo induced responses in the nerve cells. In other work, Dahan and his group [45] have reported the use of QD for tracking the diffusion dynamics of glycine receptors in neuronal membranes of living cells. Glycine is one type of aminoacid neurotransmitter and the study of the dynamics of the glycine receptors is a key point in the understanding the neural system, since they have a main role in the synaptic transmission process. The QD bioprobes were fabricated by binding a primary antibody (mAb2b, for targeting GlyR 1 subunits at the surface of the neurons) to biotinylated anti-mouse Fab fragments, and then linking the antibody conjugate to streptavidin-coated QDs. Dahan and coworkers [78] have also reported the importance of singlemolecule experiments in the investigation of intracellular transport. In this work the motion of intracellular proteins was tracked with a high sensitivity using QDs. The in vivo motion of individual QD-tagged kinesin motors was characterized in living HeLa cells. This work shows the relevance of singlemolecule measurements, which in this case provided important information about parameters of the motor. However, although the application of QDs to biology applications had contributed to enormous scientific advances, they also show some important problems. One of the main negative effects of traditional QD nanocrystals in

6 Quantum Dots for Sensing

163

biological systems is their potential cytotoxicity [110], and their impact at certain QD dose levels in the cellular activity. Nevertheless, although QD toxicity is an obvious concerning in in vitro techniques, it is not as critical as it is in in vivo applications. Even more, most of the traditional limitations of QDs are currently being overcome with the application of new materials and approaches, as it will be commented later.

6.3.3 In vivo Biological Applications The first approach of using QDs for in vivo biological applications consists in simply introducing stable QDs into living subjects and proceed to their observation, trying to extract conclusions about their natural distribution inside a living organism. Therefore it is possible to speak about non-selective in vivo imaging as the range of applications where QDs are spontaneously accumulated in an area of interest (like tumors, or lymphatic system) though they are not biofunctionalized. On the other hand we speak about selective (or targeted) imaging when biomolecules such as proteins or antibodies are attached to the QD probes in order to bind specifically a target zone. In the following subsections the main advances in these two approaches are commented.

6.3.3.1 Non-selective Imaging As far as the living bodies show an almost negligible transparency to visible light, optical in vivo diagnosis tools have been traditionally reduced to very specific applications. Nevertheless, since the first studies it is known that there are two spectral windows for optical in vivo imaging; the first one in the 700–900 nm range and the other one in the 1200–1660 nm range [113], both of them situated in the near infra-red (NIR). It is also difficult to find high efficiency NIR emitting organic dyes. Therefore QDs, with their high efficiency and NIR emitting capability, are a much better alternative to organic dyes for optical in vivo diagnosis tools [114, 115, 116] in fact, the use of these NIR-QD probes have been successfully demonstrated [114, 115] in living mice. In those works, optically quenched NIR-QDs probes consisting of a QD-organic dye conjugate (based on FRET) were used for cancer detection. In contact with some enzymes present in the tumor cells, the quenching dyes were released and the QDs recovered their emission. Another positive QDs property is that they have a very large two-photon cross efficiency, approximately two or three orders of magnitude higher than most of the organic dye fluorophores [117, 118, 119]. This property makes them ideal for multiphoton microscopy. This fluorescence microscopy technique takes advantage of the two-photon excitation concept, first described in 1931. According to this concept, two coherent photons can combine to produce one photoinduced electronic excitation double than their initial energy. This makes possible to have

164

J. Goicoechea et al.

fluorescence while illuminating with a laser with a wavelength which is double than the required excitation. This phenomenon has a low probability to occur, and therefore needs a high excitation power in order to achieve significant signal levels. On the other hand it allows to use excitation wavelengths which can penetrate into in vivo bodies making this application suitable for deep tissue imaging. This powerful microscopy tool has been widely used for vasculature imaging in deep tissue applications, as it is shown in the next lines. In [97] Larson and coworkers used two photon microscopy for blood vessel imaging, achieving deep scanning and high-contrast images with less excitation power than with organic dyes. In this work it was even possible to track the movement of particles inside the capillary vessels, as it is shown in Fig. 6.18. With this technique, Levene and coworkers [120] have achieved deep fluorescence imaging in the brain tissue of anesthetized living mice. In this work, images of the cortical layer V and hippocampus as deep as 1 mm were successfully acquired. Multiphoton microscopy was also reported by Voura and coworkers in [117] to study the extravasation of QD-labeled metastatic tumor cells in living mice. Bawendi’s group has also used multiphoton microscopy [121] to differentiate tumor vessels from perivascular cells, comparing QDs performance with standard dextran vessel markers. Smith and coworkers have reported QD

Fig. 6.18a Fluorescent capillaries containing 1 mm QDs (approx.) were clearly visible through the skin at the base of the dermis (ca. 100 mm deep). Dashed line indicates position of line scan shown in (b). b Line scan (13.7 ms per line) measurement of blood flow velocity taken across a capillary in (a). The diameter of this capillary is ca. 5 mm, and the flow is around 10 mm/s. c Zoom of section in (b) showing undulations in capillary due to heartbeat. d Comparison image at the same depth as in (a), acquired by injecting FITC-dextran at its solubility limit. e Image of the surface of adipose tissue surrounding the ovary. Dark circles are adipose cells. f Projection of capillary structure through 250 mm of adipose tissue From [97] Reprinted with permission from AAAS

6 Quantum Dots for Sensing

165

fluorescence visualization of blood vessels in chick Corioallantoic Membrane (CAM) models [122]. The CAM model is useful for studying angiogenesis processes [123]. In [122] it is reported QDs were used as angiographic contrast in CAM models and they stayed in circulation for more than 4 days. Signal levels were similar to those obtained when using dextran vessel marker, but with concentration doses around two or three orders of magnitude lower. Other approaches have contributed to blood vessel imaging. In [36] peptideconjugated QDs are used to target-specific tissue recognition. After intravenous injection in mice, histological sections after 5 and 20 min showed that QDs were internalized by endocytosis in the vessel cells but not in the surrounding tissue. Bawendi reported in [113] an approach where NIR QDs were used to image the coronary vasculature in vivo. Nevertheless, the stability of the QD conjugates may be a concern in certain applications. In order to increase the stability of such QD probes, Bruchez and his group [124] have developed some QD–polymer conjugates for long-term imaging in mice. Here, polyethylene glycol (PEG) coatings were performed over the QDs. This protective coating prevents the QD to degrade (they are visible after several months in the bone marrow and lymph nodes of living mice) and also reduces the accumulation in organs like the liver and bone marrow. Other QD protective layers have been successfully used, like phospholiphid block-copolymer micelles [42]. In this work, the micelle encapsulation reduced significantly the photobleaching and the non-specific adsorption of the QD probes. These QD conjugates were injected into living Xenopus embryos, and it was found that QD probes dispersed diffusely inside the cells in their early stages, concentrating later in the nuclei. Little or no cytotocicity evidences were found. Another work [118] studied the behavior of QDs in living zebrafish embryos. Both the works [42, 118] show the relevance of the QDs as stable contrast agent in living individuals, which can be useful in long-term studies. This may suppose a great impulse to embryology, cell biology, disease phenotyping, just to cite a few potential applications. QDs have also been used for mapping the lymphatic circulatory system of living animals. This has an extraordinary importance in cancer diagnosis and therapy because the lymphatic system is one of the main ways of propagation of tumor metastatic cells. Bawendi and coworkers reported a study in which type II QDs were synthesized and their photophysical properties were analyzed [125]. The QDs described in this latter work consisted in CdTe/CdSe core/ shell QDs which emitted around 850 nm; the NIR region of the spectrum. In [23] these NIR-QDs probes were injected in living pigs and mice. QD was monitored by infrared imaging, and it was observed that QD rapidly migrated to the lymph nodes. In medicine, the first lymph node (or group of nodes) reached by metastasizing cancer cells from a tumor receives the name of sentinel lymph node (SLN). The localization of such SLNs is not an easy task, and it is a key factor in the study and diagnosis of the spreading of tumors. The approach reported in [23] permits to acquire precise, fast, and almost background-free images of the SLNs. It was also reported that the image-guided resection of a

166

J. Goicoechea et al.

Fig. 6.19 Sentinel lymph node (SLN) detection in nude mice using NIR emitting InAsP-core QDs. The three rows correspond to different moments; postinjection (a, b, c), 3 min postinjection (d, e, f), and post-resection (g, h, i). The images were acquired using white light, NIR fluorescence, and color/NIR merge, respectively Reprinted from [126] with permission from ACS publishing group

1 cm deep lymph node in a pig with only 5 mW/cm2 of excitation power and with a low QD dose. In a similar way, Kim et al. reported the synthesis of nonCd based NIR-QDs based on an InAsP core for SLN mapping in mice [126] (see Fig. 6.19). These approaches opened the way to near-future surgical guidance applications. Other groups have reported SLN mapping approaches in living rats [127, 128, 129] and pigs [130, 131, 132]. These works report a quick and accurate lymph mapping by injecting only a 200 pmol dose of QD probes in an animal as big as an adult pig. This can give valuable information in surgery applications due to the high sensitivity and real-time imaging of the area of interest. Non-targeted QD imaging has also been reported to observe neural cells in vivo. Several works have reported successful results [133, 134, 135, 136] in this field. Water soluble QDs have been injected into the brain extracelular space (ECS) of a rat’s neocortex in vivo [133] in order to predict the ECS width in normal neuronal tissue and also in one with terminal ischemia. The ECS width is a key factor in order to study the neurons and glia access to nutrients and therapeutics. The works [134, 135, 136] also reported detectable fluorescence from QD labeled neuronal tissue of living rats, but the amount of QDs necessary to achieve neuronal QD staining is found to be very high. In fact, in order

6 Quantum Dots for Sensing

167

to reach detectable fluorescence in brain cells it is necessary to overcome the QD sequestration form of other tissues more prone to the QD accumulation, like lymph nodes, liver, bone marrow, etc. This enormous increase in the minimum QD doses produces an immediate cost problem and also may cause toxicity issues. Obviously, as well as in ex vivo applications, cytotoxicity appears as a very adverse effect for in vivo applications [35, 42, 137]. This drawback needs to be solved if an extrapolation to human medicine applications is desired.

6.3.3.2 Targeted Imaging It is clear that non-specific imaging is the simplest approach to in vivo labeling applications. Although successfully and simple approaches have been reported for example in SLNs mapping, there are more sophisticated approaches, which intend to achieve more functional labels; almost smart probes. Using some of the biomolecule conjugation techniques used in QD fluorescence immunoassays or in DNA detection, it is possible to fabricate selective indicators, which incorporate all the good photophysical advantages of the QDs (like low photobleaching, high quantum yield, long fluorescence lifetimes, high two-photon efficiency, wavelength tuning, etc.). Therefore, functionalized QD probes can be selectively accumulated in the desired targets. This is extremely important in applications such as cancer diagnosis and treatment. Attending to the biomolecule used to provide the selectivity to the QDs it is possible to classify the QDtargeted probes into peptide-QD conjugates and antibody-QD conjugates. Regarding the peptide-conjugated QDs approaches, Ackerman and coworkers [36] studied the effect of three different polypeptide-QDs conjugated in their affinity to breast carcinoma MDA-MB-435 target cells. Different sized CdSe/ZnS core–shell QDs coated with mercaptoacetic acid were conjugated with three different polypeptides: GFE, F3, and LyP-1. Afterwards, the QD probes were injected into the tail vein of xenograft-bearing living mice. The different color emitting QD-F3 and QD-LyP-1 conjugates were bound to different structures in the tumor tissue showing a good selectivity. While QDF3 accumulated in the tumor blood vessels, the QD-LyP-1 accumulated in the tumor tissue instead of the vessels. In some cases the QD was co-conjugated with the peptide and also with polyethyleneglycol (PEG) in order to reduce the aggregation between QD conjugates and also to decrease the non-specific accumulation in the reticuloendothelial system (RES). Another example of wavelength multiplexed imaging is shown in Fig. 6.20 [138]. Multiplexing can potentially allow researchers to perform real-time tracking of multiple biological molecules in live animals with minimal loss in optical detection sensitivity. The ability to spectrally distinguish unique sets of NIR QDs can lead to improvements in earlier detection of diseases such as cancer, where multiple markers can be detected simultaneously using a bioconjugated NIR-QDs cocktail.

168

J. Goicoechea et al.

Fig. 6.20 Multiplexed in vivo imaging using CdTexSe1-x/CdS NIR QDs. a Bright field images. b Fluorescence images obtained from mouse injected with 650-nm emitting QDs (arrow) using 650/50 bandpass filters. c Fluorescence images from the same mouse injected with 750-nm emitting QDs (arrow) using 750/50 bandpass filters. d Superimposed images of (a, b, and c) demonstrate multiplexed near-IR QDs at the injection site relative to normal mouse anatomy Reprinted from [138] with permission from Elsevier

A similar approach was reported by Cai [139] for tumor cell detection. This time the target was U87MG human glioblastoma cells transferred to living mice. The QDs were conjugated with QD705-RGD polypeptide. This biomolecule contains an arginine–glycerine–aspartic acid (RGD) group, which is a potent vb3 antagonist. This membrane subunit is highly expressed in tumor cells while is almost undetectable in healthy ones. The result was an excellent contrast between the tumor and the unaffected tissue which reached its maximum 6 h after the injection. Nevertheless, the QD probes concentrated in the tumor vessels because the overall size of the probes (20 nm) resulted too big for an efficient extravasation. Also several works have been reported where antibody-conjugated QDs are used as targeted probes. In these applications, similar conjugation techniques as in fluoroimmunoassays are used. However, it is difficult to find a large amount of successful works for in vivo applications, since the QD-biomolecule conjugates are still not as stable as it is desired, yielding in short circulation times of the probes, and therefore limiting their applications. In [43], Gao and his group reported the way that triblock copolymer-coated QDs can be conjugated with prostate-specific membrane antigen (PSMA) monoclonal antibody for prostate cancer detection. The PSMA is a cell surface marker for prostate epithelial cells and also for neovascular endothelial cells. Although the injected probes accumulated selectively in the tumor tissue (into bearing living mice) and wavelength multiplexed imaging was achieved, it is unclear if the tumor labeling was targeted or untargeted because no hystologic measurement was performed in order to determine if the probes were selectively

6 Quantum Dots for Sensing

169

adsorbed onto the cell membranes or if they were non-selectively accumulated into the tumor tissue. A different QD conjugate was fabricated by Yu et al. [140]. This time QDs were kinked to anti-alfa-fetoprotein (anti-AFP), a marker for hepatocellular carncinoma cells for in vivo tumor detection and tracking. Successfully labeling of the tumor tissue was achieved, allowing to acquire spectroscopic images of the hepatoma cells. Nevertheless, no preliminary ex vivo study was performed, and therefore not real evidences of selective antibody–antigen binding were provided. It is possible to find in the bibliography other works where QDantibody conjugates were used [77], but it is difficult to find works where specific probe binding is doubtless evidenced. 6.3.3.3 Drawbacks of QD Nanocrystals for Bio-applications One of the main problems when working with QD nanocrystals together with biological systems is the cytotoxicity of CdSe QDs. This toxicity may be used as an advantage in applications such as photodynamic therapy, where selectively attached QDs to the target tissue are deliberately photodegraded in order to produce a local toxic effect. These treatments are being studied in some specific cancer therapies. Nevertheless, for most of the sensing applications where living organisms are involved, toxicity is a serious concern. Several groups have demonstrated that there was no cytotoxicity or that it was very low for optimal doses [75, 90, 141] measured for long periods of time (even after several months). However, when the amount of QDs is increased, negative effects start to appear [9, 10, 11, 12]. In order to solve this problem a new generation of novel non-Cd-based QDs is being recently developed [126, 142, 143]. Researchers are also focusing their effort on achieving smaller QDs. The aim is to solve the problem that supposes the QD probe overall size for some applications. In order to develop certain applications such as an efficient extravasation of QD bioconjugates, smaller QDs are needed. The study of new synthetic routes using new materials with different Bohr radius may lead to the production of smaller QD like in the case of InAs/ZnSe core/shell QD, with an unusual size of around 2 nm [144]. These new QDs still have low efficiencies compared to the traditional CdSe/ZnS QDs, but their small size can be used for certain imaging applications. There is also interest in replacing the protective polymeric overcoatings (such as PEG) because of the consequently increase in the overall size of the QD probe. A good alternative to this polymeric coating is the use of dendromers. The dendrom-coated QDs have high stability, versatility and biochemical processability [145], and at the same time they show a small radial size compared with the polymers. Another issue of QDs when thinking about in vivo applications is the lack of transparency of the living tissues to UV excitation light. Although NIR QDs can be used for in vivo applications, and their emitted light can be collected from the outside of the living organism, the excitation of such materials is still a problem. Most of the times in vivo applications are restricted to the outer layers

170

J. Goicoechea et al.

of the specimens’ body (cutaneous tumor detection), and even in such applications, very high excitation powers are needed. It has been commented how multiphoton fluorescence microscopy is used for deep tissue imaging, but it still works in a depth range going from a hundred microns up to 1 mm. A very elegant approach was reported in [146] where self-illuminating QD probes were fabricated and successfully tested. This phenomenon is based on a FRET approach, but this time, the QD acts as an acceptor instead of being the donor. This phenomenon has been named as Bioluminescence Resonant Energy Transfer (BRET). In a BRET scheme a biomolecule with the ability of biochemically emitting light (donor) is attached very close to NIR-QDs (acceptor). This resulted in a biochemically triggered self-illuminating NIR probes. In [146] the biochemically luminescent protein renilla reniformis luciferase (Luc8) was used as donor molecule. Carboxyl-terminated QDs were linked with Luc8 through coupling of the amino groups from Luc8 and the carboxylates presented in the QDs.

6.3.4 Other Applications Finally, due to their high versatility, QDs have also been used by some researchers in other sensing applications that do not match in any of the previous classifications. It is true that most of the QD applications have been developed in the biological field. However, there are more applications with interesting approaches. One of the advantages of QD nanocrystals is their meaningful high resistance to photobleaching. This makes them ideal as active fluorophore for optical sensing applications because incredibly long lifetime sensors can be fabricated using QDs. One interesting result was reported by Bawendi and coworkers in 2003 [147]. It was found that the emission properties of CdSe/ZnS QDs showed a high dependence with temperature. This temperature induced changes consisted in a shift of the emission peak to the red (reaching more than 20 nm of displacement), and a simultaneous emission efficiency decreasing and peak widening as temperature was increased. These changes were reversible and almost linear with temperature. Several devices were successfully fabricated, and the QDs retained their temperature sensitivity even when immobilized into polymeric films. In a later work, De Bastida et al. [148] show a practical application of an optical fiber sensor based on the temperature-dependent emission from CdTe QDs. The QDs were immobilized into a polymeric thin films fabricated using the electrostatic self-assembly technique, or Layer-byLayer (LbL). Different size QDs have been used into a single LbL coating, making possible to achieve multiplexed optical measuring with only one excitation source. Some results of this optical fiber sensor are shown in Fig. 6.21. The temperature dependence of QDs has also been used to create sensors with other optical fiber arrangements. One good example of this is proposed by

6 Quantum Dots for Sensing

a

171

b

Fig. 6.21a Fluorescence spectra from the optical fiber sensor fabricated using LbL polymeric coatings incorporating red-emitting CdTe QDs. A significant fluorescence quenching and redshift are observed as the temperature of the [PDDA/CdTe]20 sensor was varied. b The CdTecontaining optical fiber sensors were submitted to different ambient conditions using a climatic chamber. A very linear response to temperature was observed while no humidity cross-sensitivity was registered a is reprinted from [148] with permission from the IEEE

Bravo et al. using a Hollow Core Fiber setup [149, 150] for evanescent wave sensing [151]. Another QD sensors family was fabricated for the detection of organophosphorous (OP) pesticides. The detection of such chemical compounds is very important in environmental and health applications since the organophosphorous pesticides may be toxic for humans and for the environment even at trace levels due to their structural similarity to nerve agents like sarin or soman. Constantine and coworkers reports a fluorescence-based sensor of diethyl 4-nitrophenyl phosphate, or paraoxon (an OP pesticide). The reported sensor takes advantage from the fluorescence-quenching effect of activated organophosphorous hydrolase (OPH) enzyme when it is close enough to QDs. The sensors were fabricated by placing together thioglycol-capped QDs with OPH into LbL polymeric thin films built up onto quartz substrates [152, 153]. A detection limit as low as 109M of paraoxon was reported in this work. In a different approach, Ji et al. report a paraoxon sensor, using the same sensing mechanism, namely CdSe/ZnS QD fluorescence quenching due to OPH activation in the presence of paraoxon. The difference is that this time the sensitive phase is not immobilized but it consists in a colloidal suspension of QD-OPH conjugates. The conjugation was performed via the electrostatic attraction between the negatively charged QD and the OPH, and the activity of the enzyme was retained after conjugation. The sensitivity limit for this latter approach was 108M. Since the versatility of QD nanocrystals is so high, it is also possible to fabricate optical sensors for other applications using similar approaches to the ones used in biological sensing and imaging. This is the case of the FRETbased biosensor. Goldman and her group have reported a QD biosensor for the detection of trinitrotoluene (TNT) [154]. This particular application field is catching the attention of lots of research groups due to the increasing

172

J. Goicoechea et al.

importance of homeland security. The QDs were conjugated with an anti-TNTspecific antibody via metal-affinity coordination. A dye-labeled TNT analogue prebound in the antibody-binding site quenches the fluorescence from QDs by FRET. When TNT is added to the solution, it displaces the quenching TNT analogue, and QD emission is recovered. A schematic plot of the FRET-based sensing mechanism is shown in Fig. 6.22a. Using this approach, a minimum concentration of 20 ng/ml of TNT was measured, and most of the fluorescence recovery was observed between 0 and 0.5 mg/ml, with a high selectivity. Other sensing applications can emerge due to the redox potential sensitivity of QDs. As it was discussed in Sec. 6.3.1 the QDs fluorescence can be quenched thanks to the electron transfer from QD to the conjugate molecule. This

Fig. 6.22 TNT biosensor based on the Forster resonance energy transfer (FRET) mechanism. a Schematic of the assay. When TNB-BHQ-10 is bound to the QD-TNB2-45 conjugate, QD fluorescence is quenched. As TNT is added to the assay, it competes for binding to the antibody fragment and the QD fluorescence increases following TNB-BHQ-10 release from the conjugate. The data resulting from the increase in QD PL are plotted as the difference signal versus concentration, both in linear (b) and in logarithmic (c) scales Reprinted from [154] with permission of the ACS publishing group

6 Quantum Dots for Sensing

173

electron transfer prevents the radiative recombination, yielding to a strong fluoresce signal decreasing. Please note that this quenching effect is very different from the FRET quenching phenomenon previously reported. While in the FRET quenching mechanism there is a fast energy transfer between the donor and the acceptor, in the present quenching mechanism a charge transfer is involved and, consequently, it is slower and depends on the redox state of the QD ligand. Using this approach it is possible to build sensors by controlling the electronic interaction between the analyte and the QD ligand that modulates its capability for this electron transfer, leading to a change in the fluorescence signal. Several works have reported using this approach. For example Yildiz and coworkers [155] have reported a system involving ferrocene, and methyl viologen as quenching redox QD ligand. As the quenching molecules were sequestered from the QD surface by adding cucurbituril, the fluorescence was recovered. Suffern et al. [156, 157] have developed a very interesting redox QD sensor by attaching a neurotransmitter like dopamine onto the QD surface. As far as the dopamine was oxidized or reduced, it varies its ability to quench the fluorescence from the QDs. Therefore those redox QD probes are used to monitor the local redox potential inside living cells; under reducing conditions, fluorescence is only seen in the cell periphery and lysosomes, however, as the cell becomes more oxidizing, QD labeling appears in the perinuclear region, including in or on mitochondria. This work suggests methods for the creation of phototoxic drugs and for redox-specific fluorescent labeling that are generalizable to any QD conjugated to an electron donor. Other aspects of the research need to be improved to reach the commercial application of such technologies. Currently the use of QDs is restricted in most of the cases to solution-based techniques. There is a need to develop appropriate immobilization techniques of QDs to substrates in a controllable way in order to fabricate active solid phases for working in flowing solutions [25]. Such new optrodes combined with microfluidic techniques may lead to applications as exciting as the development of QD sensitive microarrays for lab-on-a-chip applications. There are several techniques that have been successfully used for immobilizing QDs for sensing applications, like the creation of sol-gel matrices [158], molecularly imprinted polymers (MIP) embedded QDs [159], or Layerby-Layer thin films [148].

6.4 Conclusion As result of the wide application range in which QDs have improved significantly the existing sensing tools, it can be said that QDs have gone beyond their initial expectations especially in the biological field. If they are used simply as outstanding fluorophores, they maybe will not completely replace the organic dyes in the most well-established standard techniques where there is no need to change anything. Nevertheless, QDs will

174

J. Goicoechea et al.

replace standard fluorophores in some exigent applications where a higher photostability, NIR emission, single-molecule measurement, or long-term sensitivity is required. The possibility of achieving real and easy multitarget imaging due to the wavelength tenability opens the way to optical guided surgery. But it is their conjugation with specific biomolecules what really makes them very attractive and really versatile. This latter QD capability appears to be one of the most promising strategies in further developing bioactive fluorescent probes for sensing applications. There is also another great potential of QD for sensing. This is the ability of screening a large number of biomolecules (genes, proteins or cells), which makes them ideal for high throughput techniques, for DNA tracking or for protein detection. The promise of multicolor QDs nanoprobes (like the QD-tagged nanobeads for DNA detection) will have, without any doubt, an enormous impact in massive parallel biosensing and analytical detection applications. The combination of these QD-based sensing techniques with microfluidics and microarray fabrication (MIP) will produce shortly a new generation of nanostructured optrodes. In the biology and medical fields QDs are being used for developing new unforeseen tools which enhance the wide range existing of bioimaging techniques. Nevertheless, an improvement of the QD synthesis and bio-conjugation is still needed for in vivo applications in order to achieve longer lifetime selective NIR bioprobes. Optical imaging techniques are relevant for diagnosing tissues close to the surface of the skin, tissues accessible by endoscopy and intraoperative visualization. The main drawbacks of QDs are still inefficient probe delivery (due to the overall size of the probes), and the toxicity. Nevertheless the last improvements in synthetic routes are leading to smaller, non-Cd-based nanocrystals which can overcome in some cases the traditional behavior of these exciting materials. NIR optical imaging has been tested in patients, and the initial results are encouraging. QDs can definitively impact on cancer diagnosis and cancer patient management. Ex vivo protein applications (FISH, FRET) can be combined with in vivo diagnostics in the career toward future cancer treatment. Acknowledgments This work was funded in part by the Spanish Ministry of Education and Science - FEDER TEC2006-12170/MIC Research Grant and Government of NavarreFEDER EUROINNOVA Research Grants.

Bibliography 1. Brus LE (1984). Electron–electron and electron–hole interactions in small semiconductor crystallites – the size dependence of the lowest excited electronic state, J Chem Phys, 80, 9: 4403–4409. 2. Gaponenko SV (1998). Optical Properties of Semiconductor nanocrystals. Cambridge University Press, Cambridge.

6 Quantum Dots for Sensing

175

3. Murphy CJ and Coffer JL (2002). Quantum dots: A primer, Appl Spectrosc, 56, 1: 16A–27A. 4. Bryant GW and Solomon GS (2005). Optics of Quantum Dots and Wires. Artech House, Boston, MA. 5. Neumann W, Kirmse H, Hausler I, et al. (2004). Quantitative TEM analysis of quantum structures, J Alloy Comp, 382, 1–2: 2–9. 6. Tachibana K, Someya T, Ishida S, et al. (2000). Formation of uniform 10-nm-scale InGaN quantum dots by selective MOCVD growth and their micro-photoluminescence intensity images, J Cryst Growth, 221: 576–580. 7. Henini M, Sanguinetti S, Brusaferri L, et al. (1997). Structural and optical characterization of self-assembled InAs-GaAs quantum dots grown on high index surfaces, Microelectron J, 28: 933–938. 8. Cao YW and Banin U (2000). Growth and properties of semiconductor core/shell nanocrystals with in as cores, J Am Chem Soc, 122, 40: 9692–9702. 9. Hardman R (2006). A toxicologic review of quantum dots: toxicity depends on physicochemical and environmental factors, Environ Health Perspect, 114, 2: 165–172. 10. Colvin VL (2003). The potential environmental impact of engineered nanomaterials, Nat Biotech, 21, 10: 1166–1170. 11. Hoet PH, Nemmar A, Nemery B (2004). Health impact of nanomaterials? Nat Biotech, 22, 19. 12. Tsay JM and Michalet X (2005). New light on quantum dot cytotoxicity, Chem Biol, 12, 11: 1159–1161. 13. Peng XG, Schlamp MC, Kadavanich AV, et al. (1997). Epitaxial growth of highly luminescent CdSe/CdS core/shell nanocrystals with photostability and electronic accessibility, J Am Chem Soc, 119, 30: 7019–7029. 14. Han MY, Gao XH, Su JZ, et al. (2001). Quantum-dot-tagged microbeads for multiplexed optical coding of biomolecules, Nat Biotech, 19, 7: 631–635. 15. Bruchez M, Moronne M, Gin P, et al. (1998). Semiconductor nanocrystals as fluorescent biological labels, Science, 281, 5385: 2013–2016. 16. Koyama T, Ohtsuka S, Nagata H, Tanaka S (1992). Fabrication of microcrystallites of IIVI-compound semiconductors by laser ablation method, J Cryst Growth, 117, 1–4: 156–160. 17. Danek M, Jensen KF, Murray CB, et al. (1994). Electrospray organometallic chemicalvapor-deposition – a novel technique for preparation of II-VI quantum-dot composites, Appl Phys Lett, 65, 22: 2795–2797. 18. Sooklal K, Hanus LH, Ploehn HJ, et al. (1998). A blue-emitting CdS/dendrimer nanocomposite, Adv Mater, 10: 1083. 19. Huang JM, Sooklal K, Murphy CJ, et al. (1999). Polyamine-quantum dot nanocomposites: Linear versus starburst stabilizer architectures, Chem Mater, 11: 3595–3601. 20. Spanhel L, Haase M, Weller H, et al. (1987). Photochemistry of colloidal semiconductors 20 surface modification and stability of strong luminescing CdS particles, J Am Chem Soc, 109: 5649–5655. 21. Murray CB, Norris DJ, Bawendi MG (1993). Synthesis and characterization of nearly monodisperse CdE (E = S, Se, Te) semiconductor nanocrystallites, J Am Chem Soc, 115: 8706–8715. 22. Peng ZA and Peng XG (2001). Formation of high-quality CdTe, CdSe, and CdS nanocrystals using CdO as precursor, J Am Chem Soc, 123: 183–184. 23. Kim S, Lim YT, Soltesz EG, et al. (2004). Near-infrared fluorescent type II quantum dots for sentinel lymph node mapping, Nat Biotechnol, 22: 93–97. 24. Medintz IL, Uyeda HT, Goldman ER, et al. (2005). Quantum dot bioconjugates for imaging, labelling and sensing, Nat Mater, 4: 435–446. 25. Costa-Fernandez JM, Pereiro R, and Sanz-Medel A (2006). The use of luminescent quantum dots for optical sensing, Trends Anal Chem, 25, 3: 207–218.

176

J. Goicoechea et al.

26. Ludolph B, Malik MA, O’Brien P, et al. (1998). Novel single molecule precursor routes for the direct synthesis of highly monodispersed quantum dots of cadmium or zinc sulfide or selenide, Chem Commun, 17: 1849–1850. 27. Crouch DJ, O’Brien P, Malik MA, et al. (2003). A one-step synthesis of cadmium selenide quantum dots from a novel single source precursor, Chem Commun, 12: 1454–1455. 28. Dabbousi BO, RodriguezViejo J, Mikulec FV, et al. (1997). (CdSe)ZnS core-shell quantum dots: synthesis and characterization of a size series of highly luminescent nanocrystallites, J Phys Chem B, 101, 46: 9463–9475. 29. Hines MA and Guyot-Sionnest P (1996). Synthesis and characterization of strongly luminescing ZnS-Capped CdSe nanocrystals, J Phys Chem, 100, 2: 468–471. 30. Chan WCW and Nie SM (1998). Quantum dot bioconjugates for ultrasensitive nonisotopic detection, Science, 281, 5385: 2016–2018. 31. Wu XY, Liu HJ, Liu JQ, et al. (2003). Immunofluorescent labeling of cancer marker Her2 and other cellular targets with semiconductor quantum dots, Nat Biotechnol, 21, 1: 41–46. 32. Lim JG, Park YJ, Park YM, et al. (2005). Effect of InxGa1-xAs strain release layers on the microstructural and interband transition properties of InAs/GaAs quantum dots, J Cryst Growth, 275, 3–4: 415–421. 33. Sheehan JC, Boshart GL and Cruickshank PA (1961). Convenient synthesis of watersoluble carbodiimides, J Org Chem, 26, 7: 2525. 34. Bodanszky M (1992). The myth of coupling reagents. Pept Res, 5, 3: 134–139. 35. Kloepfer JA, Mielke RE, Wong MS, et al. (2003). Quantum dots as strain- and metabolism-specific microbiological labels. Appl Environ Microbiol, 69, 7: 4205–4213. 36. Akerman ME, Chan WCW, Laakkonen P, et al. (2002). Nanocrystal targeting in vivo, Proc Natl Acad Sci, 99, 20: 12617–12621. 37. Pinaud F, King D, Moore HP, et al. (2004). Bioactivation and cell targeting of semiconductor CdSe/ZnS nanocrystals with phytochelatin-related peptides, J Am Chem Soc, 126, 19: 6115–6123. 38. Clapp AR, Medintz IL, Mauro JM, et al. (2004). Fluorescence resonance energy transfer between quantum dot donors and dye-labeled protein acceptors, J Am Chem Soc, 126, 1: 301–310. 39. Medintz IL, Trammell SA, Mattoussi H, et al. (2004). Reversible modulation of quantum dot photoluminescence using a protein-bound photochromic fluorescence resonance energy transfer acceptor, J Am Chem Soc, 126, 1: 30–31. 40. Medintz IL, Konnert JH, Clapp AR, et al. (2004). A fluorescence resonance energy transfer-derived structure of a quantum dot-protein bioconjugate nanoassembly, Proc Natl Acad Sci, 101, 26: 9612–9617. 41. Yu WW, Chang E, Drezek R, et al. (2006). Water-soluble quantum dots for biomedical applications. Biochem Biophys Res Commun, 348, 3: 781–786. 42. Dubertret B, Skourides P, Norris DJ, et al. (2002). In vivo imaging of quantum dots encapsulated in phospholipid micelles, Science, 298, 5599: 1759–1762. 43. Gao XH, Cui YY, Levenson RM, et al. (2004). In vivo cancer targeting and imaging with semiconductor quantum dots, Nat Biotechnol, 22, 8: 969–976. 44. Edgar R, McKinstry M, Hwang J, et al. (2006). High-sensitivity bacterial detection using biotin-tagged phage and quantum-dot nanocomplexes, Proc Natl Acad Sci, 103, 13: 4841–4845. 45. Dahan M, Levi S, Luccardini C, et al. (2003). Diffusion dynamics of glycine receptors revealed by single-quantum dot tracking, Science, 302, 5644: 442–445. 46. Marshall D, Pedley RB, Boden JA, et al. (1996). Polyethylene glycol modification of a galactosylated streptavidin clearing agent: Effects on immunogenicity and clearance of a biotinylated anti-tumour antibody, Br J Cancer, 73, 5: 565–572. 47. Rogach AL, Kornowski A, Gao MY, et al. (1999). Synthesis and characterization of a size series of extremely small thiol-stabilized CdSe nanocrystals, J Phys Chem B, 103, 16: 3065–3069.

6 Quantum Dots for Sensing

177

48. Pathak S, Choi SK, Arnheim N, et al. (2001). Hydroxylated quantum dots as luminescent probes for in situ hybridization, J Am Chem Soc, 123, 17: 4103–4104. 49. Goldman ER, Anderson GP, Tran PT, et al (2002). Conjugation of luminescent quantum dots with antibodies using an engineered adaptor protein to provide new reagents for fluoroimmunoassays, Anal Chem, 74, 4: 841–847. 50. Mattoussi H, Mauro JM, Goldman ER, et al. (2000). Self-assembly of CdSe-ZnS quantum dot bioconjugates using an engineered recombinant protein, J Am Chem Soc, 122, 49: 12142–12150. 51. Mattoussi H, Mauro JM, Goldman ER, et al. (2001). Bioconjugation of highly luminescent colloidal CdSe-ZnS quantum dots with an engineered two-domain recombinant protein, Phys Status Solidi B- Basic Res, 224, 1: 277–283. 52. Moore DE and Patel K (2001). Q-CdS photoluminescence activation on Zn2+ and Cd2+ salt introduction, Langmuir, 17, 8: 2541–2544. 53. Lynes MA, Kang YJ, Sensi SL, Perdrizet GA, et al. (2007). Heavy metal ions in normal physiology, toxic stress, and cytoprotection, Ann N Y Acad Sci, 1113: 159–172. 54. Duffus JH (2002). ‘‘Heavy metals’’ – A meaningless term? Pure Appl Chem, 74, 5: 793–807. 55. Chen YF and Rosenzweig Z (2002). Luminescent CdS quantum dots as selective ion probes, Anal Chem, 74, 19: 5132–5138. 56. Li J, Bao DS, Hong X, et al. (2005). Luminescent CdTe quantum dots and nanorods as metal ion probes, Colloids Surf A, 257, 58: 267–271. 57. Chen JL and Zhu CQ (2005). Functionalized cadmium sulfide quantum dots as fluorescence probe for silver ion determination. Anal Chim Acta, 546, 2: 147–153. 58. Wang XJ, Ruedas-Rama MJ, and Hall EAH (2007). The emerging use of quantum dots in analysis, Anal Lett, 40, 8: 1497–1520. 59. Isarov AV and Chrysochoos J (1997). Optical and photochemical properties of nonstoichiometric cadmium sulfide nanoparticles: Surface modification with copper(II) ions, Langmuir, 13, 12: 3142–3149. 60. Liang JG, Ai XP, He ZK, et al. (2004). Functionalized CdSe quantum dots as selective silver ion chemodosimeter, Analyst, 129, 7: 610. 61. Xie HY, Liang HG, Zhang ZL, et al. (2004). Luminescent CdSe-ZnS quantum dots as selective Cu2+ probe, Spectrochim Acta Part A, 60, 11: 2527–2530. 62. Bo C and Ping Z (2005). A new determining method of copper(III) ions at ng ml(-1) levels based on quenching of the water-soluble nanocrystals fluorescence, Anal Bioanal Chem, 381, 4: 986–992. 63. Fernandez-Arguelles MT, Jin WJ, Costa-Fernandez JM, et al. (2005). Surface-modified CdSe quantum dots for the sensitive and selective determination of Cu(II) in aqueous solutions by luminescent measurements, Anal Chim Acta, 549, 1–2: 20–25. 64. Gattas-Asfura KA and Leblanc RM (2003). Peptide-coated CdS quantum dots for the optical detection of copper(II) and silver(I), Chem Commun, 21: 2684–2685. 65. Lakowicz JR, Gryczynski I, Gryczynski Z, et al. (1999). Luminescence spectral properties of CdS nanoparticles, J Phys Chem B, 103, 36: 7613–7620. 66. Sarkar SK, Chandrasekharan N, Gorer S, et al. (2002). Reversible adsorption-enhanced quantum confinement in semiconductor quantum dots, Appl Phys Lett, 81, 26: 5045–5047. 67. Jin WJ, Fernandez-Arguelles MT, Costa-Fernandez JM, et al. (2005). Photoactivated luminescent CdSe quantum dots as sensitive cyanide probes in aqueous solutions, Chem Commun, 7: 883–885. 68. Chen CY, Cheng CT, Lai CW, et al. (2006). Potassium ion recognition by 15-crown-5 functionalized CdSe/ZnS quantum dots in H2O, Chem Commun, 3: 263–265. 69. Goldman ER, Balighian ED, Mattoussi H, et al. (2002). Avidin: A natural bridge for quantum dot-antibody conjugates, J Am Chem Soc, 124, 22: 6378–6382. 70. Goldman ER, Balighian ED, Kuno MK, et al (2002). Luminescent quantum dot-adaptor protein-antibody conjugates for use in fluoroimmunoassays, Phys Stat Solidi B, 229, 1: 407–414.

178

J. Goicoechea et al.

71. Makrides SC, Gasbarro C, and Bello JM (2005). Bioconjugation of quantum dot luminescent probes for Western blot analysis, Biotechnology, 39, 4: 501–506. 72. Hoshino A, Fujioka K, Manabe N, et al. (2005). Simultaneous multicolor detection system of the single-molecular microbial antigen with total internal reflection fluorescence microscopy. Microbiol Immun, 49, 5: 461–470. 73. Bakalova R, Zhelev Z, Ohba H, et al (2005). Quantum dot-based western blot technology for ultrasensitive detection of tracer proteins, J Am Chem Soc, 127, 26: 9328–9329. 74. Goldman ER, Clapp AR, Anderson GP, et al. (2003). Multiplexed toxin analysis using four colors of quantum dot fluororeagents, Anal Chem, 76, 3: 684–688. 75. Michalet X, Pinaud FF, Bentolila LA, et al. (2005). Quantum dots for live cells, in vivo imaging, and diagnostics, Science, 307, 5709: 538–544. 76. RG. Neuhauser RG, Shimizu KT, Woo WK, et al. (2000). Correlation between fluorescence intermittency and spectral diffusion in single semiconductor quantum dots, Phys Rev Lett, 85, 15: 3301–3304. 77. Tada H, Higuchi H, Wanatabe TM, et al. (2007). In vivo real-time tracking of single quantum dots conjugated with monoclonal anti-HER2 antibody in tumors of mice, Cancer Res, 67, 3: 1138–1144. 78. Courty S, Luccardini C, Bellaiche Y, et al. (2006). Tracking Individual Kinesin Motors in Living Cells Using Single Quantum-Dot Imaging, Nano Lett, 6, 7: 1491–1495. 79. Xiao Y and Barker PE (2004). Semiconductor nanocrystal probes for human metaphase chromosomes, Nucl Ac Res, 32, 3: E28. 80. Wu SM, Zha X, Zhang ZL, et al. (2006). Quantum-dot-labeled DNA probes for fluorescence in situ hybridization (FISH) in the microorganism Escherichia coli, Chemphyschem, 7, 5: 1062–1067. 81. Xu HX, Sha MY, Wong EY, et al. (2003). Multiplexed SNP genotyping using the Qbead (TM) system: a quantum dot-encoded microsphere-based assay, Nucl Ac Res, 31, 8: E43. 82. Gerion D, Chen FQ, Kannan B, et al. (2003). Room-temperature single-nucleotide polymorphism and multiallele DNA detection using fluorescent nanocrystals and microarrays, Anal Chem, 75, 18: 4766–4772. 83. Patolsky F, Gill R, Weizmann Y, et al. (2003). Lighting-up the dynamics of telomerization and DNA replication by CdSe-ZnS quantum dots, J Am Chem Soc, 125, 46: 13918–13919. 84. Kim JH, Morikis D, and Ozkan M (2004). Adaptation of inorganic quantum dots for stable molecular beacons, Sens Act B, 102, 2: 315–319. 85. Tyagi S and Kramer FR (1996). Molecular beacons: Probes that fluoresce upon hybridization, Nat Biotech, 14, 3: 303–308. 86. Cai W, Hsu AR, Li ZB, et al (2007). Are quantum dots ready for in vivo imaging in human subjects? Nanoscale Res Lett, 2, 6: 265–281. 87. Tokumasu F and Dvorak J (2003). Development and application of quantum dots for immunocytochemistry of human erythrocytes, J Microsc, 211: 256–261. 88. Lacoste TD, Michalet X, Pinaud F, et al. (2000). Ultrahigh-resolution multicolor colocalization of single fluorescent probes, Proc Natl Acad Sci, 97, 17: 9461–9466. 89. Smith AM, Dave S, Nie SM, et al. (2006). Multicolor quantum dots for molecular diagnostics of cancer, Exp Rev Mol Diagn, 6, 2: 231–244. 90. Jaiswal JK, Mattoussi H, Mauro JM, et al. (2003). Long-term multiple color imaging of live cells using quantum dot bioconjugates, Nat Biotech, 21, 1: 47–51. 91. Lidke DS, Nagy P, Heintzmann R, et al. (2004). Quantum dot ligands provide new insights into erbB/HER receptor-mediated signal transduction, Nat Biotech, 22, 2: 198–203. 92. Sukhanova A, Devy J, Venteo L, et al. (2004). Biocompatible fluorescent nanocrystals for immunolabeling of membrane proteins and cells, Anal Biochem, 324, 1: 60–67. 93. Minet O, Dressler C, and Beuthan J (2004). Heat stress induced redistribution of fluorescent quantum dots in breast tumor cells, J Fluoresc, 14, 3: 241–247.

6 Quantum Dots for Sensing

179

94. Rosenthal SJ, Tomlinson A, Adkins EM, et al. (2002). Targeting cell surface receptors with ligand-conjugated nanocrystals, J Am Chem Soc, 124, 17: 4586–4594. 95. Yao J, Larson DR, Vishwasrao HD, et al. (2005). Blinking and nonradiant dark fraction of water-soluble quantum dots in aqueous solution, Proc Natl Acad Sci, 102, 40: 14284–14289. 96. Hohng S and Ha T (2004). Near-complete suppression of quantum dot blinking in ambient conditions, J Am Chem Soc, 126: 1324–1325. 97. Larson DR, Zipfel WR, Williams RM, et al. (2003). Water-soluble quantum dots for multiphoton fluorescence imaging in vivo, Science, 300: 1434–1436. 98. Dahan M, Levi S, Luccardini C, et al. (2003). Diffusion dynamics of glycine receptors revealed by single-quantum dot tracking, Science, 302: 442–445. 99. Hanaki K, Momo A, Oku T, et al. (2003). Semiconductor quantum dot/albumin complex is a long-life and highly photostable endosome marker, Biochem Biophys Res Commun, 302, 3: 496–501. 100. Chen FQ and Gerion D (2004). Fluorescent CdSe/ZnS nanocrystal-peptide conjugates for long-term, nontoxic imaging and nuclear targeting in living cells, Nano Lett. 4, 10: 1827–1832. 101. Kaul Z, Yaguchi T, Kaul SC, et al. (2003). Mortalin imaging in normal and cancer cells with quantum dot immuno-conjugates, Cell Res, 13, 6: 503–507. 102. Mansson A, Sundberg M, Balaz M, et al. (2004). In vitro sliding of actin filaments labelled with single quantum dots, Biochem Biophys Chem Commun, 314, 2: 529–534. 103. Ishii D, Kinbara K, Ishida Y, et al. (2003). Chaperonin-mediated stabilization and ATPtriggered release of semiconductor nanoparticles, Nature, 423, 6940: 628–632. 104. Zhu L, Ang S, and Liu WT (2004). Quantum dots as a novel immunofluorescent detection system for Cryptosporidium parvum and Giardia lamblia, App Environ Microbiol, 7, 1: 597–598. 105. Tully E, Hearty S, Leonard P, et al. (2006). The development of rapid fluorescence-based immunoassays, using quantum dot-labelled antibodies for the detection of Listeria monocytogenes cell surface proteins, Int J Biol Macromol, 39, 1–3: 127–134. 106. Hahn MA, Tabb JS, and Krauss TD (2005). Detection of single bacterial pathogens with semiconductor quantum dots, Anal Chem, 77, 15: 4861–4869. 107. Yang LJ and Li YB (2006). Simultaneous detection of Escherichia coli O157 : H7 and Salmonella Typhimurium using quantum dots as fluorescence labels, Analyst, 131, 3: 394–401. 108. Mattheakis LC, Dias JM, Choi YJ, et al. (2004). Optical coding of mammalian cells using semiconductor quantum dots, Anal Biochem, 327, 2: 200–208. 109. Ramachandran S, Merrill NE, Blick RH, et al. (2005). Colloidal quantum dots initiating current bursts in lipid bilayers, Biosens Bioelectr, 20, 10: 2173–2176. 110. Maysinger D, Lovric J, Eisenberg A, et al. (2007). Fate of micelles and quantum dots in cells, Eur J Pharma Biopharma, 65: 270–281. 111. Pellegrino T, Parak WJ, Boudreau R, et al. (2003). Quantum dot-based cell motility assay, Differentiation, 71: 542–548. 112. Winter JO, Liu TY, Korgel BA, et al. (2001). Recognition molecule directed interfacing between semiconductor quantum dots and nerve cells, Adv Mat, 13, 22: 1673–1677. 113. Lim YT, Kim S, Nakayama A, et al. (2003). Selection of quantum dot wavelengths for biomedical assays and imaging, Mol Imaging, 2, 1: 50–64. 114. Bremer C, Tung CH, and Weissleder R (2001). In vivo molecular target assessment of matrix metalloproteinase inhibition, Nat Med, 7: 743–748. 115. Weissleder R, Tung CH, and Mahmood U (1999). In vivo imaging of tumors with protease-activated near-infrared fluorescent probes, Nat Biotech, 17: 375–378. 116. Cerussi AE, Berger AJ, Bevilacqua F, et al. (2001). Sources of absorption and scattering contrast for near-infrared optical mammography, Acad Radiol. 8, 3: 211–218.

180

J. Goicoechea et al.

117. Voura EB, Jaiswal JK, Mattoussi H, et al. (2004). Tracking metastatic tumor cell extravasation with quantum dot nanocrystals and fluorescence emission-scanning microscopy, Nat Med, 10, 9: 993–998. 118. Rieger S, Kulkarni RP, Darcy D, et al. (2005). Quantum dots are powerful multipurpose vital labeling agents in zebrafish embryos, Dev Dyn, 234, 3: 670–681. 119. Serrano EE and Knight VB, (2005). Multiphoton imaging of quantum dot bioconjugates in cultured cells following Nd: YLF laser excitation, Proc SPIE, 5705: 225–234. 120. Levene MJ, Dombeck DA, Kasischke KA, et al. (2004). In Vivo Multiphoton Microscopy of Deep Brain Tissue, J Neurophysiol, 91: 1908–1912. 121. Stroh M, Zimmer JP, Duda DG, et al. (2005). Quantum dots spectrally distinguish multiple species within the tumor milieu in vivo, Nat Med, 11, 678–682. 122. Smith JD, Fisher GW, Waggoner AS, et al. (2007). The use of quantum dots for analysis of chick CAM vasculature, Microvasc Res, 73, 2: 75–83. 123. Leng T, Miller JM, Bilbao KV, et al. (2004). The chick chorioallantoic membrane as a model tissue for surgical retinal research and simulation, Retina 24, 3: 427–434. 124. Ballou B, Lagerholm BC, Ernst LA, et al. (2004). Noninvasive imaging of quantum dots in mice, Bioconj Chem, 15, 1: 79–86. 125. Kim S, Fisher B, Eisler HJ, et al. (2003). Type-II Quantum Dots: CdTe/CdSe (Core/ Shell) and CdSe/ZnTe(Core/Shell) Heterostructures, J Am Chem Soc, 125, 38: 11466–11467. 126. Kim SW, Zimmer JP, Ohnishi S, et al. (2005). Engineering InAsxP1-x/InP/ZnSe III-V alloyed core/shell quantum dots for the near-infrared, J Am Chem Soc, 127, 30: 10526–10532. 127. Parungo CP, Colson YL, Kim SW, et al. (2005). Sentinel lymph node mapping of the pleural space, Chest, 127, 5: 1799–1804. 128. Parungo CP, Soybel DI, Colson YL, et al. (2007). Lymphatic drainage of the peritoneal space: a pattern dependent on bowel lymphatics, Ann Surg Oncol, 14, 2: 286–298. 129. Frangioni JV, Kim SW, Ohnishi S, et al. (2007). Sentinel lymph node mapping with typeII quantum dots, Methods Mol Biol, 374: 147–159. 130. Soltesz EG, Kim S, Laurence RG, et al. (2005). Intraoperative sentinel lymph node mapping of the lung using near-infrared fluorescent quantum dots, Ann Thorac Surg, 79, 1: 269–277. 131. Parungo CP, Ohnishi S, Kim SW, et al. (2005). Intraoperative identification of esophageal sentinel lymph nodes with near-infrared fluorescence imaging, J Thorac Cardiovasc Sur, 125, 4: 844–850. 132. Soltesz EG, Kim S, Kim SW, et al (2006). Sentinel lymph node mapping of the gastrointestinal tract by using invisible light, Ann Surg Oncol, 13, 3: 386–396. 133. Thorne RG and Nicholson C (2006). In vivo diffusion analysis with quantum dots and dextrans predicts the width of brain extracellular space, Proc Natl Acad Sci, 103, 14: 5567–5572. 134. Muhammad O, Popescu A, and Toms SA (2007). Macrophage-mediated colocalization of quantum dots in experimental glioma, Methods Mol Biol, 374: 161–171. 135. Popescu MA and Toms SA (2006). In vivo optical imaging using quantum dots for the management of brain tumors, Exp Rev Mol Diagn, 6, 6: 879–890. 136. Jackson H, Muhammad O, Daneshvar H, et al. (2007). Quantum dots are phagocytized by macrophages and colocalize with experimental gliomas, Neurosurgery, 60, 3: 524–529. 137. Derfus AM, Chan WCW, and Bhatia SN (2004). Probing the cytotoxicity of semiconductor quantum dots, Nano Lett, 4, 1: 11–18. 138. Jiang W, Singhal A, and Kim BY (2008). Assessing near-infrared quantum dots for deep tissue, organ, and animal imaging applications, J Assoc Lab Autom, 13, 1: 6–12. 139. Cai W, Shin DW, Chen K, et al. (2006). Peptide-Labeled Near-Infrared Quantum Dots for Imaging Tumor Vasculature in Living Subjects, Nano Lett, 6, 4: 669–676.

6 Quantum Dots for Sensing

181

140. Yu X, Chen L, Li K, et al. (2007). Immunofluorescence detection with quantum dot bioconjugates for hepatoma in vivo, J Biomed Opt, 12, 1: 014008. 141. Alivisatos AP, Gu WW, and Larabell C (2005). Quantum dots as cellular probes, Annual Rev Biomed Eng, 7: 55–76. 142. Landin L, Miller MS, Pistol ME, et al. (1998). Optical studies of individual InAs quantum dots in GaAs: few-particle effects, Science, 280: 262–264. 143. Tanaka N, Yamasaki J, Fuchi S, et al. (2004). First Observation of InxGa1-xAs Quantum Dots in GaP by Spherical-Aberration-Corrected HRTEM in Comparison with ADF-STEM and Conventional HRTEM, Microsc Microanal, 10, 1: 139–145. 144. Zimmer JP, Kim SW, Ohnishi S, et al. (2006). Size series of small indium arsenide-zinc selenide core-shell nanocrystals and their application to in vivo imaging, J Am Chem Soc, 128, 8: 2526–2527. 145. Liu Y, Kim M, Wang Y, et al. (2006). Highly Luminescent, Stable, and Water-Soluble CdSe/CdS Core-Shell Dendron Nanocrystals with Carboxylate Anchoring Groups, Langmuir, 22, 14: 6341–6345. 146. So MK, Xu C, Loening AM, Gambhir SS, et al. (2006). Self-illuminating quantum dot conjugates for in vivo imaging, Nat Biotech, 24: 339–343. 147. Walker GW, Sundar VC, and Rudzinski CM (2003). Quantum-dot optical temperature probes, Appl Phys Lett, 83: 3555–3557. 148. de Bastida G, Arregui FJ, Goicoechea J, et al. (2006). Quantum dots-based optical fiber temperature sensors fabricated by layer-by-layer, IEEE Sens J, 6, 6: 1378–1379. 149. Bravo J, Matias IR, Del Villar I, et al. (2006). Nanofilms on hollow core fiber-based structures: an optical study, J Lightwave Technol, 24, 5: 2100–2107. 150. Matias IR, Arregui FJ, and Corres JM (2007). Evanescent field fiber-optic sensors for humidity monitoring based on nanocoatings, IEEE Sens J, 7, 1: 89–95. 151. Bravo J, Goicoechea J, Corres JM, et al. (2007). Fiber optic temperature sensor depositing quantum dots inside hollow core fibers using the layer by layer technique, Proc SPIE, doi: 10.1117/12.738388. 152. Constantine CA, Gatta´s-Asfura KM, Mello SV, et al. (2003). Layer-by-layer biosensor assembly incorporating functionalized quantum dots, Langmuir, 19, 23: 9863–9867. 153. Constantine CA, Gatta´s-Asfura KM, Mello SV, et al. (2003). Layer-by-Layer films of chitosan, organophosphorus hydrolase and thioglycolic acid-capped CdSe quantum dots for the detection of paraoxon, J Phys Chem B, 107, 50: 13762–13764. 154. Goldman ER, Medintz IL, Whitley JL, et al. (2005). A hybrid quantum dot-antibody fragment fluorescence resonance energy transfer-based TNT sensor, J Am Chem Soc, 127, 18: 6744–6751. 155. Yildiz I, Tomasulo M, and Raymo FM (2006). A mechanism to signal receptor-substrate interactions with luminescent quantum dots, Proc Natl Acad Sci, 103, 31: 11457–11460. 156. Suffern D, Clarke SJ, Hollmann CA, et al. (2006). Labeling of subcellular redox potential with dopamine-conjugated quantum dots, Proc. SPIE, 6096: 60960O. 157. Clarke SJ, Hollmann CA, Zhang Z, et al. (2006). Photophysics of dopamine-modified quantum dots and effects on biological systems, Nat Mater, 5, 5: 409–417. 158. Jiang H, Yao X, Che J, et al. (2004). Preparation of ZnSe quantum dots embedded in SiO2 thin films by sol-gel process, Ceram Int, 30, 7: 1685–1689. 159. Lin I, Joseph AK, Chang CK, et al. (2004). Synthesis and photoluminescence study of molecularly imprinted polymers appended onto CdSe/ZnS core-shells, Biosens Bioelectr, 20, 1: 127–131.

Chapter 7

Nanostructured Magnetic Sensors Manuel Va´zquez, Aqustina Asenjo, Maria del Puerto Morales, Kleber Roberto. Pirota, Giovanni Badini-Confalonieri, and Manuel Herna´ndez-Ve´lez

7.1 Introduction: Magnetic Sensors and Nanostructures The world of sensors is characterized by a large variety of devices, appropriately selected to answer the needs to detect different external conditions. For instance, multitudes of thermal sensors are present as contact and non-contact sensors, devices based on the measure of heat flux or temperature changes, and basic sensing technologies for temperature probes include thermocouple [1], thermistor [2], solid state [3] and fibre optic [4]. Magnetic sensors are equally variegated in their types and shapes according to the different physical principles they make use of: from those that are actuated by the presence of a permanent magnet, where their operating principle is based on the use of reed contacts, to fluxometric sensors which use the voltage induced by the changes in permeability of a ferromagnetic core. Most common types of magnetic sensors are fluxgate Sensors [5], magnetoresistors making use of the anisotropic behaviour of the magnetoresistance in the core [6], Hall-effect magnetic sensors [7], magneto-optical sensors based on Faraday and Kerr effects, resonance magnetometers, SQUID and spin valve-based sensors [8]. Chemical sensors include solid-state gas sensors [9] that measure a physical property changed by a reaction at the surface, and solid electrolytes, which measure electrical conductivity changes. Catalytic sensors, such as the pellistor, measure temperature changes due to heat of reaction at the surface [10]. On the other hand biosensors contain a part that has a biological origin, such as an enzyme or an antibody. The biological component is in contact with a suitable physical transducer that converts the biological signal into an electrical one [11]. Stress, strain, pressure [12] and torque sensors [13] are also represented by a large variety of devices, ranging from strain gauges, making use of light beams and photocell detector, or detecting electromagnetic properties of a material [14], to amorphous ribbons, which have been used to measure changes in curvatures by observing its magnetic response to applied stresses and have been earlier proved a valuable way to develop a stress sensor device. Finally, it should be mentioned orientation and position sensors which are relevant particularly for aerospace and automative industries [15]. M. Va´zquez Instituto de Ciencia de Materiales, CSIC 28049, Madrid, Spain

F.J. Arregui (ed.), Sensors Based on Nanostructured Materials, DOI: 10.1007/978-0-387-77753-5_7, Ó Springer ScienceþBusiness Media, LLC 2009

183

184

M. Va´zquez et al.

What has been just mentioned is only the tip of an iceberg in the large panorama of sensor families. In this scenario, the presence of a sensor family able to detect multiple sensing environments brings immediate advantages in terms of simplifying the technology, improving measurement rate efficiencies and reducing costs. The main candidate to this role is, at present, represented by magnetic sensor and, more specifically, nanostructure-based magnetic sensors. The high interest from scientific and industrial communities in nanostructure-related materials and technologies arises from the fact that some structures, usually smaller than 100 nm, have new properties and behaviour that are not exhibited by the bulk matter of the same composition; new effects appear and play an important role that is often related to quantum mechanics and quantum mechanisms. These new properties depend on the size as well as the geometry of the nanostructures and can be used to the great advantage of many technological applications. These materials can be used in sensors whose employment is found in a wide range of environments, from biomedical, optical to various physical and chemical sensors, including, of course, the category of magnetic sensors.

7.1.1 About Magnetic Nanostructures The scientific and technological development in magnetism and magnetic materials within the last two decades has reached that point that is nowadays possible to fabricate, study and manipulate the matter at the nanoscale in a controlled way. Usually, one speaks of nanoscale when thinking in the range 1–100 nm, although the limits should be kept flexible. This search to reduce the size of investigated materials has been made possible, thanks to the particular development of experimental equipments. Such advance is enabled first by the optimization of the equipments to fabricate in a controlled way novel magnetic nanoscaled materials, magnetic nanomaterials or magnetic nanostructures. This is particularly evident, for example, when considering the sophisticated facilities that are required to prepare magnetic heterostructures (i.e. where clean rooms are needed) by a number of techniques, or the complex multi-step tasks required for multicomponent composite magnetic magnets. The development of techniques to prepare magnetic nanostructures is nevertheless not sufficient since their nanoscale structure has to be unveiled. Geometrical, compositional, morphological and particularly, magnetic characteristics of what has been fabricated need to be known. The development of a number of techniques has allowed one that achievement. For example, while chemistry routes to synthesize magnetic materials (i.e. nanoparticles) follow similar trends as those classically employed, their full characterization is being only possible by specific advanced techniques (i.e. high-resolution scanning electron microscopy). The parallel evolution of equipments to fabricate and characterize magnetic nanostructured materials has permitted the study of their magnetic behaviour, which together with parallel theoretical studies has given rise to the discovery of new magnetic phenomena directly ascribed to the nanoscale dimensions of the

7 Nanostructured Magnetic Sensors

185

magnetic nanomaterials (i.e. quantum Hall effect). The term nanoscience has been connected to these phenomena in magnetic nanostructures. The final technological development, or nanotechnology, is related to the possibility of manipulating the magnetic matter in a controlled way. This is the final step at the laboratory scale before the commercial use of scientific advances. Possibly, the best example of such conjunction between nanomaterials, nanoscience and nanotechnology has been awarded this year with the acknowledgement by the Swedish Academy to the discovery and understanding of the giant magnetoresistance, GMR, effect to Albert Fert and Peter Gru¨nberg. This is certainly a clear evidence of the particular correlation between the magnetic mentioned terms: the magnetic nanomaterials, a new effect in nanoscience and the magnetic recording-related technological development. The GMR phenomenon was discovered by 1988 in particular magnetic nanostructures consisting of layered films with different magnetic nature. It was only possible after the controlled fabrication of such magnetic nanostructures and the characterization of their magnetic-dependent electrical resistance behaviour. A new scientific effect was discovered: the different scattering of electrons when traversing layers with different magnetic nature which should be controllable by the action of applied magnetic fields. The final technological development has been quite rapidly achieved: the high variation of resistivity owing to the presence of a neighbouring magnetic field modifies the resistance of a multilayer which is used as magnetic field sensor produced by or at the border between magnetic bits of information. By the second half of the 1990s magnetic recording heads were already commercially available enabling a faster increase of the areal density of information. It was allowed, thanks to the increased sensitivity of GMR heads with regard to previously used reading heads. The present chapter is devoted to sensor applications of magnetic nanostructures. As we have described before, novel applications are closely related to parallel advances in related topics. Particularly, nanoscience and nanotechnology is a multidisciplinary topic involving disciplines as physics, chemistry, biology, materials sciences or engineering materials. Magnetic sensors based on magnetic nanostructures actually profit by the development in parallel achievements. In this chapter we summarize some of the most attractive applications making use of magnetic nanostructures. We should emphasize here that apart from the mentioned importance of scientific and technological components, the economical issue should never be forgotten. Most applications require an optimization with regard to prize reduction: only those technological achievements with suitable economical fittings overcome the commercial barrier. The chapter is divided into several sections. After these introductory comments, a first section is devoted to the main techniques that are currently employed for the fabrication of magnetic nanostructures. Various methods to synthesize magnetic nanoparticles are first mentioned, which is followed by a general description of the experimental techniques to prepare magnetic nanowires and 2D magnetic nanostructures. In the subsequent section, some important methods for the magnetic characterization of magnetic nanostructures are described.

186

M. Va´zquez et al.

Most relevantly, the techniques to study the magnetization process are described as well as the magnetic domain structure of magnetic nanostructures. In the main section of the chapter, different families of magnetic sensor families are presented. We first introduce some technological developments dealing with magnetic nanoparticles: applications dealing with biomedical issues, as, for example, detection and magnetic separation as well as magnetic resonance imaging. One-dimensional magnetic objects, as magnetic nanowires and nanotubes, are mainly employed as probes for a number of magnetic imaging techniques. Concerning the sensor applications of 2D magnetic nanostructures, one should mention the magnetic recording matters as the main fueling for its development. Magnetic recording is presently attracting more than 60% of the global market involving magnetic applications in the commercial activities. The main application regarding magnetic storage is that related to magnetic writing and particularly reading the magnetic information stored in the recording media. But there are many other applications where magnetic nanostructures are employed in different technological developments and sensors involved in nanotechnology. Other applications related to ultrasoft and hard magnetic materials consisting of composite magnetic nanostructures of relevance for soft magnetic applications (i.e. core of transformers and motors) and permanent magnets (i.e. advanced spring magnets) are not considered here. In those cases, magnetic nanostructures are useful in technological applications but not properly as magnetic sensor elements and devices. Magnetic nanostructures are characterized by dimensions somehow in between atomic and bulk scales. From a magnetic viewpoint this is quite important since the magnetic behaviour is determined mainly by the magnetic anisotropy apart from the exchange interaction. In a simple way, magnetic anisotropy can be considered to arise either from spin lattice or from magnetostatic origins. The first one involves magnetocrystalline and magnetoelastic contributions, while the second one is mainly dependent on the geometrical shape of the magnetic material. In our magnetic nanostructures, the shape is actually determinant of the overall magnetic behaviour. In fact, the shape of the magnetic nanostructure is quite important. In the case of isometric magnetic nanoparticles they exhibit spherical symmetry with no predominant direction, but in the case of nanowires the axial orientation determines an easy magnetization direction, while in thin films and layered nanostructures in-plane orientations are favoured. The actual control of shape anisotropy of magnetic nanostructures is really decisive when trying to use them in sensor applications.

7.1.2 About Technological Applications of Magnetic Nanostructures An important number of applications related to magnetic nanoparticles are dealing with nanobiomedicine. It includes mainly diagnosis and therapeutic

7 Nanostructured Magnetic Sensors

187

treatments of different diseases such as cancer. In the last decades diagnostics has evolved towards smart biosensors for quicker, specific and more sensitive detection methods which drive the demand for new multifunctional materials based on magnetic nanoparticles. Other applications of magnetic nanoparticles comprise not only diagnosis but also therapeutic treatments, which are based on targeted drug delivery and hyperthermia. The increased surface area of nanoparticles compared with films or micrometric particles, which enables the immobilization of larger amounts of biologically active molecules on the particles, is one of the main advantages. The second one is the ability to disperse the bioactive magnetic nanoparticles in the analyte solution, which enables rapid contact between active molecules, resulting in a lower limit of detection and faster analysis time. And finally, their magnetic properties permit easy separation and re-use by cycles of magnetic separation, sample replacement and redispersion. Particles should behave as superparamagnetic at room temperature, which means that they behave as magnetic under a magnetic applied field but become non-magnetic as soon as the field is removed. Consequently, once they are attracted by a magnet, particles get no longer attached to each other and disperse themselves back into suspension. Particles developed for this kind of applications generally consist of a magnetic core (usually magnetite Fe3O4 or maghemite -Fe2O3) with a polymer coating, such as PVA or dextran, which can be functionalized so that drugs can be attached to it Fig. 7.1. Particles can be also composed of a porous polymer or block copolymer in which the magnetic particles are dispersed. Finally, magnetic nanoparticles can be incorporated into artificial capsules such as liposomes with a core/shell structure where the magnetic component is precipitated (a)

(b)

Magnetic nanoparticle

Antibody Dextran

Polymer or SiO2

Magnetic nanoparticle probe Conjugated with antibodies (d)

CdSe

S-Dye CO O-

(c)

Magnetic beads

SH COO-S-D ye

OCO SH C O O -

O-

CO SH

Fig. 7.1 Schematic representation of magnetic nanoparticles for biomedical applications consisting of the following: (a) a magnetic core with a polymer coating functionalized to attach drugs; (b) porous polymer or block copolymer particles in which the magnetic particles are dispersed; and (c) and (d) magnetic and fluorescent nanohybrid materials

Polymer or SiO2

Magnetic and fluorescent nanoparticles

188

M. Va´zquez et al.

within a spherical lipid membrane or embedded within a hydrogel along with the therapeutic drug or gene. The preparation of bioactive magnetic nanoparticles from high-quality starting materials is really important because the performance of these particles largely depends on their physical properties, particularly on the shape and size of the particles and their size distribution [16, 17, 18]. Last developments on the synthesis of uniform nanoparticles and nanohybrid materials with controlled composition and complex microstructure are presented here together with their structural and magnetic properties Fig.7.1. Finally some of the most interesting applications in sensing will be described, such as biomolecule separation and NMR imaging. Another interesting field of application involving magnetic nanoparticles, which will not be discussed here in detail, is drug delivery. The use of magnetic micro- and nanoparticles as carriers for in vivo targeting of therapeutic compounds was first proposed 25 years ago. However, delivering these agents to specific targets within the body remains a difficult task and several obstacles need to be overcome before translating it into an effective clinical treatment [19]. Particularly, the development of novel particles with enhanced magnetic properties is required to facilitate a better capture. Nickel-embedded carbon nanotubes have been used coupled with DNA for in vitro gene transfection [20]. Although the results are promising, toxicity effects of carbon nanotubes need to be evaluated. Other proposed materials are cobalt nanoparticles with a gold shell and mesoporous silica nanoparticles containing 80% iron oxide [21]. Finally, magnetic nanoparticles are also tested nowadays for hyperthermia treatment of cancer. The critical challenge for their successful application of magnetic nanoparticle in vivo treatments is to raise the working temperature in the range between 42 and 46˚C under a physiological tolerance range of applied frequencies and magnetic fields, that is f ¼ 50–100 kHz and H ¼ 0–15 kAm–1 [22]. Recently, nanotechnology and nanostructure-related applications have been at the centre of strong debate in and outside the scientific world. The appropriateness of using nanomaterials in today’s devices has been extensively discussed in all its aspects from health to ethic, from economic to performance. While there is no doubt that all the possible precautions must be taken in order to guarantee a safe and risk-free use of nanostructures, it is equally true that these materials can be exploited to very good advantages being versatile to and compatible with a huge number of different applications. Thanks to this versatility and excellent performances, nanoscale structures and materials have recently found a fruitful ground in the field of magnetic sensor and, among the various nanostructures, to 1D and 2D magnetic systems. As will be later discussed, there are quite a number of applications making use of magnetic nanowires and films. In many cases, nanowires are used as sensors for field detection, and related applications. Many sensors are based on scanning a magnetic probe so ‘‘reading’’

7 Nanostructured Magnetic Sensors

189

the magnetic information. Important physical phenomena are used as for example the Josephson effect or tunnel junctions. Magnetic nanosensors require very specific preparation techniques as will be described. The particular nature of wires, with a defined orientation, determines a directional magnetic anisotropy in opposition to nanoparticles, mostly with spherical shape where no specific orientation is favoured. Such directionality in nanowires and nanotubes makes them attractive to be employed as probes in magnetic imaging techniques. A number of magnetic microscopies allow one to determine magnetic characteristics of surfaces and films. Thin films and heterostructures have been proved to be very successful in sensing applications. In this case, the planar configuration determines in-plane anisotropy. In turn, applications of 2D magnetic nanostructures are mainly useful in connection with magnetic recording technologies, the most important from the market point of view. That is why we will devote an important part of the section on technological applications of 2D nanostructures to magnetic recording. The preparation of 2D nanostructures needs sophisticated fabrication and characterization techniques, quite necessary to achieve full control of quality of the final materials in magnetic recording. That involves not only recording media itself but also the development of magnetic sensors.

7.2 Sensing Elements: Synthesis and Magnetic Characterization 7.2.1 Magnetic Nanoparticles: Synthesis Two different approaches can be considered to synthesize magnetic nanoparticles: the top-down approach, which utilizes physical methods, and the bottomup approach, which employs solution-phase chemistry [23]. The advantage of the first methods is the capability of production in large quantities whereas the synthesis of uniform-sized nanoparticles is difficult to achieve. In contrast, chemical methods can be used to synthesize uniform nanoparticles with controlled size and shape but usually only milligrams are produced [24]. In this section, we focus on the chemical synthesis methods for producing uniform nanoparticles.

7.2.1.1 Chemical Synthesis of Magnetic Nanoparticles Co-precipitation under alkaline conditions is a currently used method for the synthesis of iron oxide nanoparticles applied in biotechnology and medicine. Through this method particles are produced with sizes between 5 and 10 nm in diameter and broad size distribution Fig. 7.2a [25]. Larger particles, with average size of 20–40 nm, can be obtained in aqueous solutions by oxidation of Fe(II) salts with H2O2 [26] or KNO3 [27] under mild alkaline conditions (Fig. 7.2b) or by hydrothermal growing of freshly prepared magnetite particles

M. Va´zquez et al.

190 Fig. 7.2 Fe3O4 nanoparticles prepared by different methods: (a) coprecipitation; (b) oxidation of a Fe(II) salt; (c) decomposition of organic precursor; and (d) laser pyrolysis

a

b

100 nm

c

d

100 nm

20 20nm nm

by the Massart route [28]. Recently, a bioinspired route for the synthesis of uniform magnetite nanocrystals has been proposed. The synthesis is performed in the presence of a recombinant protein forming a polymeric gel, which is responsible for slowing down the diffusion rates of the reagents and results in uniform 30 nm Fe3O4 nanoparticles [29]. Important advances have been achieved by using organic solvent surfactant methodology which leads to great improvement in particle size distribution and crystallinity (Fig. 7.2c) [30, 31]. By seed-mediated growth of previously synthesized monodispersed nanoparticles seeds, 1 nm diameter control of uniform magnetite nanoparticles can be achieved [32]. Nanoparticles of pure metals such as Co [33] and alloys like FePt [34], CoPt, CoNi have been synthesized by codecomposition and by sequential decomposition of two different metal ion complexes, respectively (Fig. 7.3a and b). However, the resulting particles, always smaller than 20 nm, were capped with surfactant molecules and therefore miscible in non-polar solvents. Different strategies have been described in the literature for replacing the surfactant including ligand exchange, surface silanization and polymer or micelle coating [35, 36, 37, 38, 39, 40]. Alternative routes for the preparation of magnetic nanoparticles are laser and thermal pyrolysis of vapours and aerosols [41, 42, 43, 44]. Thus, laser-induced pyrolytic reactions have been reported to produce pure maghemite nanoparticles with narrow particle size distributions, non-aggregated and free of impurities [41, 42] (Fig. 7.2d). This method of synthesis involves heating a flowing mixture of gases (C2H4 and Fe(CO)5) with a continuous wave carbon dioxide laser, which initiates and sustains a chemical reaction. The C2H4 absorbs the laser energy and a portion of this energy is transferred and absorbed by Fe(CO)5, which is rapidly heated and decomposed resulting in an atomic Fe saturated

7 Nanostructured Magnetic Sensors Fig. 7.3 Alloys and core/ shell nanoparticles: (a) FePt, (b) CoNi, (c) FePt/Au and (d) Fe3O4/SiO2. Inset in (b) shows the metallic-core/ oxide-shell structure of a nanoparticle

191

a

b

20 nm

5.0 nm

100 nm

d

c

20 nm

100 nm

vapour leading to the nucleation and growth of iron metal nucleus. Either uniform iron oxide or metal iron nanoparticles can be obtained in a continuous way by controlling the oxidation procedure [45]. In order to impart unique biological function and targeting to the nanoparticles, it is often necessary to attach affinity ligand to such materials. A number of different covalent conjugation strategies using amine (–NH2), carboxyl (COOH), aldehyde (CHO) and thiol (SH) groups exposed on the surface of nanoparticles have been developed [46]. Conjugation should be robust, cost effective, biocompatible and fast. For example, if conjugation is possible at the same time than synthesis, in one-pot reaction, scaling up the process will be rapid. In this regard, the use of acid and alkyne chemistries has been recently proposed for rapid, site-specific modification of nanoparticles with small molecules which has been labelled ‘‘click chemistry’’ [47]. Non-covalent absorption of ligands does not guarantee that the ligand will remain bound to the nanoparticle surface at various stages of surface modification or during application. Covalent linkage can be accomplished by first coating the surface of the nanoparticles with a functionalized alkoxysilane reagent such as APTES through the silanization reaction [48].

7.2.1.2 Synthesis of Hybrid Nanoparticles Future developments are expected in the synthetic creation of nanohybrid materials, inorganic/organic or inorganic/inorganic, with controlled composition [49].

192

M. Va´zquez et al.

The synthesis of these materials can be carried out by chemical synthesis, i.e. growing one material on top of another, through linkage with biological molecules or packing them in a container, for example using hollow capsules.

Organic/Inorganic Coating with linear dextrans, polymers of –-glucose units, was the earliest attempt to stabilize the particles in solution, with ligand molecules forming hidrophylic brushes around the particle surface [50]. More recently, complementary regulatory proteins such as heparine and most sulphate glycosaminoglycans have been used. Coating the particles has a double function, to stabilize the particles in aqueous solutions and to make them active and selective for a specific receptor. An alternative successful approach is to create a steric surface barrier of sufficient density with non-ionic copolymers such as poloxamers and poloxamines, or polyethyleneglycol (PEG) and derivates. PEG is a linear polyether diol that exhibits a low-degree inmunogenicity. Surface modification with PEG can be performed by adsorption, by incorporation during the production of nanoparticles or by covalent attachment to the surface of the particles. Recently, lipophilic and magnetic materials, or magnetocerasomes, have been prepared by grafting long-chain surfactants to the surface of magnetite particles through siloxane or phosphonic functional groups [51]. Currently, magnetic beads are the most common commercially available nanoparticles for magnetic cell separation. These beads contain small iron oxide particles embedded in a spherical polymer matrix with sizes between 1 and 100 mm. They were first prepared by mixing small grains of magnetic oxides with natural or synthetic polymers, followed by procedures to achieve appropriate particle size [52]. Magnetic iron oxide might also be produced in situ in the polymer phase of the process or by dispersing magnetic oxides in a mixture of highly waterinsoluble compounds and vinyl monomers which is further polymerized. These methods lead to the formation of quite heterogeneous materials. Most uniform particles commonly used are made by mixing Fe(II) with monosized porous polymer beads and its subsequent coating with a new polymer layer to close the pores and keep the iron inside the particles [53]. These magnetic colloids can also be obtained from oil in water magnetic emulsion by a two-step polymer immobilization procedure [54].

Inorganic/Inorganic Coating with inorganic layers such as silica, carbon or gold also provides steric repulsion between particles (Fig. 7.3c and d). The coating by biocompatible materials like gold seems to be the better way to diminish the cytotoxicity grade of magnetic nanoparticles such as FePt [55] (Fig. 3c). In the case of metallic particles, those coatings also prevent oxidation and degradation of the magnetic properties during or after the synthesis.

7 Nanostructured Magnetic Sensors

193

Some commercial magnetic beads are based on silica spheres containing magnetic nanoparticles and they can be prepared either by precipitation of magnetite within long, narrow pores in silica nanoparticles [56] or by coating the magnetic particles with a silica layer through a modified aqueous sol–gel route [57, 58] (Fig. 3d). A similar method has been applied for the preparation of FePt@SiO2 core–shell nanoparticles. In this approach, oleic acid and olyel amine-stabilized FePt nanoparticles are first encapsulated through an aminopropoxysilane (APS) monolayer and then subsequent condensation of triethoxysilane (TEOS) on FePt particle surface. These well-defined FePt@SiO2 core–shell nanoparticles with narrow size distribution become colloidal in aqueous media and can thus be used as carrier fluid for biomolecular complexes. In comparison, the scarce hydrophilic nature of oleic acid monolayers on FePt particle surface yields an edgy partial coating of silica when only TEOS is applied for surface modification [59]. Aerosol pyrolysis is also a highly attractive method for the direct synthesis of magnetic hollow spheres of -Fe2O3 with a surface layer enriched in silica (Fig. 7.2c) [60]. Important advantages of this method are reproducibility, simplicity, continuity and very rapid times involved in the generation of the powders in comparison to methods based on the sequential adsorption of components onto colloidal templates (i.e. layer-by-layer self-assembly) and the further removal of the organic matrix [61]. It should be noted that surface silanol groups on a silica layer surrounding a particle provide an ideal anchorage for the covalent binding of specific ligands, which is very important for the main application of this material as magnetic carrier. Magnetic mesoporous silica spheres were also obtained following an aerosol-assisted route and the adsorption and release performance tested for several drugs [62]. The sol–gel process has been shown to be a useful route for the synthesis of magnetic SiO2 nanocomposites. They can be obtained either by dispersing magnetic nanoparticles in different matrices [63, 64], or by obtaining the magnetic nanoparticles in situ within the matrix [65, 66, 67]. An important advantage of the latter is that it is a direct method, i.e. the porous nature of the matrix formed by sol–gel provides the sites for nucleation of the nanoparticles, minimizes their aggregation and imposes an upper limit on their size. Thus, -Fe2O3 nanoparticles can be embedded in an inert, inorganic, transparent and temperature-resistant silica matrix with excellent optical characteristics. Superparamagnetic porous carbons with tunable magnetic properties have been also prepared by forming superparamagnetic nanoparticles of iron oxide ferrites, which are highly dispersed throughout the porous structure of a widely available commercial activated carbon [68]. The preparation procedure involves, first, filling the porosity of the activated carbon with an appropriate amount of Fe(NO3) dissolved in ethanol and, then, impregnating the dried sample with an organic reducing agent (ethylene glycol). Carbon nanotubes have being coated with magnetic nanoparticles by combining the polymer wrapping and layer-by-layer (LbL) assembly techniques. The particlecoated multiwall carbon nanotubes, MWNTs, are superparamagnetic and can

194

M. Va´zquez et al.

be aligned at room temperature on any substrate by deposition from an aqueous solution in an external field. This composite seems to be an excellent candidate to be used as building blocks for the fabrication of novel composite materials with a preferential orientation of the magnetic carbon nanotubes, CNTs [69]. Gold seems to be an ideal coating owing to its low reactivity by which gold surface can be further functionalized with thiol groups. However, direct coating is very difficult because of the dissimilar nature of the two surfaces. Some progress has been achieved by partial replacement reaction in a polar aprotic solvent and by a reverse microemulsion method leading to gold-coated iron nanoparticles [70]. Sequential decomposition methods and reduction reaction have been used for the synthesis of core–shell heterostructured nanoparticles such as FePt/Au (Fig. 7.3d ) [71] and Co/Au [72]. Other interesting nanohybrid materials are those with interesting bifuntional properties such as Co/CdSe, which can be also prepared by a sequential decomposition method [73]. These hybrid materials based on magnetic and fluorescent nanoparticles have been developed with very complex microstructures [74, 75, 76, 77, 78]. Silica composites containing fluorescent and magnetic particles are also obtained by mixing those particles previously synthesized with tetraethoxisilane. A silica shell trapped several iron oxide and CdTe particles yielding an average total diameter of 50 nm. Liposomes prepared from self-assembling amphiphilic molecules provide a recent technology to encapsulate materials. Their ability to encapsulate watersoluble materials makes them attractive devices for transporting a whole spectrum of molecules, including drugs. The majority of the liposomes are based on phospholipids but also fatty acids and their salts. Derivatives of polymers such as polyoxyethylene and polyglycerol and recently palmitic acid together with cholesterol have been used in the preparation of liposomes with great stability and pH sensitivity [79]. Self-Assembling Supracrystals Some techniques have been recently developed for organizing monodispersed colloidal spheres into 2D and 3D ordered lattices [80]. The morphology of the aggregates changes by varying the size, the concentration, the surface hydrophobicity of the solid support, the charge density of the colloids and the electrolytic properties of the underlying liquid. The collective physical properties of an assembly of nanocrystals are neither those of the isolated particles nor those of the corresponding bulk phase. They may depend upon the shape and the nanocrystal assembly at the mesoscopic scale [81]. Concerning the magnetic properties, the squareness ratio of hysteresis loops for nanocrystals is higher when they are deposited in compact hexagonal networks in comparison to isolated nanocrystals. Experimental and modelled hysteresis loops for a chain-like structure are squarer than those of a well-ordered array of nanocrystals. The linear chains of nanocrystals behave thus as nearly

7 Nanostructured Magnetic Sensors

195

homogeneous nanowires. The flipping of the spins could require higher fields when the nanocrystals are ordered in a long-range superlattice compared to the disordered system, where we have only very short-range order [82]. Collective optical and magnetic properties are observed in Co nanocrystals when organized in 2D superlattices [84] and 3D supracrystals [85, 86] because of dipolar interactions. The magnetic properties also markedly depend on the shape of the arrangement of nanocrystals in the mesoscopic structure. This has also been demonstrated by using 10 nm -Fe2O3 nanocrystals, differing by the number of carbon atoms forming the surfactant used to coat them [87, 88]. Self-assembled FePt magnetic nanoparticles offer the potential to store data at areal densities > 1 Tbit/in.2 [89] (Fig. 7.4). This potential arises from the highanisotropy (L10) phase of approximately equi-atomic FePt that allows particles of 3 nm diameter to be thermally stable at room temperature. However, in order to create the L10 phase of FePt it is necessary to anneal at temperatures in excess of 5008C, typically for times of 30 min. This gives rise to a number of thermally activated processes including the desired phase transformation, oxidation and particle agglomeration. Coercivities of up to 13 kOe at room temperature were readily obtained, demonstrating that self-assembled nanoparticles do indeed offer significant potential as recording media. However, most applications derived from these materials are still in an early stage of technical development.

7.2.2 Magnetic Nanowires and Films: Fabrication Techniques Experimental techniques to grow functional nanostructured materials are briefly described here. We first consider techniques to prepare 1D nanostructures followed by those currently used in the preparation of 2D nanostructures. Some of the techniques employed for nanowires, also common for films, are described in more detail in the second section. 7.2.2.1 Fabrication of Nanowires The controlled production of magnetic nanowires and their arrays is recently attracting great interest owing to their applications in emerging nanotechnologies

Fig. 7.4 Self-assembled FePt magnetic nanoparticles [83]

FePt nanoparticles

50 nm

196

M. Va´zquez et al.

related with magnetic storage, magneto-transport, controlled opto-magnetic response and, in general, for multifunctional sensor applications [90]. Nanowires can be prepared by means of more or less sophisticated techniques involving ultrahigh vacuum techniques, e.g. e-beam lithography, similar to advanced techniques used for the production of nanostructured magnetic thin films and multilayers. Alternative techniques based on self-assembly and filling of templates are being increasingly used (for example, by electroplating filling of ordered nanopores in alumina membranes) [91, 92]. Lithography Methods Top-down and bottom-up strategies have been used and sometimes combined to obtain 1D functional materials. In top-down approach highly ordered nanolines can be obtained by different lithography methods (e.g. by electron beam lithography, EBL [93] or nanoimprint lithography) which can produce features characteristic in the order down to 25 nm [94]. Lithography involves a number of related processes like resist coating, or exposure [95, 96, 97]. Generally speaking, a uniform layer of resist is first deposited onto the surface of an unpatterned film or substrate. The resist, with thickness typically from a few nanometres to a micrometre, can be deposited by spin coating. Selected areas of the resist are subsequently exposed to a radiation source, often through a mask. Upon sufficient exposure, the polymer chains in the resist are either broken or cross-linked leaving to a positive or negative resist, respectively. As the lithography process transforms a 2D pattern into a 3D structure in the resist and eventually in the unpatterned film, the depth profiles in both materials are very important. By selecting a suitable developer, temperature and developing time, one can obtain different tailored profiles in the resist. The pattern transfer can be realized in two general processes: from the resist to the unpatterned film by etching or by post-deposition onto patterned resist by lift-off or electrodeposition. The lithography resolution limit is ultimately determined by the radiation wavelength. Hence lithography is usually categorized by the radiation source as optical, electron-beam, ion-beam or X-ray lithography. In the optical lithography, generally UV light is used ( ¼ 193 nm for ArF laser). In the electronbeam lithography, an electron-sensitive resist is exposed on an electron beam. This exposure is usually done using the electron source of the scanning electron microscopy (SEM) or the transmission electron microscopy (TEM). This method combined with, for example, electroplating technique enables the fabrication of patterned elements of high aspect ratio as nanowires. In the case of X-ray lithography, the key point is the exposure of a resist to X-ray radiation in a parallel replication process. Similar to the electron-beam lithography, the sample is covered by a resist layer with high sensitivity in the X-ray wavelength zone. Figure 7.5 shows the SEM image of an array of Ni nanowires fabricated using combined electron-beam lithography and electrodeposition techniques [98].

7 Nanostructured Magnetic Sensors

197

Fig. 7.5 Ni nanocolumns prepared by combined e-beam lithography and electrodeposition [98]

Ni

It is well known that the costs of lithographic techniques increase when the circuit dimensions are diminished which is related with the wavelengths used to define the features through predetermined masks. The improvement of the masks is one of the most significant innovations introduced to increase the resolution of imprint processes on the silicon wafers. For this reason masks constitute an important device which raises the costs of lithographic processes; therefore, nowadays much effort is centred on the production of maskless lithography and nanoimprint lithography [99, 100]. Lithography-based techniques have been very successful and powerful for the fabrication of a large variety of circuit components at nanometric scale. Nevertheless, owing to the limit of the molecule sizes of photosensitive materials [101] and the focusing lengths of the ion beams [102], lithographic methods fail to produce nanostructures with feature sizes smaller than 20 nm. The main determinant process for a self-assembly process in the relief of a solid surface is the creation of a desirable surface structure on which nanowire growth will take place. In that sense, several top-down techniques like molecular beam epitaxy (MBE), electron beam evaporation (EBE), phase-shift optical lithography and sputtering have been used [103, 104] by which continuous and insulated wires could be obtained. Viernow et al. [105] have obtained linear arrays of CaF2 stripes of around 7 nm by self-assembly on silicon <111>. Au nanowires fabrication has been also reported on a previously prepared Si <557> surface [106]. The theoretical explanation of the obtained results predicts the formation of bands on the metallic surface with gaps similar to those described for bulk semiconductors. Other surface manipulation for nanowire growth is based on the stress sources created by misfit dislocations at an interphase [107]. Preferential orientations can appear on the surface due to electric polarity [108]. In this case, the surfaces serve as template as well, determining anisotropic growth of different material species. The processes can be addressed so that vapour or liquid species surrounding the solid surface could be attached to it to compensate daggling bonds or to minimize the surface stress and therefore give place to thermodynamically stable surface morphology. Recently, Wang et al. [109] taking advantage of these facts have reported interesting results on the fabrication of nanocantilever arrays and a varied group of ZnO 1D nanostructures including wires, belts, springs and helices.

198

M. Va´zquez et al.

The development of patterning nanometre scale lines over large areas using nanowires as masks has been recently reported [110]. Core–shell Si–SiO2 nanowires uniaxially compressed on a Langmuir-Blodgett trough were aligned with controlled pitch and then transferred to silicon substrate. Reactive ion etching (RIE) was used to remove the SiO2 shell from the sides and the tops of the nanowires. After the deposition of metals or other materials, the nanowires mask is removed to get parallel lines all over the substrate surface. This constitutes an ingenious nanolithographic method with great potential for the development of integrated nanosystems. Recently the same group using a similar strategy produced assembly of nanowires into integrated device arrays [111]. Based on the advantages offered by this approach such as formation of ordered monolayers over large areas, the facility for transferring the organized monolayers to substrates and the possibility to obtain multilayer by repeating these processes, they have been able to grow aligned nanowires with strict control on several important parameters, as the nanowire pitch and orientation, and the array size. On the other hand, this approach allows one room temperature growth of the arrays and multilayered devices which make it a compatible technique with the required low costs and with the use of flexible substrates for applications in integrated functional nanosystems [112, 113]. Methods based on vapour liquid solid growth (VLS) are commonly used for semiconductor nanowire growth, although laser ablation, chemical vapour deposition (CVD) or template-based synthesis [114, 115, 116, 117] have been also used. Recently, single-electron transistor devices (SET) based on InP have been used which show carrier transport behaviour similar to that shown by carbon nanotubes (CNTs) in this kind of devices [118].

Using Templates and Self-Assembly Self-assembly and filling of templates is a quite versatile technique to fabricate nanoscaled arrays with systematically reproducible properties. A number of novel materials with new and optimized properties can be envisaged, thanks to the capability of manipulating the nanowire nature (i.e. its magnetic properties) and the templates using, i.e., metallic, semiconductor or insulating, which can be tailored by suitable replica–antireplica processes [119]. For example, in the case of composite magnetic materials new phenomena and interactions can be studied in ideally ordered nanomagnets embedded into magnetic/non-magnetic matrices. Low-cost preparation of nanowires is possible with the bottom-up approach by self-assembly of atoms and molecules although it is hard to get them patterned and well arranged [120]. Recently, an excellent review by Y. Xia et al. [121] reports on chemical methods to synthesize 1D nanostructures including wires, belts, rods and tubes. There, molecular wires exhibiting natural growth into 1D are reported with high-anisotropic crystalline structure so that it is not necessary to use surfactant molecules or other elements to co-direct the formation of

7 Nanostructured Magnetic Sensors

199

1D structures. The cleanness of this strategy should be useful for nanoelectronic device fabrication. The template based synthesis strategy has been used in combination with self-assembly processes in a number of works to synthesize nanoparticles, nanowires and thin films. The most successful methods to obtain nanowires and nanowire arrays by using the synthesis on templates and self-assembly are (i) self-assembly of one material in the roughness or relief present on the solid surface, (ii) self-organization within the channels and cavities of porous material and (iii) the formation of inorganic mesoscopic material by self-assembled on surfactants organic molecules. The formation of metallic nanowires and nanochannels through guided selfassembly has recently been reported [122]. In this method, the initiation and termination points of the nanostructure are pre-designed so that the pattern evolution is dictated by stress-assisted cracking on a dielectric film attached to a silicon substrate and then their cracks are used like moulds for the nanowires growth by filling them with a desired material. The technique used to produce the dielectric films that later on will be cracked is plasma enhanced chemical vapour deposition (PECVD). The use of DNA as scaffolds for nanostructure fabrication has already been explored [123] in the last years as well and it has been proposed as a way for nanoelectronic devices construction. Particularly, the fabrication of metallic nanowires has been reported based on the self-assembling of complementary DNA used as localized templates on specific surface sites [124, 125]. Anodization techniques are increasingly being used to create ordered nanoporous structures which can lately be used as templates for growing functional nanowires of different species by using mainly self-assembling process and electroplating techniques. Up to now, aluminium constitutes the unique metal which allows one to fabricate highly ordered nanoporous anodic films (NAFs) with hexagonal symmetry in a densely packed array of nanopores [126, 127]. These NAFs have been specially used to fabricate functional nanowire arrays. The template-synthesis strategy for nanofabrication has been treated in details by Hulteen et al. [128] considered one of the pioneer groups in this subject, particularly in the fabrication of functional nanowire arrays. Recently, similar strategies of synthesizing the anodization of titanium foils have been followed to obtain thin films of self-aligned TiO2 nanotube arrays which have also acquired great interest [129, 130]. The anodization and electroplating processes for tailoring functional 1D nanostructures somehow fulfil some of the demanding requirements of new technologies, such as low costs, repeatability and potential compatibility with silicon-installed technologies which make these nanostructure synthesis routes very attractive. Anodization and electrodeposition processes are detailed and largely described in the current literature [131]. In general, by means of these strategies it is possible to take control of (i) ordering degree, i.e. the size of crystalline single domains (up to several square micrometres), (ii) the single 1D structure diameter (from 15 to 200 nm) and length (from tens to thousands of

M. Va´zquez et al.

200

nanometres) and (iii) the lattice parameter of ordered arrays (between 65 and 500 nm). One of the most interesting characteristics of NAFs is the parallelism of the pores with a symmetry axis perpendicularly oriented to the substrate surfaces. Taking advantage of this particular property, the electrochemical growth technique, either galvanostatic, potentiostatic or a mixture of both methods, is the most used procedure to obtain functionalized NAFs [132, 133, 134]. Thus, NAFs have become an important material in nanotechnology developments within research fields such as optoelectronic, high-density magnetic storage devices and biotechnology, in which the main attention has been focused on the use of NAFs as templates for self-assembly processes of a wide variety of species [135, 136, 137]. The ordering of a given NAF can be transferred to other media by suitable replica/antireplica processes [119]. Such a technique has been successfully employed to prepare metallic and polymeric membranes or antidot arrays [139]. In addition, hard nanostructures can be obtained which can be used as stamps for reproducing a given ordering in precursor Al disk, an example of which is given in Fig. 7.6. 7.2.2.2 Fabrication of Thin Films The main techniques to prepare thin films and heterostructures are described in this section. Physical Vapour Deposition (PVD) Physical vapour deposition (PVD) is a technique whereby physical processes, such as evaporation, sublimation or ionic impingement on a target, facilitate the transfer of atoms from the source to a substrate. A schematic representation of main PVD techniques is given in Fig. 7.7. Evaporating and sputtering are the

Fig. 7.6 HRSEM image from nanostructured TiN obtained by sputtering and using nanoporous anodic alumina films as substrates [138]

100 nm

7 Nanostructured Magnetic Sensors Fig 7.7 Scheme of the different physical vapour deposition, PVD, techniques

201

Laser Ablation

Sputtering

MBE

two most widely used PVD methods for depositing films. The principal difference between the chemical vapour deposition (CVD) and the PVD is that in the case of PVD the source target is in solid state and there is no chemical reaction of the evaporated gases, while in the CVD methods the source is in the gaseous state and chemical reactions occur on or very close to the substrate surface. The laser ablation method is a particular case of PVD where a laser beam is used to evaporate the source material. Although pulsed laser method is commonly used, continuously laser beam can also be employed. The amount of material removed by a single laser pulse, just at the surface and below, where the laser is active depends on the optical properties of the material and on the laser wavelength.

Sputtering This is one of the most used techniques along the last decades both in research laboratories and at industrial level for producing a lot of materials with a broad variety of compositions [138, 140, 141]. Sputtering techniques are basically based on the bombardment of a selected material, acting as a target, with ions coming from electric discharge produced in plasma form. The incident ions and the atoms on the surface of the target interact very strongly to produce vapours of the latest. In a vacuum chamber, those vapours can be deposited on the highly clean surface of determined substrates. Argon is the most used discharge gas. The target acts as a cathode and the substrate is located, in general, on a conductor anode. Among others, the following advantageous characteristics should be mentioned: (a) low temperature deposition, (b) growth of a wide variety of materials with very different nature and properties, (c) reproducibility of the target composition on the substrate surfaces with high efficiency and (d) high deposition rate and environment control which allow high quality in the obtained materials. A number of magnetic nanostructures are fabricated by sputtering, with applications profiting from their various magnetic and magneto-optical properties [142].

202

M. Va´zquez et al.

Evaporation In this process, thermal energy is supplied to a source from which atoms are evaporated for deposition onto a substrate. The vapour source configuration is intended to concentrate heat near the source material and to avoid heating the surroundings. The source material can be heated by several methods. The simplest way is the Joule heating by a resistance of wire or stripe of refractory metal to which the material to be evaporated is attached. The evaporated atoms travel through reduced background pressure in the evaporation chamber and condense on the substrate surface. Molecular Beam Epitaxy (MBE) This is a technique to fabricate thin films where molecular or atomic beams fall upon substrates at constant temperature in an ultrahigh vacuum (UHV) chamber. It can be considered as a particular case of evaporation, already mentioned. The main advantage of this technique is the possibility to use moderate deposition rates and temperatures allowing, among others, a high spatial control of the impurities and doping. On the other hand, the UHV atmosphere allows one to obtain epitaxial layers with very high purity. In addition, such experimental conditions enable the use of different in situ characterization techniques of the grown materials so that the growth processes can be checked and controlled. By means of this complex and sophisticated technology it is possible to prepare low-dimensional magnetic nanostructures (0D, 1D and 2D) with high homogeneity and control on their chemical composition. In this case, the main characteristic of such structures is their epitaxial nature. Thus, layer-by-layer growth homoepitaxial and heteroepitaxial are possible to be prepared. In the first case, the grown layers have the same lattice parameter as that of the substrate, while in the second one the layers grow on the substrate surface with different lattice parameter. Although MBE was initially used for semiconductor growth in the microelectronic industries, today it is further employed for fabrication of a wide range of functional materials particularly magnetic nanostructures [143, 144, 145]. The high control of the thickness of layers enables the study, for example, of magnetic coupling through non-magnetic layers, or the giant magnetoresistance effect [146]. In Fig. 7.8, an example of a particular multilayered system of magnetic/non-magnetic [(NixFe12x)yAu12y]/Au multilayers is shown [147]. Chemical Vapour Deposition, CVD From the seventies, chemical vapour deposition (CVD) techniques have played a crucial role in microelectronic industries, specially in the fabrication of highly homogeneous thin films. In these techniques, chemical reactions of precursor gases or liquids are caused to produce the formation of pre-designed materials on the surface of determined substrate. By this method it is possible to obtain single and multicomponent materials with high stoichiometric control. Homogeneous

7 Nanostructured Magnetic Sensors

203

Fig. 7.8 Antiferromagnetically coupled FeNi/Au multilayer growth by MBE [147]

or heterogeneous reactions can occur in CVD experiments, taking place in the gaseous phase or just in contact with the substrate surface, respectively. The first one can produce the formation of particles with diameters in the nanometric scale, while the second gives rise to the formation of thin films with highly controlled thickness allowing the formation of nanostructured 2D systems. One important shortcoming of traditional CVD techniques is the high temperature required for chemical reactions (typically above 4508C) which restricts the type of substrate that can be used for the growth of specific materials. In the last decades, plasma enhanced chemical vapour deposition (PECVD) has been developed which particularly enables to obtain different thin films with high stoichiometry at lower temperatures than those used in CVD processes where thermal activation is only applied [148, 149, 150].

7.2.3 Characterization of Magnetic Nanostructures 7.2.3.1 Structure Characterization Before magnetic properties of nanostructures are properly determined, their structure characterization is really necessary to be performed particularly since nanoscale dimensions play a determinant role in many of those magnetic

204

M. Va´zquez et al.

structures. That has to be performed in many cases through sophisticated techniques [151, 152, 153, 154]. In the following, a brush image of the different methods is illustrated. Most used methods to characterize nanostructured materials are those systematically used for bulk materials. From structural, morphological and composition points of view, due to the size scale diminution, some of those methods have acquired special relevance, such as high-resolution scanning electron microscopy (HRSEM), high-resolution transmission electron microscopy (HRTEM) and synchrotron radiation-based techniques. Both scanning and transmission microscopy are essentially based on the same physical principles. In the former, the images are mainly formed from secondary electrons generated on the material surfaces, while in the case of transmission the electrons transmitted through the samples are responsible for the formation of images. Another difference between these microscopies is the invasive character of the HRTEM in which it is necessary to use very high accelerating voltage values in addition to the sample slimming. Nevertheless, as a positive characteristic, we should mention that the material structure can be determined by means of the electron diffraction spectra produced from the interaction with the samples. In the last decades high-technology microscopes have been developed with very high resolutions to solve nanometric features in the samples and micro-probes which, based on the X-ray fluorescence phenomena, allow one to determine the elemental chemical composition of the samples. In general, the synchrotron-based techniques needed for a complete characterization (structure and composition) of nanostructured materials are X-ray fluorescence (XRF), X-ray absorption spectroscopy (XAS), extended X-ray absorption fine structure (EXAFS), X-ray absorption near-edge spectroscopy (XANES) and X-ray diffraction (XRD). A brief description of these techniques is given. X-Ray Diffraction From advanced semiconductor technology to new magnetic materials, crystallography using hard X-ray diffraction techniques at synchrotron radiation facilities plays a key role in our ability to understand and control the material’s properties. The research issues that are addressed by this technique are structural studies of crystalline materials, size determinations, texture identification, crystallographic orientations, phase transitions, residual stress fields or in situ environments. X-Ray Fluorescence It provides one of the simplest, accurate and straightforward analytical methods for the chemical composition determination. It can be used for a wide range of elements providing detection limits well below the ppm level. It is specially recommended when the sub-micrometre scales are too small.

7 Nanostructured Magnetic Sensors

205

X-Ray Absorption Spectroscopy It is applied in a wide variety of fields as quantum heterostructures, high-pressure studies, magnetic systems and doping issues in order to investigate geometric and electronic structures. This technique includes two kinds of analyses: (i) extended X-ray absorption fine structure (EXAFS) and (ii) X-ray absorption near-edge structure (XANES). EXAFS analysis is very useful to determine local environments of specific atoms to correlate them with the macroscopic physical properties, while XANES generates determinant information on the dopant charge state. 7.2.3.2 Techniques to Determine Magnetic Properties of Nanostructures In this section we consider some of the most relevant techniques to determine different magnetic properties of interest for magnetic applications and sensors. Hysteresis Properties: Vibrating Sample and SQUID Magnetometers and Kerr Effect The magnetic susceptibility and the reversal magnetization process determine the sensibility and the range of the sensor particularly in the case of the magnetic field sensor, SPM tips and MR-based sensors. The vibrating sample magnetometer, VSM [155, 156], is a conventional and reliable instrument to measure the magnetic moment of any ferromagnetic material. The sample is placed inside a magnet and vibrated perpendicularly to the field direction. The magnetometer is based on the induction law by which the magnetic moment is determined through the change of flux received by a system of coils when the sample vibrates. The signal from a pair of pickup coils is compared with that induced in a pair of reference coils by a permanent magnet. Figure 7.9a shows as an example the hysteresis loop measured with 40 FC

20

0,8

H

= –5.5 T

M (emu/g)

0

M/Ms

0,4 0,0

–20 –40 –1.0

–0.5

0.0

0.5

1.0

0.0

0.5

1.0

40

–0,4

20

–0,8

–20

H

FC

= +5.5 T

0

–100

a)

–75

–50

–25

0

Field (mT)

25

50

75

100

–40 – 1.0

b)

– 0.5

H (T)

Fig. 7.9 (a) VSM hysteresis loop of a FePt thin film partially irradiated with Cl ions. A double critical field is observed [159]. (b) Exchange bias effect observed in the SQUID measured hysteresis loop of Co80Ni20 nanoparticles (shown in Fig. 7.3b) at 5 K after field cooling under indicated magnetic field [160]

206

M. Va´zquez et al.

VSM for a FePt thin film [159]. Some VSM instruments use superconducting magnets creating fields that are able to saturate even hard magnetic materials. Cooling and heating devices can be implemented. The sensitivity is limited by the mechanical noise induced by the vibrator. The superconducting quantum interferometer device, SQUID [157, 158], is a very sensitive magnetometer used to measure extremely small magnetic fields. SQUIDs are based on superconducting coils containing two Josephson junctions, consisting of two weakly coupled superconductors separated by a very thin insulating barrier, where the current flowing across the junctions is sensitive to the magnetic flux. The two superconductors separated can experience tunnelling of Cooper pairs of electrons through the junction. The Cooper pairs on each side of the junction can be represented by a wavefunction similar to a free particle wavefunction. In DC Josephson effect, a current proportional to the phase difference of the wavefunctions can flow in the junction in the absence of a voltage, while in the AC alternative, a Josephson junction oscillates with a characteristic frequency proportional to the voltage across the junction. Since frequencies can be measured with great accuracy, a Josephson junction device has become the standard measure of voltage. The SQUID is sensitive to fields near 1010 G, near the theoretical limit of energy sensitivity given by the uncertainty principle, and has a low l/f noise. The SQUID is more sensitive than regular VSMs although in some cases external perturbations may cause spurious signals. Magnetic nanoparticles exhibit some peculiar properties that can be resolved by sensitive magnetometry as SQUIDs. As an example, Fig. 7.9b depicts the SQUID low-temperature asymmetric hysteresis loops observed for core/shell CoNi nanoparticles. In this case, the shift in the hysteresis loop is determined by the exchange bias effect introduced by the antiferromagnetic shell to the ferromagnetic core. The shell of these nanoparticles, also represented in Fig. 7.3b, represents a very significant fraction of the whole nanoparticle volume. When compared to bulk material, magnetic behaviour of nanomaterials, particularly in the case of magnetic nanoparticles, is characterized by two main features that finally derive from the larger fraction of atoms at the surface compared to those at the volume: an effective decrease of the magnetic moment and the enhancement of the magnetic anisotropy. The first one is ascribed to the existence of a magnetically dead layer at the surface, spin canting or spin-glasslike behaviour of the surface spins. Concerning the second one, the magnetic anisotropy is larger than that of crystalline and shape origin, and it has been shown to vary with the adsorption of different molecules, which means that surface anisotropy is the main source. Therefore, magnetic properties of nanoparticles are highly dependent on particle size, shape and in some cases surface coating [161]. A complex interplay between particle core and coating determines the magnetic properties of the resulting composite. In the case of inorganic coatings, for example, silica coating has been used to tune the magnetic properties of nanoparticles by controlling the distance between particles and therefore dipolar interactions. Gold-coated Co nanoparticles have a lower magnetic anisotropy than uncoated particles, but

7 Nanostructured Magnetic Sensors

207

Au-coated Fe enhances the anisotropy probably due to the formation of a Fe–Au alloy. Organic ligands can also modify the anisotropy and magnetic moment of metal atoms located at the surface of the particles. Cobalt nanoparticles show a reduction of magnetic moment and a large anisotropy; however, magnetite shows an enhancement of the saturation magnetization and reduction in anisotropy when coated with oleic acid [162]. Finally, a magnetic coating has a dramatic influence on the magnetic properties. The exchange coupling across the interface between a ferromagnetic core and an antiferromagnetic or ferromagnetic shell or vice versa causes the shift of the hysteresis loop along the field axis [163]. VSM and SQUID magnetometers are macroscopic techniques that supply information about the magnetization process and parameters as saturation magnetization, coercive field and susceptibility, as well as temperature or time dependences of magnetic moment of the nanomaterial as a whole. In turn, the hysteretic magnetic behaviour arising from a tiny depth at the surface of materials, particularly relevant in the case of bidimensional nanostructures, can be determined through magneto-optical methods. They are based on the small rotation of the polarization plane of the light when it is reflected (Kerr effect) or transmitted (Faraday effect) by the magnetic material [164, 165]. Since the light interacts with the surface of the sample within a penetration of up to few tens of nm, the behaviour of the top layer can be studied separately from the whole sample. In addition, the hysteresis loops of the different regions of a patterned media can be characterized independently [134]. Figure 7.10 shows the Kerr effect loop of a FePt nanopatterned thin film irradiated with Cl ions. An important characteristic of the magneto-optical Kerr effect is its time resolution. Magnetization dynamics on the sub-nanosecond scale, and particularly high-speed switching phenomena, can be determined by this technique [166, 167], even for individual single-domain nanomagnets [168].

Kerr Signal (a.u.)

0.8

Fig. 7.10 Kerr effect hysteresis loop of a FePt patterned thin film (same as in Fig. 7.9a)

0.4

0.0

–0.4

–0.8 –60

–40

–20

0

Field (mT)

20

40

60

M. Va´zquez et al.

208

Magnetic Imaging A number of techniques are used to gain direct information of the orientation of magnetic moments at the surface of nanostructured samples at micron, sub-micron and nanoscale. Some of them represent the most advanced scanning magnetic sensor devices that will be considered later. Magnetic imaging techniques used to characterize the domain structure of magnetic nanostructures can be sensitive to the stray field, the magnetic polarization or the total flux density. Magneto-optical Effects Since the magnetic material used in the magnetic sensor is not usually transparent, magneto-optical characterization based on Kerr effect [169] is used to observe the domain structure. The magnetic moment distribution at the surface is imaged through the magneto-optic interactions which depend directly on the magnetization. The relative orientation between the magnetization, M, and the polarization, E, vectors determines which component of the magnetization will be visible in a particular magneto-optic image. Figure 7.11 shows schematically the different configurations to observe perpendicular and in-plane magnetization components with regard to the sample surface plane. The polar Kerr effect is used to image perpendicular magnetization. The most common arrangement is the longitudinal effect in which the magnetization lies in the scattering plane of light. A drawback of the Kerr effect technique is its limited resolution reduced to few hundreds of nm. An example of domain structure obtained by Kerr effect is given in Fig. 7.12. Electron Microscopies Higher resolution can be obtained with scanning electron microscopy with polarization analysis (SEMPA). Secondary electrons emitted by a ferromagnetic material are spin-polarized [171]. In particular, the magnetic moment is parallel to the magnetization direction of the origin region. The polarization of secondary electrons can be used for imaging magnetic domains [172]. SEMPA (a)

(b) Plane of indicence

M

(c) Plane of indicence

Surface normal

M

polar

longitudinal

Plane of indicence

M

transversal

Fig. 7.11 Scheme of the (a) polar, (b) longitudinal and (c) transversal Kerr effect modes

7 Nanostructured Magnetic Sensors

209

Fig. 7.12 Kerr effect image the free layer (Ni81Fe19/ Co90Fe10) of a spin valve showing the effective anisotropy distribution after annealing under magnetic field perpendicular to the magnetic anisotropy axis. Two-axis anisotropy is observed [170]

directly provides an image of the surface magnetization while the finely focused (unpolarized) beam scans over the sample, as shown schematically in Fig. 7.13. SEMPA therefore produces a direct image of the magnitude and the direction of the magnetization in the region probed by the incident electron beam [174]. SEMPA depends on the fact that the polarization of the secondary electrons reflects the net spin density of the material. For the purposes of SEMPA, it is sufficient to treat each component of the vector polarization separately. The polarization along the x-direction is Px ¼ (N"  N#)/(N" þ N#), where N" (N#) are the number of electrons with spins parallel (antiparallel) to the x-direction. As an example, Fig. 7.14 shows SEMPA images for a nanoscaled trilayer under indicated conditions. SEMPA has several unique capabilities that set it apart from other magnetic imaging techniques: it measures the magnitude and direction of the magnetization with high spatial resolution (about 10 nm); topographic maps are measured simultaneously; it is a relatively surface-sensitive technique (1 nm). The drawbacks of the technique are the long exposure time, the UHV conditions’ requirement and that it is limited to metallic materials. Alternative electron microscopes are used to characterize the magnetic properties of nanostructures, all of them require UHV working conditions and thus they are less extended for sensor characterization. In the Lorentz methods,

Fig. 7.13 Scheme of SEMPA microscopy [173]

210

M. Va´zquez et al.

Fig. 7.14 SEMPA images displaying the magnetization observed in different layers of a Fe/ Cr/Fe trilayer. (a) My magnetization component of a Fe whisker substrate; (b) the same region after deposition of 1 nm thick Cr layer; (c) the same region subsequent addition of 2 nm thick Fe layer. (df) Mx magnetization component for the same conditions as in ac [175]

based on the transmission electron microscope (TEM), electrons in the beam are viewed as particles deflected by the Lorentz force produced by the magnetic field resulting from nearby magnetic material. There are three Lorentz microscopy modes, which are frequently referred to as Fresnel, Foucault and differential phase contrast (DPC) microscopy. In all modes, the deflection of the electron beam is sensed as it travels through a magnetic field. On the other hand, spin-polarized low-energy electron microscope (SPLEEM) is a surface-sensitive technique using very slow electrons (0–20 eV). Its main advantage is its capability to simultaneously record images showing topographic (low-energy electron microscopy, LEEM) and magnetic contrast. SPLEEM images are obtained by subtraction and subsequent normalization of two images taken with antiparallel polarization of the incident electron beam. Magnetic contrast in such images is a consequence of the exchange-scattering asymmetry. Magnetic transmission X-ray microscope (MTXM) [176] is based on the X-ray magnetic circular dichroism effect which occurs in the vicinity of elementspecific inner-core absorption edges. Depending on the relative orientation between magnetization and helicity of the circularly polarized photon, the absorption coefficient varies. Scanning Techniques A new imaging technique has become widely extended, the magnetic force microscope (MFM) [177]. It is based on the magnetic interaction between the stray field of the sample and the magnetization of the MFM tip. The high spatial resolution of this technique [178] is based on the low-dimension tip radius and the short tip–sample working distance. Topographic contribution is avoided from the magnetic signal by a double scan technique: the first scan, usually obtained in dynamic mode [179], relates to the topography while the magnetic signal is recorded during the second scan. In such second scan, the distance between the

7 Nanostructured Magnetic Sensors Fig. 7.15 Scheme of the MFM system. Two images can be obtained simultaneously: the topography and the domain distribution

211

Photodetector

Laser 1st scan Topography Oscillator

2nd scan Magnetic Signal

tip and the sample is of few tens of nm. When the tip–sample interaction is negligible, the MFM signal corresponds to the distribution of the magnetic poles which concentrates in the centre of the domains or at the domain walls in materials with perpendicular or in-plane anisotropy, respectively. A schematic view of the MFM system is presented in Fig. 7.15, while an example of MFM images is given in Fig. 7.16 for an array of triangular Ni nanostructures. The advantages of this technique are its high resolution, the possibility of obtaining simultaneous topographic and magnetic images, the easy sample preparation, not having restrictions for measuring at ambient conditions and the possibility to apply magnetic fields during the microscope operation, i.e. to study reversal magnetization process [180]. The drawbacks include the reciprocal tip–sample influence and the lack of quantitative information. In some works, the use of novel elements to be employed as MFM probes is proposed, in particular, ferromagnetic nanowires tips could improve the spatial resolution and avoid the always annoying tip-induced changes (see subsequent section) [181]. Spin-polarized scanning tunnelling microscope (SPSTM) can also be used to characterize magnetic nanostructures useful for sensor applications as in the

Fig. 7.16 Topography and MFM signals of an array of triangular Ni nanostructures fabricated by nanolithography [182]

600nm

212

M. Va´zquez et al.

Fig. 7.17 Map of the dI/dU signal of a single 8 nm high Fe island recorded with a Cr-coated W tip. The vortex domain pattern can be recognized. Arrows illustrate the interpreted orientation of the domains [183]

example given in Fig. 7.17. The current between the tip and the sample in a scanning tunnelling microscope can be used to obtain the magnetic state of the surface. High-resolution images can be obtained with this technique that should be used in UHV and low temperature [183]. Magnetic Resonance Imaging When the nuclear magnetic moment associated with a nuclear spin is placed in an external magnetic field, the different spin states are in different magnetic potential energies. In the presence of the static magnetic field that produces a small spin polarization, a radio-frequency signal of the proper frequency can induce a transition between spin states. This process is called nuclear magnetic resonance (NMR), and its resonant frequencies for each particular substance are directly proportional to the strength of the applied magnetic field, in agreement with the precession Larmor equation. This is used to great advantage in the medical imaging process with micrometre resolution known as magnetic resonance imaging (MRI). MRI has been regarded as a powerful imaging tool with non-invasive nature, high spatial resolution and tomographic capabilities, low signal sensitivity being its major limitation. The image contrast is due to the response of water protons to an external magnetic field. Energy is applied to the protons (in the radio-frequency range), exciting the water protons and when the radio-frequency source is

7 Nanostructured Magnetic Sensors

213

removed, water protons relax or return to a state of equilibrium. During relaxation to lower energy states, energy associated with the spin flip is emitted at the resonant frequency resulting in the signal observed in MRI. The difference in relaxation rates and proton densities within the various tissues is responsible for the native soft-tissue contrast. Despite the inherent contrast of MRI, there are situations where contrast agents are required to enhance the relaxation of water protons in specific tissues. Contrast agents are either paramagnetic complexes such as gadolinium chelates (chemicals that control the concentration or effects of metal ions) or superparamagnetic particles of iron oxide with various surface modifications. Today’s NMR microscopy has reached voxel resolution of 3.5 mm3 at 400 MHz, with 30 h acquisition time [184]. Alternative electron spin resonance is used in electron spin resonance (ESR) microscopy [185]. Some theoretical derivations have been shown that ESR imaging methods should achieve voxel resolution better than 1110 mm in several minutes of acquisition (at 35–60 GHz) for bio/liquid samples [186].

Magnetoresistance Magneto-transport effects (Hall effect and magnetoresistance) appear as a consequence of the influence of magnetic field onto electrical current carriers so modifying the resistance of the material. They have been observed and studied traditionally in semiconductors in connection with the development of magnetic field sensors. The study of the magneto-transport properties of ferromagnetic nanostructures is particularly important for their applications, and they supply in addition information about the magnetic microstructure and reversal magnetization process. The present relevance of various magnetotransport effects is related to their wide use in advanced magnetic recording heads. Besides conventional magnetoresistance, MR, effect observed in semiconducting layers, two effects can be distinguished: the anisotropic magnetoresistance (AMR) and the giant magnetoresistance (GMR). The anisotropic magnetoresistance [187] is based on the dependence of the resistance on the angle between the current density and the magnetization. AMR is a relatively weak effect, where about 4% resistance variation can be reached. However, it is enough to use it in reading heads read. The GMR effect appears when two ferromagnetic material layers are separated by a non-magnetic or spacer metallic layer. The discovery, understanding and technological applications of this phenomenon have resulted in the awarding of the Nobel Prize to Fert and Gru¨nberg [188, 189]. If the metallic interlayer is thinner than the spin diffusion length and the ferromagnetic layers are magnetized in different directions, the electrons coming from the source suffer an additional scattering process in the second as schematically shown in Fig. 7.18. This is also observed in multilayer systems consisting of a large number of bilayer systems as that described here.

M. Va´zquez et al.

214

No magnetic field

Magnetic field applied

Ferromagnetic layer

Ferromagnetic layer

Low current

Spin-polarized current

Non-magnetic layer

Non-magnetic layer

Ferromagnetic layer

Ferromagnetic layer

Fig. 7.18 Scheme of the GMR effect. The resistance varies as a function of the relative magnetization of both ferromagnetic materials

In spin valves, both ferromagnetic layers have different hysteresis behaviour. One layer is magnetically soft, and consequently sensitive to small applied fields, while the other layer is magnetically harder. As the magnetization of the soft ‘‘free’’ layer changes due to an applied magnetic field, the resistance of the whole structure will vary. The hard magnetic layer is typically pinned or exchange biased by an antiferromagnetic material as shown in Fig. 7.19. An electrical current can flow through local barriers as nanoconstrictions or insulating layers by tunnelling effect (see Fig. 7.20a). A variant of the GMR elements is the tunnelling junction where the two ferromagnets are separated by a thin insulating film. In this case, the electrons travel from one ferromagnet to the other by tunnelling effect through the insulator layer [190]. Higher tunnelling magnetoresistance, TMR, effects can be obtained with this configuration (see Fig. 7.20b). Finally, the giant magneto-impedance (GMI) is being increasingly used in microsensor applications [191]. This effect occurs when a high-frequency

Free layer Spacer layer Pinned layer AF pinning layer

Fig. 7.19 Schematic view of the different nanoscaled layer in a spin valve

h w

7 Nanostructured Magnetic Sensors

215

Fig. 7.20 A magnetic nanoconstriction (a) and magnetoresistance effect in a tunnel junction (b)

current flows along a soft magnetic metallic conductor, typically in the form of microwire or thin film. The current actually flows along a thin skin-depth penetration below the surface of the material. Under the action of a stationary magnetic field, the skin-penetration depth increases which induces a decrease in the impedance, both real and imaginary components. The GMI effect is particularly effective in ultrasoft magnetic materials where the action of that stationary field strongly modifies their ac permeability. In the case of magnetic microwires exhibiting bamboo domain pattern, when an alternating current flows through the wire, the induced circular magnetic field displaces the domain walls reversible. As a result strong impedance is measured in the wire that is suppressed by applying a longitudinal field. The working frequencies range from 1 kHz to 1 MHz. Nowadays some sensors based on nanostructures use this GMI effect [192].

7.3 Magnetic Sensors and Applications This section is divided into three main sections attending to the dimensionality of the magnetic nanostructures on which the magnetic sensors and technological applications are in general based. Owing to the above-mentioned large variety of those applications, we focus on each section towards most relevant applications. The first one is devoted to technological applications where magnetic nanoparticles are involved, which are mainly related to magnetobiological and biosensor applications. The second is devoted to magnetic nanowires and their use as magnetic probes in different scanning techniques. Finally, the third related to thin films is mainly devoted to magnetic sensing in connection with magnetic storage of information.

216

M. Va´zquez et al.

7.3.1 Biological Applications Based on Magnetic Nanoparticles In molecular biology, a strong interest is being devoted to the use of magnetic nanoparticles as an efficient tool for a fast, efficient and easy biomolecule extraction [193]. Today, these processes can advantageously replace other techniques such as centrifugation, filtration and chromatography. The large surface area per volume of the nanoparticles provides rapid reaction kinetics, while the magnetic separation allows easy and rapid handling, and is extremely gentle to the target, preserving protein nativity and viability of fragile cells. On the other hand, the use of magnetic nanoparticles as contrast agents for nuclear magnetic resonance imaging will be addressed. One of the current challenges in the biomedical sciences is the ultrasensitive imaging of biological targets under non-invasive in vivo conditions, and NMR imaging is a very powerful technique for this purpose [194].

7.3.1.1 Biosensors for Detection and Separation The major application of magnetic nanoparticles nowadays concerns the extraction of biomolecules such as proteins, antibodies and nucleic acids [195, 196], although in addition, they are used for specific bacteria, virus captures and cell recognition [197]. The most important factors are the magnetic particles and the used ligand. Magnetic beads should have high magnetic oxide content for their fast magnetic separation and a good compatibility with other biomolecules such as those enzymes used for nucleic acid amplification. Geometry characteristics, as size and shape, are important parameters to be considered. Some commercial products, such a Dynabeads1, are based on monosized polystyrene beads which have an even dispersion of magnetic nanoparticles inside (Fe2O3 or Fe3O4) coated with a thin polymer shell. A schematic description of the cell separation process is given in Fig. 7.21. A wide range of surface activated and pre-coated Dynabeads1 products are available [198]. For example, by using Dynabeads1 with covalently coupled streptavidin together with a biotinylated probe/ligand, any target molecule can be captured, isolated and further manipulated. This product has also been used for detection of various metastatic cancers in tissues and blood [199]. These commercial products are used on a wide variety of automated devises. Polyamine particles of 1 mm diameter have also been used in the development of an ultrasensitive method for detecting protein analytes [195]. In general, a biosensor is a compact analytical device incorporating a biological-sensitive element associated with a physicochemical transducer. Such transducer systems can be based on electrochemical, piezoelectric or magnetic principles. A schematic representation of the enzyme-linked immunomagnetic electrochemical assay (ELIME) is presented in Fig. 7.22. An immunogenic analyte (bacteria, for example) is sandwiched between an antibody-coated magnetic bead and an antibody-enzyme conjugate. The bead is

7 Nanostructured Magnetic Sensors

217

Fig. 7.21 The principle of dynabeads work for cell separation. Dynabeads are added to a starting sample in a tube. The beads bind to the target cells after 15–30 min incubation. A magnet is placed near the tube containing sample and beads. The beads and bound cells migrate to the magnet and the supernatant is removed with a pipette. Dynabeads can be detached from isolated cells if required. Cells isolated with dynabeads can be used in many assays/applications including cell culture, functional and proliferation studies, flow cytometry, molecular studies, cytokine secretion, phenotyping and further sorting

O

Antibody-Enzyme Conjugate

AP

R-O-P-O + H2O O R-OH + HPO 3+ 4

Bacterium

Antibody R-OH

Magnetic Bead Fig. 7.22 Schematic representation of the enzyme-linked immunomagnetic electrochemical assay (ELIME)

R=O H+ e–

N N MAGNET

S S

218

M. Va´zquez et al.

trapped magnetically on the electrode surface, exposed to the enzymatic substrate and the electroactive product is detected electrochemically. This type of assay has been applied for different analytes with very good reproducibility, i.e. 2% standard deviation [200, 201, 202, 203]. For the detection of antigen concentration in biological samples, the electrical properties of an antibody immobilized on a gold electrode can be followed by electrochemical impedance measurements. Beads are modified with streptavidin and the antibody with biotin which have a high affinity for streptavidin. In this way, a small pesticide molecule has been detected with a detection limit of 10 ng/ml [204]. A piezoelectric immunosensor based on a 5 MHz quartz crystal resonator has been used for the detection of biological pathogens such as Salmonella typhinium [205]. Salmonella cells are captured by antibody-coated magnetic beads and then moved magnetically to the sensing quartz. An analyser measured the impedance behaviour of the oscillating quartz crystal exposed to various concentrations of Salmonella (102–108 cells per ml). Enzymes immobilized on magnetic beads can be trapped by magnets and retained on an electrode surface at a specific location in flow analysis devices and further analysed by amperometric, potentiometric or conductimetric measurements. An example of enzyme immobilization strategy is given in Fig. 7.23. For environmental toxicity analysis, detection limits as low as 1011 M for heavy metal ions and 1012 M for organophosphates and carbamates have been achieved [206]. The practice of DNA sequence detection has become more and more ubiquitous in genetics, pathology, criminology, food safety and many other fields. Magnetic nanoparticles, as biomolecule carriers via a suitable immobilization process, offer good potential for sensitive sensors. Nanoparticles, prepared by co-precipitation and then coated with a carboxylic acid containing polymer layer, have been used in electrochemical nucleic acid sensor systems [207]. Nanoparticles prepared by decomposition in organic media have been also modified using a combination of alkylphosphonate surfactants and ethoxylated fatty alcohols. These particles can be selectively hybridized to DNA functionalized gold surfaces and used in biomolecule detection [208].Moreover, if magnetic nanoparticles are provided with a gold coating, the combined benefits of the robust chemistry for gold surfaces and the uniqueness of magnetic nanoparticles could be realized [209].

Fig. 7.23 Preparation of magnetic beads for a tyrosinase-based biosensor

7 Nanostructured Magnetic Sensors

219

On the other hand, Reich et al. demonstrated the utility of high aspect ratio magnetic nanowires for cell manipulation [210]. Nanowires were synthesized by electrodeposition of Ni into nanoporous alumina membranes and isolated as a suspension. Finally, novel magnetic and fluorescent nanocomposites prepared by layer-by-layer self-assembly approach have shown to be sensitive probe for the determination of proteins [211]. A genomagnetic electrochemical bioassay (GEME) for the separation of DNA based on the immobilization of DNA on magnetic beads and electrochemical detection has been proposed and successfully used with detection limits of 100 ppb for breast cancer gene [212]. For high-sensitivity detection of protein and DNA, bio-bar code assay appeared in the early 2000s as a promising analytical tool. It relies on a sandwich structure based on specific biological interaction between magnetic beads and nanoparticles (gold and polystyrene) with an immobilized oligonucleotide, which is call bio-bar code complement. Magnetic beads are added to the solution and allow interaction with the biological target to be detected, generally via DNA/DNA interaction or antibody/antigen interaction. Then, nanoparticles are added and interact with the biological target to form a sandwich-like structure. Beads are then separated and the sandwich redispersed in water. The bio-bar code DNA is then dehybridized and captured on a DNA chip and detected. Three main detection methods are used: scanometric detection, fluorescent detection and rolling circle amplification (RCA). Scanometric detection leads to detection as low as 5001021 M (molar), in the case of DNA and 31018 M for protein (PSA). The detection limit of the fluorescent method is of a few hundreds of 1018 M, while for the RCA assay it is of 1 pg/ml for DNA. In the case of the magnetic biosensors, a magnetic field sensor can be used in combination with magnetic nanoparticles, which act as magnetic labels to detect low concentration of the target of interest [213] (see Fig. 7.24). Recently developed giant magnetoresistance (GMR) or tunnelling magnetoresistance (TMR) magnetic sensors have been developed (see a later section for details). Then, the magnetic nanoparticles are specifically attached to the target molecules, and their magnetic stray field is picked up by an embedded magnetoresistive sensor as a change of the electrical resistance. In comparison to previously mentioned methods, i.e. fluorescent, these magnetic biosensors have a number of advantages, including low-molecular detection limits, flexibility and the direct availability of an electronic signal suitable for further automated analysis. This makes them a promising choice for the detection units of future widespread and easy-to-use lab-on-a-chip systems or biochips [214]. Another novel method of detecting either protein binding or DNA hybridization at room temperature has been developed based on magnetic nanoparticles of manganese ferrite and a magnetic tunnel junction-based biosensor situated in orthogonal magnetic fields [215]. Spin valve sensors have been used for single bead detection of 2 mm diameter [216]. These sensors, described in a previous section, consist of an antiferromagnet layer next to a pair of spaced ferromagnetic layers, where the antiferromagnet serves as exchange biaser for

220

M. Va´zquez et al.

Fig. 7.24 Some examples of magnetic biosensors developed in different laboratories where magnetic field sensors are used in combination with magnetic nanoparticles

the adjacent magnetic layer, pinning its magnetization direction. The other magnetic layer is free to rotate. This leads to a linear magnetic field dependence of the resistance. Spin valves are sensitive not only to the magnitude but also to the direction of the field in the plane. GMR and spin valves are the most common sensor types. Biosensors based on tunnelling magnetoresistence and giant magneto-impedance are still very new. Biomagnetic sensor can be also based on Brownian relaxation of magnetic nanoparticles suspended in liquids. The characteristic time scale of the Brownian relaxation can be determined directly by ac susceptibility measurements as a function of frequency. This is a consequence of the shift of peak of the imaginary component of the ac susceptibility to lower frequencies upon binding the target molecules to the magnetic nanoparticles. The frequency shift is consistent with an increase of the hydrodynamic radius corresponding to the size of the target molecule [217]. Measurements based on the variation of magnetic permeability of a compound using inductance measurements have been applied for detection and quantification of DNA with detection limits of around 50 mg/ml for plasmid DNA in buffered solutions [218]. For human albumin detection in undiluted urine, a detection limit of 5 mg/ml was obtained [219]. Further work should be focused on optimizing biological molecule immobilization in order to reduce crossreactivity and non-specific adsorption.

7 Nanostructured Magnetic Sensors

221

7.3.1.2 Magnetic Nanoparticles as Contrast Agents in MRI Imaging One of the current challenges in the biomedical sciences is the ultrasensitive imaging of biological targets under non-invasive in vivo conditions, and NMR imaging is a very powerful technique for this purpose [220]. The success of the in vivo imaging techniques is highly dependent on the molecular imaging agents, which could lead us to precisely detect early-stage diseases, monitor the responses to drug therapies and track cell migrations. Low-molecular-weight agents are extravascular and distribute with blood flow but provide no cell-specific or process-specific information. They cannot be used for cell tracking. Polysaccharide-coated iron oxide nanoparticles have been largely investigated because of their good inherent signal differentiation and ability to be internalized by cell with phagocytic capacity. Iron oxide nanoparticles have been used as imaging agents in cell migration, gene expression, angiogenesis, apoptosis, cancer detection [221] and also as marker of inflammatory status [222, 223]. Effect of particle size and coating on biodistribution has been explored leading to the conclusion that smaller and neutral particles circulate longer and have a limited uptake by the reticuloendothelial system (RES). The signal enhancement of magnetic resonance imaging, MRI, however is still unsatisfactory compared to that obtained with other imaging modalities. One problem is the lack of a clear relationship between nanoscale material characteristics and MR signal enhancement effects. Magnetic iron oxide nanoparticles used as contrast agents are generally synthesized in water. However, high-quality nanoparticles prepared by decomposition in organic media and possessing a variety of metallic dopants with distinct magnetic spin magnitude have been recently checked for NMR imaging and proposed for ultrasensitive detection of target biological molecules. It was observed that composition controls the magnetic spin magnitude and that this is critical for modulating the spin–spin relaxation processes of protons in the water molecules surrounding the nanoparticles. In addition, there was a marked size dependency of MR signals with a gradual increase in contrast as the size increases. A faster spin–spin relaxation process of water molecules is induced by materials with a larger magnetization. Combining both effects, composition and particle size, manganese ferrite particles, 12 nm in diameter, have been shown to enhance sensitivity for cancer cell detection and to enable in vivo imaging of tumours as small as 50 mg [224]. Bi-functional contrast agents with both optical and magnetic contrast have been demonstrated to serve as good molecular imaging probes for in vitro and in vivo experiences [225]. Dual modality detections can be simultaneously achieved using a single material, i.e. by MRI and fluorescence microscopy. The material consists of CdSe/ZnS core/shell QDs, coated with a pegylated phospholipids and a gadolinium compound. Dye molecules have also been incorporated into a silica shell coating iron oxide nanoparticles [226] and the material has been probed to serve as a superior multifunctional tracking agent.

222

M. Va´zquez et al.

MRI also offers the potential of in vivo tracking of cells using innovative approaches for cell labelling and image acquisition. Cell labelled with magnetic nanoparticles can be imaged from weeks to months after intravenous administration or direct injection in order to track migration into the target tissue, for example in cardiovascular diseases [227]. Cell labelling with magnetic nanoparticles is now widely used as a means of in vitro cell separation, to isolate cells of interest, or in vivo for delivering drugs and locally concentrating them at the desired site. If cells are labelled with nanoparticles, they can be located, tracked and recovered and those processes can be followed by imaging techniques such as NMR. Magnetic cell targeting opens up new possibilities for vascular tissue engineering and for delivery localized cell-based therapies [228]. Various organic coatings have been used. Amphiphilic coatings such as polyethylene glycol and dendrimers and also transfection agents such as HIV-derived TAT proteins allow their free passage into cells. Nanoparticles with anionic charge result in non-specific adsorption at the surface and endocytosis into the cell of a large number of particles, i.e. up to more than 3–4 orders of magnitude (see Fig. 7.25). However, there is a certain concern about the effect of intracellular iron oxide nanoparticles on the normal cell behaviour. Some works have shown that nanoparticles and surface coating can have a dramatic effect on cellular morphology, uptake efficiency, cytotoxicity and cell mobility [229]. Moreover, moderate levels of anionic iron oxide nanoparticles modified with dimercaptosuccinic acid (DMSA) adversely affect cell function in growing neurons [230] although DMSA by itself has been demonstrated that in the same concentrations or larger have no measurable effect. More studies on the acute and long-term effects of cellular Fe2O3 internalization are then necessary and warranted. Non-invasive imaging of myocardial macrophage infiltration has been shown to be possible by both fluorescence tomography and magnetic resonance imaging using magnetofluorescent nanoparticles, which are taken up by macrophages in infrarcted myocardium. An increase in magnetic resonance imaging contrast-to-noise ratio, indicative of myocardial probe accumulation, has been observed in the anterolateral walls of the infarcted mice together with significantly greater fluorescence intensity over the heart. The

Fig. 7.25 HeLa cells labelled with anionic magnetic nanoparticles (courtesy of Dr. A. Villanueva and M. Can˜ete, Faculty of Biology,UAM, Madrid)

7 Nanostructured Magnetic Sensors

223

uptake of magnetofluorescent nanoparticles by macrophages infiltrating the infarcted myocardium was confirmed by fluorescence microscopy and immunohistochemistry [231]. Other Applications A magnetic hydrogen sensor has been proposed based on the variation of saturation magnetization and remanence of nanoscale antiferromagnetic hematite with hydrogen flow [232]. The study of transparent magnets, among other magnetic composites, is an interesting challenge because of their novel potential applications in magnetooptical switches, modulators, optical circulators, laser isolators, magnetic field and electric current sensors based on Faraday effect. It is important to reduce the size of the magnetic particles in the composite in order to obtain superparamagnetic behaviour allowing the sample to be used as a low magnetic field sensor. Guerrero et al. [233] have observed Faraday rotation in -Fe2O3/SiO2 samples in which the iron oxide particles grew during the formation of the sol–gel silica matrix. Further improvement of this composite was achieved by impregnation of slices of porous Vycor glass (VG) rods with an iron nitrate solution followed by a thermal treatment and a reduction process [234]. The resulting -Fe2O3/Vycor glass composites exhibit remarkable Faraday rotation and can find application in a large range of magneto-optical devices such as a magnetic field sensor. Finally, very recently a magnetic hydrogen sensor has been proposed based on the variation of saturation magnetization and remanence of nanoscale antiferromagnetic hematite with hydrogen flow [235].

7.3.2 Magnetic Nanowires and Sensors for Magnetic Scanning Techniques In this section we review the use of cylindrical nanostructures, such as nanowires and nanotubes, and thin films in magnetic sensors technologies. We particularly focus on those technologies related to magnetic scanning techniques where various kinds of tips with elongated shape and nanowires are employed. Nevertheless, in few examples sensing elements include 2D nanoelements. 7.3.2.1 Sensors Based on Nanowires Grown into Ordered Membranes Earlier in this chapter, the preparation of magnetic nanowires has been described, covering processes such as nanolithography, electroplating or sputtering. Of particular interest are nanowires grown into ordered hexagonal membranes, either polymeric or alumina made. Such nanowires can be removed

224

M. Va´zquez et al.

from the template by suitable methods in case sensing requires individual nanowires as for example in scanning techniques which will be considered in more detail in the next section. One of the most common applications of nanowires is in technologies related to magneto-transport properties. In this regard, magnetoresistance, MR, and magneto-impedance, MI, sensors are proposed to detect magnetic fields, where a large change in the impedance is obtained when the susceptibility of the sample changes. An interesting example of this type of sensors is given by Enculescu et al. [236]. who makes use of a nanowire consisting of alternating layers of Co and non-magnetic Cu obtained by electrochemical filling and etched-track nanochannels. The single ion track is prepared by irradiating a 30 mm thick polycarbonate foil with heavy ions and then the damaged trail is transformed into a narrow channel by selective etching, interrupted once the desired size is reached. By changing the potential between anode and cathode during the electroplating process it is possible to force selective growth of Co or Cu, as required. Therefore a pulse voltage leads to the growth of multilayers nanowires. The contacts of the sensor are established by sputtering the top of the membrane with a gold film and mounting it on a circuit board. An application of nanowires for magnetic biosensing has been proposed by Anguelouch et al. [237] that makes use of Ni nanofilament with 170 nm diameter and length ranging from 5 to 30 mm. The nanowires are obtained using template electrodeposition and the GMR sensor consists of two ferromagnetic electrodes separated by a non-magnetic metal. The magnetization direction of the top layer, being fixed, defines the axis of maximum sensitivity, i.e. the sensing direction. The magnetoresistance of the sensor depends therefore on the divergence of the magnetization direction of the bottom layer relative to the top layer. The nanowires are suspended in an aqueous solution and introduced over the sensor chip and the bridge output is recorded, as discrete increase of the output, as the wire settled onto the sensor, as shown in Fig. 7.26. This sensor has the ability to detect a single nanowire touching the GMR sensor, widening its application to liquid flow tracking and biosensing applications where single particle sensitivity is needed. A more exotic application is represented by the use of magnetic nanowires for acoustic sensor [238]. Also in this case, nanowires are grown into nanoporous template such as nuclear track etched, block polymer or anodic aluminium oxide membranes. A combination of anodic aluminium oxide and nanoimprinting is proposed to obtain a nanoporous membrane with large ordered areas. With this technique, the aluminium precursor foil was electrochemically polished prior to imprinting. A Si3N4 stamp was used to transfer the ordered pattern to the Al foil, the average depth of the prepatterned imprint having value of 30 nm, followed by an anodization process. The resultant nanoporous membranes present areas of ordered hexagonally distributed pores, where the imprinting was performed, and disordered porous areas, as shown in Fig. 7.27. To grow the nanowires, a metal electrode was

7 Nanostructured Magnetic Sensors

225

Fig. 7.26 Discrete response of a GMR sensor using nanowires in the case of (a) two 5 mm nanowires and (b) one 30 mm approaching the surface of the sensor [237]

deposited on the back of the membranes and Galfenol wires have to be grown by electrochemical route. The magnetic nanowires are exposed by partially removing the alumina membrane via selective acid attack. Acoustic measurements are obtained by bounding the matrix of nanowires to the surface of an integrated circuit die containing a commercial giant magnetoresistive, GMR, element. According to the authors this sensor requires further optimization in terms of reducing the distance between the nanowire arrays and the GMR sensor, so increasing the sensitivity.

Fig. 7.27 SEM micrograph of surface view of anodized pretextured aluminium [238]

226

M. Va´zquez et al.

7.3.2.2 Magnetic Sensors Based on Scanning Techniques Scanning techniques are very useful to determine the nanoscale characteristics. Here, we describe some advanced techniques that make use of magnetic nanostructures and associated phenomena which enable the measuring of magnetic fields.

Magnetic Force and Magnetic Resonance Force Microscopies As described in a previous section, magnetic force microscopy, MFM, is a very useful tool to determine the surface magnetic characteristics of a nanostructured material. This microscopy requires of a magnetic tip to sense the stray field created by that surface. Recently, a number of studies have been performed trying to incorporate magnetic nanowires and nanotubes to substitute conventional magnetic tips. For example, Yang et al. [239]. have proposed the use of Ni nanowires as MFM tips. In this case, the nanowires are also obtained by filling the pores of anodized alumina membranes with Ni by electrodeposition. The diameter of the nanowires ranges typically between 20 and 40 nm, their lengths around 500 nm and presents saw tooth morphology. The nanowires are separated from the alumina membrane by attacking the latest with phosphoric acid at room temperature and subsequently dispersed in deionized water. Ni nanowires are then attached to a commercial AFM probe by a dielectrophoresis process, as shown in Fig. 7.28. The assembled tip is straight and aligned along the cone axis of the Si cantilever, having length of around 2 mm and a diameter at the very end of approximately 40 nm. Carbon nanotubes have been shown to work successfully when employed to image magnetic domains [240]. A conventional MFM magnetic probe was prepared using iron-filled multiwall carbon nanotube. The nanotubes are prepared by pyrolysis of ferrocene in a quartz furnace. Their diameters range between 30 and 100 nm while the Fe filling has a mean diameter between 10 and 20 nm and lengths up to several microns. A single nanotube is attached to the cantilever by pinning it with the electron beam of a scanning electron microscope that induces the deposition of carbon-containing contaminations.

Fig. 7.28 SEM images of a Ni bundle attached to a commercial AFM tip [239]

7 Nanostructured Magnetic Sensors

227

The magnetic resonance force microscopy (MRFM) [241, 242] actually represents a hybrid of magnetic resonance imaging (MRI) and atomic force microscopy (AFM), which is widely used in nanotechnology and nanoscience. Gene sequencing, drug development, analysis of storage media and many other commercial, medical and industrial applications can benefit from this new imaging technology. The resolution of this technique is so high that the "spin" of a single electron can be detected [243] and high spatial resolution has been obtained [244]. A schematic view of a MRFM is given in Fig. 7.29 [245]. Composite nanowires have been proposed by Barbic [246] for the fabrication of probes for atomic resolution magnetic resonance force microscopy (MRFM). The fabrication technique combines electrochemical deposition of multifunctional metals into nanoporous polycarbonate membranes and chemically selective electroless deposition of optical nanoreflector onto the nanowire. The completed composite nanowire structure contains all the required elements for an ultrahigh sensitivity and resolution MRFM sensor with a magnetic nanowire segment providing atomic resolution magnetic field imaging gradients as well as large force gradients for high sensitivity, and a noble metal enhanced nanowire segment providing efficient scattering cross-section from a sub-wavelength source for optical readout of nanowire vibration. A nonmagnetic/non-plasmonic nanowire segment acts as a cantilever structure for mechanical detection of magnetic resonance. In MRFM, a ferromagnetic tip is brought near the sample being the other end fixed to a cantilever. The cantilever, with very low spring constant, is in a perpendicular configuration to determine high force sensitivity. They stick electrostatically to the surface if used in a parallel configuration. The nuclear spins in the sample are polarized by the inhomogeneous magnetic field. A second oscillating magnetic field is applied by an RF coil, which excites a spin resonance in the atoms of the sample. By slow frequency or amplitude modulation of the RF field, a modulation in the nuclear magnetization of the resonant

Fig. 7.29 Scheme of the magnetic resonance force microscopy. See the coil system that produces the field gradient; the tip is fixed to the end of the cantilever. The changes in the cantilever oscillation due to the magnetic resonance are optically detected [245]

228

M. Va´zquez et al.

fraction of the sample occurs, leading to a modulation in the force between the sample and the magnetic tip. This force produces a measurable oscillation in the deflection of the cantilever, which is optically detected. For a Fe tip with a 100 nm radius of curvature, the field gradient at the surface is approximately 110 G/cm. Three-dimensional elementally specific images can be constructed by scanning the magnetic tip in three dimensions and scanning the RF frequency. An example is given in Fig. 7.30. In (a), a schematic view is given of a CaF2 structure used for the imaging test that represents a thin film evaporated onto a template with a focused ion beam. The dimensions, in nanometres, are taken from electron micrographs. A simulated image for the cyclic-CERMIT protocol using a conical tip model is given in (b), while the magnetic resonance image taken at a tip–sample spacing of 45 nm is shown in (c), where the image colours represent the spatially varying mean-square force signal. The resonant field was 2.89 T, with an applied field of 2.83 T. Generally, good correlation with the expected morphology is observed, although the island at the right of the image is only barely visible. This discrepancy is perhaps due to a slight tilt of the sample with respect to the plane of the scan. The data were acquired with a

Fig. 7.30 (a) Schematic view of a CaF2 structure used for the imaging test. (b) Simulated image for the cyclic-CERMIT protocol using a conical tip model. (c) Magnetic resonance image taken at a tip–sample spacing of 45 nm, where the image colours represent the spatially varying meansquare force signal. (d) Line scan showing raw image data taken from the location of the dotted line in c [247]

7 Nanostructured Magnetic Sensors

229

measurement bandwidth of 0.44 Hz. In (d), the line scan shows raw image data taken from the location of the dotted line in (c). The 100 nm and 250 nm gaps in the test sample are both resolved with essentially 100% contrast [247]. MicroSQUIDs Superconducting quantum interference devices (SQUIDs) are currently used as ultrasensitive magnetic flux detectors in research and industrial applications. Sensors for detection of magnetic field below 100 pT in the frequency range between 1 and 1 kHz are largely dominated by the technologies of fluxgates and superconducting quantum interference devices (SQUIDs). Unfortunately, these sensors are bulky and expensive. Above 100 pT, small inexpensive solidstate thin-film sensors provide a more attractive alternative. A major challenge is thus to extend the range of thin-film sensors below 100 pT. A promising approach is to develop a magnetically softer sensing layer. Recently, ultrasoft magnetic materials have been developed with coercivity as small as 102 mT in the easy axis magnetization and very large magnetic permeability (105) in the magnetization hard axis direction. The major driving force for this development has been the demand for higher efficiency in electrical transformers, and these soft materials have been developed in bulk form. The hard saturation field of the soft layer in commercial thin-film sensors is typically several tenths of mT which corresponds to susceptibility values of the order of 103. Thus, there is a potential to improve permeability of the order of 103 if the properties of the best soft materials can be integrated in these sensors. Technological improvements require the understanding of dynamical magnetization reversal processes at nanosecond time scales. New strategies are needed to overcome the limitations of current devices. For example, the application of high fields to reverse the magnetization of high-anisotropy nanoparticles as well as the measurement of such reversal process. Thirion et al. [248] have proposed a new method to overcome these limitations. A constant applied field, well below the switching field, combined with a radiofrequency (RF) field pulse can reverse the magnetization of a nanoparticle. The efficiency of this method is demonstrated on a 20 nm diameter cobalt particle by using the microSQUID technique. The miniaturization of these devices has been possible, thanks to the use of the electron-beam lithography. Reducing the size of SQUIDs to the micrometre regime (microSQUIDs) has already given important scientific results in the characterization of magnetic nanoparticles. Figure 7.31 shows a Josephson junction (microbridge) of a microsuperconducting quantum interference device, on which a 20 nm-diameter h.c.p. cobalt particle was placed. The microbridge of the SQUID is used like a strip line. An injected RF supercurrent IRF induces an RF field HRF that is directly coupled to the nanoparticle on the microbridge. In addition, microSQUID can be used as a particular scanning microscopy as reported by Tsuei et al. [249]. In 1994, the use of a high-resolution scanning SQUID microscope made possible the first direct observation of the half-flux

230

M. Va´zquez et al.

Fig. 7.31 (a) Scanning electron microscope image of the microbridge junction. The SQUID is patterned from a 20 nm thick niobium film. (b) Schematic of the microbridge junction. The microbridge of the SQUID is used like a strip line. An injected RF supercurrent IRF induces an RF field HRF that is directly coupled to the nanoparticle on the microbridge [248]

quantum effect in YBa2Cu3O7 (YBCO). The presence of a spontaneously generated half-flux quantum in the three-junction ring centred at the tricrystal meeting point and the fact that there is no magnetic flux in the other three rings (as shown in Fig. 7.32) represent the first definitive evidence for dx2y2-wave pairing symmetry in a cuprate superconductor. Magnetoencephalography (MEG) is an imaging technique used to measure the magnetic fields produced by electrical activity in the brain via extremely sensitive devices such as microSQUIDs. It measures the magnetic fields generated by the intercellular currents of neurons in the brain. The magnetic field generated by a single neuron is extremely low in amplitude, and when several thousands of closely packed cells are synchronously active, the resultant extracranial magnetic field takes values of the order of picoTesla.

Fig. 7.32 Micrograph of a scanning microSQUID probe (a) and scanning SQUID image in a YBCO (b) [249]

7 Nanostructured Magnetic Sensors

231

The SQUID magnetometer systems with small diameter pickup coils have been developed to measure in vivo MEG on small animals. The diameter of a pickup coil was chosen according to the size of objectives. The system with the pickup coils of 4 mm in diameter was used for MEG measurement of a swine, 5 mm for MEG measurement of a rat and 200 for MCG measurement of a frog [250].

Hall Probe Microscopy Hall micromagnetometry and scanning Hall probe microscopy, SHPM, have been shown to be very sensitive and versatile experimental techniques for detection of extremely small magnetic fields. The traditional materials applied for the fabrication of this kind of sensors are GaAs/AlGaAs heterostructures, with a two-dimensional electron gas (2DEG) buried below the surface due to the large Hall coefficient. However, this material is extremely sensitive to patterning that induces charge depletion that limits the applications of these devices in lateral size and temperature. To overcome such 2DEG limitations, different materials are thus currently investigated to fabricate sub-micron Hall probes. Sandhu et al. [251] have demonstrated that Bi and InSb thin films are practical alternative materials for fabricating sub-micron Hall effect probes of high spatial resolution for room temperature scanning Hall probe microscopy. As mentioned in their work, GaAs/AlGaAs 2D electron gas Hall probes become impractical for sub-micrometric dimensions at room temperature, mainly due to surface depletion effects that limit the Hall driving current and magnetic sensitivity. The authors fabricated the InSb probes using photolithography, while Bi probes are prepared by optical and focused ion-beam lithography. Fabricating Hall magnetometers by means of focused ion-beam milling or focused electron-beam-induced deposition, active areas as small as 104 nm2 can be obtained [252, 253] (see Fig. 7.33) [254]. Three different classes of materials as metallic Au, semi-metallic Bi and doped bulk Si doped GaAs semiconductor were selected. It was found that Au nanoprobes can work from room temperature down to liquid helium temperature with a magnetic flux sensitivity less than 0.21015 Tm2 (10% of 0).

Fig. 7.33 SEM micrographs of Hall devices with an active area of about 500 nm2 fabricated by focused electron-beam-induced deposition [254]

232

M. Va´zquez et al.

Fig. 7.34 SHPM images of a YBCO thin film in (a) 0.1 Oe (gray scale spans 0.73 G), (b) 20.1 Oe (gray scale spans 0.8 G), (c) 1.1 Oe (gray scale spans 1.1 G), (d) 20.9 Oe (gray scale spans 1.1 G) [255]

An example of the resolution of the SHPM is shown in Fig. 7.34, where a thin film of YBCO can be observed with precipitates after field cooling to 77 K [255]. The advantages of the use of Hall sensors made of InSb type thin films in the cryogenic range, between 1.8 and 250 K, have been recently analysed [256]. The sensors are grown epitaxially on insulating GaAs substrates, and for specified temperature region, the intrinsic electron concentration is well below 1016 cm3. To achieve magnetic sensors with the temperature coefficient of the magnetic sensitivity  as low as 0.001%/deg, the InSb films have to be doped to the effective donor concentration of about 1018 cm3. The investigations have shown that epitaxial n-InSb/GaAs thin-film structures are an excellent material for the preparation of very high quality Hall sensors in cryogenics applications. By changing the film thickness and the technological parameters, such as the doping level, sensors with various adjustable parameters can be manufactured. Their temperature coefficient of the input/output resistance  ¼ 104/K in the temperature range from 1.8

7 Nanostructured Magnetic Sensors

233

Fig. 7.35 Comparison of sensitivities for various magnetic sensors with different size of sensing element (MFM, microSQUIDs and Hall probes)

to 250 K is smaller than those reported earlier for InSb magnetic sensors [257]. This is very important for the accuracy of measurements and/or magnetic field stabilization. They also show lack of any Shubnicov-de Haas oscillations, both in the Hall voltage and in the resistance. Finally, Fig. 7.35 summarizes the sensitivity for various magnetic field sensors based on MFM, microSQUID and Hall probes of different size.

7.3.3 Magnetic Sensors Based on Bidimensional Magnetic Nanostructures The magnetic quality of thin films is being improved, thanks to the developing of manufacturing techniques and optimization of the magnetic properties through the control of preparation methods. Here, in a first section, we summarize some of the applications currently used in connection to magnetic recording, mainly reading the information. Afterwards, we consider sensors for other applications.

234

M. Va´zquez et al.

7.3.3.1 Magnetic Recording and Related Sensors Magnetic recording is the most extended application for the 2D magnetic nanostructures. It essentially involves a recording media where magnetic information is stored, being each bit of information typically a single-domain structure, and the writing and reading processes by which such information is stored and read, respectively. The recording media is made out of a relatively hard magnetic material, and the main tasks in storage process are the cost reduction and development of faster, more compact and less power-consuming memory systems, with greater storage capacity. In a first approach, those benefits are obtained by reducing the size of the basic storage units; hence it is natural to assume that nanotechnology will play a fundamental role in this process. However, it is imperative that the whole system must also include a head for writing and reading the units efficiently, entity to which we will pay attention in this section due to its intrinsic sensing character. The applications of nanomaterials as patterned media for hard disks, miniature magnetic sensors using GMR effect and magnetic memory cells which are being developed are outlined in several text books. Conventional magnetic recording has been based on longitudinal recording media where bits are stored in an in-plane configuration of the planar recording media. The recording media consists of a regular array of isolated singledomain magnetic element, and the requirements for a thin film with high application potential lie in a good response to the recording head (coercivity not too high) and to retaining the magnetization in spite of the magnetostatic field of adjacent bits, stray fields and ambient temperature fluctuations (coercivity not too low). The coercivity of the medium is typically in the range of 500–3000 Oe. A bit ideally should be composed of a single-domain, isolated magnetic particle. In practice, approximately 1000 particles are required to constitute a bit in order to ensure a sharp transition between two information units. The thin-film media also must have sufficient high remanent magnetization and saturation magnetization to be easily detected by the head. Figure 7.36 shows schematically a system for conventional magnetic recording device. The writing process involves passing a current through the coil of the writing head. This current generates a field in the air gap of the U-shaped core and a fringing field in the plane of the tape or disk that extends out of the gap or disk that is moving past it. The fringing field will change the magnetic state of the media, and if the magnetic properties of the media are appropriate, the remanence of the tape in that region will be proportional to the coil current. For digital signals only two remanent states are required for the material and hence the material requirements are not as stringent as for analogue recording, although smaller particle size is desired for high storage capacity and faster access time. The reading process, when carried out with an inductive head, is similar to the writing process; the magnetic field extending out from the tape or disk induces a field in the core of the head that, in turn, generates a voltage in the reading coil. When

7 Nanostructured Magnetic Sensors

235

Fig. 7.36 Writing/reading magnetic device that uses a GMR reading head [258]

using a GMR head, the reading process takes place measuring the changes in the resistance of the GMR probe due to the mentioned extending magnetic field out from the bit information in the disk or tape. The areal density of magnetic recording has been increased over the past few years at 60%/year due to the current optimization of successive generations of thin-film heads, miniaturization of the planar write head and improvement of the Co-based in-plane thin-film media. However, the effects associated with the finite grain size of magnetic alloys used for storage limit the miniaturization [259]. A conventional hard disk is constituted by a thin film of granular Co-alloys. The thin films contain exchange decoupled grains, each grain reversing individually its magnetization. Intergranular exchange coupling is suppressed by adding a non-magnetic element (Cr, for example) that segregates at the grain boundaries. In order to increase the storage capacity, it is necessary to reduce the size of the grains and to guarantee the stabilization of those grains. Recording media is reaching nowadays its limits for stored information: from a basic viewpoint it deals with the superparamagnetic limit indicating a saturation in the reduced size of bits (thermal energy overcoming the magnetic anisotropy stored in the volume of the particle). Other technical limitations are related to signal-to-noise ratio, proportional to the number of particles in a bit, the dispersion of magnetic characteristics or protective coating. The new generation of media involves perpendicular storage media where bits are in an out-of-plane configuration of a hard media, while at the bottom a soft layer allows the closing of the magnetic flux. Perpendicular recording enables a significant increase of areal density of information, where the superparamagnetic limit is no more a restriction due to the volume increase of the bit due to the perpendicular dimension. Present density reached at the laboratory is

M. Va´zquez et al.

236 “Ring” InductiveWriteElement

Longitudinal Recording

Perpendicular Recording

Fig. 7.37 Comparision between longitudinal and perpendicular magnetic recording [260]

around 200 Gb/in.2 Figure 7.37 shows schematically a comparison between longitudinal and perpendicular recording media [260]. Future trends that are currently being investigated are heat-assisted magnetic recording/writing, bit patterning of media and eventually self-organized media technologies [261]. Heat-assisted recording essentially consists of increasing locally the temperature of the bit in order to reduce the field required to print that information. Bit patterning considers the possibility of fabricating single bits consisting of a single magnetic unit (nanowire, nanocolumn or nanoobject in general) that enables an increase of density of storage since neighbouring bits are no more exchange coupled. Self-organized media includes arrays of self-assembled magnetic nanoparticles (i.e. FePt around 3 nm size) which at the moment present some technical difficulties to reach the hard phase and exhibit ordering simultaneously. Alternatively, arrays of magnetic nanowires as bits are obtained by electroplating filling of self-organized pores in non-magnetic media as alumina membranes. Continuous magnetic thin films containing arrays of antidots have been also suggested as innovative recording media. In Fig. 7.38, an array of Ni antidots is shown prepared by sputtering onto a nanoporous alumina membrane: the Ni film exhibits an in-plane uniaxial magnetic anisotropy, and the presence of nanostructured antidots determines arrays of small in-plane single domains or bits of information. These different attempts are being searched nowadays with a final objective of reaching an areal density of information of up to 1 Tbit/in.2

7 Nanostructured Magnetic Sensors

237

Fig. 7.38 Ni antidot array as a potential magnetic recording media (50 nm antidot diameter)

In all the cases, magnetic recording essentially involves detecting changes in the direction of magnetization in the storage medium. The corresponding stray fields are read generally by resistance changes in the head as the stored medium goes through them. The most commonly used technology to read the information is based on the magnetoresistance phenomenon observed in thin films. More recently, with the discovery of the giant magnetoresistance, GMR, in multilayers [188], the efficiency of the magnetic recording systems, especially the reading head, was significantly enhanced. Specific deposition technology has been developed for the fabrication of thinfilm magnetoresistive magnetic head readers. Deposition of sensor materials involves the deposition of a stack consisting of a large number of individual layers, whereby the individual layers are extremely thin (1 nm). Thickness control over such depositions can be achieved by a number of methods. Deposition of sensor materials also involves a strict control over the microstructure (grain structure and interfaces). Detailed description of the deposition technology of thin films for magnetic recorder read head can be found elsewhere [262]. Apart from applications in magnetic recording, various commercial magnetoresistive sensors using conventional MR materials, such as NiFe and NiCo, are available on the market, used in rotation, angular and position sensing. For certain uses, the enhancement of the sensor output is required rather than enhance the sensitivity to the magnetic field. In such cases, conventional MR sensors can be replaced by GMR sensors of the superlattice type (not spin valve) which can produce a larger signal than spin valve and conventional MR sensors. They are strongly resistant to electromagnetic interference, and the

238

M. Va´zquez et al.

preparation of the superlattice is easier than that of the spin valve, resulting in high production throughput. Therefore, GMR superlattice sensors have many advantages for application to rotation, angular and position sensing in automobiles and aircraft. However, there is the drawback that superlattices may not be durable enough to be used under such severe conditions. On the other hand, for practical applications, durability for a much longer period in air is required. Suzuki et al., 10 years ago, have shown that Co/Cu superlattice GMR sensors passivated from the air by a SiO2 layer show extremely high durability against high temperature air [263]. The exchange bias is one of the phenomena associated with the exchange anisotropy created at the interface between an antiferromagnetic and a ferromagnetic material when the system is cooled through the Ne´el temperature of the antiferromagnetic material [264, 265]. Possible applications of the exchange bias effect include permanent magnets, magnetic recording media and, which is most interesting for this propose, sensors based on giant magnetoresistance since the exchange bias systems show a reduction of the saturation fields. Spin valves (SVs) constitute a major type of metal-based spintronics which utilizes both the spin and the charge properties of an electron. It has been extensively investigated since the last decade because of its important application in hard disk drives and magnetic random access memories (MRAM). One of the common driven forces in magnetic storage and memory devices is the ever-growing demand of higher storage density. It requires the ever small and in particular more sensitive read sensor. For example, sensors with a feature size of sub-100 nm and sensitivity of more than 15 mV/mm are required for hard disk drives with an areal density of more than 100 Gb/in.2 This means that the SVs must be electrically and magnetically robust; thus the magnetoresistance ratio (MR) of the spin valve should be as large as possible in order to increase the sensitivity. In this sense, nano-oxide layer (NOL) added SVs is very promising in terms of MR ratio because of the enhancement of the specular reflectivity. Li et al. [266] systematically studied the effect of NOL on the electrical and magnetic properties of the SVs. The NOL layer can smooth the surface topography of the interface so that it can suppress the ferromagnetic Ne´el magnetostatic coupling and, at the same time, enhance the RKKY exchange coupling between the free and pinned layers. About 4.5% of MR ratio has been observed for a particular sensor configuration with laminated CFe/Cu used as the free layer and NOL layer inside the pinned layer. Liu et al. [267] have investigated the effect of Ta buffer layer in IrMn top spin valve. They conclude that, using 3 nm Ta buffer layer, spin valves with high MR (9.24%), high exchange bias field (255 Oe) and low coercivity (2.43 Oe) are obtained. Utilizing these spin valve thin films and standard IC process, mass production of robust GMR sensors could be achieved. Magnetic tunnel junctions (MTJ) [268] have attracted considerable attention in the last years due to their high potential in applications such as magnetic random access memory (MRAM), magnetic read head in hard disks (HD) and

7 Nanostructured Magnetic Sensors

239

highly sensitive magnetic sensors [269]. During the last years new advances using MTJ materials have been carried out in order to increase the tunnel magnetoresistance (TMR) ratio [270]. The standard fabrication process for MTJ involves both e-beam and optical lithography. For large-scale production of relatively large devices this process has many advantages. However, on the nanometre scale, difficulties are encountered with the lift-off step of both the insulator and the removing of the resist. Wei et al. [271] have reported a controlled fabrication method for nanoscale double barrier magnetic tunnel junctions (DBMTJs) that showed good characteristic properties with TMR ratio, resistance-area product and switching field. This method consists in the deposition of the MTJ film with double barriers on Si/SiO2 wafer using magnetron sputtering system. A platinum nanoscale pillar was deposited by focused ion beam on the metal stack to act as a patterning mask. UV lithography with Ar-ion etching was used to pattern the top and bottom electrodes of the DBMTJs. A non-volatile magnetic random access memory (MRAM) (see Fig. 7.39) essentially consists of an array of individual magnetic memory cells, each one being a magnetic tunnel junction (MTJ) element and a diode electrically connected in series. Each MTJ is formed by a pinned ferromagnetic layer whose magnetization direction is prevented from rotating by some mechanism like antiferromagnetic coupling due to an antiferromagnetic additional layer, a free ferromagnetic layer whose magnetization direction is free to rotate between states of parallel and antiparallel to the fixed magnetization layer and an insulating tunnel barrier between and in contact with the two ferromagnetic layers. Each memory cell in the array is controlled by only two lines. The write line applies the currents to change the magnetic state of a MTJ by use of the write current inherent magnetic fields to rotate the magnetization of the free layer. The writing lines are used to read the ‘‘information’’ by

Fig. 7.39 Scheme of an array of magnetic memory cells used for MRAM

M. Va´zquez et al.

240

(a)

(b)

Fig. 7.40 (a) Scheme of the device. (b) The TMR changes sign with Vg. Linear response resistance R as a function of H at temperature T ¼ 1.85 K for different Vg. The dark grey (light grey) arrow indicates the up (down) magnetic field sweep direction, respectively. Inset: SEM picture of a carbon nanotube (NT) in contact with ferromagnetic PdNi strips

measuring the resistance across the MTJ junction that will depend on the relative magnetization direction between the ferromagnetic layers. Spintronics [272] aims to develop electronic devices whose resistance is controlled by the spin of the charge carriers that flow through them. The most basic spintronic device is the spin valve [273] formed by two ferromagnetic electrodes separated by a thin tunnelling barrier. As shown in a previous section, in most cases, its resistance is greater when the two electrodes are magnetized in opposite directions than when they are magnetized in the same direction, i.e. the magnetoresistance is positive [188, 189]. However, if the transport of carriers inside the device is spin- or energy dependent, the opposite can occur and the magnetoresistance is negative [274]. Sahoo et al. have tried to construct an analogous device to a field-effect transistor by using this effect to control spin transport and magnetoresistance with a voltage applied to a gate [275]. They achieve a pronounced gate-field-controlled magnetoresistance response in carbon nanotubes (see Fig. 7.40) connected by ferromagnetic leads. Both the magnitude and the sign of the magnetoresistance in the resulting devices can be tuned in a predictable manner. This opens an important route to the realization of multifunctional spintronic devices.

7.3.3.2 Other Sensors Based on Bidimensional Nanostructures Magnetic Computer Sensors for Biomolecules Studies As described in a recent paper [276], NIST researchers found that arrays of ‘‘spin valves’’ switches, commonly used as magnetic sensors in the read heads of high-density disk drives, also show promising use as tools for controlled trapping of single biomolecules. The arrays might be used in chip-scale, low-power

7 Nanostructured Magnetic Sensors

241

microfluidic devices for stretching and uncoiling, or capturing and sorting, large numbers of individual biomolecules simultaneously for massively parallel medical and forensic studies, a sort of magnetic random access memory (MRAM) for biosciences. Spin valves are made by stacking thin layers of materials with different magnetic properties. Their net magnetization can be switched ‘‘on’’ and ‘‘off’’ by applying an external magnetic field of sufficient strength to align the electron ‘‘spins’’ in the magnetic layers in the same (on) or opposite (off) directions. NIST researchers made an array of spin valves, each about 1  4 mm in size, patterned on a 200 nm thick silicon nitride membrane in fluid. When the spin valves are turned on, a local magnetic field is created that is stronger near the ends of the magnetic stack below the membrane, a field strong enough to trap nanoscale magnetic particles. The NIST experiments demonstrated that the spin valves not only can trap magnetic particles but also can be used as the pivot point for rotating strands of particles when a rotating magnetic field is applied (see Fig. 7.41). These experimental results, combined with computer modelling, suggest that if biomolecules such as proteins or DNA strands were attached to the magnetic particles, the spin valve array could apply torsional forces strong enough to alter the structure or shape of the biomolecules. The NIST group is now working on a microfluidic chip that will accomplish this electronically, which would be a significant milestone for applications. Parallel processing of single biomolecules would represent a significant advance over existing techniques limited to studying one molecule at a time. Optical tweezers, which use lasers to trap and manipulate biomolecules, tend to

Fig. 7.41 Image (taken from a video) shows a strand of magnetic particles trapped by a ‘‘spin valve’’, highlighted in white and rotated by the application of a rotating magnetic field. The use of spin valve arrays for parallel processing of biological molecules is presently under study

242

M. Va´zquez et al.

be slow and limited in force, and the particles need to be micrometre sized or larger. Existing magnetic tweezers can trap smaller particles and apply torque, but typically require permanent immobilization of biomolecules, which is time consuming and prevents subsequent analysis. Sensors Based on Magneto-optical Effects Other kinds of magnetic sensors are those based on magneto-optical effects. Magneto-optical sensors provide the opportunity to combine the advantages of optical methods (i.e. contactless, wide dynamic range, absence of electrical connections) with those of magnetic methods, thus easing the requirements placed at the overall set-up. Magneto-optical devices are mainly used for switching, polarization and amplitude modulation of light. Industrially available magneto-optical sensor is based in the magneto-optical current transformer measuring magnetic field of strong currents by the Faraday rotation in diamagnetic fibres. Particularly, orthoferrites are good candidates for this purpose since they are ferrimagnetic, transparent in the visible and nearinfrared regions of the spectrum, present high domain wall (DW) mobility and high magneto-optical effect which provide rich opportunities for using in sensor application utilizing the modulation of light [277, 278]. The combination of such magneto-optical properties with the low-dimensional effects has been proposed [279], a new magneto-optical surface plasmon resonance (MOSPR) sensor which can improve the sensitivity of the conventional surface plasmon resonance (SPR) sensors (see Fig. 7.42). This MOSPR sensor is based on the combination of the surface plasmon resonance in thin metallic layers and the magneto-optic (MO) activity of ferromagnetic metallic materials. Such device generates a large enhancement of the MO effects closely localized at the surface plasmon resonance. The sensor device uses Co/Au multilayers of nanometric thicknesses as transducers, a prism-coupling configuration and p-polarized light to excite the surface plasmon, rotating magnets or magnetic coils to apply a modulating magnetic field, and detects the

Fig. 7.42 Schematic representation of the surface plasmon resonance system developed by Sensia [280, 281]. In the magneto-optic sensors the gold layer is substituted by a gold matrix with magnetic nanostructures

7 Nanostructured Magnetic Sensors

243

magneto-optic effects of the reflected light as a function of the angle of incidence. A large and sharp enhancement of the MO effect is produced at the angle of incidence in which SPR is excited and depends also on the refraction index of the adjacent dielectric medium. The layers are prepared by physical methods as sputtering and molecular beam epitaxy. The experimental characterization of the MOSPR sensor has shown an increase in the limit of detection in a factor of three in changes of refractive index and in the adsorption of biomolecules as compared to the standard SPR sensors. An improvement of the limit of detection up to one order of magnitude can be achieved by an adequate combination of the magnetic metallic layers and by decreasing the noise of the experimental set-up.

Gas and Humidity Sensors Applications Sandu et al. [282] have shown that cobalt–manganese–ferrite thin films appear to be quite suitable for an application as gas sensors. A sensitive layer for gas sensor application requires a high surface activity which is strongly dependent on the magnitude of their surface area and the nature of their porosity. The necessary properties are a low density and a high surface area, which imply a small crystallite size. The effective index N of a guided mode is changed by adsorption and desorption of molecules on the surface of a planar waveguide or inside the volume of the waveguiding film itself. The principle of integrated optical grating couplers and Bragg reflector switches or gas sensors is as follows. An adsorbate increases the effective refractive index N of the guided mode in a planar waveguide. A change in N is very sensitively detected in an input grating coupler as a change in the power of the in-coupled mode and in a Bragg reflector as a change in its transmission or reflectance. Tiefenthaler and Lukosz [283] showed that SiO2/TiO2 waveguides are highly sensitive sensors or switches actuated by adsorption of a few monolayers of water either on the surface or in the micropores of the waveguide itself. Therefore these integrated optical grating devices can be employed as adsorption–desorption actuated directional switches or gas sensors.

7.4 Final Remarks The investigation of nanoparticles, nanowires and thin-film magnetic nanostructures, from their preparation by different routes to their experimental characterization and modelling of magnetic behaviour, and finally to their manipulation at the nanoscale, has become during the last decade one of the hot topics in multidisciplinary research. The manipulation of such nanostructures in a controlled way has enabled the development of a series of applications and particularly of magnetic sensor devices that make use or are based on the

244

M. Va´zquez et al.

particular properties of those magnetic nanostructures. Such development has been related to the parallel optimization of experimental equipments to face the real knowledge and even direct observation of the magnetic nanoscale nature. The perspectives of technological use of such magnetic nanostructures in many different fields are really very broad; nevertheless some restrictions are still evident in connection with the ‘‘nanodimensions’’. In the present chapter we have reviewed some of the most relevant and useful aspects of magnetic nanostructures starting by the methods of fabrication and magnetic characterization. Sensing devices and technologies in general make use of the particular shape of magnetic nanostructures. In that way, we could summarize by saying that most outstanding technologies of magnetic nanoparticles are related to biomedical applications as for magnetic resonance imaging, contrast agents or separation media which is at least partially due to their smallest size. In turn, magnetic nanowires are particularly useful in those applications making use of their directional character, for example, as magnetic probes in a number of magnetic imaging devices. Finally, bidimensional magnetic nanostructures are particularly important in magnetic field reading in magnetic recording applications. This is probably the most relevant application of all magnetic nanostructures in connection with the huge funding resources invested to develop novel families of recording media and phenomena ascribed to the nanoscale. In this regard, we believe that the real challenge is in reducing size in a controlled way, and most promising effects are probably connected to new recording media and sensing devices for magnetic storage of information.

References 1. T. Hutter and B. Danhamer: Thermoanalytical sensor, and method of producing the thermoanalytical sensor, patent US2005169344 (August 2005). 2. R. Welch, Y. Tung-Sheng and W. Miller: Thermistor sensor probe, patent USD507977S (August 2005). 3. J. Feder, R. Beyerle, S. Byers and T. Jones: Thermal control of a DUT using a thermal control substrate, patent US2005007136 (January 2005). 4. J. Laurence and N. Corner: Optical temperature, patent US5180227 (January 1993). 5. G. Bartington: Fluxgate sensor, patent GB2411964 (September 2005). 6. N. Holmes: Anisotropic magnetoresistive sensor, patent GB2388915 (November 2003). 7. Ch. Rumenin: Microsensor for magnetic field, patent BG108430 (June 2005). 8. N. Hasegawa: Spin-valve type magnetoresistive sensor and method of manufacturing the same, patent US6913836 (July 2005). 9. C. Salter and R. Pendergrass: Thin film gas sensor configuration, patent US2005183967 (August 2005). 10. H.V. Markovych, A.I. Bondarchurk, V. Bubleinyk, V. Medvediev, V. Koptikov and L. Lupo: Catalytic thermal sensor for detecting combustible gases or vapors, patent UA73019 (January 2005). 11. Ch.-K. Kim, Ch.-S. Yoon, J.-H. Lee and Ch.-S. Yu: Micro-magnetoelastic biosensor array for detection of DNA hybridization and fabrication method thereof, patent US2004014201 (January 2004).

7 Nanostructured Magnetic Sensors

245

12. J. Viola and W. Moore: Magnetoelastic pressure sensor, patent US2004093951 (May 2004). 13. J. Viola and J. Laidlaw: Magnetoelastic torque sensor assembly, patent GB2395568 (May 2004). 14. H. Morimoto: Strain gauge type sensor and gauge type sensor unit using the same, patent WO2005045388 (May 2005). 15. S. Masuda and K. Sumino: Position sensor, patent JP8313237 (November 1996). 16. P. Tartaj, M.P. Morales, S.Veintemillas-Verdaguer, T. Gonza´lez-Carren˜o and C.J. Serna (2003). J. Phys D, Appl. Phys., 36, R182. 17. E. Duguet, S.Vasseur, S. Mornet, G. Goglio, A. Demourgues, J. Portier, F. Grasset, P. Veverka and E. Pollert (2006). Bull. Mater. Sci., 29, 581. 18. A.-H. Lu, E.L. Salabas and F. Schuth (2007). Angew. Chem. Int. Ed., 46, 1222. 19. J. Dobson (2006). Nanomedicine, 1, 31. 20. D. Cai, J.M. Mataraza, Z.-H. Qin, Z. Huang, J. Huang, T.C. Chiles, D. Carnahan, K. Kempa and Z.C. Ren (2005). Nat. Methods, 2, 449. 21. Y. Bao and K.M. Krishnan (2005). J. Magn. Magn. Mater., 293, 15. 22. A. Jordan, P. Wust, H. Fahling, W. John, A. Hinz and R. Felix (1993). Int. J. Hyperthermia, 9, 51. 23. P. Tartaj, M.P. Morales, S. Veintemillas-Verdaguer, T. Gonzalez-Carren˜o and C.J. Serna, ‘‘Synthesis, properties and biomedical applications of magnetic nanoparticles’’, Handbook of Magnetic Materials, ed. K.H.J. Buschow # 2006 Elsevier B.V., 16, 403–483. 24. J. Park, J. Joo, G.K. Soon, Y. Jang and T. Hyeon (2007). Angew. Chem. Int. Ed. 46, 4630. 25. R. Massart and V. Cabuil (1987). J. Chem. Phys., 84, 967. 26. M. Tada, S. Hatanaka, H. Sanbonsugi, N. Matsushita and M. Abe (2003). J. Appl. Phys., 93, 7566. 27. T. Sugimoto and E. Matijevic´ (1980). J. Colloid Interface Sci., 74, 227. 28. T.J. Daou, G. Pourroy, S. Be´gin-Colin, J.M. Grene`che, C. Ulhaq-Bouillet, P. Legare´, P. Bernhardt, C. Leuvrey and G. Rogez (2006). Chem. Mater., 18, 4399. 29. T. Prozorov, S.K. Mallapragada, B. Narasimhan, L. Wang, P. Palo, M. NilsenHamilton, T.J. Williams, D.A. Bazylinski, R. Prozorov and P.C. Canfield (2007). Adv. Funct. Mater., 17, 951. 30. J. Rockenberger, E.C. Scher and A. Paul Alivisatos (1999). J. Am. Chem. Soc., 121, 11595. 31. Y.-W. Jun, J. Choi and J. Cheon (2006). Angew. Chem. Int. Ed., 45, 3414. 32. J. Park, E. Lee, N.-M. Hwang, M. Kang, C.K. Sung, Y. Hwang, J.-G. Park, H.-J. Noh, J.-Y. Kim, J.-H. Park and T. Hyeon (2005). Angew. Chem. Int. Ed., 44, 2872. 33. V.F. Puntes, K.M. Krishnan and A.P. Alivisatos (2001). Science, 291, 2115. 34. S. Sun, C.B. Murray, D. Weller, L. Folks and A. Moser (2000). Science, 287, 1989. 35. A. Sukhanova, M. Devy, L. Venteo, H. Kaplan, M. Artemyev, V. Oleinikov, D. Klinov, M. Pluot, J.H.M. Cohen and I. Nabiev (2004). Anal. Biochem., 324, 60. 36. M. Darbandi, R. Thomann and T. Nann (2005). Chem. Mater., 17, 5720. 37. T. Pellegrino, L. Manna, S. Kudera, T. Liedl, D. Koktysh, A.L. Rogach, S. Keller, J. Radler, G. Natile and W. J. Parak (2004). Nano Lett., 4, 703. 38. J. Lee, A.O. Govorov and N.A. Kotov (2005). Angew. Chem. Int. Ed., 117, 7605 (2005) and Angew. Chem. Int. Ed., 44, 7439. 39. V. Salgueirinn˜o-Maceira, L.M. Liz-Marzan and M. Farle (2004). Langmuir, 20, 6946. 40. V. Salgueirin˜o-Maceira, M.A.Correa-Duarte and M. Farle (2005). Small, 1, 1073. 41. S. Veintemillas-Verdaguer, M.P. Morales and C.J. Serna (1998). Mater. Lett., 35, 227. 42. S. Veintemillas-Verdaguer, M.P. Morales and C.J. Serna (2001). Appl. Organomet. Chem., 15, 1. 43. T. Gonza´lez-Carren˜o, M.P. Morales, M. Gracia and C.J. Serna (1993). Mater. Lett., 18, 151.

246

M. Va´zquez et al.

44. T. Gonza´lez-Carren˜o, A. Mifsud, C.J. Serna and J.M. Palacios (1993). Mater. Chem. Phys., 27, 287. 45. O. Bomatı´ -Miguel, P. Tartaj, M.P. Morales, P. Bonville, U. Golla-Schindler, X.Q. Zhao and S. Veintemillas-Verdaguer (2006). Small, 2, 1476. 46. E.Y. Sun, L. Josephson, K.A. Kelly and R. Weissleder (2006). Bioconj. Chem, 17, 109. 47. E.Y. Sun, L. Josephson and R. Weissleder (2006). Mol. Imag., 5, 122. 48. W. Wu, Q. He, H. Chen, J. Tang and L. Nie (2007). Nanotechnology, 18, 145609. 49. Y.-W. Jun, J.-S. Choi and J. Cheon (2007). Chem. Commun., 12, 1203. 50. S.M. Moghimi, A.C. Hunter and J.C. Murray (2001). Pharmacol. Rev., 53, 283. 51. L. Burgos-Asperilla, M. Darder, P. Aranda, L. Vazquez, M. Vazquez and E. Ruiz-Hitzky (2007). J. Mater. Chem., 17, 4233. 52. Ø. Olsvik, T. Popovic, E. Skjerve, K.S. Cudjoe, E. Hornes, J. Ugelstad and M. Uhle´n (1994). Clin. Microbiol. Rev., 7, 43. 53. J. Ugelstad, A. Berge, T. Ellingsen, R. Schmid, T.-N. Nilsen, P.C. Mørk, P. Sienstad, E. Hornes and Ø. Olsvik (1992). Progr. Poly. Sci. (Oxford), 17, 87. 54. R. Veyret, A. Elaissari, P. Marianneau, A.A. Sall and T. Delair (2005). Anal. Biochem., 346, 59. 55. Nanotechnologies for life sciences, Vol.1: Biofunctionalization of Nanomaterials, Ed. Challa Kumar, Wiley-VCH, Weinheim, Germany (2005). 56. X. Gao, K.M.K. Yu, K.Y. Tam and S.C. Tsang (2003). Chem. Commun, 2998. 57. Y. Kobayashi, M. Horie, M. Konno, B. Rodr´ guez-Gonza´lez, and L.M. Liz-Marza´n (2003). J. Phys. Chem. B, 107, 7420. 58. Y. Lu, Y. Yin, B. T. Mayers and Y. Xia (2002). Nano Lett., 2, 183. 59. M. Aslam, L. Fu, S. Li and V.P. Dravid (2005). J. Colloid Inter. Sci., 290, 444. 60. P. Tartaj, T. Gonza´lez-Carren˜o and C.J. Serna (2001). Adv. Mater., 13, 1620. 61. F. Caruso, R.A. Caruso and H. Mohwald (1998). Science, 282, 1111. ¨ ´ 62. E. Ruiz-Herna´ndez, A. Lopez-Noriega, D. Arcos, I. Izquierdo-Barba, O. Terasaki and M. Vallet-Regı´ (2007). Chem. Mater., 19, 3455. 63. L. D’Amico, F. D’Orazio, J. L. Dormann, D. Fiorani, F. Lucari and E. Tronc (1995). Mater. Sci. Forum., 195, 173. 64. L. Zhang, G.C. Papaefthymiou, R.F. Ziolo and J.Y. Ying (1997). Nanostruct. Mater., 9, 185. 65. C. Cannas, D. Gatteshi, A. Musinu, G. Piccaluga and C. Sangregorio (1998). J. Phys. Chem B., 102, 7721. 66. F. Del Monte, M.P. Morales, D. Levy, A. Fernandez, M. Ocan˜a, A. Roig, E. Molins, K. O’Grady and C.J.Serna (1997). Langmuir, 13, 3627. 67. S. Solinas, G. Piccaluga, M.P. Morales and C.J. Serna (2001). Acta Mater., 49, 2805. 68. B.A. Fuertes and P. Tartaj (2006). Chem. Mater., 18, 1675. 69. M.A.Correa-Duarte, M. Grzelczak, V. Salgueirino-Maceira, M. Giersig, L.M. Liz-Marza´n, M. Farle, K. Sierazdki and R. Diaz (2005). J. Phys. Chem. B, 109, 19060. 70. Z. Ban, Y.A. Barnakov, F. Li, V.O. Golub and C.J. O’Connor (2005). J. Mater. Chem., 15, 4660. 71. P. de la Presa, M. Multigner, M.P. Morales, T. Rueda, E. Ferna´ndez-Pinel and A. Hernando (2007). J. Magn. Magn. Mater, 316, E753. 72. Z. Lu, M.D. Prouty, Z. Quo, V.O. Golub, C.S.S.R. Kumar and Y.M. Lvov (2005). Langmuir, 21, 2042. 73. H. Kim, M. Achermann, L.P. Balet, J.A. Hollingsworth and V.I. Klimov (2005). J. Am. Chem. Soc., 127, 544. 74. D. Wang, J. He, N. Rosenzweig and Z. Rosenzweig (2004). Nano Lett., 4, 409. 75. Y. Lu, Y. Yin, B.T. Mayers and Y. Xia (2002). Nano Lett., 2, 183. 76. F. Bertorelle, C. Wilhelm, J. Roger, F. Gazeau, C. Me´nager and V. Cabuil (2006). Langmuir, 22, 5385.

7 Nanostructured Magnetic Sensors

247

77. T. Pellegrino, S. Kudera, T. Liedl, A.M. Javier, L. Manna and W.J. Parak (2005). Small, 1, 48. 78. V. Salgueirin˜o-Maceira, M.A. Correa-Duarte, M. Spasova, L.M. Liz-Marza´n and M. Farle (2006). Adv. Funct. Mater., 16, 509. 79. G. Bastiat, P. Oligere, G. Karlsson, K. Edwards and M. Lafleur (2007). Langmuir, 23, 7695. 80. Y. Xia, B. Gates, Y. Yin and Y. Lu (2000). Adv. Mater., 12, 693. 81. M.-P. Pileni (2007). Acc. Chem. Res., 40, 685. 82. M.-P. Pileni (2001). J. Phys. Chem. B, 105, 3358. 83. C. Luna, P. Morales, C. Serna and M. Va´zquez (2003). Nanotechnology, 14, 268. 84. C. Petit, A. Taleb and M.-P. Pileni (1998). Adv. Mater., 10, 259. 85. D. Parker, I. Lisiecki, C. Salzemann and M.-P. Pileni (2007). J. Phys. Chem. C, 111, 12632. 86. I. Lisiecki, C. Salzemann, D. Parker, P.-A. Albouy and M.-P. Pileni (2007). J. Phys. Chem. C, 111, 12625. 87. Y. Lalatonne, L. Motte, V. Russier, A.T. Ngo, P. Bonville and M.-P. Pileni (2004). J. Phys. Chem. B, 108, 1848. 88. A.T. Ngo and M.-P. Pileni (2002). J. Appl. Phys., 92, 4649. 89. S. Sun, C.B. Murray, D. Weller, L. Folks and A. Moser (2000). Science, 287, 1989. 90. A.-P. Li, F. Mu¨ller, A, Birner, K. Nielsch and U. Gosele (1999). Adv. Mater., 11, 483. ¨ 91. M. Va´zquez, K. Pirota, J. Torrejon, D. Navas and M. Hernandez-Velez (2005). J. Magn. Magn. Mater., 294, 174. 92. M. Va´zquez, K. Pirota, M. Herna´ndez-Ve´lez, V. M. Prida, D. Navas, R. Sanz and F. Batalla´n (2004). J. Appl. Phys., 95, 6642. 93. A.M. Haghiri-Gosnet, C. Vieu, G. Simon, M. Mejı´ as, F. Carcenac and H. Launois (1999). J. Phys. IV, 9(2), 133. 94. C.R.K. Marrian and D.M. Tennant (2003). J. Vac. Sci. Technol. A, 21, S207. 95. J.I. Martı´ n, J. Nogue´s, K. Liu, J.L. Vicent and I.K. Schuller (2003). J. Magn. Magn. Mater., 256, 449. 96. C.Ross (2001). Ann. Rev. Mater. Res., 31, 203. 97. W.O. Rosa, M. Knobel, L. Cescato, A.L. Gobbi and M. Vazquez (2007). Solid State Communications 142, 228. 98. C.A. Ross, H.I. Smith, T. Savas, M. Schattenburg, M. Farhoud, M. Hwang, M. Walsh, M.C. Abraham and R.J. Ram (1999). J. Vac. Sci. Technol. B, 17, 3168. 99. S.Y. Chou, P.R. Krauss and P.J. Renstrom (1996). Science, 272, 85. 100. M. Li, L. Chen and S.Y. Chou (2001). Appl. Phys. Lett., 78, 3322. 101. S.P. Beaumount (1991). in ‘‘Low-dimensional Structures in Semiconductors’’. Plenum Press, New York, 109. 102. P.R. Krauss and S.Y. Chou (1997). Appl. Phys. Lett., 71, 3174. 103. T. Mu¨ller, K.-H. Heining and B. Schmidt (2001). Nuc. Instrum. Methods Phys. Res., 175, 468. 104. Y. Sun, D.-Y. Khang, F. Hua, K. Hurley, R.G. Nuzzo and J.A. Rogers (2005). Adv. Funct. Mater., 15, 30. 105. J. Viernow, D.Y. Petrovykh, F.K. Men, A. Kirakosian, J.-L. Lin and F.J. Himpsel (1999). Appl. Phys. Lett., 74, 2125. 106. J.N. Crain, A. Kirakosian, K.N. Altmann, C. Bromberger, S.C. Erwin, J.L: McChesney, J.-L. Lin and F.J. Himpsel (2003). Phys. Rev. Lett., 90, 176805. 107. Q. Xie, A. Madhukar, P. Chen and N. P. Kobyashi (1995). Phys. Rev. Lett., 75, 2542; V.A. Shchukin and D. Bimberg (1999). Rev. Mod. Phys., 71, 1125. 108. P.W. Tasker (1979). J. Phys. C: Solid State Phys., 12, 4977. 109. Z.L. Wang, X.Y. Kong and J.M. Zuo (2003). Phys. Rev., Lett., 91(18), 185502. 110. D. Whang, S. Jin, and Ch.M. Lieber, (2003). Nano Lett., 3(7), 951. 111. D. Whang, S. Jin and Ch.M. Lieber, (2004). J. Appl. Phys., 43(7B), 4465.

248

M. Va´zquez et al.

112. 113. 114. 115. 116.

Y. Cui, Q. Wei, H. Park and Ch.M. Lieber (2001). Science, 293, 1289. A. DeHon (2003). IEEE Trans. Nanotech., 2, 23. G.T. Hu, T.W. Odom and C.M. Lieber (1999). Acc. Chem. Res., 32, 435. W.Q. Han, S.S. Fan, Q.Q. Li and Y.D. Hu (1997). Science, 277, 1278. M.T. Bjork, B.J. Ohlsson, C. Thelander, A.I. Persson, K. Deppert, L.R. Wallenberg and L. Samuelson (2002). Appl. Phys. Lett., 81, 4458. C.R. Martin (1994). Science, 266, 1961. S. de Franceschi (2003). Appl. Phys. Lett., 83, 344. D. Navas, M. Herna´ndez-Ve´lez, and M. Va´zquez, W. Lee and K. Nielsch (2007). Appl. Phys. Lett., 90, 192501. T. Hanrath and B.A. Korgel (2002). J. Am. Chem. Soc., B 104, 1424. Y. Xia, P. Yang, Y. Sun, Y. Wu, B. Mayers, B. Gates, Y. Yin, F. Kim and H. Yan (2003). Adv. Mater., 15(5) 353. B. Erdem Alaca, H. Sehitoglu and T. Saif (2004). Appl. Phys. Lett., 84, 4669. A.P. Alivisatos, K.P. Johnsson, X. Peng, T.E. Wilson, C.J. Loweth, M.P. Bruchez and P.G. Schultz (1996). Nature, 382, 609. B.D. Martin, D.J. Dermody, B.D. Reiss, M. Fang, L.A. Lyon, M.J. Natan and T.E. Mallouk (1999). Adv. Mater., 11, 1021. E. Braun, Y. Eichen, U. Sivan and G.B. Yoseph (1998). Nature, 391, 775. O. Jessensky, F. Mu¨ller and U. Gosele (1998). Appl. Phys. Lett., 72(10), 1173. ¨ M. Herna´ndez-Ve´lez, K.R. Pirota, F. Pa´szti, D. Navas, A. Climent, and M. Va´zquez (2005). Appl. Phys. A, 80, 1701. J.C. Hulteen and C.R. Martin (1997). J. Mater. Chem., 7, 1075. G.K. Mor, O.K. Varghese, M. Paulose, N. Mukherjee and C.A. Grimes (2003). J. Mater. Res., 18, 2588. V.M. Prida, M. Herna´ndez-Ve´lez, K.R. Pirota, A. Mene´ndez and M Va´zquez (2005). Nanotechnology, 16, 2696. K. Nielsch and B. Stadler (2007). Template-based synthesis and characterization of high-density ferromagnetic nanowire arrays, in ‘‘Handbook Magn. Adv. Magn. Mater., 4, 2227–2255. J. Choi, K. Nielsch, M. Reiche, R.B. Wehrspohn and U. Gosele (2003). J. Vac. Sci. ¨ Technol. B, 21(2), 763. H. Masuda, A. Abe, M. Nakao, A. Yokoo, T. Tamamura and K. Nishio (2003). Adv. Mater., 15(2), 161. M. Herna´ndez-Ve´lez (2006). Thin Solid Films, 495, 51. H. Masuda, M. Ohya, H. Asoh and K. Nishio (2001). Jpn. J. Appl. Phys., 40, 217. E.J. Bae, W.B. Choi, K.S. Jeong, J U. Chu, G.-S. Park, S. Song and I.K. Yoo (2002). Adv. Mater., 14, 277. V.M. Prida, K.R. Pirota, D. Navas, A. Asenjo, M. Herna´ndez-Ve´lez and M. Va´zquez (2007). J. Nanosci. Nanotechnol., 7, 272. O. Sa´nchez, M. Herna´ndez-Ve´lez, D. Navas, M.A. Auger, J.L. Baldonedo, R. Sanz, K.R. Pirota and M. Va´zquez (2006). Thin Solid Films, 495, 149. M. Va´zquez, M. Herna´ndez-Ve´lez, A. Asenjo, D. Navas, K.R. Pirota, V. Prida, O. Sa´nchez and J.L. Baldonedo (2006). Physica B, 384, 36. W.D. Westwood (1988). MRS Bulletin 46. P.J. Kelly and R.D. Arnell (2000). Vacuum, 56, 159. J.L. Mene´ndez, G. Armelles, A. Cebollada, C. Quintana, D. Ravelosona, C. Chappert, F. Peiro´ and A. Cornet (2002). Appl. Phys. Lett., 81, 1603. A. Ruiz, L. Gonza´lez, A. Mazuelas and F. Briones (1989). Appl. Phys. A, 49, 729. D.A. Glocker and S.I. Shah (Eds.) ‘‘Handbook of thin film process technology’’, 1, Inst. Phys. Pub. (1996). Y. Cho (1971). ‘‘Film deposition by MBE techniques’’.. J. Vac. Sci. Tehcnol., 8.

117. 118. 119. 120. 121. 122. 123. 124. 125. 126. 127. 128. 129. 130. 131.

132. 133. 134. 135. 136. 137. 138. 139. 140. 141. 142. 143. 144. 145.

7 Nanostructured Magnetic Sensors

249

146. A. Cebollada, J.L. Martinez, J.M. Gallego, J.J. de Miguel, R. Miranda, S. Ferrer, F. Batallan, G. Fillion and J.P. Rebouillat (1989). Phys. Rev. B 39, 9726. 147. R.F.C. Farrow, S.S.P. Parkin, R.F. Marks, K.M. Krishnana and N. Thangaraj (1996). Appl. Phys. Lett., 69, 13. 148. W. Kern and V.S. Van (1978). ‘‘Chemical vapour deposition of inorganic thin films’’ in Thin Film Processes, Ch. 3, Academic Press. 149. W. Kern (1989). ‘‘Chemical Vapour Deposition’’ in Microelectronic Materials and Processing, Ch. 6, NATO ASI Series, E164. 150. R. Reif (1990). ‘‘Plasma Enhanced Chemical Vapour Deposition of Thin Films for Microelectronic’’ in Handbook of Plasma Processing Technology’’, Noyes Pub. 151. I. Goldstein (1992). in ‘‘Scanning Electron Microscopy and X-ray Microanalysis’’, Plenum Press., N.Y. 152. G. Margaritondo (1988). in ‘‘Introduction to Synchrotron Radiation’’, Oxford Univ. Press, N.Y. 153. G. Martı´ nez-Criado et. al. (2005). Appl. Phys. Lett., 87, 061913. 154. D.C. Koenigsberger et. al. (1988). in ‘‘X-Ray Absorption, Principles and Applications. Tehcniques of EXAFS, SEXAFS and XANES’’, Wiley. 155. F. Foner (1959). Rev. Sci. Instrum., 30, 548. 156. S. Foner (1981). IEEE. Trans. Magn., 17, 3358. 157. S. Foner (1967). J. Appl. Phys., 38, 1510. 158. J.S. Philo and W. M. Fairbank (1977). Rev. Sci. Instr., 48, 1529. 159. M. Jaafar, R. Sanz, M. Va´zquez and A. Asenjo, J. Jensen and K. Hjort, S. Flohrer, J. McCord and R. Scha¨fer (2007). Phys. Stat. Sol. (a), 204, 1724. 160. C. Luna, P. Morales, C. Serna and M. Va´zquez (2004). Nanotechnology 15, 293. 161. X. Batlle and A. Labarta (2002). J. Phys. D: Appl. Phys., 35, R15. 162. A.G. Roca, M.P. Morales, K. O’Grady and C.J. Serna (2006). Nanotechnology, 17, 2783. 163. C. Luna, M.P. Morales, C.J. Serna and M. Vazquez (2004). Nanotechnology, 15, S293. 164. D. Paccard, C. Schlenker, O. Massanet and R. Montmory. A. Yelon (1966). Phys. Stat. Sol., 16, 301. 165. R. Jungblut, R. Cowhoorn, M.T. Johnson, J. aan de Stegge and A. Reinders (1994). J. Appl. Phys., 75, 6659. 166. D. Atkinson and R.P. Cowburn (2004). Appl. Phys. Lett., 85, 1386. 167. B.C. Choi, M. Below, W.K. Heibert, G.E. Ballentine and M.R. Freeman (2001). Phys. Rev. Lett., 86, 728. 168. N. Qureshi, S. Wang, M.A. Lowther, A.R. Hawkins, S.Kwon, A. Liddle, J. Bokor and H. Schmidt (2005). Nano Lett., 5, 1413. 169. A. Hubert and R. Sha¨fer (1998). ‘‘Magnetic Domains’’ Springer, New York. 170. L. Baril, D. Mauri, J. McCord, S. Gider and T. Lin (2001). J. Appl. Phys., 89, 1320. 171. G. Chrobok and M. Hofmann(1976). Phys. Lett. A, 57, 257. 172. K. Koike and K. Hayakawa (1984). Appl. Phys. Lett., 45, 585. 173. See for example cnst.nist.gov/Facilities/sempa.html. 174. A. Gavrin and J. Unguris (2000). J. Magn. Magn. Mate., 213, 95. 175. R.J. Celotta, J. Unguris and D.T. Pierce (2002). in Magnetic Interactions and Spin Transport Electronics, edited by S. Wolf and Y. Idzerda, Hingham, MA:Kluwer Academic/Plenum Publishers, 341. 176. P. Fischer, T. Eimu¨ller, G. Schu¨tz, P. Guttmann, G. Schmahl, K. Pru¨gel and G. Bayreuther (1998). J. Phys. D: Appl. Phys., 31, 649. 177. J.J. Sa´enz, N. Garcı´ a, P. Gru¨tter, E. Meyer, H. Heinzelmann, R. Wiesendanger, L. Rosenthaler, H.R. Hidber and H.-J. Guntherodt (1987). J. Appl. Phys., 62, 4293. 178. P. Gru¨tter, T. Jung, H. Heinzelmann, A. Wadas, E. Meyer, H.R. Hidber and H.J. Gu¨ntherodt (1990). J. Appl. Phys., 67, 1437. 179. S. Morita, R. Wiesendanger and E. Meyer (2002). Springer, Berlin. 180. A. Asenjo, M. Jaafar, D. Navas and M. Va´zquez (2006). J. Appl. Phys., 100, 023909.

250

M. Va´zquez et al.

181. J.M. Garcı´ a, A. Thiaville and J. Miltat (2002). J. Magn. Magn. Mat., 249, 163. 182. A. Asenjo, M. Jaafar, E.M. Gonza´lez, J.I. Martin, M. Va´zquez and J.L. Vicent (2007). J. Magn. Magn. Mater., 310, e936. 183. A. Wachowiak, J. Wiebe, M. Bode,O. Pietzsch, M. Morgenstern and R. Wiesendanger (2002). Science, 298, 577. 184. L. Ciobanu, D.A. Seeber and C.H. Pennington (2002). J. Magn. Reson., 158, 178. 185. A. Feintuch, G. Alexandrowicz, T. Tashma, Y. Boasson, A. Grayevsky and N. Kaplan (2000). J. Magn. Reson., 142, 382. 186. A. Blank, C. R. Dunnam, P.P. Borbat and J.H. Freed (2004). Rev. Sci. Inst., 75 (9) 3050. 187. U. Dibbern (1989). ‘‘Magnetoresistive sensors’’, in Magnetic Sensors, 5, ed. by R. Boll, K. Overshott (VCH, Weinheim), 341–380. 188. M.N. Baibich, J.M. Broto, A. Fert, F. Nguyen Van Dau, F. Petroff, P. Eitenne, G. Creuzet, A. Friederich, and J. Chazelas (1988). Phys. Rev. Lett., 61(21), 2472–2475. 189. G. Binasch, P. Gru¨nberg, F. Saurenbach and W. Zinn (1989). Phys. Rev., B 39(7) 4828–4830. 190. J.S. Moodera, L.R. Kinder, T.M. Wong and R. Meservey (1995). Phys. Rev. Lett., 74, 3273. 191. M. Va´zquez and A. Hernando (1996). J. Phys. D: Appl. Phys., 29, 939. 192. G.V. Kurlyandskaya, V. Fal Miyar, A. Saad, E. Asua and J. Rodrı´ guez (2007). J. Appl. Phys., 101, 054505. 193. N. Jaffrezic-Renault, C. Martelet, Y. Chevolot and J.-P. Cloarec (2007). Sensors, 7, 589. 194. S. Pe´rez-Rial, I. Rodrı´ guez, J. Gonza´lez, J.M. Pe´rez-Sa´nchez F. Herranz, N. Beckmann and J. Ruı´ z-Cabello (2007). J. Pharm. Sci., (in press). 195. J.-M. Nam, C.S. Thaxton and C.A. Mirkin (2003). Science, 301, 1884. 196. L.M., Rossi, A.D. Quach and Z. Rosenzweig (2004). Anal. Bioanal. Chem., 380, 606. 197. Ø. Olsvik, T. opovic, E. Skjerve, K.S. Cudjoe, E. Hornes, J. Ugelstad and M. Uhle´n (1994). Clin. Microbiol. Rev., 7, 43. 198. www.dynalbiotech.com. 199. I. Miyashiro, C. Kuo, K. Huynh, A. Iida, D. Morton, A. Bilchik, A. Giuliano and D.S.B. Hoon (2001). Clin. Chem., 47, 505. 200. M. Santandreu, S. Sole´, E. Fa`bregas and S. Alegret (1998). Biosens. Bioelectron., 13, 7. 201. A.G. Gehring, J.D. Brewster, P.L. Irwin, S.-I. Tu and L.J. Van Houten (1999). J. Electroanal. Chem., 469, 27. 202. S. Sole´, S. Alegret, F. Ce´spedes, E. Fa`bregas and T. Dı´ ez-Caballero (1998). Anal. Chem., 70, 1462. 203. M. Dequaire, C. Degrand and B. Limoges (1999). Anal. Chem., 71, 2571. 204. S. Helali, C. Martelet, A. Abdelghani, M.A. Maaref and N. Jaffrezic-Renault (2006). Electrochim. Acta, 51, 5182. 205. G.-H. Kim, A. Rand, Garth and S.V. Letcher (2003). Biosens. Bioelectr., 18, 91. 206. S. Sole´, A. Merkoc¸I and S. Alegret, TrAC - Trends Anal. Chem., 20, 102. 207. A. Erdem, F. Sayar, H. Karadeniz, G. Guven, M. Ozsoz and E. Piskin (2001). Electroanalysis, 19, 798. 208. D.B. Robinson, H.H.J. Persson, H. Zeng, G. Li, N. Pourmand, S. Sun, S.X. Wang (2005). Langmuir, 21, 3096. 209. H. Zhang and M.E. Meyerhoff (2006). Anal. Chem., 78, 609. 210. A. Hultgren, M. Tanase, C.S. Chen, G.J. Meyer and D.H. Reich (2003). J. Appl. Phys. 93, 7554. 211. L. Wang, S. Hong, L. Wang, L. Dong, G. Bian, T. Xia and H. Chen (2006). Spectrochim. Acta - Part A: Mol. Biomol. Spectrosc., 65, 439. 212. J. Wang, A.-N. Kawde, A. Erdem and M. Salazar (2001). Analyst, 126, 2020. 213. M. Megens and M. Prins (2005). J. Magn. Magn. Mater. 293, 702. 214. J. Schotter, P.B. Kamp, A. Becker, A. Pu¨hler, G. Reiss and H. Bru¨ckl (2004). Biosens. Bioelectron., 19, 1149. 215. S.G. Grancharov, H. Zeng, S. Sun, S.X. Wang, S. O’Brien, C.B. Murray, J.R. Kirtley and G.A. Held (2005). J. Phys. Chem. B, 109, 13030.

7 Nanostructured Magnetic Sensors

251

216. D.L. Graham, H.A. Ferreira, P.P. Freitas and J.M.S. Cabral (2003). Biosens. Bioelectron., 18, 483. 217. S.H. Chung, A. Hoffmann, S.D. Bader, C. Liu, B. Kay, L. Makowski and L. Chen (2004). Appl. Phys. Lett., 85, 2971. 218. K. Kriz, J. Gehrke and D. Kriz (1998). Biosens. Bioelectron., 13, 817. 219. M. Lu, F. Ibraimi, D. Kriz and K. Kriz (2006). Biosens. Bioelectron., 21, 2248. 220. S. Pe´rez-Rial, I. Rodrı´ guez, J. Gonza´lez, J.M. Pe´rez-Sa´nchez F. Herranz, N. Beckmann and J. Ruı´ z-Cabello (2007). J. Pharm. Sci. (in press). 221. I. Hilger, R. Trost, J.R. Reichenbach, W. Linß, M.-R. Lisy, A. Berndt and W.A. Kaiser (2007). Nanotechnology, 18, 135103. 222. K. Mu¨ller, J.N. Skepper, M. Posfai, R. Trivedi, S. Howarth, C. Corot, E. Lancelot, P.W. Thompson, A.P. Brown and J.H. Gillard (2007). Biomaterials, 28, 1629. 223. C. Corot, P. Robert, J.-M. Ide´e and M. Port (2006). Adv. Drug Deliv. Rev., 58, 1471. 224. J.-H. Lee, Y.-M. Huh, Y.-W. Jun, J.-W. Seo, J.-T. Jang, H.-T. Song, S. Kim, E.-J. Cho, H.-G. Yoon, J.-S. Suh and J. Cheon (2007). Nat. Med., 13, 95. 225. W.J.M. Mulder, R. Koole, R.J. Brandwijk, G. Storm, P.T.K. Chin, G.J. Strijkers, C. De Mello Donega´, K. Nicolay and A.W. Griffioen (2006). Nano Lett., 6, 1. 226. C.-W. Lu, Y. Hung, J.-K. Hsiao, M. Yao, T.-H. Chung, Y.-S. Lin, S.-H. Wu and Y.-C. Chen (2007). Nano Lett., 7, 149. 227. W.J. Rogers, C.H. Meyer and C.M. Kramer (2006). Nat. Clin. Pract. Cardiovasc. Med., 3, 554. 228. C. Wilhelm, L. Bal, P. Smirnov, I. Galy-Fauroux, O. Cle´ment and F. Gazeau (2007). J. Emmerich, Biomater., 28, 3797. 229. C. Wilhelm, F. Gazeau and J.-C. Bacri (2002). Eur. Biophys. J., 31, 118. 230. T.R. Pisanic II, J.D. Blackwell, V.I. Shubayev, R.R. Fin˜ones and S. Jin (2007). Biomaterials, 28, 2572. 231. D.E. Sosnovik, M. Nahrendorf, N. Deliolanis, M. Novikov, E. Aikawa, L. Josephson, A. Rosenzweig, R. Weissleder and V.Ntziachristos (2007). Circulation, 115, 1384. 232. A. Punnoose, K.M. Reddy, A. Thurber, J. Hays and M.H. Engelhard (2007). Nanotechnology, 18, art. no. 165502. 233. H. Guerrero, G. Rosa, M.P. Morales, F. Del Monte, E.M. Moreno, D. Levy, R. Pe´rez del Real, T. Belenguer and C.J. Serna (1997). Appl. Phys. Lett., 71, 2698. 234. M. Zayat, F. del Monte, M.P. Morales, G. Rosa, H. Guerrero, C.J. Serna and D. Levy (2003). Adv. Mater., 15, 1809. 235. A. Punnoose, K.M. Reddy, A. Thurber, J. Hays and M.H. Engelhard (2007). Nanotechnology, 18, 165502. 236. I. Enculescu, M.E. Toimil-Molares, C. Zet, M. Daub, L. Westerberg, R. Neumann and R. Spohr (2007). Appl. Phys. A, 86, 43. 237. A. Anguelouch, D.H. Reich, C.L. Chien and M. Tondra (2004). IEEE Trans. Magn., 40, 2997. 238. P.D. McGary, L.W. Tan, J. Zou, B.J.H. Stadler, P.R. Downey and A.B. Flatau (2006). J. Appl. Phys., 99, 08B310. 239. G. Yang, J. Tang, S. Kato, Q. Zhang, L.C. Qin, M. Woodson, L. Liu, J.W. Kim, P.T. Lillehei, C. Park and O. Zhou (2005). Appl. Phys. Lett., 87, 123507. 240. A. Winkler, T. Muhl, S. Menzel, R. Kozhuharova-Koseva, S. Hampel, A. Leonhardt and B. Buchner (2006). J. Appl. Phys., 99, 104905. 241. J.A. Sidles, J.L. Garbini and G.P. Drobny (1992). Rev. Sci. Instrum., 63, 3881. 242. S.R. Garner, S. Kuehn, J.M. Dawlaty, N.E. Jenkins and J.A. Marohn (2004). Appl. Phys. Lett., 84, 5091. 243. D. Rugar, O. Zueger, S. Hoen, C.S. Yannoni, H.-M. Vieth and R.D. Kendrick (1994). Science, 264, 1560. 244. H.J. Mamin, M. Poggio, C.L. Degen and D. Rugar (2007). Nat. Nanotechnol., 2, 301. 245. D. Rugar, R. Budakian, H.J. Mamin and B.W. Chui (2004). Nature 430, 329. 246. M. Barbic (2005). Nano Lett. 5, 187. 247. H.J. Mamin, M. Poggio, C.L. Degen and D. Rugar (2007). Nat. Nanotechnol., 2, 301.

252

M. Va´zquez et al.

248. C. Thirion, W. Wernsdorfer and D. Mailly (2003). Nat. Mater., 2, 524. 249. C.C. Tsuei, J.R. Kirtley, C.C. Chi, L.S. Yujahnes, A. Gupta, T. Shaw, J.Z. Sun and M.B. Ketchen (1994). Phys. Rev. Lett. 73, 593. 250. Y. Ono, A. Ishiyama, N. Kasai and A. Odawara, IEEE Trans. Appl. Super., 15, 640. 251. A. Sandhu, A. Okamoto, I. Shibasaki and A. Oral (2005). Microel. Eng. 73, 524 (2004). 252. A. Candini, G. C. Gazzadi, A. Di Bona, M. Affronte, D. Ercolani, G. Biasiol and L. Sorba (2007). J. Magn. Magn. Mater. 310, 2752. 253. G. Boero, I. Utke, T. Bret, N. Quack, M. Todorova, S. Mouaziz, P. Kejik, J. Brugger, R.S. Popovic and P. Hoffmann (2005). Appl. Phys. Lett., 86, 42503. 254. S. Mouaziz, G. Boero, R.S. Popovic and J. Brugger (2006). J. Microelectromech. Syst. 15, 4. 255. A.N. Grigorenko and S.J. Bending G.D. Howells and R.G. Humphreys (2000). Phys. Rev. B, 62, 721. 256. M. Oszaldowski and T. Berus (2006). Thin Solid Films, 515, 2692. 257. T. Berus, M. Oszwaldowski and J. Grabowski (2004). Sens. Actuat. A, 116, 75; M. Oszwaldowski (1998). Sens. Act. A, 68, 234. 258. www.aacg.bham.ac.uk/magnetic_materials/magnetic_recording.htm. 259. K. Kirk (2000). Contemp. Phys., 41, 61. 260. Hitachi Global Storage Technologies web page. 261. D. Weller and T. McDaniel (2006). ‘‘Media for extremely high density recording’’ in ‘‘Advanced Magnetic Nanostructures’’ Ed. D. Sellmyer and R. Skomski, Springer 295. 262. J.C.S. Kools, K. Rook, H. Hegde, S.B. Sant, J. Wong, W. Xiong, B. Druz, A. Lam, A. Devayasaham and I. Wagner (2000). Thin Solid Films, 377, 705. 263. M. Suzuki, T. Ohwaki and Y. Taga (1997). Thin Solid Films, 304, 333. 264. W.H. Meiklejohn and C.P. Bean (1957). Phys. Rev., 105, 904. 265. J. Nogue´s and I.K. Schuller (1999). J. Magn. Magn. Mater., 192, 203. 266. K. Li, Y. Wu, G. Han, J. Qiu, Y. Zheng, Z. Guo, L. An and P. Luo (2006). Thin Solid Films, 505, 22. 267. H. R. Liu, B. J. Qu, T. L. Ren, L. T. Liu, H. L. Xie, C. X. Li and W. J. Ku (2003), J. Magn. Magn. Mater. 267, 386. 268. T. Miyazaki and N. Tezuka (1995). J. Magn. Magn. Mater., 139, L231. 269. S. Tehrani, B. Engel, J. M. Slaughter, E. Chen, M. de Herrera, M. Durlam, P. Naji, R. Whig, J. Janesky and J. Calder (2000). IEEE Trans. Magn. 36, 2752; E. Nakashio, J. Sugawara, S. Onoc and S. Kumagai (2001). J. Appl. Phys. 89, 7356. 270. S.S.P. Parkin, C. Kaiser and A. Panchula (2004). Nat. Mater. 3, 862. 271. H.X. Wei, T.X. Wang, Z.M. Zeng, X.Q. Zang, J. Zhao and X.F. Han (2006). J. Magn. Magn. Mater. 303, e208. 272. S.A. Wolf, D.D. Awschalom, R.A. Buhrman, J.M. Daughton, S. von Molna´r, M.L. Roukes, A.Y. Chtchelkanova and D. M. Treger (2001). Science, 294, 1488. 273. J.S. Moodera, L.R. Kinder, T.M. Wong and R. Meservey (1995). Phys. Rev. Lett., 74, 3273. 274. J.M. George, L.G. Pereira, A. Barthe´le´my, F. Petroff, L. Steren, J.L. Duvail, and A. Fert, R. Loloee, P. Holody and P.A. Schroeder (1994). Phys. Rev. Lett., 72, 408. 275. S. Sahoo, T. Kontos, JU RG Furer, C. Hoffmann, M. Gra¨ber, A. Cottet and C. Schonenberger (2005). Nature Physics, 1, 99. ¨ 276. E. Mirowski, J. Moreland, S. Russek, M. Donahue and K. Hsieh (2007). J. Magn. Magn. Mater., 311, 401. 277. H. Hauser and M. Evanzin (2006). Sens. Actuat. A: Phys., 129, 231. 278. P. Ripka and G. Ve´rtesy (2000). J. Magn. Magn. Mat., 795 (215–216). 279. B. Sepu´lveda, A. Calle, L.M. Lechuga and G. Armelles (2006). Opt. Lett. 31, 1085. 280. S.L. Sensia at http://www.sensia.es. 281. B. Sepu´lveda, G. Armelles and L.M. Lechuga (2007). Sens. Actuat. A, 134, 339. 282. I. Sandu, L. Presmanes, P. Alphonse and P. Tailhades (2006). Thin Solid Films, 495, 130. 283. K. Tiefenthaler and W. Lukosz (1985). Thin Solid Films, 126, 205.

Chapter 8

Encapsulated Probes Michael J. McShane

8.1 Introduction and Rationale The term ‘‘encapsulated probes’’ refers to a class of molecules or multimolecular cocktails that are responsive to their environment – typically to a certain target analyte found in that environment – and have been somehow physically encased within a protective package. These ‘‘probes’’ are often chemical assays that produce an optical change (absorbance or fluorescence) in proportion to the concentration of their target, and the encapsulation matrix serves as a means to protect them from interfering materials in the surrounding environment or to protect the surrounding environment from the probes themselves. In the former case, the matrix must be engineered to allow rapid penetration of the target while excluding interfering species. In the latter case, the probes may contain toxic materials that must be sequestered away from living systems to avoid unintended damage.

8.2 Brief Overview of Optical Probes Three primary classes of probes (receptors) will be considered for this discussion: direct, competitive, and indirect (Fig. 8.1). Direct probes are those that respond directly to analyte through a physicochemical interaction between the target molecule and a probe molecule or material, producing a change in fluorescence or absorption magnitude or wavelength. A large array of ‘‘molecular probes’’ have been developed and commercialized, including molecules responsive to ions, reactive oxygen species, and molecular oxygen, as well as those that selectively partition into and report from lipids and membranes, organelles, and ion channels. Examples of encapsulated probes have recently surfaced as (or within) retail products, including fluorescent polymer beads [1] and fiber optic oxygen sensors [2]. Other examples of research on developing direct probes have M.J. McShane Biomedical Engineering Department, Texas A&M University, College Station, TX 77843, USA

F.J. Arregui (ed.), Sensors Based on Nanostructured Materials, DOI 10.1007/978-0-387-77753-5_8, Ó Springer ScienceþBusiness Media, LLC 2009

253

254

M.J. McShane

Fig. 8.1 Illustration of generic receptor types (R, E) and typical interactions with analyte (A). For the indirect case, the receptor is an enzyme (E) that catalyzes the conversion of analyte and a co-substrate (C) to products (P1, P1). The concentration of a co-substrate or one of the products that is then directly transduced as an indirect measure of the analyte

been reported for sugars, including both synthetic and engineered protein probes [3, 4, 5, 6, 7, 8, 9, 10]. Competitive assays rely on competition between the target and an analogous ligand for binding sites on the probe; thus, these assays are bimolecular and require free dissociation of target and ligand from the probe according to binding kinetics [8, 11, 12]. Transduction in competitive systems generally involves a change in the optical signal due to dissociation of a labeled ligand from the labeled receptor. A common approach to this is based on resonance energy transfer (RET) [8, 13, 14], which exhibits a distance-dependent efficiency in transfer between two molecules with complementary spectral properties and proper dipole alignments [15, 16]. To satisfy the conditions for RET, the emission spectrum of one fluorophore, termed the ‘‘donor’’, must overlap with the excitation spectrum of the second fluorophore, termed the ‘‘acceptor’’, such that the donor-excited state is nonradiatively transferred to the acceptor, resulting in photon emission. RET efficiency (E) is highly dependent  on the
distance (r) between the donor and the acceptor, following E ¼ R60 R60 þ r6 , where R0 represents the Forster distance, the distance at which 50% efficiency is obtained ¨ (typically in the range of 50A) [15]. RET-based transduction in competitive probes relies simply on the decreasing energy transfer between the fluorescent

8 Encapsulated Probes

255

competitive ligand and the labeled receptor, as the ligand is displaced from the receptor when in the presence of the target analyte [17, 18]. It is noteworthy that sensors based on competitive binding are research-level systems that have not found broad use outside the labs of the developers. A number of successful examples for glucose sensing have been demonstrated based on this concept. Finally, indirect probes rely on a cascade of effects to transduce the interactions; typically, a probe for one parameter (e.g., pH) is used to respond to pH changes induced by a chemical reaction between a second probe molecule (e.g., enzyme) and the target (e.g., the enzymatic substrate). This approach is useful when sufficiently selective and sensitive direct probes are unavailable. An example that illustrates the progression from protein-based molecular recognition agents to encapsulated probes can be seen in the development of fluorescence-based glucose sensors. In the late 1980s, it was reported that glucose-binding events triggered measurable changes in the intrinsic flavine fluorescence of glucose oxidase (GOx) from Aspergillus niger [19]; as a result, it was proposed that the enzyme may be used as a ‘‘probe’’ in a quantitative assay for glucose. By labeling the enzyme with a fluorescein derivative, this intrinsic fluorescent response was later converted to a single-molecule RET system, with energy transfer occurring between the flavine group and fluorescein derivative; the fluorescence spectrum was observed to shift in proportion to glucose concentration [20]. Later studies investigated reagentless sensors using the deactivated form of the enzyme (apoenzyme) [21], where apo-GOx was labeled with the environmentally sensitive fluorophore 8-anilino-1-anpthalene sulfonic acid (ANS). This work showed that apo-GOx retained its high selectivity to glucose and in many ways paved the way for advances in the development of a new biosensor genre [22, 23, 24, 25]. Similar work was performed using yeast hexokinase (HEX), which exhibits a decrease in intrinsic tryptophan fluorescence during exposure to glucose [26, 27]. It was later determined that HEX-based transduction schemes are vulnerable to static quenching when exposed to serum, suggesting that a separation from the biological environment – an encapsulation of the probe – is necessary for in vivo applications [28]. Following this notion, HEX was immobilized in a sol–gel matrix, resulting in enhanced quenching resistance while retaining glucose sensitivity [29]. Thus, the encapsulation of the probes met the requirements for sensing applications: the matrix allowed permeation of the target molecules while providing a stable framework for the probe molecules that protected from the environment.

8.3 Toxicity of Probe Materials It is noteworthy that many small molecules and nanomaterials possess significant toxicity [30, 31, 32, 33, 34, 35]. Therefore, packaging must be given due consideration for nanostructured sensors that may be in some form or fashion exposed to living systems. In particular, optical sensors as described above typically

256

M.J. McShane

contain small-molecule fluorescent probes, protein-based receptors, and/or nanophotonic elements such as quantum dots or metal nanoparticles. When recognizing the potential risks associated with exposure, and identifying requirements for encapsulation strategies (beyond the functional aspects of the immobilization scheme), it is important to note that the materials used must be sufficiently biocompatible to avoid severe acute or sustained long-term host response to the system for the expected route of exposure. For aerosolized or surface-adherent materials, the respiratory tract, eyes, and skin are potentially affected sites; the gastrointestinal tract could receive exposure to materials ingested via contamination of the food or water supply. For sensors intended for implantation, the site of implantation will determine specific requirements. In general, an acute inflammatory response to the implant is expected due to local tissue trauma during the procedure, though the steady-state response should be limited to minimal fibrous tissue formation, and activation of the immune response by antibodies recognizing foreign material must be avoided. A key point related to nanostructured sensors, however, is that the response to the foreign material is determined by the surface properties [36] and, so long as the contents are not released or degraded over time, a stable interface will be formed between the implant surface and the host. Thus, toxic sensor components may potentially be employed in systems with high exposure risk, as long as the encapsulation adequately separates the dangerous entities from the living system. The purity of the materials used in sensor fabrication can be a key factor in determining biocompatibility.

8.4 Immobilization Requirements When developing nanostructured sensors based on a combination of carrier materials and exogenous indicator materials such as fluorimetric or colorimetric probes, the primary goals involve entrapment of the assay chemistry in a way that physically restricts the transduction molecules yet allows sufficient transport of the target analyte [37]. Recent research from our group has reported advances in these areas, particularly in co-immobilizing enzymes and dyes or competitive-binding reagents within microcapsules [38, 39, 40, 41, 42, 43, 44 ,45, 46, 47, 48, 49, 50, 51, 52, 53, 54]; while the focus of our investigations has been on noninvasive glucose biosensing via fluorescence spectroscopy, the findings are more generally applicable to encapsulation of probes in nanostructured materials. In this context, it is especially noteworthy that use of certain transduction schemes requires dramatically different and more complex encapsulation strategies. Take, for example, competitive-binding reactions, which are common to many bioassays. These require free movement of two molecules, which must be able to associate and dissociate quickly in accordance with binding kinetics and the relative affinity to the target analyte; and this must be maintained within a restricted and protected interaction volume.

8 Encapsulated Probes

257

The permeability of the encapsulating matrix is also a key parameter influencing the performance of encapsulated probes. In direct and competitive systems, permeability to the analyte affects response time; in addition, the partitioning of the analyte into the matrix can affect the relative forward and reverse response times. In indirect sensing schemes, however, mass transport considerations can become much more crucial in controlling both the response time and the magnitude of the signal generated; in fact, the type of encapsulating materials can ultimately dictate whether or not a particular design is successful. A prime example of this can be seen for enzyme-based glucose sensors, which operate on the principle of glucose-limited oxygen consumption within the local environment of the sensor. As glucose diffuses into the sensor, oxygen is consumed and rapidly replaced, allowing indirect glucose monitoring through an oxygen reporter. To achieve this fundamental condition, the relative diffusion rate of oxygen into the sensor must be greater than that of glucose; therefore, the local oxygen levels within the sensor are in excess and the sensor is operating in a glucose-limited regime. As bulk glucose levels are increased, the rate of glucose delivery into the sensor begins to eclipse the rate of oxygen replenishment, preventing the relative recovery of oxygen levels within the sensor. In this state of operation, the sensor response begins to deviate from linearity, thus implicating the onset of oxygen-limited glucose consumption. Further increases in glucose levels elicit no sensor response, due to complete internal depletion of oxygen levels (oxygen is readily consumed upon diffusion into the sensor). This condition is elicited by the surplus of local glucose within the sensor and signifies absolute oxygen-limited catalysis and response saturation. Therefore, the mass transport properties of the encapsulating materials effectively control the operational range of enzymatic-based encapsulated probes.

8.5 Encapsulation Strategies A number of different approaches to encapsulation of functional probes have been developed, ranging from ‘‘brute force’’ methods relying on filling of prefabricated containers to self-assembled, spontaneously loaded micro/nanocapsules produced and filled in situ. An example of the former, demonstrated by pioneering work in translating wet reagents into reusable semi-solid-state components, involved the filling of hollow fiber dialysis membranes [11, 55, 56]. These relatively large structures are ideal for interfacing with the tip of optical fibers, but are disadvantageous for reasons of susceptibility to fouling and poor mechanical stability. Standalone materials such as polymeric microspheres are preferable due to the large surface area to volume ratio, which maximizes transport, and low absolute area, which enhances mechanical integrity. Polymer matrices are good encapsulation solutions when the entrapped molecules need not be mobile; in contrast, liposomes, polymersomes, and other microcapsules are superior encapsulation of competitive assays when free association and dissociation of

258

M.J. McShane

reagents and target are required. Furthermore, nanocomposite materials are often preferred, as the precise control over transport properties offers solution to a number of existing problems in creating nanostructured sensors based on encapsulated optical sensing probes. The following sections will focus on encapsulation strategies based on hollow vesicles and nanoengineered composites that enable tunable sensor responses.

8.6 Progress and Opportunities for Encapsulated Probes 8.6.1 Liposomes Liposomes are phospholipid vesicles that contain a hydrophilic interior surrounded by a protective outer phospholipid shell (Fig. 8.2); many types have been prepared from a wide array of natural and synthetic materials. These nanostructured systems have been considered for applications such as drug delivery and bioreactors as well as biosensors, and are particularly attractive for biomedical applications because of their chemical similarity to the cell membrane. For sensors, the aqueous interior of the liposome provides a hydrophilic environment that is preferred for many probes and ensures high permeability for hydrophilic species. To achieve sufficient stability for use in sensing applications, however, phospholipid vesicles typically must be crosslinked through use of polymerizable phospholipids, incorporation of hydrophilic polymers, or polymerization of monomer units incorporated into the bilayer. Encapsulation of fluorescent probes with liposomes has led to demonstration of sensors for molecular oxygen [57], acetylthiocholine chloride [58], calcium [59], and pH [60]. It is noteworthy that phospholipid bilayers are attractive as biocompatible coatings, and therefore have been studied as potential solutions to overcome immunogenicity or toxicity issues for other materials used to encapsulate optical probes. As an example, fluorescent microspheres intended as intracellular monitors for pH, O2, and Cl– have been coated with phospholipids to increase biocompatibility, providing a protective barrier between the probe chemistry and the cell. It is also noteworthy that polymersomes – synthetic vesicles comprising self-assembling block copolymers – are newer materials that

Fig. 8.2 Schematic of enzyme and probe encapsulation within liposomes. (Reproduced with permission) [58]

8 Encapsulated Probes

259

are similar in structure to liposomes and are generally more versatile [61, 62, 63, 64]; they have not yet been used for sensing purposes, but are useful for encapsulation and therefore may impact sensor architectures in the near future.

8.6.2 PEBBLEs (Probes Encapsulated by Biologically Localized Embedding) PEBBLEs are a class of polymeric or organically modified silica particles (20–500 nm) originally developed as a means of protecting sensing chemistry from the ‘‘harsh’’ internal environment of the intracellular space (Fig. 8.3) [65, 66]. Sensing PEBBLEs have been demonstrated for many relevant analytes, including molecular oxygen [67] as well as univalent (K+) and divalent (Ca2+, Zn2+, Mg2+) cations. A key advantage shown in this work was the ability to assemble sensors targeting ‘‘difficult’’ analytes by co-localizing multiple chemistries (e.g., probe + ionophore) within the confined volume of the sensor. Nanoscale glucose-sensing particles have also been reported, utilizing a combination of GOx and O2 indicators [68].

8.6.3 Polyelectrolyte Multilayers Layer-by-layer self-assembly (LbL) of multilayer films [69] has been demonstrated to be a practical and versatile approach to surface modification with nanocomposite materials [70]. LbL has recently been reviewed extensively [71, 72], and therefore will only be introduced here in the context of constructing sensors. The LbL process is general, relying primarily on the attractive force between oppositely charged molecules; following polyion adsorption to an oppositely charged surface, the terminal charge is reversed after every

Fig. 8.3 Various embodiments of PEBBLE sensors. (Reproduced with permission from [37])

260

M.J. McShane

Fig. 8.4 Schematic illustration of layer-by-layer nanofabrication process, applied to (top) planar surfaces and (bottom) colloidal templates. (Reproduced with permission from [71])

subsequent layer deposition (see Fig. 8.4). The resulting films have a thickness on the order of a few nanometers per layer deposited, with the exact thickness depending on the adsorption material and reaction parameters such as pH and ionic strength. Materials employed in the assembly process cover a wide variety of synthetic and natural materials [73, 74], typically charged polymers (‘‘polyelectrolytes’’ or ‘‘polyions’’) or proteins, but also including small molecules such as dyes and even inorganic nanoparticles [75, 76, 77]. In addition to assembly of films on bulk planar templates (1D) or cylindrical (2D) templates, the LbL procedure has also been employed for the modification of three-dimensional surfaces by using charged substrates with micrometer and nanometer dimensions [78, 79, 80]. Specifically, objects such as microspheres, nanoparticles, nanotubes, and platelets [81, 82] have been extensively studied because of the wide availability and attractive surface properties of spherical particles. Functional nanocomposite films containing encapsulated probes can be deposited on these carriers, which are attractive for controlling

8 Encapsulated Probes

261

biocatalytic reactions and other interfacial phenomena due to the high surfaceto-volume ratio [40, 80, 83, 84, 85, 86, 87]. Relating to the focus of this chapter, a number of different LbL structures containing sensing probes embedded within a multilayer have been reported, and several different processes for encapsulation have been employed, as illustrated in Fig. 8.5. It is also noteworthy that sensing films can be prototyped by constructing LbL films with sensing materials on macrotemplates and then, once the desired

Fig. 8.5 Schemes for encapsulation of optical molecular probes in multilayer films: (top) direct electrostatic assembly of anionic (left) and cationic (right) indicators; (middle) covalent attachment of indicators to charged polymers, followed by incorporation via electrostatic adsorption; and (bottom) post-assembly indicator immobilized by electrostatics (left) and precipitation (right)

262

M.J. McShane

Fig. 8.6 Representative templates used for LbL assembly of fluorescent sensing films: (a) glass slides, (b) optical fibers, and (c) latex microspheres. (Reproduced from [38])

(a)

(b)

(c)

behavior has been achieved, the same process parameters may be used to deposit the same materials onto micro/nanotemplates with confidence that very similar sensing behavior will be observed [38]. This has been demonstrated using glass slides, optical fibers, and nanoparticles. These findings point to the versatility of the approach, as well as its ‘‘portability’’ to the point that functional films can be assembled on nearly any surface using the same process (Fig. 8.6).

8.6.4 Multilayer Capsules In the past decade, the multilayer assembly process applied to colloids has been used as the first step in a technique for fabrication of hollow micro/nanocapsules [88]. Following deposition of multilayer coatings onto the templates, the core material is removed using organic solvents or chemical etching to arrive at hollow capsules (Fig. 8.7) [78, 89, 90, 91]. The versatility in construction of these tiny capsules and control over their properties make them attractive for use in sensor applications, especially those in which encapsulation of active molecules and control over transport properties are critical to proper function. Microcapsules based on the layer-by-layer assembly technique have been demonstrated as effective systems for encapsulation in many different potential applications, with a heavy emphasis on drug loading and delivery [92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107]. Early investigations of the properties of polymer capsules revealed that permeability depends directly on the wall composition, and some architectures allow for dynamic control over permeability to open and close pores [108, 109, 110, 111]. For example, pHinduced formation of pores ( 100 nm) was observed for a combination of

Fig. 8.7 Illustration of multilayer capsule fabrication process using sacrificial templates to create hollow microcapsules. (Reproduced with permission from [71])

8 Encapsulated Probes

263

strong and weak polyelectrolytes [112]. Furthermore, changes in solvents have also been shown to induce encapsulation [113, 114]. A more recent development has demonstrated permanent changes in capsule walls’ permeability by employing a photosensitive cationic resin (diazoresin, DAR) [115, 116, 117] to enable photo-induced crosslinking via UV irradiation; such capsules can be used providing permanent and stable encapsulation of macromolecules that permeate the capsules prior to photocrosslinking [48]. In principle, similar strategies based on chemical, thermal, or other crosslinking strategies may also be employed [106, 108, 109, 118–134], though it is desirable to limit crosslinking to the walls and avoid undesired crosslinking of assay components; hence, the DAR-based approach is advantageous. Based on these findings, it has been proposed that such capsules have attractive properties for sensing applications, where reagent chemistry may be packaged within the capsule interior. A number of embodiments for chemical sensing and biosensing can be envisioned, as illustrated in Fig. 8.8. This approach, involving entrapment of assay molecules in a hollow capsule, is particularly attractive for competitive assays, which require mobility of competitive ligands when they are displaced from the receptor. A prime example of this can be seen in the recent literature on glucose-sensing ‘‘smart tattoo’’ systems [135], which can be seen as generic prototypes for similar chemical using optical probes encapsulated in multilayer capsules.

Polyanion Polycation

Catalytic Biosensor

Dye-filled Capsules= Chemical Sensors Reference Dye Indicator

Competitive-binding Biosensor

Enzyme 5µm

Labeled Protein Labeled Ligand

FITC

Transmission

TRITC

Overlay

Fig. 8.8 General description of microcapsule-based sensors. Top left: polyelectrolyte capsule, hollow or filled with polymer matrix; Top right: capsule filled with indicator and reference dye for ratiometric monitoring of a target analyte; Bottom left: enzyme-based microcapsule sensor containing enzyme and dyes; Bottom right: microcapsule sensor employing competitivebinding FRET assay, comprising glucose-binding protein labeled with acceptor and glucose analog labeled with donor. (Reproduced with permission from [52])

264

M.J. McShane

Following demonstration of glucose sensitivity in solution phase, RET-based competitive assay components (apo-glucose oxidase and dextran, labeled with complementary donor–acceptor pair) were encapsulated within a hollow polymeric shell [44, 51]. These polyelectrolyte microcapsules, templated on 5 mm MnCO3 cores with walls comprising multilayer nanofilms including a photosensitive diazoresin, were suspended in a solution of the RET assay and then irradiated with UV light. A key advantage of this approach is the absence of polymer or other matrix in the capsule interior, maintaining a nonrestrictive environment for the molecular components of the competitive-binding assay to easily move relative to one another. The hollow shells act in a similar way to a dialysis membrane, as a semipermeable barrier that allows glucose to diffuse in and out while retaining the sensing components inside. This expected behavior was verified via fluorescence spectroscopy, whereby intensity ratio measurements performed at varying glucose levels revealed a completely reversible response to glucose, matching the observations for the solution-phase assay [44, 51].

8.6.5 Enzymatic Sensors In addition to the obvious encapsulation possibilities afforded by capsules constructed via nanoassembly methods, multilayer nanofilms have also been used to directly entrap sensing probes, or as designed transport-controlling materials deposited on top of prefabricated sensing materials with encapsulated probes. Layer-by-layer construction of enzyme multilayers has been demonstrated for many different catalytic proteins with relevance to sensors; for example, multilayer enzyme nanoreactors can be coupled with multilayers entrapping chemical nanosensors for pH, oxygen, etc., resulting in an exquisitely simple stratified film comprising reactive and monitoring regions. Alternatively, the sensing chemistry can be integrated within the matrix of the colloidal particles, and multilayers deposited on the surface aid in stability, interface interactions with the environment, and transport control. For example, enzymes and fluorescent probes for pH have been combined in calciumcrosslinked alginate microspheres via emulsion and in organically modified silica (ormosil) particles via charge-enhanced absorption (described below). In these cases, a nanofilm coating is easy to apply, serving as a simple and effective barrier to loss of encapsulated macromolecules as well as a diffusion barrier to enable control of relative flux rates for enzymatic substrates [47]. A number of different material combinations are being compared to assess potential for stable immobilization of active enzyme as well as control of relative diffusion of enzymatic substrates; glucose and oxygen have been studied extensively due to the interest in developing glucose-sensing particles [45, 46]. Diffusion-loaded capsules and emulsion-based systems have advantages in stability and ease of use and appear to be the most viable solutions for encapsulation of competitive assays. However, these approaches have limitations

8 Encapsulated Probes

265

in the concentrations of encapsulated materials that can be achieved. Higher encapsulation concentration is more critical to construction of enzyme-based systems, where high enzyme concentrations are required to maintain diffusionlimited behavior. Furthermore, more loaded enzyme allows longer operational lifetimes as the enzyme becomes deactivated through spontaneous or other pathways. One possibility to increase loading capacity is the use of electrostatic interactions between the matrix of a particle and the desired loading material; this has been successfully demonstrated to result in high concentrations of oppositely charged molecules from the surrounding solution. These examples constitute only a limited view of many possibilities for microcapsule-based sensor construction; they show promise for building stable systems with entrapped glucose-sensing chemistry and therefore provide sufficient basis for discussion of the different sensing systems that can be achieved using them. Enzymatic sensing typically relies upon monitoring either a product or cosubstrate of a specific reaction between the enzyme and the target. A common example is the oxidation of glucose driven by glucose oxidase, described as Gox

follows: glucose þ O2 þ H2 O ! gluconic acid þ H2 O2 [136]. This reaction is discussed in more detail in a subsequent section; here, it is sufficient to note that oxygen is a co-substrate with glucose, and both hydrogen peroxide and gluconic acid are produced. Fluorescence monitoring of oxygen, discussed below, is typically preferred, as it provides an indirect measure of glucose level. Alternatively, pH or peroxide could similarly be used. Drawbacks specific to using enzymes include (1) changes in activity over time, leading to drifting calibration curves; (2) dependence on local oxygen levels; and (3) consumption of analyte and co-substrates, accompanied by production of by-products. Even if these issues can be overcome, simply creating a system where diffusion and reaction are sufficiently balanced for a sensitive response is a difficult task. Because layer-by-layer self-assembly allows deposition of ultrathin polyelectrolyte multilayer films on the surface of colloidal templates, such a nanofilm coating may be used to perform several functions: (1) provide a diffusion barrier to inhibit leaching of encapsulated material out of the spheres, (2) provide a transport barrier to slow inward diffusion of substrates, allowing control over the response of the sensor, and (3) introduce an internal intensity reference complementary to the oxygen-sensitive fluorophore by the use of polyelectrolyte–fluorophore conjugates, allowing ratiometric measurements. Figure 8.8 contains an illustration and image of a prototype smart tattoo microsphere using this approach; this is further elaborated in the following sections. Oxygen is one of the best known collisional quenchers of fluorescence; therefore, many fluorophores exhibit, to some degree, oxygen sensitivity. In collisional quenching, the quencher contacts the fluorophores while in the excited state, returning the fluorophore to ground state without photon emission [15]. The process of collisional quenching is typically characterized by the Stern–Volmer equation F0 =F ¼ 0 = ¼ 1 þ kq 0 ½Q ¼ 1 þ KD ½Q. In this equation, F0 and F are the fluorescence intensities in the absence and presence of the quencher,  0

266

M.J. McShane

and  are the lifetimes of the fluorophore in the absence and presence of the quencher, kq is the biomolecular quenching constant, and [Q] is the concentration of the quencher (in this case, molecular oxygen). The Stern–Volmer quenching constant is KD and is calculated as the product of kq and  0. Metal porphyrin compounds are highly sensitive to oxygen; two of the more popular examples of these materials are the Pt(II) and Pd(II) octaethylporphines, both of which possess lifetimes on the order of 100 ms, exhibiting excitation peaks at ca. 375, 500, and 535 nm and emission at 640–660 nm [137]. Pt(II) complexes are highly sensitive to low oxygen levels; however, these complexes are poorly soluble in water, making aqueous application difficult. To overcome this difficulty, surface adsorption to a carrier, usually silicon, organic glassy, cellulose derivatives, or fluoropolymers is common practice; however, the immobilization matrix has been shown to significantly affect O2 sensitivity [138, 139], and it is understood that the difference in reported sensitivities for various immobilization media is due to the solubility and diffusivity of oxygen in the host matrix [137]. Thus, the encapsulation of the probe is a key factor in affecting sensor performance. Given this observation, the highest O2 sensitivities with Pt(II) complexes have been reported with Pt(II) porphyrin adsorption onto silicacontaining matrices [67, 137]. It is hypothesized that oxygen molecules adsorb to the surface substrate then rapidly diffuse across the surface; interestingly, silica-containing materials exhibit high oxygen-binding affinities and surface diffusion rates [137, 138]. While the layer-by-layer self-assembly process coupled with other chemical production techniques such as emulsification enables the construction of a wide variety of microspheres from a seemingly infinite selection of materials, creating useful devices that function well at the microscale demands careful consideration of the system of interest. In the case of microsphere glucose sensors employing enzymes to drive a reaction that will be monitored with an oxygen indicator, this requires a balance between reaction (consumption) and diffusion (supply) of the co-substrates glucose and oxygen. This balance must be engineered to arrive at a measurable signal change for the expected glucose concentrations. A mathematical model has been used to predict the expected behavior of a microsphere system containing homogeneously distributed enzyme and a nanoscale polymer coating, using glucose sensing as a model application. The model requires solving a set of coupled partial differential equations that describe the reaction and diffusion kinetics for the system. A detailed explanation of the model is provided elsewhere [140], and the reader is encouraged to review details of these previous theoretical and experimental investigations to fully understand the implications for enzymatic sensing at the microscale. Here, the major conclusions are summarized to describe the expected functional properties of enzyme-encapsulating microspheres with nanofilm coatings. Some key variables for these nanostructured enzymatic sensors are the composition, density, and thickness of the coating. It is well established that the membrane applied to enzyme-based sensors is critical in determining response

8 Encapsulated Probes

267

sensitivity, range, and time [19, 141, 142, 143, 144, 145]. When working at the microscale, practical realization of transport control requires (1) a deposition method amenable to efficient coating and (2) a highly precise means to control substrate transport, either through precise depositions or by having a wide variety of candidate materials. Multilayer films meet both of these criteria. Transport of various molecular species through polyelectrolyte multilayers has been studied, with very interesting results [146, 147, 148, 149, 150, 151]. In one example, glucose diffusivity through multilayer films of only 22–40 nm thickness was decreased by four orders of magnitude when compared to the glucose diffusivity in water [146]. Since multilayers are easily scaled up, it is easy to see that varying nanofilm thickness through simply adjusting the amount of bilayers adsorbed onto the surface enables precise control over molecular flux rates. For the example of the glucose sensor (Fig. 8.9), glucose relative to oxygen is modulated with film thickness, thereby adjusting the average steady-state oxygen levels observed within the sensor and altering the effective sensitivity of the encapsulated probes. More convincingly, excellent agreement between simulated response and experimental response was observed, validating the predictions of the model [152]. It is noteworthy that, while these comments relate to enzymatic sensors in which oxygen was monitored, a similar approach could be taken for microspheres monitoring pH [153], or other local variables, that are changed as a result of reactions occurring within the confined environment of the microsphere/capsule. These theoretical predictions were followed with experimental work to investigate the potential for using nanofilm coatings to control transport into microsphere sensors. Sensor particles were prepared by immobilizing Pt(II) octaethylporphine (PtOEP), a phosphorescent dye readily quenched by molecular oxygen, into hybrid silicate microspheres, followed by loading and subsequent covalent immobilization of glucose oxidase (GOx) and reference dye, rhodamine B (RITC). Multilayer nanofilms were subsequently assembled on

Fig. 8.9 Predicted average oxygen concentration at steady-state within enzymeloaded alginate particles with varying glucose levels and nanofilm thicknesses. (Reproduced from [152])

268

M.J. McShane

the surfaces of the particles to provide critical control of glucose transport into the particle. The final particles were characterized as a population to determine the optical response to changes in glucose concentration. The sensors were observed to respond rapidly (90 s) and were fully reversible. The sensitivity to glucose, defined as the change in intensity ratio per unit concentration over the linear response range, was determined to be 4.16  0.57%/mg dl–1, over a range of 2–120 mg/dl [154]. The ability to controlling sensor response by adjusting the nanofilm properties was further investigated. Specifically, alterations in the surface-adsorbed polyelectrolyte nanofilms were employed to modulate the relative fluxes of glucose and oxygen into the sensor, allowing the analytical range and sensitivity to be tuned. It was found that the nanofilm thickness, ionic strength of assembly conditions, and outermost constituents played key roles in determining substrate flux. In general, increasing film thickness through additional cycles of adsorption resulted in consistently decreased glucose flux, correspondingly decreasing sensitivity and increasing range. It was shown that the sensor response may be customized to cover the hypo- (0–80 mg/dl), normo- (80–120 mg/dl), and hyperglycemic levels (>120 mg/dl) from a single batch of particles through appropriate selection of coating structure and assembly conditions. These findings demonstrate the key influence of the encapsulation system on sensor response.

8.7 Summary and Conclusions This chapter highlighted the key requirements and progress in nanostructured sensors based on encapsulated probes. For the limited examples available, a review has revealed positive outcomes in the use of nanomaterials to entrap probes for stability, protection, and even to control nanoscale diffusion in order to adjust sensor properties. With the proliferation of new materials and nanofabrication methods, it is likely that the future will see more common research into these areas, and development of standard approaches and products based on probes encapsulated in nanostructured materials. Acknowledgments This work was funded in part by NIH (R01 EB000739), NSF (0210298 and 0640037), and the Louisiana Board of Regents (LEQSF(2001-04)-RD-A-18). The majority of the work and illustrations cited are the result of dissertation research conducted by J. Quincy Brown, Rohit Srivastava, Erich Stein, Patrick Grant, Swetha Chinnayelka, Suman Nayak, and Saurabh Singh and the postdoctoral research work of Huiguang Zhu and Jinshu Mao.

References 1. Invitrogen, Inc., www.probes.com. 2. Ocean Optics, Inc., www.oceanoptics.com.

8 Encapsulated Probes

269

3. Cordes, D. B.; Miller, A.; Gamsey, S.; Singaram, B. (2007). Analytical and Bioanalytical Chemistry 387, 2767–2773. 4. Cordes, D. B.; Gamsey, S.; Singaram, B. (2006). Angewandte Chemie – International Edition 45, 3829–3832. 5. Phillips, M. D.; James, T. D. (2004). Journal of Fluorescence 14, 549–559. 6. Hao, F.; Kaur, G.; Binghe, W. (2004). Journal of Fluorescence 14, 481–489. 7. Piletsky, S. A.; Piletskaya, E. V.; Panasyuk, T. L.; El’skaya, A. V.; Levi, R.; Karube, I.; Wulff, G. (1998). Macromolecules 31, 2137–2140. 8. Kataoka, K.; Hisamitsu, I.; Sayama, N.; Okano, T.; Sakurai, Y. (1995). Journal of Biochemistry (Tokyo) 117, 1145–1147. 9. Salins, L. L. E.; Shrestha, S.; Daunert, S. (2000). ACS Symposium Series 762, 87–101. 10. Ye, K.; Schultz, J. S. (2003). Analytical Chemistry 75, 3451–3459. 11. Schultz, J. S.; Mansouri, S.; Goldstein, I. J. (1982). Diabetes Care 5, 245–253. 12. Rolinski, O. J.; Birch, D. J. S.; McCartney, L.; Pickup, J. C. (2001). Spectrochimica Acta, Part A (Molecular and Biomolecular Spectroscopy) 57A, 2245–2254. 13. Meadows, D. L.; Schultz, J. S. (1991). Biotechnology and Bioengineering 37, 1066. 14. Meadows, D. L.; Schultz, J. S. (1993). Analytica Chimica Acta 280, 21–30. 15. Lakowicz, J. R. (1999). Principles of Fluorescence Spectroscopy, 2nd ed.; Kluwer Academic/Plenium, New York. 16. Chinnayelka, S.; McShane, M. J. (2004). Biomacromolecules 5, 1657–1661. 17. Ballerstadt, R.; Polak, A.; Beuhler, A.; Frye, J. (2004). Biosensors and Bioelectronics 19, 905. 18. Ballerstadt, R.; Schultz, J. S. (2000). Analytical Chemistry 72, 4185. 19. Trettnak, W.; Wolfbeis, O. S. (1989). Analytica Chimica Acta 221, 195–203. 20. Sierra, J. F.; Galban, J.; De Marcos, S.; Castillo, J. R. (2000). Analytica Chimica Acta 414, 33–41. 21. D’Auria, S.; Herman, P.; Rossi, M.; Lakowicz, J. R. (1999). Biochemical and Biophysical Research and Communication 263, 550–553. 22. Chinnayelka, S. (2005). Louisiana Tech University. 23. Chinnayelka, S.; McShane, M. J. (2004). Journal of Fluorescence 14, 585. 24. Chinnayelka, S.; McShane, M. J. (2005). Analytical Chemistry 77, 5501–5511. 25. Chinnayelka, S.; McShane, M. J. (2004). Biomacromolecules 5, 1657. 26. Maity, H.; Maiti, N. C.; Jarori, G. K. (2000). Journal of Photochemistry and Photobiology B: Biology 55, 20–26. 27. Maity, H.; Kasturi, S. R. (1998). Journal of Photochemistry and Photobiology B 47, 190–196. 28. Pickup, J. C.; Hussain, F.; Evans, N. D.; Rolinski, O. J.; Birch, D. J. S. (2005). Biosensors and Bioelectronics 20, 2555. 29. Hussain, F.; Birch, D. J. S.; Pickup, J. C. (2005). Analytical Biochemistry 339, 137–143. 30. Kopelman, R.; Bakker, E.; Barker, S.; Clark, H.; Dourado, S.; Parus, S.; Rosenzweig, Z.; Shi, Z. Y.; Song, T.; et al. (1997). Book of Abstracts, 213th ACS National Meeting, San Francisco, April 13–17, ANYL-233. 31. Feron, V. J.; Groten, J. P. (2002). Food and Chemical Toxicology 40, 825–839. 32. Choi, A. O.; Cho, S. J.; Desbarats, J.; Lovric, J.; Maysinger, D. (2007). J Nanobiotechnology 5, 1. 33. Ryman-Rasmussen, J. P.; Riviere, J. E.; Monteiro-Riviere, N. A. (2007). Journal of Investigative Dermatology 127, 143–153. 34. Wang, H. Q.; Zhang, H. L.; Li, X. Q.; Wang, J. H.; Huang, Z. L.; Zhao, Y. D. (2007). Journal of Biomedical Material and Research A 86, 833–841. 35. Zhang, Y.; Chen, W.; Zhang, J.; Liu, J.; Chen, G.; Pope, C. (2007). Journal of Nanoscience and Nanotechnology 7, 497–503. 36. Anderson, J. M. (1996). In Biomaterials Science; Ratner, Hoffman, Schoen, Lemons, Eds.; Academic Press, 1, 165–173.

270

M.J. McShane

37. Clark, H. A.; Barker, S. L. R.; Brasuel, M.; Miller, M. T.; Monson, E.; Parus, S.; Shi, Z. Y.; Song, A.; Thorsrud, B.; Kopelman, R.; Ade, A.; Meixner, W.; Athey, B.; Hoyer, M. ; Hill, D.; Lightle, R.; Philbert, M. A. (1998). Sensors and Actuators B-Chemical 51, 12–16. 38. Grant, P. S.; McShane, M. J. (2003). IEEE Sensors Journal 3, 139. 39. McShane, M. J. (2003). Proceedings of IEEE Sensors 2003. IEEE International Conference on Sensors, 2nd, Toronto, ON, Canada, Oct. 22–24, 1, 689–693. 40. Stein, E. W.; McShane, M. J. (2003). IEEE Transactions on Nanobioscience 2, 133–137. 41. Brown, J. Q.; Chopra, S.; Grant, P. S.; McShane, M. J. (2004). Proceedings of SPIE-The International Society for Optical Engineering 5325, 21–30. 42. Brown, J. Q.; Srivastava, R.; McShane, M. J. (2005). Biosensors and Bioelectronics 21, 212–216. 43. Guice, K. B.; Caldorera, M. E.; McShane, M. J. (2005). Journal for Biomedical Optimal 10, 064031. 44. Chinnayelka, S.; McShane, M. J. (2005). Analytical Chemistry 77, 5501–5511. 45. Srivastava, R.; Brown, J. Q.; Zhu, H.; McShane, M. J. (2005). Biotechnology and Bioengineering 91, 124–131. 46. Srivastava, R.; Brown, J. Q.; Zhu, H.; McShane, M. J. (2005). Macromolecules Bioscience 5, 717–727. 47. Srivastava, R.; McShane, M. J. (2005). Journal of Microencapsulation 22, 397–411. 48. Zhu, H.; McShane, M. J. (2005). Langmuir 21, 424–430. 49. Zhu, H.; Srivastava, R.; Brown, J. Q.; McShane, M. J. (2005). Bioconjugation Chemistry 16, 1451–1458. 50. Zhu, H.; Srivastava, R.; McShane, M. J. (2005). Biomacromolecules 6, 2221–2228. 51. Chinnayelka, S.; McShane, M. J. (2006). Diabetes Technology and Therapeutics 8, 269–278. 52. McShane, M. J. (2006). In Topics in Fluorescence; Lakowicz, J. R., Geddes, C. D., Eds., 10. 53. Stein, E. W.; Volodkin, D. V.; McShane, M. J.; Sukhorukov, G. B. (2006). Biomacromolecules 7, 710–719. 54. Nayak, S. R.; McShane, M. J. (2007). Journal of Biomedical Nanotechnology 3, 170–177. 55. Meadows, D. L.; Schultz, J. S. (1992). Proceedings of SPIE – The International Society for Optical Engineering, Los Angeles, CA, USA; Publ by Int Soc for Optical Engineering, Bellingham, WA, USA; 202. 56. Ballerstadt, R.; Schultz, J. S. (2000). Analytical Chemistry 72, 4185–4192. 57. Cheng, Z. L.; Aspinwall, C. A. (2006). Analyst 131, 236–243. 58. Vamvakaki, V.; Fournier, D.; Chaniotakis, N. A. (2005). Biosensors and Bioelectronics 21, 384–388. 59. Nguyen, T.; Rosenzweig, Z. (2002). Analytical and Bioanalytical Chemistry 374, 69–74. 60. Nguyen, T.; McNamara, K. P.; Rosenzweig, Z. (1999). Analytica Chimica Acta 400, 45–54. 61. Discher, B. M.; Won, Y. Y.; Ege, D. S.; Lee, J. C. M.; Bates, F. S.; Discher, D. E.; Hammer, D. A. (1999). Science 284, 1143–1146. 62. Discher, B. M.; Bermudez, H.; Hammer, D. A.; Discher, D. E.; Won, Y. Y.; Bates, F. S. (2002). Journal of Physical Chemistry B 106, 2848–2854. 63. Meng, F. H.; Engbers, G. H. M.; Feijen, J. (2005). Journal of Controlled Release 101, 187–198. 64. Yow, H. N.; Routh, A. F. (2006). Soft Matter 2, 940–949. 65. Clark, H. A.; Kopelman, R.; Tjalkens, R.; Philbert, M. A. (1999). Analytical Chemistry 71, 4837–4843. 66. Clark, H. A.; Hoyer, M.; Philbert, M. A.; Kopelman, R. (1999). Analytical Chemistry 71, 4831–4836. 67. Koo, Y.-E. L.; Cao, Y.; Kopelman, R.; Koo, S. M.; Brasuel, M.; Philbert, M. A. (2004). Analytical Chemistry 76, 2498–2505.

8 Encapsulated Probes 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79.

80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91.

92. 93. 94. 95. 96. 97. 98. 99.

271

Xu, H.; Aylott, J. W.; Kopelman, R. (2002).The Analyst 127, 1471–1477. Decher, G. (1997). Science 277, 1232–1237. Decher, G.; Lvov, Y.; Schmitt, J. (1994). Thin Solid Films 244, 772–777. McShane, M. J.; Lvov, Y. (2004). In Dekker Encyclopedia of Nanoscience and Nanotechnology; Schwarz, J. A., Contescu, C. I., Putyera, K., Eds.; Marcel Dekker, 1–20. Decher, G.; Schlenoff, J. B., Eds. (2006). Multilayer Thin Films: Sequential Assembly of Nanocomposite Materials, Wiley-VCH. Keller, S. W.; Johnson, S. A.; Brigham, E. S.; Smith, K. M.; Mallouk, T. E. (1995). Abstracts of Papers of the American Chemical Society 210, 83–Coll. Keller, S. W.; Kim, H. N.; Mallouk, T. E. (1994). Journal of the American Chemical Society 116, 8817–8818. Decher, G.; Lehr, B.; Lowack, K.; Lvov, Y.; Schmitt, J. (1994). Biosensors & Bioelectronics 9, 677–684. Sukhorukov, G. B.; Mohwald, H.; Decher, G.; Lvov, Y. M. (1996). Thin Solid Films 285, 220–223. Lvov, Y.; Ariga, K.; Ichinose, I.; Kunitake, T. (1995). Journal of the American Chemical Society 117, 6117–6123. Sukhorukov, G. B.; Donath, E.; Davis, S.; Lichtenfeld, H.; Caruso, F.; Popov, V. I.; Mohwald, H. (1998). Polymers for Advanced Technologies 9, 759–767. Sukhorukov, G. B.; Donath, E.; Lichtenfeld, H.; Knippel, E.; Knippel, M.; Budde, A.; Mohwald, H. (1998). Colloids and Surfaces A – Physicochemical and Engineering Aspects 137, 253–266. Caruso, F.; Caruso, R. A.; Mohwald, H. (1998). Science 282, 1111–1114. Ai, H.; Fang, M.; Jones, S. A.; Lvov, Y. M. (2002). Biomacromolecules 3, 560–564. Ai, H.; Jones, S. A.; Lvov, Y. M. (2003). Cell Biochemistry and Biophysics 39, 23–43. Caruso, F.; Lichtenfeld, H.; Giersig, M.; Mohwald, H. (1998). Journal of the American Chemical Society 120, 8523–8524. Caruso, F.; Mohwald, H. (1999). Journal of the American Chemical Society 121, 6039–6046. Caruso, F.; Schuler, C. (2000). Langmuir 16, 9595–9603. Lvov, Y.; Caruso, F. (2001). Analytical Chemistry 73, 4212–4217. Fang, M.; Grant, P. S.; McShane, M. J.; Sukhorukov, G. B.; Golub, V. O.; Lvov, Y. M. (2002). Langmuir 18, 6338–6344. Mohwald, H. (2000). Colloids and Surfaces a – Physicochemical and Engineering Aspects 171, 25–31. Sukhorukov, G.; Fery, A.; Mohwald, H. (2005). Progress in Polymer Science 30, 885–897. Sukhorukov, G. B.; Brumen, M.; Donath, E.; Mohwald, H. (1999). Journal of Physical Chemistry B 103, 6434–6440. Sukhorukov, G. B.; Rogach, A. L.; Garstka, M.; Springer, S.; Parak, W. J.; MunozJavier, A.; Kreft, O.; Skirtach, A. G.; Susha, A. S.; Ramaye, Y.; Palankar, R.; Winterhalter, M. (2007). Small 3, 944–955. Caruso, F.; Yang, W. J.; Trau, D.; Renneberg, R. (2000). Langmuir 16, 8932–8936. Qiu, X. P.; Leporatti, S.; Donath, E.; Mohwald, H. (2001). Langmuir 17, 5375–5380. Antipov, A.; Sukhorukov, G.; Moehwald, H. (2002). Abstracts of Papers of the American Chemical Society 223, p. 446. Khopade, A. J.; Caruso, F. (2002). Biomacromolecules 3, 1154–1162. Radtchenko, I. L.; Sukhorukov, G. B.; Mohwald, H. (2002). International Journal of Pharmaceutics 242, 219–223. Ai, H.; Jones, S. A.; de Villiers, M. M.; Lvov, Y. M. (2003). Journal of Controlled Release 86, 59–68. De Geest, B. G.; Dejugnat, C.; Sukhorukov, G. B.; Braeckmans, K.; De Smedt, S. C.; Demeester, J. (2005). Advanced Materials 17, 2357. Liu, X. Y.; Gao, C. Y.; Shen, J. C.; Mohwald, H. (2005). Macromolecular Bioscience 5, 1209–1219.

272

M.J. McShane

100. Pargaonkar, N.; Lvov, Y. M.; Li, N.; Steenekamp, J. H.; de Villiers, M. M. (2005). Pharmaceutical Research, 22, 826–835. 101. Shutava, T. G.; Lvov, Y. M. (2006). Journal of Nanoscience and Nanotechnology 6, 1655–1661. 102. Zelikin, A. N.; Li, Q.; Caruso, F. (2006). Angewandte Chemie-International Edition 45, 7743–7745. 103. De Geest, B. G.; Sanders, N. N.; Sukhorukov, G. B.; Demeester, J.; De Smedt, S. C. (2007). Chemical Society Reviews 36, 636–649. 104. De Geest, B. G.; Skirtach, A. G.; De Beer, T. R. M.; Sukhorukov, G. B.; Bracke, L.; Baeyens, W. R. G.; Demeester, J.; De Smedt, S. C. (2007). Macromolecular Rapid Communications 28, 88–95. 105. De Geest, B. G.; Skirtach, A. G.; Mamedov, A. A.; Antipov, A. A.; Kotov, N. A.; De Smedt, S. C.; Sukhorukov, G. B. (2007). Small 3, 804–808. 106. Kohler, K.; Sukhorukov, G. B. (2007). Advanced Functional Materials, 17, 2053–2061. 107. Sukhorukov, G. B.; Mohwald, H. (2007). Trends in Biotechnology 25, 93–98. 108. Antipov, A. A.; Sukhorukov, G. B. (2004). Advances in Colloid and Interface Science 111, 49–61. 109. Antipov, A. A.; Sukhorukov, G. B.; Leporatti, S.; Radtchenko, I. L.; Donath, E.; Mohwald, H. (2002). Colloids and Surfaces A– Physicochemical and Engineering Aspects 198, 535–541. 110. Antipov, A. A.; Sukhorukov, G. B.; Mohwald, H. (2003). Langmuir 19, 2444–2448. 111. Sukhorukov, G. B.; Donath, E.; Moya, S.; Susha, A. S.; Voigt, A.; Hartmann, J.; Mohwald, H. (2000). Journal of Microencapsulation 17, 177–185. 112. Mendelsohn, J. D.; Barrett, C. J.; Chan, V. V.; Pal, A. J.; Mayes, A. M.; Rubner, M. F. (2000). Langmuir 16, 5017. 113. Moya, S.; Sukhorukov, G. B.; Auch, M.; Donath, E.; Mohwald, H. (1999). Journal of Colloid and Interface Science 216, 297–302. 114. Lvov, Y.; Antipov, A. A.; Mamedov, A.; Mohwald, H.; Sukhorukov, G. B. (2001). Nano Letters 1, 125–128. 115. Wang, J. F.; Jia, X. R.; Zhong, H.; Luo, Y. F.; Zhao, X. S.; Cao, W. X.; Li, M. Q. (2002). Chemistry of Materials 14, 2854–2858. 116. Chen, J. Y.; Huang, L.; Ying, L. M.; Luo, G. B.; Zhao, X. S.; Cao, W. X. (1999). Langmuir 15, 7208–7212. 117. Chen, J. Y.; Luo, H.; Cao, W. X. (2000). Polymer International 49, 382–386. 118. Dai, Z. F.; Dahne, L.; Mohwald, H.; Tiersch, B. (2002). Angewandte Chemie-International Edition 41, 4019–4022. 119. Gao, C. Y.; Moya, S.; Donath, E.; Mohwald, H. (2002). Macromolecular Chemistry and Physics 203, 953–960. 120. Georgieva, R.; Moya, S.; Hin, M.; Mitlohner, R.; Donath, E.; Kiesewetter, H.; Mohwald, H.; Baumler, H. (2002). Biomacromolecules 3, 517–524. 121. Ibarz, G.; Dahne, L.; Donath, E.; Mohwald, H. (2002). Chemistry of Materials 14, 4059–4062. 122. Ibarz, G.; Dahne, L.; Donath, E.; Mohwald, H. (2002). Macromolecular Rapid Communications 23, 474–478. 123. Nayak, S. R.; Stein, E. W.; Gupta, N.; Palmer, J.; McShane, M. J. (2003). Abstracts of Papers, 226th ACS National Meeting, New York, NY, United States, September 7–11, PMSE-204. 124. Petrov, A. I.; Gavryushkin, A. V.; Sukhorukov, G. B. (2003). Journal of Physical Chemistry B 107, 868–875. 125. Glinel, K.; Prevot, M.; Krustev, R.; Sukhorukov, G. B.; Jonas, A. M.; Mohwald, H. (2004). Langmuir 20, 4898–4902. 126. Kohler, K.; Shchukin, D. G.; Sukhorukov, G. B.; Mohwald, H. (2004). Macromolecules 37, 9546–9550.

8 Encapsulated Probes

273

127. Lulevich, V. V.; Vinogradova, O. I. (2004). Langmuir 20, 2874–2878. 128. Shutava, T.; Zheng, Z. G.; John, V.; Lvov, Y. (2004). Biomacromolecules 5, 914–921. 129. Dong, W. F.; Ferri, J. K.; Moehwald, H. (2005). Abstracts of Papers of the American Chemical Society 230, U1231–U1232. 130. Dong, W. F.; Liu, S. Q.; Wan, L.; Mao, G. Z.; Kurth, D. G.; Mohwald, H. (2005). Chemistry of Materials 17, 4992–4999. 131. Dong, W.-F.; Liu, S.; Wan, L.; Mao, G.; Kurth, D. G.; Mohwald, H. (2005). Chemistry of Materials 17, 4992–4999. 132. Kohler, K.; Shchukin, D. G.; Mohwald, H.; Sukhorukov, G. B. (2005). Journal of Physical Chemistry B 109, 18250–18259. 133. Tong, W. J.; Doug, W. F.; Gao, C. Y.; Mohwald, H. (2005). Journal of Physical Chemistry B 109, 13159–13165. 134. Shchukin, D. G.; Gorin, D. A.; Moehwald, H. (2006). Langmuir 22, 7400–7404. 135. McShane, M. J. (2002). Diabetes Technology and Therapeutics 4, 533–538. 136. Wilson, R.; Turner, A. P. F. (1992). Biosensors & Bioelectronics 7, 165–185. 137. Han, B.-H.; Manners, I.; Winnik, M. A. (2005). Chemistry of Materials 17, 3160–3171. 138. Lu, X.; Manners, I.; Winnik, M. A. (2001). Macromolecules 34, 1917–1927. 139. Papkovsky, D. B.; O’Riordan, T. C. (2005). Journal of Fluorescence 15, 569–584. 140. Brown, J. Q.; McShane Michael, J. (2006). Biosensors & Bioelectronics 21, 1760–1769. 141. Wolfbeis, O. S.; Oehme, I.; Papkovskaya, N.; Klimant, I. (2000). Biosensors & Bioelectronics 15, 69–76. 142. Rosenzweig, Z.; Kopelman, R. (1996). Analytical Chemistry 68, 1408–1413. 143. Choi, H. N.; Kim, M. A.; Lee, W.-Y. (2005). Analytica Chimica Acta 537, 179–187. 144. Wang, B.; Li, B.; Deng, Q.; Dong, S. (1998). Analytical Chemistry 70, 3170–3174. 145. Pandey, P. C.; Upadhyay, S.; Pathak, H. C. (1999). Sensors and Actuators, B: Chemical B60, 83–89. 146. Liu, X.; Bruening, M. L. (2004). Chemistry of Materials 16, 351–357. 147. Bruening, M. L.; Sullivan, D. M. (2002). Chemistry – A European Journal 8, 3832–3837. 148. Miller, M. D.; Bruening, M. L. (2004). Langmuir 20, 11545–11551. 149. Stanton, B. W.; Harris, J. J.; Miller, M. D.; Bruening, M. L. (2003). Langmuir 19, 7038–7042. 150. Harris, J. J.; Stair, J. L.; Bruening, M. L. (2000). Chem. Mater. 12, 1941–1946. 151. Miller, M. D.; Bruening, M. L. (2005). Chem. Mater. 17, 5375–5381. 152. Brown, J. Q.; McShane, M. J. (2006). Biosensors and Bioelectronics 21, 1760–1769. 153. Nayak, S. R.; McShane, M. J. (2006). Sensor Letters 4, 433–439. 154. Stein, E. W.; Grant, P. S.; Zhu, H.; McShane, M. J. (2007). Analytical Chemistry 79, 1339–1348.

Chapter 9

Optical Fiber Sensors Based on Nanostructured Coatings Francisco J. Arregui, Ignacio R. Matias, Javier Goicoechea, and Ignacio Del Villar

9.1 Introduction Optical fiber sensors have been developed from the late 1960s, when optical fiber was proposed as a practical medium for communication [1, 2]. Since then, a great effort has been dedicated for the design and development of optical fiber sensors. In fact, the use of this technology to fabricate sensors is very attractive because optical fibers make possible large sensor data capacities over long distances (kilometers). This implies that the sensing head can be very far from the electronic unit that processes the information. In addition to this, the optical fiber is made of dielectric materials which make possible to incorporate these devices in circumstances where high electromagnetic fields are applied, such as in medical magnetic resonance or in situations with high radiation doses [3]. Besides, optical fibers are made of biocompatible materials. Therefore, this technology is very suitable to develop biomedical instrumentation. Other advantages with respect to conventional sensors are that several sensors can be multiplexed in the same optical fiber or even distributed sensing along a fiber cable can be achieved [1, 2, 4]. Basically, a standard communications optical fiber is a cylindrical dielectric waveguide which consists of a core surrounded by a cladding. The fiber can transmit light along its axis due to the phenomenon of total internal reflection, which promotes the confinement of the optical signal in the core. This total internal reflection happens because the refractive index of the core is greater than that of the cladding. From an electromagnetic waveguide perspective it can be analyzed and seen that the light energy in the fiber is not completely confined to the core. Instead, a significant fraction of the energy travels in the cladding as an evanescent wave. Some optical fiber sensors are fabricated by coating the optical fiber core with a sensing cladding. This sensitive cladding is easy to excite by means of the evanescent field [5]. Thus, the interaction between light and matter is governing the sensing mechanisms of these devices. F.J. Arregui Electric and Electronic Engineering Department, Universidad Pu´blica de Navarra, Edificio de los Tejos, Campus Arrosadı´ a, 31006 Pamplona, Navarra, Spain

F.J. Arregui (ed.), Sensors Based on Nanostructured Materials, DOI: 10.1007/978-0-387-77753-5_9, Ó Springer ScienceþBusiness Media, LLC 2009

275

276

F.J. Arregui et al.

Therefore, these sensors are based on physical phenomena like optical absorption, fluorescence or scattering. In fact, the penetration depth of the evanescent field, dp, which is defined as the distance where the evanescent field is reduced to 1/e of its interface value, can be calculated from the solution of Maxwell’s equations [5]: dp ¼

 qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2 2p n1 sin2   n22

(9:1)

In this equation  is the free-space wavelength of incident light, n1 and n2 are the refractive indices of the core and the sensing cladding of the fiber, respectively, and  is the angle of the incident light at the interface between both optical media. Considering that the refractive index of the core, fused silica, is around 1.5 at these wavelengths and that the sensing coatings used have refractive indices similar to the core, for instance 1.49, the penetration depth of the evanescent field is of the order of magnitude of one wavelength. Since the wavelength range of standard optical fibers goes from the visible to the near infrared band, typically until what is called the U band, 1,675 nm, this implies that for devices with these characteristics, sensing coatings with thickness greater than 2 mm could be considered as an infinitum media and the phenomena to be studied would be restricted to this supposition. The assumption of an infinitum medium is valid when the sensing coatings used in optical sensors are fabricated by means of standard procedures that only allow the synthesis and arrangement of sensitive coatings on the micrometer scale. On the contrary, new phenomena have to be considered when the coatings are under the micron scale. Fortunately, with the appearance of techniques that permit the development of nanostructured coatings, it is possible to fabricate devices that take advantage of the new observed phenomena. These nanostructured coatings have opened the door to new sensors that have been studied and presented in the last years. This chapter intends to show a collection of some of them.

9.2 Methods of Fabrication of Nanostructured Films on Optical Fibers: the Layer-by-Layer Technique The fabrication of proper sensing films is a complex discipline that requires a deep multidisciplinary study which has to take into account multiple variables in order to achieve the optimal performances with respect to sensitivity, response time, working range, hysteresis or cross-sensitivity of the sensing devices. In addition to this, if the sensing devices are based on optical fibers, the special geometry of the fiber devices requires also the ability of depositing not only on flat surfaces but on cylindrical or conical substrates as well. Classic deposition techniques such as physical vapor deposition or spin coating are intended usually for flat semiconductor substrates and cannot deposit uniform

9 Optical Fiber Sensors Based on Nanostructured Coatings

277

films easily on complex geometries. In fact, there are only three different techniques that have been successfully used for the deposition of uniform coatings onto the cylindrical, and sometimes conical, shape of optical fibers such as the dip-coating (DC) technique, the Langmuir–Blodgett (LB) technique and the layer-by-layer electrostatic self-assembly (LbL) technique. The DC technique is usually associated to sol-gel or hydrogel coatings and, although suitable for cylindrical shapes, it is not useful for controlling the thickness of the coatings on the nanometer scale. On the contrary, the LB and LbL techniques can be used for the fabrication of nanostructured films. Unfortunately, the LB technique is limited to very specific molecules with combinations of lipophilic and hydrophilic parts [6]. On the other hand, the LbL process has been successfully probed as a useful tool for the fabrication of nanostructured materials that include many diverse species, as was anticipated in Chapter 8, such as colorimetric dyes, fluorescent indicators, inorganic semiconductors, conducting polymers, ceramics, metals, quantum dots, enzymes, antibodies or even DNA strands [7–14]. Since the LbL technique can summarize by itself the different devices fabricated by means of nanostructured sensing films, this chapter will mainly focus on the LbL technique. The subsequent adsorption of oppositely charged colloids, the LbL technique, was suggested for the first time by R. Iler in the mid-1960s [14]. Unfortunately, nobody followed this line of research until almost 40 years later when the technique was rediscovered by G. Decher and coworkers [9], and extended to the layering of polyelectrolytes and many other systems [7–14]. In the last years the number of works on this topic has increased exponentially and some reviews permit to understand the current state of the art [15–17]. Authors have referred to this technique in different ways: ionic self-assembly monolayer (ISAM) process, electrostatic self-assembly (ESA), layer-by-layer process (LbL). Henceforward we will refer to this technique as the LbL method. Basically, the LbL method is based on the electrostatic attraction between oppositely charged polyelectrolytes in each monolayer deposited, and involves several steps. The LbL film deposition method is described schematically in Fig. 9.1. First, a substrate (in this case the optical fiber) is cleaned and treated to create a charged surface. Then, the substrate is exposed to a solution of a polyion of opposite charge for a short time (minutes) and by adsorption a monolayer of polyions is formed on the surface. In this way, the substrate is alternately dipped into solutions of cationic and anionic polymers (or appropriately charged inorganic clusters) to create a multilayer thin film, a polyanion–polycation multilayer. The molecular species of the anionic and the cationic components and the long-range physical order of the layers determine the resulting coating properties. It is important to notice that the polyanions and polycations overlap each other at the molecular level and this produces a homogeneous optical material [18–20]. The pair of one anionic monolayer and one cationic monolayer will be called bilayer henceforward. The composition and thickness of an individual bilayer can be controlled by adjusting the deposition parameters. Moreover, these coatings can be formed in many different substrates; for instance, metals, plastics, ceramics and semiconductors. Additional details

278

F.J. Arregui et al. NEGATIVE

POSITIVE

CHARGED

CHARGED

MONOLAYER

MONOLAYER

CHARGED SUBSTRATE

−

−

−

−

−

−

−

−

− − − − − − − − − − −

−

−

POLYANION DIPPING & WASH

−− −

−

POLYCATION

−

−

−

− −

− − − − − − − − − −

DIPPING & WASH

−− −

− − −

−

REPEAT

Fig. 9.1 LbL schematic for construction of multilayer assemblies by consecutive adsorption of anionic and cationic molecule-based polyelectrolytes. After the clean substrate is treated to obtain an appropriate charged surface (positive in the figure) the substrate is immersed in solutions containing the anionic and cationic polyelectrolytes. Thus, the coating is built up monolayer by monolayer. At each dipping step, the surface charge gets reversed, allowing the adsorption of each monolayer. The symbols are idealized and are not intended to represent exactly the conformation of the polyelectrolyte chains

concerning the deposition process and its possible applications can be found in the literature [7–20] and also in the chapters about ‘‘Encapsulated probes sensors’’ and ‘‘Nanostructured flexible materials’’ of this book. Taking advantage of the versatility of the LbL method for the synthesis of materials, this technique has been applied to fabricate different optical fiber devices such as Fabry–Perot interferometers on the nanometer scale, in-fiber optical gratings and different types of sensors, such as humidity, harmful gas, pH, hydrogen peroxide or glucose sensors, just to mention a few. Some of these devices will be described in the following paragraphs. For a given combination of cationic and anionic materials, the LbL technique also permits the possibility of tuning some of the deposition parameters such as the concentration of the solutions, their temperature, ionic strength or pH. This will help to tailor the properties of the fabricated films. For instance, by means of changing the pH of the polyelectrolytes solutions involved in the deposition process, it is possible to fabricate bilayers whose thickness, roughness or even the amount of the active material (a colorimetric or fluorescent indicator) can be modified. This can be clearly appreciated in Fig. 9.2 where two atomic force microscopy (AFM) images of the surface of [PAH+NR/PAA]50 coatings fabricated at two different pH values, 5.5 and 7, are shown. PAH stands for the cationic poly(allylamine hydrochloride), while PAA means the anionic poly(acrylic acid). NR is an anionic pH indicator, Neutral Red, whose

9 Optical Fiber Sensors Based on Nanostructured Coatings

279

Fig. 9.2 AFM images of [PAH+NR/PAA]50 coatings built up at pH 5.5 (left) and pH 7 (right)

color changes from red (acid form) to yellow (basic form) and 50 is the number of bilayers. An AFM Digital Instruments NanoScope IIIa multimode in tapping mode was used in order to measure the thickness of the coatings by making a scratch in the films with a razor blade and then measuring the depth of the line. The roughness of the films was also measured with the AFM in tapping mode, with silicon cantilever (RTESP, Veeco) in a region near the central zone of the sample to avoid border effects. From Fig. 9.2 it is possible to observe the great difference in roughness with values of the rms roughness (Rq) of 45.3 and 1.38 nm for the samples fabricated at pH 5.5 and 7, respectively. The thickness of these two samples was also very different, 35.2 and 10.6 nm per bilayer. This is only an example of the tunability of the coatings just by changing the pH of the solutions used for the fabrication of the films. For this particular combination of materials, [PAH+NR/PAA], the evolution of thickness and roughness of the coatings when the pH of the polyelectrolyte solutions is varied in the fabrication process is plotted in Fig. 9.3. This information is very useful to determine the optimal conditions of thickness or roughness for the final sensitive films.

9.3 Types of Devices and Sensing Mechanisms 9.3.1 Interferometric Cavities: NanoFabry–Perots The simplest structure that can be formed using nanostructured coatings on optical fibers is by depositing a coating on the cleaved or polished end of an optical fiber. The final parallel layer structure of the coating forms an interferometric cavity on the optical wavelength scale and the behavior of this structure is the same as that of a nano Fabry–Perot cavity, as shown in Fig. 9.4. In this interferometric cavity the mirrors are formed by the refractive index differences between the different optical materials: n1, n2 and n3 are the refractive indices of the optical fiber, the sensing coating and the surrounding

280

F.J. Arregui et al.

Fig. 9.3 Left: the bars show the bilayer thickness as the pH of the polyelectrolyte solutions was varied. Right: the dotted line is the square roughness of the surface of [PAH+NR/PAA]25 coatings. Graph from [21]

Fig. 9.4 Schematic of a nanoFabry–Perot formed by the nanostructured coating at the end of an optical fiber

medium, respectively. The reflectance of each mirror will be approximately that determined by Fresnel’s law for the case of normal incidence: R1 ¼

R2 ¼

ðn1  n2 Þ2 ðn1 þ n2 Þ2 ðn2  n3 Þ2 ðn2 þ n3 Þ2

;

(9:2)

(9:3)

where R1 is the reflection coefficient at the first interface (optical fiber–coating) and R2 is the reflection coefficient at the second interface (coating–air). Usually,

9 Optical Fiber Sensors Based on Nanostructured Coatings

281

these coatings incorporate some indicator and are optically denser than the core. Then the refractive index of the coating is higher than core refractive index. Then, n2>n1> n3, and for this case we can determine an expression for the reflected optical power by following the procedure of Lee and Taylor et al. to obtain the reflected optical power of a Fabry–Perot resonator [22]. Following this procedure, the phase shift in reflector 1 is p rad, because n1n3. The successive reflections in the two mirrors are indicated in Fig. 9.5. InpFig. ffiffiffiffiffiffiffi 9.5, E0 is the field amplitude of the incident lightwave on this nanocavity; E0  T12 is the field amplitudepof lightwave when it passes from ffiffiffiffiffiffiffithe transmitted d j=2 the fiber to the coating; E0  T12  e e is the field amplitude of the pffiffiffiffiffiffiffi lightwave after traveling a distance d in the coating; T12 is the transmission coefficient in the first mirror (from the fiber to the coating, from media 1 to pffiffiffiffiffiffiffi media 2);  is the absorption coefficient of the coating; and T21 is the transmission coefficient from media 2 to media 1. The value  is the round-trip phase shift and assuming that  is the wavelength in the free space then  is defined by ¼

4  p  n2  d 

(9:4)

Therefore, the reflected field amplitude ER is the addition of all the reflected terms, 0 pffiffiffiffiffiffi

1

 R1 þ pffiffiffiffiffiffi pffiffiffiffiffiffiffi B þpffiffiffiffiffiffiffi C T12  R2  T21  e2d  ej þ C (9:5) ER ¼ E0  B p ffiffiffiffiffiffiffi p ffi pffiffiffiffiffiffi pffiffiffiffiffiffi pffiffiffiffiffiffiffi @ þ T12  ffiffiffiffiffi A R2  R1  R2  T21  e4d  ej2 þ pffiffiffiffiffiffiffi pffiffiffiffiffiffi pffiffiffiffiffiffi pffiffiffiffiffiffi pffiffiffiffiffiffi pffiffiffiffiffiffi pffiffiffiffiffiffiffi 6d j3 þ T12  R2  R1  R2  R1  R2  T21  e e þ 

coating n2

optical fiber n1 E0

E 0 ⋅ T1

E 0 ⋅ T1 ⋅ e − α d ⋅ e

E 0 ⋅ R1 ⋅ e −π

E 0 ⋅ R 2 ⋅ T1 ⋅ e − 2αd ⋅ e E 0 ⋅ R 2 ⋅ T1 ⋅ R 1 ⋅ R 2 ⋅ e − 4αd ⋅ e

surrounding medium n3 − jφ 2

E 0 ⋅ T1 ⋅ T2 ⋅ e − αd ⋅ e

− jφ 2

− jφ E 0 ⋅ T1 ⋅ R 2 ⋅ R 1 ⋅ T2 ⋅ e − 3αd ⋅ e

− j2φ

. . .

. . . d

Fig. 9.5 Optical reflections in the nanoFabry–Perot

− j⋅ 3 φ 2

282

F.J. Arregui et al.

Assuming that the transmission coefficients are similar then

pffiffiffiffiffiffiffi pffiffiffiffiffiffiffi pffiffiffiffiffiffi T12 ffi T21 ¼ T1 ,

0 pffiffiffiffiffiffi 1  R1 þ p ffiffiffiffiffi ffi B þ R  T  e2d  ej þ C 2 1 B C ER ¼ E0  B pffiffiffiffiffiffi pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 4d j2 C @ þ R2  T 1  R1  R2  e e þ A pffiffiffiffiffiffi þ R2  T1  R1  R2  e6d  ej3 þ   

(9:6)

therefore,   1 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi n X pffiffiffiffiffiffi pffiffiffiffiffiffi ER ¼ E0   R1 þ  R2  T1  e2d  ej  R1  R2 e2d  ej

(9:7)

n¼0

In addition, it is clear that pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi     R1  R2 e2d  ej 51

(9:8)

Then, the next sum is the sum of a geometric series with a common ratio lower than 1, hence we can write 1 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi n X R1  R2 e2d  ej ¼ n¼0

1 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2d j 1  R1  R2  e e

(9:9)

Since it is possible to define Ai based on the equation, Ai þ Ti þ Ri ¼ 1

(9:10)

(where Ai would be associated with the losses due to scattering), it is possible to simplify the expression of the reflected field amplitude: 

 1 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ¼ 1  R1  R2  e2d  ej pffiffiffiffiffiffi pffiffiffiffiffiffi  pffiffiffiffiffiffi   R1 þ R1  R2  e2d  ej þ R2  T  e2d  ej pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ¼E0  ¼ (9:11) 1  R1  R2  e2d  ej  pffiffiffiffiffiffi pffiffiffiffiffiffi   R1 þ R2  ð1  A1 Þ  e2d  ej pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ¼E0  1  R1  R2  e2d  ej

pffiffiffiffiffiffi pffiffiffiffiffiffi ER ¼E0   R1 þ R2  T1  e2d  ej 

Since what is experimentally measured is the optical power instead of the field amplitude, it is more interesting to know the expression of the reflected optical power that will be defined as

9 Optical Fiber Sensors Based on Nanostructured Coatings

283

1 IR ¼ ER ER ¼ 2

 pffiffiffiffiffiffi pffiffiffiffiffiffi  1  R1 þ R2  ð1A1 Þe2d ej pffiffiffiffiffiffiffiffiffiffiffiffiffi 2d j ¼ E20  2 1 R1 R2 e e  pffiffiffiffiffiffi pffiffiffiffiffiffi  (9:12)  R1 þ R2  ð1A1 Þe2d ej pffiffiffiffiffiffiffiffiffiffiffiffiffi 2d j  ¼ 1 R1 R2 e e pffiffiffiffiffiffiffiffiffiffiffiffiffi 2 4d 1 R1 þR2  ð1A1 Þ e 2 R1 R2  ð1A1 Þe2d cos pffiffiffiffiffiffiffiffiffiffiffiffiffi ¼ E20  2 1þR1 R2 e4d 2 R1 R2 e2d cos

Since the incident optical power is I0 ¼

1 1  E0  E0 ¼  E20 2 2

(9:13)

then, the ratio between reflected optical power and incident optical power will be finally, RFP ¼

pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi IR R1 þ R2  ð1  A1 Þ2 e4d  2  R1  R2  ð1  A1 Þ  e2d  cos  pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ¼ I0 1 þ R1  R2  e4d  2  R1  R2  e2d  cos 

(9:14)

This theoretical model can predict the reflected optical power from the cavity depending on the thickness of the coating [18–20, 23]. Similar curves can be also obtained experimentally just by using a very simple optical arrangement, see Fig. 9.6. In this experimental setup a low coherence light source, like a LED which usually has a coherence length of some tens of microns, can be used. This is possible because, in order to observe the interferometric behavior of the cavity, it is necessary to use light sources with coherence lengths longer than the round-trip optical path length. Since the cavities studied here are under the micron thickness the utilization of inexpensive LEDs or even incandescent white light sources can be used for monitoring this phenomenon. An example of one of these curves is plotted in Fig. 9.7. nanoFabry-Perot

light source coupler detector

index matching gel

Fig. 9.6 Basic experimental setup for monitoring the reflected optical power from the nanoFabry–Perot

284

F.J. Arregui et al. 9 Theoretical

Reflected Optical Power (%)

8

Experimental data

7 6 5 4 3 2 1 0 0

50

100 150 Number of bilayers

200

250

Fig. 9.7 Experimental and theoretical results of the optical power reflected by the nanointerferometric cavity as the number of bilayers increases, the theoretical results have been obtained using the model described earlier. Reproduced with permission from [18]

The curves obtained here, see Fig. 9.7, are similar to a dumped sine where the dumping factor of this curve would be determined by Ai and  parameters that have to be estimated and adjusted experimentally: Ai depends on the scattering of the coating, therefore it depends on the roughness of the coating and this value is affected by the fabrication conditions as was shown in Fig. 9.3,  is also determined by the fabrication conditions. But perhaps the more interesting thing is that the oscillation of these curves, the position in number of bilayers of peaks and valleys, defines the optical thickness, the product of the refractive index multiplied by the thickness of the coating. This is because the maximum reflected power will happen when the round-trip phase shift is an integer multiple of p. Then, from Equation (9.4), the optical thickness n2_d can be calculated. Since it is easy to have an estimation of the refractive index of a material with an error lower than 2% (i.e., by ellipsometry), then the thickness of these coatings could be deduced. For instance, usually most of the polymeric coatings have refractive indices around 1.5–1.8 and it is easy to obtain a measurement of the refractive index of these materials with a precision of 2 decimals or higher by using different methods, therefore, after obtaining the experimental curves of the reflected optical power it is easy to determine the approximate thickness of the nanoFabry–Perot; more details can be found in [18–20, 23]. These nanoFabry–Perots can be very useful for sensing. The classical optical fiber Fabry–Perot etalon operates based on the variation of the length of the cavity, d, to produce a phase shift, and thus, a change in the intensity of reflected or transmitted optical power. This can be possible if the sensing coating is made

9 Optical Fiber Sensors Based on Nanostructured Coatings

285

with materials that swell, for instance hydrogels. In addition, this new Fabry– Perot etalon can take advantage of changes in the reflection of the external mirror at the endface of the optical fiber, R2, that can produce a change in the reflected optical power. For instance, changes in R2 with the humidity could be achieved because the coating material is hygroscopic. Other sensing mechanism of the nanoFabry–Perots could be based on the incorporation of dyes or indicators sensitive to a target, for example, pH. Some of these applications will be shown later.

9.3.2 Microgratings Since the combination of polyanions and polycations overlap each other at the molecular level and this produces a homogeneous optical material with a given refractive index A, then with other polyanions and polycations it is possible to build stacks of materials with different refractive indices, A, B, C, etc. Following this philosophy, a stack of multiple quarter-wavelength-thickness films with alternating high and low refractive indices can be deposited on the end of an optical fiber to form a grating, see Fig. 9.8. This structure materializes a micrograting whose spectrum changes when the number of quarter-wavelengththickness films in the stack increases. The more quarter-wavelength-thickness films are built on the stack the narrower are the resonance bands of the resultant optical filter. See Fig. 9.9. where the optical response of two of these microgratings is plotted. The spectral response of these microgratings to the sensing target is such that it provides some wavelengths where the reflected optical power does not change when the sensing target varies and that could be taken as reference wavelengths for normalizing the signal, in a similar way to the isosbestic points in pH sensors [24, 25].

9.3.3 Coatings on Conical Surfaces Sometimes, especially when we are dealing with fluorescent indicators, it is very interesting to have surfaces that minimize the reflected optical power from the

Fig. 9.8 Schematic of the micrograting fabricated by means of a stack of quarter-wavelengththickness films. The higher refractive index material stacks are noted as ‘‘H’’, and the lower refractive index material stacks as ‘‘L’’

F.J. Arregui et al. 5

HLHLH

4 3

H

2 1 0 –1

Reflected Optical Power (dB)

Reflected Optical Power (dB)

286

–2 1150 1200 1250 1300 1350 1400 1450 1500 1550 1600 1650

Wavelength (nm)

5 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0 –0.5

HLHLH H

1150 1200 1250 1300 1350 1400 1450 1500 1550 1600 1650

Wavelength (nm)

Fig. 9.9 Spectral response of microgratings fabricated with the same materials and experimentally adjusted for achieving a maximum reflectance at 1,310 nm (left) or at 1,550 nm (right). These gratings consist of an alternating stack of 5 quarter-wavelength-thickness films of materials with high refractive index (H) and low refractive index (L). In the figure, the H curve is the almost flat spectrum of a quarter-wavelength-thickness coating (a nanoFabry–Perot) and HLHLH is the spectrum of the micrograting. The fabrication procedure and materials employed are described in [24, 25].

excitation light and maximize the fluorescence from the indicator. This can be achieved with different approaches. On one hand, perpendicular cuts at the end of the fiber made easy the reproducibility of the devices because they can be performed using a fiber cleaver with negligible human participation. However, this perpendicular cut provides less surface area for deposition than an oblique cut at the end of the fiber (a slanted endface) or a tapered end surface. Even worse, a perpendicular cut is the optimum configuration for a maximum reflection, thus, most of the excitation light sent through the fiber to the fluorescence coating is reflected back masking the fluorescence spectrum and making more difficult the study of the devices’ properties. On the other hand, the conical shape at the tapered end maximizes the depositing surface and minimizes the reflected signal which makes it an optimum candidate for the fabrication of sensors based on fluorescence [26], see Fig. 9.10. The difficulty of depositing materials on a conical surface by classic deposition techniques is easily overcome by the LbL technique as has been reported in the literature [27, 28]. In addition to this, biconically tapered optical fibers of single-mode optical fibers, see Fig. 9.11, are very sensitive to changes of the surrounding refractive index and the wavelength of the input light and can be used for evanescent fieldbased sensing schemes. In this configuration, when a fiber is tapered, the core/ cladding interface is redefined in such a way that the single-mode fiber in the

Fig. 9.10 Schematic of an optical fiber tapered end coated with a LbL sensing coating

9 Optical Fiber Sensors Based on Nanostructured Coatings

287

Fig. 9.11 Schematic of a biconically tapered optical fiber sensor with a LbL sensing overlay

central region of the taper acts as a multimode fiber and the light becomes guided through the cladding of the fiber, which plays the role of the new core, and the new cladding is the surrounding external medium. Theoretical aspects of this phenomenon have been exhaustively studied in the literature [29–32]. Classical deposition techniques such as physical vapor deposition (PVD) cannot deposit azimuthally symmetrical palladium, other techniques such as sol-gel dip coating do not allow the overlay thickness to be tuned in the nanometer scale. By means of the LbL, the idea here is that a nanostructured film can be deposited on the thinner part of the biconically tapered optical fiber with the possibility of controlling its thickness in real time and in a nanometer scale, providing the possibility to stop the nanodeposition process when the optimum working point is reached [33, 34, 35].

9.3.4 Coatings onto Long-Period Gratings Long-period gratings (LPG) consist of a periodic index modulation of the refractive index of the core of a single-mode fiber, with a much longer period than the more known fiber Bragg gratings (FBG). If FBGs have periods typically on the sub-micron scale, LPGs have periods in the range 100–1,000 mm. This grating induces attenuation bands in the transmission spectrum based on the coupling between the core mode and the copropagating cladding modes. Consequently, the influence of the surrounding medium on the properties of LPGs is more important than in FBGs, where there is a contrapropagative coupling between core modes. In this way LPGs have extended its applications to both optical communications and sensor fields [36]. The sensing mechanism is based on the spectral shift of the attenuation bands which are generated due to the coupling of light from the core mode to the cladding modes. Since the conditions of this coupling can be affected by the temperature, strain, curvature and, what is more relevant in this chapter, the refractive index of the surrounding medium, then temperature, strain, curvature and refractometric sensors based on LPGs can be fabricated [37–40]. In fact, LPGs exhibit a high

288

F.J. Arregui et al.

sensitivity for refractive indices similar or slightly lower than the one of the fiber, usually around 1.45 at 1,550 nm, and show negligible sensitivity to refractive indices higher than 1.48 [41]. This means that it is possible to coat a LPG with a sensitive material that changes its refractive index with a target parameter (gas, chemical specie, biological agent). Once again, if the coating is thin enough some very interesting properties appear. For instance, the central wavelengths and the minimum transmission of the LPG attenuation bands can exhibit a dependence on both the thickness and the refractive index of an overlay material, even when the overlay has a refractive index higher than that of the cladding [41, 42]. Moreover, the sensitivity of these devices can be enhanced by adjusting the exact thickness of the coatings [43, 44] or by combining a two-overlay structure where the sensitivity can be improved by a factor of 70 [45, 46]. James et al. have also proposed the utilization of cascaded LPGs [47]. These concatenated LPGs act as a Mach–Zehnder interferometer: light coupled into the cladding by the first LPG is re-coupled into the core by the second LPG, where it interferes with the light that propagated in the core, producing interference fringes within the attenuation bands. Very recently, the selective removal of the LbL coating of Fig. 9.12 to form a periodic coating onto the LPG has also been proposed as a potential device for sensing applications [48].

9.3.5 Coatings on Hollow Core Fibers Very recently, it has been proposed a structure for evanescent field sensing which consists of a short segment of hollow core fiber (HCF) spliced between two standard multimode fibers (MMFs) [35, 49]. For the sake of simplicity, it is designated MMF-HCF-MMF (MHM) throughout the text (see Fig. 9.13). Early works reported about a MHM structure which was used for the surrounding medium

LbL overlay cladding core nh

0

–10

–20

–30 1000

1100 1200 1300 1400 Wavelength (nm)

1500

output light nl

period

Transmission (dB)

Transmission (dB)

input light

0

–10

–20

–30 1000

1100 1200 1300 1400 Wavelength (nm)

1500

Fig. 9.12 Schematic of a sensor based on a nanostructured overlay onto a long-period grating (LPG). In the graph the input light from a broadband light source is represented with a flat spectrum (left) and the output light from the LPG has attenuation bands inherent to the LPG

9 Optical Fiber Sensors Based on Nanostructured Coatings

289

Fig. 9.13 Schematic of a MHM structure with a sensing coating in the inner part of the hollow core fiber

fabrication of strain sensors based on in-line fiber etalons; these interferometric cavities were formed by the segment of HCF spliced between two sections of multimode fiber [50]. In the cited works, light is confined in the air core of the HCF. In contrast, the aim here is to spread out the light into the silica cladding of the HCF in order to achieve a higher evanescent field ratio with respect to the total transmitted optical power. This evanescent field intensity will be modulated as a function of the coating deposited onto the external part of such a HCF segment. The HCF-based structure used is schematically shown in Fig. 9.13. It consists of one short segment (10–20 mm) of HCF spliced between two standard MMFs. The jacket of the HCF has previously been removed. If the HCF and the MMF are spliced together using the appropriate electric arc conditions, the HCF collapses, forming a tapered solid fiber in the interface between both fibers (see photograph in Fig. 9.14). In these devices, the light that is guided into the core of the lead-in MMF can be coupled to the cladding of the HCF due to the tapered region instead of being confined in the air core. When the light reaches the lead-out MMF, it is coupled into the silica core again. Because the light is guided by the silica cladding in the HCF region, these devices can be used as evanescent field sensors that are sensitive to any coating deposited onto this region. Plus, using the LbL technique it is possible to adjust the thickness of the coating to its optimum value for sensing [35, 49]. (a)

(b) (d)

(c)

(e)

Fig. 9.14 (a) Collapsed end of the HCF spliced to the lead-in MMF section. (b) Collapsed end of the HCF spliced to the lead-out MMF section. (c–e) Pictures of the light projected by different parts of the MHM on one screen, using one green laser at 530 mm. (c) Light projected by the lead-in MMF. (d) Light projected by the HCF. (e) Light leaving the lead-out MMF

290

F.J. Arregui et al.

9.4 Sensing Applications After reviewing the different structures that have been fabricated by LbL assembly on optical fibers, we will summarize the sensors that have been reported so far using these structures.

9.4.1 Humidity A humidity sensor based on a nanoFabry–Perot, see Fig. 9.4, was the first optical fiber sensor reported using the LbL technique [20]. The sensing mechanism of this sensor approach relies on the change in the reflected optical power of the specific coating materials, used to form the interferometric cavity of the nanoFabry–Perot, that is induced by humidity or a target chemical. That is, some molecules of the target can be trapped at the surface of the sensing coating, altering the optical reflectivity. For instance, in the case of relative humidity measurements, it is clear that, for a given relative humidity value, the choice of either a hydrophilic or hydrophobic sensing material will determine the size of the water drops trapped on the surface due to the respective affinity or repulsion of the water to the sensing material. In fact, choosing properly the polyelectrolytes that form the nanostructured coating, for instance a given combination of hydrophilic and hydrophobic materials, it is possible to fabricate humidity sensors with a very fast response, shorter than a tenth of a second. Because the resulting nanostructured coatings have mixed properties of the hydrophilic and hydrophobic materials, these films are sensitive to humidity with little water absorption. Due to this the response time is so fast that these sensors can be used even for human breathing monitoring [19, 20, 51], see Fig. 9.15. In addition, as can be seen in Fig. 9.16, the sensors showed a repetitive response. Another device for simultaneously measuring the humidity and temperature is based on a fiber Bragg grating (FBG) cascaded with a low-finesse Fabry–Perot [52]. The sensing scheme is depicted in Fig. 9.17. Light from a broadband light source is launched to the combined FBG and Fabry–Perot fiber sensor head through a 3-dB optical fiber coupler. The FBG element reflects optical power with maximum reflection coefficient centered at the Bragg wavelength. This wavelength will shift due to changes in temperature, and the measurement of this shift in wavelength can be used to give a reading of the corresponding change in temperature at the physical location of the sensor head. The FBG also behaves as a stop-band optical filter, so optical wavelengths outside of the Bragg grating window are transmitted with negligible attenuation. These transmitted wavelengths thus pass through the Bragg grating and reach the humidity sensor. For these wavelengths, reflection at the Fabry–Perot changes depending on the humidity. In this way, the reflected optical power again passes through the Bragg grating, this time in the opposite direction with respect to that of the incident light. Because

9 Optical Fiber Sensors Based on Nanostructured Coatings

291

90

–41

80

–42

70

–43

RH (%)

60

–44

50

–45

40 –46

30

–47

20

RH (climatic chamber) OFHS

10 0 0

50

100 150 200 Time (minutes)

250

Optical Power (dB)

Fig. 9.15 Response to human breathing of an optical fiber humidity sensor based on a LbL nanoFabry–Perot formed. The materials used were Poly R-478 and poly(diallyldimethyl ammonium chloride). Reprinted with permission from [20]

–48 –49 300

Fig. 9.16 Response of the optical fiber humidity sensor (OFHS) to changes of relative humidity (RH) at a constant temperature of 258C, inside a climatic chamber for 96 hours of continuous monitoring. Reprinted with permission from [20]

the optical power spectrum of this reflected signal matches the transmission window of the Bragg grating, the optical power transmitted at the Bragg wavelength is not altered by the presence of the grating. Therefore, by measuring the reflected optical power within this unaltered transmission window of the Bragg

292

F.J. Arregui et al.

optical fiber

d

BRAGG GRATING

WHITE LIGHT SOURCE

HUMIDITY SENSOR

TRANSMISSION

OPTICAL POWER

INCIDENT LIGHT

WAVELENGTH

OPTICAL FIBER

OPTICAL FIBER REFLECTION

TEMPERATURE

REFLECTION

WAVELENGTH

%RH WAVELENGTH

WAVELENGTH

REFLECTED LIGHT

Fig. 9.17 Schematic of the sensing system of the dual temperature–humidity sensor. Reprinted with permission from [52] # 2002 IEEE

grating, it is possible to measure intensity changes corresponding to variations in the humidity. In order to corroborate this, the concatenated FBG and nanoFabry–Perot sensors described above were evaluated in a climatic chamber. The results obtained after a test cycle involving a constant temperature of 258C with relative humidity values of 11, 33, 52, 68, 85 and 97% are plotted in Fig. 9.18. In the region labeled ‘‘Zone A’’, where the six curves overlap, it is noticeable that the reflected optical power at the Bragg wavelength was not affected by humidity changes. In contrast, in the region labeled as ‘‘Zone B’’, the reflected optical power is observed to change as a function of the relative humidity. More details about this dual sensor are in [52].

9.4.2 Temperature In [53] the deposition of CdTe Quantum Dots (QD) on tapered ends of optical fibers has been reported. A very interesting property of QDs is that they can be excited in a broad range of wavelengths and, at the same time, have a narrow emission spectrum. Moreover, the center wavelength of the emission peak depends on the geometrical size of the QD; therefore, the emission wavelength can be tuned by changing the size of the nanocrystals that gives a large choice of emission wavelengths. In addition, QDs offer an exceptional photostability and

9 Optical Fiber Sensors Based on Nanostructured Coatings 16

293 Temperature = 25ºC

Reflected Optical Power (dB)

Zone A

14 12 10 8 6 Zone B

4 RH = 11% 2 0 RH = 97% –2 1540

1545

1550

1555

1560

Wavelength (nm)

Fig. 9.18 Spectral response of the dual temperature–humidity sensor with a coating of 350-nm thickness at a constant temperature of 258C and different humidity rates. Reprinted with permission from [52] # 2002 IEEE

have a high quantum yield compared to regular fluorescent dyes used for sensing applications: higher than 50%. In [53] QDs of different sizes, 4 and 5 nm, with a maximum fluorescence around 570 or 620 nm, respectively, were successfully tested for temperature monitoring, and showed a linear and reversible variation of the emission wavelength for a temperature range from 30 to 1008C, with a sensitivity of 0.2 nm/8C. One of the main drawbacks of this sensor is the photobleaching effect of the QDs due to photooxidation. In order to minimize this effect, a good strategy could be the encapsulation of the sensing coatings. In [54] the fabrication of a QD nanofilm inside a hollow core fiber (HCF) has been reported. This portion of HCF with an inner sensing coating is then spliced between two multimode fibers as was plotted in Fig. 9.13. This way the material is protected against the environmental conditions and is still able to sense temperature changes since it is conducted through the thin glass cylinder of the HCF segment (50 mm of diameter). Using this self-encapsulated structure, photobleaching behavior is enhanced: if previous devices with no encapsulation showed an intensity decay of 80% after 1.5 hours of continuous illumination, with this new device a decrease of only 6.3% of the fluorescence emission intensity after 4 hours of continuous illumination has been reported.

9.4.3 Gas and Volatile Organic Compounds An example of an optical fiber sensor for the detection of volatile organic compounds (VOCs), such as acetone, dichloromethane or ethanol, has been reported based on the deposition of Al2O3 and polymer ultra-thin films on the

294

F.J. Arregui et al.

ends of optical fibers forming a nanoFabry–Perot [55]. These sensors are designed to operate at the standard communications wavelengths with no cross-sensitivity to temperature from at least 10 to 708C. A humidity crosssensitivity of less than 1.4% from 11 to 85% of relative humidity was achieved and this cross-sensitivity was still negligible after 1 year of the fabrication. A similar approach which consists of a nanoFabry–Perot of polymeric films doped with a vapochromic material, Au2Ag2(C6F5)4(C6H5N)2, has been reported for the detection of some VOCs such as methanol, ethanol and isopropanol. This material changes its color when exposed to some organic vapors, recovering its original state when vapors get disappeared [56]. A more complex scheme for the detection of dichloromethane is based on a multilayered dielectric stacks on the ends of optical fibers (microgratings, see Fig. 9.8). This technique allows the selection of the sensing wavelength, as well as the reference wavelength, prior to sensor fabrication. The sensors exhibit negligible cross-sensitivity with temperature between 10 and 708C, a fast response time (less than 2 s), and negligible hysteresis [24, 25]. As is usual in generic VOC sensors, the main drawback of all these sensors is the low selectivity to a single compound. In a different approach, multilayer fluorescent films containing ruthenium complexes were also deposited on optical fibers and additional layers of fluorescein 5(6)-isothiocyanate (FITC) were also deposited to serve as an internal reference and allow ratiometric measurements of dissolved oxygen in aqueous solutions [57].

9.4.4 pH and Chemical Species Different techniques have been studied for pH sensing. The simplest ones are based on the colorimetric behavior of a pH indicator embedded in a polymeric matrix [21, 58, 59]. Very recently, it has been discovered that the utilization of poly(allylamine hydrochloride) (PAH) and poly(acrylic acid) (PAA), two of the most used polymers in LbL coatings, under certain circumstances can lead to the fabrication of polymeric films with a high swelling behavior. Moreover, devices which incorporated a pH indicator, for instance Neutral Red (NR), in the (PAH/PAA) multilayer structure were sensitive to pH changes with a high non-linear response. In [60] the polymeric matrix without the indicator, the (PAH/PAA) multilayer structure, was tested against pH variations and showed a high sensitivity to pH changes mainly due to swelling. Therefore, this matrix is not the optimum solution for the fabrication of optical fiber pH sensors based on colorimetric reagents because the swelling behavior of the polymeric matrix can affect the optical response of the nanoFabry–Perot and mask the changes in color due to pH. In order to solve this, the substitution of PAH by some pH indicator, i.e., (NR/PAA) films, has been reported as a successful strategy for the fabrication of pH sensors with a response curve more similar to the optical

9 Optical Fiber Sensors Based on Nanostructured Coatings

295

response of the sensing pH indicator. Besides, these last devices showed a fast response time (shorter than 1 s), a good resolution of 0.03 pH units, negligible drift in 30 min of continuous monitoring and a low cross-sensitivity to temperature[60]. Sometimes, conversely, it is possible to take advantage of the swelling behavior of the (PAH/PAA) matrix. In fact, these coatings can be deposited onto long-period gratings, see Fig. 9.12. Since these coatings swell depending on pH, the refractive index and the thickness of the coatings change with pH. Therefore, the wavelengths of the attenuation bands of the LPG will shift with pH [61]. Other schemes, which can overcome the possible pH-dependent behavior of the polymeric matrix due to swelling, are based on the incorporation of fluorescent dyes. Among them, 8-hydroxypyrene-1, 3, 6-trisulfonic acid trisodium salt (HPTS), a well-studied pH fluorescent dye can serve as an optimum candidate for pH monitoring. The continuous illumination of the fluorescent dye can provoke photobleaching which can be minimized with the incorporation of antifading agents in the nanostructured coatings. In fact, an improvement has been achieved from the original photobleaching rate of 58% of the non-optimized devices after only one hour and half under continuous illumination to a decrease of only 4.7% after 72 hours using the same continuous excitation source in the case of the optimized sensors, as shown in Fig. 9.19 [28]. In these fluorescent pH sensors fabricated by the LbL method the response time of the devices can be limited sometimes due to the hydrophobic nature of the polymeric matrix. In [27] the response time is minimized by modifying the hydrophilic properties of the sensitive nanocoatings, this strategy helps the diffusion of water molecules and ions through the multilayer structure. In this work, LbL multilayer polymeric pH-sensitive and non-sensitive coatings were successfully

Fig. 9.19 Lifetime enhancement of fluorescence-based optical fiber pH sensors. In this case HPTS was used as active pH-sensitive fluorophore, embedded into a LbL polymeric nanostructured coating. (a) The response of a (PAH/PAA+HPTS)10 multilayer coating shows a very high photobleaching rate. (b) The incorporation of the antifading agent DABCO into the LbL structure significantly improves the sensor lifetime, stabilizing the fluorescent signal. Reprinted from [28] with permission from Elsevier

296

F.J. Arregui et al.

built up onto both glass slides and tapered optical fiber ends. Two different structures were used to compare the results: first, fabricating only the sensitive coating and second, adding a high hydrophilic coating prior to the sensitive one. The hydrophilicity of two different structures, S1 (sensitive coating without a previous hydrophilic matrix) and S2 (sensitive coating with the previous hydrophilic matrix) were studied by means of water contact angle measurements and the results obtained were 508 and 108, respectively, as can be seen in Fig. 9.20. Using this alternative construction technique, the rise time response was five times minimized from 15 to 3 min between pH 3 and 7 and fall time response was three times minimized from 3 to 1 min between pH 7 and 3. In general rise time responses were considerably reduced to less than a fourth while fall time response reductions had less significance. In [62] a LPG, see Fig. 9.12, is coated with a LbL film sensitive to copper. The sensing coating is formed by incorporating the reagent Cibacron Blue with generation 4-poly(amidoamine) dendrimer. In that work, it is reported that a detection of 1.3 mg Cu2+ L-1 is observed when six bilayers comprised the coating. In addition, a stable response is achieved with 0.6 mg L-1 in less than 1 min and when 0.1 M HCl was used as the rinsing solution, this LPG sensor was reversible and the signal response to similar concentrations of Cu2+ reproducible [62]. The immobilizing of a redox indicator, Prussian Blue, in a nanoFabry–Perot makes possible the detection of H2O2 in the range from 10-6 to 10-3 M [63]. For higher values, the sensor saturated. The recovery of oxidized Prussian Blue was successful after immersion in a reductive agent such as ascorbic acid. In this device, the reflected power depends highly on the pH, as is usual in sensors based on redox indicators, but the slope of the change in the reflected power

Fig. 9.20 Contact angle measurements for S1 (structure with a sensitive coating deposited) and S2 (structure with a hydrophilic coating deposited before the sensitive one) devices. Reprinted from [27]

9 Optical Fiber Sensors Based on Nanostructured Coatings

297

produced by different concentrations of H2O2 presents a low dependence on the pH in the range between 6 and 7.4. The range of pH values where the sensor detects hydrogen peroxide is located between 4 and 7.4. In addition to this, the sensor is immune to interferences of a large number of products like chloride, bromide, iodide, oxalate, thiocyanate, tartrate, urea, sulfate, phosphate and glucose except for ascorbic acid, which is the reductive agent of the sensor.

9.4.5 Biological Recognition An approach presented in [64] is based on the immobilization by LbL of singlestranded capture DNA on the surface of a silica optical fiber tip. When there is a presence of the target (complementary) DNA sequence in the tested sample, the optical thickness of the fiber probe increases. This is mainly due to the change in density of the DNA monolayer. In this study, the change in optical thickness is measured by a multicavity Fabry–Perot interferometer formed by a short piece of hollow fiber sandwiched between two pieces of optical fiber. The interference of light reflected from the three fiber endfaces results in periodic oscillations in the reflection spectrum. This method can be very appealing because in contrast to other direct DNA methods, this does not require labels as indicators [64]. Based on the same sensing scheme of the multicavity Fabry–Perot, an immunosensor was proposed based on the immobilization of proteins such as immunoglobulin G (IgG) [65]. A similar approach that enhances the response of the device above described combines the utilization of microgaps in optical fiber and thin films. The microgaps are fabricated by splicing an etched fiber (the chemical etching can be achieved using hydrofluoric acid) with a cleaved end of another fiber. Then, a microgap is generated inside the fiber. After a short length of fiber (some tens of microns) the process is repeated and low-finesse FP cavities are formed between the microgap reflectors and the cleaved fiber end. This structure is a multicavity sensor which uses a cavity as temperature sensor and the other cavity for measuring the changes in thickness of the thin film. A temperature resolution of 0.18C has been reported with the microgap sensor. In this way it can be used for measuring the changes in thickness of a nanostructured LbL film deposited at the end of the fiber with an error of less than 0.2 nm from 0 to 1008C [66]. A device which takes advantage of the shift observed in the resonance peak of a surface plasmon resonance (SPR) sensor due to changes in refractive index of external medium is reported in [67]. In the device, the deposition of thin films or particles of silver or gold (usually less than 50 nm thick) combined with fluorescent anti-immunoglobulin G (IgG) antibody for the detection of IgG is necessary. A similar approach is proposed in [68] for the detection of nitric oxide using extracellular enzymes like lignin peroxidase. Very recently the detection of anti-gliadin antibodies has also been reported by means of the deposition of gliadin LbL films onto biconically tapered optical

298

F.J. Arregui et al.

fibers (Fig. 9.12); this device could be very useful as a diagnostic test for the detection of celiac disease [69].

9.5 Conclusions This chapter summarizes the applications of the layer-by-layer electrostatic selfassembly method for the fabrication of optical fiber sensors using novel sensing architectures. The suitability of this procedure to complex geometries makes this technique an optimum tool for the deposition of nanostructured sensing coatings on optical fiber devices. Humidity, temperature, oxygen, VOCs, pH, H2O2, copper or glucose optical fiber sensors have been fabricated by means of the LbL method. The possibility of incorporating proteins, enzymes or antibodies as well makes this technique especially useful for the fabrication of biosensors for biological recognition. Acknowledgments This work was funded in part by the Spanish Ministry of Education and Science-FEDER TEC2006-12170/MIC Research Grant and Government of Navarre-FEDER Euroinnova Research Grants.

References 1. Culshaw B (2000). Fiber optics in sensing and measurement. IEEE J Sel Top Quant Electron, 6(6): 1014–1021. 2. Dakin J, Culshaw B (1988, 1989, 1996, 1997). Optical Fiber Sensors. Vol I, II, III, and IV. Artech House Publishers, Massachusetts, USA. 3. Gusarov A, Fernandez Fernandez A, Vasiliev S, et al. (2002). Effect of gamma-neutron nuclear reactor radiation on the properties of Bragg gratings written in photosensitive Ge-doped optical fiber. Nucl Instrum Methods Phys Res, B Beam Interact Mater Atoms, 187(1): 79–86. 4. Matias IR, Arregui FJ, Claus RO (2006). Optical Fiber Sensors. In: Grimes CA, Dickey EC, Pishko MV (eds), Encyclopedia of Sensors. American Scientific Publishers, New York, USA. 5. Bunganaen Y, Lamb DW (2005). An optical fibre technique for measuring optical absorption by chromophores in the presence of scattering particles. J Phys: Conf Ser, 15: 67–73. 6. Davis F, Hodge P, Tredgold RH, et al. (2005). Langmuir–Blodgett films of preformed polymers containing biphenyl groups. Langmuir, 21(20): 9199–9205. 7. Lvov Y, Ariga K, Ichinose Y, et al. (1995). Layer-by-layer architectures of concanavalin A by means of electrostatic and biospecific interactions. J Am Chem Soc, 117(22): 6117–6123. 8. Ariga K, Lvov Y, Kunitake T (1997). Assembling alternate dye-polyion molecular films by electrostatic layer-by-layer adsorption. J Am Chem Soc, 119: 2224–2231. 9. Decher G (1997). Fuzzy nanoassemblies: toward layered polymeric multicomposites. Science, 277: 1232–1237. 10. Liu YJ, Wang AB, Claus RO (1997). Molecular self-assembly of TiO2/polymer nanocomposite films. J Phys Chem B, 101: 1385–1388.

9 Optical Fiber Sensors Based on Nanostructured Coatings

299

11. Lenahan KM, Wang AB, Liu YJ, Claus RO (1998). Novel polymer dyes for nonlinear optical applications using ionic self-assembled monolayer technology. Adv Mater, 10(11): 853–855. 12. Bertrand P, Jonas A, Laschewsky A, et al. (2000). Ultrathin polymer coatings by complexation of polyelectrolytes at interfaces: suitable materials, structure and properties. Macromol Rapid Commun, 21: 319–348. 13. Shiratori SS, Rubner MF (2000). pH-dependent thickness behavior of sequentially adsorbed layers of weak polyelectrolytes. Macromolecules, 33(11): 4213–4219. 14. Iler RK (1966). Multilayers of colloidal particles. J Colloid Interface Sci, 21: 569–594. 15. Pastoriza-Santos I, Scholer B, Caruso F (2001). Core-shell colloids and hollow polyelec¨ trolyte capsules based on diazoresins. Adv Mat, 11(2): 122–128. 16. Schonhoff M (2003). Self-assembled polyelectrolyte multilayers. Curr Opin Colloid ¨ Interface Sci, 8(1): 86–95. 17. Hammond PT (2004). Form and function in multilayer assembly: new applications at the nanoscale. Adv Mat, 16(15): 1271–1293. 18. Arregui FJ, Matias IR, Liu Y, et al. (1999). Optical fiber nanometer-scale Fabry–Perot interferometer formed by the ionic self-assembly monolayer process. Opt Lett, 24: 596–598. 19. Arregui FJ, Liu Y, Matias IR, et al. (1999). Optical fiber humidity sensor using a nano Fabry–Perot cavity formed by the ionic selfassembly method. Sens Actuat B, 59(1): 54–59. 20. Arregui FJ, Cooper KL, Liu Y, et al. (2000). Optical fiber humidity sensor with a fast response time using the ionic self-assembly method. IEICE Trans Electron, E83C: 360–365. 21. Goicoechea J, Arregui FJ, Corres J, et al. (2008). Study and optimization of selfassembled polymeric multilayer structures with neutral red for pH sensing applications. J Sens, Article ID 142854, 7 pages, doi: 10.1155/2008/142854. 22. Lee CE, Gibler WN, Atkins RA, et al. (1992). In-line fiber Fabry–Perot interferometer with high-reflectance internal mirrors. IEEE J Lightwave Technol, 10: 1376–1379. 23. Del Villar I, Matı´ as IR, Arregui FJ, et al. (2005). Fiber-optic hydrogen peroxide nanosensor. IEEE Sensors J, 5(3): 365–371. 24. Arregui FJ, Matias IR, Cooper KL, et al. (2001). Fabrication of microgratings on the ends of standard optical fibers by the electrostatic self-assembly monolayer process. Opt Lett, 26: 131–133. 25. Arregui FJ, Claus RO, Cooper KL, et al. (2001). Optical fiber gas sensor based on selfassembled gratings. J Lightwave Technol, 19(12): 1932–1937. 26. Garcı´ a-Moreda FJ, Arregui FJ, Achaerandio M, et al. (2006). Study of indicators for the development of fluorescence based optical fiber temperature sensors. Sens Actuat B, 118(1–2): 425–432. 27. Zamarren˜o CR, Bravo J, Goicoechea J, et al. (2007). Response time enhancement of pH sensing films by means of hydrophilic nanostructured coatings. Sens Actuat B, 128(1): 138–144. 28. Goicoechea J, Zamarren˜o CR, Matias IR, et al. (2007). Minimizing the photobleaching of self-assembled multilayers for sensor applications. Sens Actuat B, 126(1): 41–47. 29. Lacroix S, Black R, Veilleux C, et al. (1986). Tapered single-mode fibers: external refractive index dependence. Appl Opt, 25(15): 2468–2469. 30. Love JD, Henry WM, Stewart WJ, et al. (1991). Tapered single-mode fibers and devices (part 1). Adiabatic criteria. IEE Proc J, 138(5): 343–353. 31. Black RJ, Bourbonnais R (1986). Core-mode cutoff for finite-cladding lightguides, IEE Proc J, 133(6): 277–384. 32. Shankar PM, Lloyd C, Bobb HD, et al. (1991). Coupling of modes in bent biconically tapered single-mode fibers. J Lightwave Technol, 9: 832–837. 33. Corres JM, Arregui FJ, Matias IR (2006). Design of humidity sensors based on tapered optical fibers. J Lightwave Technol, 24: 4329–4336.

300

F.J. Arregui et al.

34. Corres JM, Arregui FJ, Matı´ as IR (2007). Sensitivity optimization of tapered optical fiber humidity sensors by means of tuning the thickness of nanostructured sensitive coatings. Sens Actuat B, 122(2): 442–449. 35. Matias IR, Arregui FJ, Corres, et al. (2007). Evanescent field fiber-optic sensors for humidity monitoring based on nanocoatings. IEEE Sens J, 7(1): 89–95. 36. James SW, Tatam RP (2003). Optical fibre long-period grating sensors: characteristics and application. Meas Sci Technol, 14(5): R49. 37. Ng MN, Chiang KS (2002). Thermal effects on the transmission spectra of long-period fiber gratings. Opt Commun, 208(4–6): 321–327. 38. Chen X, Zhou K, Zhang L, et al. (2004). Optical chemsensors utilizing long-period fiber gratings UV-inscribed in D-fiber with enhanced sensitivity through cladding etching. IEEE Photon Technol Lett, 16(5): 1352–1354. 39. Allsop T, Dubov M, Martinez A, et al. (2005). Long period grating directional bend sensor based on asymmetric index modification of cladding. Electron Lett, 41(2): 59–60. 40. Zhang L, Liu Y, Everall L, et al. (1999). Design and realization of long-period grating devices in conventional and high birefringence fibers and their novel applications as fiberoptic load sensors. IEEE J Sel Top Quant Electron, 5(5): 1373–1378. 41. James SW, Tatam RP (2006). Fibre optic sensors with nano-structured coatings. J Opt A, Pure Appl Opt, 8(7): S430–S444. 42. Del Villar I, Corres JM, Achaerandio M, et al. (2006). Spectral evolution with incremental nanocoating of long period fiber gratings. Opt Express, 14: 11972–11981. 43. Del Villar I, Matias IR, Arregui FJ (2006). Influence on cladding mode distribution of overlay deposition on long-period fiber gratings. J Opt Soc Am A, 23: 651–658. 44. James SW, Cheung CS, Tatam RP (2007). Experimental observations on the response of 1st and 2nd order fibre optic long period grating coupling bands to the deposition of nanostructured coatings. Opt Express, 15(20): 13096–13107. 45. Del Villar I, Matias IR, Arregui FJ (2005). Enhancement of sensitivity in long-period fiber gratings with deposition of low-refractive-index materials. Opt Lett, 30: 2363–2365. 46. Del Villar I, Matı´ as IR, Arregui FJ, et al. (2005). optimization of sensitivity in long period fiber gratings with overlay deposition. Opt Express, 13: 56–69. 47. James SW, Ishaq I, Ashwell GJ, et al. (2005). Cascaded long-period gratings with nanostructured coatings. Opt Lett, 30(17): 2197–2199. 48. Del Villar I, Arregui FJ, Matias IR, et al. (2007). Fringe generation with non-uniformly coated long-period fiber gratings. Opt Express, 15: 9326–9340. 49. Bravo J, Matias IR, Del Villar I, et al. (2006). Nanofilms on hollow core fiber-based structures: an optical study. J Lightwave Technol, 24: 2100–2107. 50. Sirkis J, Berkoff TA, Jones RT, et al. (1995). In-line fiber etalon (ILFE) fiber-optic strain sensors. J Lightwave Technol, 13: 1256–1263. 51. Kang Y, Ruan H, Wang Y, et al. (2006). Nanostructured optical fibre sensors for breathing airflow monitoring. Meas Sci Technol, 17(5): 1207–1210. 52. Arregui FJ, Matı´ as IR, Cooper KL, et al. (2002). Simultaneous measurement of humidity and temperature by combining a reflective intensity-based optical fiber sensor and a fiber bragg grating. IEEE Sens J, 2(5): 482–487. 53. de Bastida G, Arregui FJ, Goicoechea J, et al. (2006). Quantum dots-based optical fiber temperature sensors fabricated by layer-by-layer. IEEE Sens J, 6(6): 1378–1379. 54. Bravo J, Goicoechea J, Corres JM, et al. (2007). Fiber optic temperature sensor depositing quantum dots inside hollow core fibers using the layer by layer technique. Proc SPIE. doi: 10.1117/12.738388. 55. Arregui FJ, Matias IR, Claus RO (2003). Optical fiber gas sensors based on hydrophobic alumina thin films formed by the electrostatic self-assembly monolayer process. IEEE Sens J, 3(1): 56–61. 56. Elosua C, Bariain C, Matas IR, et al. (2006). Volatile alcoholic compounds fibre optic nanosensor. Sens Actuat B, 115: 444–449.

9 Optical Fiber Sensors Based on Nanostructured Coatings

301

57. Grant PS, McShane MJ (2003). Development of multilayer fluorescent thin film chemical sensors using electrostatic self-assembly. IEEE Sens J, 3(2): 139–146. 58. Arregui FJ, Latasa I, Matias IR (2003). An optical fiber pH sensor based on the electrostatic self-assembly method. Sens Proc IEEE, 1(22–24): 107–110. 59. Goicoechea J, Arregui FJ, Matias IR (2007). Optical fiber pH sensors based on selfassembled multilayered neutral red. Proc SPIE. doi: 10.1117/12.738427. 60. Goicoechea J, Zamarren˜o CR, Matias IR, et al. (2008). Optical fiber pH sensors based on layer-by-layer electrostatic self-assembly. Sens Actuat B, 132(1): 305–311. 61. Corres JM, Del Villar I, Matias IR, et al. (2007). Fiber-optic pH-sensors in long-period fiber gratings using electrostatic self-assembly. Opt Lett, 32: 29–31. 62. Keith J, Hess LC, Spendel WU, et al. (2006). The investigation of the behavior of a long period grating sensor with a copper sensitive coating fabricated by layer-by-layer electrostatic adsorption. Talanta, 70: 818–822. 63. Del Villar I, Matı´ as IR, Arregui FJ, et al. (2005). ESA-based in-fiber nanocavity for hydrogen–peroxide detection. IEEE Trans Nanotechnol, 4(2): 187–193. 64. Wang X, Cooper KL, Wang A, et al. (2006). Label-free DNA sequence detection using oligonucleotide functionalized optical fiber. Appl Phys Lett, 89: 163901.1–163901.3. 65. Zhang Y, Shibru H, Cooper KL, et al. (2005). Miniature fiber-optic multicavity Fabry– Perot interferometric biosensor. Opt Lett, 30: 1021–1023. 66. Zhang Y, Chen X, Wang Y, et al. (2007). Microgap multicavity Fabry–Perot biosensor. J Lightwave Technol, 25(7): 1797–1814. 67. Kaul S, Chinnayelka S, McShane MJ (2004). Self-assembly of polymer/nanoparticle films for fabrication of fiber-optic sensors based on SPR. In: Gannot I (ed) Optical Fibers and Sensors for Medical Applications IV. SPIE, Bellingham, WA. 68. Kuila D, Tien M, Lvov Y et al. (2004). Nanoassembly of immobilized ligninolytic enzymes for biocatalysis, bioremediation and biosensing. In: Islam MS, Dutta AK (eds) Nanosensing: Materials and Devices. SPIE, Bellingham, WA. 69. Corres JM, Bravo J, Matias IR, et al. (2007). Tapered optical fiber biosensor for the detection of anti-gliadin antibodies. IEEE Sens, 28–31: 608–611.

Chapter 10

Nanostructured Flexible Materials: Metal RubberTM Strain Sensors Christelle Jullian, Jennifer Lalli, Bradley Davis, and Richard Claus

10.1 Introduction Strain sensors are fundamental building blocks in measurement of materials and structures. Conventional foil strain gages are based on macroscopic principles of bulk material deformation due to stress, and changes in the electrical resistance of deformed bulk metal geometries. Nanostructured strain sensors that operate based on very different physical principles may be envisioned. This chapter discusses such nanostructured strain sensor devices based on self-assembled Metal RubberTM materials. The first part of the chapter reviews the background on self-assembly processing. The second part of the chapter discusses Metal RubberTM manufacturing and Metal RubberTM strain sensor operation.

10.2 Molecular-Level Self-Assembly Processing: Long-Range Ordered Langmuir–Blodgett (LB) Films and Self-Assembled Monolayers (SAMs) More than eight decades ago, Langmuir published the first theory supporting that a monomolecular film can be created and self-assembled on solid surfaces [1]. However, it was only in the 1980s that the field of self-assembly started to grow exponentially, partly motivated by the need of materials and devices miniaturized and controlled at the molecular scale. Multiple monolayers of long-chain amphiphilic oil molecules were first realized by Langmuir and Blodgett based on the following theory. A sufficiently small amount of olive oil molecules dropped on clean water should maximize spreading resulting in one monolayer of oil at the surface of water [1]. Calcium ions from talc were used as a witness of the oil layer. By measuring the area of the oil layer, and knowing the volume of oil that was dropped, the layer was C. Jullian Department of Materials Science and Engineering, Virginia Tech, Blacksburg, VA 24061, USA

F.J. Arregui (ed.), Sensors Based on Nanostructured Materials, DOI 10.1007/978-0-387-77753-5_10, Ó Springer ScienceþBusiness Media, LLC 2009

303

304

C. Jullian et al.

estimated to be one molecule thick. Furthermore, the orientation and packing of the oil molecules was demonstrated. The carboxylic acid end (COO) was found to be bound to calcium ions in the water, while the hydrocarbon end (CH3) was facing up toward the air. The spreading of oil was hypothesized to be due to the presence of ‘‘active groups’’ in the oil molecule that combined with water via ‘‘secondary valence’’. Later, the transfer of a calcium stearate ((C17H35COO)2Ca) monolayer from the air/water interface to a glass substrate was demonstrated with conservation of the order and packing that existed on the water [2]. The clean hydrophilic glass was placed in the water subphase. When rose through the floating molecules, the first layer was applied. The possibility of depositing successive monolayers was then demonstrated. Following the first monolayer applied during an upward motion of the substrate, alternating layers oriented in opposite directions were successively applied as the substrate was moving downward and upward, creating a Langmuir–Blodgett (LB) film. The film’s alternating structure was deduced from the change in the wettability of the film’s outermost layer, given by the measurement of the surface’s contact angle at different stages of the process. Surface wettability is readily noticeable as the substrate emerges dry when the hydrophobic moieties are the outer moieties, whereas it emerges wet when the hydrophilic moieties are the outer moieties. The LB deposition method has given rise to self-assembled monolayers (SAMs), which are highly ordered and crystalline thin films. SAMs are very often used to mimic lipid biomembranes, but have found some applications in molecular electronics and biotechnology devices such as protein-based devices and sensors of interest here.

10.2.1 Surfactants and Floating Monolayers Approximately 50 years separated the first transferred LB film to the extensive use of this technique for very different applications in various research fields. Surfactant molecules are also known as amphiphiles because they are usually comprised of two moieties, one hydrophobic and one hydrophilic [3]. While hydrocarbon or fluorocarbon chains constitute hydrophobic segments, charged or polar groups (NH3+, PO4, COOH, OH) and even small oligomers such as hexa(ethylene glycol) are common hydrophilic moieties. Thus, by definition, surfactants are specifically comprised of an active hydrophilic head and a hydrophobic tail. The head undergoes chemisorption via the creation of a chemical bond with a liquid or solid interface, which may involve electrostatic and van der Waals interactions. The tail forms intermolecular van der Waals interactions, with neighboring tails helping the packing and ordering of the floating surfactant. Because van der Waals interactions are short-range interactions, high packing of the molecules present at the surface and thus long-range molecular order, can only be achieved if the molecules are close enough to allow molecular tails

10 Nanostructured Flexible Materials

305

interactions. This is characterized by the surface pressure, introduced in the next paragraph. Depending on the polarity of the tail, long-range electrostatic interactions may also arise between neighboring tails. Finally, the third part of amphiphile molecules is the end group. It is usually a methyl (CH3) or hydroxyl (OH) group. The physicochemical properties of amphiphiles determine their orientation and architecture at an interface or in aqueous solutions [3]. Shorter hydrophobic tails may result in more soluble amphiphiles, which may associate as micelles in aqueous solution as opposed to longer hydrophobic tails which could give rise to floating monolayers.

10.2.2 Surface Pressure One of the critical parameters in building stable LB films is the bath surface pressure which is measured as a function of the water area per molecule at a given temperature [4]. Thus, a plot of surface pressure represents various phase transitions between solid, liquid, and gaseous states of the floating monolayer. A sufficiently high surface pressure means that the floating monolayer is cohesive and highly crystalline. This highly ordered monolayer may then be transferred to a substrate conserving the same order, provided the surface pressure remains constant. In some cases molecular reorientation occurs during monolayer transfer to a substrate. Constant surface pressure can be achieved by physical motion of the bath barrier during transfer of the ordered floating monolayer to the substrate. Variable surface pressure translates into variable amphiphile orientations at the liquid interface as shown in Fig. 10.1. As a result, film defects such as boundaries between solid, liquid, and gas phases, boundaries between crystalline and collapsed areas, and pinholes generally result from lower surface pressures [4]. This translates into thickness and refractive index nonuniformity, as well as rigid, amorphous, and unstable structures. As mentioned previously, this ‘‘ideal’’ surface pressure depends on the chemical and physical structures of the amphiphile molecules and is often obtained empirically. Water is a highly ordered liquid via hydrogen bonds and thus has a fairly high surface tension (73 mN/m at 208C) [4]. Surfactants usually lower the water

Fig. 10.1 Illustration of a Langmuir isotherm showing the crystalline, liquid, and gas phases, with their corresponding orientations at the liquid interface

Surface pressure

solid

••••••••

liquid

•• • •••• •

gas Area per floating molecule

•

•••

•

•

306

C. Jullian et al.

surface tension. The surface pressure is simply the difference between , the surface tension of pure water, and  0, the water surface tension with the floating monolayer. is given as Y

¼   0

Controlling the surface pressure of the floating monolayer allows control of the state of the monolayer at the water interface. Increased surface pressure corresponds to increased packing, order, and crystallinity of the floating monolayer. As a result, as the surface pressure increases, the number of molecules per unit water surface increases, and the monolayer state at the interface may change from gas-, to liquid-, to solid-like structures. Disordered floating molecules refer to a large water surface available to floating amphiphiles. By decreasing the water area, the molecules are physically moved closer together until crystalline-like order is achieve d.

10.2.3 Monolayer Transfer on Solid Surfaces The deposition of a monolayer on a substrate is obtained by immersing the substrate down through the monolayer or/and pulled up through the monolayer [4]. When a hydrophobic substrate is immersed in the liquid phase through the monolayer on the liquid interface, the hydrophobic tails of the monolayer spontaneously adsorb on the substrate. Conversely, when a hydrophilic substrate is pulled up through the monolayer on the liquid interface, the hydrophilic heads of the monolayer adsorb on the hydrophilic substrate. Depending on the sequence chosen, various multilayer structures can be achieved [3, 4]. A head-to-tail or X-type structure is created by depositing a first monolayer where the tails are adsorbed on the substrate and the heads are sticking out, and subsequently depositing a second monolayer where the tails adsorb on the previously deposited monolayer heads. Varying the substrate hydrophilicity/ hydrophobicity and the dipping up/down sequence, tail-to-head, head-to-head, or tail-to-tail structures may also be created. The design of the LB film deposition apparatus (troughs) has been significantly improved over the years. Automated systems improving deposition speed and film quality and reducing vibrations with shock absorbers are in wide use today. A wide range of temperatures for the water bath is available. Troughs comprised of two water baths and two rotating arms are now common, decreasing bath contamination and allowing multi-functional film fabrication. Deionized and bacteria-filtered water is used to ensure maximum purity. These have led to improved structural quality and crystallinity of LB films. The deposition of monolayers on a substrate via self-assembly can therefore create long-range ordered structures. Although the nature of both the substrate and the amphiphilic molecules may differ, this particular adsorption mechanism is referred to as spontaneous chemisorption via the creation of a chemical

10 Nanostructured Flexible Materials

307

bond between surface-active amphiphile heads and a surface-active substrate. The most common chemical bonds include covalent, ionic, dipolar, and hydrogen bonds. The LB plate substrate is usually made of glass, quartz, silicon, platinum, or mica. These allow for self-assembly of alcohols, carboxylic and fatty acids, phospholipids, and amines via van der Waals or ionic interactions, as well as self-assembly of hydrophobic tails on hydrophobic substrates via hydrophobic interactions. The water pH, ionic strength, and temperature significantly affect the rate at which the floating monolayer can be transferred to a substrate.

10.2.4 From LB Films to SAMs Although self-assembled monolayers and multi-monolayers were first realized about 85 years ago, the creation of films with specified physical, mechanical, and thermal properties remains a challenge. One strategy has been to use amphiphilic polymers with hydrophilic groups present along the chain as opposed to monomers [4]. This has allowed to expend the range of materials which can be deposited with the LB technique. Reinforcement of the multilayer mechanical and thermal properties has been achieved by taking advantage of cross-linkable groups [4] and introducing additional interactions such as electrostatic interactions between amphiphiles and proteins [5]. The LB process is also limited by the deposition rate, and the substrate size and topology which can be coated. Finally, while the LB technique offers fabrication of tailored molecular architectures, it is restricted to amphiphilic molecules with appropriate physicochemical properties which make them insoluble, i.e., floating. The obtained long-range ordered structures often comport defects. Self-assembled monolayers (SAMs) carry the same high order and crystallinity as LB films do, but are generally processed by immersing a substrate in a dilute solution containing amphiphile molecules. Because the deposition process occurs in solution, single SAM with the amphiphile surfaceactive heads chemisorbed onto the surface-active substrate limits the number of feasible film architectures. Consequently, van der Waals interactions usually control the adsorption process, inducing long-range order. While the LB technique allows for multi-monolayers deposition, a SAM traditionally refers to a single, ultrathin, and crystalline monolayer. In fact, most SAMs are formed from stabilizing alkyl CH2 chains, allowing for crystalline packing of the amphiphiles. This concept allows for some freedom in designing the amphiphile molecule. For instance, Prime and Whitesides have synthesized mixed SAMs comprised of highly crystalline hydrophobic alkyl moieties and highly amorphous hydrophilic oligo(ethylene glycol) moieties [6]. Because adsorption occurs in solution, soluble amphiphiles are used, which opens up new architectural possibilities such as utilizing micelles. Unlike LB films, SAM deposition is controlled by the diffusion of amphiphiles to the substrate which has the potential to provide longer-range order with very

308

C. Jullian et al.

a) X X

X

b) X

X

X X

X

Si S

S

S

S

S

S

O

X

Si O

O

Au

X

Si O

O

Si O

O O

O

Si

Fig. 10.2 (a) Organothiol SAMs on gold substrate showing the 308 tilt angle with the surface normal and the long-range order, and (b) organosilicon SAMs on hydroxylated silicon substrate. The Xs represent the end groups attached to the tails, which are most of time hydroxyl (OH) or methyl (CH3) groups

few defects. Often, amphiphiles comprising thiol (SH) end-functionalities are used to deposit organothiol monolayers on gold substrates due to high bond stability [7]. The deposition process is known to first occur via chemisorption of the amphiphilic active heads onto the substrate which follows a first-order Langmuir isotherm. There is a linear relationship between the SAM adsorption rate and the number of unoccupied gold atoms. In addition, lateral rearrangement of the SAMs occurs in order to satisfy van der Waals attraction forces between neighboring chains and consequently leads to long-range order. Van der Waals interactions are maximized by a 308 tilt of the alkyl chains (from the gold surface normal) arising from the hexagonal structure of S–Au bonds [8]. Organosilicon SAMs deposited on hydroxylated substrates (OH functionalized) is another well-known and stable system based on Si–O covalent bonds. Organosilicon SAMs are characterized by very strong adhesion to substrates due to the high binding energy of siloxane Si–O–Si bonds, but are not as ordered as organothiol-based SAMs. Figure 10.2. represents organothiol and organosilicon SAM structures.

10.3 Layer by Layer: Toward Shorter-Range Ordered Structures 10.3.1 Electrostatic-Based Self-Assembly In contrast to LB films and SAMs, layer-by-layer (LBL) self-assembly allows for the formation of multilayers that may exhibit short-range order as opposed to long-range order seen with LB films and SAMs. This results from the common deposition of high-molecular-weight flexible polyelectrolytes, which render LBL films more amorphous in nature with a fuzzy internal structure. However, this more amorphous structure has proven to be appropriate for many applications such as organic light-emitting diodes and photovoltaics [8] due to the absence of

10 Nanostructured Flexible Materials

309

defects which are readily introduced using LB films and SAMs. Two- and threedimensional self-assembled structures have been demonstrated. In 1992, Decher and coworkers introduced this novel technique to deposit polyelectrolyte thin films on charged substrates based on electrostatic attraction [8]. Cationic and anionic polyelectrolytes in solution at relatively high concentrations were thus successively adsorbed on an oppositely charged surface, inducing a reversal of the surface charge. A rinsing step in between deposition of each cationic and anionic layer is necessary in order to remove loosely attached ionic residues and prevent solution contamination. The formation of multilayer assemblies was thus achieved by Decher and coworkers. Up to 100 layers were deposited, and the linearity of the process was demonstrated using UV/Vis spectroscopy and small angle X-ray scattering, techniques still used today. The film thickness was found to increase linearly with the number of bilayers. The LBL technique was in fact first experienced with the alternate deposition of charged inorganic colloids by Iler in 1966 [9]. Later, magnetic particles [10], polyelectrolyte polymers [8], functionalized fullerenes [11], charged DNA [12], proteins [13], gold nanoparticles [14, 15], and clay platelets [16] were selfassembled on substrates based on electrostatic interactions. The LBL deposition technique is therefore appropriate for various types of nanoparticles in addition to polyelectrolytes, which makes it considerably more versatile than LB films and SAMs. Additionally, the adsorption of higher molecular weight molecules makes the LBL technique appropriate for building supramolecular self-assembled systems. Kunitake demonstrated the fabrication of ordered multilayer free-standing films, which opened up new perspectives [17]. Casting of dispersed bilayers and polymers was found to conserve the multilayerarchitecture, and therefore, ordered free-standing composite films were demonstrated. The tailoring of molecular design and interactions translated the multilayer molecular order into macroscopically ordered materials. Multilayer composite cast films made from dispersions of amphiphiles, polymers, proteins, quantum-sized metallic particles (CdS), and metal oxide (Al2O3) have been fabricated [17]. LBL films can be highly uniform due to the conformal nature of the LBL electrostatic self-assembly process. The driving force is the electrostatic attraction between oppositely charged or polar molecules on a given substrate. Adsorption of charged or polar molecules on a surface induces a change in free energy at the interface. This change in free energy is believed to be associated with the desolvated substrate, the displacement of counterions at the interface, and the electrostatic attraction between the substrate and the charged adsorbent [18]. Indeed, the formation of LBL self-assembled architectures represents a thermodynamic equilibrium (minimum free energy) that is mainly governed by a change in entropy and a relatively small change in enthalpy. The change in enthalpy depends on the interactions between adjacent bilayer molecules and is supposed to be relatively small compared with covalent bond formation for example. The change in entropy occurs by bringing molecules close to each other, which were originally somewhat far apart. This depends on

310

C. Jullian et al.

the solute concentration and creates a decrease in translational entropy, while binding these molecules to the substrate induces a small decrease in conformational entropy. Note that the change in entropy is generally smaller for the LB technique than for the LBL process, because the LB technique is only affected by a relatively small decrease in conformational entropy since the molecules are already ordered at the liquid–gas interface. This entropy loss is outweighed by a large entropy gain due to the liberation of solvent and counterions from the interface, which drives the adsorption. Because of the change in entropy, kinetic controls the adsorption of LBL molecular layers.

10.3.2 Factors Influencing Adsorption and Structure As in the LB process, the solution pH, temperature, and concentration are known to affect the LBL film structure and deposition rate. This is due to the fact that LBL film deposition is controlled kinetically due to a great change in entropy during molecule adsorption. The influence of adsorption time and ionic strength of the solutions on the film growth and thermal properties has been studied [19]. Polyelectrolyte multilayers were found to have low roughness compared to bare glass, and increasing salt concentration or ionic strength led to thicker layers. Increased salt concentration results in increased electrostatic shielding between polyelectrolytes in solution. This implies that polyelectrolytes in high-ionic-strength solvents form denser structures with loops as opposed to more loose structures when no or less salt is added [18]. In turn, polyelectrolytes adsorb on the surface conserving the same dense structure which leads to more polyelectrolyte mass per unit area, i.e., thicker layers. Increased ionic strength also shields the attractive forces that exist between oppositely charged surface and polyelectrolyte molecules. Layer thickness and roughness have been found to increase nonlinearly with salt concentration. This means that there should be a critical salt concentration that can be added without causing too many defects or too much roughness. Additionally, polyelectrolytes often contain weak ionic groups such as amine and carboxyl groups whose charges strongly depend on the aqueous solution pH, i.e., it depends on whether the pH is above or below the polyelectrolyte isoelectric point. At a pH close to the isoelectric point, the polyelectrolyte will be uncharged and will adsorb as a relatively thick layer due to water becoming a poorer solvent and due to formation of loops. However, at a pH either well above or below the isoelectric point, the polyelectrolyte will be fully charged, and will adsorb as a thinner layer due to stronger interactions between the surface and the corresponding polyelectrolyte. Polyelectrolyte concentrations also affect the deposited layer structure. Higher concentrations may result in larger surface charge reversal, adsorption time reduction, and thicker layer due to chain contraction and loop formation. Drying steps in between layer deposition may also affect layer thickness due to swelling decrease, especially for hygroscopic molecules.

10 Nanostructured Flexible Materials

311

As demonstrated by Schmitt and others later, the internal structure of these polyelectrolyte films has often been studied with X-ray and neutron reflectivity and was found to be characterized by a fuzzy internal structure [20]. Although these polyelectrolyte films have a well-defined supramolecular internal structure; bilayer interpenetration is known to occur. Schmitt reported that about 50% of a given layer was in fact overlapping with neighboring layers, the layers close to the substrate being thinner than those closer to the film/air interface. Anionic and cationic layers have been found to have different thicknesses, again depending on electrostatic interaction shielding. Inorganic counterions, such as Na+ and Cl, from the solutions may also be adsorbed into the multilayers.

10.3.3 LBL Advantages As an alternative to LB films, the LBL process is simpler because the adsorption occurs spontaneously in solution and is kinetically controlled. Therefore, adsorption rates and film buildup are generally faster for the LBL process, usually on the order of several minutes per layer, although it has been shown to be on the order of hours for gold colloids [8]. Generally, LBL involves stronger intermolecular interactions such as long-range electrostatic interactions, leading to stable films and defect-free when amorphous polyelectrolyte layers are used. There is no restriction on substrate size and geometry. The increased range of materials that may be deposited using the LBL versus the LB method is appropriate for multiple applications ranging from electronic to photonic and biomedical devices. For instance, the use of proteins and DNA allowed the fabrication of coatings and devices useful for biomedical applications. The use of conducting polymers and magnetic nanoparticles shows promises for electronic and photonic devices. While LB films are mostly used as an interface in devices and to mimic biological membranes, LBL multilayers have had impact in various fields of research.

10.4 Metal RubberTM Manufacturing Molecular-level self-assembly processing can be used to form thin multilayer, multi-constituent coatings on the surfaces of a variety of substrate materials. For use as sensors, however, the mechanical, thermal, chemical, and electrical properties of the substrate may interfere with sensor transduction coefficient, minimum detectable signal, dynamic range, noise, and interference. Free-standing rather than rigidly attached materials may be self-assembled using similar processes and offer improved performance for sensing applications. Free-standing selfassembled materials may be formed by first coating a substrate surface with a chemical release layer and then alternately self-assembling anionic and cationic molecules until the desired thickness is obtained. The release layer may be etched away using an appropriate solvent, and a free-standing piece of material results.

312

C. Jullian et al.

Important to this manufacturing process are (1) the selection of release layer, (2) the time required to fabricate a coating of desired thickness, and (3) the resulting properties of the material formed. The release layer material must be capable of being etched using a solvent other than the one used during the selfassembly process itself. For example, release materials that can be etched in concentrated acids but not water would be appropriate for aqueous solutionbased self-assembly. The release layer needs to be applied uniformly over the substrate, although the choice of the substrate material is relatively unimportant. A low-cost material with good surface uniformity and the ability to withstand the chemical etchant used to dissolve the release layer is usually chosen. The substrate material itself may be discarded after manufacturing, or cleaned and re-used multiple times. The immersion time required for each individual monolayer and thus for the entire Metal RubberTM material is also important. This is because a sufficient number of bilayers need to be deposited in order to make a thick enough piece of material so that it is mechanically robust. Long immersion times per deposited bilayer mean that the total manufacturing time may be as long as several days. Typically, materials several hundred microns thick are formed, and production times on the order of one full day are used. The deposition time for a single monolayer obeys a normal S-curve. During the first part of the deposition, initiation of molecular coverage of the surface occurs, and the effective rate of molecules deposited per unit time is low. During the middle part of the deposition, the number of molecules deposited per unit time increases to a maximum rate. This rate decreases as saturation is achieved, and the number of additional molecules deposited to the layer is small. The key to reducing the width of this S-curve, and thus minimizing the amount of time required for the formation of a single monolayer, is to reduce the time needed for surface coverage initiation. This is a function of the molecules to be deposited, the property of the substrate, solution concentration, and others. A reduction of deposition to less than 1 min is required in order to synthesize practical materials in a reasonable time period. Another important aspect of Metal RubberTM material production is the use ofan automated manufacturing system rather than the hand dipping of substrates. Automated substrate dipping allows exact control over substrate submersion times, precise counting of the number of bilayers deposited, and continuous operation for longer than a normal work day.

10.5 Metal RubberTM Material Properties The combined mechanical, electrical, and other constitutive properties of Metal RubberTM fabricated by the above process are different from those of other materials. For sensor applications, one combination of properties that is interesting is mechanical modulus and electrical conductivity. This combination is

10 Nanostructured Flexible Materials

313

achieved by alternately self-assembling monolayers of metal nanoclusters with polymers. The mechanical modulus of the Metal RubberTM material may be controlled by varying the polymer molecules used to form every other monolayer. Since any polymer molecule that can be made water soluble, or soluble in a solvent other than water, can be self-assembled by this process, the possible range of achievable modulus properties is large. In general, modulus is controlled by the nature of the particular polymer and its molecular weight. Modulus values from approximately 0.1 MPa to 1 GPa have been obtained using different polymers as the base for Metal RubberTM, and different associated processing conditions. Electrical conductivity is achieved by incorporating electrically conducting metal nanoclusters and polymers as the opposite monolayers in each bilayer of Metal RubberTM material. Because the metal nanoclusters do not form the entire material but only fraction of it, conductivity does not occur in Metal RubberTM the same way that it does in solid metal conductors. Conductivity in solid metals is usually modeled as the transport of negatively charged electrons through a lattice of positively charged atomic nuclei. The nuclei are located at lattice points that are regularly spaced in three dimensions according to the specific crystal structure of the metal. A potential difference or voltage V applied between two points separated by a distance d in such a metal produces an electric field E, with an amplitude equal to V/d, and a vector direction oriented between the points where the potential is applied. The free electrons experience a force F with a magnitude given by F = qE, where q is the charge on the electron and E is the magnitude of the electric field, and move and accelerate in response to the force. The electrons accelerate through the lattice, collide with the fixed lattice nuclei, and are scattered from their original trajectory. Once scattered, they are again accelerated by the field, and again collide. The average electron speed in the direction of the field is defined as the drift velocity. As the temperature of the metal increases, the vibrational energy of the lattice increases, the fixed lattice points present larger physical targets to the moving electrons, and the drift velocity decreases. Since electrical current is proportional to the product of charge and velocity, this means that at higher temperature, current through the material decreases. Electrical resistance is proportional to the voltage V divided by the current, so as temperature increases, the resistance of normal metals increases. The mechanism for electrical conductivity of Metal RubberTM is very different from that for bulk solid metals. Metal RubberTM can be modeled as a polymer matrix that contains a certain volume percentage of electrically conducting inclusions. Conventional polymer blends – different from Metal RubberTM but similar in constituents – that contain conducting metal particles become conductive through their bulk when the volume percentage of such clusters in the material is increased above the percolation limit. At percolation, there are just enough particles to form a continuous interconnected path through the material and allow electron transport from one side of the material to the other. The onset of percolation in such materials is different for

314

C. Jullian et al.

conducting inclusions of different shapes. For spherical inclusions, typically tens of volume percent of filler is required. Conductivity in Metal RubberTM is conceptually similar to this behavior with several important differences. First, the volume percentage of metal nanoclusters required to achieve conductivity percolation has been observed to be less than 0.01%, or more than three orders of magnitude smaller than that needed in particle-filled polymer blends. Second, the approximately 10 nm diameter of the nanoclusters incorporated into Metal RubberTM is smaller than the particle size in most particle-filled systems. This may lead to conductivity effects that are controlled by the quantum confinement properties of the clusters and not the bulk transport properties of micron-sized particles. The size uniformity of the particles may also have an influence; the metal nanoclusters used to form Metal RubberTM are approximately monodisperse rather than randomly sized. Third, the spatial uniformity of materials fabricated by the ESA deposition process suggests that percolation may occur for a small increase in the local volume percentage of nanoclusters. Conductivity in Metal RubberTM may be modeled as being due to electron hopping or quantum mechanical tunneling between adjacent nanoclusters rather than electron transport between larger metal particles at positions where they physically touch. This suggests that the uniformity of (1) the molecular weight of the polymers, (2) the dimensions and shapes of the nanoclusters, and (3) the thicknesses of each of the monolayers in the total nanocomposite material is important to electron transfer characteristics.

10.6 Metal RubberTM Strain Sensors Metal RubberTM materials having the combined elastomeric and electrical conductivity properties described above can be used as strain sensors. The function of the sensors is to respond to mechanical force and produce a change in electrical resistance which can be measured using conventional instrumentation methods. The transfer function, F, of Metal RubberTM as a sensor element can be expressed as F¼

@R @e @R ¼ ; @e @ @

or the change in the electrical resistance of the sensor element that occurs as a function of applied stress. There are two parts to this function, namely the change in resistance that occurs due to strain or elongation of the material, and the change in strain or displacement of the material that occurs due to stress. The latter quantity is the slope of the strain/stress curve, or the inverse of the elastic modulus of the material, 1/E, in units of 1/Pa, or m2/N. A low modulus for Metal RubberTM sensor materials is desirable because proportional decreases in modulus produce equally proportional increases in sensor sensitivity. Thus,

10 Nanostructured Flexible Materials

315

Metal RubberTM materials with modulus on the order of 0.1 MPa should have a sensitivity to strain that is many orders of magnitude larger than similar materials with a modulus of 1 GPa, assuming that all other material properties are identical. The first term in the transfer function expression is related to how the individual nanoclusters physically separate within the material due to strain and how that separation affects the hopping or tunneling of electrons between all of the nanoclusters and the resulting resistance of the material. This is complicated due to the multilayered self-assembly geometry of the Metal RubberTM material and Poisson effects that couple multiple geometrical displacement effects in the material. The layer-by-layer geometry of Metal RubberTM inherently leads to anisotropic properties. Without special design, the electrical conductivity of the material is higher in the two-dimensional plane than through the thickness. This suggests that the uniformity of nanocluster separation in the plane of each of the deposited monolayers is perhaps more regular than the mean separation distance between clusters deposited in alternating cluster monolayers. When strained axially in a direction in the two-dimensional plane of the sheet material, the nanoclusters separate and electrical resistance increases. Figure 10.3 shows a typical result of a piece of Metal RubberTM material being strained axially and the measurement of the end-to-end resistance of the material measured as strain is increased. Metal RubberTM strain sensor materials have been strained to more than 1,000% without failure and cycled at 10% strain to more than 100,000 cycles without fatigue failure. ‘‘Metal RubberTM’’ represents a manufacturing technology

Resistance ( Ω) 90

Stress (psi)

500

70

50

0

2

4

6 Strain %

Fig. 10.3 Resistance versus strain response of Metal RubberTM material

8

30 10

316

C. Jullian et al.

and not a single material having a single set of material properties. Thus, by changing the properties of the Metal RubberTM material, the transduction properties of Metal RubberTM sensors may be affected. Specifically, the hysteresis properties may be reduced by adjusting the viscoelastic properties of the polymer used during LBL production.

10.7 Summary This chapter has summarized self-assembly processes, the use of such processes to produce Metal RubberTM nanocomposite materials, and the use of Metal RubberTM for the sensing of mechanical strain.

References 1. Langmuir, I., The constitution and fundamental properties of solids and liquids. II. Liquids. J. Am. Chem. Soc., 1917. 39: pp. 1848–1906. 2. Langmuir, I., The mechanism of the surface phenomena of flotation. Trans. Faraday Soc., 1920. 15: pp. 62–74. 3. K. S. V. Instruments LTD, Langmuir and Langmuir-Blodgett films, Application Note #117. 4. Hann, R. A., Molecular Structure Monolayer Properties, in Langmuir-Blodgett Films, ed. G. Roberts, 1990, Plenum Press. 5. Hamachi, I.; Noda, S.; Kunitake, T., Layered arrangement of oriented myoglobins in cast films of a phosphate bilayer membrane. J. Am. Chem. Soc., 1990. 112: pp. 6744–6745. 6. Prime, K. L.; Whitesides, G. M., Adsorption of proteins onto surfaces containing endattached oligo(ethylene oxide): A model system using self-assembled monolayers. J. Am. Chem. Soc., 1993. 115: pp. 10714–10721. 7. Ventra, M. D.; Evoy, S.; Heflin, J. R. Jr., Introduction to Nanoscale Science and Technology, ed. M. D. Ventra; S. Evoy; J. R. Heflin Jr., 2004, Springer. 8. Decher, G.; Hong, J. D.; Schmitt, J., Buildup of ultrathin multilayer films by self-assembly process: III. Consecutively alternating adsorption of anionic and cationic polyelectrolytes on charged surfaces. Thin Solid Films, 1992. 210–211: pp. 831–835. 9. Iler, R. K., Multilayers of colloidal particles. J. Colloid Inter. Sci , 1966. 21: pp. 569–594. 10. Liu, Y.; Claus, R. O., Layer-by-layer electrostatic self-assembly of nanoscale Fe3O4 particles and polyimide precursor on silicon and silica surfaces. Appl. Phys. Lett., 1997. 71: pp. 2265–2267. 11. Wang, Y.-X.; Du, W.; Spillman, W. B., Jr.; Claus, R. O., Biocompatible thin film coatings fabricated using the electrostatic self-assembly process. Proc. SPIE, 2001: 4265: pp. 142–151. 12. Lvov, Y.; Decher, G.; Sukhorukov, G., Assembly of thin films by means of successive deposition of alternate layers of DNA and poly(allylamine). Macromolecules, 1993. 26: pp. 5396–5399. 13. Keller, S. W.; Kim, H.-N.; Mallouk, T. E., Layer-by-layer assembly of intercalation compounds and heterostructures on surfaces: Toward molecular ‘‘Beaker’’ epitaxy. J. Am. Chem. Soc., 1994. 116: pp. 8817–8818. 14. Feldheim, D. L.; Grabar, K. C.; Natan, M. J.; Mallouk, T. E., Electron transfer in selfassembled inorganic polyelectrolyte/metal nanoparticle heterostructures. J. Am. Chem. Soc., 1996. 118: pp. 7640–7641.

10 Nanostructured Flexible Materials

317

15. Liu, Y.; Wang, Y.; Y.-X.; Claus, R. O., Layer-by-layer ionic self-assembly of Au colloids into multilayer thin-films with bulk metal conductivity. Chem. Phys. Lett., 1998. 298: pp. 315–319. 16. Eckle, M.; Decher, G., Tuning the performance of layer-by-layer assembled by controlling the position of isolating clay barrier sheets. Nano Lett., 2001. 1: pp. 45–49. 17. Kunitake, T., Anisotropic incorporation of functional molecules and synthesis of lowdimensional clusters in cast multibilayer films. Mol. Cryst. Liq. Cryst., 1994. 255: pp. 7–16. 18. Malmsten, M., Biopolymers at Interfaces, ed. M. Malmsten, 2003, Marcel Dekker, Inc. 19. Lvov, Y.; Decher, G.; Mohwald, H., Assembly, structural characterization, and thermal behavior of layer-by-layer deposited ultrathin films of poly(vinyl sulfate) and poly(allylamine). Langmuir, 1993. 9: pp. 481–486. 20. Schmitt, J.; Grunewald, T.; Decher, G.; Pershan, P. S.; Kjaer, K.; Losche, M., Internal structure of layer-by-layer adsorbed polyelectrolyte films: A neutron and X-ray reflectivity study. Macromolecules, 1993. 26: pp. 7058–7063.