Water Properties in Food, Health, Pharmaceutical and Biological Systems: ISOPOW 10

WATER PROPERTIES IN FOOD, HEALTH, PHARMACEUTICAL AND BIOLOGICAL SYSTEMS: ISOPOW 10

Edited by DAVID S. REID TANABOON SAJJAANANTAKUL PETER J. LILLFORD SANGUANSRI CHAROENREIN

I

S O

P A John Wiley & Sons, Inc., Publication

O

W

WATER PROPERTIES IN FOOD, HEALTH, PHARMACEUTICAL AND BIOLOGICAL SYSTEMS: ISOPOW 10

WATER PROPERTIES IN FOOD, HEALTH, PHARMACEUTICAL AND BIOLOGICAL SYSTEMS: ISOPOW 10

Edited by DAVID S. REID TANABOON SAJJAANANTAKUL PETER J. LILLFORD SANGUANSRI CHAROENREIN

I

S O

P A John Wiley & Sons, Inc., Publication

O

W

Edition first published 2010 © 2010 Blackwell Publishing Chapter 18 and 28 copyrights held by Anne-Marie Hermansson. Blackwell Publishing was acquired by John Wiley & Sons in February 2007. Blackwell’s publishing program has been merged with Wiley’s global Scientific, Technical, and Medical business to form Wiley-Blackwell. Editorial Office 2121 State Avenue, Ames, Iowa 50014-8300, USA For details of our global editorial offices, for customer services, and for information about how to apply for permission to reuse the copyright material in this book, please see our website at www.wiley.com/ wiley-blackwell. Authorization to photocopy items for internal or personal use, or the internal or personal use of specific clients, is granted by Blackwell Publishing, provided that the base fee is paid directly to the Copyright Clearance Center, 222 Rosewood Drive, Danvers, MA 01923. For those organizations that have been granted a photocopy license by CCC, a separate system of payments has been arranged. The fee codes for users of the Transactional Reporting Service are ISBN-13: 978-0-8138-1273-1/2010. Designations used by companies to distinguish their products are often claimed as trademarks. All brand names and product names used in this book are trade names, service marks, trademarks or registered trademarks of their respective owners. The publisher is not associated with any product or vendor mentioned in this book. This publication is designed to provide accurate and authoritative information in regard to the subject matter covered. It is sold on the understanding that the publisher is not engaged in rendering professional services. If professional advice or other expert assistance is required, the services of a competent professional should be sought. Library of Congress Cataloguing-in-Publication Data International Symposium on the Properties of Water (10th : 2007 : Bangkok, Thailand) Water properties in food, health, pharmaceutical and biological systems : ISOPOW 10 / edited by David Reid, Tanaboon Sajjaanantakul, et al. p. cm. Includes bibliographical references and index. ISBN-13: 978-0-8138-1273-1 (alk. paper) ISBN-10: 0-8138-1273-9 (alk. paper) 1. Food–Water activity–Congresses. 2. Food–Moisture–Congresses. 3. Pharmaceutical chemistry– Congresses. I. Reid, David. II. Sajjaanantakul, Tanaboon. III. Title. TX553.W3I57 2007 664–dc22 2008054012 A catalog record for this book is available from the U.S. Library of Congress. Set in 10/12 pt Times by Toppan Best-set Premedia Limited Printed in Singapore Disclaimer The publisher and the author make no representations or warranties with respect to the accuracy or completeness of the contents of this work and specifically disclaim all warranties, including without limitation warranties of fitness for a particular purpose. No warranty may be created or extended by sales or promotional materials. The advice and strategies contained herein may not be suitable for every situation. This work is sold with the understanding that the publisher is not engaged in rendering legal, accounting, or other professional services. If professional assistance is required, the services of a competent professional person should be sought. Neither the publisher nor the author shall be liable for damages arising herefrom. The fact that an organization or Website is referred to in this work as a citation and/or a potential source of further information does not mean that the author or the publisher endorses the information the organization or Website may provide or recommendations it may make. Further, readers should be aware that Internet Websites listed in this work may have changed or disappeared between when this work was written and when it is read. 1

2010

Table of Contents

Preface Editorial Note Acknowledgments Contributors PART 1 Session 1:

Invited Speakers and Oral Presentations Water Mobility/Dynamics and Its Application in Food and Pharmaceutical Systems

xiii xv xvii xix 3

5

Invited Speakers 1. Complementary Aspects of Thermodynamics, Nonequilibrium Criteria, and Water Dynamics in the Development of Foods and Ingredients M. P. Buera 2. Water Mobility in Solid Pharmaceuticals as Determined by Nuclear Magnetic Resonance, Isothermal Sorption, and Dielectric Relaxation Measurements S. Yoshioka and Y. Aso

9

25

Oral Presentations 3. The Effect of Water and Fat Contents on the Enthalpy of Dissolution of Model Food Powders: A Thermodynamic Insight A. Marabi, A. Raemy, A. Burbidge, R. Wallach, and I. S. Saguy

41

4. “Solvent Water” Concept Simplifies Mathematical Modeling in Fermenting Dough, a Multiphase Semisolid Food S. M. Loveday and R. J. Winger

49

5. Microdomain Distribution in Food Matrices: Glass Transition Temperature, Water Mobility, and Reaction Kinetics Evidence in Model Dough Systems Y. Kou

59

v

vi

Table of Contents

Session 2: Water Essence and the Stability of Food and Biological Systems

67

Invited Speakers 6. Effect of Combined Physical Stresses on Cells: The Role of Water J.-M. Perrier-Cornet, M. Moussa, H. Simonin, L. Beney, and P. Gervais 7. Soft Condensed Matter: A Perspective on the Physics of Food States and Stability T. P. Labuza, T. J. Labuza, K. M. Labuza, and P. S. Labuza 8. Antiplasticization of Food Polymer Systems by Low Molecular Mass Diluents C. C. Seow

71

87

115

Oral Presentations 9. Freeze Drying of Lactobacillus coryniformis Si3: Focus on Water Å. Schoug, J. Schnürer, and S. Håkansson

141

10. Water-Sorption Properties and Stability of Inclusion Complexes of Thymol and Cinnamaldehyde with β-Cyclodextrins P. A. Ponce, M. P. Buera, and B. E. Elizalde

149

11. Beyond Water: Waterlike Functions of Other Biological Compounds in a Waterless System B. R. Bhandari

157

12. Water Sorption and Transport in Dry, Crispy Bread Crust M. B. J. Meinders, N. H. van Nieuwenhuijzen, R. H. Tromp, R. J. Hamer, and T. van Vliet

165

13. Water State and Distribution During Storage of Soy Bread with and without Almond A. Lodi and Y. Vodovotz

175

14. Phase Separation of Ice Crystals in Starch-Based Systems During Freezing and Effects on Moisture Content and Starch Glass Transition T. Tran, K. Piyachomkwan, and K. Sriroth

185

15. Carrot Fiber as a Carrier in Spray Drying of Fructose K. Cheuyglintase and K. R. Morison

191

Session 3:

199

Microstructured and Nanostructured Changes in Food

Invited Speakers 16. Taking the Measure of Water D. S. Reid

203

Table of Contents

17. Rehydration Modeling of Food Particulates by Using Principles of Water Transport in Porous Media I. S. Saguy, O. Troygot, A. Marabi, and R. Wallach 18. Protein Hydration in Structure Creation P. J. Lillford and A.-M. Hermansson

vii

219 237

19. Water Partitioning in Colloidal Systems as Determined by Nuclear Magnetic Resonance P. Chinachoti and P. Chatakanonda

251

20. Physical Changes in Confectionery Products Caused by the Availability of Water, with a Special Focus on Lactitol Crystallization M. H. Lim, B. Lampen, L. F. Siow, and T. Rades

271

Oral Presentations 21. Entrapment of Probiotic Bacteria in Frozen Cryoprotectants and Viability in Freeze Drying Y. H. Roos and K. S. Pehkonen

285

22. Fracture Behavior of Biopolymer Films Prepared from Aqueous Solutions 291 I. Yakimets, S. S. Paes, N. Wellner, and J. R. Mitchell Session 4:

Biomaterial Sciences: Water in Stability and Delivery of Active Biomolecules

297

Invited Speakers 23. The Plasticization-Antiplasticization Threshold of Water in Microcrystalline Cellulose: A Perspective Based on Bulk Free Volume S. P. Chamarthy, F. X. Diringer, and R. Pinal 24. Understanding the Role of Water in Nonaqueous Pharmaceutical Systems B. D. Anderson, S. S. Rane, and T.-X. Xiang

301 315

25. Crystallization, Collapse, and Glass Transition in Low-Water Food Systems Y. H. Roos

335

26. Carbohydrates in Amorphous States: Molecular Packing, Nanostructure, and Interaction with Water J. Ubbink

353

27. Ice Crystallization in Gels and Foods Manipulated by the Polymer Network N. Murase, S. Yamada, and N. Ijima

373

28. Marine-Inspired Water-Structured Biomaterials A.-M. Hermansson, P. Olofsson, S. Ekstedt, M. Pihl, and P. Gatenholm

385

viii

Table of Contents

PART 2 Poster Presentations

397

Session 5: Role of Water Mobility/Dynamics in Food and Pharmaceutical Systems

399

29. Another Unusual Property of Water: It Increases the Glass Transition Temperature of a Glassy Polymer S. P. Chamarthy and R. Pinal

401

30. Molecular Mobility Interpretation of Water-Sorption Isotherms of Food Materials by Means of Gravimetric Nuclear Magnetic Resonance W. P. Weglarz, M. Witek, C. Inoue, H. Van As, and J. van Duynhoven

411

31. Kinetics of Enthalpy Relaxation in Corn Syrup–Sucrose Mixtures B. R. Bhandari and R. W. Hartel 32. Development of a Novel Phase Transition Measurement Device for Solid Food Materials: Thermal Mechanical Compression Test (TMCT) Y. Liu, P. Intipunya, T. T. Truong, W. Zhou, and B. R. Bhandari

419

429

33. Proton Nuclear Magnetic Resonance Studies of Molecular Mobility in Potato Systems in Relation to Nonenzymatic Browning N. C. Acevedo, C. Schebor, and M. P. Buera

437

34. Nonenzymatic Browning Reaction and Enthalpy Relaxation of Glassy Foods K. Tsuji, K. Kawai, M. Watanabe, and T. Suzuki

445

35. Film-Forming Ability of Duck Egg White and Its Water-Vapor Barrier Property W. Garnjanagoonchorn, A. Yimjaroenpornsakul, N. Poovarodom, and S. Praditdoung 36. Water-Vapor Permeability of Chitosan and Methoxy Poly(ethylene glycol)-b-poly(ε-caprolactone) Blend Homogeneous Films N. Niamsa, N. Morakot, and Y. Baimark 37. Ice Formation in Concentrated Aqueous Glucose Solutions P. Thanatuksorn, K. Kajiwara, N. Murase, and F. Franks

453

459 465

38. Effects of Sodium and Potassium Ions on the Viscosities in the Sodium/Potassium-Glucose-Water Ternary System M. Soga, K. Kurosaki, and K. Kajiwara

473

39. Comparison of Water Sorption and Crystallization Behaviors of Freeze-Dried Lactose, Lactitol, Maltose, and Maltitol K. Jouppila, M. Lähdesmäki, P. Laine, M. Savolainen, and R. A. Talja

477

Table of Contents

40. Sorption Behavior of Extruded Rice Starch in the Presence of Glycerol J. Enrione, S. Hill, J. R. Mitchell, and F. Pedreschi 41. Water State and Mobility Affect the Mechanical Properties of Coffee Beans P. Pittia, G. Sacchetti, P. Rocculi, L. Venturi, M. Cremonini, and M. Dalla Rosa

ix

483

491

42. Effect of Water Activity on the Release Characteristics of Encapsulated Flavor A. Soottitantawat, H. Yoshii, and T. Furuta

499

43. Water and Protein Modifier Effects on the Phase Transitions and Microstructure of Mung-Bean Starch Granules P. Hongsprabhas and K. Israkarn

507

44. Evaluation of the Disintegration and Diffusion of Pharmaceutical Solid Matrices by Image Processing and Nonlinear Dynamics D. I. Téllez-Medina, A. Ortíz-Moreno, J. J. Chanona-Pérez, L. Alamilla-Beltrán, and G. F. Gutiérrez-López

Session 6:

Properties and Stability of Food and Biological Systems

515

523

45. Effect of Water Content on Physical Properties of Potato Chips F. Pedreschi and P. Moyano

525

46. Predicting Water Migration in Starchy Food During Cooking S. Thammathongchat, M. Fukuoka, T. Hagiwara, T. Sakiyama, and H. Watanabe

533

47. Nonenzymatic Browning May Be Inhibited or Accelerated by Magnesium Chloride According to the Level of Water Availability and Saccharide-Specific Interactions P. R. Santagapita, S. B. Matiacevich, and M. P. Buera

539

48. Combined Effect of Cinnamon Essential Oil and Water Activity on Growth Inhibition of Rhizopus stolonifer and Aspergillus flavus and Possible Application in Extending the Shelf Life of Bread S. Nanasombat, N. Piumnoppakun, D. Atikanbodee, and M. Rattanasuwan

545

49. From Water to Ice: Investigation of the Effect of Ice Crystal Reduction on the Stability of Frozen Large Unilamellar Vesicles L. F. Siow, T. Rades, and M. H. Lim

551

50. Does Microencapsulation Improve Storage Stability of Cloudberry (Rubus chamaemorus) Ellagitannins? P. Laine, P. Kylli, M. Heinonen, and K. Jouppila

563

x

Table of Contents

51. Nonenzymatic Browning Reaction of Glassy Foods: Characterization of Local Reactions Independent of the Glassy Matrix K. Kawai, T. Suzuki, and K. Kajiwara

571

52. Physical Properties of Protein-Carbohydrate Sheets Produced by a Twin-Screw Extruder R. A. Talja, K. S. Pehkonen, K. Jouppila, and Y. H. Roos

577

53. Thermal Transitions, Mechanical Properties, and Molecular Mobility in Cornflakes as Affected by Water Content A. Farroni, S. B. Matiacevich, S. Guerrero, S. Alzamora, and M. P. Buera 54. Texture of Glassy Tapioca-Flour–Based Baked Products as a Function of Moisture Content R. Kulchan, P. Suppakul, and W. Boonsupthip 55. Effects of Excipients on the Storage Stability of Freeze-Dried Xanthine Oxidase P. Srirangsan, K. Kawai, N. Hamada-Sato, M. Watanabe, and T. Suzuki 56. Water Properties in Bread Produced with an Innovative Mixer E. Curti, E. Vittadini, A. Di Pasquale, L. Riviera, F. Antoniazzi, and A. Storci 57. Evaluation of Deformation and Shrinking of Potato Slabs During Convective Drying R. Campos-Mendiola, C. Gumeta-Chávez, J. J. Chanona-Pérez, L. Alamilla-Beltrán, A. Jiménez-Aparicio, and G. F. Gutiérrez-López 58. Effects of Different Cut-Induced Microstructural and Macrostructural Arrays on Convective Drying of Agave atrovirens Karw C. Gumeta-Chávez, J. J. Chanona-Pérez, L. Alamilla-Beltrán, G. Calderón-Domínguez, A. Vega, P. Ligero, J. A. Mendoza-Pérez, and G. F. Gutiérrez-López 59. Study of White-Bread Structural Evolution by Means of Image Analysis and Associated Thermal History and Water-Loss Kinetics A. Pérez-Nieto, J. J. Chanona-Pérez, G. Calderón-Domínguez, R. Farrera-Rebollo, L. Alamilla-Beltrán, and G. F. Gutiérrez-López

583

591

599

605

613

619

627

60. Effect of Hydrothermal Treatment on the Rheological Properties of High-Amylose Rice Starch P. Khunae, T. Tran, and P. Sirivongpisal

635

61. Influence of Glass Transition on Oxygen Permeability of Starch-Based Edible Films D. Thirathumthavorn, S. Charoenrein, and J. M. Krochta

641

Table of Contents

62. Molecular Mobility and Seed Longevity in Chenopodium quinoa M. Castellión, S. Maldonado, and M. P. Buera 63. Analyzing the Effect of Freeze-Thaw Cycle on the Off-Aroma of Pineapple by Using an Electronic Nose Technique S. Charoenrein and T. Kaewtathip 64. Water Uptake and Solid Loss During Soaking of Milled Rice Grains P. Chatakanonda and K. Sriroth 65. Microstructural, Physical, and Rehydration Properties of Maltodextrin Powders Obtained by Spray Drying A. L. Muñoz-Herrera, V. Tejeda-Hernández, A. Jiménez-Aparicio, J. Welti-Chanes, J. J. Chanona-Pérez, L. Alamilla-Beltrán, and G. F. Gutiérrez-López 66. Nanostructures and Minimum Integral Entropy as Related to Food Stability L. A. Pascual-Pineda, E. Flores-Andrade, C. I. B. Guevara, L. Alamilla-Beltrán, J. J. Chanona-Pérez, E. Azuara-Nieto, and G. F. Gutiérrez-López Index

xi

647

657 663

673

681

689

Preface

Water plays an important role in the structure, functionality, and stability of food and biomaterials. The ubiquitous water molecules are small and simple, yet they possess unusual properties and develop complex interactions with surrounding molecules and compounds. An increasing understanding of water properties and their significance in interacting and regulating chemical and biological systems has led to in-depth research to better understand water ’s role in food structure and stability. Water-sorption isotherms of foods were first published in 1943, and the concept of water activity as a major control variable in food spoilage was introduced in 1953. ISOPOW—the International Symposium on the Properties of Water—was first organized in Glasgow, Scotland, in 1974 (see the Editorial Note for details) to promote the exchange of knowledge between scientists in the field of food science and scientists whose interests in water derived from different disciplines. Since then, ISOPOW has become an important focal point for scientific presentations and discussion on water properties, such as water activity, aqueous glass transitions, and water mobility as related to food, pharmaceutical, biological, and biomaterial systems. This volume is based on lectures, oral presentations, and posters presented at the 10th ISOPOW in Bangkok, Thailand, on 2–7 September 2007. The title Water Properties in Food, Health, Pharmaceutical and Biological Systems: ISOPOW 10 emphasizes the context of research findings presented at the symposium. Part 1 of the book is from full manuscripts of invited lectures and oral presentations and is divided into four sessions. Session 1 deals with water dynamics and its application in food and pharmaceutical systems, with some examples in food powders and dough systems. Session 2 involves water and its influence on food and biological systems. Soft condensed matter, antiplasticization of food polymers, and physical stress on cells are among the topics presented. Session 3 examines the microstructured and nanostructured changes in food, including the measurement of water properties, rehydration modeling of food particulates, water in colloids, and examples of water ’s effects in confectionary products. Session 4 discusses biomaterial science aspects of water, such as its properties considered by bulk free volume concepts and its role in nonaqueous pharmaceutical systems. The behavior of water in phase transition, molecular packing, and nanostructure in food systems, along with water in structured biomaterials (as elucidated by marine jellyfish), is discussed in the session. Part 2 of the book has been compiled from research posters presented at the symposium in two sessions. The role of water mobility/dynamics in various food products xiii

xiv

Preface

and systems is presented in Session 5. Session 6 represents different aspects of current research on the understanding of chemical and physical changes in food and food stability control as affected by water. These two sessions reflect the vast array of investigations and applications of water worldwide. It is hoped that these proceedings will provide a useful reference on water, its properties, and applications to the scientific community according to the spirit of ISOPOW.

Editorial Note

ISOPOW (International Symposium on the Properties of Water) is a nonprofit scientific organization and a standing committee under the International Union of Food Science and Technology (IUFoST). The first ISOPOW was organized in Glasgow, Scotland, in 1974 through the initiative of Dr. Ron B. Duckworth and Dr. Louis Rockland. Their goals were to present the state of knowledge on water and its application in food science and related disciplines, to organize meetings and stimulate discussions between academic and industrial scientists, to bring together participants under conditions conducive to the greatest interactions, and to publish symposium proceedings of high scientific quality. In addition to food scientists, biological and pharmaceutical scientists have recognized that water plays an important role influencing structure, functionality, and stability of biomaterials. ISOPOW symposia always include delegates from other fields for cross-understanding and multidisciplinary approaches to the study of water. Each symposium provides multiple opportunities for speakers and participants to share perspectives, address challenges, and develop collaborations to advance understanding in the field of water properties. ISOPOW meetings, held in various locations, reflect the worldwide dimension of ISOPOW and the interdisciplinary characteristics of the subject. Most of the meetings have resulted in the publication of books of the proceedings. Previous ISOPOWs were ISOPOW 1 Glasgow, UK, 1974 ISOPOW 2 Osaka, Japan, 1978 ISOPOW 3 Beaune, France, 1983 ISOPOW 4 Banff, Canada, 1987 ISOPOW Practicum I Penang, Malaysia, 1987 ISOPOW 5 Peniscola, Spain, 1992 ISOPOW Practicum II Puebla, Mexico, 1994 ISOPOW 6 Santa Rosa, USA, 1996 ISOPOW 7 Helsinki, Finland, 1998 ISOPOW 8 Zichron Yaakov, Israel, 2000 ISOPOW 9 Mar del Plata, Argentina, 2004 The 10th ISOPOW’s success was based on strong support from the ISOPOW Central Committee. Members of the central committee at the time were: xv

xvi

Editorial Note

Dr. David S. Reid, University of California, USA, President Dr. María del Pilar Buera, Universidad de Buenos Aires, Argentina, President-Elect Dr. Louis B. Rockland, FoodTech Research and Development, USA, Honorary President Dr. Imad A. Farhat, Firmenich, Switzerland Dr. Patrick Gervais, Université de Bourgogne, France Dr. Miang Hoong Lim, University of Otago, New Zealand Dr. David Lechuga-Ballesteros, Nektar Therapeutics, USA Dr. Jim Leslie, Consultant, UK Dr. Peter J. Lillford, University of York, UK Dr. Norio Murase, Tokyo Denki University, Japan Dr. Yrjö H. Roos, University College Cork, Ireland Dr. Tanaboon Sajjaanantakul, Kasetsart University, Thailand Dr. Denise Simatos, Université de Bourgogne, France Dr. Jorge Welti-Chanes, Universidad de las Américas, Mexico With suggestions and support from the Central Committee, the scientific program was formulated. The 10th ISOPOW in Bangkok, Thailand included 22 invited lectures from renowned food scientists, pharmacists, and physical chemists from academic and research institutions and related industries. From the abstracts submitted, 18 oral presentations were selected and 58 poster presentations were displayed for discussion. The symposium attracted 120 participants from 25 different countries worldwide. The participation of young scientists is one of the key successes in expanding the knowledge and enhancing the spirit of the ISOPOW meeting. For the 10th ISOPOW, the Central Committee has granted five travel bursaries to assist young scientists in presenting their findings at the symposium. The proceedings of the 10th ISOPOW are the product of the symposium, and all author contributions are thankfully acknowledged. The symposium was made possible by cosponsorship by the Thailand Commission on Higher Education, Kasetsart University, the ISOPOW Central Committee, the National Science and Technology Development Agency (Thailand), the Thailand Convention and Exhibition Bureau, Nestlé (Thai) Ltd., and several international and Thai food industry allies, detailed in the 10th ISOPOW book of abstracts published by the Department of Food Science and Technology, Kasetsart University, at the time of the symposium. David S. Reid Tanaboon Sajjaanantakul Peter J. Lillford Sanguansri Charoenrein

Acknowledgments

It is our pleasure to acknowledge the ISOPOW Central Committee, all session chairpersons, and the local scientific committee for their efforts in reviewing the scientific program and the proceedings. Special thanks are due to Dr. Denise Simatos, Dr. Pilar Buera, Dr. David Reid, and Dr. Peter Lillford for valuable suggestions early in the preparation for ISOPOW 10. Appreciation also goes to all local organizing committees, particularly faculty members and staff of the Department of Food Science and Technology, Kasetsart University, for their due diligence in activities required for the success of the symposium. Kasetsart Food Science’s undergraduate and graduate students provided symposium attendees with a warm welcome and outstanding hospitality. The students’ liveliness and eagerness contributed to a pleasant atmosphere that promoted interaction among the symposium participants. We express our gratitude to Dr. Sanguansri Charoenrein, Dr. Utai Klinkesorn, Dr. Parichat Hongsprabhas, Miss Wasaporn Chanput, and Miss Kunwadee Kaewka of the Department of Food Science and Technology, Kasetsart University, for their assistance in preparation of manuscripts for this volume. Finally, the Central Committee congratulates the local organizers, lead by Dr. Tanaboon Sajjaanantakul, for a delightful, stimulating meeting.

xvii

Contributors

Acevedo, Nuria Cristina Departamento de Industrias, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Buenos Aires, Argentina Alamilla-Beltrán, Liliana Departamento de Ingeniería, Bioquímica y Graduados en Alimentos, Escuela Nacional de Ciencias Biológicas, Instituto Politénico Nacional, México City, México Alzamora, Stella Departamento de Industrias, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Buenos Aires, Argentina Anderson, Bradley D. Department of Pharmaceutical Sciences, College of Pharmacy, University of Kentucky, Lexington, Kentucky, USA Antoniazzi, F. Food Science and Technology, Department of Industrial Engineering, University of Parma, Parma, Italy Aso, Yukio National Institute of Health Sciences, Tokyo, Japan

Atikanbodee, Dusita Department of Applied Biology, Faculty of Science, King Mongkut’s Institute of Technology Ladkrabang, Bangkok, Thailand Azuara-Nieto, Ebner Instituto de Ciencias básicas, Universidad Veracruzana, Xalapa, Vercruz, México Baimark, Yodthong Department of Chemistry, Faculty of Science, Mahasarakham University, Mahasarakham, Thailand Beney, Laurent Laboratoire de Génie des procédés Alimentaires et Biotechnologiques, ENSBANA, Université de Bourgogne, Dijon, France Bhandari, Bhesh R. School of Land, Crop, and Food Sciences, University of Queensland, Brisbane, Australia Boonsupthip, Waraporn Department of Food Science and Technology, Faculty of AgroIndustry, Kasetsart University, Bangkok, Thailand xix

xx

Contributors

Buera, María del Pilar Departamentos de Industrias y de Química Orgánica, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Buenos Aires, Argentina Burbidge, A. Nestlé Research Center, Nestec Ltd., Lausanne, Switzerland Calderón-Domínguez, Georgina Departamento de Ingeniería, Bioquímica y Graduados en Alimentos, Escuela Nacional de Ciencias Biológicas, Instituto Politénico Nacional, México City, México Campos-Mendiola, R. Escuela Nacional de Ciencias Biológicas, Instituto Politénico Nacional, México City, México Castellión, Martina Departamento de Biodiversidad y Biología Experimental, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Buenos Aires, Argentina Chamarthy, Sai Prasanth Department of Industrial and Physical Pharmacy, Purdue University, West Lafayette, Indiana, USA; and Respiratory Product Development, Schering-Plough Research Institute, Summit, New Jersey, USA Chanona-Pérez, Jose Jorge Departamento de Graduados e Investigación en Alimentos, Escuela Nacional de Ciencias Biológicas, Instituto Politénico Nacional, México City, México

Charoenrein, Sanguansri Department of Food Science and Technology, Faculty of AgroIndustry, Kasetsart University, Bangkok, Thailand Chatakanonda, Pathama Kasetsart Agricultural and AgroIndustrial Product Improvement Institute, Kasetsart University, Bangkok, Thailand Cheuyglintase, Kloyjai Rajamongala University of Technology Tunyaburi, Pathumtani, Thailand; and Department of Chemical and Process Engineering, University of Canterbury, Christchurch, New Zealand Chinachoti, Pavinee Faculty of Agro-Industry, Prince of Songkla University, Songkla, Thailand Cremonini, Mauro Department of Food Science, Alma Mater Studiorum, University of Bologna, Campus of Food Science, Cesena, Italy Curti, Elena Food Science and Technology, Department of Industrial Engineering, University of Parma, Parma, Italy Dalla Rosa, Marco Department of Food Science, Alma Mater Studiorum, University of Bologna, Campus of Food Science, Cesena, Italy

Contributors

Di Pasquale, A. Food Science and Technology, Department of Industrial Engineering, University of Parma, Parma, Italy

Flores-Andrade, Enrique Escuela Nacional de Ciencias Biológicas, Instituto Politécnico Nacional, México City, México

Diringer, F. X. Department of Industrial and Physical Pharmacy, Purdue University, West Lafayette, Indiana, USA; and Faculté de Pharmacie, Université Louis Pasteur de Strasbourg, Srasbourg, France

Franks, Felix BioUpdate Foundation, London, UK

Ekstedt, S. SIK/Swedish Institute for Food and Biotechnology, Gothenburg, Sweden Elizalde, Beatriz E. Departamento de Industias, Facultad Ciencias Exactas y Naturales, Universidad de Buenos Aires, Buenos Aires, Argentina Enrione, Javier Departamento de Ciencia y Tecnologías de los Alimentos, Facultad Tecnológica, Universidad de Santiago de Chile, Santiago, Chile

xxi

Fukuoka, Mika Department of Food Science and Technology, Tokyo University of Marine Science and Technology, Tokyo, Japan Furuta, Takeshi Department of Biotechnology, Tottori University, Tottori, Japan Garnjanagoonchorn, Wunwiboon Department of Food Science and Technology, Faculty of AgroIndustry, Kasetsart University, Bangkok, Thailand Gatenholm, Paul Department of Chemical and Biological Engineering, Chalmers University of Technology, Gothenburg, Sweden

Farrera-Rebollo, R. Departamento de Ingeniería, Bioquímica y Graduados en Alimentos, Escuela Nacional de Ciencias Biológicas, Instituto Politénico Nacional, México City, México

Gervais, Patrick Laboratoire de Génie des procédés Alimentaires et Biotechnologiques, ENSBANA, Université de Bourgogne, Dijon, France

Farroni, Abel Instituto Nacional de Tecnología Agropecuaria, Pergamino, Pcia de Buenos Aires, Buenos Aires, Argentina

Guerrero, Sandra Departamento de Industrias, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Buenos Aires, Argentina

xxii

Contributors

Guevara, César Ignacio Beristain Instituto de Ciencias básicas, Universidad Veracruzana, Xalapa, Vercruz, México Gumeta-Chávez, Carolina Escuela Nacional de Ciencias Biológicas, Instituto Politécnico Nacional, México City, México Gutiérrez-López, Gustavo F. Departamento de Graduados e Investigación en Alimentos, Escuela Nacional de Ciencias Biológicas, Instituto Politécnico Nacional, México City, México Hagiwara, Tomoaki Department of Food Science and Technology, Tokyo University of Marine Science and Technology, Tokyo, Japan Håkansson, Sebastian Department of Microbiology, Swedish University of Agricultural Sciences, Uppsala, Sweden

Heinonen, Marina Department of Applied Chemistry and Microbiology, University of Helsinki, Helsinki, Finland Hermansson, Anne-Marie SIK/Swedish Institute for Food and Biotechnology, Gothenburg, Sweden; and Department of Chemical and Biological Engineering, Chalmers University of Technology, Gothenburg, Sweden Hill, Sandra Division of Food Sciences, School of Biosciences, University of Nottingham–Sutton Bonington Campus, Leicester, UK Hongsprabhas, Parichat Department of Food Science and Technology, Faculty of AgroIndustry, Kasetsart University, Bangkok, Thailand

Hamada-Sato, Naoko Course of Safety management in Food Supply Chain, Tokyo University of Marine Science and Technology, Tokyo, Japan

Ijima, Noriyuki Division of Life Science and Engineering, School of Science and Engineering, Tokyo Denki University, Saitama, Japan

Hamer, R. J. T.I. Food and Nutrition, Wageningen, The Netherlands; and TNO Quality of Life, Zeist, The Netherlands

Inoue, Chiharu Unilever Food and Health Research Institute, Vlaardingen, The Netherlands

Hartel, Richard W. Department of Food Science, University of Wisconsin, Madison, Wisconsin, USA

Intipunya, Pilairuk School of Land, Crop and Food Sciences, University of Queensland, Brisbane, Australia

Contributors

Israkarn, Kamolwan Department of Food Science and Technology, Faculty of AgroIndustry, Kasetsart University, Bangkok, Thailand Jiménez-Aparicio, Antonio Departamento de Graduados e Investigación en Alimentos, Escuela Nacional de Ciencias Biológicas, Instituto Politécnico Nacional, México City, México Jouppila, Kirsi Department of Food Technology, University of Helsinki, Helsinki, Finland Kaewtathip, Thipthida Department of Food Science and Technology, Faculty of AgroIndustry, Kasetsart University, Bangkok, Thailand Kajiwara, Kazuhito School of Bionics, Tokyo University of Technology, Tokyo, Japan Kawai, Kiyoshi Department of Biofunctional Science and Technology, Graduate School of Biosphere Science, Hiroshima University, Hiroshima, Japan Khunae, Parida Department of Food Technology, Faculty of Agro-Industry, Prince of Songkla University, Songkhla, Thailand

xxiii

Kou, Yang General Mills, Inc., Riverside Technical Center, Minneapolis, Minnesota, USA Krochta, John M. Department of Food Science and Technology, University of California, Davis, California, USA Kulchan, Ratchaneewan Department of Packaging Technology, Faculty of Agro-Industry, Kasetsart University, Bangkok, Thailand Kurosaki, Kousuke School of Bionics, Tokyo University of Technology–Hachioji, Tokyo, Japan Kylli, Petri Department of Applied Chemistry and Microbiology, University of Helsinki, Helsinki, Finland Labuza, Katherine M. St. Paul Academy, St. Paul, Minnesota, USA Labuza, Peter S. St. Paul Academy, St. Paul, Minnesota, USA Labuza, Ted P. St. Paul Academy, St. Paul, Minnesota, USA Labuza, Theodore J. Department of Food Science and Nutrition, University of Minnesota, St. Paul, Minnesota, USA

xxiv

Contributors

Lähdesmäki, Maarit Department of Food Technology, University of Helsinki, Helsinki, Finland Laine, Pia Department of Food Technology, University of Helsinki, Helsinki, Finland Lampen, Ben Department of Food Science, University of Otago, Dunedin, New Zealand Ligero, Pablo Universidad de la Coruña, A Coruña, Spain Lillford, Peter J. Centre for Formulation Engineering, Chemical Engineering, University of Birmingham, Birmingham, UK Lim, Miang Hoong Department of Food Science, University of Otago, Dunedin, New Zealand Liu, Yeting Food Science & Technology Programme, National University of Singapore, Singapore Lodi, A. Department of Radiology and Biomedical Imaging, University of California San Francisco, San Francisco, California, USA

Loveday, Simon M. Riddet Institute, Massey University, Palmerston North, New Zealand Maldonado, Sara Departamento de Biodiversidad y Biología Experimental, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Buenos Aires, Argentina Marabi, Alejandro Nestlé Research Center, Nestec Ltd., Lausanne, Switzerland; and Institute of Biochemistry, Food Science and Nutrition, Faculty of Agricultural, Food and Environmental Quality Sciences, The Hebrew University of Jerusalem, Rehovot, Israel Matiacevich, Silvia Beatriz Departamento de Industrias, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Buenos Aires, Argentina Meinders, Marcel B. J. T.I. Food and Nutrition, Wageningen, The Netherlands; and Wageningen University and Research Centre, Wageningen, The Netherlands Mendoza-Pérez, Jorge A. Secretaría de Marina México, México City, México Mitchell, John R. Division of Food Sciences, School of Biosciences, University of Nottingham–Sutton Bonington Campus, Leicester, UK

Contributors

xxv

Morakot, Nongnit Department of Chemistry, Faculty of Science, Mahasarakham University, Mahasarakham, Thailand

Olofsson, P. SIK/Swedish Institute for Food and Biotechnology, Gothenburg, Sweden

Morison, K. R. Department of Chemical and Process Engineering, University of Canterbury, Christchurch, New Zealand

Ortíz-Moreno, A. Escuela Nacional de Ciencias Biológicas, Instituto Politécnico Nacional, México City, México

Moussa, Marwen Laboratoire de Génie des procédés Alimentaires et Biotechnologiques, ENSBANA, Université de Bourgogne, Dijon, France

Paes, Sabrina S. Division of Food Sciences, School of Biosciences, University of Nottingham–Sutton Bonington Campus, Leicester, UK

Moyano, Pedro Departamento de Ingeniería Química Universidad de Santiago de Chile, Santiago, Chile Muñoz-Herrera, Ana Laura Departamento de Graduados e Investigación en Alimentos, Escuela Nacional de Ciencias Biológicas, Instituto Politécnico Nacional, México City, México Murase, Norio Division of Life Science and Engineering, School of Science and Engineering, Tokyo Denki University, Saitama, Japan Nanasombat, Suree Department of Applied Biology, Faculty of Science, King Mongkut’s Institute of Technology Ladkrabang, Bangkok, Thailand Niamsa, Noi Department of Chemistry, Faculty of Science, Mahasarakham University, Mahasarakham, Thailand

Pascual-Pineda, Luz A. Escuela Nacional de Ciencias Biológicas, Instituto Politécnico Nacional, México City, México Pedreschi, Franco Pontificia Universidad Católica de Chile, Departmento de Ingeniería Química y Bioprocesos, Santiago, Chile Pehkonen, Kati S. Department of Food and Nutritional Sciences, University College Cork, Cork, Ireland Pérez-Nieto, A. Universidad de Guanajuato–Unidad de Estudios Superiores de Salvatierra, Salvatierra, México Perrier-Cornet, Jean-Marie Laboratoire de Génie des procédés Alimentaires et Biotechnologiques, ENSBANA, Université de Bourgogne, Dijon, France

xxvi

Contributors

Pihl, M. Department of Chemical and Biological Engineering, Chalmers University of Technology, Gothenburg, Sweden Pinal, Rodolfo Department of Industrial and Physical Pharmacy, Purdue University, West Lafayette, Indiana, USA Pittia, Paola Department of Food Science, University of Teramo, Teramo, Italy

Rades, Thomas School of Pharmacy, University of Otago, Dunedin, New Zealand Raemy, A. Nestlé Research Center, Nestec Ltd., Lausanne, Switzerland Rane, Sagar S. Department of Pharmaceutical Sciences, College of Pharmacy, University of Kentucky, Lexington, Kentucky, USA

Piumnoppakun, Nattaya Department of Applied Biology, Faculty of Science, King Mongkut’s Institute of Technology Ladkrabang, Bangkok, Thailand

Rattanasuwan, Metavee Department of Applied Biology, Faculty of Science, King Mongkut’s Institute of Technology Ladkrabang, Bangkok, Thailand

Piyachomkwan, Kuakoon Cassava and Starch Technology Research Unit, National Center for Genetic Engineering and Biotechnology, Kasetsart University, Bangkok, Thailand

Reid, David S. Department of Food Science and Technology, University of California, Davis, California, USA

Ponce, Peggy A. Departamento de Industias, Facultad Ciencias Exactas y Naturales, Universidad de Buenos Aires, Buenos Aires, Argentina Poovarodom, Ngamtip Department of Packaging Technology, Faculty of Agro-Industry, Kasetsart University, Bangkok, Thailand Praditdoung, Saisanom Department of Food Science and Technology, Faculty of AgroIndustry, Kasetsart University, Bangkok, Thailand

Riviera, Luca Storci Spa, Lemignano di Collecchio (PR), Italy Rocculi, Pietro Department of Food Science, Alma Mater Studiorum, University of Bologna, Campus of Food Science, Cesena, Italy Roos, Yrjö H. School of Food and Nutritional Sciences, University College Cork, Cork, Ireland Sacchetti, Giampiero Department of Food Science, University of Teramo, Teramo, Italy

Contributors

Saguy, I. Sam Institute of Biochemistry, Food Science and Nutrition, The Robert H. Smith Faculty of Agriculture, Food and Environment, The Hebrew University of Jerusalem, Rehovot, Israel Sakiyama, Takaharu Department of Food Science and Technology, Tokyo University of Marine Science and Technology, Tokyo, Japan Santagapita, Patricio Román Departamento de Industrias, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Buenos Aires, Argentina Savolainen, Marja Division of Pharmaceutical Technology, Faculty of Pharmacy, University of Helsinki, Helsinki, Finland Schebor, Carolina Departamento de Industrias, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Buenos Aires, Argentina Schnürer, Johan Department of Microbiology, Swedish University of Agricultural Sciences, Uppsala, Sweden Schoug, Åsa Department of Microbiology, Swedish University of Agricultural Sciences, Uppsala, Sweden Seow, Chee Choon FoodTech Consultancy, Penang, Malaysia

xxvii

Simonin, Hélène Laboratoire de Génie des procédés Alimentaires et Biotechnologiques ENSBANA, Université de Bourgogne, Dijon, France Siow, Lee Fong Department of Food Science, University of Otago, Dunedin, New Zealand Sirivongpisal, Piyarat Department of Food Technology, Faculty of Agro-Industry, Prince of Songkla University, Songkhla, Thailand Soga, Makoto School of Bionics, Tokyo University of Technology–Hachioji, Tokyo, Japan Soottitantawat, Apinan Center of Excellence in Particle Technology, Department of Chemical Engineering, Chulalongkorn University, Bangkok, Thailand Srirangsan, Paveena Department of Food Science and Technology, Tokyo University of Marine Science and Technology, Tokyo, Japan Sriroth, Klanarong Cassava and Starch Technology Research Unit, National Center for Genetic Engineering and Biotechnology, Kasetsart University, Bangkok, Thailand; Kasetsart Agricultural and Agro-Industrial Product Improvement Institute, Kasetsart University, Bangkok, Thailand; and Department of Biotechnology, Faculty of AgroIndustry, Kasetsart University, Bangkok, Thailand

xxviii

Contributors

Storci, Alfio Storci Spa, Lemignano di Collecchio (PR), Italy Suppakul, Panuwat Department of Packaging Technology, Faculty of Agro-Industry, Kasetsart University, Bangkok, Thailand Suzuki, Toru Department of Food Science and Technology, Tokyo University of Marine Science and Technology, Tokyo, Japan Talja, Riku A. Department of Food Technology, University of Helsinki, Helsinki, Finland Tejeda-Hernández, Violeta Departamento de Graduados e Investigación en Alimentos, Escuela Nacional de Ciencias Biológicas, Instituto Politécnico Nacional, México City, México Téllez-Medina, D. I. Escuela Nacional de Ciencias Biológicas, Instituto Politécnico Nacional, México City, México Thammathongchat, Savitree Department of Food Science and Technology, Tokyo University of Marine Science and Technology, Tokyo, Japan Thanatuksorn, Pariya School of Bionics, Tokyo University of Technology–Hachioji, Tokyo, Japan

Thirathumthavorn, Doungjai Department of Food Technology, Silpakorn University, Nakhon Pathom, Thailand Tran, Thierry Cassava and Starch Technology Research Unit, National Center for Genetic Engineering and Biotechnology, Kasetsart University, Bangkok, Thailand Tromp, R. Hans T.I. Food and Nutrition, Wageningen, The Netherlands; and NIZO Food Research, Ede, The Netherlands Troygot, Oranit Institute of Biochemistry, Food Science and Nutrition, The Robert H. Smith Faculty of Agricultural, Food and Environment, The Hebrew University of Jerusalem, Rehovot, Israel Truong, Tuyen Thuc School of Land, Crop and Food Sciences, University of Queensland, Brisbane, Australia Tsuji, Kaori Department of Food Science and Technology, Tokyo University of Marine Science and Technology, Tokyo, Japan Ubbink, Job Nestlé Research Center, Lausanne, Switzerland Van As, Henk Wageningen University and NMR Centre, Wageningen, The Netherlands

Contributors

van Duynhoven, John Unilever Food and Health Research Institute, Vlaardingen, The Netherlands Van Nieuwenhuijzen, Neleke H. T.I. Food and Nutrition, Wageningen, The Netherlands; and Wageningen University and Research Centre, Wageningen, The Netherlands Van Vliet, Ton T.I. Food and Nutrition, Wageningen, The Netherlands; and Wageningen University and Research Centre, Wageningen, The Netherlands Vega, Alberto Universidad de la Coruña, A Coruña, Spain Venturi, Luca Department of Food Science, Alma Mater Studiorum, University of Bologna, Campus of Food Science, Cesena, Italy Vittadini, Elena Food Science and Technology, Department of Industrial Engineering, University of Parma, Parma, Italy Vodovotz, Yael Department of Food Science and Technology, Ohio State University, Columbus, Ohio, USA Wallach, Rony Department of Soil and Water Sciences, The Robert H. Smith Faculty of Agricultural, Food and Environment, The Hebrew University of Jerusalem, Rehovot, Israel

xxix

Watanabe, Hisahiko Department of Food Science and Technology, Tokyo University of Marine Science and Technology, Tokyo, Japan Watanabe, Manabu Department of Food Science and Technology, Tokyo University of Marine Science and Technology, Tokyo, Japan Weglarz, Wladyslaw P. Unilever Food and Health Research Institute, Vlaardingen, The Netherlands; and Department of Magnetic Resonance Imaging, Institute of Nuclear Physics, Polish Academy of Sciences, Kraków, Poland Wellner, Nikolaus Institute of Food Research, Norwich Research Park, Norwich, UK Welti-Chanes, Jorge Departamento de Graduados e Investigación en Alimentos, Escuela Nacional de Ciencias Biológicas, Instituto Politécnico Nacional, México City, México Winger, Ray J. Institute of Food, Nutrition and Human Health, Massey University, Palmerston North, New Zealand Witek, Magdalena Wageningen University and NMR Centre, Wageningen, The Netherlands

xxx

Contributors

Xiang, Tian-xiang Department of Pharmaceutical Sciences, College of Pharmacy, University of Kentucky, Lexington, Kentucky, USA Yakimets, Iryna Division of Food Sciences, School of Biosciences, University of Nottingham–Sutton Bonington Campus, Leicester, UK Yamada, Shunsuke Division of Life Science and Engineering, School of Science and Engineering, Tokyo Denki University, Saitama, Japan Yimjaroenpornsakul, Achana Department of Food Science and Technology, Faculty of AgroIndustry, Kasetsart University, Bangkok, Thailand

Yoshii, Hidefumi Department of Biotechnology, Tottori University, Tottori, Japan Yoshioka, Sumie National Institute of Health Sciences, Tokyo, Japan Zhou, Weibiao Food Science and Technology Programme, National University of Singapore, Singapore

WATER PROPERTIES IN FOOD, HEALTH, PHARMACEUTICAL AND BIOLOGICAL SYSTEMS: ISOPOW 10

PART 1 Invited Speakers and Oral Presentations

Session 1 Water Mobility/Dynamics and Its Application in Food and Pharmaceutical Systems

Invited Speakers

1 Complementary Aspects of Thermodynamics, Nonequilibrium Criteria, and Water Dynamics in the Development of Foods and Ingredients M. P. Buera

Abstract Temperature and water content of the systems have been the variables most widely employed to define and predict the kinetic coefficients of desirable and undesirable changes in foods. Supplemented temperature vs composition phase diagrams have been demonstrated to be helpful in determining the feasibility of occurrence of phase or state transitions. These diagrams include the glass transition temperature (Tg) curve and the equilibrium liquidus curves. The inclusion of the nonequilibrium curves enables relationships with the time coordinate and, thus, with the dynamic behavior of the systems to be established and helps to predict whether the systems are under thermodynamic or kinetic control for given composition vs temperature conditions, provided that the thermal history of the samples is known. The present work analyzed how complementary aspects of thermodynamics and nonequilibrium criteria and water dynamics can be assembled in order to demonstrate the formulation and processing strategies to optimize the stability of food products and ingredients, especially in dry systems. A wide variety of kinetic data from several chemical reactions (in vegetable and animal tissues, dairy products, ingredients, and pharmaceutical formulations or model systems) from the published literature on the results from specifically designed experiments were distributed in supplemented phase diagrams. The results indicated that both solid-water interactions and structural characteristics of the systems governed the dependence of reaction rates on relative humidity. In addition to the supplemented phase diagrams, structural aspects of the matrices where the reaction takes place, water-sorption properties, and water mobility itself were key aspects for a complete interpretation to describe the dynamics of the chemical reactions.

Introduction The kinetic control of desirable and undesirable aspects of chemical reactions represents a challenge in basic and applied areas of chemistry and food sciences. Therefore, the impact of the variables and mechanisms that determine reaction rates has been given much attention. Temperature and water content of the systems have been the variables most widely employed to define and predict the kinetic coefficients. The significance of state and phase transitions in the stability of amorphous food materials 9

10

PART 1: Invited Speakers and Oral Presentations

and also their impact on chemical and enzymatic reactions have been evaluated since the 1980s (Slade and others 1989; Roos and Karel 1991; Karmas and others 1992; Levine and Slade 1992; Bell and Hageman 1994; Bell 1995; Buera and Karel 1995; Lievonen and others 1998; Bell and White 2000; Kouasi and Roos 2000). Since an equilibrium state does not exist in these systems, the conservation of desirable properties in foods and ingredients is governed by conditions of metastability, often based on the maintenance of the systems in an amorphous state (Levine and Slade 1992): although the system components are not thermodynamically stable, they are kinetically stabilized. Supplemented temperature composition phase diagrams have proven helpful in determining the potential of phase or state transitions (Slade and others 1989; Levine and Slade 1992). These diagrams include the nonequilibrium glass transition temperature (Tg) curve and the equilibrium liquidus curves. The inclusion of the nonequilibrium curves enables the relationships with the time coordinate and, thus, with the dynamic behavior of the systems to be established and helps to predict whether the systems are under thermodynamic or kinetic control for given temperature/composition conditions, providing that the thermal history of the samples is known. Structural aspects of the matrices where the reactions occur, sorption, and water mobility itself were also detected as key aspects in describing the dynamic of this reaction. The present work analyzed how complementary aspects of thermodynamics and nonequilibrium criteria and water dynamics can be assembled in order to point out formulations or processing strategies to optimize the stability of food products and ingredients, especially in dry systems.

Reactions, Materials, and Methods The stability of selected dehydrated systems toward the Maillard reaction, the loss of enzymatic activity, or carotene degradation was analyzed. Data obtained in freezedried vegetable tissues, dairy products, ingredients, and pharmaceutical formulations or model systems as the result of specifically designed experiments (Mazzobre and others 2001; Longinotti and others 2002; Prado and others 2006; Acevedo and others 2006, 2008b; Sutter and others 2007) were plotted in supplemented phase diagrams. Additional information on water dynamics and water-sorption properties was obtained. Differential scanning calorimetry (DSC) was helpful in analyzing thermal transitions and generating temperature- vs composition-supplemented state diagrams. Timeresolved proton nuclear magnetic resonance (1H-NMR) was used to complement DSC with the aim of obtaining a better understanding of the mobility of water and food solids in the systems (Schmidt and Lai 1991; Kou and others 2000; Tang and others 2000; Chatakanonda and others 2003). X-ray diffraction and microscopy provided information on molecular and microscopic structural changes.

Maillard Reaction The Maillard reaction may have a positive or a negative contribution to the quality of food products, and its complexity necessitates a deep kinetic analysis of reaction

Water Dynamics in the Development of Foods and Ingredients

11

media and the variables involved to direct the reaction with the desirable outcome. In dehydrated foods, the reaction is the cause of off-flavors, off-colors, and loss of nutritional value. The kinetics of the Maillard reaction must be controlled also as a requirement of new food technologies: in the development of natural flavors, pigments, emulsifiers, antimicrobials, and antioxidants; in the formulation of protective media for biological systems and ingredients; and for controlled modification of biomolecular functionality or structure. The reaction rate is strongly dependent on the concentration, ratio, and chemical nature of reactants, temperature, water content, pH, and water activity (aw) (Labuza and Baisier 1992), but it is also influenced by the physical properties of the media. In liquid systems, the Maillard reaction rate diminishes continuously as relative humidity (RH) increases, mainly because water is a product of the reaction (Hodge 1953; Eichner and Karel 1972; Labuza and Saltmarch 1981). However, in solid or quasi-solid systems, in which reactants are constrained by mobility restrictions, a maximum rate of nonenzymatic browning (NEB) is observed at a given intermediate RH value. Thus, the presence of a maximum in the plot of rate versus water content (or aw) is a consequence of the low reaction rates due to mobility limitations of the reactants (at low water content) and inhibition by the product (at high water content) (Buera and Karel 1995; Van Boekel 2001). Systems with different structural characteristics were analyzed to elucidate the relationship among browning rate, water-solid interaction, and water mobility and the incidence of structural changes accompanying phase or state transitions: highly collapsible polymeric (polyvinylpyrrolidone [PVP40]) matrices, crystallizing lactose and milk systems, and vegetable tissues exhibiting an intermediate degree of collapse because of the presence of water-insoluble polymers capable of providing residual structure. The shaded areas in Figures 1.1 and 1.2 represent the temperature vs composition conditions at which the Maillard reaction was analyzed, and the circled regions show the conditions at which the maximum rates were observed. The corresponding Tg curves are shown in Figures 1.1 and 1.2 also as a function of water content. Since sugars are the main soluble components determining Tg values in vegetable and dairy systems, solubility curves for the main sugars present were included as a reference. In PVP systems (Figure 1.1) at a given temperature, the Maillard reaction rate increased as water content increased, and the maximum rate occurred at a water content at which Tg was close to the storage temperature. Above this point, the samples presented a fully collapsed fluid aspect, and the Maillard rate decreased when water content increased (behavior similar to that of liquid systems). In the crystallizable lactose- or trehalose-containing systems (Figure 1.1), the maximum rate occurred at conditions in which a considerable degree of sugar crystallization had occurred, but when the samples were totally crystalline the rate decreased. Figure 1.3 shows the rate of the Maillard reaction as a function of water content for lactose systems compared with the rates with milk and lactose-starch. It should be noted that, in milk or lactose-starch systems, the presence of proteins or biopolymers retarded lactose crystallization. Consequently, the maximum rate of

Figure 1.1. State diagrams showing the glass transition temperatures (Tg) as a function of the mass fraction of water (w) for (a) polymeric (PVP), (b) lactose, and (c) trehalose matrices in which the Maillard reaction was developed. Shaded areas represent the regions in which the experiments were performed. The conditions at which fast-collapse or crystallization phenomena were observed are indicated by dotted lines. Circled regions indicate the conditions at which the maximum browning rate was observed.

12

Figure 1.2. State diagrams showing the glass transition temperatures (Tg) as a function of the mass fraction of water (w) for (a) apple, (b) cabbage, and (c) potato in which the Maillard reaction was developed. Shaded areas represent the regions in which the experiments were performed. The solubility curves (S) for the sugars fructose (fru), glucose (glu), and sucrose (suc) are indicated as a reference. Circled regions indicate the conditions at which the maximum browning rate was observed.

13

14

PART 1: Invited Speakers and Oral Presentations

150 Crystallization zone

0.3

k lactose

50

Tg

k milk

0.2 0

k lactose-starch

-50 -100

K

Tg (°C)

100

0.1 0

10

20

30

40

50

W Figure 1.3. State diagrams showing the lactose glass transition temperature (Tg) as a function of the mass fraction of water (W). Browning rate constants (k) for the lactose, milk, and lactose-starch systems are also shown as a function of W.

browning occurred with a higher water content, and the maximum rate was maintained in higher water contents than in the sugar-only matrices (complete crystallization occurred at higher water content values in those samples). In vegetable tissues containing structure-maintaining water-insoluble biopolymers and presenting an intermediate degree of collapse, the maximum rate of NEB occurred at RHs in a range of 50%–80% when the samples were well above Tg at the storage temperature (Figure 1.2). In all cases analyzed, the Maillard reaction occurred below the Tg and required a minimum of water content (w in Figure 1.2) to mobilize the reactants, but the reaction was immediately inhibited by water in excess of this requirement. The minimum water content was directly related neither to the monolayer Guggenheim-Anderson-de Boer (GAB) value nor to the Tg value of the systems. The browning rate of the vegetables and food models analyzed was very low in the glassy state, but at temperatures above the Tg, in addition to the decreasing viscosity and increasing rate, other changes such as crystallization and collapse affected the browning rate. As already discussed, the maximum reaction rates were reached either close to or well above the Tg, depending on the system structure. The observation of the maximum rate as a function of water content indicated that the reaction rate decreased at a point at which the matrix was unable to adsorb water either because of crystallization (as in the case of lactose-containing systems), saturation of the most active sites of the water-adsorbing matrix, or capillary condensation. I thus analyzed the location of the beginning of the third water-sorption stage, the transverse spin-spin 1H-NMR relaxation times by spin echo after the Hahn sequence pulse, and the detection of frozen water by DSC. The location of the beginning of the third sorption stage was analyzed by the inverse plot of the GAB model, as proposed by Timmermann and Chirife (1991). The analysis of proton spin-spin relaxation times by the Hahn spin-echo pulse sequence in the samples at different water contents showed two T2 components. A fast-decaying component, with T2-Hahn-1 values on the order of 30–40 ms was present

Water Dynamics in the Development of Foods and Ingredients

15

at all the RHs analyzed and corresponds to both protons of solids and protons of water molecules that are strongly influenced by their proximity to the solid components. (It could also be detected by free-induction decay [FID] measured after a single 90° pulse.) Above a given water content (in the proximity of aw = 0.22), a slow-relaxing component (T2-Hahn-2) in the range of 400–950 ms was also observed, the value of which increased linearly up to the beginning of the third sorption stage, as previously defined, and then became constant. The inhibitory water concentration was associated with the second inflection point of the water-sorption isotherm up to the appearance of freezable water as determined by DSC and up to the saturation of the second sorption stage as determined by proton spin-spin transverse relaxation times (T2) analyzed by 1HNMR with the Hahn pulse sequence (Acevedo and others 2008b). Upon the appearance of freezable water and highly mobile water (determined by the T2-Hahn-2 value), the NEB rate decreased as the water molecules inhibited the reaction and/or diluted the reactants. As the scheme in Figure 1.4 shows, the water content at a maximum Maillard reaction rate results from the compromise between water plasticization and its inhibitory effect. These conditions can be predicted on the basis of sorption and structural properties and of thermal transitions of the matrices where the reaction takes place and through the analysis of water mobility.

Matrix crystallization (lactose)

Matrix collapse (PVP)

3rd sorption stage (T2-2 ~ 1 ms) (vegetables)

Tem

per

atur

e

kb

Figure 1.4. Schematic tridimensional plot showing the browning rate (kb, in arbitrary units) as a function of the temperature and mass fraction of water (w). The conditions for the decreasing rate constant after the maximum were different for each type of system and are indicated by the arrows. PVP, polymeric matrix.

16

PART 1: Invited Speakers and Oral Presentations

Enzyme Stability Due to chemical and physical changes, most proteins lose their activity when stored for extended periods in aqueous solution, and they are generally freeze-dried to achieve a stable product. The dry state slows chemical degradation and alleviates physical problems, such as protein unfolding and aggregation. However, protein stability in the solid state can be worse than that in liquid state if adequate components are not present during the process to form suitable matrices (Crowe and others 1998). Sugars, and particularly trehalose, have been found to be optimal in protecting enzymes during drying and later storage (Leslie and others 1995; Suzuki and others 1997; Crowe and others 1998). Much knowledge about protein stability is derived from understanding how organisms survive thermal and hydric stresses (Sun and Leopold 1997). Vitrification of protective sugars is possibly the main strategy in nature to avoid the crystallization of these sugars in anhydrobiotic organisms. Mazzobre and others (2003, 2008) analyzed the stability of freeze-dried enzymes (β-galactosidase, invertase, honey amylase, soy urease, and soy transaminase) over a wide range of temperature and water-content conditions. Materials capable of forming amorphous matrices, but with different physicochemical characteristics, were chosen to compare their efficiency in protecting the enzymes. In the polymeric glassy matrices, the enzyme stability was diminished either by increasing water content at a certain temperature or by increasing the storage temperature at fixed water content. The good glass–former polymers maltodextrin and PVP were not effective in protecting the enzyme during heat treatment, the enzyme stability being affected mainly by effects of temperature. Crowe and others (1998) reported that although vitrification of the structure is necessary for improved enzyme stability, it is not the only condition required for the protection of molecules, since specific hydrogen-bond interactions between the matrix and the protein are also needed. According to this, sugars, and especially trehalose, were more effective protectants despite their lower Tg values, even at relatively high temperatures. The enzymes retained quite good activity in the trehalose supercooled region, but their activity decreased drastically when the sugar crystallized at a mass fraction of water greater than 0.1. Starch protected the enzymes adequately at high water content and low temperature where stability was completely lost in other systems. Enzyme interactions with the surface of starch granules and the high Tg value of this matrix may have played a role in enzyme stability. When sugar crystallizes, the protein is excluded from the sugar crystals, and is exposed to a matrix where, besides lacking the stabilizing effect of hydroxyl groups, the changes in pH, concentration of reactive groups, and ionic strength may also negatively affect its stability. However, it has been observed that, for many enzymes, if sugar crystallization is inhibited or conveniently delayed, the protective action of sugars may be extended to the supercooled-liquid state (Suzuki and others 1997; Buera and others 2005; Mazzobre and others 2008). The addition of polymers, other sugars, or salts extended the protective effect of sugars to the supercooled region by delaying crystallization (Mazzobre and others 1997; Gabarra and Hartel 1998). In dairy systems, compared with pure-lactose systems

Water Dynamics in the Development of Foods and Ingredients

17

Figure 1.5. Supplemented phase diagram showing the sucrose glass transition temperature (Tg) as a function of the mass fraction of water (W): Ts, sucrose solubility curve; and Tm, water-melting curve. The effect of magnesium chloride on the sucrose curves is indicated by arrows. The dotted lines indicate the freezing-point depression for increasing sugar/salt proportions (R), from R1 to R3: w ′g, the mass fraction of water for the maximum cryoconcentrated matrix; and Tg′, the glass transition temperature of the maximally concentrated matrix. Adapted from Mazzobre and others (2001).

(Jouppila and Roos 1994), the presence of proteins delayed lactose crystallization. Gelatin inhibited crystallization of raffinose and, in the presence of bovine serum albumin, of sucrose, and raffinose crystallization was inhibited even at high RH (84%) (Espinosa and others 2006). Additives like polymers that raise the overall Tg of pure sugars reduced molecular mobility and inhibited crystallization, although delayed crystallization in many studies was related to the modification of the molecular environment of the crystallizing sugar in those mixtures, affecting thermodynamic and geometric factors that control nucleation without changing Tg significantly (Mazzobre and others 2001; Longinotti and others 2002). A special set of studies was dedicated to the analysis of the effect of salts on the phase diagram of sugars (Mazzobre and others 2001, 2008; Longinotti and others 2002), which are presented in Figure 1.5. Water sorption and water or sugar crystallization behavior indicated that the effect of salts on the sugar phase diagram was directly related to the charge/mass ratio of the cations present (Mg > Ca > Na > K). Measurements of electrical conductivity in concentrated sugar-salt-water systems revealed a high population of local inhomogeneities, which were induced by preferential solvation of the ions as a consequence of the larger ion-water interactions as compared with the ion-disaccharide interactions. Therefore, whereas the ion mobility is enhanced by a low-viscosity local environment, sugar molecular mobility should

18

PART 1: Invited Speakers and Oral Presentations

be depressed by a high local viscosity. Ediger (2000) described spatially heterogeneous dynamics in supercooled liquids, as resulted from analysis by different techniques. The dynamics in regions separated by a few nanometers could be different by several orders of magnitude. The short-range dynamics of sugar-water systems could change dramatically without modifying the Tg of the system, which is a result of suprastructural relaxation. Thus, the presence of salts could retard the sugar crystallization even when Tg remains unchanged. This effect should be more pronounced for ions with stronger interactions with water. In frozen systems, on the other hand, the salts have a colligative effect (Figure 1.5) and promote an increase of the water associated with the maximally concentrated nonfrozen phase (w in Figure 1.5). Consequently, the Tg of this phase decreases. Decreased enzyme activity was observed in these saltcontaining systems. Mazzobre and others (2008) observed that the effect of salts, with regard to sugar crystallization kinetics and enzyme inactivation, seems also to be associated with the magnitude of their effect on disrupting the tetrahedral hydrogen-bond network of water. Water structure-maker ions (citrate > acetate; Mg+2) enhance the tetrahedral coordinated hydrogen-bond structure of water, and water structure-breaker ions (K+) disrupt the tetrahedral coordination of water. Some ions, such as Na+ or Cl−, are considered neutral (Calligaris and Nicoli 2006). Trehalose acts as a structure breaker but provides enzyme stabilization by strong hydrogen-bonding interactions (Patist and Zoerb 2005). In restricted water environments, such as the dehydrated (or frozen) systems analyzed in the present work, the amount of water determines the kinetics of phase changes and enzyme inactivation. Thus, the type of water-ion interactions are manifested in those dynamic changes, and the use of the so-called Hofmeister series could offer great help in their description. To optimize the efficiency of biomolecular dehydroprotectant agents, the development of state diagrams is a good starting point for the analysis of the dynamics of quality changes, but the diagrams must be complemented by knowledge of the intermolecular interactions that may occur. Besides supramolecular aspects, like Tg and crystallinity, the density of the molecular packing of the matrices, reducing power, and hydrogen-bonding capacity determined the effectiveness of agents as biomolecular protectants. Electrolytes commonly present in biological media or food and pharmaceutical formulations modified metastable systems, affecting the kinetics of water and sugar crystallization and enzyme inactivation, by both molecular and supramolecular interactions.

Degradation of Carotenes Stability and retention of labile biomolecules during drying and later storage often depend on encapsulation of the biomolecules in the amorphous matrix formed during dehydration processes (Constantino and others 1998; Pyne and others 2003). The high degree of unsaturation in the structure of carotenes renders them extremely susceptible to oxidation. Amorphous sugars are effective encapsulating agents. However, sugar crystallization as a consequence of storage above the Tg promotes not only the loss of the stabilizing effect on biomolecules such as enzymes, as previously discussed, but

Water Dynamics in the Development of Foods and Ingredients

19

also the release of encapsulated lipids (Shimada and others 1991; Labrousse and others 1992). The changes in the physical structure of the matrix may also lead to increased permeability and diffusivity of gases (water and oxygen) that affect reaction rates and decrease stability of encapsulated active materials (Karel 1991; Karel and Saguy 1991). Figure 1.6 presents the phase diagram of PVP (a) and crystallizable trehalose (b) matrices. The shaded area indicates the magnitude of the kinetic constants of carotene degradation.

Figure 1.6. State diagram showing the glass transition temperatures (Tg) as a function of the mass fraction of water (w) for (a) polymeric (PVP) and (b) trehalose matrices in which β-carotene was encapsulated. The conditions at which fast-collapse or crystallization phenomena were observed are indicated by dotted lines. Shaded areas represent the regions in which the experiments were performed; within those areas, the darker regions indicate a high β-carotene loss rate. Arrows show the direction of the kinetic constant increase for β-carotene loss.

20

PART 1: Invited Speakers and Oral Presentations

Elizalde and others (2002) showed that the rate of encapsulated β-carotene loss in a trehalose matrix was affected mainly by the moisture in excess of that necessary for trehalose dihydrate crystallization. In fact, when water content was low but higher than 10%, the available water crystallized with trehalose, forming the dihydrate, and less water was free to liberate the encapsulated β-carotene. Once crystallization was completed, the kinetics of β-carotene loss were strongly accelerated (Figure 1.6a). Prado and others (2006) demonstrated that β-carotene loss in a freeze-dried polymeric (PVP-40) matrix was observed mainly in the glassy state (below Tg), where the highporosity matrix allowed oxygen diffusion and then fast degradation of β-carotene. To the contrary, the lower degradation rate constants were observed under conditions in which the structural collapse caused the disappearance or dramatic decrease of matrix micropores (Figure 1.6b). Although maltodextrins improved the shelf life of β-carotene in spray-dried carrot juice (Desobry and others 1998), and the release of encapsulated material has been qualitatively related to structural collapse or shrinkage caused by storage above the matrix Tg (Omatete and King 1978; Levi and Karel 1995; Selim and others 2000; Serris and Biliaderis 2001), in the case of oxidizable compounds the sample porosity in freeze-dried amorphous systems may negatively affect the stability of encapsulated compounds. Mannitol is a popular excipient used in freeze-dried formulations to stabilize biomolecules (proteins, enzymes, hormones, and vitamins). Obtaining amorphous pure mannitol is difficult because of the great tendency of this polyol to crystallize. Various solutes that remain amorphous in frozen solutions and during freeze drying (sodium chloride, dipotassium hydrogen phosphate [K2HPO3], and glycine) were reported to inhibit mannitol crystallization (Pikal and others 1991; Constantino and others 1998; Pyne and others 2003; Yoshinari and others 2003). As in the case of sugars, discussed previously with regard to enzyme stability, the effect of salts on carotene encapsulation was studied in mannitol systems (Sutter and others 2007). Phosphate salts significantly delayed mannitol crystallization during freeze drying, and consequently the degree of β-carotene encapsulation increased (Sutter and others 2007). This effect was maintained for quite a long time during storage of the freeze-dried samples at 25°C. The divalent cations showed a synergistic effect and also modified β-carotene degradation kinetics during storage, increasing carotene’s stability. The mechanism of crystallization inhibition likely includes a change in the hydrogen-bond network and/or a change in molecular mobility in the presence of divalent cations and phosphate anions. The degradation rate of β-carotene in a mannitol–potassium dihydrogen phosphate matrix increased as the %RH increased until reaching a value at which the samples collapsed (75% RH) and then the degradation rate decreased.

Structural Effects In systems with very low water content, water released during the Maillard reaction or sugar crystallization accelerated enzyme inactivation and browning (Kim and others 1981; Burin and others 2000), and the magnitude of the effect depended on the degree of collapse or porosity, which affected the retention of water in the systems (Buera

Water Dynamics in the Development of Foods and Ingredients

21

and Karel 1995; Burin and others 2004). The water-content increase resulting from the Maillard reaction was also reflected in a Tg depression (Roos and others 1996; Burin and others 2004). Acevedo and others (2008a) showed that the different structure generated during drying of apple discs, depending on the type of drying method used (air convection or the use of different freezing rates before freeze drying) affected sorption properties of the dried material and, consequently, the rate of browning development, which in turn were also different from those of dried-apple powdered samples. An inverse correlation was observed between degradation rate constants for βcarotene and degree of collapse. Thus, matrix collapse under controlled conditions during product processing may improve the stability of encapsulated biomolecules. These observations demonstrated that factors such as microstructure and matrix porosity may be important modifiers of reaction kinetics, and the sample structure is another variable that has to be considered when the kinetics of deteriorative reactions are analyzed.

Concluding Remarks The rates of three different chemical reactions were analyzed from the perspective of phase diagrams of the matrices where the reactions occurred. Both solid-water interactions and structural characteristics of the systems governed the dependence of reaction rates on RH. In addition to affecting chemical reactions via aw and by plasticizing amorphous systems, water mobility itself was demonstrated to have a direct impact on chemical reactivity in low-moisture and intermediate-moisture systems. In this way, besides the valuable information provided by localizing the plausible system conditions (compositions and temperatures) on supplemented phase diagrams, structural aspects of the matrices where the reaction occurs, water-sorption properties, and water mobility itself were detected also as key aspects that must be considered for a complete interpretation in describing the dynamics of the chemical reactions. Product formulation, process, and storage may be managed through knowledge of reaction kinetics, solid and water dynamic properties, transition temperatures, and process variables (mainly water content and temperature). Potential topics for further research include the study of macroscopic and molecular properties of the materials, such as the effect of sub-Tg relaxations on the kinetics of chemical reactions, or local heterogeneities in water distribution at microscopic scales. Also, the quantification of structural effects such as collapse and compression would be valuable in the complete interpretation of deteriorative reaction kinetics.

Acknowledgments The author acknowledges financial support from the University of Buenos Aires (EX226), the Argentine National Agency of Scientific and Technological Promotion (Agencia Nacional de Promoción Científica y Tecnológica [PICT 20545 and 3066]), and the Argentine National Scientific and Technical Research Council (Consejo

22

PART 1: Invited Speakers and Oral Presentations

Nacional de Investigaciones Científicas y Técnicas [CONICET]). The findings in this work are part of the technical report of the International Union of Pure and Applied Chemistry (IUPAC) Project (2003-036-2-100).

References Acevedo NC, Briones V, Buera MP, Aguilera JM. 2008a. Microstructure affects the rate of chemical, physical and color changes during storage of dried apple discs. J Food Eng 85:222–31. Acevedo NC, Schebor C, Buera MP. 2006. Water-solids interactions, matrix structural properties and the rate of non-enzymatic browning. J Food Eng 77:1108–15. Acevedo NC, Schebor C, Buera MP. 2008b. Non-enzymatic browning kinetics analyzed through watersolids interactions and water mobility. J Agric Food Chem 108:900–6. Bell L. 1995. Kinetics of non-enzymatic browning in amorphous solid systems: distinguishing the effects of water activity and the glass transition. Food Res Int 28:591–7. Bell LN, Hageman MJ. 1994. Differentiating between the effects of water activity and glass transition dependent mobility on a solid state chemical reaction: aspartame degradation. J Agric Food Chem 42:2398–401. Bell LN, White KL. 2000. Thiamin stability in solids as affected by the glass transition. J Food Sci 65:498–501. Buera MP, Karel M. 1995. Effect of physical changes on the rates of nonenzymic browning and related reactions. Food Chem 52:167–73. Buera MP, Schebor C, Elizalde B. 2005. Carbohydrate crystallisation phenomena in dehydrated food and ingredient formulation: involved factors, consequences and prevention. J Food Eng 67:157–65. Burin L, Jouppila K, Roos Y, Kansikas J, Buera MP. 2000. Color formation in dehydrated modified whey powder systems as affected by compression and Tg. J Agric Food Chem 48:5263–8. Burin L, Jouppila K, Roos Y, Kansikas J, Buera MP. 2004 Retention of β-galactosidase activity as related to Maillard reaction, lactose crystallization collapse and glass transition in low moisture whey systems. Int Dairy J 14:517–25. Calligaris S, Nicoli MC. 2006. Effect of selected ions from lyotropic series on lipid oxidation rate. Food Chem 94:130–5. Chatakanonda P, Chinachoti P, Sriroth K, Piyachomkwan K, Chotineeranat S, Tang HR, Hills B. 2003. The influence of time and conditions of harvest on the functional behavior of cassava starch: a proton NMR relaxation study. Carbohydr Polym 53:233–40. Constantino HR, Andya JD, Nguyen PE, Dasovich N, Crowe JH, Carpenter JF, Crowe LM. 1998. The role of vitrification in anhydrobiosis. Annu Rev Physiol 60:73–103. Crowe JH, Carpenter JF, Crowe LM. 1998. The role of vitrification in anhydrobiosis. Annu Rev Physiol 60:73–103. Desobry SA, Netto FM, Labuza TP. 1998. Preservation of β-carotene from carrots. Crit Rev Food Sci Nutr 38:381–96. Ediger MD. 2000. Spatially heterogeneous dynamics in supercooled liquids. Annu Rev Phys Chem 51:99–128. Eichner K, Karel M. 1972. The influence of water content and water activity on the sugar-amino browning reaction in model systems under various conditions. J Agric Food Chem 20:218–23. Elizalde BE, Herrera L, Buera MP. 2002. Retention of β-carotene encapsulated in a trehalose matrix as affected by moisture content and sugar crystallization. J Food Sci 57:3039–45. Espinosa L, Schebor C, Buera MP, Moreno S, Chirife J. 2006. Inhibition of trehalose crystallization by cytoplasmic yeast components. Cryobiology 52:157–60. Gabarra P, Hartel W. 1998. Corn syrup solids and their saccharide fractions affect crystallization of amorphous sucrose. J Food Sci 63:523–8. Hodge JE. 1953. Chemistry of browning reactions in model systems. J Agric Food Chem 1:928–43. Jouppila K, Roos YH. 1994. Glass transitions and crystallization in milk powders. J Dairy Sci 77:2907–15.

Water Dynamics in the Development of Foods and Ingredients

23

Karel M. 1991. Physical structure and quality of dehydrated foods. In: Mujumdar AS, Filkova I, editors. Drying ’91. Amsterdam: Elsevier. p 26–35. Karel M, Saguy I. 1991. Effects of water on diffusion in food systems. In: Levine H, Slade L, editors. Water relationships of foods. New York: Plenum. p 157–73. Karmas R, Buera MP, Karel M. 1992. Effect of glass transition on rates of non-enzymatic browning in food systems. J Agric Food Chem 40:873–9. Kim M, Saltmarch M, Labuza TP. 1981. Non-enzymatic browning of hygroscopic whey powders in open versus sealed pouches. J Food Proc Preserv 5:49–57. Kou Y, Dickinson LC, Chinachoti P. 2000. Mobility characterization of waxy corn starch using wide-line 1 H nuclear magnetic resonance. J Agric Food Chem 48:5489–95. Kouasi K, Roos YH. 2000. Glass transition and water effects on sucrose inversion by invertase in a lactosesucrose system. J Agric Food Chem 48:2461–6. Labrousse S, Roos Y, Karel M. 1992. Collapse and crystallization in amorphous matrices with encapsulated compounds. Sci Aliments 12:575–769. Labuza T, Baisier WM. 1992. The kinetics of nonenzymatic browning. In: Schwartzberg H, Hartel R, editors. Physical chemistry of foods. New York: Marcel Dekker. p 595–649. Labuza T, Saltmarch M. 1981. Kinetics of browning and protein quality loss in whey powders during steady state and nonsteady state storage conditions. J Food Sci 41:92–6. Leslie SB, Israeli E, Lighthart B, Crowe JH, Crowe LM. 1995. Trehalose and sucrose protect both membranes and proteins in intact bacteria during drying. Appl Environ Microbiol 61:3592–7. Levi G, Karel M. 1995. The effect of phase transitions on release of n-propanol entrapped in carbohydrate glasses. J Food Eng 24:1–13. Levine H, Slade L. 1992. Glass transitions in foods. In: Schwartzberg H, Hartel R, editors. Physical chemistry of foods. New York: Marcel Dekker. p 83–221. Lievonen SM, Laaksonen TJ, Roos YH. 1998. Glass transition and reaction rates: nonenzymatic browning in glassy and liquid systems. J Agric Food Chem 46:2778–84. Longinotti MP, Mazzobre MF, Buera MP, Corti HR. 2002. Effect of salts on the properties of aqueous sugar systems in relation to biomaterial stabilization. 2. Sugar crystallization rate and electrical conductivity behaviour. Phys Chem Chem Phys 4:533–40. Mazzobre MF, Buera MP, Chirife J. 1997. Protective role of trehalose on thermal stability of lactase in relation to its glass and crystal forming properties and effect of delaying crystallization. Lebensm Wiss Technol 30:324–9. Mazzobre MF, Hough G, Buera MP. 2003. Phase transitions and functionality of enzymes and yeasts in dehydrated matrices. Food Sci Technol Int 9:163–72. Mazzobre MF, Longinotti MP, Buera MP, Corti HR. 2001. Effect of salts on the properties of aqueous sugar systems in relation to biomaterial stabilization. 1. Water sorption behavior and ice crystallization/ melting. Cryobiology 43:199–210. Mazzobre MF, Santagapita PR, Gutiérrez N, Buera MP. 2008. Consequences of matrix structural changes on functional stability of enzymes as affected by electrolytes. In: Gutiérrez-López GF, Barbosa-Cánovas GV, Welti-Chanes J, Parada-Arias E, editors. Food engineering: integrated approaches. New York: Springer. p 73–87. Omatete OO, Judson King C. 1978. Volatiles retention during rehumidification of freeze-dried food models. J Food Technol 13:265–80. Patist A, Zoerb H. 2005. Preservation mechanisms of trehalose in foods and biosystems. Colloids Surf [B] 40:107–13. Pikal MJ, Dellerman KM, Roy MI, Riggin RM. 1991. The effects of formulation variables on the stability of freeze-dried human growth hormone. Pharm Res 8:427–36. Prado SM, Buera MP, Elizalde BE. 2006. Structural collapse prevents β-carotene loss in a super-cooled polymeric matrix. J Agric Food Chem 54:79–85. Pyne A, Koustov CH, Suryanarayanan R. 2003. Solute crystallization in mannitol-glycine systems: implications on protein stabilization in freeze-dried formulations. J Pharm Sci 92:2272–83. Roos Y, Jouppila K, Zielasko B. 1996. Non-enzymatic browning-induced water plastification: glass transition temperature depression and reaction kinetics determination using DSC. J Thermal Anal 47:1437–50.

24

PART 1: Invited Speakers and Oral Presentations

Roos YH, Karel M. 1991. Plasticizing effect of water on thermal behavior and crystallization of amorphous food models. J Food Sci 56:38–43. Schmidt SJ, Lai H. 1991. Use of NMR and MRI to study water relations in foods. In: Levine H, Slade L, editors. Water relationships in foods. New York: Plenum. p 405–52. Selim K, Tsimidou M, Biliaderis CG. 2000. Kinetic studies of saffron carotenoids encapsulated in amorphous polymer matrices. Food Chem 71:199–206. Serris GS, Biliaderis CG. 2001. Degradation of beetroot pigment encapsulated in polymeric matrices. J Sci Food Agric 81:691–700. Shimada Y, Roos Y, Karel M. 1991. Oxidation of methyl linoleate encapsulated in amorphous lactose-based food model. J Agric Food Chem 39:637–41. Slade L, Levine H, Finlay JW. 1989. In: Phillips R, Findlay LAW, editors. Protein quality and the effects of processing. Amsterdam: Elsevier Science. p 9–124. Sun WQ, Leopold AC. 1997. Cytoplasmatic vitrification and survival of anhydrobiotic organisms. Plant Physiol 89:767–72. Sutter SC, Buera MP, Elizalde BE. 2007. β-Carotene encapsulation in a mannitol matrix as affected by divalent cations and phosphate anion. Int J Pharm 332:45–54. Suzuki T, Imamura K, Yamamoto K, Satoh T, Okazaki M. 1997. Thermal stabilization of freeze-dried enzymes by sugars. J Chem Eng Jpn 30:609–13. Tang HR, Godward J, Hills B. 2000. The distribution of water in native starch granules: a multinuclear NMR study. Carbohydr Polym 43:375–87. Timmermann EO, Chirife J. 1991. The physical state of water sorbed at high activities in starch in terms of the GAB sorption equation. J Food Eng 13:171–9. Van Boekel MA. 2001. Kinetic aspects of the Maillard reaction: a critical review. Nahrung/Food 45:150–9. Yoshinari T, Forbes RT, York P, Karawahisma Y. 2003. Crystallization of amorphous mannitol is retarded using boric acid. Int J Pharm 258:109–20.

2 Water Mobility in Solid Pharmaceuticals as Determined by Nuclear Magnetic Resonance, Isothermal Sorption, and Dielectric Relaxation Measurements S. Yoshioka and Y. Aso

Abstract Nuclear magnetic resonance, dielectric relaxation spectroscopy, differential scanning calorimetry, and water-sorption isotherm data are presented to demonstrate a wide range of molecular mobilities observed for water molecules present in solid active pharmaceutical ingredients and water molecules coexisting with pharmaceutical excipients.

Introduction It is widely recognized that the presence of water molecules in solid pharmaceuticals can affect the chemical and physical stability of these solids. For example, water can act as a reactant or medium for chemical degradation, or as a plasticizer, which enhances chemical and physical degradation, in amorphous pharmaceuticals (Shalaev and Zografi 1996). In contrast, small amounts of water can stabilize solid peptide and protein drugs. Because the significance of these effects is often related to the mobility of water molecules, an understanding of water mobility in solid pharmaceuticals should be of great value in the development of stable solid-dosage forms. This chapter presents the experimental data on the mobility of water molecules present in solid active pharmaceutical ingredients (APIs) and water molecules coexisting with various pharmaceutical excipients, which demonstrate a wide range of water mobilities in the solids. Nuclear magnetic resonance (NMR), dielectric relaxation spectroscopy (DRS), differential scanning calorimetry (DSC), and water-sorption isotherm analysis were used to determine the molecular mobility of water. This chapter includes the interpretation of NMR data partially modified according to the comments of Prof. Peter Lillford at the Tenth International Symposium on the Properties of Water (ISOPOW 10).

Apparent Correlations Between the Stability of Solid APIs and the Molecular Mobility of Water Correlations between stability and water mobility are suggested for various APIs in the solid state. For example, the hydrolysis rate of cephalothin in freeze-dried formulations containing one of several excipients such as starch and methylcellulose appears 25

26

PART 1: Invited Speakers and Oral Presentations

0.02

kapp (h-1)

0.015 MC

0.01

SSTA STA

0.005

0

MCC

5

10

15

20

25

T1 (ms)

Figure 2.1. Correlation between the T1 of deuterium (2H2O) and the hydrolysis rate of cephalothin in freeze-dried formulations. MC, methyl cellulose; SSTA, water-soluble starch; STA, starch; and MCC, microcrystalline cellulose.

to be correlated with the molecular mobility of water as determined by the spin-lattice relaxation time (T1) of deuterium in the water molecule, as shown in Figure 2.1 (Aso and others 1997). The hydrolysis rate of flomoxef in gelatin gel containing kanamycin also exhibited an apparent correlation with the molecular mobility of water as determined by the T1 of 17O in the molecule (Yoshioka and others 1992). Not only chemical stability but also physical stability, such as crystallization of solid APIs during storage, appear to be correlated with the molecular mobility of water. The crystallization rate of amorphous nifedipine in solid dispersion formulations exhibited an apparent correlation with the T1 of deuterium in the water molecule, as shown in Figure 2.2. Similar correlations are observed for the stability of lyophilized protein formulations.

Molecular Mobility of Water in API Hydrates Water molecules in API hydrates exhibit a variety of physical states, suggesting a wide range of molecular mobilities. Because water of hydration plays important roles in determining the physical characteristics—such as solubility and flowability—of the API hydrate, an understanding of the molecular mobility of hydration water is critical in the formulation of API hydrates. Molecular Mobility of Hydration Water as Determined by NMR NMR has been used to determine the molecular mobility of water in the solid state and to examine the various mechanisms by which solids interact with water. However, there have been few studies in which the molecular mobility of water in API hydrates was determined using NMR. This may be because proton nuclear magnetic resonance (1H-NMR), even high-resolution 1H-NMR, cannot separate the peaks of the water

Water Mobility in Solid Pharmaceuticals

27

1.0

1/t 50 (h-1)

0.8

0.6

STA MCC

0.4

MC PVA

0.2 0.0

7

9

11

13

15

17

T1 (ms)

Figure 2.2. Correlation between the T1 of deuterium (2H2O) and the crystallization rate of amorphous nifedipine in solid dispersion formulations. STA, starch; MCC, microcrystalline cellulose; MC, methyl cellulose; and PVA, polyvinyl alcohol.

protons from those of the protons in other components, which prevents specific determination of water mobility. Although the preparation of API hydrate samples by using 17 O-labeled water enables the molecular mobility of the water molecules to be specifically determined by 17O-NMR, unaffected by the other components this approach is expensive and labor intensive. Thus, determination of the molecular mobility of hydration water in API hydrates by using NMR presents some challenges. However, the molecular mobility of hydration water in API hydrates can be determined by spin-spin relaxation measurement, providing the spin-spin relaxation time (T2) of the water protons is significantly different from that of the API protons. Furthermore, the T1 of the water protons may be a useful indicator of water mobility if the ratio of water protons to API protons is sufficiently large or if the water protons have a correlation time (τc) corresponding to the T1 minimum, such that the T1 of the water proton is sensitively reflected in the measured T1 value without being affected by spin diffusion between the water and the API protons. For example, in Na2HPO4 · 12H2O and Na2HPO4 · 2H2O, water protons are predominant (24/25 and 4/5, respectively). Na2HPO4 · 12H2O exhibited slower spin-spin relaxation (larger T2), as shown in Figure 2.3, and faster spin-lattice relaxation (smaller T1) (Figure 2.4), compared with Na2HPO4 · 2H2O, which indicates that either T1 or T2 can be used as an indicator for approximate comparison of the molecular mobility of hydration water (Yoshioka and others 2008). Moreover, even if the ratio of water protons to API protons is not particularly large, and even if the water proton does not have a τc corresponding to the T1 minimum, it may be possible to compare the molecular mobility of hydration water in API hydrates

28

PART 1: Invited Speakers and Oral Presentations

1400 1200

Intensity

1000 800 600

12H2O

400 200 0

2H2O 0

20

40

60

Time (μs)

Figure 2.3. Free-induction decay for Na2HPO4 · 12H2O and Na2HPO4 · 2H2O. 1.5 12H2O

1.0

Intensity

0.5

2H2O

0.0 -0.5 -1.0 -1.5

0

50

100

150

200

Pulse interval (s)

Figure 2.4. Spin-lattice relaxation for Na2HPO4 · 12H2O and Na2HPO4 · 2H2O.

based on measured T1 values if both the T1 of the API proton and the ratio of water protons to API protons are similar for all of the API hydrates compared. We measured spin-spin and spin-lattice relaxation for the 11 API hydrates listed in the Japanese Pharmacopoeia by using pulsed 1H-NMR to examine the possibility of determining the molecular mobility of hydration water in API hydrates by NMR relaxation measurement (Yoshioka and others 2008). Of the four antibiotic hydrates

Water Mobility in Solid Pharmaceuticals

Intensity

1500

29

ceftazidime

1000

500

0

0

20

40

60

80

100

80

100

Time (μs)

Intensity

1500

cefazolin sodium

1000

500

0

0

20

40

60

Time (μs)

Figure 2.5. Free-induction decay for ceftazidime and cefazolin sodium hydrates.

(cefazolin sodium, ceftazidime, amoxicillin, and ampicillin), all exhibited both Gaussian-type decay and Lorentzian decay, as exemplified by ceftazidime and cefazolin sodium hydrates (Figure 2.5). The other seven API hydrates exhibited only Gaussian-type decay, as exemplified by quinidine sulfate and scopolamine hydrobromide hydrates (Figure 2.6). The time courses of spin-spin relaxation observed for the four antibiotic hydrates were well fitted to Equation 2.1 by using the proportion of water protons calculated from the water content measured by the Karl Fischer method, as shown by the regression curve in Figure 2.5. I ( t ) = I 0 ( PG exp ( − (1 2 ) (t T2(G) ) ) + PL exp ( − t T2(L ) )) 2

(2.1)

30

PART 1: Invited Speakers and Oral Presentations

1500

Intensity

quinidine sulfate 1000

500

0

0

20

40

60

80

100

Time (μs)

1500

Intensity

scopolamine hydrobromide 1000

500

0

0

20

40

60

80

100

Time (μs)

Figure 2.6. Free-induction decay for quinidine sulfate and scopolamine hydrobromide hydrates.

where I(t) and I0 are signal intensity at time t and time 0, respectively. T2(G) and T2(L) are T2 for Gaussian decay and Lorentzian decay, respectively, and PG and PL are the proportion of protons that show Gaussian decay and Lorentzian decay, respectively. Therefore, all of the water protons in the molecule are considered to show Lorentzian decay, and the Gaussian decay is attributed to the drug protons. The other seven API hydrates did not exhibit Lorentzian decay, indicating that all water protons and drug protons in the molecule showed Gaussian decay. The T2 value of the water protons was calculated according to Equation 2.2, assuming that the T2 of the water protons is similar to that of the drug protons. I ( t ) = I 0 exp ( − (1 2 )( t T2 )

2

)

(2.2)

Water Mobility in Solid Pharmaceuticals

31

Ease of Evaporation for Hydration Water as Determined by DSC and Water-Sorption Isotherm Measurements The four antibiotic hydrates, which exhibited Lorentzian decay upon spin-spin relaxation, showed a single endothermic peak due to water evaporation, as shown in Figure 2.7a. The onset temperature was determined as a parameter for approximate comparison of ease of evaporation among the API hydrates, along with ease of evaporation under isothermal conditions as determined by water-vapor-sorption analysis (Figure 2.8). Onset temperature is known to depend on various factors, such as the heating rate, the shapes of the pan and lid, the surface area of the sample, and the flow rate of nitrogen gas. In this study, controllable factors such as the heating rate and the flow rate of nitrogen gas were kept constant, and a pan without a lid was used. The ease of evaporation for the four antibiotic hydrates as determined based on the observed onset temperature was in this order: ampicillin < amoxicillin < ceftazidime < cefazolin sodium. Berberine chloride, quinine hydrochloride, scopolamine hydrobromide, and saccharin sodium hydrates, which did not exhibit Lorentzian decay, showed two endothermic peaks, indicating the presence of two water populations: molecules that evaporate at high temperature and others that evaporate at lower temperature (Figure 2.7b). Pipemidic acid, sulpyrine, and quinidine sulfate hydrate, which did not exhibit Lorentzian decay, showed a single endothermic peak at a relatively high temperature (Figure 2.7c). The water-sorption isotherms observed for cefazolin sodium hydrates, which showed Lorentzian decay upon spin-spin relaxation, indicate that, during the desorption process, the water content decreased with decreasing humidity in the range of 90% relative humidity (RH) to 0% RH, with a significant slope in the plot of the water content versus humidity (Figure 2.8a). Pipemidic acid hydrate gave a water desorption isotherm in which the water content was constant over a wide range of humidity, as shown in Figure 2.8b. The water desorption isotherm observed for scopolamine hydrobromide showed flat lines at two levels of water content (Figure 2.8c). Correlation of Water Mobility as Determined by NMR with That as Determined by DSC and Water-Sorption Isotherm Measurements The T2 values determined based on the Lorentzian decay observed for hydration water in the four antibiotic hydrates increased as the onset temperature of the endothermic peak due to water evaporation decreased, as shown in Figure 2.9. This indicates that hydration water, which evaporates at lower temperatures, has greater molecular mobility as determined by T2, suggesting that ease of evaporation under nonisothermal conditions is correlated with T2. As exemplified by ceftazidime hydrate shown in Figure 2.10, T2 increased significantly with increasing temperature, indicating that T2 reflects the increases in molecular mobility associated with increases in temperature. Thus, molecular mobility can be considered to correlate with T2. As shown in Figure 2.11, antibiotic hydrates with

Figure 2.7. Differential scanning calorimetric (DSC) thermograms for active pharmaceutical ingredient (API) hydrates. W/g, watt per gram of sample.

32

(a)

Number of water molecules per hydrate molecule

6 5 4

cefazoline sodium

3 2 1 0

0

20 40 60 80 Relative humidity (%RH)

100

(b)

Number of water molecules per hydrate molecule

6 5 pipemidic acid

4 3 2 1 0

0

20 40 60 80 Relative humidity (%RH)

100

(c)

Number of water molecules per hydrate molecule

6

4 3 2 1 0

Figure 2.8. hydrates.

scopolamine hydrobromide

5

0

20 40 60 80 Relative humidity (%RH)

100

Water-sorption isotherms for active pharmaceutical ingredient (API)

33

34

PART 1: Invited Speakers and Oral Presentations

1.10

1.80

1.05

1.75

1.00

1.70

0.95

1.65

0.90 3.1

3.2

3.3

3.4

3.5

log T2 (μs)

log T2 (μs)

Figure 2.9. Correlation between onset temperature and T2 for four antibiotic hydrates. DSC, differential scanning calorimetry.

1.60 3.6

1000/T

Figure 2.10. Temperature dependence of T2 for ceftazidime hydrates (circles) and pipemidic acid (triangles) hydrates.

smaller T2 values showed a smaller change in T2 with temperature change. This finding suggests that lower values of T2 reflect a smaller scale of molecular motion of water, with lower activation energies. The values of T2 for the Gaussian decay observed for hydration water in the API hydrates other than the antibiotic hydrates did not vary significantly. Furthermore, the onset temperatures of the single endothermic peaks due to water evaporation for quinidine sulfate, pipemidic acid, and sulpyrine hydrates (Figure 2.7c), as well as

Water Mobility in Solid Pharmaceuticals

35

Figure 2.11. Correlation between T2 and temperature dependence of T2 for four antibiotic hydrates.

each of the two peaks due to water evaporation observed for quinine hydrochloride, scopolamine hydrobromide, saccharin sodium, and berberine chloride hydrates (Figure 2.7b), were not correlated with T2. These findings suggest that the molecular mobility of hydration water that shows Gaussian decay is too low to be reflected in T2. The endothermic peaks shown in Figure 2.7c seem to be due to hydration water with low molecular mobility, as suggested by the finding that their onset temperatures were not correlated with T2. Changes in T2 associated with changes in temperature were much smaller than those observed for the antibiotic hydrates that exhibited Lorentzian decay, as exemplified by pipemidic acid in Figure 2.10. This finding also indicates that the molecular mobility of hydration water is too low to be reflected in T2; thus, Gaussian decay rather than Lorentzian decay is observed, in contrast to the antibiotic hydrates. The water content versus humidity plot for pipemidic acid showed a flat line at three water molecules per hydrate molecule, and evaporation of these water molecules was observed under only very low humidity conditions (Figure 2.8b). This also supports the conclusion that the hydration water molecules show low molecular mobility. For the double endothermic peaks presented in Figure 2.7b, the peak at a higher temperature may be attributable to hydration water with low mobility, as observed for pipemidic acid hydrate, whereas the one observed at a lower temperature may be attributable to hydration water with higher mobility. The water content versus humidity plot for scopolamine hydrobromide showed flat lines at two levels of water content (Figure 2.8c), suggesting the presence of two water populations: molecules that evaporate at high humidity, and others that evaporate at lower humidity. This seems to be consistent with the observation of two endothermic peaks in DSC. The onset temperatures of evaporation for water molecules that easily evaporate at low humidity were lower than those observed for water molecules that exhibited Lorentzian decay, such as those of ceftazidime hydrate, indicating greater ease of evaporation. Such ease of

36

PART 1: Invited Speakers and Oral Presentations

evaporation may explain why Lorentzian decay was not observed in NMR relaxation measurements. Although the T1 of protons in the four antibiotic hydrates was determined, correlation between T1 and T2 was not observed. This finding indicates that for API hydrates containing a significant amount of drug protons, such as these antibiotic hydrates, the molecular mobility of hydration waters is not reflected in T1. Usefulness and Limitation of Determination of Water Mobility by NMR It was found that T2 is a useful parameter that can indicate the molecular mobility of water of hydration, which has relatively high mobility and shows Lorentzian decay upon spin-spin relaxation. For these water molecules, molecular mobility as determined by T2 is correlated with ease of evaporation under both nonisothermal and isothermal conditions, such that water molecules with greater ease of evaporation have higher T2 values. In contrast, for hydration water that has low mobility and shows Gaussian decay, T2 was found not to correlate with ease of evaporation under nonisothermal conditions, suggesting that molecular motion that determines the ease of evaporation is not reflected in T2; in this case, T2 cannot be used as a parameter to indicate molecular mobility. The water molecules in the API hydrates were found to have wide-ranging molecular mobilities, from low molecular mobility that could not be evaluated by NMR relaxation times, such as the water molecules in pipemidic acid hydrate, to high molecular mobility that could be evaluated by NMR relaxation times, such as the water molecules in ceftazidime hydrate.

Molecular Mobility of Water Coexisting with Various Pharmaceutical Excipients Pharmaceutical excipients exhibit a variety of water-sorption behaviors, as shown in Figure 2.12, which compares water contents at 60% RH and 10% RH determined for widely used pharmaceutical excipients (Yoshioka and others 2007). Alpha-cornstarch and alpha-potato starch absorb a similar amount of water as povidone (PVP) at 10% RH but much less at 60% RH than PVP. Thus, this figure indicates a variety of watervapor-sorption behaviors among excipients. The molecular mobility of water absorbed in dextran, methylcellulose, and PVP was compared by DRS (Yoshioka and others 1999). For each of the water-excipient mixtures, the water content of which was adjusted to be approximately the same, two relaxation processes were observed at frequencies of 108–109 Hz and 109–1010 Hz, indicating the presence of two water populations, one with high mobility and one with lower mobility. As shown in Table 2.1, the ratio of high-mobility water to lowermobility water calculated by curve-fitting of spectra was the smallest for dextran and the largest for PVP. These data indicate that pharmaceutical excipients contain water molecules with widely ranging ratios of high and low mobilities, even at a similar water content.

PVP cros-PVP CMC-Na CMS-Na cros-CMC-Na CMC-Ca Potato starch Cornstarch Pullulan alpha-Cornstarch 60% RH 10% RH

alpha-Potato starch L-HPC Dextrin CMC PVP/VA Dextrin MC Powdered cellulose MCC HPC HPMC 50

0

100

150

200

250

Water sorbed (mg/g)

Figure 2.12. Water contents of various pharmaceutical excipients at 10% RH and 60% RH. cros, cross-linked; Na, sodium; Ca, calcium; CMC, carboxymethyl cellulose (carmellose); CMS, carboxymethyl starch; HPC, hydroxypropyl cellulose; HPMC, hydroxypropylmethyl cellulose; L-HPC, low-density hydroxypropyl cellulose; MC, methyl cellulose; MCC, microcrystalline cellulose; PVP, povidone; and PVP/VA, copovidone. Table 2.1. The ratio of high-mobility water to low-mobility water in water-excipient mixtures Water content (g/g)

% RH

[H2O]h/[H2O]l

Dextran

0.29

86

0.93

MC

0.23

86

1.7

MC

0.38

98

3.3

PVP

0.34

75

4.4

MC, methylcellulose; and PVP, povidone.

37

38

PART 1: Invited Speakers and Oral Presentations

Concluding Remarks Water molecules present in solid APIs and those coexisting with pharmaceutical excipients have widely ranging molecular mobilities, as determined by NMR, DRS, DSC, and water-sorption isotherm measurements. Therefore, it is very important to gain insight into the molecular mobility of water in the formulation of solid APIs.

References Aso Y, Sufang T, Yoshioka S, Kojima S. 1997. Amount of mobile water estimated from 2H spin-lattice relaxation time, and its effects on the stability of cephalothin in mixtures with pharmaceutical excipients. Drug Stability 1:237–42. Shalaev EY, Zografi G. 1996. How does residual water affect the solid-state degradation of drugs in the amorphous state? J Pharm Sci 85:1137–41. Yoshioka S, Aso Y, Kawanishi T. 2007. Water sorption isotherms of pharmaceutical excipients listed in Japanese Pharmacopoeia. Pharm Regul Sci 38:228–34. Yoshioka S, Aso Y, Kojima S. 1999. The effect of excipients on the molecular mobility of lyophilized formulations, as measured by glass transition temperature and NMR relaxation-based critical mobility temperature. Pharm Res 16:135–40. Yoshioka S, Aso Y, Osako T, Kawanishi T. 2008. Wide-ranging molecular mobilities of water in active pharmaceutical ingredient (API) hydrates as determined by NMR relaxation times. J Pharm Sci 97:4258–68. Yoshioka S, Aso Y, Terao T. 1992. Effect of water mobility on dug hydrolysis rates in gelatin gels. Pharm Res 9:607–12.

Oral Presentations

3 The Effect of Water and Fat Contents on the Enthalpy of Dissolution of Model Food Powders: A Thermodynamic Insight A. Marabi, A. Raemy, A. Burbidge, R. Wallach, and I. S. Saguy Abstract The dissolution process of food powders is a topic of vast practical and commercial importance. However, the physical and chemical processes involved are far from being fully understood. Calorimetric determination of dissolution enthalpies (ΔHdiss) of food powders enables the extent of solute-solvent interactions to be assessed in terms of thermodynamic parameters. In this context, this work studied the effect of water and fat content representing typical food powders on the enthalpy of dissolution measured by isothermal calorimetry. The model food powders were obtained by freeze-drying a mixture of skim-milk powder, maltodextrin, and variable amounts of butter fat. The moisture content (MC) was determined gravimetrically, and the enthalpy of dissolution was measured by isothermal calorimetry at 30°C. The powders were studied in the dry state and after equilibration at different water activity (aw) values. Near-infrared spectroscopy and X-ray powder diffraction data indicated that the powders were in an amorphous state after freeze-drying. Crystallization of lactose was observed after exposure of the powders to the higher aw investigated. The enthalpy of dissolution decreased (less exothermic) with increasing fat content in the dry samples (MC = ∼1% dry basis [db]), ranging from −62 to −34 J/g for powders containing 0.7 and 45% fat (wet basis [wb]), respectively. The MC also had a significant effect on the enthalpy of dissolution. The exothermic enthalpy decreased from −62 to −5 J/g for powders containing 0.7% fat (wb) in the dry state (MC = 1.7% db) and after equilibration at 0.54 aw (MC = 11.4% db), respectively. A similar trend was observed for all other samples, for which the exothermic ΔHdiss of the equilibrated samples decreased with an increase in the MC.

Introduction The study of the dissolution process of powders started more than a century ago, possibly with the classic work of Noyes and Whitney (1897). The dissolution kinetics of powders are of critical importance in many applications, including food and pharmaceutical products. In the food industry, in particular, it is one of the significant factors that defines the quality of the product (Marabi and others 2007b). Frequently, the aim is to obtain a powder with good dissolution and reconstitution properties. The thermodynamic aspects of the dissolution of powders can be addressed by isothermal solution calorimetry, in which the heat of solution is measured. The output 41

42

PART 1: Invited Speakers and Oral Presentations

is a composite of wetting responses, liquid penetration, dissolution phenomena, and any other interaction process that might occur (Buckton 1995; Gao and Rytting 2006). The significantly different dissolution behavior of crystalline and amorphous saccharides has also been studied (Miller and others 1997; Miller and de Pablo 2000; Salvetti and others 2007). All of the crystalline samples examined presented endothermic enthalpies of dissolution. Conversely, the same samples in the amorphous state showed an exothermic response. The effect of the moisture content (MC) on the thermodynamic response of dissolving powders was studied, and significantly less exothermic values were reported for amorphous lactose samples stored at increasing humidity (Hogan and Buckton 2000). As composition and especially moisture and fat contents are expected to affect the enthalpy of dissolution, the objective of the present study was to quantify the effects of increasing moisture and fat contents on the thermodynamic response measured during the dissolution process of a model food powder. It should be noted that the term dissolution is used in the present work, but other possible terms for this process have been suggested and discussed during this symposium. It was argued that dissolution applies to the process in which a crystalline material is put into solution and does not apply for amorphous materials because they can be defined as a solid solution. Some of the suggested terms included liquefying, rehydration, dilution of the glass, and redissolution or redispersion. However, there was no clear agreement, so we retained the term dissolution for this work. Clearly, this issue should be discussed, including questions regarding, for example, what would be the term to use for a powder containing crystalline and amorphous domains.

Experimental Methods Five samples containing different amounts of fat were produced by freeze drying (FD). The materials used included maltodextrin DE 21 (MD), skim-milk powder (SMP), and butter fat. The composition of the wet and dry mixes is listed in Table 3.1. The materials were mixed in water at 50°C and then homogenized in two steps at 50 and 250 bar, respectively. The samples were frozen at −55°C, FD at 10−6 bar for 48 h with

Table 3.1. Characterization of the model food powders Sample

A

B

C

D

E

Wet mix before freeze drying Butter fat (% wt/wt) Maltodextrin (% wt/wt) Skim-milk powder (% wt/wt) Water (% wt/wt)

0

5

10

15

20

14

14

14

14

14

7

7

7

7

7

79

74

69

64

59

14.3

29.3

35.7

45.0

Freeze-dried powder Fat (% wt/wt)

0.7

Water activity

0.06

0.05

0.05

0.05

0.05

Water content (g/100 g dry solids)

1.71

0.96

0.70

0.64

0.58

Effect of Water and Fat on Enthalpy of Dissolution

43

a stepwise temperature increase from −50° to 25°C and ground in a mill yielding a powder passing 30 mesh. All the samples were transferred to hermetically sealed aluminum pouches to avoid moisture uptake, and stored for at least 1 week prior to the calorimetric studies for internal equilibration of water within the matrix. The FD samples were then exposed to different saturated salt solutions (water activity [aw] values: 0.11, 0.22, 0.33, 0.43, and 0.54) in desiccators at 20° ± 1°C until no significant weight change was observed (∼4 weeks). The fat content of the different samples of the model powder was determined by the Mojonnier method (AOAC 2006). The MC was determined by exposing the samples at 102°C in vacuum to phosphorus pentoxide. The enthalpy of dissolution was quantified isothermally (30°C) in a Calvet calorimeter (Marabi and others 2007a, 2007b). The results are expressed in J/g of total sample weight (i.e., including the MC). The calorimetric curves obtained are integrated, yielding the heat (J/g) adsorbed or released during the complete dissolution of the solid sample. A negative value for the enthalpy of dissolution indicates the release of heat (exothermic process), and positive values indicate an absorption of heat (endothermic process).

Results and Discussion The freeze-dried particles presented a typical flakelike shape resembling broken glass, and increasing amounts of fat were observed at the surface of the particles (i.e., by scanning electron microscopy; data not shown). The X-ray powder diffraction (XRPD) patterns of the powders confirmed that all the samples were in the amorphous state after FD, as observed by the broad and diffuse halo with no sharp peaks (data not shown). The dissolution of all the powders at all the conditions tested produced exothermic responses. The effect of increasing the amount of fat in the samples is clearly related to a decrease in the amount of heat released during the dissolution process (Figure 3.1). The decrease in the enthalpy of dissolution (i.e., a less exothermic process) ranged from about −62 to −34 J/g for samples with 0.7% and 45% fat, respectively. The enthalpy of dissolution of pure butter fat was also measured, and a slight endothermic response was observed (1.7 J/g). This small value is expected because, when two immiscible compounds are mixed, only a very small calorimetric response representing the heat of immersion is measured. Therefore, the immersion of this ingredient is responsible for lowering the enthalpy of dissolution of the powders. To elucidate the effect of the fat content further, the enthalpy of dissolution measured for all aw was normalized by the amount of nonfat solids in the samples. Figure 3.2 clearly shows that the enthalpy of dissolution is independent of the amount of fat in the powders, because a single curve was obtained for all the conditions tested. Consequently, the exothermic response can be assumed to arise from the dissolution of the MD and the SMP ingredients. We have also shown the effects of the physical state of MD and SMP on the enthalpy of dissolution (Marabi and others 2007b). For FD amorphous samples, large exothermic responses were observed, whereas both

Figure 3.1. Typical dissolution calorimetry curves of the freeze-dried samples (aw ≤ 0.06) containing different amounts of fat. EXO, exothermic responses.

Figure 3.2. Enthalpy of dissolution of all the samples tested as a function of the water activity. Note that the enthalpy of dissolution (ΔHdiss) is normalized by the amount of nonfat content. Exo, exothermic responses.

44

Effect of Water and Fat on Enthalpy of Dissolution

45

Figure 3.3. Typical dissolution calorimetry curves of samples containing 14.3% fat equilibrated at various water activities. EXO, exothermic responses.

increased MC and recrystallized lactose in the SMP resulted in decreased exothermic responses. The different behavior of crystalline and amorphous powders is attributed to the higher entropy and internal free energy of the metastable amorphous state, leading to an enhanced dissolution rate and chemical reactivity, relative to the thermodynamically more favorable and stable crystalline state (Hancock and Zografi 1997; Hancock and Parks 2000; Hancock 2002; Wong and others 2006). When the enthalpy of dissolution was measured for the powders equilibrated at increasing aw, a decrease in the exothermic response was also observed (Figure 3.3). The ΔHdiss for the samples with 0.7% fat decreased from about −62 for the sample with an MC of 1.7% (FD) to −5 J/g for the sample with an MC of 11.2% (aw 0.54). Similar trends were observed for all other fat contents (Figure 3.4). The highly exothermic ΔHdiss values for the FD samples are possibly due to the hydration of hydrophilic groups in the powders. This effect was explained in terms of the exothermic nature of the wetting process, resulting in a less exothermic response for samples with higher MC as compared to a completely dry sample. It is often argued that the sorption of water by a dry powder is the first stage of wetting and that the first few molecules adsorbed on the surface are responsible for the greatest part of the overall exothermic wetting response (Buckton 1995; Hancock and Shamblin 1998; Hancock and Dalton 1999; Hogan and Buckton 2000). Increasing the fat concentration on the particles’ surface and the amount of adsorbed water leads to a reduced number of hydrogen-bonding sites available for additional sorption of water molecules during the dissolution process. Consequently, less

46

PART 1: Invited Speakers and Oral Presentations

Figure 3.4. Enthalpy of dissolution (ΔHdiss) of all the samples tested as a function of their water and fat contents. db, dry basis; and EXO, exothermic responses.

exothermic values of ΔHdiss were observed for samples equilibrated at higher aw. Furthermore, the recrystallization of lactose at aw values higher than 0.43 (as observed by near-infrared spectroscopy and XRPD; data not shown) also contributes to a less exothermic overall response. In another publication (Marabi and others 2007b), we have shown that less exothermic responses are related to slower dissolution rates. Apart from the common wettability problems arising from the presence of fat on the surface of the powders, the current results indicate that the thermodynamics of the process might play a key role and could therefore be used to better characterize and predict the dissolution of food powders.

Conclusions The usefulness of the enthalpy of dissolution for studying the thermodynamic behavior of soluble food powders was demonstrated. Valuable information previously unavailable was obtained, providing a new insight into the dissolution process of food powders. These data complement those obtained with in situ kinetic measurements, providing the basis for better understanding the coupling between heat and mass transfer during the dissolution process. The effects of the fat and MC on the enthalpy of dissolution were elucidated. Higher water and fat contents led to less exothermic responses. This behavior is due to the negligible endothermic enthalpy of the fat, and the decrease in exothermic enthalpy measured as fewer hydrogen bonds are formed

Effect of Water and Fat on Enthalpy of Dissolution

47

during the dissolution in powders with higher MC. In addition, the recrystallization of lactose in the samples contributes to an endothermic response, thus lowering the overall exothermic value measured. The present data highlight that the dissolution is a complex phenomenon in which local heat transfer can play an important role.

References AOAC (Association of Official Analytical Chemists) 2006. Official methods of analysis, 16th ed. Washington, DC: Association of Official Analytical Chemists. Buckton G. 1995. Applications of isothermal microcalorimetry in the pharmaceutical sciences. Thermochim Acta 248:117–29. Gao D, Rytting JH. 2006. Use of solution calorimetry to determine the extent of crystallinity of drugs and excipients. Int J Pharm 151:183–92. Hancock BC. 2002. Disordered drug delivery: destiny, dynamics and the Deborah number. J Pharm Pharmacol 54:737–46. Hancock BC, Dalton CR. 1999. The effect of temperature on water vapor sorption by some amorphous pharmaceutical sugars. Pharm Dev Technol 4:125–31. Hancock BC, Parks M. 2000. What is the true solubility advantage for amorphous pharmaceuticals? Pharm Res 17:397–404. Hancock BC, Shamblin SL. 1998. Water vapour sorption by pharmaceutical sugars. Pharm Sci Technol Today 1:345–51. Hancock BC, Zografi G. 1997. Characteristics and significance of the amorphous state in pharmaceutical systems. J Pharm Sci 86:1–12. Hogan SE, Buckton G. 2000. The quantification of small degrees of disorder in lactose using solution calorimetry. Int J Pharm 207:57–64. Marabi A, Mayor G, Burbidge A, Wallach R, Saguy IS. 2007a. Assessing dissolution kinetics of powders by a single particle approach. Chem Eng J 139:118–27. Marabi A, Mayor G, Raemy A, Bauwens I, Claude J, Burbidge A, Wallach R, Saguy IS. 2007b. Solution calorimetry: a novel perspective into the dissolution process of food powders. Food Res Int 40:1286–98. Miller DP, de Pablo JJ. 2000. Calorimetric solution properties of simple saccharides and their significance for the stabilization of biological structure and function. J Phys Chem [B] 104:8876–83. Miller DP, de Pablo JJ, Corti H. 1997. Thermophysical properties of trehalose and its concentrated aqueous solutions. Pharm Res 14:578–90. Noyes A, Whitney WR. 1897. The rate of solution of solid substances in their own solutions. J Am Chem Soc 19:930–4. Salvetti G, Tognoni E, Tombari E, Johari GP. 2007. Excess energy of polymorphic states or glass over the crystal state by heat of solution measurement. Thermochim Acta 285:243–52. Wong SM, Kellaway IW, Murdan S. 2006. Enhancement of the dissolution rate and oral absorption of a poorly water soluble drug by formation of surfactant-containing microparticles. Int J Pharm 317:61–8.

4 “Solvent Water” Concept Simplifies Mathematical Modeling in Fermenting Dough, a Multiphase Semisolid Food S. M. Loveday and R. J. Winger

Abstract The shelf life of frozen dough depends on retention of yeast activity through frozen storage. It is difficult to study yeast cells in situ within the dough because they are embedded in a fine semisolid network structure. Mathematical modeling can simulate the environment around yeast cells and provide information about their interaction with that environment. This mathematical model predicts sugar consumption in fermenting dough, based on a mechanistic model of yeast cell fermentation in liquid broth. Adapting the liquid model to dough required knowledge of the water distribution and availability in dough. The solvent-water concept assumes that water in the presence of biopolymers and solutes can be divided into (a) solvent water, which has physical properties equivalent to pure liquid water; and (b) nonsolvent water, which has altered physical properties because of its interactions with solids and solutes. This assumption facilitated construction of an accurate model for sugar consumption in fermenting dough.

Introduction The yeast-leavened frozen-dough segment of the bakery industry supplies foodservice, commercial, and domestic consumers with products that require only thawing, proofing, and baking. In the proofing (AKA proving) stage, the dough is held under warm, humid conditions and expands due to the production of carbon dioxide by yeast cells fermenting sugars. The shelf life of frozen dough is limited by the decline of post-thaw proofing power. This is primarily because of a loss of yeast activity and carbon dioxide production. Inadequate proofing power means a frozen dough takes too long to proof or will not proof to adequate volume in the required time. If doughs are fermented (deliberately or accidentally) before freezing, the loss in proofing power is much faster and shelf life is dramatically shortened. The mechanistic basis for this effect is poorly understood, partly because of the colloidal-scale complexity, semisolid texture, and dynamic nature of dough. In most baking research, sugar concentrations in dough are expressed as weight percentages, which give no indication of concentration in different phases or locations within a dough piece. Comparisons with industrial fermentation literature would be easier if concentrations could be expressed as molarities. 49

50

PART 1: Invited Speakers and Oral Presentations

Detailed mechanistic knowledge of dough fermentation requires an understanding of the physicochemical environment around yeast cells. Here we have sought to understand that environment via knowledge of the distribution and state of water in dough. We have mathematically modeled yeast substrate flux in fermenting dough in a way that facilitates comparisons with kinetic studies in other liquid and semisolid systems.

Dough Structure Dough constitutes a hierarchy of dispersed gaseous, liquid, and solid phases (Schiraldi and Fessas 2003). On the macroscopic scale, it is a foam of gas cells dispersed in a continuous semisolid viscoelastic phase. The semisolid phase is further phaseseparated on the colloidal scale into a bicontinuous network of hydrated gluten proteins and an aqueous phase containing water-soluble proteins and oligosaccharides, low molecular weight solutes, starch granules, and yeast cells (Eliasson and Larsson 1993). The colloidal phase separation is driven by hydrophobic association of gluten proteins and the immiscibility of gluten proteins with polysaccharides, soluble proteins, and starch (Tolstoguzov 1997). Water is the continuous medium of both phases, and its partitioning between phases is affected by dissolved solutes such as salts and sugars.

Solvent-Water Concept In the middle of the 20th century, the term nonsolvent water emerged from the realization that “the concentration of a solute in an aqueous solution is increased by the immersion of a solid [because] a certain amount of water, preferentially sorbed by the solid, has not become available as solvent” (Lindenberg and others 1963). This formed the basis of a widespread method for measuring water sorption by a solid material (Lindenberg and others 1963; Gary-Bobo 1967). The effective concentration of a probe solute in a solid-solute-water system was determined via accurate measurement of a colligative property such as vapor pressure. Since the amount of solute and its effective concentration were known, the amount of water in the solutewater part of the tertiary system could be calculated. This approach relied on the working assumption that total water could be cleanly delineated into solvent and nonsolvent fractions. That assumption was recognized as an oversimplification (Gary-Bobo 1967), but was useful in the absence of detailed knowledge about how and where water interacted with polymers or solutes. It is now thought that water interacts with solids and solutes via hydrogen bonding, dipole-dipole and ion-dipole forces, van der Waals interactions, and hydrophobic hydration (Franks 2000). The multiplicity of interaction types, strengths, time scales, and orientations create an extremely complex picture on the molecular scale. Most foods contain water, hydrophilic or amphiphilic biopolymers, and low molecular weight solutes. In intermediate-moisture and high-moisture foods, a proportion of total water behaves identically to pure water; this fraction is often termed bulk water

“Solvent Water” Concept Simplifies Mathematical Modeling

51

(Garti and others 2001). The remaining water exhibits different physical properties than pure water (e.g., fugacity, heat capacity, and freezing point) by virtue of its interactions with solutes or polymers. Moreover, it is not homogeneous, as previously thought. The water fraction that does not behave like pure water has attracted names such as unfreezable and bound. These are ill-defined and misleading terms (Franks 1991), but nonetheless have gained widespread use. Although the term bulk water has attracted relatively little criticism, it gives the impression of excluding water that cannot be readily separated from a sample by physical means; for example, water in capillaries. Solvent water is a less ambiguous alternative. For the purposes of mathematical modeling, dividing water into two fractions based on similarity to pure water, although not strictly representative of the real situation, reduces the degree of complexity required in model equations.

Solvent Water in Dough Studies of water in dough have found several populations of water molecules with different physical properties. Solvent water appears only above a certain looselydefined water-content threshold. Below 20–25 g H2O per 100 g dry matter (g H2O · [100 g DM]−1) or water activity (aw) of ∼0.8, the latent heat of water sorption by flour is greater than the latent heat of condensation (Bushuk and Winkler 1957; Rückold and others 2003), indicating that water interacts with polymers in flour (proteins, starch, pentosans). The latent heat of adsorption approaches the latent heat of condensation at 20–30 g H2O · (100 g DM)−1, indicating capillary condensation of solvent water (Bushuk and Winkler 1957; Ruckold and others 2003). Moisture-sorption isotherms curve sharply upward in this region (Bushuk and Winkler 1957; Roman-Gutierrez and others 2002a). Recent nuclear magnetic resonance studies have fitted proton-relaxation data from dough with continuous spectra featuring three distinct peaks (Ruan and Chen 1998; Ruan and others 1999; Kou and others 2002; Esselink and others 2003). Only two proton populations with low and intermediate mobility are observed below 23 g H2O · (100 g DM)−1 (Ruan and others 1999). Increasing moisture content to 35 g H2O · (100 g DM)−1 leads to (a) the disappearance of the low-mobility proton population; (b) enlargement of the intermediate-mobility population, which shifts to slightly lower mobility; and (c) the appearance of one or more high-mobility populations (Ruan and Chen 1998; Ruan and others 1999). The high-mobility population fits the description of solvent water. Differential thermal analysis (DTA) (Davies and Webb 1969; Bushuk and Mehrotra 1977) and differential scanning calorimetry (DSC) experiments (Roman-Gutierrez and others 2002b) have reported the appearance of water freezable at −50°C only where the water content of dough is more than 30–33 g H2O · (100 g DM)−1. These findings should be interpreted cautiously in light of objections raised by Hatley and others (1991) to such methodologies. Davies and Webb (1969) concluded that all water

52

PART 1: Invited Speakers and Oral Presentations

above 33 g H2O · (100 g DM)−1 was freezable at −50°C, in agreement with the magnetic resonance study by Toledo and others (1968). These results from independent research groups using different methodologies are in quite good agreement that solvent water, as described here, is not present until the water content of dough exceeds 0.3 g of water per gram of flour DM. In this study, it was assumed that all water above 0.3 g per g DM was solvent water.

Materials and Methods Dough Manufacture Full details of dough manufacture and sugar extraction are provided by Loveday and Winger (2007). Each 1-kg batch of dough contained 600 g high-grade white wheat flour (11.5% protein, 12.0% moisture), 330 g water, 20 g compressed yeast (67.3% water), 20 g canola oil, 10 g iodized salt, and 20 g glucose, fructose, or sucrose. Ingredients were mixed for 18 min in a domestic breadmaker. Doughs were divided by hand into 50-g pieces that were shaped into slabs, sealed in plastic bags, and fermented at 30°C in a temperature-controlled room for 0–180 min. Sugar Extraction and Analysis Approximately 30 g of dough was quench-frozen in liquid nitrogen and shattered to a powder in a precooled laboratory Waring blender. A 2.5-g sample of frozen dough powder was added to 25 mL of water at room temperature and homogenized for 30 s. An aliquot of homogenate was filtered through a 0.8-μm filter into a 1.5-mL plastic tube. Extracts were held at 85°C in a water bath for 1 h to inactivate enzymes, cooled on ice for 10 min, and centrifuged. The supernatant was analyzed for glucose, fructose, and sucrose content by enzymatic assays (BioAnalysis kits; Roche Diagnostics, Mannheim, Germany). Results were expressed as millimoles (mmol) per 100 g dough (fresh weight). Duplicate analyses were performed on three independent doughs. Sugar uptake rates were expressed as differential equations and solved numerically with Matlab version 6.5 release 13 (Mathworks, Natick, MA, USA).

Results and Discussion The aqueous phase of dough is analogous to the liquid medium in broth culture; that is, a solution of sugars and other nutrients in which yeast cells are suspended. Glucose uptake was described with a hyperbolic equation (Equation 4.1) derived from a mechanistic description of sugar-carrier proteins traversing the cell membrane (CornishBowden 1995). VG = Vmax G ⋅ X ⋅

G G + KG

(4.1)

VG is the rate of glucose uptake (mmol · L−1 · h−1), Vmax G is the maximum specific rate of glucose uptake (mmol · g biomass−1 · h−1), X is yeast biomass concentration

“Solvent Water” Concept Simplifies Mathematical Modeling

53

(g biomass · L−1), G is the concentration of glucose (mmol · L−1), and KG is the affinity constant (mmol · L−1). Saccharomyces cerevisiae cells consume glucose in preference to fructose when both are available (Serrano and De la Fuente 1974). In broth cultures fed with a mixture of glucose and fructose, the influence of glucose on fructose uptake is well fitted by Equation 4.2 (Barford and others 1992), which is derived from mechanistic enzyme competitive inhibition equations commonly seen in enzymology textbooks (e.g., Cornish-Bowden 1995). VF = Vmax F ⋅ X ⋅

F G ⎞ ⎛ F + KF ⎜ 1 + ⎟ ⎝ K GF ⎠

(4.2)

KGF is a constant describing competitive inhibition of fructose uptake by glucose (mmol · L−1), while F, VF, Vmax F, and KF are analogous to G, VG, Vmax G, and KG in Equation 4.1. The total moisture present in doughs, calculated from the moisture content of ingredients, was 41.55 g · (100 g dough)−1. Doughs contained 60 g flour per 100 g dough, and flour contained 12.0% moisture. According to the literature, nonsolvent water comprised 0.3 g per 100 g flour solids; that is, 0.3 × 60 (1 − 0.12) = 15.84 g · (100 g dough)−1. Solvent water was therefore 41.55 − 15.84 = 25.71 g · (100 g dough)−1. It was assumed that all the solvent water was in the aqueous phase and that the density of solvent water was 103 g · L−1. The latter assumption enabled conversion from molalities to molarities so that kinetic constants were in the same units as those in the fermentation literature. Glucose and fructose contents expressed in mmol · (100 g dough)−1 were converted to molarities in the aqueous phase by dividing by 25.71 × 10−3 L · (100 g dough)−1. It was assumed that the solids in compressed yeast (32.7% solids) comprised biomass only. Biomass concentration in the solvent water (X) was calculated from the amount of compressed yeast and the volume of solvent water in dough. Cell numbers were assumed to remain constant, in agreement with the low biomass yield from anaerobic fermentation (Chen and Chiger 1985). In yeasted doughs made with 2% added glucose or fructose only, sugar was consumed at a constant rate during fermentation at 30°C. Vmax G and Vmax F were calculated from the rate of decline of each sugar during 90 min of fermentation at 30°C. Sucrose was almost completely hydrolyzed during mixing (Loveday and Winger 2007). Initial values of G and F (designated Gi and Fi) were calculated from glucose and fructose concentrations immediately after mixing in doughs made with 2% added sucrose. Experimental parameters are summarized in Table 4.1. KG, KF, and KGF were initially set at the values reported by Barford and others (1992) and then modified to improve fit to experimental results. Equations 4.1 and 4.2 fitted glucose and fructose consumption well (Figure 4.1). KG was two orders of magnitude higher than the value used by Barford and others (1992) but of similar magnitude to inhibition coefficients

Figure 4.1. Concentration of glucose (top) and fructose (bottom) in dough made with 2% yeast and 2% added sucrose fermenting at 30°C. Points are experimental data (mean ± SE of six assay results) and lines are fits from Equations 4.1 and 4.2. Sugar concentrations are expressed as mmol · L−1 in the solvent water (left, y-axis) or as a weight percentage of dough on a wet basis (right, x-axis). KG, affinity constant; and KGF, a constant describing competitive inhibition of fructose uptake by glucose.

54

“Solvent Water” Concept Simplifies Mathematical Modeling

55

Table 4.1. Summary of mathematical model parameters Symbol

Description

Unit

Value

Gi

Initial glucose concentration

mmol · L−1

250

Fi

Initial fructose concentration

mmol · L−1

389

X

Biomass concentration

(g biomass) · L−1

Vmax G

Maximal glucose uptake rate

mmol · (g biomass)−1 · h−1

6.53

Vmax F

Maximal fructose uptake rate

mmol · (g biomass)−1 · h−1

6.26

25.4

Table 4.2. Affinity constants for glucose uptake by Saccharomyces cerevisiae reported in various studies Reference

KG (mmol · L−1)

Glucose concentration (mmol · L−1)

Maier and others 2002 (HXT1)

129

111

Serrano and De la Fuente 1974

100

100

Elbing and others 2004

76

111

This work

50

250

Reifenberger and others 1997

46

100

Maier and others 2002 (HXT3)

34.2

111

Barford and others 1992

0.5

55.6

KG, affinity constant; HXT1 and HXT3, acronyms for specific glucose transporter proteins in S. cerevisiae.

in other reports (Table 4.2) (Serrano and De la Fuente 1974; Bisson and Fraenkel 1983; Reifenberger and others 1997; Maier and others 2002; Elbing and others 2004). At 0.35 mmol · L−1, KFG was similar to other reports of 0.27 (Herwig and others 2001) and 1.0 (Barford and others 1992).

Conclusions Dividing water in semisolid foods into solvent and nonsolvent is a simplification, but a useful one. There is quite good consensus among researchers that in dough, ∼0.3 g water per gram of dry matter has the characteristics of nonsolvent water. The assumption that the remainder was solvent water simplified mathematical modeling and facilitated the use of liquid system models to predict sugar-fermentation rates in dough.

References Barford JP, Phillips PJ, Orlowski JH. 1992. A new model of uptake of multiple sugars by S. cerevisiae (part 1). Bioproc Eng 7:297–302. Bisson L, Fraenkel DG. 1983. Involvement of kinases in glucose and fructose uptake by Saccharomyces cerevisiae. Proc Natl Acad Sci USA 80:1730–4. Bushuk W, Mehrotra VK. 1977. Studies of water binding by differential thermal analysis. II. Dough studies using the melting mode. Cereal Chem 54:320–5.

56

PART 1: Invited Speakers and Oral Presentations

Bushuk W, Winkler CA. 1957. Sorption of water vapor on wheat flour, starch, and gluten. Cereal Chem 34:73–86. Chen SL, Chiger M. 1985. Production of baker ’s yeast. In: Moo-Young M, editor. Comprehensive biotechnology: the principles, applications, and regulations of biotechnology in industry, agriculture, and medicine. Oxford: Pergamon. p 429–62. Cornish-Bowden A. 1995. Fundamentals of enzyme kinetics. London: Portland. Davies RJ, Webb T. 1969. Calorimetric determination of freezable water in dough. Chem Ind (London) 1969:1138–9. Elbing K, Larsson C, Bill RM, Albers E, Snoep JL, Boles E, Hohmann S, Gustafsson L. 2004. Role of hexose transport in control of glycolytic flux in Saccharomyces cerevisiae. Appl Environ Microbiol 70:5323–30. Eliasson AC, Larsson K. 1993. Cereals in breadmaking: a molecular colloidal approach. New York: Marcel Dekker. Esselink EFJ, Van Aalst H, Maliepaard M, Van Duynhoven JPM. 2003. Long-term storage effect in frozen dough by spectroscopy and microscopy. Cereal Chem 80:396–403. Franks F. 1991. Hydration phenomena: an update and implications for the food processing industry. In: Levine H, Slade L, editors. Water relationships in foods: advances in the 1980s and trends for the 1990s. New York: Plenum. p 1–19. Franks F. 2000. Water: a matrix of life. 2nd ed. Cambridge: Royal Society of Chemistry. Garti N, Asering A, Fanun M, Leser ME, Ezrahi S. 2001. Sub-zero temperature behavior of water in W/O microemulsions. In: Berk Z, Leslie RB, Lillford PJ, Mizrahi S, editors. Water science for food, health, agriculture and environment. International Symposium on the Properties of Water in Foods 8. Lancaster, PA: Technomic. p 97–124. Gary-Bobo CM. 1967. Nonsolvent water in human erythrocytes and hemoglobin solutions. J Gen Physiol 50:2547–64. Hatley RHM, Van den Berg C, Franks F. 1991. The unfrozen water content of maximally freezeconcentrated carbohydrate solutions: validity of the methods used for its determination. Cryo Lett 12:113–24. Herwig C, Doerries C, Marison I, von Stockar U. 2001. Quantitative analysis of the regulation scheme of invertase expression in Saccharomyces cerevisiae. Biotechnol Bioeng 76:247–58. Kou Y, Ross EW, Taub IA. 2002. Microstructural domains in foods: effect of constituents on the dynamics of water in dough, as studied by magnetic resonance spectroscopy. In: Levine H, editor. Amorphous food and pharmaceutical systems. Cambridge: Royal Society of Chemistry. p 48–58. Lindenberg AB, Dang-Vu-Bien, Castan-Rechencq E. 1963. Constancy of extrapolated amount of nonsolvent water in cellulose, immersed in different salt solutions of varying concentration. Nature 200:358–9. Loveday SM, Winger RJ. 2007. Mathematical model of sugar uptake in fermenting yeasted dough. J Agric Food Chem 55:6325–9. Maier A, Volker B, Boles E, Fuhrmann GF. 2002. Characterisation of glucose transport in Saccharomyces cerevisiae with plasma membrane vesicles (countertransport) and intact cells (initial uptake) with single Hxt1, Hxt2, Hxt3, Hxt4, Hxt6, Hxt7 or Gal2 transporters. FEMS Yeast Res 2:539–50. Reifenberger E, Boles E, Ciriacy M. 1997. Kinetic characterization of individual hexose transporters of Saccharomyces cerevisiae and their relation to the triggering mechanisms of glucose repression. Eur J Biochem 245:324–33. Roman-Gutierrez AD, Guilbert S, Cuq B. 2002a. Distribution of water between wheat flour components: a dynamic water vapour adsorption study. J Cereal Sci 36:347–55. Roman-Gutierrez AD, Guilbert S, Cuq B. 2002b. Frozen and unfrozen water contents of wheat flours and their components. Cereal Chem 79:471–5. Ruan RR, Chen PL. 1998. Water in foods and biological materials: a nuclear magnetic resonance approach. Lancaster, PA: Technomic. Ruan RR, Wang XA, Chen PL, Fulcher RG, Pesheck P, Chakrabarti S. 1999. Study of water in dough using nuclear magnetic resonance. Cereal Chem 72:231–5.

“Solvent Water” Concept Simplifies Mathematical Modeling

57

Rückold S, Isengard H-D, Hanss J, Grobecker KH. 2003. The energy of interaction between water and surfaces of biological reference materials. Food Chem 82:51–9. Schiraldi A, Fessas D. 2003. The role of water in dough formation and bread quality. In: Cauvain SP, editor. Breadmaking: improving quality. Cambridge, UK: Woodhead. p 306–20. Serrano R, De la Fuente G. 1974. Regulatory properties of the constitutive hexose transport in Saccharomyces cerevisiae. Mol Cell Biochem 5:161–71. Toledo R, Steinberg MP, Nelson AI. 1968. Quantitative determination of bound water by NMR. J Food Sci 33:315–7. Tolstoguzov V. 1997. Thermodynamic aspects of dough formation and functionality. Food Hydrocolloids 11:181–93.

5 Microdomain Distribution in Food Matrices: Glass Transition Temperature, Water Mobility, and Reaction Kinetics Evidence in Model Dough Systems Y. Kou

Abstract The concept of distribution of microstructural domains within an amorphous food matrix provides a basis for explaining the dispersion of its physical properties. These spatially separated domains correspond to differing molecular arrangements within and about the proteinaceous and aqueous phases of flour-based and meat-based food matrices. The volume averaging of these localized properties determines the global values of such properties as glass transition temperature (Tg), water mobility as reflected in proton spin-spin relaxation time (T2), and reaction rate constants (k). The changes in Tg, T2, and k (for a reduction reaction) as moisture content and temperature were changed were measured using primarily electron spin resonance and timedomain nuclear magnetic resonance (NMR) techniques. These measured values were then correlated where appropriate, and distribution functions for the microdomains were obtained from the Tg, T2, and k values.

Introduction Unlike many synthetic polymers, most food materials are complex in chemical composition, heterogeneous in structure, and reactive. The stability of a food matrix depends strongly on its microstructure, local viscosity, and associated molecular mobility (Slade and Levine 1991; Fennema 1996). Moisture content, concentration of constituents, and temperature are key factors that determine the structure and distribution of microstructural domains in an amorphous food matrix, and control local viscosity and molecular mobility. Microstructural domains are defined as regions of differing local viscosity and molecular mobility, which are randomly distributed. The concept of a distribution of microstructural domains within an amorphous food matrix provides a basis for explaining the dispersion in its physical properties (Lillford and others 1980; Kou and others 2002; Kou 2006). These spatially separated domains correspond to differing molecular arrangements within and about the proteinaceous and aqueous phases of flour-based and meat-based food matrices. Such arrangements influence local viscosity and molecular mobility. Consequently, the volume averaging of these localized properties in turn determines the global values of such properties as glass transition temperature (Tg), water mobility as reflected in proton spin-spin relaxation time (T2), and reaction rate constant (k). 59

60

PART 1: Invited Speakers and Oral Presentations

To understand the nature of these microdomains, their distributions were either discerned or inferred from measurements made on dough matrices of different moisture contents and at different temperatures. The changes in Tg, T2, and k (for a reduction reaction) as moisture content and temperature are changed were measured using primarily electron spin resonance (ESR) and time-domain nuclear magnetic resonance (NMR) techniques. These measured values were then correlated where appropriate, and distribution functions for the microdomains were obtained from the Tg, T2, and k values. The main implication in the microdomains concept is that such localized differences in structure and properties need to be quantitatively considered to ensure that the physically, chemically, or microbiologically least-stable part of the matrix does not compromise quality or safety.

Materials and Methods Sample Preparation Wheat flour (12% protein, 11% moisture, 0.49% ash; Manildra Milling, Shawnee Mission, KS, USA) was used to make dough (flour + water) samples of varying water content (28%, 30%, 33%, and 35%, total weight basis). In the NMR experiment, samples were prepared prior to use, and ∼10 g of sample was weighed and placed into a glass test tube. In the kinetic study using ESR spectroscopy, erythorbic acid (Aldrich Chemical, Milwaukee, WI, USA), and TEMPO (tetramethylpiperidine nitroxide; Aldrich Chemical) were added to separate dough samples that were then mixed together by kneading equal volumes just prior to use. To ensure that the kneading uniformly distributed the reactants, aliquots of dough containing either TEMPO or erythorbic acid in judiciously chosen concentrations were kneaded together in 1 : 1, 1 : 2, 1 : 4, and 1 : 9 proportions, respectively; no differences in the results were found. NMR Experiments A 20-MHz PCT 20/20 NMR analyzer (Process Control Technology, Ft. Collins, CO, USA) was used in all NMR experiments. A 90° pulse sequence and a Carr-PurcellMeiboom-Gill (CPMG) pulse sequence were combined and used for acquisition of free-induction decay (FID) data for T2. Experimental parameters were set appropriately to maximize signal-to-noise ratio and to cover the entire relaxation range as completely as possible. ESR Experiments An ESR spectrometer equipped with a VT-2000 temperature controller (model EMX; Bruker Instruments, Billerica, MA, USA) was used to study the spectral pattern and chemical reduction of TEMPO. All samples were placed in a special plastic holder that fit into the resonator, the temperature of which was controlled by cold nitrogen gas. Temperature accuracy is ±0.1 K. The equilibrium time, even for the largest change in temperature, was less than 2 min. Microwave power was set to 0.63 mW. Scan

Microdomain Distribution in Food Matrices

61

range, scan rate, time constant, and field modulation amplitude were adjusted so that distortion of the spectra was avoided. Mathematical Model The dispersive model (Kou and others 2002; Kou 2006) describes a mathematical procedure for analyzing data obtained from experiments on the time-dependent behavior of dough samples subjected to various moisture and temperature conditions. The data consisted of either proton NMR or ESR signal arising from the reduction of TEMPO with erythorbic acid in the dough. The dough is viewed as an assemblage of many randomly distributed domains, each with specific physical properties. In each domain, a first-order or a pseudo–first-order reaction is assumed to take place, and the change in the NMR or ESR signal is taken as the sum of all changes associated with these reactions. The spatially distinct reactions are assumed to occur at different, random rates and to be affected by moisture and temperature, whose levels may be constant or variable over time. As described in the dispersive model, the mathematical procedure can be generalized for cases where more than one reaction or process of this sort is occurring simultaneously. Similar extensions can be made when two or more reactions are thought to take place. However, the parameter estimates become less reliable as the number of parameters increases, so there is a limit (essentially governed by the noise level of the data) beyond which applying this model is not useful.

Results and Discussion Glass Transition Temperature Representation of a system using an average value will be inadequate if the system is not sufficiently homogeneous in terms of physical structure and chemical composition (Ruan and others 1999). In fact, many food systems are heterogeneous, and such heterogeneity may be caused by ingredients’ incompatibility, poor mixing, or their redistribution during processing and storage. Therefore, there will be a nonuniform distribution of physical structure and moisture content within such a food matrix, and a variable spatial distribution of Tg values is expected. The NMR method for determination of Tg is based on the fact that polymers, when undergoing glass transition, experience a dramatic change in segmental mobility that can be measured by the spin-lattice and spin-spin reaction time constants from NMR experiments. The Tg can vary at the local or microscopic level and should be described in terms of an average and a dispersion. Ruan and others (1999) demonstrated this by measuring the Tg of a piece of bread (25 mm) while using magnetic resonance imaging with the spin-lattice (T1) mapping technique to achieve a resolution of 200 μm. The results indicate that microscopic distribution of Tg exists in bread matrix. The average Tg value is −10°C, with a standard deviation of 11°C. As in similar measurements on maltodextrin samples (Ruan and others 1999), the spread in Tg implies that microdomains exist, because most foods matrices are complex in chemical composition, heterogeneous in microstructure, and nonuniform in reaction reactivity.

62

PART 1: Invited Speakers and Oral Presentations

Water Mobility (Proton Spin-Spin Relaxation Time, T2) Since spin-spin relaxation is an entropy relaxation process, it is directly linked to the environment of water molecules and can be used to measure their rotational mobility. Higher T2 values of protons on water imply an increased water mobility and, in turn, less viscous locality. Accordingly, measurements of T2 in dough matrices show that mobility decreases as moisture content and temperature decrease. It is widely believed that water molecules in foods are in some way associated with different sites on constituents or are exchanged with such site-associated water (socalled bound water) and relax faster than the molecules in bulk water and thus have much lower T2 than does bulk water (Ablett and others 1991). Therefore, use of a single relaxation time constant to describe complex and heterogeneous food systems may be an oversimplification. In a heterogeneous system, spins exist in a large variety of environments, giving rise to a wide range of relaxation time constants. Thus, it is reasonable to assume that a continuous relaxation time constant would arise from a continuum of different environments and exchange rates (Ablett and others 1991; Ruan and others 1999; Kou and others 2002). In the present experiment, a 90° or CPMG pulse sequence set the proton magnetization signals at their maximum values, which began to decay immediately after the pulse was removed. The resulting decay curve is usually termed free-induction decay (FID), representing both the fast (rigid proton signals) and slow (mobile proton signals) relaxation processes in samples, and can be described by a single-exponential or multiple-exponential equation that can yield a single T2 or several discrete T2 values. Rigid proton signals (i.e., in a microsecond range) originated from solid molecules (i.e., gluten or starch in this case) or from water molecules tightly associated with solids, whereas mobile proton signals (i.e., in a millisecond range) originated from water molecules with relatively high mobility (Kou and others 2000). The experimental data were fitted to our mathematical model in order to obtain predicted values and the probability density function for T2. There was good agreement between the predicted and experimental values in both the microsecond and millisecond ranges. Figure 5.1 shows the distribution of values of the probability density function for T2, in the microsecond range (i.e., rigid proton signals), for 30% moisture dough samples at different temperatures (258–298 K). Model fitting resulted in one T2 peak with a different dispersion for each dough sample. As expected, because of the lower overall average viscosity, the average T2 increases with increasing temperature. Also, peak dispersion increases with decreasing temperature, which indicates that the degree of homogeneity in sample decreases with decreasing temperature. Figure 5.2 shows the distribution of values of the probability density function for T2, in the millisecond range (i.e., mobile proton signals), for 30% moisture dough samples at differing temperatures (258–298 K). Model fitting resulted in two T2 peaks with a different dispersion for each dough sample. The appearance of a second peak in the mobile proton signal suggested that new physical or chemical environments were formed within the system. Average T2 values ranged from about 2–9 ms, which suggested that the signals detected were from water molecules with relatively high mobility.

Microdomain Distribution in Food Matrices

63

3.5

2.5

Wt (1/T2)

Temperature (K)

298

3.0

288

2.0

278

1.5 1.0

268

0.5 0.0 0.00

258 0.02

0.04

0.06

1/T2

0.08

0.10

0.12

(μs-1)

Figure 5.1. The distribution of values of probability density function for T2, in the microsecond range (i.e., rigid proton signals), for 30% moisture dough at different temperatures (258–298 K).

Wt (1/T2 )

0.3

0.2

0.1

0.0 0.0

Temperature (K) 298

0.4

288

0.3

278 268 258

Wt (1/T2)

0.4

0.2 0.1 0.0 0.0

0.1

0.2

0.3

1/T2 (ms-1)

0.5

1.0

1/T2

1.5

(ms-1)

Figure 5.2. The distribution of values of probability density function for T2, in the millisecond range (i.e., mobile proton signals), for 30% moisture dough at different temperatures (258–298 K).

A comparison of these averaged T2 values with values obtained from an exponential model showed good agreement between the two approaches. However, the “continuum” approach used here is more consistent with the “continuum” nature of water mobility in food systems (Lillford and others 1980; Slade and Levine 1991). Furthermore, additional information may be obtained from a continuum model. For

64

PART 1: Invited Speakers and Oral Presentations

example, a spectrum with a larger number of peaks or with broader peaks would be expected for heterogeneous samples rather than for relatively homogeneous samples. In other words, the number of peaks and their breadth could be used as a measure of the homogeneity of a sample under analysis. Reaction Kinetics To investigate further the microscopic distribution of local viscosity and the inhomogeneous reaction kinetics in a dough matrix, the spin probe (TEMPO) was dissolved in the dough matrix and then monitored with ESR spectroscopy. In the kinetic study, an excess of erythorbic acid was used as the reductant. Dough samples were prepared right before an experiment by mixing two equal-weight portions, one containing TEMPO and the other the reductant. Both the shape of the ESR spectral patterns for TEMPO and the changes in signal intensity for TEMPO, upon reduction to its spin-inactive form through reaction with erythorbic acid, were monitored at several combinations of temperature and moisture content. Although the ESR spectral pattern for TEMPO was, in all cases, affected by temperature and moisture content, under appropriate conditions there was only a slight change in the spectral pattern because the ESR signal for TEMPO was reduced by reacting with the erythorbic acid. This finding suggested that the decreases in fast-line intensity could be used to monitor the reaction. Since localized differences in matrix viscosity can affect the rates of reaction between two diffusing molecules, the reduction in spin-active TEMPO by erythorbic acid was monitored in dough matrices at different moisture contents and temperatures. Such reduction converts TEMPO to a spin-inactive state, so the spectral intensity decreases as the reaction proceeds. To simplify the kinetics, excess erythorbic acid was used, so a pseudo–first-order process should have been observed. However, a semilogarithmic plot of normalized signal intensity, It/Io, against time was curved rather than straight. This result indicated that the reduction of TEMPO by erythorbic acid did not conform to homogeneous kinetics. The ever-decreasing rate of reaction with increasing time was characteristic of dispersive kinetics, which could be modeled by taking into account the distribution of microstructural domains. In other words, the data could be analyzed on the basis of a microscopic distribution of local domains, each with characteristic local viscosity and associated rate constant. Experimental data for the ESR signal intensity of TEMPO as a function of reaction time were obtained at combinations of six temperatures (273–298 K) and four moisture contents (28%–35%). As expected, the higher the temperature, the faster was the reduction of TEMPO because of the lower overall average viscosity. The same trend was observed for the effect of moisture content (i.e., higher moisture content resulted in faster reduction). The distribution of values for the reaction rate constant as a function of temperature for 35% moisture dough samples is presented in Figure 5.3. The average reaction rate constant, kave, increased with increasing temperature because of lower overall average viscosity, while the dispersion about the peak also increased with increasing temperature, which indicates that the degree of sample inhomogeneity increased with increas-

Microdomain Distribution in Food Matrices

65

Probability density function, Wt (k)

20 35% moisture

273K 15 278K 10

283K 288K

5

293K

0 0.0

0.2

298K

0.4

0.6

Reaction rate constant (k,

0.8

min-1)

Figure 5.3. Distribution of the reaction rate constant, k, as a function of temperature for 35% moisture dough.

Probability density function, Wt (k)

8 35%

288K

33%

6

30% 4

28%

2

0

0.0

0.1

0.2

0.3

0.4

0.5

Reaction rate constant (k, min-1)

Figure 5.4. Distribution of the reaction rate constant, k, as a function of moisture content in dough at 288 K.

ing temperature. Figure 5.4 shows the distribution of values of the probability density function, Wt(k), for reaction rate constant as a function of moisture content in dough samples at 288 K. The results indicate that the average reaction rate constant, kave, increased with increasing moisture content because of lower overall average viscosity. It was also observed that the dispersion about the peak decreased with increasing moisture content, which indicates that the degree of sample inhomogeneity decreased with increasing moisture content.

66

PART 1: Invited Speakers and Oral Presentations

When the resulting kave values obtained from these analyses were plotted semilogarithmically against reciprocal temperature, the observed reduction reaction apparently followed the Arrhenius relationship, and the activation energies corresponded to ∼12 kcal/mol. This finding suggests that there are barriers to both diffusion and electron transfer (following encounter and complex formation). Overall, the kinetic results were consistent with other rheological and relaxational phenomena, all of which reflect dispersion in the data associated with spatial inhomogeneities in microstructure.

Conclusions The concept of a distribution of microdomains within a complex and heterogeneous food matrix provides a basis for explaining the observed dispersion in its volumeaveraged properties. Glass transition temperature, molecular mobility, local viscosity, and reactivity are properties that can vary at the local or microscopic level. They are best described in terms of an average and dispersion. Since physical, chemical, and microbiological changes can occur in regions of low viscosity or high reactivity, the distribution of microstructural domains should be considered in assessing food safety, quality, and stability.

References Ablett S, Darke AH, Lillford PJ. 1991. The effect of mechanical deformation on the movement of water in foods. In: Levine H, Slade L, editors. Water relationships in foods. New York: Plenum. p 453–63. Fennema R. 1996. Food chemistry. 3rd ed. New York: Marcel Dekker. Kou Y. 2006. Microdomain distributions in food matrices: kinetic evidence from the reduction of spinactive TEMPO in model dough systems. Paper presented at the Amorph 2006 Conference, University of Cambridge. Kou Y, Dickinson LC, Chinachoti P. 2000. Mobility characterization of waxy corn starch using wide-line 1 H nuclear magnetic resonance. J Agric Food Chem 48:5489–95. Kou Y, Ross EW, Taub IA. 2002. Microstructural domains in foods: effect of constituents on the dynamics of water in dough, as studied by magnetic resonance spectroscopy. In: Levine H, editor. Amorphous food and pharmaceutical systems. London: Royal Society of Chemistry. p 48–58. Lillford PJ, Clark AH, Jones DV. 1980. Distribution of water in heterogeneous food and model systems. In: Comstock MJ, editor. Water in polymers. American Chemical Society Symposium Series 127. Washington, DC: American Chemical Society. p 177–95. Ruan R, Long Z, Chang K, Chen PL, Taub I. 1999. Glass transition temperature mapping using magnetic resonance imaging. Trans ASAE 42:1055–9. Slade L, Levine H. 1991. Beyond water activity: recent advances based on an alternative approach to the assessment of food quality and safety. Crit Rev Food Sci Nutr 30:115–360.

Session 2 Water Essence and the Stability of Food and Biological Systems

Invited Speakers

6 Effect of Combined Physical Stresses on Cells: The Role of Water J.-M. Perrier-Cornet, M. Moussa, H. Simonin, L. Beney, and P. Gervais

Abstract The role of water in microorganism viability was envisaged through the application of combined physical perturbations. The combination of different physical parameters could enable one to balance the property variations (especially water related) resulting from the increase of one parameter alone. The first example shows that the combination of osmotic level and temperature can enable yeast cell survival to be optimized when following membrane fluidity variations. This analysis has enabled a better comprehension of cell inactivation during rehydration and dehydration. The second example deals with the combination effect of high hydrostatic pressure, low temperature, and medium water activity (aw) on Escherichia coli resistance. The synergetic effect of high pressure and low temperature was observed only at a pressure level lower than 300 MPa and with high water content. Otherwise, low temperature, as well as low aw, protect the microorganisms from inactivation even at an extreme pressure level (P > 600 MPa).

Introduction The change in physical environment (e.g., hydrostatic pressure and temperature) or physicochemical environment (e.g., water activity [aw] and pH) induces significant stress on eukaryotic and prokaryotic cells. Depending on how the level of perturbation and the kinetics of conditions change, this stress could lead to the inactivation of relevant cells. When high-level and rapid perturbations are used in a poor medium, cells have only few active systems for adaptation, and their response is, in this case, essentially passive. Under these conditions, cell resistance can be attributed to the cell’s constitution (e.g., a robust cell wall, adaptability of the cell membrane, its cytoskeleton) and the repair systems that the cell can use after its return to more favorable conditions. A combination of physical treatments can modulate the effect of each stress and provide interesting information about the mechanisms implicated. The biological basis of these interactions is not yet clearly understood. Numerous experiments have shown the role of cell osmotic balance and of cell membrane passive and active permeability. Membrane structure and fluidity seem to play an important role during dehydration, rehydration, and generally in all stress conditions. 71

72

PART 1: Invited Speakers and Oral Presentations

The effects of these intense perturbations on cell survival are highly important when food processes like drying, freezing, sterilization, and pasteurization are considered. In these processes, very drastic perturbations are applied to food products and their microorganisms. These perturbations were intended to inactivate pathogens (for food preservation) or preserve food and/or cells (by drying or freezing). These industrial processes generally combine different drastic physical modifications including temperature, osmotic pressure, and/or hydrostatic pressure. Combinations of physical perturbations have been experimented with in a model medium to understand the mechanisms leading to microbial inactivation. A better perception of such mechanisms would enable not only these food processes to be optimized but also would have other applications, like the conservation of human cells at ambient temperature or in a frozen state. We approach this research theme through two examples of physical perturbation combinations. The first example deals with the effect of hyperosmotic stress and the possibility of combining it with temperature. The second is centered on the effect of high hydrostatic pressure on cells and the effect of a combination of low temperature and/or low aw.

Example 1: Effects of Combined Hyperosmotic and Temperature Perturbation Sequence of Hyperosmotic Perturbation of Yeast Cells During the first part of dehydration, sudden exposure to a hyperosmotic stress causes rapid equilibration of the osmotic pressures of the cytoplasm and the external medium. During the transitional step of the passive osmotic response, water flows out of the cells, leading to cell shrinkage, and permeant solutes, such as glycerol, penetrate into the cells. This exchange is rapid (Berner and Gervais 1994) and ends in a stationary step when osmotic pressures are equilibrated. As shown in Figure 6.1, cell volume decreased exponentially between an aw of 0.99 and 0.8, before reaching a constant volume corresponding to 40% of the initial volume, generally called nonosmotic volume. Cell volume was evaluated from light-microscopic images, and thus the total envelope of the yeast was taken into account; that is, cell wall and membrane. In contrast to plant cells and bacteria, in which the plasma membrane shrinks away from the cell wall, in yeast the entire cell volume shrinks when cells are placed in hypertonic solutions (Morris and others 1986). Considering the poor compressibility of biological membranes, this strong cell shrinkage must be associated with wrinkling of the membrane. According to Adya and others (2006), the cell shape became irregular. In decreasing aw to below 0.55, the cells became more permeable as indicated by the ratio of propidium iodide (PI)-stained cells in Figure 6.2. Therefore, this osmotic pressure interval appears to be critical for membrane permeability during dehydration. Phase transitions of phospholipids have been proposed as the main cause of the increase in membrane permeability in both phospholipid vesicles (Yamazaki and others 1989) and yeasts (Laroche and others 2001) under osmotic stress. Water loss from phospholipid head groups may lead to phase transition in some lipids, resulting

Figure 6.1. Variations in average cell volume (open circles) and cell viability (open squares) of Saccharomyces cerevisiae after an osmotic shock from a culture medium (water activity = 0.99) to a binary medium (water + glycerol) at different water activity levels. Volume data were obtained by the analysis of confocal images and viability was determined by the colony-forming unit method.

% of stained cells (or viability)

100

80

60

40

20

0 0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Water activity

Figure 6.2. Viability and membrane permeability of Saccharomyces cerevisiae versus water activity by use of two probes: Lucifer yellow (LY) and propidium iodide (PI). Shown are cells with increased permeability (solid circles), which are doubled marked; cells with endocytosis (open circles), which are marked with LY; and intact cells after rehydration (open squares).

73

74

PART 1: Invited Speakers and Oral Presentations

in a lateral phase separation (Lehtonen and Kinnunen 1995) that allows the leakage of intracellular content (Crowe and others 1992). A phase transition in yeast membrane lipids from aw 0.64 to 0.38 in glycerol solution at an average temperature of 22°C (Laroche and others 2001) could explain the permeabilization observed. During hyperosmotic treatments, the number of Lucifer yellow (LY) stained cells also increased with increasing osmotic pressures (Figure 6.2). LY is a membraneimpermeant anionic dye. This polar tracer is usually loaded by microinjection, pinocytosis, or scrape loading. It has been used to characterize endocytosis in plant cells (Roszak and Rambour 1997) and yeasts (Wiederkehr and others 2001) in which the presence of a cell wall prevents the access of high molecular weight molecules to the plasma membrane. In PI/LY double-stained cells, LY probably penetrated the cells because the plasma membranes were made permeable. In cells stained only with LY, plasma membrane endocytic vesiculation under hyperosmotic conditions seems possible. Endovesicles have already observed by Mille and others (2002) with E. coli. Slaninova and others (2000) reported deep plasma membrane invaginations filled from the periplasmic side with an amorphous cell wall material when Saccharomyces cerevisiae cells were transferred to hyperosmotic growth medium. Such invaginations, when associated with lipid phase separation induced by dehydration, could lead to the formation of endocytic vesicles. In fact, Liu and others (2006) showed that the scission of membrane invaginations could be promoted by lipid phase separation to form endovesicles. The percentage of increased permeability (PI stained) cells was constant below an aw of 0.86 during rehydration and increased strongly at the upper levels of rehydration, showing that most of the cells that had reached a critical aw (0.35) could not recover their permeability. Therefore, the aw interval between 0.86 and 0.99 appears to be critical for membrane permeability during rehydration. The existence of this critical step could be related to membrane events that occur during dehydration. Indeed, cells labeled with LY may have suffered from a reduction in surface area associated with the formation of endovesicles, as has already been proposed by Shalaev and Steponkus (1999) and is supported by our observations. Therefore, exposing these cells to rehydration levels that impose significant increases in volume (cf. Figure 6.1) may cause their lysis during volume expansion. Okada and Rechsteiner (1982) reported that endovesicles that form under hyperosmotic conditions swell and burst upon rehydration of the cytosol. In fact, we show that, for an aw lower than 0.6, the removal of a portion of water from the cells may lead to changes in the cell permeability resulting from the phase separation of phospholipids. In fact, lipid phase transition affects the resistance of membranes to shear forces (Sparr and Wennerstrom 2001), and volume contraction may thus be critical when this occurs. Plasma membrane changes are strongly implicated in the mechanism leading to cell death during osmotic dehydration and rehydration. In particular, permeabilization resulting from lipid phase transitions and severe volume contractions could explain the observed sequence of events. Moreover, the changes that occur during the dehydration step and the rehydration step are interdependent.

Effect of Combined Physical Stresses on Cells: The Role of Water

75

Figure 6.3. Isoviability (N/N0) diagram of Saccharomyces cerevisiae versus temperature (4–40°C) and water activity. The viability data were obtained after 1 h at the physical conditions indicated and rehydration to optimum conditions.

Effect of Combined Osmotic and Thermal Stress To show the link between yeast survival following combined osmotic and thermal treatments and membrane-fluidity variations induced by the treatments, a cell-viability diagram (aw and temperature) and a membrane-fluidity diagram (aw and temperature) are presented in Figures 6.3 and 6.4. According to these figures, we see that, without a phase change during dehydration, we can expect to have a higher survival percentage. Thus, the fluidity diagram appears to be a potential tool for controlling membrane fluidity during cell dehydration and rehydration by simultaneously and independently managing aw and temperature over time. The fact that cell death provoked by osmotic shocks depends on temperature is well established. Dried-yeast recovery is optimal if rehydration is performed at 38°– 40° or 50°C (Poirier and others 1999). Furthermore, the temperature at which dehydration shock occurs in liquid medium has been shown to affect cell viability greatly (Laroche and Gervais 2003). The results presented in Figure 6.4 show that the yeast

76

PART 1: Invited Speakers and Oral Presentations

Figure 6.4. Isoanisotropy of Saccharomyces cerevisiae membrane versus temperature and water activity. Anisotropy was measured by using diphenylhexatriene fluorescence polarization.

has an enhanced resistance to osmotic shock at temperatures lower than 10°C and higher than 22°C. However, resistance to osmotic shock based on temperature is strain dependent, and each strain may have a specific behavior. A strain-dependent response to glycerol osmotic stresses has also been reported by Blomberg (1997). Laroche and Gervais (2003) proposed that mortality following rapid dehydration or rehydration was related to water flow through an unstable membrane. Guyot and others (2006) confirmed this assumption and hypothesized that change in the fluidity of the plasma membrane was the critical event leading to cell death and that water flow was not necessarily involved in the cell-death mechanism. Our present work confirms this latest assumption. However, here, if water outflow is insufficient to provoke cell death, fluidity variation in the case of thermal stress alone in the range of 4°–40°C—that is, without osmotic stress—did not provoke cell death. Thus, change in membrane fluidity is the critical event, but must be accompanied by osmotic stress and certainly subsequent volume contraction. In the case of hyperosmotic shock, not

Effect of Combined Physical Stresses on Cells: The Role of Water

77

only are the membranes in phase transition, but also the cells are contracted. It is well known that cells shrink in response to osmotic stress. Adya and others (2006) have associated the surface topology of S. cerevisiae (irregular shape, surface roughness) and volume loss to the cell’s resistance to both thermal and osmotic stress. Slaninova and others (2000) reported deep invaginations in the plasma membrane and bulges in osmotically stressed yeasts. Such conditions of shrinkage associated with lipid phase separation occurring before and/or during the dehydration-rehydration steps (Δr2 provoked by the osmotic stress) probably leads to plasma membrane permeabilization and leakage of cellular components. Our hypotheses regarding the mechanism leading to cell death during dehydration and rehydration are developed in two recent investigations (Simonin and others 2007a, 2007b). Conclusions Based on the First Example: Combination of Osmosis and Temperature Membrane state and survival of osmotic stresses are linked. Particularly, changes in membrane fluidity before and/or during osmotic treatment influence yeast survival, and lipid phase transition in membranes is disadvantageous for cells submitted to osmotic shock. The use of the membrane-fluidity diagram enabled control of the membrane fluidity of cells during dehydration and rehydration. To understand the plasma membrane changes occurring during dehydration and rehydration, complementary techniques of membrane study should be used. Actually, it must be taken into account that membranes are complex organelles composed of a variety of lipids structured in membrane domains, and a global coefficient related to membrane fluidity is insufficient to appreciate all the changes occurring in them. In fact, complex lipid phase behavior is known to occur when water content is low (Milhaud 2004). Particularly, nonlamellar phases are suspected to arise at low water concentrations, as observed in model biomembranes (Shalaev and Steponkus 2001). We show that such a diagram should be a useful tool for improving yeast survival in dehydration-rehydration processes. Such processes are involved not only when drying food or fermenting, but also in freezing. In fact, a freezing process at a moderate temperature (Δt < 1000°C/min) consists essentially for microorganisms in a hyperosmotic perturbation at low temperature (near 0°C, during water crystallization). Cell inactivation in this process might mainly be attributed to the combination of osmotic and temperature perturbation (Dumont and others 2006).

Example 2: Effects of Combined High Hydrostatic Pressure, Low Temperature, and Hyperosmotic Perturbations Combination of High Hydrostatic Pressure and Low Temperature on Escherichia coli Survival Numerous studies have demonstrated that the antimicrobial effects of high pressure depend on temperature (Sonoike and others 1992). Moreover, the efficiency of highpressure treatments is controlled by other process parameters such as the pressure applied and the kinetics of pressurization (Palou and others 1998), as well as by the

78

PART 1: Invited Speakers and Oral Presentations

physicochemical properties of the medium being treated, such as pH (Alpas and others 2000) and aw (Van Opstal and others 2003). These parameters must be controlled precisely to ensure efficient treatment. With appropriate combinations of these parameters, a synergistic effect might be achieved, reducing the pressures and treatment times required. The combined effects of high pressure and low or subzero temperatures on microbial inactivation have been studied. A synergistic effect between these parameters has generally been reported in the inactivation of vegetative microorganisms (Hashizume and others 1995; Perrier-Cornet and others 2005). In some cases, the initial microbial populations were completely inactivated with a combined treatment of high pressure and low or subzero temperature, whereas only a slight microbial inactivation was achieved under the same pressure conditions at room temperature (Perrier-Cornet and others 2005). The magnitude of this synergistic effect depends strongly on the type of microorganism (Takahashi 1992). The interaction of high pressure and subzero temperature in microbial inactivation is complex, and possible phase-transition phenomena must be taken into account. Some authors have demonstrated that freezing under hyperbaric conditions is effective in reducing microbial contamination (Luscher and others 2004). In addition to the antimicrobial effects of combining high-pressure and subzero-temperature treatments, the combination offers various processing advantages, such as rapid freezing and thawing and cold storage of foods under liquid conditions (Cheftel and others 2002). Figure 6.5 shows the effect of a 10-min treatment at different pressure and temperature levels on the logarithmic inactivation of E. coli K12TG1. At −20°C, in the supercooled region, the pressure sensitivity was greater than at 25°C for pressure lower than 350 MPa. This synergism between high pressure and subzero temperature made it possible to reduce the pressure and/or improve pressure-mediated inactivation. Irrespective of the inactivation rate, our findings corroborate the observations reported by Takahashi (1992), who examined the inactivation of E. coli after pressure treatment (200 MPa, 20 min) at −20°C and at room temperature. More recently, we reported that, at a fixed pressure of 150 MPa, an initial population of S. cerevisiae was completely inactivated at −20°C (>8 log cycles under liquid conditions), whereas it was only slightly inactivated at 25°C (<0.5 log cycles) (Perrier-Cornet and others 2005). The viability of E. coli K12TG1 was affected less by the synergism between high pressure and subzero temperature than were the viabilities of Lactobacillus plantarum and S. cerevisiae. Above 350 MPa, the synergistic effect was completely neutralized by an antagonistic effect. Accordingly, E. coli K12TG1 cells were more resistant to higher than to lower pressure levels. A similar observation was described by Pagán and Mackey (2000) for E. coli H1071 cells in the stationary phase of growth after pressure treatment at room temperature. The unusual pattern of survival of E. coli K12TG1 cells after combined high-pressure and subzero-temperature treatments was observed consistently in many experiments. These observations reflected a baroprotective effect at subzero temperatures and very high pressure. This effect has never been observed before and could be combined with the atypical comportment of water molecules under

Effect of Combined Physical Stresses on Cells: The Role of Water

79

Figure 6.5. Isoinactivation (log N0/N) of Escherichia coli versus pressure and temperature in a binary medium (water + glycerol) at a water activity of 0.99. The sample was maintained for 10 min at the conditions indicated before being grown under optimum conditions.

pressure. At pressure lower than 300 MPa, because of hydrogen bonds, water exhibits atypical properties (e.g., maximum density, phase change, viscosity) especially at low temperature. At higher pressure (P > 400 MPa), water behavior becomes more regular. To confirm this relation with water thermodynamic comportment at low temperature, the aw of the medium has been modulated. Effect of Low Temperature and Hyperosmotic Perturbation on Escherichia coli Baroresistance As shown in Figure 6.6, the pressure sensitivity of E. coli K12TG1 was highly dependent on the aw of the system. When the bacterium was suspended in a water-glycerol solution with an aw of 0.85, it was more pressure resistant than at an aw of 0.99. This finding underscores the baroprotective effect of solutes, previously described for E. coli (Satomi and others 1995; Van Opstal and others 2003), Rhodotorula rubra (Oxen and Knorr 1993), and Zygosaccharomyces bailii (Palou and others 1997). The combination of subzero temperature and high pressure at an aw of 0.85 caused a cumulative

80

PART 1: Invited Speakers and Oral Presentations

Figure 6.6. Isoinactivation (log N0/N) of Escherichia coli versus pressure and temperature in a binary medium (water + glycerol) at a water activity of 0.85. The sample was maintained for 10 min at the conditions indicated before being grown under optimum conditions.

protective effect of solute and subzero temperature against pressure-induced inactivation. Only the protective effect of low temperature appears in Figure 6.6 in a medium with an aw of 0.85. The protection by the solute led to the higher pressure level necessary to inactivate E. coli, and then inactivation occurs only in the pressure-temperature domain where the antagonistic effect was dominant. When pressurized in distilled water (aw = ∼1), E. coli K12TG1 showed a much higher pressure sensitivity than at lower aw, especially at −20°C (Figure 6.7). In this case, only the synergistic effect of low temperature is observed, probably because the entire microbial population is inactivated at a pressure lower than 400 MPa. Parallel Changes with Pressure and Temperature of Protein Behavior, Microbial Inactivation, and Water Structure Several studies have highlighted the crucial role of water in the pressure-induced denaturation of biological systems. Oliveira and others (1994) reported that protein denaturation decreased linearly with a decrease in water concentration. Similarly,

Effect of Combined Physical Stresses on Cells: The Role of Water

81

Figure 6.7. Isoinactivation (log N0/N) of Escherichia coli versus pressure and temperature in distilled water. The sample was maintained for 10 min at the conditions indicated before being grown under optimum conditions.

Kinsho and others (2002) observed that the removal of water by the addition of polyols or small cationic ions had an efficient protective effect against enzyme inactivation at high pressures and subzero temperatures. The latter authors also reported that coldinactivation mechanisms were pressure dependent and differed at pressures below 200 MPa from those at pressures above 200 MPa. Moreover, a maximum stability temperature was evidenced for different proteins, and a bell-shaped dependence of protein stability on temperature was observed (Smeller 2002). A parallel has been proposed between the structure of water and the thermal denaturation of proteins (Klotz 1999). In fact, among other similarities, the graph of liquid water density follows a bell-shaped curve at atmospheric pressure, with a maximum at 4°C. Some authors have emphasized the effect of pressure on water density as a key for understanding cold denaturation of proteins at high pressure (Marques and others 2003). The properties of water under pressure vary and are largely a function of the pressure range (Cavaille and others 1996). Indeed, the effect of increasing pressure on the behavior of cold water is to push the temperature of maximum density systematically to lower and lower temperatures. The so-called atypical properties are observed for

82

PART 1: Invited Speakers and Oral Presentations

pressures below 200 MPa. However, above 400 MPa pressure, water loses its particular characteristics and behaves like a classic hydrogen-bonded liquid. The addition of solutes causes the formation of hydration shells, leading to a new organization of water molecules. This phenomenon is strongly enhanced when the pressure is increased and, accordingly, it cancels out the particular properties of pure water in the pressure range of 0.1–200 MPa (Kanno and Angell 1979). The variation in water properties with pressure, temperature, and the presence of solutes reflects changes in the arrangement of water molecules. From a biological perspective, this could explain the baroprotective effects of solutes on proteins and microorganisms under denaturing conditions. The mechanisms of pressure-induced microbial inactivation may involve denaturation of some critical life processes such as enzyme reactions, as suggested by some authors (Hashizume and others 1995; Perrier-Cornet and others 2005). Also, a parallel between water properties and microbial inactivation can be identified. For a known set of hydration conditions, a synergistic effect was observed at pressures up to a critical level (250 MPa for an aw of 0.992), whereas antagonism occurred at pressures higher than this critical level. The consequence of increasing the hydration rate at a fixed pressure was to enhance the synergism and increase the pressure threshold that marked the crossover between synergism and antagonism. Below this threshold, pressure and temperature affect microbial viability in a similar manner and, in the same way, water behaves as a singular liquid. Above this threshold, pressure and temperature have roughly opposite effects on microbial viability and, at the same time, water behaves as a classic hydrogen-bonded liquid. Conclusions Based on the Second Example: Combination of High Pressure, Temperature, and Osmosis This work shows that combined high-pressure and subzero-temperature treatment is a promising way to optimize high-hydrostatic-pressure processes, since the combination made it possible to reduce the pressure magnitude and/or improve the pressuremediated inactivation. Nevertheless, the interaction between high pressure and subzero temperature appears to be complex. Indeed, it was pointed out that, depending on pressure level and aw of the medium being treated, subzero temperature counteracted the inactivation caused by high pressure. This unexpected phenomenon leads to the necessity to take into account the process parameters to ensure efficient treatment. Considering the structure of water in relation to the stability of proteins and to microbial inactivation led to the suspicion of a crucial role of water in this phenomenon. Further work should be undertaken with a view to better elucidate this phenomenon.

Conclusions These two examples show the critical role of thermodynamic properties of water in the survival of microorganisms. Maintenance of living structures by water is effective only if water retains its specific properties. Modifying thermodynamic properties

Effect of Combined Physical Stresses on Cells: The Role of Water

83

by pressure, temperature, or osmotic solutes changes cell equilibrium. This change in thermodynamic conditions is also accompanied by mechanical constraints (water efflux with hyperosmotic stress or hydrostatic compression with pressure) that contribute to the destabilization of the cell and especially the cell membrane. In minimizing such perturbations and using correct thermodynamic properties, the viability of live cells can be maintained even under drastic conditions. The osmotic dehydration at a controlled temperature can enable viable dehydrated cells to be obtained even for sensitive organisms. High-pressure processing at low temperature can also enable cell viability to be maintained at very high pressure. This process would be useful in maintaining cells at low temperature and high pressure in liquid water. Thus, understanding the role of water in the maintenance of cell viability would enable specific promising processes combining different thermodynamic parameters to be developed to preserve or inactivate organisms and microorganisms

References Adya AK, Canetta E, Walker GM. 2006. Atomic force microscopic study of the influence of physical stresses on Saccharomyces cerevisiae and Schizosaccharomyces pombe. FEMS Yeast Res 6:120–8. Alpas H, Kalchayanand N, Bozoglu F, Ray B. 2000. Interactions of high hydrostatic pressure, pressurization temperature and pH on death and injury of pressure-resistant and pressure-sensitive strains of foodborne pathogens. Int J Food Microbiol 60:33–42. Berner JL, Gervais P. 1994. A new visualization chamber to study the transient volumetric response of yeast cells submitted to osmotic shifts. Biotechnol Bioeng 43:165–70. Blomberg A. 1997. The osmotic hypersensitivity of the yeast Saccharomyces cerevisiae is strain and growth media dependent: quantitative aspects of the phenomenon. Yeast 13:529–39. Cavaille D, Combes D, Swick A. 1996. Effect of high hydrostatic pressure and additives on the dynamics of water: a Raman spectroscopy study. J Raman Spectrosc 27:853–7. Cheftel J-C, Thiebaud M, Dumay E. 2002. High pressure–low temperature processing of foods: a review. In: Winter R, editor. Advances in high pressure bioscience and biotechnology II. Heidelberg, Germany: Springer. p 327–40. Crowe JH, Hoekstra FA, Crowe LM. 1992. Anhydrobiosis. Annu Rev Physiol 54:579–99. Dumont F, Marechal PA, Gervais P. 2006. Involvement of two specific causes of cell mortality in freezethaw cycles with freezing to −196°C. Appl Environ Microbiol 72:1330–5. Guyot S, Ferret E, Gervais P. 2006. Yeast survival during thermal and osmotic shocks is related to membrane phase change. J Agric Food Chem 54:8450–5. Hashizume C, Kimura K, Hayashi R. 1995. Kinetic analysis of yeast inactivation by high pressure treatment at low temperatures. Biosci Biotechnol Biochem 59:1455–8. Kanno H, Angell CA. 1979. Water: anomalous compressibilities to 1.9 kbar and correlation with supercooling limits. J Chem Phys 70:4008–16. Kinsho T, Ueno H, Hayashi R, Hashizume C, Kimura K. 2002. Sub-zero temperature inactivation of carboxypeptidase Y under high hydrostatic pressure. Eur J Biochem 269:4666–74. Klotz IM. 1999. Parallel change with temperature of water structure and protein behavior. J Phys Chem [B] 103:5910–6. Laroche C, Beney L, Marechal PA, Gervais P. 2001. The effect of osmotic pressure on the membrane fluidity of Saccharomyces cerevisiae at different physiological temperatures. Appl Microbiol Biotechnol 56:249–54. Laroche C, Gervais P. 2003. Achievement of rapid osmotic dehydration at specific temperatures could maintain high Saccharomyces cerevisiae viability. Appl Microbiol Biotechnol 60:743–7. Lehtonen JY, Kinnunen PK. 1995. Poly(ethylene glycol)-induced and temperature-dependent phase separation in fluid binary phospholipid membranes. Biophys J 68:525–35.

84

PART 1: Invited Speakers and Oral Presentations

Liu J, Kaksonen M, Drubin DG, Oster G. 2006. Endocytic vesicle scission by lipid phase boundary forces. Proc Natl Acad Sci USA 103:10277–82. Luscher C, Balasa A, Frohling A, Ananta E, Knorr D. 2004. Effect of high-pressure-induced ice I-to-ice III phase transitions on inactivation of Listeria innocua in frozen suspension. Appl Environ Microbiol 70:4021–9. Marques MI, Borreguero JM, Stanley HE, Dokholyan NV. 2003. Possible mechanism for cold denaturation of proteins at high pressure. Phys Rev Lett 91:138103-1-4. Milhaud J. 2004. New insights into water-phospholipid model membrane interactions. Biochim Biophys Acta 1663:19–51. Mille Y, Beney L, Gervais P. 2002. Viability of Escherichia coli after combined osmotic and thermal treatment: a plasma membrane implication. Biochim Biophys Acta 1567:41–8. Morris GJ, Winters L, Coulson GE, Clarke KJ. 1986. Effect of osmotic stress on the ultrastructure and viability of the yeast Saccharomyces cerevisiae. J Gen Microbiol 129:2023–34. Okada CY, Rechsteiner M. 1982. Introduction of macromolecules into cultured mammalian cells by osmotic lysis of pinocytic vesicles. Cell 29:33–41. Oliveira AC, Gaspar LP, Da Poian AT, Silva JL. 1994. Arc repressor will not denature under pressure in the absence of water. J Mol Biol 240:184–7. Oxen P, Knorr D. 1993. Baroprotective effects of high solute concentrations against inactivation of Rhodotorula rubra. Lebensm Wiss Technol 26:220–3. Pagán R, Mackey B. 2000. Relationship between membrane damage and cell death in pressure-treated Escherichia coli cells: differences between exponential- and stationary-phase cells and variation among strains. Appl Environ Microbiol 66:2829–34. Palou E, López-Malo A, Barbosa-Cánovas GV, Welti-Chanes J, Davidson PM, Swanson BG. 1998. High hydrostatic pressure come-up time and yeast viability. J Food Prot 61:1657–60. Palou E, López-Malo A, Barbosa-Cánovas GV, Welti-Chanes J, Swanson BG. 1997. Effect of water activity on high hydrostatic pressure inhibition of Zygosaccharomyces bailii. Lett Appl Microbiol 24: 417–20. Perrier-Cornet JM, Tapin S, Gaeta S, Gervais P. 2005. High-pressure inactivation of Saccharomyces cerevisiae and Lactobacillus plantarum at subzero temperatures. J Biotechnol 115:405–12. Poirier I, Marechal PA, Richard S, Gervais P. 1999. Saccharomyces cerevisiae viability is strongly dependant on rehydration kinetics and the temperature of dried cells. J Appl Microbiol 86:87–92. Roszak R, Rambour S. 1997. Uptake of Lucifer Yellow by plant cells in the presence of endocytic inhibitors. Protoplasma 199:198–207. Satomi M, Yamagushi T, Okuzumi M, Fujii T. 1995. Effect of conditions on the barotolerance of Escherichia coli. J Food Hyg Soc Jpn 36:29–34. Shalaev EY, Steponkus PL. 1999. Phase diagram of 1,2-dioleoylphosphatidylethanolamine (DOPE):water system at subzero temperatures and at low water contents. Biochim Biophys Acta 1419:229–47. Shalaev EY, Steponkus PL. 2001. Phase behavior and glass transition of 1,2-dioleoylphosphatidylethanolamine (DOPE) dehydrated in the presence of sucrose. Biochim Biophys Acta 1514:100–16. Simonin H, Beney L, Gervais P. 2007a. Cell death induced by mild physical perturbations could be related to transient plasma membrane modifications. J Membr Biol 216:37–47. Simonin H, Beney L, Gervais P. 2007b. Sequence of occurring damages in yeast plasma membrane during dehydration and rehydration: mechanisms of cell death. Biochim Biophys Acta 1768:1600–10. Slaninova I, Sestak S, Svoboda A, Farkas V. 2000. Cell wall and cytoskeleton reorganization as the response to hyperosmotic shock in Saccharomyces cerevisiae. Arch Microbiol 173:245–52. Smeller L. 2002. Pressure-temperature phase diagrams of biomolecules. Biochim Biophys Acta 1595:11–29. Sonoike K, Setoyama T, Kuma Y, Kobayashi S. 1992. Effects of pressure and temperature on the death rates of Lactobacillus casei and Escherichia coli. In: Balny C, Hayashi R, Heremans K, Masson P, editors. High pressure and biotechnology. London: John Libbey Eurotext. p 297–301. Sparr E, Wennerstrom H. 2001. Responding phospholipid membranes-interplay between hydration and permeability. Biophys J 81:1014–28.

Effect of Combined Physical Stresses on Cells: The Role of Water

85

Takahashi K. 1992. Sterilization of microorganisms by hydrostatic pressure at low temperature. In: Balny C, Hayashi R, Heremans K, Masson P, editors. High pressure and biotechnology. London: John Libbey Eurotext. p 303–7. Van Opstal I, Vanmuysen SCM, Michiels CW. 2003. High sucrose concentration protects E. coli against high pressure inactivation but not against high pressure sensitization to the lactoperoxidase system. Int J Food Microbiol 88:1–9. Wiederkehr A, Meier KD, Riezman H. 2001. Identification and characterization of Saccharomyces cerevisiae mutants defective in fluid-phase endocytosis. Yeast 18:759–73. Yamazaki M, Ohnishi S, Ito T. 1989. Osmoelastic coupling in biological structures: decrease in membrane fluidity and osmophobic association of phospholipid vesicles in response to osmotic stress. Biochemistry 28:3710–5.

7 Soft Condensed Matter: A Perspective on the Physics of Food States and Stability T. P. Labuza, T. J. Labuza, K. M. Labuza, and P. S. Labuza

Abstract Several foods are complex examples of soft condensed matter (SCM) that can undergo phase transition from an amorphous glass to an amorphous rubber state and vice versa. The concept of SCM in terms of the state diagram is presented for an understanding of physical state changes in foods. Transformation of a food matrix from its glassy to rubbery state enables gravity-induced collapse, flow, and possible recrystallization of a food component producing drastic changes in food textural property and affecting its quality as perceived by consumers. Discussion is focused on cotton candy, hard-ball candy, and crisp snacks in terms of the effect of water content and temperature on their state behavior. The sensory crispness paradigm of hard cookies and crisp snacks can be predicted from changes in SCM states from brittle material into truly soft SCM.

Introduction In 1991, Roos and Karel (1991a, 1991b, 1991c) introduced the concept of applying the glass transition temperature (Tg) curve for sugar on top of the phase diagram of the sugar-water mixture. Although an extremely useful tool that can combine water activity (aw) and Tg in one graph, it has had limited applicability, likely because of misunderstanding of the basic physics, thermodynamics, and kinetics of food systems by many researchers. In 2005, a group of Nestle/University of Fribourg (Switzerland) physicists stirred up the pot of understanding of food physics by publishing a paper on “soft condensed matter” (Mezzenga and others 2005, p 729), in which they noted, Foods make up some of the most complex examples of soft condensed matter (SCM) with which we interact daily. Their complexity arises from several factors: the intricacy of components, the different aggregation states in which foods are encountered, and the multitude of relevant characteristic time and length scales. Because foodstuffs are governed by the rules of SCM physics but with all the complications related to real systems, the experimental and theoretical approaches of SCM physics have deepened our comprehension of their nature and behaviour, but many questions remain. … With their complexity, heterogeneity and multitude of states, foods provide SCM physics with a challenge of remarkable importance. 87

88

PART 1: Invited Speakers and Oral Presentations

What they said is very relevant to what has occurred in the field of watermanagement science and the physical state of foods; that is, material science concepts applied to SCM. We will present this concept of SCM in terms of the state diagram of Karel and Roos for understanding physical state changes in foods, which they did not cover in much depth because the study of glass transition was in its infancy. The work in this chapter was initiated by my children because I wanted to clarify a base understanding so that it could be used to educate our undergraduates and graduate students in food science. The focus will be on cotton candy, hard-ball candy, and crisp snacks in terms of the effect of water content and temperature on the change in physical state; that is, an amorphous glass either to/from an amorphous rubber state or to a crystalline state. For example, crystalline sucrose when melted and spun forms a dry amorphous glass (cotton candy) that upon humidification above the glass transition line enters the rubbery state, another form of SCM. The higher the humidity, the faster the amorphous glass transforms into an amorphous rubber, which allows for gravityinduced collapse, flow, and recrystallization. With a hard-ball candy (mostly sucrose with maltodextrins), which is a more dense glass, both temperature and humidity induce surface stickiness above the Tg line. At high enough temperature (55°–65°C) at all humidities, the hard ball begins to flow (collapse) under gravity and, if confined in a can, re-forms into a hockey-puck shape when cooled. Hard cookies and crisp snacks also show changes in SCM states, starting out as brittle material that transforms into truly soft SCM in the rubbery state for which we can project a sensory crispness paradigm. Why this occurs is a lesson in the physics of foods. With hard cookies, the transition from a glass to a rubber causes a loss of crispness, which can be tied to the brittle-ductile line. Understanding this will show why food science is as fascinating as E = mc2 and Feynman string theory. (We won’t get into pasta texture at this time even though good spaghetti simulates a string.) The Background When water interacts with a dissolved solute or with amorphous or crystalline materials, the thermodynamics of the system change such that the free energy of the water is decreased. This is manifested by decreased vapor pressure in the gas phase, a reduced freezing-point curve as the amount of interacting dissolved solute increases, and an increase in boiling and melting points of the solutes. These three curves represent the thermodynamic equilibrium conditions and, at any one point of water/ temperature, the water has a specific free energy that can be represented as the water activity through the first law of thermodynamics (Glasstone 1946; Barbosa-Cánovas and others 2007).

μ = μo + RT Ln ( aw )

(7.1)

In Equation 7.1, μ is the free energy of water in a given state, μ0 is the free energy of pure water at the system temperature T, R is the gas constant, and aw is the measured water activity of the system at those conditions. This concept of water activity (aw) or equilibrium relative humidity (ERH) (i.e., ERH = 100 × aw = 100 × [p/p0]),

Soft Condensed Matter: A Perspective on the Physics of Food States and Stability

Texture

Lipid oxidation

Mobility point

Enzyme activity

Vitamins

Mold

Moisture 0.1

ISOPOW IX 2004

0.2

0.3

0.4

Moisture content

Relative reaction rate

NEB

0.0

89

0.5 0.6 0.7 Water activity

Yeast 0.8

0.9

Bacteria 1.0

Figure 7.1. Water activity (aw) stability diagram of relative reaction rate vs aw. ISOPOW, International Symposium on the Properties of Water; and NEB, nonenzymatic browning.

where p is pressure, was brought to the forefront of food science by Scott (1957) and the early leaders who formed ISOPOW (the International Symposium on the Properties of Water). This led to an understanding of water management in foods in terms of molecular mobility and its relationship to microbial and chemical reactions that cause deterioration, as well as physical state changes, the focus of this review (Labuza and others 1970; Duckworth 1975). What resulted is the aw stability map (Figure 7.1), first presented in 1970 (Labuza and others 1970), and the formation of the ISOPOW group that had its first meeting on water in foods in Glasgow in 1971. The research that followed led to an important list of aw criteria for chemical, physical, and microbiological guide points in assessing the stability of dry, semimoist, and high-moisture foods (Labuza 1975). In practice, this also led to the development of several instruments for reliable aw measurement and the incorporation of aw into government regulations related to food safety. An example is the US Food and Drug Administration’s regulation, 21 CFR 113, which requires that foods with an aw of ≥0.85 coupled with a pH of ≥0.46 that are heated to sterility in sealed hermetic containers, so they are stable at room temperature and pose no pathogenic risk, must have a heat process that delivers a ≥12 log-cycle reduction in the number of Clostridium botulinum spores present. The research by hundreds of researchers in the food science field has led to a series of general rules with respect to the measured aw and food stability, which are listed here: aw ∼ 0.2–0.3: Brunauer-Emmett-Teller (BET) monolayer. At and below this value, reactions requiring a water phase do not occur. aw ∼ 0.2–0.3: Below this, the rate of lipid oxidation increases.

90

PART 1: Invited Speakers and Oral Presentations

aw ∼ 0.35–0.4: Onset of stickiness. aw ∼ 0.4–0.5: Onset of a loss of crispness on gain of moisture by dry foods. aw ∼ 0.5: Onset of hardening (e.g., raisins) on loss of moisture. aw ∼ 0.6: Onset of microbial growth. aw ∼ 0.65–0.85: Maximum reaction rate in amorphous systems as a function of aw (e.g., enzymes, nonenzymatic browning [NEB], lipid oxidation). aw ∼ 0.85: Onset of growth of bacterial pathogens and the limit of aw in thermal processing as already noted. Also of importance is that the difference in the chemical potential of water (μ), and thus the aw difference between two systems, results in moisture exchange (Hyman and Labuza 1998), a critical problem in multidomain systems (e.g., a cheese-and-cracker sandwich). Slade and Levine (1989, 1990) after almost 30 years of the food industry’s practical use of aw as a guiding principle stirred the pot by noting that foods generally are never at true equilibrium and thus a polymer science approach based on Tg phenomena was a better way to understand stability. This Tg was a function of moisture content and represented a moisture vs T line that separated dry and intermediate moisture foods (IMFs) into two states: an amorphous glassy state and an amorphous rubbery state. Thus, we can represent states of matter as in the chart in Figure 7.2, which we will return to. Slade and Levine (1989, 1990) and Levine and Slade (1993), early on when Tg was being introduced to food scientists, suggested that there was no need to invoke any concept of aw at all. The proven use of its principles, however, vis-à-vis the whole IMF revolution in the 1970s; the rapid development of reliable aw-measuring devices by Rotronics, Nova-Sina, and Decagon; and its measured value lending itself to predict potential stability problems, allowed that one should not throw out the baby with the Solid System States Solid Food Matrix State High-moisture elastic gel, colloidal dispersion, dough or cell trapped

Moisture removal or freezing

Amorphous (unstructured)

Concentration or freezing

Crystallization of ice and sugars structured mininum free energy

Moisture loss or T decrease below Tg

Glassy or brittle dry

Hold below Tm but above Tg

Rubbery or ductile semimoist

Moisture gain or T above Tg

Hardening in storage

Figure 7.2. Hypothetical states of matter of foods, drugs, and biologics. T, absolute temperature; Tg, glass transition temperature; and Tm, melting temperature.

Soft Condensed Matter: A Perspective on the Physics of Food States and Stability

91

bathwater; that is, perhaps the two concepts were compatible. It should be noted that, in a 1948 chemical engineering textbook on polymer science and technology, Schmidt and Marles suggested that glass transition could probably explain the physical states of confectionaries, but this was overlooked until 1987 even though the senior author of this chapter used that very textbook in 1959 for a polymer science course in chemical engineering while an undergraduate at MIT. Of course, only one paragraph was devoted to Tg in that book, so this oversight can be forgiven. Given this and the better understanding of physical states, a consolidating approach first presented by Roos and Karel (1991b, 1991c) and reviewed by Rahman (2006) 15 years later is the theoretical development of a state diagram, as shown in Figure 7.3, that combines equilibrium curves with the Tg line as a hypothetical way to understand food states. The state diagram, which combines equilibrium freezing-point, melting-point, and boiling-point curves on the same graph with the glass transition curve (i.e., Tg vs % solids), supplements the aforementioned aw stability map to better explain physical state changes in processing and storage of foods, drugs, and biologics. This is particularly useful when considering frozen foods, and also foods that exist in the dry glassy state either at or slightly above room temperature, occupying a small region on the right side close to the vertical center of the diagram (the small triangular area in Figure 7.3), as well as semimoist foods such as soft candies and meat and fish jerkys. Raising temperature and humidity moves a dry product from the glassy state into the rubbery state (i.e., above the Tg line), thereby allowing for increases in mobility of both water and other molecules and a local, reduced storage modulus related to texture. The consequence of this is that if temperature or moisture content are increased for products in the glassy state, like rigid or brittle dry foods (e.g., crackers), such that they cross the Tg line into the rubbery state, crisps become soft, losing crispness; sugary foods, such as hard candies and infant formula, can crystallize with stickiness and caking that occur; and soft cookies become hard. In addition, reactions requiring an aqueous phase increase in rate exponentially with an increase in the ERH or aw (Labuza and others 2004). The position in either the glassy or the rubbery region thus becomes very important to shelf-life control and is a tool along with the aforementioned aw paradigms for both the understanding and the management of moisture in foods. This can be further illustrated by a simple depiction of the changes in the physical state of a food undergoing processing and storage, as is shown in Figure 7.2. In essence, we begin at the top with a high-moisture system with an aw close to 1 and remove water by some means; for example, through evaporation (boiling), drying, or freezing. On the right side, water can be converted to ice by lowering the temperature or it can be removed by boiling or evaporation as is used in the process to crystallize out solutes such as sugar or salt. The change in state on the drying side on the left brings us into the amorphous state, which can be either a rubbery (semimoist) or glassy (dry) state. For systems high in sugars, such as a pre-candy mix or a cookie dough (state 1 [solid circle] in Figure 7.4 at room temperature), the system during processing by partial evaporation (boiling or baking) can end up in a rubbery state once cooled to room temperature (state 2 [solid square] in Figure 7.4) (e.g., a gummy candy, a soft cookie,

92

PART 1: Invited Speakers and Oral Presentations

or a caramel). By further water removal in baking or boiling, the system will end up as hard and glassy, as designated by the small triangular area in Figure 7.3 (Dry-food region), which is the final place at room temperature for state 3 (solid triangle) in Figure 7.4 (e.g., crispy cracker, hard-ball candy, or peanut brittle). Thus, the state diagram can be used to follow the physical state changes as a function of moisture content (or plasticizer)

Vapor Solution crystal-melt line

Boiling-point line 0 °C

Dry Tm

Freezing-point line

Dry Tg Tamb

Te

Rubber

Ice and super sat solution

Dry-food region

Tg′

Glass

–135 °C 0

100%

% solids

Figure 7.3. A hypothetical state diagram. sat, saturated; Tamb, ambient temperature; Te, eutectic temperature; Tg, glass transition temperature; Tg′ , solute-specific glass transition temperature; and Tm, melting temperature.

Vapor Boiling line

Solution crystal-melt line

Freezing line

0 °C

Tm Tg

23 °C

Rubbery state

Te

Ice and super sat solution

dry

Glassy state

–140 °C

Tg′ 0

100% % solids

Figure 7.4. Hypothetical state diagram showing equilibrium and amorphous regions: state 1 (solid circle), initial nonequilibrium sugar mix or cereal dough; state 2 (solid square), final rubbery state (e.g., for soft cookie or caramel); and state 3 (solid triangle), final glassy state for hard-ball candy, cotton candy, or crisp cookie. Te, eutectic temperature; Tg, glass transition temperature; Tg′, solute-specific glass transition temperature; Tm, melting temperature; and sat, saturated.

Soft Condensed Matter: A Perspective on the Physics of Food States and Stability

93

in relationship to the theoretical equilibrium-state functions during any water-removal process such as drying, baking, extrusion, or evaporation, as well to explain changes occurring in storage.

Physical State Changes in Foods in Storage Glass transition theory from the study of polymer science also helps in understanding textural properties of food systems and explains changes that occur during food storage, such as stickiness, caking, collapse under the force of gravity, softening, and hardening (Stearne and Ward 1969; Levine and Slade 1989; Slade and Levine 1989; Roos and Karel 1991a, 1991b, 1991c; Sperling 1992). Figure 7.5 shows a glass-rubber transition diagram, the line representing the glass temperature, Tg. The glass transition point is the temperature (Tg) at a given moisture content where a transition from a glassy stable amorphous solid state to a rubbery amorphous solid state can begin (Sperling 1992). As seen in Figure 7.5, this can take place by either the temperature being increased or the plasticizer amount (e.g., water or polyols in foods) being increased or both. In the amorphous state of a solid the molecules are randomly distributed and mobile, whereas in a crystal the molecules are in a distinct arrangement with little mobility. This results in a distinct X-ray fingerprint pattern (Suryanarayanan 1995). The glassy state, which is below the Tg line, has brittlelike properties similar to stiff, hard plastics, glass wool, or a crisp cracker. In the rubbery state above the Tg line at constant temperature and thus increasing RH, the material can pick up water and collapse (flow) under the force of gravity. If the amorphous material is composed of small molecules like sugars, the sugars become mobile as they absorb water and then can recrystallize, becoming very hard and losing weight as moisture is lost (Roos and Karel 1991a, 1991b; Chuy and Labuza 1994). This is explained by there being a dramatic increase in the local mobility of both monomers like sugars and polymer chains just above the glass transition temperature

RUBBERY

C

¥ stickiness ¥ caking ¥ collapse

Temperature

B

A

Tg curve GLASSY

% solids

Figure 7.5. Influence of temperature and moisture increase on state change. Tg, glass transition temperature.

94

PART 1: Invited Speakers and Oral Presentations

over the whole moisture range. As a system moves from the glassy into the rubbery state when the temperature or moisture content is increased (relative to the external RH), the local viscosity (or storage modulus G′) drops dramatically at the molecular level from approximately 1012 down to 109 Pa · s just above the glass transition temperature (Sperling 1992). The reduced local viscosity allows for greater polymer chain and reactant mobility. The amount of free volume, defined as the amount of system-associated space that is not taken up by polymer chains themselves, also changes between the glassy and rubbery states. The free volume available within a glassy system has been estimated to be between 2% and 11.3% of the total volume and is believed to result from an increase in the thermal expansion coefficient (Ferry 1980). This increase in free volume should allow for faster diffusion of reactants. Based on the free volume required for diffusion, the size of a diffusing molecule may also be an important factor affecting diffusion rates. Diffusion is a function of the probability of creating a matrix hole that is sufficiently large enough for a molecule to occupy. When a molecule is large compared to the available free-volume pore or voids, the probability of creating such a hole is low. However, in addition to translational mobility, short-range mobility may also be important for reactions such as crystallization. The diffusivity of water is obviously greater in the rubbery than in the glassy state, but one must be careful in stating this since, in the glassy phase, diffusion of both water and oxygen most likely occurs in the vapor phase, with a diffusivity about 10,000 times greater than in the liquid state (2.5 × 10−5 m2/s in the gas phase vs 2 × 10−9 m2/s for the liquid state). Thus, crystallization and reactions could occur, although more slowly below Tg, such as nucleation and crystallization in the microregions. Using electron spin resonance, Roozen and coworkers (1990, 1991) found a significant increase in the rotational mobility of spin probes within sucrose-water, glycerol-water, and maltodextrin-water mixtures at a temperature that corresponded to the glass transition temperature. The mobility of protons, as measured by nuclear magnetic resonance (NMR), has also been found to be higher in the rubbery state compared to the glassy state (Kalichevsky and others 1992). Sherwin and others (2002) and Sherwin and Labuza (2003) used cross-polarization magic-angle spinning NMR to show similar effects in the Maillard reaction at limited aw. In a multidomain food (e.g., with a soft moist center and crisp outer layer), where domain states 2 and 3 (Figure 7.3) are in contact, they will exchange moisture if their water activities (and most likely moisture contents) are different (Hyman and Labuza 1998). Given this, products at state 2 will lose moisture and become more leathery (move to the right), whereas those in state 3 will gain moisture (moving to the left) and become softer, creating another storage problem.

Examples of Use of the State Diagram and Glass Transition Curve To illustrate the use of the state diagram and state changes, three food systems will be discussed: (1) crystallization of cotton candy, (2) stickiness of hard candy, and (3) softening of crisp foods.

Soft Condensed Matter: A Perspective on the Physics of Food States and Stability

95

1. Sugar Recrystallization in Storage of Foods: Sucrose and Cotton Candy The physical structure of a food, important from both functional and sensory standpoints, is often altered by changes in aw due to moisture gain resulting in a transition from the glassy to the rubbery state. Moisture pickup also causes crisp hard baked, fried puffed, or extruded (glassy) foods to become soft due to the plasticization of carbohydrate polymers into the rubber zone (Katz and Labuza 1981). When powders are made by spray drying or freeze drying, the sugars after drying are generally in the amorphous glassy state since they are dried quickly to low moisture so as to prevent recrystallization during drying (Roos and Karel 1991a, 1991b, 1991c). Thus, they end up in the small triangular zone at room temperature shown in Figure 7.2. Since amorphous sugars are very hygroscopic, if they are exposed to high humidity they will cross over into the rubbery zone as a function of the temperature and humidity, and caking, collapse, and crystallization will occur, resulting in state changes forming either a sticky texture or a hard, coarse, or grainy texture (Saltmarch and Labuza 1980; Downton and others 1982). Both loss of crispness and caking generally begin to happen when the aw is raised to above 0.3–0.4 at room temperature via moisture pickup through the package or moisture redistribution from other microdomains. Caking interferes with the powder ’s ability to dissolve or be free flowing, and phase transitions can lead to volatile loss or oxidation of encapsulated lipids. Cotton candy made from pure sucrose is one of the most simple forms of food; in this case, a carbohydrate confection. It has been made for over a century, but little has been studied about its stability during distribution and storage (Labuza and Labuza 2004). Cotton candy is made by a spin-melt-cool and air-drying process that uses crystalline sucrose, possibly with minor additional ingredients (flavoring and coloring). This process has not changed much in the 169 years since the product’s introduction to the American public at the 1830 World’s Fair (Hetzler 2001). There are earlier reports of it, termed angel floss, in 16th-century Europe, but no details could be found. It was commercially introduced sometime in the late 1890s when it was produced by the Nashville candy makers William Morrison and John C. Wharton (Bishop 1998; Davis 2001). They later designed the world’s first electric machine that allowed crystallized sugar to be poured onto a heated spinning plate and then be pushed by centrifugal force through a series of tiny holes. The product was first sold as fairy floss at the 1904 St. Louis World’s Fair. Though the Nashville maker ’s first production is well documented, many believe that cotton candy was invented earlier in the 1900s by a vendor at Ringling Brothers and Barnum & Bailey Circus. Makower and Dye (1956) performed the first experiments on state changes of pure amorphous sucrose by using a freeze-dried frozen saturated sucrose solution rather than a spin-melt-cool process as in the making of cotton candy. They determined the moisture uptake for sucrose at 25°C at different relative humidities (%RHs). At <12% RH (∼2% moisture), the freeze-dried mass was stable over the 3 years of their study, whereas, at 33% RH, it gained about 5% moisture and then recrystallized and lost that moisture in about 3 days. Their group (Palmer and others 1956) also conducted some early X-ray diffraction (XRD) studies to show that crystallization occurred as the water gain over time above 28% RH switched to water loss. The initial gain in weight

96

PART 1: Invited Speakers and Oral Presentations

Onset of sucrose crystallization

Time in days

1000

Days = 10447e–25.054a R2 = 0.9744

100 10 1 0.1

0

0.1

0.2 0.3 Water activity

0.4

0.5

Figure 7.6. Time to crystallization of sucrose as a function of water activity (∼25°C). From Makower and Dye (1956).

is due first to adsorption of moisture by the hydroxyl groups of the amorphous sucrose, which is then followed by a loss in weight due to the increased mobility in the available water, followed by recrystallization of the sugar thereby increasing the local plasticizer phase, which is then lost to the vapor phase as equilibrium occurs. The crystals hold less water (only by hydrogen bonds on the surface -OH groups) and, since the product was stored in a constant-humidity chamber, the local aw in the sugar is raised by the released water, which in turn evaporates into the chamber at the lower aw. Figure 7.6 shows our calculated time to onset of crystallization vs aw from the Makower and Dye (1956) sucrose data (semilog plot). It is obvious that a shelf life of longer than 1 year can be obtained (i.e., no collapse or recrystallization) if un-crystallized sucrose is kept below 15% ERH (aw = 0.15). Figure 7.7 shows the XRD pattern for fresh cotton candy. The y-axis is counts per second, and the x-axis is the Bragg 2-theta angle. The cotton candy was processed in our lab and measured at about 4 h after it was made. The vertical dotted lines are the theoretical angles where peaks should appear for the powder XRD pattern of crystalline sucrose (Labuza and Labuza 2004). Figure 7.7 demonstrates no specific crystalline pattern for the cotton candy as compared to the crystalline sucrose pattern (vertical lines at specific 2-theta values). The broad, low intensity at all angles means that the molecules in the amorphous solid are randomly oriented and diffract X-rays in all directions, giving the typical amorphous “halo” pattern. Thus, cotton candy immediately after being made is a completely glassy amorphous solid material as was the freeze-dried material reported by Makower and Dye (1956). The XRD pattern for the cotton candy stored at both 0.05% RH and 11% RH for up to 5 years showed no crystalline sucrose peaks; instead it has a completely amorphous halo pattern as if just made. The XRD pattern of cotton candy stored at 33% RH (∼5.7% moisture) showed that on days 1 and 2 it was in an amorphous glassy state, whereas on day 3 it went into an amorphous rubbery state and crystalline sucrose peaks appeared. Based on data for the Tg diagram of sucrose, this will be at the moisture content of 5 g/100 g solids, which

Soft Condensed Matter: A Perspective on the Physics of Food States and Stability

97

Log counts 3

2

10

12

14

16

18

20

22

24

26

2-theta degrees

Figure 7.7. Powder X-ray diffraction of initial dry cotton candy. Vertical dotted lines represent the theoretical crystalline sucrose Bragg 2-theta angle peaks.

Log counts 4

3

2 10

12

14

16 18 20 2-theta degrees

22

24

28

Figure 7.8. X-ray diffraction powder pattern of cotton candy stored for 2 h at 45% relative humidity and 23°C. Vertical dotted lines represent the theoretical crystalline sucrose Bragg 2-theta angle peaks.

is the moisture content when the Tg is at room temperature, so one should suspect mobility and crystallization to occur. The cotton candy collapsed into a grainy and sticky mass at day 3 and showed the equivalent moisture gain/loss pattern over the first 3 days as was found by Makower and Dye (1956). The XRD of cotton candy stored for 5 h at 45% RH (Figure 7.8) contained both amorphous and crystalline sucrose, as evidenced by the XRD pattern at 2 h, and quickly collapsed into a sticky grain in 5 h, close to the 2.4 h predicted from Figure 7.6. Note that the peaks range between 1000 and 10,000 counts/s, whereas the amorphous pattern in Figure 7.5 shows a range from 100 to about 1000 counts/s, with no peaks. At 75% RH, this occurred in less than 1 h.

98

PART 1: Invited Speakers and Oral Presentations

It is interesting to note that adding raffinose, a known sucrose crystallization inhibitor, to cotton candy prolonged the onset time of crystallization significantly (Leinen and Labuza 2004). Belcourt and Labuza (2007) found that this also helped to reduce the hardening of soft cookies that is induced by sucrose recrystallization. It is important to understand the temperature dependence of physical and chemical reactions in order to predict product shelf life. Often, shelf-life studies of foods are conducted at elevated temperatures in order to speed data collection. To predict the rates of degradative chemical reactions or time to crystallization at other temperatures, a relationship between the reaction rate and temperature must be established. The Arrhenius relationship (ln rate constant vs 1/T absolute) has traditionally been used to describe the temperature dependence of chemical reactions (Glasstone 1946) and has very extensive applicability for reactions limiting the storage stability of dehydrated and semimoist foods (Labuza 1980, 1985; Labuza and Schmidl 1985). However, when reactions are diffusion limited, the alternative approach by Williams and others (1955) may be appropriate. This latter rate dependence is represented as a function of the temperature above the glass transition temperature (T − Tg) and is also affected by the amount of material dissolved (Roos and Karel 1991b). Thus, the rate of crystallization for sugars in foods depends on the position in the rubbery or glassy states in terms of both temperature and moisture content or aw (Roos and Karel 1990, 1992; Slade and Levine 1990; Nelson and Labuza 1994). For example, Krusch (2003) has shown that the time to crystallization of lactose at different moisture contents follows the Arrhenius equation at values much lower than the 100 K above Tg at each moisture content suggested by Slade and Levine (1990) as the upper region where Arrhenius kinetics would apply. At 1.22% moisture, the Tg is 90°C or 1/T = 0.00276, indicated by the leftmost dashed vertical line, and the data fit neatly into a straight line at 127°C, starting at about 40 K above Tg, as shown in Figure 7.9. At 6.04% moisture, the measured Tg was 55°C, which is off the right y-axis on this plot (rightmost dashed vertical line); again the Arrhenius fit begins at only 16 K above Tg. Thus, one must be careful about generalization of trends. The prior work by Roos and Karel (1990) on lactose crystallization where they presented the theorem that the Williams-Landel-Ferry (WLF) Equation 7.1 was the proper approach has not been supported from two standpoints. First, they combined all data from all temperatures and moistures into the WLF equation to do the nonlinear regression of the WLF Equation 7.2, whereas the seminal report by William and others (1995) indicates that the equation fit should be done separately for different plasticizer content similar to the application of the Arrhenius equation to reaction rates at different moisture content as indicated in Figure 7.9. log

kref −C1 (T − Tref ) = k C2 + (T − Tref )

(7.2)

A different approach to predict time to crystallization near Tg is to rearrange the WLF equation to get a linear fit (Nelson and Labuza 1994) as shown by Equation 7.3:

Soft Condensed Matter: A Perspective on the Physics of Food States and Stability

99

Log of time to crystallization

100000

1.22 3.44 6.04

10000

1000

100

10

1 0.00220 0.00230

0.00240

0.00250 0.00260 0.00270 0.00280 0.00290 0.00300

1/T

Figure 7.9. Arrhenius plot for crystallization of lactose at three different moisture contents from the data of Krusch (2003). Numbers represent dry-basis % moisture.

0 0

0.01

0.02

Linearized WLF plot 0.03 0.04 0.05

0.06

0.07

0.08

–0.1 –0.2

1/ln (tc/tg)

–0.3

y = -8.1962x - 0.0141 R2 = 0.9849

–0.4

1.22 3.44 6.04

–0.5 –0.6

y = -9.2238x - 0.1125 R2 = 0.8126

–0.7 –0.8

y = -11.489x - 0.0166 R2 = 0.9363

–0.9 –1

1/(T - Tg)

Figure 7.10. Modified Williams-Landel-Ferry (WLF) plot for lactose crystallization.

−1

kref ⎤ ⎡ ⎡ 1 ⎤ ⎡ C2 ⎤ ⎡ 1 ⎤ ⎢⎣ log k ⎥⎦ = − ⎢⎣ C ⎥⎦ − ⎢⎣ C ⎥⎦ ⎢ T − T ⎥ 1 1 ref ⎦ ⎣

(7.3)

A straight line with a slope equal to −C2/C1 and an intercept of −1/C1 is found if the WLF model is applicable. As shown in Figure 7.10, the same data from Figure 7.9 fit Equation 7.3, but with different lines at each moisture content, illustrating the

100

PART 1: Invited Speakers and Oral Presentations

prohibition against combining all moisture and plasticizer data into one line, as was noted during the original equation development (Williams and others 1955). The reason for both equations fitting is that the temperature range of the data is small enough that it fails to show deviations. A test of which one is the right fit means determining the actual times to crystallization at Tg, which is a difficult task but possible, at least in cotton candy as shown earlier. 2. Stickiness of Hard Candy Before the establishment of the concept of glass transition, numerous studies investigated the temperature and humidity conditions in which sticking or caking can occur during storage of powders containing amorphous sugars, such as dairy powders, infant formula, and drink mixes (White and Cakebread 1966; Downton and others 1982). Chuy and Labuza (1994) and Jouppila and Roos (1997) clearly showed that the key to the understanding of state changes is related to the temperature of storage in relationship to the glass transition temperature; that is, T − Tg. This was followed by many studies, including reports by Lloyd and others (1996), Roos and others (1998), Roozen and others (1990, 1991), and Roe and Labuza (2005). They have all used the concept of the temperature difference by which the glass transition temperature is exceeded (T − Tg) to describe the sticking and caking behavior of amorphous powders. In general, the critical state for powders during storage is reached at a temperature of about 10°–20°C above Tg, the latter as determined by differential scanning calorimetry (DSC), as shown in Figure 7.11 from Chuy and Labuza (1994), where they used the

120

Tg Temperature °C

100

Tac = -9.3681m + 154.85 R2 = 0.3522 Tsc = -11.622m + 149.84 R2 = 0.9278

80 60

Tg = -7.5901m + 94.477 R2 = 0.9531

40 20 0 1

2

3 4 5 6 7 Moisture content (db) g H2O/100 g solids

8

Figure 7.11. Tg by differential scanning calorimetry (open circles); Tsc (solid triangles), sticky point determined by the ampoule method; and Tac (open triangles), advance caking point determined by the propeller method, as a function of moisture content for an infant formula at a heating rate of 10°C/min. Tg, glass transition temperature; Tsc, sticky point temperature; and Tac, advance caking point temperature.

Soft Condensed Matter: A Perspective on the Physics of Food States and Stability

101

ampoule method to determine onset of stickiness and the propeller method to determine advanced caking. Hard-ball candy is made by boiling a solution of sugar and corn syrup to >180°C, which is then rapidly cooled in a mold without crystallizing. At room temperature and low humidity, the hard ball exists as a hard glasslike amorphous solid of the SCM state noted earlier. This is a much more compact solid amorphous state (∼1200 g/L) then found for cotton candy (50–100 g/L) discussed previously. Sucrose crystals have a density of 1587 g/L; thus, the glass density is less, but depends on the method used to make it. An observation made by Katherine Labuza when she inadvertently left a commercial hard-ball product (Altoids) in the family car in the direct summer sun in the central valley of California was that the product became unattractive by virtue of being completely stuck together. This suggested that the physical state of hard-ball candy may depend on both temperature and RH at least on the surface. A composite T vs % sucrose transition line formed from the data of Roe and Labuza (2005) and Sun and Zografi (1996) is presented in Figure 7.12. Figure 7.12 represents the amorphous states of a sucrose-based hard candy that would be glassy below the line and rubbery above the line. In addition, the data of Labuza and Labuza (2004), noted previously, used this plot to show that cotton candy entered the sticky state and collapsed (i.e., flow of the material under the force of gravity) if held at or above the Tg line at ≥33% (i.e., higher moisture and lower % sucrose). This is very close to the moisture content that gives a Tg of 23°C. Based on this work, one could speculate that if a hard candy is abused at high temperatures or increased humidity or both, the candy could become sticky and clump together, causing consumer complaints. Hard candy may be exposed to very warm temperatures. For example, McLaren and others (2005) demonstrated that the interior of a car,

100

Experimental Tg

Glass transition temperture

80

Gordon Taylor

60 40 20 0 –20 –40 –60 –80

60

70

80 % Sucrose

90

100

Figure 7.12. Glass transition line for sucrose. Composite of data by Sun and Zografi (1996) and Roe and Labuza (2005). Tg, glass transition temperature.

102

PART 1: Invited Speakers and Oral Presentations

Vehicle Temperature Rise Over Time

Vehicle temperature (°F)

140 130 120 110

Ambient temperature 71 74 77 78 84 88 93

100 90 80 70

0

5

15

10

20

25 30 35 40 Elapsed time (min)

45

50

55

60

Glass transition temperature (°C)

Figure 7.13. Car interior temperature rise over time (McLaren and others 2005). 50.0 Tg onset Tg end Tg half

40.0 30.0 20.0 10.0 0.0 –10.0 –20.0 –30.0

0

20

40 RH (%)

60

80

Figure 7.14. Glass transition temperature as a function of % relative humidity (RH) for hard-ball candy (surface material).

a typical place where candy might be stored, left in the sun can reach temperatures over 60°C (140°F) (Figure 7.13). To determine why this occurs, the hard-ball candy was held at 5 different %RHs for 3 weeks and then the outer surface was shaved off and a DSC scan was run on the scrapings. This resulted in obtaining both the Tg and crystal melt of the sucrose since, during the run at 10 K/min, the candy crystallized first from the amorphous rubbery state, as temperature increased, and then melted. Figures 7.14 and 7.15 show these results as a function of %RH. As noted, the T at room temperature is at about 30% RH, similar to the prior studies already noted. Using this as the basis for state changes, the Altoids were then humidified to 4 humidities over saturated salt solutions for 3 weeks. Surface material was scrapped

Soft Condensed Matter: A Perspective on the Physics of Food States and Stability

103

Melting temperature (°C)

140

120

100

80

60

10

0

20

30

40

50

60

70

80

RH (%)

Figure 7.15. Onset temperature for crystallization of amorphous-state hard-ball candy as a function of % relative humidity (RH).

Glass transition temperature (°C)

70 60 50

5

5

4

4

30 20

5

4

5

3.5

3.5

1

1.5

3.5

4.5

1

1

3.5

3.5

2.5

40

5

4.5

10 0 –10

Tg end

–20

Tg start

–30

0

20

40

60

80

RH (%)

Figure 7.16. States of hard-ball candies stored at different abuse conditions as compared to storage condition at or above Tg. A PerkinElmer (PE) Model 7 differential scanning calorimeter was used with a scanning rate of 10°C/min. 1, free flowing and separate; 2, stuck pieces break apart with gentle shaking; 3, very sticky pieces that pull apart by hand; 4, pieces stuck in one mass; and 5, pieces all flow into one mass in shape of hockey puck. RH, relative humidity.

off samples from each humidity and then subjected to DSC to obtain the onset and end-point Tg values shown in Figure 7.16. The cover was put on, and then the product was subjected to 12-h storage at 25°, 35°, 45°, 55°, and 65°C, cooled, and then evaluated for obvious physical state changes on a five-point scale as shown in Figure 7.16.

104

PART 1: Invited Speakers and Oral Presentations

Figure 7.17. Visualization of hard-ball candy states after humidification at 33% relative humidity and then temperature abused.

These results show that if stored at about 7°–10°C above the Tg line, the candy begins to become sticky, whereas at or below Tg it remains free flowing after cooling to room temperature. This follows the pattern shown for powders containing amorphous-phase sugars mentioned earlier. Once the storage condition is >10°C above Tg, the candies stick together such that they do not fall out of an open can when it is turned upside down. At >55°C storage temperature, the force of gravity overcomes the internal hydrogen-bonding force within the hard ball such that the mass flows, completely disrupting the original structure reacting in the material becoming a hockey-puck shape, as in the bottom right of Figure 7.17. Note that at 65°C at all humidities tested, the product is below the melting point of sugar (Figure 7.15), so the flow is not a crystalline melt. It is obvious from this study that increasing either the temperature or the moisture content in the region above Tg increases molecular mobility, allowing the product to become sticky and then flow. Mazzobre and others (1997) found that raising the Tg of trehalose in a mixture inhibited the rate of crystallization of lactose by decreasing mobility; that is, resulted in a smaller T − Tg. Review of these results produced the following observations: (1) Hard-ball candy (e.g., Altoids) is an amorphous glassy material, which initially shows a physical state in which the candy is free flowing and can be easily picked out of a can. (2) Such candies are subject to stickiness when exposed to abusive temperatures or stored at higher humidities, if they enter the rubbery zone above Tg. This can occur when stored in a closed car directly exposed to the sun, where the inside air temperature can increase to 60°C. (3) The glass transition curve (Tg vs %RH) can be used to define the dividing line for optimal storage conditions where below this line the

Soft Condensed Matter: A Perspective on the Physics of Food States and Stability

105

candies are separate and nonsticky, whereas above the line the hard candies go through various sticky and flow states. (4) Closed cans with or without the original seal provide some protection against %RH increases at temperatures below 45°C; that is, the candies may become a little sticky, but shaking the can separates the individual hard candies. However, unopened cans with or without a seal demonstrate that a temperature increase to 55° or 65°C has a greater detrimental effect than an increase in %RH to 75%. (5) At all %RHs, temperatures of 55° and 65°C cause serious physical state changes, including complete flow and collapse to form a hockey-puck-shaped structure (solid mass), well below the melt temperature at ≤58% RH; thus the caution on the product package to “store in a cool dry place.” 3. Loss of Crispness or Hardness Texture is an important sensory attribute for many cereal-based foods, and the loss of desired texture leads to a loss in product quality and a reduction in shelf life (AC Nielsen Company 1979). Recently, the First International Symposium Crispy Cracks: Creating and Retaining the Crispiness of Food was held at the University of Wageningen in the Netherlands (Crispy Cracks 2008). The texture of dry crisp foods was first studied as a function of aw by Katz and Labuza (1981), although work 5 years earlier at Cornell University by Vickers and Bourne (1976) indicated that the perception of the crispness of dry cereal snacks was the result of sounds generated when chewed, which diminished as the aw was increased. Note that this work on crispness occurred about 10 years before an understanding of the application of glass transition to food. Vickers and Bourne concluded that the perception of crispiness was due to the sound generated in the mouth, moving through the jaw bone to the ear. Katz and Labuza (1981) determined that saltine crackers (baked), popcorn (hot-oil puffed), extruded corn curls, and deep-fat-fried potato chips lost crispness if the aw exceeded the range of 0.35–0.50. The crispness was attributed to intermolecular hydrogen bonding of starch forming small crystalline-like regions when little water was present. These regions require force to break apart, which gives the food a crisp texture as fractures perpetrate through the structure during chewing. Above a certain aw, the water was presumed to disrupt these bonds, allowing the starch molecules to slip past one another when chewed. Hsieh and others (1990) also observed that puffed-rice cakes lost crispness at an aw just above 0.44. Sauvageot and Blond (1991) suggested that some type of physicochemical change other than lubrication by water was occurring in the cereal systems because the crispness intensity decreased sharply instead of gradually as aw increased. They were on the right track, but both they and Roudaut and others (1998) conducted only cursory sensory studies that did not show a direct relationship with thermomechanical properties. Glass transition provides a clearer approach to understanding the physical and texture changes of crisp cereals or snacks as water content increases. If an amorphous material exists in the glassy state, it is hard and brittle; for cereal-based snack foods, it would represent a crisp or hard product because the elastic modulus, which is related to flow, extendibility, and bending, is approximately 103 times higher in the glassy state than it is in the rubbery state (Sperling 1992). In the rubbery state, the material is soft and elastic; for

106

PART 1: Invited Speakers and Oral Presentations

a fried snack or cereal, this would be an undesirable soggy state. The desirable crispiness of crackers and dry snack products such as hard cookies, crackers, potato chips, and breakfast cereals is lost if the moisture gained is enough to exceed the moisture content of a matrix that is just at the glass temperature for that material. In Figure 7.4, state 3 represents the crisp glass product, so it could lose crispness by gaining moisture at a constant temperature (i.e., moving toward state 2) or by being stored at a temperature above the Tg for the given moisture content. If textural changes in a cereal system can be correlated with a glass transition, and the state diagram for the cereal food is known, then the processing and environmental conditions can be controlled such that the desired state for the food is achieved and retained during distribution and storage. Unfortunately, measurement of the Tg of composite complex foods like crackers or cookies is not easy. DSC is problematic because the sample size (10–20 mg) is too small. Nikoladis and Labuza (1996) were among the first to use dynamic mechanical thermal analysis (DMTA) on cookies and crackers. They noted the main difficulties were in mounting the sample, the loss of moisture during the temperature scan, and the fact that the recorded temperature was not exactly the product temperature. About the same time, both Roudaut and others (1998), in France, and Roos and others (1998), in Finland, suggested that crispness was related to Tg, they but offered no exact sensory proof as to where the sensory loss of crispness occurs in terms of a moisture-temperature curve like the Tg vs moisture curve. An easy alternative to the measure of glass transition is brittle ductile transition (Vincent 1960; Young and Lovell 1991; ASTM 1997; Folk and others 1998; Dobraszczyk and Vincent 1999; Payne and Labuza 2004a). The brittle fracture is distinguished by the capability of an object broken into pieces to be returned to its original size and dimensions. As temperature or moisture increases, the food system begins to yield before fracture occurs. When the temperature of the object is raised, the object yields to the stress being applied; that is, it begins to deform and change shape. When the sample becomes ductile, there is no clean break. Upon fracture, the sample cannot be put back together in its original dimensions. A brittle material breaks cleanly and can be put back together in its same initial shape, whereas a ductile material deforms during stress, does not break cleanly and, if put back together, has a different shape. Brittle fracture and yielding (deformation or ductile behavior) are separate processes and have a different dependence on temperature (Vincent 1960; Stearne and Ward 1969; Young and Lovell 1991; Ward and Hadley 1993). In general, the maximum stress before breaking or deformation changes little as a function of either temperature at constant plasticizer content, or change in plasticizer amount at constant temperature. After a certain point, the property measured decreases dramatically. The point where the two lines cross corresponds to the brittle-ductile transition temperature (Tb) (Vincent 1960; Stearne and Ward 1969; Young and Lovell 1991; Ward and Hadley 1993). The brittle-ductile transition in food systems has been shown to be comparable to the glass transition (Dobraszczyk and Vincent 1999) and to the decrease in the sensory textural attribute, crispness (Nicholls and others 1995; Le Meste and others 1996). In

Soft Condensed Matter: A Perspective on the Physics of Food States and Stability

107

most of these studies, the brittle-ductile transition has generally been determined over a range of moisture contents at only one temperature. Another way to determine the brittle-ductile transition is by keeping the amount of plasticizer constant and testing the material as a function of temperature (Folk and others 1998). Nicholls and others (1995) suggested that the change in crispness in snack-food systems correlates better with the Tb than the Tg but, again, did not offer quantitative proof. The fracture response (brittle-ductile transition) of a food material is commonly determined by a three-point bend method (Dobraszczyk and Vincent 1999). At a predetermined cross-head speed, a probe is used to apply force in the center of a sample that is supported on each side. From this, the peak force and the apparent Young’s modulus (stress divided by deformation at that distance) can be measured. Very simply, to determine Tb, product is stored at 23°C at five or six different water activities. After equilibration, some of this product is transferred, after being wrapped tightly with foil, to several temperature cabinets to equilibrate. The cookies are then tested on a texture analyzer for the stress-strain relationship. Following this procedure, one can the get the brittle-ductile transition as the intersection point between the two regions. Labuza (2002) in some preliminary studies applied this to brittle, chocolateflavored (not chocolate-coated) sugar snap cookies in a series of tests at both constant T with variable moisture and constant moisture with variable T based on the aforementioned principles. As shown in Figure 7.18, a clear break point is seen in each case. What is useful about this procedure is that the actual measurement of Tb takes less than 30 min after equilibration of the sample, and additional samples can be used for sensory evaluation. Based on this preliminary work, Payne and Labuza (2004a) conducted a more extensive study with 5 moistures and 8–10 temperatures. Figure 7.19a shows an example stress-strain plot for the same sugar snap wafer cookie stored at 0.05% RH and 23°C. The stress rises steeply and then, at the peak, falls to zero as the cookie breaks. In Figure 7.19b, the same cookie was stored at 23°C and 75% RH. At 75% RH in Figure 7.19b, there is no clean break, with an extended deformation to 5 mm vs the break at 0.5-mm deformation. In addition, the peak force went from about 23 Newtons for the dry wafer down to 1 Newton, a factor of 25-fold. These data were then used to determine a Tb vs moisture-content line. Payne and Labuza (2004a) showed that the Tb curve closely followed the Tg curve measured by dynamic mechanical analysis (DMA) or DMTA for a brittle sugar snap cookie. It was about 10° to 30°C lower than the Tg line. Payne and Labuza (2004b) then took this work further and used 10 trained panelists to determine sensory crispness intensity at 8–10 temperatures for each of five different moisture contents. In addition to sensory analysis, they also measured the elastic properties of the sugar snap cookie by DMTA, DMA, and the brittle-ductile test, as noted earlier. Figure 7.20 shows the actual measured values determined by an individual tester in triplicate at nine different temperatures. As shown, at a certain temperature, crisp-

108

PART 1: Invited Speakers and Oral Presentations

(a) 30.00

Force in Newtons

25.00 20.00 15.00 10.00 5.00 0.00 0.00

20.00

40.00

60.00

80.00

% Relative humidity (b) 30.00

Force in Newtons

25.00 20.00 15.00 10.00 5.00 0.00 –40.00

–20.00

0.00

20.00

40.00

60.00

Temperature (°C)

Figure 7.18. (a) Brittle-ductile transition temperature (Tb) plot as a function of moisture (i.e., % equilibrium relative humidity) at constant temperature (25°C) for a sugar snap cookie (Labuza 2002). (b) Tb plot as a function of temperature for a sugar snap cookie at constant moisture (∼6%) content (Labuza 2002).

ness intensity decreases dramatically. The Tci (crispness intensity temperature) for that moisture was determined from the data by applying the Fermi equation using the procedure of Peleg (1992, 1994). The end result of the comparison is presented in Figure 7.21. As shown, the sensory line falls about 7°C above the brittle-ductile line as a function of moisture content and at about the same distance below the elastic storage modulus (G′) as measured by DMTA. This remarkable correlation suggests that, indeed, sensory crispness is related to the ability of water to plasticize the polymer structure. It also suggests that measuring the Tb, which is a lot simpler, would be a useful tool for bakeries in terms of optimizing moisture loss during baking. It

(a)

Force (N) 25.00

20.00

15.00

10.00

5.00

–1.5

–1.0

0.00 –0.5 0.0

0.5

1.5

2.0

Distance (mm)

–5.00 (b)

1.0

Force (N) 1.200 1.000 0.800 0.600 0.400 0.200

–5.0

–2.5

0.000 0.0

–0.200

2.5

5.0

7.5

10.0

Distance (mm)

–0.400

Figure 7.19. (a) Stress/strain plot for a wafer cookie at ∼0.05% relative humidity (RH) and 23°C (Payne and Labuza 2004a). (b) Stress/strain plot for a wafer cookie at ∼75% RH and 23°C (Payne and Labuza 2004a). N, Newton.

109

120

Sensory score

100 80

Modeled by Fermi’s equation

60 40 20 0

–20

0

40

20

60

80

Temperature (°C)

Figure 7.20. Sensory crispness intensity (solid circles) as a function of temperature for a wafer cookie at ∼5% moisture content. The intersection point represents the critical point for onset of crispness loss.

60 50

G′ midpoint

Temperature (°C)

40 30 20 10 0 –10

3

4

5

6

7

8

9

10 G′ onset

–20

Sensory

–30

Brittle ductile

–40 Moisture (g H2O/ 100 g solids)

Figure 7.21. Transition temperatures of chocolate wafers as a function of moisture content: sensory crispness (open circles); brittle-ductile method (X); G′ onset, Tg by dynamic mechanical thermal analysis (DMTA) (solid diamonds); and G′ midpoint, Tg by DMTA (open triangles). G′, elastic storage modulus; and Tg, glass transition temperature.

110

Soft Condensed Matter: A Perspective on the Physics of Food States and Stability

111

further suggests that loss of sensory crispness is very close to the Tg for any moisture content.

Conclusions This chapter has reviewed the influence of moisture and temperature on physical state properties and food changes related to the crystalline, glassy, and rubbery states; that is, the properties of soft condensed matter. The results point out the need for a state diagram that includes the glass transition line. Data for cotton candy and hard-ball candy show that caking and crystallization occur in real time at or near the Tg value. Projection of time to crystallization from higher-temperature data can be modeled with either the WLF equation or the Arrhenius equation, but one must be careful not to combine data at different moisture contents. We also showed that loss of sensory brittleness of foods is related to the plasticizing nature of water and seems to follow the brittle-ductile temperature transition curve (Tb) as a function of moisture content. Finally, on July 29, 2008, the day this chapter was submitted, the article “The Nature of Glass Remains Anything But Clear” appeared in the New York Times. In it was the following statement by Philip W. Anderson, a Nobel Prize–winning physicist at Princeton University, who wrote in 1995, “The deepest and most interesting unsolved problem in solid state theory is probably the theory of the nature of glass and the glass transition.” The entire article concerned physicists around the world working on this sticky problem but mentioned nothing about food science and Tg. It seems to me we are somewhat ahead in this area of SCM and have proven its use in solving many problems.

References AC Nielsen Company. 1979. Product and package performance: the consumers view. Chicago: AC Nielsen. ASTM (American Society for Testing and Materials). 1997. Standard test methods for flexural properties of unreinforced and reinforced plastics and electrical insulating materials. In: Annual book of ASTM standards, vol 8.01: standard method 790. Philadelphia: ASTM. p 141–51. Barbosa-Cánovas GV, Fontana AJ, Schmidt SJ, Labuza TP. 2007. Water activity in foods: fundamentals and applications. Ames, IA: Blackwell and IFT. Belcourt L, Labuza TP. 2007. Effect of raffinose on sucrose recrystallization and textural changes in soft cookies. J Food Sci 72:C65–71. Bishop M. 1998. Cotton candy. Times-Delphic, May 5. Available from: http://www.timesdelphic.com/. Chuy L, Labuza TP. 1994. Caking and stickiness of dairy based food powders related to glass transition. J Food Sci 59:43–6. Crispy-Cracks. 2008. First international symposium Crispy Cracks: creating and retaining the crispiness of food, March 19–20, Wageningen, The Netherlands. Available from TI Food and Nutrition: http://www. tifn.nl/webdb/MirrorID/MC4758A65B995AC7FC12573E00036F3B1?OpenDocument&Prog=Home. Davis B. 2001. The history of cotton candy: dollars, sense, and you. Available from: http://www.pacul.org/ communications/Dollars_Sense_and_You/2000/0008_Doll.htm (accessed May 2002). Dobraszczyk BJ, Vincent JFV. 1999. Measurement of mechanical properties of food materials in relation to texture: the materials approach. In: Rosenthal AJ, editor. Food texture: measurement and perception. Gaithersburg, MD: Aspen. p 99–151. Downton GE, Flores-Luna JL, King CJ. 1982. Mechanism of stickiness in hygroscopic, amorphous powders. Ind Eng Chem Fundam 21:447–51.

112

PART 1: Invited Speakers and Oral Presentations

Duckworth RB. 1975. Water relations in foods. London: Academic. Ferry JD. 1980. Viscoelastic properties of polymers. 3rd ed. New York: John Wiley. p 264–320. Folk II RH, Khantha M, Pope DP, Vitek P. 1998. Temperature-dependent onset of yielding in dislocationfree silicon: evidence of a brittle-to-ductile transition. In: Beltz GE, Blumberg-Selinger RL, Kim K-S, Marder MP, editors. Fracture and ductile vs. brittle behavior: theory, modelling and experiment. Boston: Materials Research Society. p 161–7. Glasstone S. 1946. Textbook of physical chemistry. 2nd ed. Princeton, NJ: Van Nostrand. Hetzler R. 2001. The history of sugar. Available from Monitorsugar.com: http://www.monitorsugar.com/ htmtext/%20HISTORY.htm. Hsieh F, Hu L, Huff HE, Peng IC. 1990. Effects of water activity on textural characteristics of puffed rice cake. Lebensm Wiss Technol 23:471–3. Hyman C, Labuza TP. 1998. Moisture migration in multidomain systems. Trends Food Sci Technol 9:47–55. Jouppila K, Roos J. 1997. The physical state of amorphous corn starch and its impact on crystallization. Carbohydr Polym 32:95–105. Kalichevsky MT, Jaroszkiewicz EM, Ablett S, Blanshard JMV, Lilleford PJ. 1992. The glass transition of amylopectin measured by DSC, DMTA and NMR. Carbohydr Polym 18:77–88. Katz EE, Labuza TP. 1981. The effect of water activity on the sensory crispness and mechanical deformation of snack food products. J Food Sci 46:403–9. Krusch L. 2003. Lactose crystallization kinetics [MSc thesis]. St Paul: University of Minnesota. Available from: https://filenet.software.umn.edu:8458/research_discovery/centers/mndak/publications/theses. html. Labuza PS, Labuza TP. 2004. Influence of temperature and relative humidity on the physical states of cotton candy. J Food Process Preserv 28:274–87. Labuza TJ. 2002. Brittle ductile behavior of a sugar snap cookie. Minnesota State Science Fair competition paper. Available from: T. P. Labuza and T. J. Labuza. Labuza TP. 1975. Sorption phenomena in foods. In: Rha C, Reidel D, editors. Theory, determination and control of physical properties of foods. Dordrecht, The Netherlands: Reidel. p 197–219. Labuza TP. 1980. The effect of water activity on reaction kinetics of food deterioration. Food Technol 34:36–41, 59. Labuza TP. 1985. Applications of chemical kinetics to deterioration of foods. J Chem 61:348–58. Labuza TP, Roe K, Payne C, Panda F, Labuza TJ, Labuza PS, Krusch L. 2004. Storage stability of dry food systems: influence of state changes during drying and storage. In: Silva M, Rocha S, editors. Proceedings of the 14th International Drying Symposium (IDS 2004), vol A: Drying 2004. São Paulo: Ourograf Grafica Campinas. p 48–68. Available from: IDS 2004. Labuza TP, Schmidl MK. 1985. Accelerated shelf-life testing of foods. Food Technol 39:57–62. Labuza TP, Tannenbaum SR, Karel M. 1970. Water content and stability of low-moisture and intermediatemoisture foods. Food Technol 24:543–50. Leinen K, Labuza TP. 2004. Influence of raffinose on collapse and crystallization of cotton candy. J Zhejiang Univ Sci [B] 7:79–83. Le Meste M, Roudaut G, Davidou S. 1996. Thermomechanical properties of glassy cereal foods. J Therm Anal 47:1361–75. Levine H, Slade L. 1989. Influences of the glassy and rubbery states on thermal, mechanical, and structural properties of dough and baked products. In: Faridi H, Faubion JM, editors. Dough rheology and baked product texture. New York: Van Nostrand Reinhold/AVI. p 157–300. Levine H, Slade L. 1993. The glassy state in applications for the food industry, with an emphasis on cookie and cracker production. In: Blanshard JMV, Lillford PJ, editors. The glassy state in foods. Loughborough, UK: Nottingham University Press. p 333–74. Lloyd RJ, Chen XD, Hargreaves JB. 1996. Glass transition and caking of spray dried lactose. Int J Food Sci Technol 31:305–11. Makower B, Dye WB. 1956. Equilibrium moisture content and crystallization of amorphous sucrose and glucose. J Agric Food Chem 4:72–7.

Soft Condensed Matter: A Perspective on the Physics of Food States and Stability

113

Mazzobre MF, Burea MP, Chirife J. 1997. Protective role of trehalose on thermal stability of lactose in relation to its glass and crystal forming properties and effect of delaying crystallization. Lebensm Wiss Technol 30:312–29. McLaren C, Jan Null CCM, Quinn J. 2005. Heat stress from enclosed vehicles: moderate ambient temperatures cause significant temperature rise in enclosed vehicles. Pediatrics 116:e109–12. Mezzenga R, Schrutenberger P, Burbridge A, Michel M. 2005. Understanding foods as soft materials. Nat Mater 4:729–40. Nelson K, Labuza TP. 1994. Arrhenius vs WLF kinetics in the rubber and glassy state. In: Fito P, Mulet A, McKenna B, editors. Water in foods. London: Elsevier Applied Science. p 271–90. Nicholls RJ, Appelqvist IAM, Davies AP, Ingman S, Lillford PJ. 1995. Glass transitions and the fracture behaviour of gluten and starches within the glassy state. J Cereal Sci 21:25–36. Nikoladis A, Labuza TP. 1996. Glass transition state diagram of a baked cracker and its relationship to gluten. J Food Sci 61:803–6. Palmer KJ, Dye WB, Black D. 1956. X-ray diffractometer and microscopic investigation of crystallization of amorphous sucrose. J Agric Food Chem 4:77–81. Payne C, Labuza TP. 2004a. The brittle-ductile transition of an amorphous food system. J Dry Technol 23:1–16. Payne C, Labuza TP. 2004b. Correlating perceived crispness intensity to physical changes in an amorphous snack food. J Dry Technol 23:17–36. Peleg M. 1992. On the use of the WLF model in polymers and foods. Crit Rev Food Sci Nutr 32:59–66. Peleg M. 1994. A model of mechanical changes in biomaterials at and around their glass transition. Biotechnol Prog 10:385–8. Rahman MS. 2006. State diagram of foods: its potential use in processing and product stability. Trends Food Sci Technol 17:29–141. Roe K, Labuza TP. 2005. Glass transition of amorphous trehalose-sucrose systems. J Food Properties 8:559–74. Roos Y, Karel M. 1990. Differential scanning calorimetry study of phase transitions affecting the quality of dehydrated materials. Biotechnol Prog 6:159–63. Roos Y, Karel M. 1991a. Application of state diagrams to food processing and development. Food Technol 45:66–71. Roos Y, Karel M. 1991b. Phase transitions of mixtures of amorphous polysaccharides and sugars. Biotechnol Prog 7:49–53. Roos Y, Karel M. 1991c. Plasticizing effect of water on thermal behavior and crystallization of amorphous food models. J Food Sci 56:38–43. Roos Y, Karel M. 1992. Crystallization of amorphous lactose. J Food Sci 57:775–7. Roos YH, Roininen K, Jouppila K, Tuorila H. 1998. Glass transition and water plasticization effects on crispness of a snack food extrudate. Int J Food Properties 1:163–80. Roozen MJGW, Hemminga MA. 1990. Molecular motion in sucrose-water mixtures in the liquid and glassy state as studied by spin probe ESR. J Phys Chem 94:7326–29. Roozen MJGW, Hemminga MA, Walstra P. 1991. Molecular motion in glassy water-malto-oligosaccharide (maltodextrin) mixtures as studied by conventional and saturation-transfer spin-probe ESR spectroscopy. Carbohydr Res 215:229–37. Roudaut G, Dacremont C, Le Meste M. 1998. Influence of water on the crispness of cereal-based foods: acoustic, mechanical, and sensory studies. J Texture Stud 29:199–213. Saltmarch M, Labuza TP. 1980. Influence of relative humidity on the physicochemical state of lactose in spray-dried sweet whey powders. J Food Sci 45:1231–6, 1242. Sauvageot F, Blond G. 1991. Effect of water activity on crispness of breakfast cereals. J Texture Stud 22:423–42. Schmidt A, Marles C. 1948. Principles of high polymer theory and practice. New York: McGraw-Hill. Scott WJ. 1957. Water relations of food spoilage microorganisms. Adv Food Res 7:83–127. Sherwin C, Labuza TP. 2003. Role of moisture in Maillard browning reaction rate in intermediate moisture foods: comparing solvent phase and matrix properties. J Food Sci 68:558–94.

114

PART 1: Invited Speakers and Oral Presentations

Sherwin C, Labuza TP, McCormack A, Chen B. 2002. Cross-polarization/magic angle spinning NMR to study glucose mobility in a model intermediate-moisture food system. J Agric Food Chem 50:7677–83. Slade L, Levine H. 1989. A food polymer science approach to selected aspects of starch gelatinization and retrogradation. In: Millane RP, BeMiller JN, Chandrasekaran R, editors. Frontiers in carbohydrate research, 1: food applications. Proceedings of a Conference on Frontiers in Carbohydrates Research held at Purdue University, 13–15 September 1988. London: Elsevier Applied Science. p 215–70. Slade L, Levine H. 1990. Beyond water activity: recent advances based on an alternative approach to the assessment of food quality and safety. Crit Rev Food Sci Nutr 30:115–360. Sperling LH. 1992. Introduction to physical polymer science. New York: John Wiley. p 224–95. Stearne JM, Ward IM. 1969. The tensile behaviour of polyethylene terephthalate. J Mater Sci 4:1088–96. Sun WQ, Zografi G. 1996. Stability of dry liposomes in sugar glasses. Biophys J 70:1769–76. Suryanarayanan R. 1995. X-ray powder diffractometry. In: Harry GB, editor. Physical characterization of pharmaceutical solids. New York: Marcel Dekker. p 157–300. Vickers ZM, Bourne MC. 1976. A psychoacoustical theory of crispness. J Food Sci 41:1158–64. Vincent PI. 1960. The tough-brittle transition in thermoplastics. Polymer 1:425–44. Ward IM, Hadley DW. 1993. An introduction to the mechanical properties of solid polymers. Chichester, UK: John Wiley. White GW, Cakebread SH. 1966. The glassy state in certain sugar-containing food products. J Food Technol 1:73–82. Williams ML, Landel RF, Ferry JD. 1955. The temperature dependence of relaxation mechanisms in amorphous polymers and other glass-forming liquids. J Chem Eng 77:3701–7. Young RJ, Lovell PA. 1991. Introduction to polymers. 2nd ed. London: Chapman & Hall.

8 Antiplasticization of Food Polymer Systems by Low Molecular Mass Diluents C. C. Seow

Abstract Low molecular mass (LMM) compounds or diluents, acting as external plasticizers, are generally used to enhance the workability, flexibility, ductility, and/or distensibility of synthetic glassy polymers. However, at low concentrations, the presence of many of such diluents may instead cause a polymer-diluent blend to become stiffer or more brittle than the neat polymer—a phenomenon termed antiplasticization—even though the glass transition temperature (Tg) is depressed. Such a phenomenon appears not to be confined to synthetic polymers. Most food and other biomaterials may be regarded as partially amorphous, partially crystalline metastable polymeric systems, that display remarkable fundamental and generic similarities to synthetic polymers. Thus, LMM diluents (notably water, sugars, and polyols), which normally act as plasticizers, may exert effects on the mechanical properties of reduced-moisture food systems similarly in ways that reflect antiplasticization rather than plasticization. LMM diluents are therefore neither intrinsically plasticizers nor antiplasticizers. Besides alterations in properties of the polymer-diluent blend, abrupt changes in properties of the diluent itself may be indirect reflections of mechanical antiplasticization of the polymer. Although the physical manifestation of antiplasticization by LMM diluents is undisputed, the actual mechanisms involved remain unresolved. This review presents an overview of this important fundamental phenomenon that can simultaneously and profoundly influence various physical and functional properties, and hence quality and acceptability, of certain food products. Attention is focused on water (which is undeniably the most important diluent in foods) and glycerol (a commonly used humectant), and their interactive plasticizing-antiplasticizing effects.

Introduction Interactions of water molecules with one another and with other food components are yet to be fully understood despite extensive study and refinements in analytic techniques. In particular, that fraction of water close to biomacromolecular surfaces, generally referred to as bound water, in earlier food science studies, still remains somewhat of an enigma. The condition and properties of water closely associated with macromolecules clearly are detectably different from those of bulk water over definite time frames. Hence the classic differentiation between bound and free water based on sharp changes or discontinuities in some such properties (e.g., unfreezable water and nonsolvent water), whether or not true water binding actually prevails (Franks 1983, 1986; Slade and Levine 1991). 115

116

PART 1: Invited Speakers and Oral Presentations

The food polymer science approach to the study of water relationships in foods has, over the past 30 years or so, provided new perspectives to our basic understanding of structure-property relationships, thereby enhancing our ability to manipulate food properties to our advantage. The basic premise of food polymer science rests on the fundamental and generic similarities between synthetic polymers and food molecules. Most food materials may be regarded as partially amorphous, partially crystalline metastable polymeric systems, with water presumably acting as a ubiquitous or universal plasticizer (Levine and Slade 1988; Slade and Levine 1991, 1995). The primary effect of a plasticizer is to increase the workability, flexibility, ductility, and/or distensibility of a polymer (Sears and Darby 1982). However, initial humidification of certain partially crystalline food polymeric systems from the dry state is known, in many cases, to instead increase order, compactness, rigidity, modulus, and toughness. Further hydration beyond a critical limit decreases the same properties. Seow and others (1995, 1999) have suggested that such maxima in certain mechanical properties of such food polymer systems observed over low ranges of moisture content may be attributed to external mechanical antiplasticization by water, a phenomenon commonly observed in many synthetic glassy polymers where increased rigidity rather than flexibility results from the presence of very small amounts of added plasticizers. In some cases, concomitant decreases or minima in certain related properties may also result from antiplasticization, such as elongation at break and permeability to gases. Besides water, other low molecular mass (LMM) nonelectrolyte solutes, when present at low concentrations, may act as antiplasticizers in reduced-moisture food systems. Antiplasticization is essentially a polymer science–based interpretation to explain so-called anomalous mechanical behavior of glassy food systems in the presence of low concentrations of water (and other small molecules). The present review provides an updated overview of this fundamental phenomenon that not only influences the physical and textural, and hence eating properties of food systems at low- to intermediate-moisture levels, but may play an important role in the preservation of foods, biological tissues, antibodies, proteins, and drugs. Studies of the plasticizingantiplasticizing phenomena in foods and other biomaterials have undoubtedly enhanced our basic understanding of food polymer-diluent interactions.

Polymer-Diluent Interactions: Plasticization versus Antiplasticization Before extensive consideration of the antiplasticization effects of water (and other LMM diluents) in food polymer systems, it would be useful to recall briefly the types of relaxations that can occur within the glassy state. Such relaxations apparently exert profound influences on the physical properties of polymeric systems. Polymers exhibit a multiplicity of relaxational processes in the glassy state. It is customary to designate the primary or highest temperature transition (other than the crystallite melting temperature [Tm]) as an α transition. The α-relaxation process is generally recognized as the main glass transition, occurring over a limited range of temperatures. This transforms a relatively rigid or brittle glass into a comparatively

Antiplasticization of Food Polymer Systems by Low Molecular Mass Diluents

117

pliant “rubber,” resulting in sharp changes in free volume, molecular mobility, and physicochemical and structural properties of the polymeric system (Roberts and White 1973; Slade and Levine 1991; Simatos and others 1995; Roos 1996). Below the normal glass transition temperature (Tg), amorphous or partially crystalline polymers also exhibit secondary relaxation regions that are normally given the notations β, γ, δ, etc. (in decreasing order of temperature of occurrence). These sub-Tg or secondary relaxation processes usually have amplitudes smaller than the α relaxation associated with major backbone-chain movements (i.e., the primary glass transition). Sub-Tg or secondary relaxations appear to be associated in some way with the brittle-ductile transition of synthetic amorphous or partially crystalline polymers (Roberts and White 1973; Wu 1992; Simatos and others 1995; Roos 1996). The existence of secondary relaxations is indicative of mobility at temperatures below the glass transition temperature (T < Tg). The physical reason or origin of a particular secondary relaxation, however, remains unclear. Secondary relaxation processes have been reported for certain food materials (Kalichevsky and others 1992; Le Meste and others 1992; Noel and others 1992, 1996, 2000; Gidley and others 1993; Lourdin and others 1998), but their relationships to mechanical properties of these materials have been increasingly explored and acknowledged (Le Meste and others 1999, 2002). The addition of a LMM compound or diluent to a glassy polymer generally increases free volume and segmental mobility, which leads to a lowering of the Tg (Sears and Darby 1982). Selective interpolymer bonds may be weakened or broken, thereby leading to a reduction in melt viscosity or elastic modulus, which is the classic macroscopic effect of a plasticizer on physical properties of a polymer. The amorphous regions become swollen, and the whole mass becomes softer. If the solvent power of the added diluent (plasticizer) is great enough, crystallites, if present, may disappear, resulting in a very soft gel or a viscous liquid. In many cases, however, the incorporation of small amounts of many types of plasticizers in polymers has been reported to produce effects opposite to those expected (Jackson and Caldwell 1967a, 1967b; Sears and Darby 1982). The polymer-plasticizer system, in the glassy state at temperatures sometimes well below its Tg, becomes harder and less flexible than the neat polymer, even though a decrease in Tg may be evident. This has been termed antiplasticization—it occurs over a range of diluent concentrations or below a plasticization threshold that must be exceeded before the conventional external plasticizing effects on physical properties can be manifested (Sears and Darby 1982). Figure 8.1 shows schematically the difference between an idealized modulus or tensile strength vs diluent concentration curve and one that exhibits antiplasticization. Sears and Darby (1982) pointed out that, in general, the more polar the diluent is, the broader is the antiplasticization range (or plasticization threshold). In general, antiplasticization has been hypothesized to involve a combination of several factors, but the actual mechanisms involved in mechanical antiplasticization are not exactly known. Various theories and models of antiplasticization in synthetic polymer systems have been proposed including the following:

118

PART 1: Invited Speakers and Oral Presentations

(a) IDEALIZED

(b) ANTIPLASTICIZED

Glassy (Log) modulus

Antiplasticization range

Rubbery

0

0 Diluent concentration

Figure 8.1. Representations of (a) an idealized and (b) an antiplasticized modulus versus diluent content curve of solid polymer-diluent blends. From Seow and others (1999).

1. Increased crystallinity induced by an increase in free volume (Sears and Darby 1982; Guerrero 1989). 2. “Hole filling” by the diluent, causing a decrease in free volume and a suppression of motion, particularly at the polymer-chain ends (Maeda and Paul 1987; Vrentas and others 1988; Anderson and others 1995). In the more specific case of polymerwater interactions, the increase in dynamic tensile modulus on humidification of nylon films, while the systems were still in the glassy state, has been ascribed to free-volume reduction (Prevorsek and others 1971). 3. Formation of mechanically stable bridges between water and amide groups postulated to explain antiplasticization by water in nylon at T < Tα (Starkweather 1980). 4. Relaxation of low-density–high-energy regions in a glass, referred to as “islands of mobility” by Johari (1985), by a low-Tg diluent, which enables the rearrangement of the polymer chains and causes densification of the glass (Liu and others 1990). 5. β-Suppression effect leading to the disappearance or decrease in intensity or amplitude of the β transition and a concomitant reduction in free volume and an increase in modulus of the polymer in the temperature region between Tβ and Tα (or Tg) (Butzbach and Wendorff 1991; Ngai and others 1991). The concept of antiplasticization should be strictly applicable only at low diluent concentrations (i.e., in the context of a polymer-rich, rather than a diluent-rich, matrix)

Antiplasticization of Food Polymer Systems by Low Molecular Mass Diluents

119

and restricted to observable changes in the glass, determined at T < Tg. Note that food polymer systems containing high concentrations of water (above Wg′, the maximum content of plasticizing water rendered unfreezable), such as food gels, would have their Tg’s depressed to subzero temperatures (Slade and Levine 1991). Therefore, such systems would normally be in the leathery or rubbery state at ambient temperatures. Any occurrence of maximum hardness or modulus in such systems over the intermediate-moisture to high-moisture or water activity (aw) range, as has been reported (Kapsalis and others 1970; Hsieh and others 1990), may be attributed to increased crystallinity through association of polymer chains (such as starch retrogradation) induced or facilitated by enhanced macromolecular mobility in the rubbery state at intermediate moisture contents, prior to dilution at very much higher hydration levels. This should not be confused with antiplasticization effects. The present review does not delve into the question of whether, from a conceptual viewpoint, antiplasticization of food polymer systems could occur at Tg < T < Tm (i.e., in the rubbery state).

Water, the Ubiquitous Diluent, as Antiplasticizer at T < Tg A reduction in moisture content of a food polymer system would cause an increase in Tg. The Tg of reduced-moisture food systems is generally very much higher than the range of ambient temperature (T) over which they are normally stored and their properties determined. As such, most low-moisture foods at ambient temperatures are glassy or brittle solids. Even so, the glassy state (T < Tg) in foods, as in synthetic polymeric systems, need not necessarily be homogeneous in terms of mechanical and related properties (e.g., dielectric relaxation and gas barrier properties). These may be modified by the presence of diluents (such as water and sugars) acting as external plasticizers or antiplasticizers. Concomitant changes in the physicochemical properties of the diluents are also expected. Properties of the Food Polymer-Water Blend Mechanical Properties The rheological and textural properties of solid foods are inextricably linked. We would expect that addition of water should induce plasticization of polymeric chains and generally facilitate deformation, thereby decreasing hardness and crispness while increasing softness and extensibility. However, many published reports, covering a wide variety of food systems, have indicated that this has not necessarily always been the case, and that maxima or minima in mechanical properties do occur over the lowto-intermediate moisture (or aw) range. For water-compatible glassy polymers, the first-sorbed and earlier-sorbed water molecules appear more likely to cause a mechanical antiplasticizing effect independent of their kinetic effect in lowering the Tg of the polymer-water blend below the Tg of the neat polymer. For complex glassy food materials, moisture-induced antiplasticization typically increases hardness, stiffness, or toughness (usually referred to as moisture toughening) before the expected softening as moisture content is further elevated beyond a critical point where the material transforms from the glassy to the rubbery state (Kapsalis and others 1970; Reidy and

120

PART 1: Invited Speakers and Oral Presentations

Heldman 1972; Katz and Labuza 1981; Bourne 1986; Halek and others 1989; Harris and Peleg 1996; Le Meste and others 1996; Fontanet and others 1997; Li and others 1998; Roudaut and others 1998; Suwonsichon and Peleg 1998; Van Hecke and others 1998; Chang and others 2000a; Moraru and others 2002; Braga and Cunha 2004; Gondek and Lewicki 2006; Mandala and others 2006; Marzec and Lewicki 2006; Pittia and others 2007). Table 8.1 lists certain food products and mechanical properties antiplasticized by water, as well as the relevant references, even though some of the researchers may not have interpreted the effects observed as moisture-induced antiplasticization. Several factors (such as the type of food, the mechanical parameter measured, and the testing method employed) may affect the physical manifestation of the antiplasticization-plasticization phenomena, thus resulting in disparate responses of different mechanical properties of particular materials to moisture sorption. The following points regarding antiplasticization by water in food polymer-water blends are noteworthy: 1. The exact moisture content or aw location of maxima or minima in mechanical properties shows variations, not only with mechanical properties studied, but also with the food product and evaluation method (Table 8.1). 2. A high-density material may be more susceptible to antiplasticization than is a low-density one (Halek and others 1989). 3. Variations in mechanical behavior in response to moisture sorption may arise as a result of different mechanical testing methods employed. For example, Chang and others (2000a) observed that water exerted plasticizing effects on all mechanical parameters measured by the three-point bend test, but clearly caused antiplasticization of fracture stress and strain determined by uniaxial compression, suggesting that the latter method might be more sensitive in detecting the more subtle changes in fracture behavior arising from moisture-induced antiplasticization in a cellular, brittle glassy product such as dried bread. 4. The antiplasticization range or plasticization threshold depends on the type of system studied and on the physical property measured. It may extend over a relatively wide range of moisture content or aw and, in some cases, possibly beyond the glassy state into the rubbery domain. Some properties are more or less sensitive than others to antiplasticization, thereby altering the magnitude of the effect observed in a given system. For example, the antiplasticization range in extruded flat breads stretched over the range of aw from 0 to 0.8 (Marzec and Lewicki 2006) (Figure 8.2). Similarly, water apparently induced an antiplasticizing effect on fracture force and energy of raw and roasted coffee beans over a range of aw from 0.3 to 0.8 (Pittia and others 2007). 5. Certain mechanical parameters may be antiplasticized on humidification from the dry state, whereas others, determined using the same test method on the same material, may be plasticized immediately or remain practically unaffected until the glass transforms into rubber. Each mechanical property apparently has its own specific moisture dependency, as pointed out by Borges and Peleg (1997).

Table 8.1. Complex food products wherein water acts as an antiplasticizer on mechanical properties Food product

Property antiplasticized

Water activity or moisture content of antiplasticization maximum or minimum

Reference

Precooked, freeze-dried beef

Modulus, toughness

0.2 aw (max)

Kapsalis and others (1970)

Precooked, freeze-dried beef

Hardness, chewiness

0.4 aw (max)

Reidy and Heldman (1972)

Snack-food products

Compression work

0.2–0.4 aw (max)

Katz and Labuza (1981)

Air-dried apple

• Hardness • Springiness

0.11 aw (max) 0.33 aw (max)

Bourne (1986)

Cornmeal extrudates

Compressive strength

8.9%–15.3% moisture (max)

Halek and others (1989)

Brittle cereal foods

Compressive force • Cheese balls • French bread croutons

0.4–0.5 aw (max) ∼0.6 aw (max)

Harris and Peleg (1996)

Extruded flat breads

• Compressive fracture stress • Young’s modulus

∼9% moisture (max)

Fontanet and others (1997)

Crispy breads

• Stiffness modulus and fracture stress • Acoustic emission intensity

11% moisture (max)

Roudaut and others (1998)

Puffed cereals

Extruded corn-based snacks Corn cakes

Compressive force • Cheese balls • Cocoa puffs • Puncturing force • Specific force of Structural ruptures • Compressive peak force • Work required to break cake

9% moisture(max)

7%–8% moisture 6%–7% moisture 0.56 aw (max) 0.35 aw (max)

Suwonsichon and Peleg (1998) Van Hecke and others (1998)

9%–10% moisture (∼0.6 aw)

Li and others (1998)

Dried bread

• Compressive fracture stress • Compressive fracture strain

0.32 aw (max) 0.32 aw (min)

Chang and others (2000a)

Extruded meatstarch product

Storage modulus

∼0.2 aw (max)

Moraru and others (2002)

Extruded flat breads (wheat and rye)

• Compressive force • Compression work • Breaking force

0.53–0.59 aw (max)

Marzec and Lewicki (2006)

Corn and wheatbran flakes

• Compressive force • Compression work

∼0.65 aw (max)

Gondek and Lewicki (2006)

“Petite beurre” biscuits

Puncture force

0.32 aw (max)

Mandala and others (2006)

Raw and roasted coffee beans

Compression fracture force, energy, and strain • Raw coffee beans • Roasted beans

∼0.50 aw (max)

Pittia and others (2007)

∼0.75 aw (max) ∼0.84 aw (max)

aw, water activity; max, maximum; min, minimum.

121

122

PART 1: Invited Speakers and Oral Presentations

Figure 8.2. Compression-force–water-activity relationships for extruded flat breads showing a wide antiplasticization range. N, Newton. Redrawn from Marzec and Lewicki (2006).

6. It has been demonstrated that, in certain puffed cereal products, moisture hardening or toughening, detected instrumentally, can be perceived sensorily (Roudaut and others 1998; Suwonsichon and Peleg 1998). The so-called anomalous antiplasticization effects of water in relation to some mechanical properties have also been clearly demonstrated in biopolymer-based films, including wheat gluten (Gontard and others 1993), tapioca starch (Chang and others 2000b, 2006), konjac glucomannan (KGM) (Cheng and others 2002, 2006, 2007), poly(lactide-co-glycolide) (Blasi and others 2005), ethylene vinyl alcohol, copolymeric food packaging (Cabedo and others 2006), as well as pullulan and caseinate films and their composites (Kristo and others 2007). Gontard and others (1993) have reported that, during hydration of wheat-gluten films, the first fraction of water molecules to be sorbed improved film elasticity and puncture resistance, probably due to the formation of supplementary hydrogen bonds between protein chains. The later-sorbed water fraction made it easier to break such bonds (plasticization), and the behavior of the gluten film changed from elastic to viscous. Maximum puncture resistance of the films occurred at a progressively lower moisture content as temperature was increased. This is to be expected since elevating the temperature adds free volume to the system, often referred to as plastication. More detailed studies of tapioca-starch films in the glassy state conducted by Chang and others (2000b, 2006) confirmed that water plays a role either as a plasticizer or an antiplasticizer, depending on the physical property measured. Whereas Tg and tensile modulus were evidently plasticized on humidification of the films from the dry state, maxima in tensile strength, strain at break, and toughness were observed over an intermediate moisture content or aw range.

Antiplasticization of Food Polymer Systems by Low Molecular Mass Diluents

123

Figure 8.3. Changes in (a) tensile strength and (b) strain at break of tapioca starch (TS), konjac glucomannan (KGM), and konjac glucomannan–carboxymethyl cellulose (KGM-CMC) films as a function of water activity (aw). Plotted from data of Chang and others (2000b) and Cheng and others (2002).

Of particular interest is that the same mechanical property of different biopolymer systems can respond differently to different levels of hydration. Figure 8.3 compares changes in (a) tensile strength and (b) tensile elongation as a function of aw for different biopolymer films (with data obtained from Chang and others 2000b; Cheng and others 2002). Minima in tensile elongation (normally taken as symptomatic of antiplasticization) were evident in KGM and KGM–carboxymethyl cellulose films, whereas a maximum in elongation was found in the case of tapioca-starch films. Attenburrow and others (1992) and Nicholls and others (1995) noted the occurrence of maxima in flexural fracture stress and strain of amorphous starch systems at 8%– 14% moisture, whereas no such peaks were exhibited by glassy gluten systems over the same range of moisture content studied (Figure 8.4). These findings reinforce the view that composition greatly influences the mechanical response of glassy biomaterials to moisture sorption. However, a cogent explanation for the aforementioned discrepancies has yet to be forthcoming. The antiplasticizing effects of water on other biomaterials have been similarly observed. In the 1970s, Hiltner and coworkers (Nomura and others 1977; Hiltner and others 1978) demonstrated that the first 16%–25% water was able to act as an antiplasticizer by increasing the rigidity of poly-L-hydroxyproline and native collagen to a maximum value. In a study on hot-pressed pullulan-starch blends containing 10%

124

PART 1: Invited Speakers and Oral Presentations

150

(a)

5

(b)

WHEAT STARCH WHEAT STARCH

100

Fracture strain (%)

Fracture stress (MPa)

4

50 GLUTEN

0

0

5

10

15

Moisture content (%)

3

2 GLUTEN

1

0

0

5

10

15

20

Moisture content (%)

Figure 8.4. Changes in (a) fracture stress and (b) fracture strain for gluten and starch as a function of moisture content. From Attenburrow and others (1992).

sorbitol or xylose, it was apparent that flexural or tensile stress was antiplasticized by water whereas Young’s modulus was plasticized (Biliaderis and others 1999). Dynamic mechanical thermal analysis (DMTA) data obtained by Pommet and others (2005) suggested that both water and 1,4-butanediol had antiplasticizing effects on the glassy storage modulus of wheat-gluten thermoplastic materials, although statistical significance of the data was not established. Thermal Properties Thus far, we have focused our attention on the antiplasticizing effects of water on macroscopic mechanical properties of the systems. We believe that such antiplasticizing effects may also lead to manifestations where thermal properties are concerned. For example, could the presence of the small endothermic event, in the range 50°– 70°C, typically observed on heating low-moisture (5%–25% wt/wt water) polysaccharide systems (Gidley and others 1993), be viewed as antiplasticization by water? Whereas the temperature of the endotherm exhibits no moisture dependency, the enthalpy shows a systematic increase with moisture content. Gidley and others (1993) suggested that this endothermic event might be due to the “cooperative thermal disruption of enthalpically-favourable interactions in the hydrogen-bonded network of solvent and solute in systems” (p. 308) where “polysaccharide chains are effectively immobilized either kinetically (below the glass transition) or thermodynamically (below the temperature of local order loss)” (p. 311). The intriguing question arises as to whether this thermal phenomenon is in any way related to the macroscopic

Antiplasticization of Food Polymer Systems by Low Molecular Mass Diluents

450

(a)

400

400

350

350 Melting enthalpy (J g–1)

Melting enthalpy (J g–1)

450

300 250 200 150 100 50 0 0

0% Glyc 10% Glyc 20% Glyc

10 20 30 40 Moisture content (% db)

250 200 150

50 50

(b)

300

100

30% Glyc 40% Glyc 50% Glyc

125

0 0

0% Sorb 10% Sorb 20% Sorb 30% Sorb 40% Sorb 50% Sorb

10 20 30 40 Moisture content (% db)

50

Figure 8.5. Melting-enthalpy–moisture-content relationships of konjac glucomannan films plasticized with (a) glycerol (glyc) and (b) sorbitol (sorb). Broken lines indicate the critical moisture content level. db, dry basis. Redrawn from Cheng and others (2006).

manifestations of antiplasticization by water commonly associated with such polysaccharide systems. Should sub-Tg rotational motions of the main chains of polymers be affected, it is to be expected that significant changes in mechanical properties (such as in brittle-ductile transitions) would be observed (Wu 1992; Le Meste and others 1999). Cheng and others (2006) investigated the thermal properties of KGM films, but restricted their attention to melting enthalpy. They discovered that the melting enthalpy (ΔH) (determined by differential scanning calorimetry) of KGM films, containing from 0 to 30% glycerol (weight of glycerol to weight of dry KGM), increased with increasing moisture content up to a peak or critical point before decreasing (Figure 8.5a). The ΔH peak occurred at progressively higher moisture contents (from 18% to 26%) as the glycerol content was increased from 0 to 30%, possibly reflective of the lowering of Tg and enhanced chain mobility with increasing additions of glycerol. The ΔH peak disappeared as glycerol content was further increased to 40% and 50%. Thus, at any hydration level below a critical level of ∼20% (dry basis), successive addition of glycerol progressively lowered ΔH, whereas the opposite was true at moisture levels higher than the critical level. In contrast to glycerol-containing films, sorbitolcontaining films exhibited no ΔH peak over the range of moisture contents studied (Figure 8.5b). However, the opposing effects of increasing amounts of sorbitol

126

PART 1: Invited Speakers and Oral Presentations

above and below a critical moisture content were still very much evident. The mechanisms behind such opposing influences of glycerol and sorbitol on thermal properties of cast KGM films remain unclear. It is possible that such polyols, depending on circumstances, could weaken or strengthen the long-range order of partially crystalline KGM films. Gas Transport Properties Gas transport behavior is an important practical consideration particularly relevant to biopolymer-based packaging films and coatings. Could mechanical antiplasticization of such materials by certain diluents affect their permeability to gases, such as carbon dioxide and oxygen? Antiplasticization of some synthetic polymers by certain diluents was accompanied by substantial reductions in permeability to gases (e.g., helium, carbon dioxide, and methane), consistent with reduced mobility in the glass (Sefcik and others 1983; Maeda and Paul 1987). Evidence of this in relation to biomaterials has more recently come to light. Benczédi and others (1998) provided evidence that, in amorphous potato-starch extrudates, water acted as an antiplasticizer, with the density of the starch passing maximum value at a low water concentration. It was suggested that antiplasticization may reduce gas-sorption and permeation rates and that such knowledge can be used to optimize gas barrier properties of polymers. Charmathy and others (2006) observed that low amounts of water (at a relative humidity equivalent to ∼5%) impeded its own permeation, as well as that of octane, through microcrystalline cellulose (MCC) before higher humidification began to facilitate permeation (Figure 8.6). Antiplasticizing effects of water were similarly observed for tensile strength and Young’s modulus (determined by three point-bending tests) of compacted MCC samples whereas crystallinity index (determined by Fourier transform infrared spectroscopy [FTIR]) increased with moisture content to a plateau (Figure 8.6b). Thus, in line with the theory of antiplasticization expounded by Guerrero (1989), those authors opined that small amounts of water increased free volume, thereby stimulating molecular rearrangement in the amorphous regions. The subsequent increase in crystallites translated into restricted mobility of the polymer, thereby leading to antiplasticization in mechanical properties and permeability to gases. The difference in plasticizer threshold observed was attributed to the different timescales of the properties measured, a macroscopic property such as tensile strength giving a higher or broader threshold than a microscopic property such as gas permeability. Considering that MCC is so commonly used in the pharmaceutical industry, especially for tablet manufacturing, it may be of some concern should its mechanical performance be altered significantly by moisture uptake during handling and storage. Properties of the Diluent (Water) Conceptually, antiplasticization refers strictly to the mechanical properties of a glassy polymer-diluent blend. These properties would be affected predominantly by the mobility of the polymer molecules in polymer-rich systems. Interactions between the polymer and the diluent must inevitably affect the mobility and physicochemical properties of the diluent itself. Thus, abrupt changes in properties of the diluent that result from such interactions may indirectly reflect mechanical antiplasticization of

Antiplasticization of Food Polymer Systems by Low Molecular Mass Diluents

127

5

4 0.05 3 Permeability of octane Permeability of water

2 0

5

10

15 20 25 Relative humidity %

30

35

(b) 700

3.0

500

2.5

400

2.0 Young’s modulus

300

Tensile strength Crystallinity index

200 0

20

40 60 Relative humidity %

80

0.94 Tensile strength (MPa)

Young’s modulus (MPa)

0.96

3.5

600

Permeability (Dp) of octame (cm2 s–1)

0.06

0.92 0.90 0.88 0.86 0.84

Crystallinity index

Permeability (Dp) of water × 1012 (cm2 s–1)

(a) 6

0.82 1.5 1.0 100

0.80 0.78

Figure 8.6. Moisture-induced antiplasticization of microcrystalline cellulose: (a) permeability of water and octane and (b) tensile strength, Young’s modulus, and crystallinity. Dp, bulk diffusion permeability. From Charmathy and others (2006).

the polymer. As pointed out earlier, such discontinuities in the physicochemical properties of water that are closely associated with macromolecules have given rise to the various classical definitions of bound water (e.g., unfreezable water and nonsolvent water). Would it be reasonable to equate the antiplasticization range of moisture content to bound water?

128

PART 1: Invited Speakers and Oral Presentations

In certain cases, abrupt changes in mechanical properties of the system and physicochemical properties of the diluent may show good correlation, particularly in terms of critical moisture contents at which such changes occur. Such was the case for freeze-dried beef, where maxima in negative entropy (−ΔS0) of water sorption and certain mechanical properties have been reported to occur at about the same critical moisture content (Kapsalis and others 1970). However, information on published correlations of this kind is scarce. Also note that, in many cases, because of different timescales in the experimental methods, changes in properties of the system need not exactly parallel each other. Thermodynamics of Water Sorption Complex foods have been treated as binary systems (i.e., solids and water) to facilitate thermodynamic analyses of water-sorption isotherms at different temperatures. Figure 8.7 shows the calculated isosteric heat of sorption (Qs)–moisture content plots of several food materials (Duckworth 1972). Any excess enthalpy over the heat of vaporization of free water was assumed to be indicative of a water-binding effect (Berlin and others 1970; Bettelheim and others 1970; Leung and Steinberg 1979). Simplistically, there is thus a division between two different fractions of water molecules. On a more fundamental basis, note that water-sorption studies are not reflective of true thermodynamic equilibrium situations (Kuntz and Kauzmann 1974; Levine and Slade 1988). Nevertheless, such an approach, despite its fundamental limitations and sizeable errors in the calculations, may be useful (practically) in the drying of foods. Quite substantial

Heat of binding of water (kcal mol–1)

Agar Gelatin

18

Starch Cellulose

16 14 12 10 Heat of condensation 0

0

10 20 Moisture content (% db)

30

40

Figure 8.7. Isosteric heat of sorption (Qs) versus moisture-content plots of food colloids exhibiting a maximum Qs value. db, dry basis. From Duckworth (1972).

Antiplasticization of Food Polymer Systems by Low Molecular Mass Diluents

129

energy, over and above that for the evaporation of bulk water, is required to desorb the last traces of water. It is interesting to note that many food materials have been reported to exhibit maxima in Qs and minima in entropy (ΔS0) of sorption at some low moisture level. This would suggest a most stable pseudo-thermodynamic state, resulting from strong binding of adsorbate to the adsorbent surface at that hydration level (Berlin and others 1970; Bettelheim and others 1970; Duckworth 1972; Leung and Steinberg 1979; Rizvi and Benado 1984). Also, water in propinquity to macromolecular surfaces has been suggested to be more structured than that in the bulk (Etzler 1991). The existence of maximum enthalpy and minimum entropy of sorption, and hence possibly of mechanical antiplasticization, would depend on the nature of the polymer substrate and its ability to interact with water molecules. Such effects are unlikely to be manifested to any great extent when polymer-diluent interactions are weak. Water Diffusivity Seow and others (1999) hypothesized that mechanical antiplasticization of biopolymers by water may be accompanied by parallel changes in diffusivity of the diluent. The effective diffusivity (Deff) of water in starch-based systems has been observed to exhibit maximum values over the low-moisture range, although porosity of the materials was found to increase linearly with a decrease in moisture content (Marousis and others 1989; Leslie and others 1991; Kostaropoulos and Saravacos 1997). Qualitatively, the form of the Deff–moisture-content curve, as shown in Figure 8.8, resembles that of the modulus–moisture-content curve of an antiplasticized system (Figure 8.1b). This pattern is apparently neither affected by the method of sorption (adsorption or desorption) nor by the incorporation of sugars (Leslie and others 1991), but the nonlinearity was apparently suppressed at lower porosities (Marousis and others 1989; Kostaropoulos and Saravacos 1997). Deff values for any specific material have been, however, generally lower in the presence of sugars (Marousis and others 1989; Leslie and others 1991) and when determined from water adsorption, rather than desorption measurements (Leslie and others 1991). Leslie and others (1991) suggested that the observed maxima in Deff in the moisture-content region below 0.2 g/g dry solids, where vapor-phase diffusion predominates, were due to strong binding of water molecules by high-affinity binding sites of the macromolecules. They hypothesized that “as the material begins to be rather dry (moisture content, <10%), the energy required to transfer the ‘strongly held’ water to the vapor phase increases, resulting in Deff falling sharply despite a now very porous, open structure” (p. 387). Kostaropoulos and Saravacos (1997), who found similar effects of moisture on thermal diffusivity and water diffusivity, attributed these “anomalies” to the state of water in the porous structure of the low-moisture systems. Although no experimental data are available, food materials over this low range of moisture content should obviously be highly susceptible to mechanical antiplasticization. As pointed out by Leslie and others (1991), physicochemical interactions between macromolecules and water are also likely to be involved in the overall mass transport mechanisms.

130

PART 1: Invited Speakers and Oral Presentations

40

Effective diffusivity, Deff × 10–10 (m2 s–1)

Hydrated hylon 7

30

Hylon 7 gel Hydrated amioca Amioca gel

20

10

0 0

1 Moisture content (% db)

2

Figure 8.8. Comparison of effective diffusivities for granular hydrates and gelatinized amioca and hylon 7, obtained from drying-curve data at 60°C. db, dry basis. From Leslie and others (1991).

Nonwater Diluents as Antiplasticizers at T < Tg Obviously, actual food systems seldom, if ever, contain water as the sole diluent. A complex mixture of water and other LMM compounds (e.g., sugars, polyols, amino acids, fatty acids, and acylglycerols) is usually present. These LMM diluents interact with each other and with different macromolecules, such interactions giving rise to effects that are extremely difficult to predict. Explaining such interactions and their influence on macroscopic properties that affect the quality and acceptability of foods is complex and difficult. Water and other LMM diluents, when present at low concentrations in food systems, are neither intrinsically plasticizers nor antiplasticizers. They may serve as antiplasticizers or plasticizers, depending on prevailing conditions. Some earlier findings have been likely caused by the antiplasticization effects of such nonwater diluents. For example, Shogren and others (1992) reported that starch extrudates rapidly became very brittle and shattered like glass at low levels of urea and glycols (e.g., glucose and glycerol). Fructose was reported to function as an antiplasticizer by increasing the stiffness (based on measurements of storage modulus [E′]) of amylopectin, but also to act as a plasticizer by lowering its Tg (Peleg 1996). Lourdin and others (1997) reported that a glycerol-containing potato-starch film, at ambient temperature (25°C), exhibited minimum elongation at break at a glycerol content of 12%, although Tg was found to

Antiplasticization of Food Polymer Systems by Low Molecular Mass Diluents

131

decrease progressively as glycerol content increased from 0 to 25%, with no change in crystallinity observed by X-ray diffraction. Supportive evidence for an antiplasticization effect was provided by DMTA. The sample containing 12% glycerol was revealed to exhibit a weak and broad β-relaxation peak as compared with the rather intense peaks formed at other glycerol concentrations. Note that the Tg of a film with a glycerol content of 15% was estimated to be close to ambient temperature. As such, a film containing 12% glycerol would be in the glassy state at ambient temperature. More detailed studies have since provided additional evidence of the ability of sugars and polyols to act as antiplasticizers in a variety of biopolymer systems. Chang and others (2006) observed that the tensile modulus of glassy tapioca-starch films attained a maximum value when glycerol was present at a low concentration (∼2.5%, dry-starch basis) at aw ≤ 0.22 (equivalent to a moisture content of 7%, dry-starch basis) (Figure 8.9).

Figure 8.9. Tensile modulus of tapioca-starch films at different water activity (aw) as a function of glycerol content. db, dry basis; and E, tensile modulus. From Chang and others (2006).

132

PART 1: Invited Speakers and Oral Presentations

On the other hand, glycerol (up to a concentration of 20%) had little or no effect on strain at break of the completely dry films, but exhibited minimum values at 0.11 and 0.22 aw. Glycerol reverted to its role as a typical plasticizer at aw ≥ 0.32. The plasticizer threshold or glycerol antiplasticization range for tensile modulus was considerably narrower than that for strain at break. In contrast, antiplasticizing effects of glycerol, if any, on tensile strength were not obvious. The tensile strength of starchbased films at low moisture contents were also found to be slightly increased by sorbitol (Gaudin and others 1999) and maltose (Follain and others 2006). Shimazu and others (2007) have since provided additional evidence to support the contention that glycerol and sorbitol can act as antiplasticizers of cast cassava films when used at low levels (≤15 g/100 g starch) at low aw (≤0.58). Da Róz and others (2006) observed that certain plasticizers (notably 1,4-butanediol, diethylene glycol, and sorbitol) influenced the mechanical properties of thermoplastic starch materials in opposite ways, viz. softening due to plasticization of the amorphous phase and stiffening due to antiplasticization. In terms of other properties, Gaudin and others (2000) found that the addition of sorbitol at concentrations up to 21% (wt%) to starch-water films caused a decrease in oxygen permeability. Above 21%, sorbitol exerted a plasticizing effect by increasing oxygen permeability. Those researchers were of the opinion that the antiplasticizing effects observed were related to a suppression of secondary relaxations because of the interactions between starch and sorbitol. The findings by Cheng and others (2006) on the effects of water-glycerol and water-sorbitol interactions on the thermal properties of konjac glucomannan films were discussed earlier (see Figure 8.5). An interesting aspect of the role of glycerol as an antiplasticizer relates to its enhancement of the bioprotective properties of sugar glasses, particularly those of the “special” sugar, trehalose, which apparently plays an important role in anhydrobiosis. Glass formation is an effective way of preserving and maintaining the activity of foods, biological tissues and agents, vaccines, organs, proteins, antibodies, and drugs by slowing molecular mobility and chemical reactivity. As an antiplasticizer, glycerol would prevent phase separation, thereby making the mixture more thermodynamically stable. Cicerone and others (2003) and Cicerone and Soles (2004), using incoherent neutron scattering, demonstrated that adding small quantities of a low-Tg diluent (such as glycerol) to a bioprotective glass (such as trehalose) lowered Tg but enhanced the stability of proteins (enzymes) sequestered within the glass. According to those authors, the Tg of the glass alone cannot predict the degree of stability that the glass can impart to the protein. Instead, they found that stiffening or suppression of short-length scale, high-frequency dynamics within the glass, caused by the addition of small amounts of diluent, led to increased protein stabilization. Using dielectric spectroscopy, Anopchenko and others (2006) have confirmed that the addition of small amounts of glycerol to glassy trehalose slows small-amplitude secondary relaxations of trehalose. Earlier, Lourdin and others (1998, 2003) had shown that glycerol has the ability to antiplasticize amorphous maltose-glycerol systems.

Antiplasticization of Food Polymer Systems by Low Molecular Mass Diluents

133

Conclusion The fact that many food polymer systems and other biomaterials can be mechanically antiplasticized by LMM diluents clearly illustrates the inhomogeneous and metastable nature of the glasses within such materials, at least over the scale and time frame of customary experimental methods. Unfortunately, there have been instances (reported especially in the early literature) where homogeneity in the mechanical properties of the glass has been assumed, or experimental data suggesting an antiplasticization effect have been overlooked or ignored, so that an idealized mechanical property vs moisture-content curve (such as Figure 8.1a) was portrayed instead. Fortunately, as the current review shows, this fundamental phenomenon is being recognized more often. Nevertheless, much scientific effort is required to explore further, in more explicit terms, the practical significance of antiplasticization by LMM diluents in complex food and biological systems and to elucidate the underlying mechanisms of antiplasticization by different LMM diluents. It is uncertain why some diluents are better antiplasticizers or plasticizers than are others. No universal hypothesis has been proposed that can account for the variation in effects exhibited by different diluents. Those diluents that lead to greater antiplasticizing effects may very likely be those that can cause substantial reductions in free volume through their ability to fill holes in the amorphous matrix and/or through their ability to interact with polymer molecules. The average size of the free-volume voids as compared with the average size of the diluent molecules would be an important consideration. Experimental evidence to support such ideas has yet to be provided regarding food polymer systems.

References Anderson SL, Grulke EA, Delassus PT, Smith PB, Kocher CW, Landes BG. 1995. A model for antiplasticization in polystyrene. Macromolecules 28:2944–54. Anopchenko A, Psurek T, VanderHart D, Douglas JF, Obrzut J. 2006. Dielectric study of the antiplasticization of trehalose by glycerol. Phys Rev [E] 74:031501–10. Attenburrow GE, Davies AP, Goodband RM, Ingman SJ. 1992. The fracture behaviour of starch and gluten in the glassy state. J Cereal Sci 16:1–12. Benczédi D, Tomka I, Escher F. 1998. Thermodynamics of amorphous starch-water systems. 1. Volume fluctuations. Macromolecules 31:3055–61. Berlin E, Klimann PG, Pallansch MJ. 1970. Changes in state of water in proteinaceous systems. J Colloid Interface Sci 34:488–94. Bettelheim FA, Block A, Kaufman LJ. 1970. Heats of water vapor sorption in swelling biopolymers. Biopolymers 9:1531–8. Biliaderis CG, Lazaridou A, Arvanitoyannis I. 1999. Glass transition and physical properties of polyolplasticised pullulan–starch blends at low moisture. Carbohydr Polym 40:29–47. Blasi P, D’Souza SS, Selmin F, DeLuca PP. 2005. Plasticizing effect of water on poly(lactide-co-glycolide). J Control Release 108:1–9. Borges A, Peleg M. 1997. Effect of water activity on the mechanical properties of selected legumes and nuts. J Sci Food Agric 75:463–71. Bourne MC. 1986. Effect of water activity on texture profile parameters of apple flesh. J Texture Stud 17:331–40.

134

PART 1: Invited Speakers and Oral Presentations

Braga ALM, Cunha RL. 2004. Plasticization and antiplasticization by small molecules in brittle cellular foods: TMDSC and mechanical properties. Int J Food Properties 7:105–20. Butzbach GD, Wendorff JH. 1991. Polycarbonate-poly(methyl methacrylate) blends: the role of molecular interactions on miscibility and antiplasticization. Polymer 32:1155–9. Cabedo L, Lagarón JM, Cava D, Saura JJ, Giménez E. 2006. The effect of ethylene content on the interaction between ethylene–vinyl alcohol copolymers and water. II. Influence of water sorption on the mechanical properties of EVOH copolymers. Polym Test 25:860–7. Chang YP, Abd Karim A, Seow CC. 2006. Interactive plasticizing-antiplasticizing effects of water and glycerol on the tensile properties of tapioca starch films. Food Hydrocolloids 20:1–8. Chang YP, Cheah PB, Seow CC. 2000a. Variations in flexural and compressive fracture behavior of a brittle cellular food (dried bread) in response to moisture sorption. J Texture Stud 31:525–40. Chang YP, Cheah PB, Seow CC. 2000b. Antiplasticization-plasticization effects of water on physical properties of tapioca starch films in the glassy state. J Food Sci 65:445–51. Charmathy SP, Carvajal T, Pinal R. 2006. Moisture-induced antiplasticization of microcrystalline cellulose. Paper presented at the IUFoST [International Union of Food Science & Technology] 13th World Congress of Food Science and Technology, Nantes, France, 2006. Cheng LH, Abd Karim A, Norziah MH, Seow CC. 2002. Modification of the microstructural and physical properties of konjac glucomannan-based films by alkali and sodium carboxymethylcellulose. Food Res Int 35:829–36. Cheng LH, Abd Karim A, Seow CC. 2006. Effects of water-glycerol and water-sorbitol interactions on the physical properties of konjac glucomannan films. J Food Sci 71:E62–7. Cheng LH, Abd Karim A, Seow CC. 2007. Effects of acid modification on physical properties of konjac glucomannan (KGM) films. Food Chem 103:994–1002. Cicerone MT, Soles CL. 2004. Fast dynamics and stabilization of proteins: binary glasses of trehalose and glycerol. Biophys J 86:3836–45. Cicerone MT, Tellington A, Trost L, Sokolov AP. 2003. Substantially improved stability of biological agents in dried form. Bioproc Int 1:36–47. Da Róz AL, Carvalho AJF, Gandini A, Curvelo AAS. 2006. The effect of plasticizers on thermoplastic starch compositions obtained by melt processing. Carbohydr Polym 63:417–24. Duckworth RB. 1972. The properties of water around the surfaces of food colloids. Proc Inst Food Sci Technol 5:60–7. Etzler FM. 1991. A comparison of the properties of vicinal water in silica, clays, wood, cellulose, and other polymeric materials. In: Levine H, Slade L, editors. Water relationships in foods. New York: Plenum. p 805–22. Follain N, Joly C, Dole P, Roge B, Mathlouthi M. 2006. Quaternary starch-based blends: influence of a fourth component addition to the starch/water/glycerol system. Carbohydr Polym 63:400–7. Fontanet S, Davidou S, Dacremont C, Le Meste M. 1997. Effect of water on the mechanical behaviour of extruded flat bread. J Cereal Sci 25:303–11. Franks F. 1983. Bound water: fact and fiction. Cryo Lett 4:73–4. Franks F. 1986. Unfrozen water: yes; unfreezable water: hardly; bound water: certainly not. Cryo Lett 7:207. Gaudin S, Lourdin D, Forssell PM, Colonna P. 2000. Antiplasticization and oxygen permeability of starchsorbitol films. Carbohydr Polym 43:33–7. Gaudin S, Lourdin D, Le Botlan D, Ilari JL, Colonna P. 1999. Plasticization and mobility in starch-sorbitol films. J Cereal Sci 29:273–84. Gidley MJ, Cooke D, Ward-Smith S. 1993. Low-moisture polysaccharide system: thermal and spectroscopic aspects. In: Blanshard JMV, Lillford P, editors. The glassy state in foods. Nottingham, UK: Nottingham University Press. p 303–16. Gondek E, Lewicki PP. 2006. Antiplasticization of cereal-based products by water. Part II: Breakfast cereals. J Food Eng 77:644–52. Gontard N, Guilbert S, Cuq JL. 1993. Water and glycerol as plasticizers affect mechanical and water vapor barrier properties of an edible wheat gluten film. J Food Sci 58:206–11.

Antiplasticization of Food Polymer Systems by Low Molecular Mass Diluents

135

Guerrero SJ. 1989. Antiplasticization and crystallinity in poly(vinyl chloride). Macromolecules 22:3480–5. Halek GW, Paik SW, Chang KLB. 1989. The effect of moisture content on mechanical properties and texture profile parameters of corn meal extrudates. J Texture Stud 20:43–55. Harris M, Peleg M. 1996. Patterns of textural changes in brittle cellular cereal foods caused by moisture sorption. Cereal Chem 73:225–31. Hiltner A, Shiraishi H, Nomura S, Baer E. 1978. Nature of biopolymer-water association by dynamic mechanical spectroscopy. In: Meier DJ, editor. Molecular basis of transitions and relaxations. New York: Gordon and Breach Science. p 249–59. Hsieh F, Hu L, Huff HE, Peng IC. 1990. Effects of water activity on textural characteristics of puffed rice cake. LWT Food Sci Technol 23:471–3. Jackson WJ, Caldwell JR. 1967a. Antiplasticization. II. Characteristics of antiplasticizers. J Appl Polym Sci 11:211–26. Jackson WJ, Caldwell JR. 1967b. Antiplasticization. III. Characteristics and properties of antiplasticizable polymers. J Appl Polym Sci 11:227–44. Johari GP. 1985. Structural changes and molecular mobility during physical ageing of glasses. J Phys Colloq 46 (C8):567–72. Cited in Liu and others (1990). Kalichevsky MT, Jaroszkiewicz EM, Ablett S, Blanshard JMV, Lillford PJ. 1992. The glass transition of amylopectin measured by DSC, DMTA and NMR. Carbohydr Polym 18:77–88. Kapsalis JG, Walker JE, Wolf M. 1970. A physicochemical study of the mechanical properties of low and intermediate moisture foods. J Texture Stud 1:464–83. Katz EE, Labuza TP. 1981. Effect of water activity on the sensory crispness and mechanical deformation of snack food products. J Food Sci 46:403–9. Kostaropoulos AE, Saravacos GD. 1997. Thermal diffusivity of granular and porous foods at low moisture content. J Food Eng 33:101–9. Kristo E, Biliaderis CG, Zampraka A. 2007. Water vapour barrier and tensile properties of composite caseinate-pullulan films: biopolymer composition effects and impact of beeswax lamination. Food Chem 101:753–64. Kuntz ID, Kauzmann W. 1974. Hydration of proteins and polypeptides. Adv Protein Chem 28:239– 345. Le Meste M, Champion D, Roudaut G, Contreras-Lopez E, Blond G, Simatos D. 1999. Mobility and reactivity in low-moisture and frozen food. In: Roos YH, Leslie RB, Lillford P, editors. Water management in the design and distribution of quality foods. Lancaster, PA: Technomic. p 267–84. Le Meste M, Champion D, Roudaut G, Simatos D. 2002. Glass transition and food technology: a critical appraisal. J Food Sci 67:2444–58. Le Meste M, Huang V, Panama J, Anderson G, Lentz R. 1992. Glass transition of bread. Cereal Foods World 37:264–7. Le Meste M, Roudaut G, Davidou S. 1996. Thermomechanical properties of glassy cereal foods. J Therm Anal 47:1361–75. Leslie RB, Carillo PJ, Chung TY, Gilbert SG, Hayakawa K, Marousis S, Saravacos GD, Solberg M. 1991. Water diffusivity in starch-based systems. In: Levine H, Slade L, editors. Water relationships in foods. New York: Plenum. p 365–90. Leung HK, Steinberg MP. 1979. Water binding of food constituents as determined by NMR, freezing, sorption and dehydration. J Food Sci 44:1212–6. Levine H, Slade L. 1988. Water as a plasticizer: physico-chemical aspects of low-moisture polymeric systems. In: Franks F, editor. Water science reviews 3. Cambridge, MA: Cambridge University Press. p 79–185. Li Y, Kloeppel KM, Hsieh F. 1998. Texture of glassy corn cakes as a function of moisture content. J Food Sci 63:869–72. Liu Y, Roy AK, Jones AA, Inglefield PT, Ogden P. 1990. An NMR study of plasticization and antiplasticization of a polymeric glass. Macromolecules 23:968–77. Lourdin D, Bizot H, Colonna P. 1997. “Antiplasticization” in starch-glycerol films? J Appl Polym Sci 63:1047–53.

136

PART 1: Invited Speakers and Oral Presentations

Lourdin D, Colonna P, Ring SG. 2003. Volumetric behaviour of maltose-water, maltose-glycerol and starch-sorbitol-water systems mixtures in relation to structural relaxation. Carbohydr Res 338:2283–7. Lourdin D, Ring SG, Colonna P. 1998. Study of plasticizer-oligomer and plasticizer-polymer interactions by dielectric analysis: maltose-glycerol and amylose-glycerol-water systems. Carbohydr Res 306:551–8. Maeda Y, Paul DR. 1987. Effect of antiplasticization on gas sorption and transport. III. Free volume interpretation. J Polym Sci [B] 25:1005–16. Mandala IG, Ioannou CA, Kostaropoulos AE. 2006. Textural attributes of commercial biscuits: effect of relative humidity on their quality. Int J Food Sci Technol 41:782–9. Marousis SN, Karathanos VT, Saravacos GD. 1989. Effect of sugars on the water diffusivity in hydrated granular starches. J Food Sci 54:1496–500, 1552. Marzec A, Lewicki PP. 2006. Antiplasticization of cereal-based products by water. Part I: Extruded flat bread. J Food Eng 73:1–8. Moraru CI, Lee T-C, Karwe MV, Kokini JL. 2002. Plasticizing and antiplasticizing effects of water and polyols on a meat-starch extruded matrix. J Food Sci 67:3396–401. Ngai KL, Rendell RW, Yee AF, Plazek DJ. 1991. Antiplasticization effects on a secondary relaxation in plasticized glassy polycarbonates. Macromolecules 24:61–67. Nicholls RJ, Appelqvist IAM, Davies AP, Ingman SJ, Lillford PJ. 1995. Glass transitions and the fracture behavior of gluten and starches within the glassy state. J Cereal Sci 21:25–36. Noel TR, Parker R, Ring SG. 1996. A comparative study of the dielectric relaxation behaviour of glucose, maltose, and their mixtures with water in the liquid and glassy states. Carbohydr Res 282:193–206. Noel TR, Parker R, Ring SG. 2000. Effect of molecular structure and water content on the dielectric behaviour of amorphous low molecular weight carbohydrates above and below their glass transition. Carbohydr Res 329:839–45. Noel TR, Ring SG, Whittam MA. 1992. Dielectric relaxations of small carbohydrate molecules in the liquid and glassy states. J Phys Chem 96:5662–7. Nomura S, Hiltner A, Lando JB, Baer E. 1977. Interaction of water with native collagen. Biopolymers 16:231–46. Peleg M. 1996. Mathematical characterization of the plasticizing and antiplasticizing effects of fructose on amylopectin. Cereal Chem 73:712–5. Pittia P, Nicoli MC, Sacchetti G. 2007. Effect of moisture and water activity on textural properties of raw and roasted coffee beans. J Texture Stud 38:116–34. Pommet M, Redl A, Stéphane Guilbert S, Morel M-H. 2005. Intrinsic influence of various plasticizers on functional properties and reactivity of wheat gluten thermoplastic materials. J Cereal Sci 42:81–91. Prevorsek DC, Butler RH, Reimschuessel HK. 1971. Mechanical relaxations in polyamides. J Polym Sci [A-2] 9:867–86. Reidy GA, Heldman DR. 1972. Measurement of texture parameters of freeze-dried beef. J Texture Stud 3:218–26. Rizvi SSH, Benado AL. 1984. Thermodynamic properties of dehydrated foods. Food Technol 38:83–92. Roberts GE, White EFT. 1973. Relaxation processes in amorphous polymers. In: Haward RN, editor. The physics of glassy polymers. London: Applied Science. p 153–222. Roos Y. 1996. Glass transitions in low moisture and frozen foods: effects on shelf-life and quality. Food Technol 50:95–108. Roudaut G, Dacremont C, Le Meste M. 1998. Influence of water on the crispness of cereal-based foods: acoustic, mechanical and sensory studies. J Texture Stud 29:199–213. Sears JK, Darby JR. 1982. The technology of plasticizers. New York: John Wiley. Sefcik MD, Schaefer J, May FL, Raucher D, Dub SM. 1983. Diffusivity of gases and main-chain cooperative motions in plasticized poly(vinyl chloride). J Polym Sci [B] 21:1041–54. Seow CC, Cheah PB, Chang YP. 1999. Antiplasticization by water in reduced moisture food systems. J Food Sci 64:576–81. Seow CC, Vasanti Nair CK, Lee BS. 1995. Effects of processing on textural properties of food phytosystems. In: Barbosa-Cánovas GV, Welti-Chanes J, editors. Food preservation by moisture control. Lancaster, PA: Technomic. p 697–728.

Antiplasticization of Food Polymer Systems by Low Molecular Mass Diluents

137

Shimazu AA, Mali S, Grossmann MVE. 2007. Plasticizing and antiplasticizing effects of glycerol and sorbitol on biodegradable cassava starch films. Sem Cienc Agrar Londrina 28:79–88. Shogren RL, Swanson CL, Thomson AR. 1992. Extrudates of cornstarch with urea and glycols: structure/ mechanical property relations. Starch/Stärke 44:335–8. Simatos D, Blond G, Perez J. 1995. Basic physical aspects of glass transition. In: Barbosa-Cánovas GV, Welti-Chanes J, editors. Food preservation by moisture control. Lancaster, PA: Technomic. p 3–31. Slade L, Levine H. 1991. Beyond water activity: recent advances based on an alternative approach to the assessment of food quality and safety. CRC Crit Rev Food Sci Nutr 30:115–360. Slade L, Levine H. 1995. Glass transitions and water-food structure interactions. Adv Food Nutr Res 38:103–269. Starkweather HW. 1980. Water in nylon. In: Rowland SP, editor. Water in polymers. Washington, DC: American Chemical Society. p 433–40. Suwonsichon T, Peleg M. 1998. Instrumental and sensory detection of simultaneous brittleness loss and moisture toughening in three puffed cereals. J Texture Stud 29:255–74. Van Hecke E, Allaf K, Bouvier JM. 1998. Texture and structure of crispy-puffed food products. Part II: Mechanical properties in puncture. J Texture Stud 29:617–32. Vrentas JS, Duda JL, Ling H-C. 1988. Antiplasticization and volumetric behavior in glassy polymers. Macromolecules 2:1470–5. Wu S. 1992. Secondary relaxation, brittle-ductile transition, and chain structure. J Appl Polym Sci 46:619–24.

Oral Presentations

9 Freeze Drying of Lactobacillus coryniformis Si3: Focus on Water Å. Schoug, J. Schnürer, and S. Håkansson

Abstract Freeze drying is commonly used to stabilize lactic acid bacteria. However, dehydration tolerance differs among strains of lactobacilli and also other factors have been reported to influence freeze-drying survival. The effects of sucrose concentration, cell density, and freezing rate on freeze-drying survival of Lactobacillus coryniformis Si3 have been evaluated. The water activity (aw) of the dry product, as well as selected thermophysical properties of importance for freeze drying, degree of water crystallization, and the glass transition temperature of the maximally freeze-concentrated amorphous phase (Tg′) were determined. Survival rates varied from 6% to 70% under different conditions, and all factors studied influenced the product’s aw and its survival of freeze drying. The most important factor for survival was the freezing rate. However, there were codependencies between freezing rate and formulation constituents, demonstrating the complexity of the system. The degree of water crystallization decreased and final aw increased as a function of sucrose concentration. The degree of water crystallization and Tg′ were not affected by cell density up to 1011 colony forming units/mL where these thermophysical values decreased possibly due to increased amounts of unfrozen water.

Introduction Microbial stabilization is important for successful commercialization of lactic acid bacteria. During the industrial stabilization process, many cells suffer damage and some die (Carvalho and others 2004). Freeze drying is considered a suitable drying process because heat and chemical degradation are minimized, and it is a wellestablished method for preservation of microorganisms. The extent to which bacterial cells succumb during the stabilization process has been shown to depend on several factors, such as bacterial strain, fermentation conditions, formulation, freeze-drying process parameters, and rehydration conditions (Carvalho and others 2004). The formulation and freeze-drying process are closely related and affect the final product quality; for example, both cellular survival and technical cake characteristics (Jennings 1999; Schoug and others 2006). The major obstacle to, and a prerequisite for stability is to remove water from living organisms without causing them substantial harm. Protective agents, such as monosaccharides or disaccharides, are very important for survival during drying and act by water replacement, glass formation, and 141

142

PART 1: Invited Speakers and Oral Presentations

depression of the melting temperature Tm in the cell membrane (Crowe and Crowe 1992a, 1992b; Crowe 2002). Sucrose is considered a well-functioning lyoprotectant, but, compared to trehalose, has a lower glass transition temperature (Tg), an increased tendency to pick up water and to hydrolyze, and can induce browning reactions upon storage (Crowe and others 2005). Many factors influence the appearance of a freeze-dried product. The degree of water crystallization is a measure of how much of the water in the formulation is available for sublimation (frozen in contrast to unfrozen), and a small amount of frozen water will result in a “tableted” structure. If the degree of crystallization is below 0.5, the formulation might “boil” and become vacuum dried. Collapse is a phenomenon that is related to a failure in the formation of a solid structure, which in turn depends on the formulation. Lactobacillus coryniformis Si3 has a potential use as an additive to silage. The strain has proven effective in defeating molds and acts as a biopreservative (Magnusson and others 2003). The survival of this strain during freeze drying varies depending on concentration of sucrose, the cell density, and the freezing rate. Here, we investigated a simple formulation consisting of cells, sucrose, and water and the effects of these parameters and freezing rate on product survival of freeze drying.

Materials and Methods Fermentation and Sample Preparation Fermentation was performed at pH 5.5 and 35°C in a 14-L fermentor (Belac, Stockholm, Sweden) with 9.6 L MRS medium and 400 mL L. coryniformis Si3 inoculum. The medium was sterilized in the fermentor at 121°C for 15 min. The pH was adjusted by addition of 4 M sodium hydroxide (Merck, Rahway, NY, USA) and 3 M potassium phosphate (Sharlau Chemie, Barcelona, Spain). After 16 h, the cell concentration was 109 colony forming units (CFU)/mL and the cells were harvested. Cells were centrifuged at 4000 g for 30 min at room temperature, washed once in bacteriologic peptone water (0.2% [wt/vol] [Oxoid, Lenexa, KS, USA] and 20 mg Tween 80 per liter), recentrifuged, and concentrated to the final cell concentrations. Sucrose (98%; SigmaAldrich, St. Louis, MO, USA) stocks were made with Millipore (nonpyrogenic) water to twice the final concentration and sterile filtered (0.45 μm; Nalgene, Rochester, NY, USA). The stock was mixed with the cell solutions in a 1 : 1 (vol/vol) ratio to produce different sucrose and cell densities. Small concentrations of Tween 80 (0.01 mg/mL, primarily to inhibit cell aggregation) and bacteriologic peptone (0.1% [wt/vol]), were also present. Thermal Analysis The formulations were stored at −70°C until analyzed by differential scanning calorimetry (DSC 220; SSC/5200H, Seiko, Tokyo, Japan). The DSC was run at 3°C/min to −50°C (in agreement with the freeze-drying program), kept isothermal for 2 min, and heated to 30°C at 1.5°C/min. Wet samples of approximately 40–50 mg were weighed and run in aluminum open sample pans (TA Instruments, Malmö, Sweden).

Freeze Drying of Lactobacillus coryniformis Si3: Focus on Water

143

An empty pan was used as reference. The heat of fusion (ΔHfusion in joules per gram) and the glass transition temperature of the maximally freeze-concentrated amorphous phase ( Tg′ ) were determined. Degree of crystallization was calculated by dividing the heat of fusion of the formulation by the heat of fusion of the total water in the formulations (Jennings 1999). Heat capacity (Cp) differs between water and ice, and the effect of different melting temperatures due to sucrose and cells was not accounted for when calculating degree of water crystallization. The Tg′ of the sucrose-cell formulations was obtained in the high-sucrose formulations and calculated as half the inflection point (½ Cp). The total amount of water in the different formulations was determined by weight-loss calculations after samples had been heated at 80°C for 2–3 days to constant weight. Freeze-Drying Protocol Freeze drying was performed in a pilot plant freeze dryer (Lyostar II; FTS kinetics, Stone Ridge, NY, USA). We used 5-mL vials (PharmaPack, Malmö, Sweden) with a final fill volume of 1.0 mL. Freezing was done, at different rates, to −50°C. The freezedrying cycle varied from approximately 44 to 52 h because of different freezing rates. After freezing, the vacuum was pulled down to 5.33 Pa (capacitance gauge) and, after 2 h, the shelf temperature was raised to −10°C. Secondary drying was done stepwise to 10°C. The program was then held at −4°C and 5.33 Pa. No instance of product collapse was observed. Cell Survival Cell viability was determined by 10-fold serial dilutions and plate assays on MRS agar (Oxoid). The plates were incubated in anaerobic jars under carbon dioxide– nitrogen gas atmosphere (GasPak System; BBL, Cockeysville, MD, USA) at 30°C for 48–72 h. Colonies in the interval 40–200 per plate were counted, and the mean values of three platings are reported. After freeze drying, rehydration was done in Millipore water at ambient temperature to the same volume as before drying. Survival was reported as percent surviving. Water Activity The water activity (aw) in the product was measured with an AquaLab CX-2 (Decagon Devices, Pullman, WA, USA). Three freeze-dried samples were pooled and aw measured at ambient conditions. The instrument was calibrated beforehand with distilled water (aw = 1.0) and with lithium chloride (high-purity standard, 13.3 molal, aw = 0.25) at ambient temperature. Statistical Analysis Freeze drying and aw were studied by varying in-dryer freezing rate (0.1°–5°C/min), cell density (108–1012), and sucrose concentration (2%–20%) with a response surface methodology (RSM) using a central composite face–centered (CCF) design (MODDE version 6; Umetrics, Umeå, Sweden) (Eriksson and others 2000). CCF designs facilitate stipulation of each factor ’s effect on the response, as well as interaction and

144

PART 1: Invited Speakers and Oral Presentations

quadratic factors. Response surface methodology is evaluated by multiple linear regressions (MLRs). The model fitness is expressed by the proportion of the total explained variation (R2, coefficient of determination), ranging from 0 to 1, and by the predicted variation (Q2, the measure of fit for the model) after cross-validation. The water content, thermophysical properties, Tg′, and degree of crystallization were, unless otherwise stated, determined in independent triplicate samples and analyzed for significance with paired two-sided Student’s t-test (95%).

Results and Discussion Important Factors for Freeze-Drying Survival The experimental design was used to determine the effects of sucrose concentration, cell density, and freezing rate on freeze-drying survival and aw of freeze-dried L. coryniformis Si3. We found that all factors influenced the survival of L. coryniformis Si3 (Figure 9.1). The survival rate varied from less than 6% to over 70% (Schoug and others 2006). There were several important interactions, as well as interactions with no statistical effect on survival. No visible collapse was observed in any product. According to the model, the freezing rate was the most important factor for survival, and statistically important interactions between freezing rate and cell density, as well as freezing rate and sucrose, were found (Figure 9.1). One of the main causes of cellular death during freeze drying is thought to be mechanical damage caused by ice formation (Mazur and others 1972). However, for successful freeze drying, some water must be crystalline and available for sublimation (Jennings 1999). A certain amount of water is always trapped in the amorphous phase at temperatures well below 0°C. The glass transition temperature varies with the formulation ingredients and is affected by, for example, electrolytes (Her and others 1995). The degree of water crystallization in formulations with different cell densities was determined, and it was found that, even at very high cellular numbers (1011 CFU/mL) in 20% sucrose, the D value (degree of water crystallization) was greater than 0.50, indicating that sublimation could occur (Schoug and others 2006). The cake appearance and aw differed among various conditions. At high sucrose and high cell densities, the degree of water crystallization was lower, and the product became more “tableted” in appearance (Jennings 1999; Schoug and others 2006). Water and Solute Properties When a cell-sucrose-water system freezes, the ice formation causes an increased solute concentration and induces an increased osmotic dehydration within the cells. The faster a system freezes, the more water will remain in the product (Ekdawi-Sever and others 2003) and the freezing rate will increase the aw of the product as shown in Figure 9.2. To control this phenomenon, an annealing step can be introduced in the freeze-drying cycle. By annealing, the amount of water can be dictated in lactic acid bacterial formulations (Ekdawi-Sever and others 2003). The maximally freezeconcentrated amorphous phase, Tg′, is considered an important process determinant for

(a)

y nsit l de l e C 9 10 11

12

5

8

Free zi

4

ng ra

te

3

2

1

0 0.3 0.2

Water activity

0.2

0.1

Water activity

0.3

0.1 0.0 0.0 5

4

3

Free zi

(b) l Cel

12

11

2 ng ra

0

12

te

sit den

10

1

9

10

11

8

ity

ens

ld Cel

Sucros

e

y

8

9

20

15

10

5 0.3

0.3

0.2 0.1

0.1

0.0

Water activity

Water activity

0.2

0.0 –0.1 –0.1 20

15

10

Sucros

e

5

12

11

9

10

8

ity

ens

ld Cel

Figure 9.1. The water activity of sucrose-cell formulations after freeze drying. The model for water activity had an R2 of 0.73, and a Q2 of 0.41. (a) Water activity of the product as a function of freezing rate and cell density at 20% sucrose. At high cell numbers, the product is very dense and dry. When the cake consists of fewer cells and more sucrose, the water activity is higher. (b) The water activity of the product at different sucrose and cell densities. The freezing rate was 2.5°C/min.

145

146

PART 1: Invited Speakers and Oral Presentations

Figure 9.2. Effects plot with error bars showing the 95% confidence level. The model for freeze-drying survival had an R2 of 0.94, and a Q2 of 0.78. Positive effects should be interpreted as a positive correlation between the factor and freeze-drying survival. The freezing rate, and interactions between the freezing rate and the formulation, significantly affect the survival. Den, cell density; Fre, freezing rate; and Suc, sucrose concentration.

maximally allowed product temperature during freeze drying. However, lactic acid bacteria have been shown to stabilize the system, and collapse might be avoided even though drying is performed above Tg′ (Fonseca and others 2004a, 2004b). We observed a slight decrease in the Tg′, accompanied with a lowering of the degree of water crystallization when high cell numbers were used (Schoug and others 2006). The decrease in Tg′ is most likely due to a plasticizing effect of water associated with the high cell number. When the cells become osmotically stressed by the sucrose, they release water and this might have a slight plasticizing effect. The effect was very small, though, and insignificant to optimization of the freeze drying. We found that the higher the sucrose concentrations that were used (up to 20 wt/ vol%), the better was the survival of L. coryniformis Si3 after freeze drying (Schoug and others 2006). Similarly, it has been shown for protein formulation that the higher the sucrose concentration is, the better is the stability, and a minimum of 0.3 M of sugars need to be present to achieve significant stability (Nema and Avis 1993; Arakawa and others 2001). The products with higher sucrose concentrations did display higher aw, and the products with high cell densities had low aw (Figure 9.2). The more sucrose there is, the more amorphicity or degree of disorder is found in freeze-dried formulations (Mosharraf and others 2007). Amorphous materials are very hygroscopic and, since the aw was determined at ambient conditions, these formulations might have adsorbed and absorbed more water. In conclusion, a high survival rate of L. coryniformis Si3 and good technical quality were achieved at approximately 1010 CFU/mL, at 15%–20% sucrose, and at a freezing rate of 2.8°C/min (Schoug and others 2006).

Freeze Drying of Lactobacillus coryniformis Si3: Focus on Water

147

Acknowledgments The authors thank Johan Olsson, Center of Human Studies of Foodstuffs, Uppsala University, for valuable discussions on experimental design. The Foundation for Strategic Environmental Research (MISTRA) is acknowledged for funding the research program DOM (Domestication of Microorganisms, http://www.mistra.org/ dom).

References Arakawa T, Prestrelski SJ, Kenney WC, Carpenter JF. 2001. Factors affecting short-term and long-term stabilities of proteins. Adv Drug Deliv Rev 46:307–26. Carvalho AS, Silva J, Ho P, Teixera P, Malcata FX, Gibbs P. 2004. Relevant factors for the preparation of freeze-dried lactic acid bacteria. Int Dairy J 14:835–47. Crowe JH, Crowe LM, Wolkers W, Oliver AE, Ma X, Auh JH, Tang M, Zhu S, Norris J, Tablin F. 2005. Stabilization of dry mammalian cells: lessons from nature. Integr Comp Biol 45:810–20. Crowe LM. 2002. Lessons from nature: the role of sugars in anhydrobiosis. Comp Biochem Physiol [A] 131:505–13. Crowe LM, Crowe JH. 1992a. Anhydrobiosis: a strategy for survival. Adv Space Res 12:239–47. Crowe LM, Crowe JH. 1992b. Stabilization of dry liposomes by carbohydrates. Dev Biol Stand 74:285–94. Ekdawi-Sever N, Goentoro LA, de Pablo JJ. 2003. Effects of annealing on freeze-dried Lactobacillus acidophilus. J Food Sci 68:2504–11. Eriksson L, Johansson E, Kettaneh-Wold N, Wikström C, Wold S. 2000. Design of experiments: principles and applications. Stockholm: Learnways. Fonseca F, Passot S, Cunin O, Marin M. 2004a. Collapse temperature of freeze-dried Lactobacillus bulgaricus suspensions and protective media. Biotech Prog 20:229–38. Fonseca F, Passot S, Lieben P, Marin M. 2004b. Collapse temperature of bacterial suspensions: the effect of cell type and concentration. Cryo Lett 25:425–34. Her LM, Deras M, Nail SL. 1995. Electrolyte-induced changes in glass transition temperatures of freezeconcentrated solutes. Pharm Res 12:768–72. Jennings TA. 1999. Lyophilisation: introduction and basic principles. Boca Raton, FL: CRC. Magnusson J, Ström K, Roos S, Sjögren J, Schnörer J. 2003. Broad and complex antifungal activity among environmental isolates of lactic acid bacteria. FEMS Microbiol Lett 219:129–35. Mazur P, Leibo SP, Chu EHY. 1972. A two-factor hypothesis of freezing injury. Exp Cell Res 71:345–55. Mosharraf M, Malmberg M, Fransson J. 2007. Formulation, lyophilization and solid-state properties of a pegylated protein. Int J Pharm 336:215–32. Nema S, Avis KE. 1993. Freeze-thaw studies of a model protein, lactate dehydrogenase, in the presence of cryoprotectants. J Parenter Sci Technol 47:76–83. Schoug Å, Olsson J, Carlfors J, Schnörer J, Håkansson S. 2006. Freeze-drying of Lactobacillus coryniformis Si3: effects of sucrose concentration, cell density, and freezing rate on cell survival and thermophysical properties. Cryobiology 53:119–27.

10 Water-Sorption Properties and Stability of Inclusion Complexes of Thymol and Cinnamaldehyde with β-Cyclodextrins P. A. Ponce, M. P. Buera, and B. E. Elizalde

Abstract β-Cyclodextrins (β-CDs), macrocyclic oligosaccharides formed by seven glucopyranose units, have a rigid lipophilic cavity and can form host-guest inclusion complexes with suitably sized hydrophobic molecules. Water content influences the structure of β-CDs and affects the chemical and physical stability of complexed compounds. This work investigated the relationship between the sorption characteristics of β-CDs and complexes formed with thymol and cinnamaldehyde and their release. The complexes were obtained by coprecipitation, filtered, freeze-dried, and stored at constant relative humidity (RH) in evacuated chambers (22%–97%) at 25°C. The release of encapsulated compounds was determined after equilibration by differential scanning calorimetry (DSC) following the enthalpy of the fusion peak at 50°C (thymol) and −7.5°C (cinnamaldehyde). DSC thermograms of complexes after freeze drying showed the disappearance of fusion peaks of thymol and cinnamaldehyde, indicating complex formation. Water-sorption isotherms for β-CD and the complexes showed constant and low water sorption at an RH of less than 80%, and then the uptake of water increased abruptly. At 95% RH, the water adsorbed was in the order β-CD (12.5 mol water/mol β-CD) > β-CD–thymol (8.5 mol water/mol complex) > β-CD–cinnamaldehyde (7 mol water/mol complex). The guest molecules displaced water molecules from inside the cavity of β-CD. At an RH of less than 84%, thymol and cinnamaldehyde were not released. The percent of released compound abruptly increased from 84% RH, coinciding with the abrupt increase of water. Water sorption significantly affected thymol and cinnamaldehyde complexes with β-CD, and complex stability was thus governed by the shape of the watersorption isotherm.

Introduction Cyclodextrins (CDs) are cyclic oligosaccharides consisting of glucose units linked by α1,4-glucoside bonds. The CDs composed of 6, 7, and 8 units are usually referred to as (α)-CD, (β)-CD, and (γ)-CD, respectively. They possess a hollow truncated cone shape with a nonpolar interior and two hydrophilic rims. Much of the interest in CDs arises from their ability to encapsulate hydrophobic molecules of suitable size inside their annulus to form inclusion complexes. The most probable mode of binding involves the insertion of the lipophilic portion of the guest molecule into the host cavity, and the 149

150

PART 1: Invited Speakers and Oral Presentations

displacement of the water molecules located inside the cavity (Uekama and others 1998). This ability has been used to increase the solubility, stability, and bioavailability of lipophilic drugs, vitamins, colorants, essential oils, and flavors in the food, cosmetic, and pharmaceutical industries (Loftsson and Brewster 1996; Uekama and others 1998; Hedges and McBride 1999; Buschmann and Schollmeyer 2002; Szente and Szejtli 2004). Also, their ability to alter physical, chemical, and biological properties of the guest molecule by the formation of inclusion complexes enabled their use as drug-carrier systems in pharmacology (Uekama and others 1998). β-Cyclodextrin (β-CD), because of its low cost and good complexation efficiency with a wide variety of drugs, has been the most used and studied among the cyclodextrins. The structure of hydrated β-CD is well known from X-ray studies (βcyclodextrin–12H2O [Lindner and Saenger 1982]), as well from neutron-diffraction studies (β-cyclodextrin–11H2O [Zabel and others 1986; Steiner and Koellner 1994]). About seven water molecules are located in the cavity. The rest of the hydration water is outside of the molecule and fills the interstices among CDs. Since most of the complexation reactions occur in an aqueous environment, the interaction between CDs and water is of fundamental importance, because the hydrophobic guest molecules compete with the water molecules in the cavity encircled by hydrogen bonding. Most research has been devoted to studying structures and the ability to encapsulate different compounds. However, no systematic studies have been conducted regarding the effects of environmental conditions on the stability of the formed complexes during storage. Thymol and cinnamaldehyde are the main constituents of thyme and cinnamon essential oils, respectively. Water content and water activity (aw) influence the structure of β-CD and affect the chemical and physical stability of complexed compounds. This work investigated the relationship between sorption characteristics of β-CD and complexes formed with thymol and cinnamaldehyde and their release.

Materials and Methods Materials β-CD was purchased from Sigma Chemical (St. Louis, MO, USA). Thymol and cinnamaldehyde were purchased from Carlo Erba (Milan, Italy). All other chemicals were of analytic grade and purchased from Mallinckrodt Chemical Works (St. Louis, MO, USA). Preparation of Solid Complexes Inclusion complexes of thymol and cinnamaldehyde (guest molecules) were prepared by the coprecipitation method (Mulinacci and others 1996). Thymol or cinnamaldehyde was added to a saturated solution of β-CD previously heated to 50°C, at equimolar concentrations with guest molecule–CD, and stirred for 4 h at that temperature. The solution obtained was allowed to cool to ambient temperature in contact with the guest

Water-Sorption Properties and Stability of Inclusion Complexes of β-Cyclodextrins

151

molecule excess and was then stored overnight at 2°C. The precipitated complexes obtained were filtered, freeze-dried, and stored under vacuum over magnesium perchlorate until they were used. Sorption Isotherms Sorption isotherms were determined by the standard isopiestic static-gravimetric method. Aliquots of about 500 mg of β-CD or of the complexes (β-CD–thymol or β-CD–cinnamaldehyde) were distributed into glass vials and exposed to atmospheres of saturated salt solutions of aw from 0.22 to 0.97 in evacuated desiccators at 25°C. The aw values were taken from Greenspan (1977). The time required for equilibration was 1–2 weeks, depending on the aw. The equilibrium moisture content of samples was determined by the difference in weight before and after drying in vacuum ovens at 98°C for 48 h. These conditions had been proven to be adequate for assessing constant weight after drying. The determinations were performed in triplicate, and the average value was reported. Storage Study The release of encapsulated thymol or cinnamaldehyde was determined by DSC following the fusion enthalpies in the ranges 48°–50°C for thymol systems and −7.5° to −10°C for cinnamaldehyde. The measurements were performed as a function of time after equilibration and storage in the vacuum desiccators at constant aw at 25°C. Differential Scanning Calorimetry A differential scanning calorimeter (Mettler TA 4000 with a TC11 TA processor and Graph Ware TA72 thermal analysis software [Mettler Instruments, Highstown, NJ, USA]) was used for all the measurements. The instrument was calibrated by using indium, zinc, and lead. Analysis in duplicate involved 40-μL hermetically sealed aluminum pans (Mettler) containing samples (of 5–10 mg). An empty pan was used as a reference. The dynamic method was used to determine melting points (Tm) and heat of fusion (ΔHm) of β-CD, thymol, cinnamaldehyde, and the complexes. A Tm was taken to be the onset of the melting peak. Each sample was heated at a rate of 10°C/min from −100° up to 110°C. The percent of pure compounds released (%R) from complexes was calculated from the ratio of the fusion enthalpy of thymol or cinnamaldehyde in the complexes (corrected according to the water content of the samples) and the fusion enthalpy of the pure compound, as indicated in Equation 10.1. % Released (% R ) =

ΔH s ΔH o

(10.1)

where ΔHs is the heat needed to melt thymol or cinnamaldehyde in the complexes, and ΔHo is the heat needed to melt pure thymol or cinnamaldehyde.

152

PART 1: Invited Speakers and Oral Presentations

Results and Discussion Sorption Moisture Studies Figure 10.1a shows the sorption isotherm of β-CD. During the sorption process, water uptake occurred up to aw 0.5. After the first sorption stage, the water content was constant up to aw 0.75, corresponding to a hydrated form of approximately 10.5 mol of water per mole of β-cyclodextrin. At aw values beyond 0.75, the sorbed-water concentration rapidly increased, reaching a final sorption level of 12.5 mol of water per mole of β-cyclodextrin at aw 0.97. This behavior agrees with that reported by Marini and others (1995), who determined that the number of water molecules changed

Figure 10.1. Water-sorption isotherms of (a) β-cyclodextrin (β-CD) and (b) β-CD– cinnamaldehyde and β-CD–thymol complexes, expressed as mol H2O/mol dry solids vs water activity (aw).

Water-Sorption Properties and Stability of Inclusion Complexes of β-Cyclodextrins

153

from 12.3 to 9.4 in a continuous and reversible way when aw changed from 1 to 0.15. Lindner and Saenger (1978, 1982), after crystallographic studies, reported that β-CD is associated with 12 water molecules forming dodecahydrate. The β-CD cavity is filled with 6.5 of those water molecules distributed statically over eight specific sites. The remaining 5.5 molecules are spread over eight other sites among β-CD molecules. The authors concluded that water in the β-CD is in an “activated” state and can be easily pushed out by a guest molecule and become part of the bulk water. Later, Fugiwara and others (1983) reported a new solvate form of β-CD with 11 molecules of water. Figure 10.1b shows the sorption isotherms of β-CD–cinnamaldehyde and β-CD– thymol complexes. It is noteworthy that, although both complexes exhibited a triphasic sorption profile globally similar to that of β-CD (Figure 10.1a), the amount of sorbed water at each aw was smaller. In fact, they formed hydrates with 7 (β-CD– cinnamaldehyde) or 8.5 (β-CD–thymol) water molecules. This indicates that the complexes were formed by displacement of water molecules from the β-CD cavity by the guest compound that was included. Differential Scanning Calorimetry The DSC thermograms for thymol and the β-CD–thymol complex are presented in Figure 10.2. Similarly, Figure 10.3 shows the DSC thermograms for cinnamaldehyde

^exo Thymol Complex b-CD–thymol

5

Complex b-CD–thymol (a w 0.97, 3 months)

wt/g

b-CD

°C -80

-60

-40

-20

0

20

40

60

80

100

120

140

160

Figure 10.2. Differential scanning calorimetry thermograms of the β-cyclodextrin (β-CD)–thymol system. From the top: pure component (thymol), inclusion compound, inclusion compound after 3 months at water activity (aw) = 0.97, and, finally, pure β-CD. Results are shown as heat flow (wt/g) vs temperature (°C). Exo, exothermic heat flow direction (upward); and wt/g.

154

PART 1: Invited Speakers and Oral Presentations

^exo

Cinnamaldehyde Complex b-CD–cinnamaldehyde freeze drying Complex b-CD–cinnamaldehyde (aw 0.97, 3 months)

2 wt/g

b-CD

°C

-80

-60

-40

-20

0

20

40

60

80

100

120

140

Figure 10.3. Differential scanning calorimetry thermograms of the β-cyclodextrin (β-CD)–cinnamaldehyde system. From the top: pure component (cinnamaldehyde), inclusion compound, inclusion compound after 3 months at water activity (aw) = 0.97, and, finally, pure β-CD. Results are shown as heat flow (wt/g) vs temperature (°C). Exo, exothermic heat flow direction (upward); and wt/g.

and the β-CD–cinnamaldehyde complex. The β-CD thermogram was included in each figure for comparative purposes. The β-CD thermogram shows an single peak a 93°C, indicating heat sorption probably caused by water evaporation. The absence of characteristic fusion peaks for thymol at 50°C (Figure 10.2) or for cinnamaldehyde at −7.5°C (Figure 10.3) in freezedried β-CD–thymol and β-CD–cinnamaldehyde samples, respectively, is strong evidence of the formation of the inclusion complexes.

Release of Thymol and Cinnamaldehyde During Storage At aw < 0.84, neither thymol nor cinnamaldehyde were released after 84 days of storage at 25°C. Figure 10.4a and b shows the %R of thymol or cinnamaldehyde from the complexes at aw 0.84 and aw 0.97, respectively. For the β-CD–thymol samples at aw 0.84, thymol release was very low after 70 days of storage and then increased slightly (Figure 10.4a). In contrast, at aw 0.97, the thymol release started from the beginning and increased almost linearly with storage of up to 70 days. For the β-CD–cinnamaldehyde samples at aw 0.84, cinnamaldehyde release began at 23 days of storage and increased linearly with time. However, after 75 days of storage, the %R was only 12. At aw 0.97, a different behavior was observed: The release started at 60 days of storage and the %R abruptly increased after that time (Figure 10.4b).

Water-Sorption Properties and Stability of Inclusion Complexes of β-Cyclodextrins

155

Figure 10.4. Percent released of (a) thymol or (b) cinnamaldehyde from their complex with β-cyclodextrin (β-CD) at water activity (aw) = 0.84 and aw = 0.97. Results shown as % released vs time (days).

The results obtained here show that the guest molecules released from β-CD complexes were detectable from aw 0.84, which coincided with the abrupt increase of water sorption observed in the corresponding isotherm. Water sorption significantly affected the stability of β-CD complexes with thymol and cinnamaldehyde, and guest molecule release was governed by the shape of the water-sorption isotherm.

References Buschmann HJ, Schollmeyer E. 2002. Applications of cyclodextrins in cosmetic products: a review. J Cosmet Sci 53:185–91.

156

PART 1: Invited Speakers and Oral Presentations

Greenspan L. 1977. Humidity fixed points of binary saturated aqueous solutions. J Res Natl Bur Stand [A] 81:89–102. Hedges A, McBride C. 1999. Utilization of cyclodextrin in food. Cereal Foods World 44:700–4. Lindner K, Saenger W. 1978. β-Cyclodextrin dodecahydrate: crowding of water molecules within a hydrophobic cavity. Angew Chem Int Ed Engl 17:694–5. Lindner K, Saenger W. 1982. Crystal and molecular structure of cyclohepta-amylose dodecahydrate. Carbohydr Res 99:103–15. Loftsson T, Brewster ME. 1996. Pharmaceutical application of cyclodextrins. I. Drug solubilization and stabilization. J Pharm Sci 85:1017–25. Marini A, Berbenni V, Bruni G, Massarotti V, Mustarelli P. 1995. Dehydration of the cyclodextrins: a model system for the interactions of biomolecules with water. J Chem Phys 103:7532–40. Mulinacci N, Melani F, Vincieri FF, Mazzi G, Romani A. 1996. 1H-NMR NOE and molecular modelling to characterize thymol and carvacrol β-cyclodextrin complexes. Int J Pharm 128:81–88. Steiner T, Koellner G. 1994. Crystalline β-cyclodextrin hydrate at various humidities: fast continuous and reversible dehydration studied by X-ray diffraction. J Am Chem Soc 116:5122–8. Szente L, Szejtli J. 2004. Cyclodextrins as food ingredients. Trends Food Sci Technol 15:137–42. Uekama K, Hirayama F, Irie T. 1998. Cyclodextrin drug carrier systems. Chem Rev 98:2045–76. Zabel V, Saenger W, Mason SA. 1986. Neutron diffraction study of the hydrogen bonding in β-cyclodextrin undecahydrate a 120 K: from dynamic flip-flops to static homodromic chains. J Am Chem Soc 108:3664–73.

11 Beyond Water: Waterlike Functions of Other Biological Compounds in a Waterless System B. R. Bhandari

Abstract This chapter focuses on the possible increased biochemical reactions in absence of water because of the presence of other polar solvents that can solubilize reactant molecules and increase their molecular mobility. Some of these solvents can be polyols. Until lately, the leading concept to describe the stability of food systems has been the water-activity theory, and lately some relationships have been found between the stability and glass transition temperature of the biological system. This chapter highlights that we need to move further from the water activity–related reaction-rate theory because a physicochemical reaction can occur in the absence of water if the nonwater liquid component can dissolve the solid components in the mixture at a given temperature. The current glass transition temperature theory has also not been able to describe the Maillard reaction occurring in waterless complex systems. Some past and present findings in the context of biological (mainly food) materials are presented in this chapter. The Maillard reaction occurring in the absence of water has been used as an example.

Introduction Almost all biological systems contain water. Water helps to increase the molecular mobility of biological systems by acting as a plasticizer, solvent, and vehicle for all biochemical movements and reactions. All biological components interact with water either directly or through mediating molecules (such as emulsifiers and surfactants). Traces of water can also be present even in dehydrated biomolecules (such as protein or hydrate crystals) to maintain their molecular structure. It has already been established that there is an adverse effect on proteins’ native structure or on other long-chain biopolymers if this trace of water is removed. In the absence of sufficient water, some materials vitrify or crystallize. A live cellular system can become inactivated or die because of overdehydration or the crystallization of solutes. Many low molecular weight food materials have been used for the protection of biological materials, such as proteins, enzymes, and even live cultures, because of the waterlike behavior of these molecules (possibly partially occupying the hydrophilic sites that supposedly interact with water). Water is also a strong polar solvent capable of weakening hydrogen, covalent, or ionic bonds in solids, thus dispersing them into the bulk water. Now we can raise 157

158

PART 1: Invited Speakers and Oral Presentations

a question: Can other relatively weak polar food materials that are in a liquid state also function as a solvent like water? Logically, they probably should be able to. Lower-chain polyols have been popular in food and pharmaceutical systems because of their plasticizing and humectant properties. They are added in conjunction or in preference to other common sugars. Do some polar polyols dissolve water-soluble solids? If so, they will promote various chemical reactions in foods, in a way similar to water, by increasing the molecular mobility of the system.

Understanding Molecular Mobility Molecular mobility is the primary parameter in causing physical or chemical change in biological materials. Molecular entities have three motions: translational (threedimensional displacement from one location to another), rotational (movement around an axis), and vibrational (stretching and bending of the bonds between the atoms, which changes the shape of the molecules). Translational motion and rotational motion relate to the movement of the entire molecule, whereas vibrational motion occurs within the molecule. The magnitude of these motions depends on the physical state of the matter (liquid, solid, or gaseous), because of the different degree of intermolecular interaction and free volume (intermolecular space) between the molecules in each of those states. The possible interactions between the molecules involve covalent and noncovalent bonds. Noncovalent bonds encompass electrostatic forces such as hydrogen, ionic, and dipole interactions and van der Waals forces. Noncovalent bonds exist more often on macromolecules and are common in biological molecules such as proteins and carbohydrates. Noncovalent bonds are important in forming the secondary and tertiary structures of the molecules. Covalent bonds are formed by the equal or unequal sharing of one or more pairs of electrons between atoms. Nonpolar bonds, which involve an equal sharing of electrons, are described as nonpolar because of the nonaccumulation of electrons and the absence of dipole movement. Covalent bonds are stronger than noncovalent bonds. The state of the matter determines the extent of these intermolecular forces and molecular mobility. It should be noted that the prerequisite for any reaction to occur is the collision between molecules. The probability of collision will be higher at higher molecular mobility and closer proximity of the reacting molecules or atoms. Reactant molecules are in very close proximity if they are miscible or soluble in the same solvent. An entire molecule does not have to be mobile for a reaction to occur. Depending on the type of molecule, if a part of the molecular functional group(s) or segment of the molecule has some kind mobility, the molecule can take part in a reaction if the orientation of the reactant groups in the molecules is right. Small molecules or plasticizers can increase this mobility. Credit goes to Slade and Levine (1991) for explaining a great deal about molecular mobility in relation to biological material stability.

Water-Activity Theory on the Stability of Biological Materials Biological materials (food, agricultural, pharmaceutical, and so on) are normally dried into solid form by removing water to limit the molecular mobility of the reactant

Beyond Water

159

molecules and extend the shelf life of the product. The degree of interaction of water with the biological components determines the activity of the water or its availability for providing mobility in the biological system. Water that is present in noncondensed state in biological matter (which means that individual molecules of water can interact with the solid components’ hydrophilic sites) may not provide molecular mobility, although water migration may occur from weaker interaction sites to stronger interaction sites in multicomponent systems. Water at this stage is normally referred as monomolecular layer water. Many studies have been published about the relationships between water activity (aw) and the rate of various reactions and growth of microbes. In water-rich biological systems, many reactions are diffusion limited, where water acts as a diffusion medium. Water has also been a diluting factor, reducing the reaction rate of nonenzymatic browning, and a protection factor against increased oxidation at lower moisture levels. Many low molecular weight biological materials have been used to protect biological materials, such as proteins, enzymes, and even live cultures, in the dry state, because of the waterlike behavior of these molecules (possibly partially occupying the hydrophilic sites that were previously interacting with water).

Waterlike Solvation Property of Polyols Food and pharmaceutical products contain many compounds that can be in a liquid state and that can act as a solvent for other polar solids. This can result in the dissolution of polar solids into polar liquids when the attractive forces between the liquids and solids exceed the attractive forces between the solid molecules. Many food systems contain components that are either in a liquid or a solid state with a certain level of polarity, as listed in Table 11.1. The polarity of a solvent can be determined based on measurement of its dielectric constant. A highly polar solvent will have a dielectric constant (ε) of greater than 50, whereas a semipolar solvent will have an ε of around 20–50, and a nonpolar solvent will have an ε of less than 20. The

Table 11.1. Dielectric constants (relative polarity) of various compounds Compounds

ε, at 20°C

Relative polarity

Water

80

High

Sorbitol (79% wt/wt)

62

High

Glycerol

46

Semi

Dimethyl sulfoxide (DMSO)

47

Semi

Methanol

33

Semi

Propylene glycol

32

Semi

Ethanol

25

Semi

Acetone

21

Semi

Olive oil

3.1

Nonpolar

Benzene

2.2

Nonpolar

ε, dielectric constant.

160

PART 1: Invited Speakers and Oral Presentations

Table 11.2. Approximate solubility of some sugars in glycerol at 60°C after an equilibration period of 10 days Sugars

Solubility (g/100 g glycerol)

Fructose

50

Glucose

20

Sucrose

15

Maltose

40

Lactose

10

From Bhandari and others (2009a).

Figure 11.1. Microscopic photographs of a sucrose crystal dissolving in anhydrous sorbitol at 100°C (sucrose/sorbitol ratio, 30 : 100).

data in Table 11.1 show that sorbitol and glycerol can act as a solvent for sugars if the sorbitol or glycerol are in a liquid state. Bhandari and Roos (2003) found sucrose crystals dissolving in sorbitol above the melting point of sorbitol (when sorbitol remained in a liquid state) (Figure 11.1). The solubility of various sugars in an aqueous solution of glycerol was reported in early publications (Segur and Miner 1953). However, no previous work has been reported on the solubility of sugars in anhydrous glycerol, which is in the liquid state at room temperature (20°C). Polyols such as glycerol or sorbitol have been widely used as plasticizers, but all plasticizers do not necessarily behave as solvents would. Bhandari and others (2009a) found that common food sugars (lactose, sucrose, fructose, glucose, and maltose) are all soluble in pure glycerol at various levels (Table 11.2). Obviously, their solubility is a function of the solvent temperature, and the equilibrium time required in glycerol was longer than in water because the high viscosity of glycerol limits the diffusivity or dispersability of solvated molecules across the solution (results not shown) and because of their relatively lower polarity than water. The solubility of various pharmaceuticals in glycerol has been reported (Seedher and Bhatia 2003). The solubility of glycerol will certainly be enhanced by the presence of a small amount of strongly polar water. The important question here regards not only the solubility, but the molecular mobility of solute in a waterless environment, because nonwater solvents can also provide molecular mobility to a complex (multicomponent) system. The role of

Beyond Water

161

glycerol as plasticizer has been recognized, but plasticizers do not necessarily solvate molecules. They increase the free intermolecular space (free volume) in the system, resulting in higher molecular mobility. The reaction rate will be much faster if the reactants are solubilized. This has implications for many low-moisture dry-food formulations that may contain sugars and polyols and other reactant molecules such as protein. The dissolution of some components, crystallization, and an antiplasticizing (toughening) effect may also result from the waterlike behavior of polyols. This implies that reaction in a low-moisture dry-food system can occur even if the water activity is close to zero if other mobile components are in the system.

Glass Transition, Molecular Mobility, and Sub-Tg Relaxation Glass transition temperature (Tg) has been the best concept related to molecular mobility. The glass transition of a multicomponent system will depend on the intrasolubility, plasticizing, and compatibility or miscibility of components with one another. At a glass transition temperature, the backbone of the molecular structure is normally frozen, which is referred as α relaxation. However, depending on the rate of this change, β relaxation of molecules continues below this glass transition temperature (α relaxation). This sub-Tg event is called molecular relaxation or structural relaxation, which occurs due to the nonequilibrium amorphous system (created during α relaxation) trying to reach the low-energy equilibrium state. It should be noted that some translational and rotational motion of smaller molecules (such as water or other solvents) or the segment or functional group of a molecule continues, though, of course, at a slower rate. Vibrational motion still exists (until the temperature of absolute zero). Dissolution of solutes in a solvent will reduce glass transition temperature. Plasticization and dissolution may influence the overall glass transition temperature of multicomponent systems differently. In a number of pharmaceutical systems, the effect of β relaxation on the stability of drugs or bioactives has been reported, but no such work has been widely published about food systems, although the effect has been identified in a few recent publications. It has also been reported that β relaxation is enhanced by the presence of water, resulting in a toughening of the product (antiplasticizing effect) below the glass transition temperature. We can not rule out the possibility of β relaxation being affected by other waterlike solvents such as glycerol.

Molecular Mobility and the Maillard Reaction Nonenzymatic browning (the Maillard reaction) is the best example illustrating the importance of molecular mobility, since this is the most common phenomenon responsible for food deterioration and is also easy to analyze because the reaction rate is very fast at higher temperature. The Maillard reaction occurs due to the condensation between amino groups and reducing sugars, and consequently many color compounds (low molecular weight polar, and high molecular weight nonpolar compounds) are produced, depending on the temperature and time. It is a bimolecular reaction; therefore, the molecular mobilities of two different molecular species are essential.

162

PART 1: Invited Speakers and Oral Presentations

The Maillard Reaction in the Absence of Water But Below the Glass Transition Temperature Relating the reactions on the basis of water activity and glass transition temperature may still be crude ways of prediction, although the use of these parameters has yielded the highest level of certainty in the majority of food and pharmaceutical systems. The Maillard reaction was found to occur in dry pasta systems (Fogliano and others 1999). Creation of some small, mobile molecules due to deamidation of glutamine was attributed to this reaction even below the Tg. Schebor and others (1999) also reported a Maillard reaction below the glass transition temperature in powdered anhydrous milk systems. In this case, the authors suspected that other types of localized mobility, such as glass relaxation, the intermixing of reactants increasing the proximity of the molecules in the concentrated state, or localized mobility near the pores might have existed, since no other components were present in the milk. However, the milk still contained some fat, which would have been mobile; and protein, which can be a part of fat globule membrane, would still have some mobility at the interface of solid oil. It is not clear though whether such factors can influence aminocarbonyl condensation. Any preexisting smaller molecules (amino compounds) can still retain translational motion within the matrix. Dissolution of some protein or peptide in lactose cannot be ruled out as causing this effect. (Localized β relaxation can accelerate the condensation of amino and carbonyl groups.) In an early study, Kamman and Labuza (1985) found that the Maillard reaction was accelerated by the presence of liquid-phase oil in a powder starch mix containing glucose and glutamate. Their assumption was that the oil may act as a solvent for the reactants. However, it was not explained how a hydrophobic oil can solubilize these polar reactants. This is an important area that needs further investigation. The results of some earlier, as well recent, studies, have indicated that the presence of glycerol can promote the Maillard reaction (Warmbier and others 1976; Mustapha and others 1998; Cerny and Guntz-Dubini 2006). Some of these studies have used some amount of water (glycerol as a cosolvent). However, Mustapha and others (1998) presented some Maillard reaction results at near zero moisture content (aw = 0.01) in glycerol. They found that the Maillard reaction between lysine and the xylose system was higher in a glycerol, than in a water medium. However, the addition of water to glycerol enhanced the reaction rate presumably because of the increased solubility of reactants. Our preliminary work found that sugars and amino acids are dissolved in glycerol, and the Maillard reaction occurs in an anhydrous glycerol + sugar + glycine system (Bhandari and others 2009b). The data clearly indicated that there is a Maillard reaction, and its rate depends on the temperature and reaction time (Figure 11.2). The compounds produced during early and later stages of Maillard reactions are now being determined and evaluated.

Concluding Remarks Some of the changes that occur in biological materials can be described accurately by neither the water-activity nor the glass transition theories. In complex systems, intra-

Beyond Water

163

100°C

115°C

0 min

30 min

60 min

130°C 120 min 180 min 240 min

24 H

Figure 11.2. Color development due to nonenzymatic Maillard browning in a waterless environment at 100°, 115°, and 130°C (fructose-glycine solution in glycerol). Samples were incubated for 24 h.

solubility and localized molecular mobility of various components are possible; in particular, the solubility of components in a waterless system warrants further study. This understanding is particularly useful in biological systems (such as food and pharmaceuticals) with limited water content. In many food products (intermediate or dried foods), polyols and sugars may be added to improve the shelf life and textural, sensory, or nutritional quality of the food products. In pharmaceutical products, polyols and sugars are added as bulking compounds, functional agents, or carriers. Not only the plasticizing effect but also the solvation property of some components may influence the chemical and physical changes in products. The solvation capacity of low-melting-point food components (such as fat, polyols, and sugars) and their role in intracomponent reactions based on the molecular mobility concept need further attention, which may answer some of the controversial questions raised about changes occurring in dry biological materials with very low water activity.

References Bhandari B, Ling OE, Yap O. 2009a. Dissolution of sugars in glycerol. J Agric Food Chem (forthcoming). Bhandari B, Roos Y. 2003. Dissolution of sucrose crystals in the anhydrous sorbitol melt. Carbohydr Res 338:361–7. Bhandari B, Wang CW, Wijayanti HB. 2009b. Maillard reaction of sugars with glycine in glycerol system. Food Chem (forthcoming). Cerny C, Guntz-Dubini R. 2006. Role of solvent glycerol in the Maillard reaction of D-fructose and L-alanine. J Agric Food Chem 54: 574–7. Fogliano V, Monti MS, Musella T, Randazzo G, Ritieni A. 1999. Formation of coloured Maillard reaction products in a gluten-glucose model system. Food Chem 66:293–9. Kamman JF, Labuza TP. 1985. A comparison of the effect of oil versus plasticized vegetable shortening on rates of glucose utilization in nonenzymatic browning. J Food Proc Preserv 9:217–22.

164

PART 1: Invited Speakers and Oral Presentations

Mustapha WAW, Hill SE, Blanshard JMV, Derbyshire W. 1998. Maillard reactions: do the properties of liquid matrices matter? Food Chem 62:441–9. Schebor C, Buera MP, Karel M, Chirife J. 1999. Color formation due to non-enzymatic browning in amorphous, glassy, anhydrous, model systems. Food Chem 65:427–32. Seedher N, Bhatia S. 2003. Solubility enhancement of cox-2 inhibitors using various solvent systems. AAPS Pharm Sci Tech 4:1–9, article 33. Segur JB, Miner CS. 1953. Sucrose and dextrose in aqueous glycerol. J Agric Food Chem 1:567–9. Slade L, Levine H. 1991. Beyond water activity: recent advances based on an alternative approach to the assessment of food quality and safety. Crit Rev Food Sci Nutr 30:115–360. Warmbier HC, Schnickels RA, Labuza TP. 1976. Effect of glycerol on nonenzymatic browning in a solid intermediate moisture model food system. J Food Sci 41:528–31.

12 Water Sorption and Transport in Dry, Crispy Bread Crust M. B. J. Meinders, N. H. van Nieuwenhuijzen, R. H. Tromp, R. J. Hamer, and T. van Vliet

Abstract Water-sorption and dynamic properties of bread crust have been studied in gravimetric sorption experiments. Water uptake and loss were measured while relative humidity (RH) was stepwise increased or decreased (isotherm experiment) or varied between two adjusted values (oscillatory experiment). Experimental results were compared with the Fickian diffusion model and empirical models like the exponential and power-law model. The sorption curves that resulted from the isotherm experiments were best described by the Fickian diffusion model for low RH and by the exponential model for high RH. Transport rates depended on moisture content and showed a maximum around RH = 70%. Adsorption and desorption curves from oscillatory experiments were best described by the exponential model. From comparison of the experimental sorption curves and the power-law model for short times it followed for all bread crust that the diffusion coefficient n is close to 1. Normally, this is associated with so-called case II diffusion and water transport that are limited by relaxation of the solid material. However, additional observations suggest that this may not be a valid explanation and that a gradual, instead of stepwise change in RH and/or a kinetic barrier for water transfer to the solid matrix may explain the observed exponential behavior.

Introduction One of the most important factors on which consumers base their appreciation of dry cellular solid-food products is crispness (e.g., see Szczesniak 1971; Roudaut and others 2002; Luyten and others 2004). A dry, crispy product that is in contact with a more humid environment takes up water, and the crispy character is quickly lost when the water activity (aw) becomes higher than about 0.5 (Labuza and Hyman 1998; Luyten and others 2004; Payne and Labuza 2005). To control crispness and increase the shelf life of composite products that consist of a dry and crispy part and a more humid and soft part, fundamental knowledge about water-sorption dynamics and its dependence on ingredients and morphology is needed. However, water sorption in these products is complex and governed by various phenomena like migration of water vapor through mesoscopic open pores and microscopic capillaries, sorption and migration of water through the solid matrix, swelling and/or relaxation of the matrix, and hysteresis. 165

166

PART 1: Invited Speakers and Oral Presentations

Most models that describe sorption dynamics in polymeric porous systems are based on diffusion (for a review, see Masaro and Zhu 1999). Two extreme cases can be distinguished in the literature: (a) Fickian diffusion, where the transport of a penetrant like water is controlled by its concentration gradient, and the dynamics can be described by Fick’s law; and (b) case II diffusion, where transport is controlled by the relaxation of the solid material. When sorption dynamics are controlled by both processes and the corresponding rates are similar, this is called anomalous diffusion. A way to distinguish between the sorption mechanisms is by analyzing the uptake behavior for short times. Then, Fickian diffusion behaves like t0.5, case II like t1, and anomalous diffusion like tn with 0.5 < n < 1. Various studies have been published about the sorption behavior in biopolymeric systems. Fickian diffusion including a water concentration–dependent diffusion coefficient was used to describe sorption dynamics of sponge cake (Guillard and others 2003), dry biscuit (Guillard and others 2004), and waxy maize starch (Enrione and others 2007). A single exponential was used to describe the sorption kinetics of wheat flours (Roman-Gutierrez and others 2002). For chitosan films, diffusion seems Fickian for aw < 0.4 and anomalous for aw > 0.4 (Despond and others 2001). Del Nobile and coworkers developed a model where Fickian diffusion, including a concentrationdependent diffusivity, and polymer relaxation were combined to describe sorption dynamics of spaghetti and films of nylon, chitosan, alginate, casein, and zein (Del Nobile and others 1997, 2000, 2003a, 2003b, 2004; Buonocore and others 2005). These studies gained much insight, but fundamental generic knowledge that enables crispness to be controlled is still lacking. Furthermore, no studies have been published about the sorption dynamics of crispy bread crust. Therefore, to obtain more insight into the water-sorption mechanisms and dynamics of bread crust, we conducted a detailed study. Gravimetric stepwise oscillatory and isotherm sorption experiments on model bread crusts were performed and compared with different sorption models such as Fickian diffusion, the exponential model, and the power-law model.

Materials and Methods Sorption experiments were performed on model bread crusts and crusts of a rusk roll. Model bread crusts were obtained by baking thin sheets of dough (diameter, 60 mm; and thickness, 2 mm) in a halogen heater. Dough was prepared from a protein-rich and starch-rich wind-sifted fraction of wheat flour. Rusk roll crusts were removed from rolls prepared from wheat flour and baked in an oven. Crust samples were milled and sieved afterward in three different fractions with sizes smaller than 63 μm, between 63 and 250 μm, and between 250 and 500 μm (called small, medium, and large, respectively). Gravimetric sorption experiments were performed on about 6 mg of these fractions by using a VTI-SGA 100 symmetric vapor-sorption analyzer (VTI, Hialeah, FL, USA) at a temperature T of 25°C. Samples were dried above phosphorus pentoxide for at least 3 days. Before sorption experiments started, a drying step was performed for 2 h at 50°C. Starch, protein, pentosan, and fat contents were also determined. A detailed description of the sample preparation and experimental methods,

Water Sorption and Transport in Dry, Crispy Bread Crust

167

as well as the results of the ingredient analysis, can be found elsewhere (van Nieuwenhuijzen and others 2007, 2008).

Models For Fickian diffusion, the weight change as a function of time t for a spherical particle having radius rs, initially having a homogeneous water concentration C0, a constant surface concentration Cs, and a diffusion coefficient D independent of concentration C is given by Crank (1975), 6 ⎛ m − m 0 = ( m ∞ − m 0 ) ⎜1 − 2 ⎝ π

∞

1

∑n n =1

2

⎞ exp ( − Dn 2 π 2 t rs2 )⎟ ⎠

(12.1)

with m = mt and mi = M0 + 兰Cidr (for i = t, 0, ∞) and M0 as the mass of the dry particle. This equation is fitted against experimental sorption curves, with D rs2 and m∞ as fit parameters. For all analysis, the first minute of the experimental sorption curve was excluded from the fit because during this time the system needs to settle after a sudden change in relative humidity (RH) and environmental conditions are uncertain. Furthermore, it is assumed that diffusion through the vapor phase is not limiting because the water diffusion coefficient in air is more than 5 orders of magnitude larger than that of the solid matrix. Sorption dynamics are also compared with single exponential behavior, where the weight change as function of time is described by m − m 0 = ( m ∞ − m 0 ) (1 − exp ( − kt ))

(12.2)

with k as the transport rate. Here k and m∞ are fit parameters. Exponential behavior may be used to describe the relaxation dynamics of the polymeric matrix (Berens and Hopfenberg 1978). For short times, sorption dynamics may also be described by the empirical powerlaw relation m − m 0 = at n

(12.3)

where the a is a constant and the n is the diffusion exponent. In the short time limit, the diffusion exponent n is evaluated from the sorption curves plotted on a log-log scale and is equal to the derivative of the first linear part in that plot. When n = 0.5, n = 1, or 0.5 < n < 1, the sorption mechanism is said to be Fickian, case II, or anomalous, respectively. Calculations were performed using Matlab (MathWorks, Natick, MA, USA).

Results and Discussion Figure 12.1a shows a typical example of the relative change in weight of bread crust (in this case, a starch-rich model crust, large fraction) together with the RH as

PART 1: Invited Speakers and Oral Presentations

(a)

100 80 60

20

40 10

1

0.98 RH [%]

(m-m0)/m0 [%]

30

(b)

R2

168

0.96

0.94 20

0

(c)

2000

0

4000 t [min]

6000

0

0.92

(d)

6

40 60 RH [%]

80

100

RH: 60% 70%

5.5

(m-m0)/m0 [%]

(m-m0)/m0 [%]

20

14

RH: 30% 40%

5

4.5

4 1800

0

13

12

11

1900

2000 t [min]

2100

2200

10 3000

3100

3200

3300

t [min]

Figure 12.1. (a) Measured relative change in weight (m − m0)/m0 (thick line, left axis) of a starch-rich model crust and the adjusted external relative humidity (RH) (thin line, right axis) as a function of time t during an isotherm experiment. (b) R2 of the best fit between the measured (m − m0)/m0 and the diffusion model (Equation 12.1, open squares) and single exponential (Equation 12.2, open circles) as a function of RH. (c) Enlargement of an adsorption curve (open circles) and the best fit of the diffusion model (Equation 12.1, dashed line) and single exponential model (Equation 12.2, continuous line) as a function of t. The external RH changed from 30% to 40%. (d) Similar, but RH changed from 60% to 70%.

a function of t. The sorption curves for each step were analyzed and fitted against the sorption models described earlier. Equilibrium values (m∞ − m0)/m0 yield the isotherm. For all bread crusts, isotherms showed hysteresis and were best described by the Guggenheim-Anderson-de Boer (GAB) equation. This corresponds well with what has been observed for most food systems (Basu and others 2006). The diffusion exponent n (Equation 12.3) turned out to be close to 1, suggesting the sorption mechanism to be case II and controlled by the relaxation rate of the polymeric matrix. Figures 12.1c and d show examples of adsorption curves for two RH steps. Also, the best fits of the diffusion model (Equation 12.1) and the exponential model (Equation 12.2) are shown. Comparison between experimental and simulated sorption curves

Water Sorption and Transport in Dry, Crispy Bread Crust

169

shows that water sorption in the low-moisture regime is better described by the Fickian diffusion equation, whereas in the high moisture regime it may be more accurately described by an exponential. This is also illustrated in Figure 12.1b, where the R2 of the fits is plotted for the adsorption curves as a function of RH. It is seen that the transition point is around RH = 40%. This is similar to results found for chitosan films (Despond and others 2001) and waxy maize systems (Enrione and others 2007), although in those cases the results were based on evaluation of the diffusion exponent n. For the desorption curves, similar behavior is found, only the change in going from more exponential to diffusive behavior happens at around RH = 60% (results not shown). Figure 12.2 (left) shows the sorption rates k obtained from the fits by using a single exponential as a function of RH. Also, diffusion coefficients D are shown (right) obtained from the fits by using the Fickian diffusion model and assuming a particle radius rs = 190 μm. The figure shows that water-adsorption rates increase with RH up to a value of about 70% and decrease at higher RH. Similar behavior is observed for the other bread crust samples (results not shown) and has also been observed for other food systems (Guillard and others 2003, 2004; Enrione and others 2007). The increase in transport rate may be associated with an increase in free volume due to the plasticizing effect of water. The decrease at high humidities may be attributed to capillary condensation and/or, more likely, to collapse and caking of the polymeric structure due to a transition from the glassy state to the rubbery state. For the bread crusts, this transition occurs at room temperature at about RH = 80% (van Nieuwenhuijzen and others 2008). Collapse might result in a reduction of the effective area exposed to the solvent and therefore a decrease in sorption rate.

-11

0.08

10

0.06 2

D [m /s]

k [1/min]

-12

10 0.04

-13

10 0.02

0

-14

0

20

40 60 RH [%]

80

100

10

0

20

40 60 RH [%]

80

100

Figure 12.2. (Left) Adsorption rates (open circles) and desorption rates (solid circles) k obtained from the best fits between the relative change of weight of a starch-rich model crust during the isotherm experiment shown in Figure 12.1a and the single exponential model (Equation 12.2) as a function of relative humidity (RH). (Right) Effective diffusion coefficients D during adsorption (open squares) and desorption (solid squares) obtained from the best fits of the Fickian diffusion model (Equation 12.1) assuming rs = 190 μm.

PART 1: Invited Speakers and Oral Presentations

6

6

5

5 (m-m0 )/m0 [%]

(m-m0 )/m0 [%]

170

4 3 2 1 0

4 3 2 1

0

1000

2000 t [min]

3000

0

0

1000

t [min]

2000

3000

Figure 12.3. (Left) Measured relative change in weight (m − m0)/m0 (thin line) of a starch-rich model crust because of an oscillatory stepwise change in adjusted relative humidity from 40% to 50% and the best fit of the diffusion model (Equation 12.1, dashed line) and single exponential model (Equation 12.2, continuous line) as a function of time t. Inset: Examples of experimental sorption curves (open circles) and best fits of the single exponential model. (Right) Simulated moisture uptake obtained by numerically solving the diffusion equation (D = 10−12 m2/s, rs = 190 μm) and an oscillatory stepwise change of surface concentration Cs between 4.4% and 6.1%. Inset: Calculated sorption curves (dashed line) and best fit of the single exponential model (continuous line).

For all RHs, desorption rates are found to be higher than adsorption rates. An explanation is that water-diffusion rates are higher than matrix relaxation rates so that during desorption, when the matrix is already swollen, the water transport is not limited by the matrix relaxation. However, then Fickian diffusion with n ≈ 0.5 is expected for the desorption steps, which is contrary to what is found in our study. It is also observed that the maximum desorption rate is at a lower RH than that of adsorption. This might be due to the hysteresis that shifts the glass-rubber transition to a lower RH. Figure 12.3 (left) shows a typical example of an oscillatory sorption experiment on bread crust. Here the relative moisture uptake is shown as a function of time for a starch-rich model crust (large fraction). The initially dry sample was exposed to an RH varying stepwise between 40% and 50% with an oscillation time of 25 min. It is seen that the moisture uptake seems to be a sum of a smooth curve that increases from zero to an equilibrium value and an oscillating one with the same frequency as the changes in RH. The figure also shows the best fits of the overall curve with the diffusion model (Equation 12.1) and the exponential model (Equation 12.2). The diffusion model turned out to describe the overall sorption curve best for low RH (oscillating between 40%–50% and 50%–60%) and the exponential model was best for higher RH (oscillating between 60%–70% and 70%–80%). This trend is similar to that found for the isotherm measurements. The fitted rate values are about a factor of 2 smaller than the rates found for the isotherm measurements because of the strong positive relation

Water Sorption and Transport in Dry, Crispy Bread Crust

171

between transport rates and moisture content. This causes an effective lower transport rate for the oscillatory experiments because the time-averaged moisture content of the crust particles during adsorption is lower compared to that during the isotherm measurements. The fitted equilibrium values correspond well to the adsorption branch of the isotherm. The inset in the left pane of Figure 12.3 shows the relative moisture uptake during an adsorption step and the best fit of the exponential model (Equation 12.2). All oscillatory experiments on bread crusts show that the diffusion exponent n is about 1 for all adsorption and desorption curves. Furthermore, all single sorption curves could be very well described by a single exponential (R2 > 0.98) and did not show the characteristic Fickian behavior with n = 0.5. This might be because during an oscillatory experiment the system is never in equilibrium and the moisture content at t0, when RH switches, is not homogeneously distributed. Thus, Equation 12.1 is not valid for describing the sorption curves of the oscillatory experiments. To check whether the exponential behavior is caused by the system not being in equilibrium when RH switches, we solved the diffusion equation numerically for a spherical particle with radius rs = 190 μm, initially dry (mt=0 = m0), and a stepwise oscillating surface concentration CS between 4.4% and 6.1% with an oscillation time of 25 min. Figure 12.3 (right) shows the results for a constant D = 10−12 m2/s. At first sight, there is a good correspondence between experimental and simulated water uptake. However, the simulated sorption curves for each oscillating step show the typical Fickian n ≈ 0.5 behavior, and poor correspondence is observed with a single exponential function, as can be seen in the inset in the figure. Calculations using a moisture-dependent diffusion coefficient taken from the isotherm measurements (Figure 12.2) show similar results and cannot explain the experimentally observed exponential behavior. Table 12.1 summarizes results of the comparisons between experiments, simulations, and model fitting. Figure 12.4 shows the mean sorption rates of starch-rich model crust during oscillatory sorption experiments in the stationary regime as a function of RH. Sorption rates kosc are obtained from best fits between experimental sorption curves and the single exponential model (Equation 12.2) for each oscillation step. The oscillatory

Table 12.1. Summary of fitted diffusional coefficient (n) and closeness of fits of the diffusion (D model) and exponential model (E model), for low and high relative humidity (RH), isotherm and oscillatory experiments, as well as isotherm and oscillatory simulation of Fickian diffusion in a sphere Experimental Isotherm

Simulation

Oscillation

Low

High

Low

High

D model

+

−

−

E model

−

+

n

∼1

∼1

RH:

Isotherm High

Low

−

++

++

++

++

+

+

−

−

−

−

∼1

∼1

∼0.5

∼0.5

∼0.5

∼0.5

++, +, and − correspond to R ≈ 1, ≈0.99, and <0.98, respectively. 2

Oscillation

Low

High

172

PART 1: Invited Speakers and Oral Presentations

k osc [1/min]

0.15

0.1

0.05

0 40

50

60 RH [%]

70

80

Figure 12.4. Adsorption rates (open circles) and desorption rates (solid circles) kosc of a starch-rich model crust during oscillatory sorption experiments. Mean values and indicated standard deviations are calculated from 20 fitted values after the stationary state has been reached. RH, relative humidity.

sorption rates are about a factor 10 larger than those obtained from the isotherm experiments. This is because, during oscillating RH, the moisture concentration varies only in an outer part of the particle while the inner part stays more or less constant at the time-averaged value. This can be seen effectively as a reduction in the particle radius, which results in an increase in the sorption rate. Sorption rates of an oscillatory sorption experiment are positively related to the oscillation frequency. The observation that the maximum in the sorption rates is shifted toward lower RH with respect to that found in the isotherm experiments is not fully understood yet, but may be related to differences in the relaxation state of the solid matrix. During the oscillatory experiments, the participating part of the matrix will probably be more relaxed than during isotherm adsorption experiments. It is therefore expected that, during oscillatory sorption experiments, water migration will not be limited by the polymer matrix relaxation. This would imply a Fickian behavior instead of the experimentally observed exponential one. From the aforementioned observations, no definite conclusion about the sorption mechanism of food systems can be drawn. Certain observations even seem contradictory. For example, for the bread crusts, n ≈ 1 for all humidities is found, suggesting water transport limited by matrix relaxation processes. For a system in the glassy state (low RH), this might be expected. However, this is contrary to the Fickian n ≈ 0.5 behavior found for chitosan and waxy maize for aw < 0.4. It is also expected that, for the stationary regime in an oscillatory experiment, the water transport rate is not limited by the relaxation rate, contrary to the exponential (n = 1) behavior that is

Water Sorption and Transport in Dry, Crispy Bread Crust

173

found. The same holds for desorption from the rubbery state. However, the association of Fickian behavior with n = 0.5 is based on the assumption that the particle surface is continuously in equilibrium with the surrounding air. It might, however, very well be that this assumption is invalid and that a sorption barrier or external resistance for water transfer exists between the air and solid matrix. This is also observed for water sorption in biscuits (Guillard and others 2004) and can account for the observed exponential behavior (Meinders and van Vliet 2009). This is also supported by the more surface-sensitive oscillatory experiments that show exponential behavior for all sorption curves. Crust particles are also not spherical or monodispersely distributed, but are cellular porous systems with irregular shapes and sizes, which also might have considerable effects on the estimation of the diffusion constant, n (Peppas and Brannon-Peppas 1994). Furthermore, n is determined from sorption curves at small t after a stepwise change in RH, at which experimental environmental conditions may be uncertain.

Conclusion Water-sorption curves of isotherm experiments, where the RH was stepwise increased or decreased, could be best described by the Fickian diffusion model for low RH and by the exponential model for higher RH. Transport rates depend strongly on moisture content and show a maximum around RH = 70%. For all RHs, desorption rates are found to be higher than adsorption rates. Moisture-uptake curves from oscillatory experiments are a sum of (a) a smooth curve increasing from zero to an equilibrium value and (b) alternating adsorption and desorption curves oscillating with the same frequency as RH. The smooth curve shows similar behavior to those measured during isotherm measurements. The oscillatory adsorption and desorption curves could be best described by the exponential model. All sorption curves behave for small times as t1, which is normally associated with so-called case II diffusion. This suggests that water transport is limited by relaxation of the solid material. However, this association may be invalid, and more information concerning the time dependency of moisture profiles in the solid material, matrix relaxation rates, available free volumes, and morphology is needed to reach a firm conclusion about the sorption mechanism of cellular solid-food systems.

References Basu S, Shivhare US, Mujumbar AS. 2006. Models for sorption isotherms for foods: a review. Drying Technol 24:917–30. Berens AR, Hopfenberg HB. 1978. Diffusion and relaxation in glassy polymer powders. 2. Separation of diffusion and relaxation parameters. Polymer 19:489–96. Buonocore GG, Conte A, Del Nobile MA. 2005. Use of a mathematical model to describe the barrier properties of edible films. J Food Sci 70:142–7. Crank J. 1975. The mathematics of diffusion. Oxford: Oxford University Press. Del Nobile MA, Buonocore GG, Altieri C, Battaglia G, Nicolais L. 2003a. Modeling the water barrier properties of nylon film intended for food packaging applications. J Food Sci 68:1334–40.

174

PART 1: Invited Speakers and Oral Presentations

Del Nobile MA, Buonocore GG, Conte A. 2004. Oscillatory sorption tests for determining the watertransport properties of chitosan-based edible films. J Food Sci 69:44–9. Del Nobile MA, Buonocore GG, Panizza A, Gambacorta G. 2003b. Modeling the spaghetti hydration kinetics during cooking and overcooking. J Food Sci 68:1316–23. Del Nobile MA, Massera M. 2000. Modeling of water sorption kinetics in spaghetti during overcooking. Cereal Chem 77:615–9. Del Nobile MA, Mensitieri G, Lostocco LR, Huang SJ, Nicolais L. 1997. Moisture transport properties of a degradable nylon for food packaging applications. Packag Technol Sci 10:311–30. Despond S, Espuche E, Domard A. 2001. Water sorption and permeation in chitosan films: relation between gas permeability and relative humidity. J Polym Sci Part [B] 39:3114–27. Enrione JI, Hill SE, Mitchell JR. 2007. Sorption behavior of mixtures of glycerol and starch. J Agric Food Chem 55:2956–63. Guillard V, Broyart B, Bonazzi C, Guilbert S, Gontard N. 2003. Evolution of moisture distribution during storage in a composite food: modelling and simulation. J Food Sci 68:958–66. Guillard V, Broyart B, Guilbert S, Bonazzi C, Gontard N. 2004. Moisture diffusivity and transfer modelling in dry biscuit. J Food Eng 64:81–7. Labuza TP, Hyman CR. 1998. Moisture migration and control in multi-domain foods. Trends Food Sci Technol 9:47–55. Luyten H, Plijter J, van Vliet T. 2004. Crispy/crunchy crust of cellular solid foods: a literature review with discussion. J Texture Stud 35:445–92. Masaro L, Zhu XX. 1999. Physical models of diffusion for polymer solutions, gels and solids. Prog Polym Sci 24:731–75. Meinders MBJ, van Vliet T. 2009. Modeling water sorption dynamics of cellular solid food systems using free volume theory. Food Hydrocolloids 23(8):2234–42. Payne C, Labuza T. 2005. The brittle-ductile transition of an amorphous food system. Drying Technol 23:871–86. Peppas NA, Brannon-Peppas L. 1994. Water diffusion and sorption in amorphous macromolecular systems and foods. J Food Eng 22:189–210. Roman-Gutierrez AD, Guilbert S, Cuq B. 2002. Distribution of water between wheat flour components: a dynamic water vapour adsorption study. J Cereal Sci 36:347–55. Roudaut G, Dacremont C, Vallès Pàmies B, Colas B, Le Meste M. 2002. Crispness: a critical review on sensory and material approaches. Trends Food Sci Technol 13:217–27. Szczesniak A. 1971. Consumer awareness of texture and of other food attributes. J Texture Stud 2:196–206. van Nieuwenhuijzen NH, Meinders MBJ, Tromp RH, Hamer RJ, van Vliet T. 2008. Water uptake mechanism in crispy bread crust. J Agric Food Chem 56:6439–46. van Nieuwenhuijzen NH, Tromp RH, Hamer RJ, van Vliet T. 2007 Oscillatory water sorption tests for determining water uptake behavior in bread crust. J Agric Food Chem 55:2611–8.

13 Water State and Distribution During Storage of Soy Bread with and without Almond A. Lodi and Y. Vodovotz

Abstract Inclusion of soy in the US diet remains relatively low despite the various epidemiological and human studies showing its health benefit. A soy bread (SB) of highly acceptable quality has been developed in our laboratory, as well as a variant of this SB, containing 5% (wt/wt, dry ingredients) of ground raw almonds (almond-enriched SB [ASB]). The latter was formulated with the objective of altering the isoflavone profile of the original product because of the high β-glucosidase content naturally present in almonds. Various factors (higher moisture content, dilution of gluten fraction, and decrease in starch content) needed to be considered to assess the impact of soy and almond addition on the quality and stability of baked bread. Since water plays a key role in both the formulation and storage of breads, this study monitored the changes in water state and distribution in SBs with and without almond during storage. SBs were made in house. The formulation consisted of 45.3% water, 6.6% soy-milk powder, 20% soy flour, 17.6% wheat flour, 2.3% gluten, 4.4% sugar, 1% yeast, 0.9% salt, 1.7% shortening, and 0.2% dough conditioner. ASB had the same formulation as SB with the addition of 5% wt/wt (dry ingredients) raw almond (substituted for the soy flour). Loaves were either analyzed at day 0 or sealed in polyethylene bags (to prevent moisture losses) and stored at 4°C for up to 10 days. Thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC), magnetic resonance imaging (MRI), and proton nuclear magnetic resonance (NMR) were used to characterize the water fraction in the SBs. Thermal analysis results showed that water was more tightly associated with the bread matrix in ASB than in SB. During storage, the easily removed water (TGA) and “freezable” water (DSC) increased with a larger change occurring in the ASB. MRI results indicated that water distribution in both breads was very homogeneous and changed little during storage. These findings are contrary to those observed in traditional wheat bread, where “freezable” water decreases during storage and water migrates from crumb to crust due to inhomogeneous water distribution. Additionally, proton NMR results showed that plasticization of the ASB matrix was first accomplished by water (up to day 3) and then by the lipid fraction. These results suggest that water state and distribution are affected by the addition of almond to the SB (especially in the first 3 days of storage), which may lead to improved loaf quality. 175

176

PART 1: Invited Speakers and Oral Presentations

Introduction The feasibility of adding soy protein to bread had first been considered in the mid 1950s as a means to enrich in high-quality protein (such as soy protein) common and relatively inexpensive food products (for example, bread), with the purpose of improving deficient diets, especially in developing countries (Ofelt and others 1954; Pomeranz and others 1969a; Hyder and others 1974). More recently, epidemiological and experimental evidence suggests that consumption of soybean products may have a significant, beneficial effect on health (Messina and others 1994; Setchell 1998; Birt and others 2001; Messina and Loprinzi 2001; Scheiber and others 2001; Brouns 2002; Hasler 2002). To that end, in 1999, the Food and Drug Administration (FDA) declared that 6.25 g of soy protein per serving (or 13% in the case of bread) included in a diet low in saturated fat and cholesterol may reduce the risk of coronary heart disease by lowering blood cholesterol levels (FDA 1999) and allowed product labeling to that effect (Henkel 2000). A patent-pending procedure and formulation of a soy bread (SB) of fully acceptable quality were developed in our laboratory (Vodovotz and Ballard 2002). A variant of the SB was also produced by incorporating 5% (wt/wt, dry ingredients) raw ground almond into the regular SB formulation (Vodovotz and others 2005). Almond addition both altered the isoflavones profile (almonds are a rich source of β-glucosidase, which helps cleave the sugar moiety from glucosides, yielding aglycones) and the lipid content (increased by ∼2.5%) of the final product. The addition of soy ingredients to bread required inclusions of larger amounts of water in the formulation (as previously reported by Ofelt and others [1954]) because of the different hygroscopic nature of soy. This can be hypothesized to affect the states of water molecules in the bread matrix of the fresh, and stored bread products. Therefore, the characterization of water distribution and mobility in SB with and without almond and their changes during storage are key to determining product stability and effect on staling rate. Various methods have been developed to measure the states and distribution of water in bread such as nuclear magnetic resonance (NMR), differential scanning calorimetry (DSC), thermogravimetric analysis (TGA), and dynamic mechanical analysis (DMA) (Chen and others 1997; Le Botlan and others 1998; Roudaut and others 1998; Ruan and others 1999; Baik and Chinachoti 2000; Wang and others 2004a, 2004b). These techniques require very small samples and therefore relate to specific portions of the bread loaf. In contrast, magnetic resonance imaging (MRI) is used to image the whole loaf of bread noninvasively and nondestructively. In this study, a combination of these methods was used to monitor the changes in water state and distribution in the SBs during storage.

Materials and Methods Preparation of Samples SB samples were prepared according to patent-pending procedures and formulations (Vodovotz and Ballard 2002; Vodovotz and others 2005). Formulation of SB and the source of the ingredients are listed in Table 13.1: 5% raw, ground almond (Wild Oats

Storage of Soy Bread with and without Almond

177

Table 13.1. Formulations of the regular soy bread Ingredients Water Soy-milk powdera

Soy-bread formulation (%wt) 45.35 6.61

Soy flourb

19.92

Wheat flourc

17.52

Glutend

2.30

Dough conditionere

0.20

Sugar

4.50

Yeastf

1.00

Salt

0.90

Shorteningg

1.70

Manufacturers of ingredients: a Soy-milk powder (Devansoy Farms, Carrol, IA, USA). b Baker ’s soy flour (ADM Protein Specialties Division, Decatur, IL, USA). c Baker ’s high-gluten-enriched bromated wheat flour, bleached (General Mills, Minneapolis, MI, USA). d Vital wheat gluten with vitamin C (Hodgson Mill, Teutopolis, IL, USA). e Dough conditioner (Caravan, Totowa, NJ, USA). f Red Star instant active dry yeast (Universal Foods, Milwaukee, WI, USA). g Crisco all-vegetable shortening, 50% less saturated fat than butter (Procter & Gamble, Cincinnati, OH, USA).

Markets, Boulder, CO, USA) was added to the regular SB to produce the almondenriched variant (ASB). After baking, the loaves were allowed to cool for about 4 h to room temperature and then were sealed in polyethylene bags to prevent moisture loss. One set of experiments was run immediately after cooling and was designated as “day 0” of the storage study. Samples were stored under accelerated staling conditions (4°C) between analyses and allowed to equilibrate for 3 h at room temperature prior to the experiments. All the experiments were performed on small-crumb portions obtained from the loaf center unless otherwise noted. Specific Loaf Volume Measurement Loaf volume was determined by using the Bread Volume Measurer BVM-L370 (TexVol Instruments, Viken, Sweden) on intact loaves. Each loaf was weighed, and the specific loaf volume was obtained from the ratio of volume and weight. The data reported are the result of at least five loaves per type of bread. Thermogravimetric Analysis About 20 mg of bread crumb, obtained from the center of the loaves, was placed in a previously tared stainless-steel pan (PerkinElmer Life and Analytical Sciences, Boston, MA, USA) inside a thermogravimetric analyzer (model 2950; TA Instruments, New Castle, DE, USA). Samples were heated from room temperature (∼20°C) up to 150°C at the rate of 5°C/min and held isothermally for 5 min. Thermograms of the

178

PART 1: Invited Speakers and Oral Presentations

sample weight as a function of temperature and its first derivative were considered for the analysis. The moisture content of the sample on a wet basis was determined by weight loss from the beginning, to the end of the experiment. The first-derivative curve was deconvoluted to a sum of three Gaussian peak functions to determine the presence of one or more water population (weight loss was presumed to be entirely due to water loss [Fessas and Schiraldi 2001]) and the range of temperature of vaporization within each population. Fitting procedures were accomplished using Matlab 6.5 (MathWorks, Natick, MA, USA). Differential Scanning Calorimetry Samples were prepared, immediately before analysis, by weighing and sealing about 10 mg of the bread crumb in large-volume stainless-steel sample pans with lids fitted with O-rings (PerkinElmer Life And Analytical Sciences) to prevent moisture loss during analysis. Samples and reference pans (left empty) were placed inside a differential scanning calorimeter (model 2920) equipped with a refrigerated cooling system (TA Instruments). The experimental procedure entailed cooling of the sample from room temperature (∼22°C) to −50°C at a rate of 5°C/min. The sample was then held isothermally for 5 min, heated at the rate of 5°C/min up to 140°C, held again isothermally for 5 min, and cooled to 25°C at the rate of 5°C/min. The percent “freezable” water (FW) (per unit weight of water in the sample) was calculated from the peak at about 0°C by using Equation 13.1: % FW =

A λ ⋅ mc

(13.1)

where A is the integrated area under the endothermic peak of water fusion, λ is the specific latent heat of fusion of water (334 J/g), and mc is the moisture content of the sample under analysis. MRI and NMR experiments are detailed in previously published work (Lodi and others 2007a, 2007b).

Results and Discussion Loaf Volume Loaf specific volume (Table 13.2) of ASB was found to be significantly larger than that of SB. This outcome is most likely due to the lipid fraction (Mizrahi and others 1967; Pomeranz and others 1969b) that is increased in the almond SB. In fact, 50% of almond kernels’ weight is comprised of lipids. This conclusion is also supported by the specific loaf volume results (data not shown [Lodi 2006]) obtained when almond lipid extract (instead of whole ground almond) is added to the traditional SB formula.

Storage of Soy Bread with and without Almond

179

Table 13.2. Moisture content and specific loaf volume of soy bread (SB) and almondenriched soy bread (ASB) samples

a

SB

ASB

Moisture contenta (%)

44.8 ± 0.3

42.8 ± 0.5

Specific loaf volume (cm3/g)

2.08 ± 0.05

2.26 ± 0.04

Wet basis.

Figure 13.1. Magnetic resonance proton-intensity signal images of soy bread on days 0 and 6 of storage at 4°C.

Moisture Loss and Migration During Storage Moisture content, measured on samples obtained from the central part of each loaf ’s crumb, was not significantly different from the value obtained in fresh bread (Table 13.2) at any point during storage. The fact that the moisture content, measured in the center of loaves, did not change during storage indicates both that minor moisture losses take place, and that moisture migration is minimal. A similar trend can be observed in these breads by using MRI, as shown in Figure 13.1 for the SB (the almond variant was very similar). The extremely homogeneous water distribution is considered the main reason for the lack of moisture migration, since no driving force is present even immediately after baking. Furthermore, it can be inferred that, although the moisture-content values included in Table 13.2 were measured on the central part of the crumb, these can be regarded as being representative of the overall moisture content in the crumb. Compared to standard wheat-bread formulas, the homogeneous water distribution in the crumb and the lack of moisture migration suggest that the inclusion of high amounts of soy in bread formula may slow the staling rate of bread (Baik and Chinachoti 2000).

180

PART 1: Invited Speakers and Oral Presentations

Water State Significantly larger amounts of water need to be added to the formulation of soycontaining breads to improve dough handling and loaf homogeneity. The state of water molecules in the bread matrix is important because it affects storage stability of the product. TGA and DSC were used to investigate the states of water in the SBs.

(b) 8

0.9 0.8

58 56

7

0.7 Area

0.6 0.5 0.4

54 52

6

50 5

0.3 0.2

48 46

4

0.1 0.0 20

40

60 80 100 Temperature (°C)

120

3

Peak temperaturer (°C)

(a) Weight-loss first derivative (mg/°C)

Thermogravimetric Analysis The analysis of weight-loss first-derivative curves, acquired using a thermogravimetric analyzer, helps to discriminate and quantify different water populations. These populations were characterized by varying binding strength to the macromolecular bread matrix, resulting in water evaporating at different temperature ranges (Fessas and Schiraldi 2005). Deconvolution of the TGA weight-loss first-derivative data of SB and ASB samples resulted in three Gaussian peaks (which were found to provide the best fit), as seen in Figure 13.2a for SB. The area under each peak obtained from the deconvolution procedure was proportional to the amount of each population’s water evaporating in the correspondent temperature range (Schiraldi and others 1996). Peak temperatures obtained for the different water populations of SB were found to be about 45°, 65°, and 80°C, whereas for ASB these were about 50°, 70°, and 80°C. Although the values of temperature and area on each day of storage were different for SB and ASB, trends were similar between these samples. The two peaks at lower temperature (<70°C) exhibited a shift toward higher temperatures and larger areas (as seen in Figure 13.2b for the lowest peak temperature curve) during storage. Moreover, in ASB, water molecules that were easily removed (two peaks at lower temperatures) evaporated later (at higher temperature) than in SB, indicating a stronger binding to the bread matrix in ASB than in SB. In both SB products, the peak around 80°C was not observed to undergo any major shift in temperature (Lodi 2006), but the water

44 0

2

4 6 8 Storage time (days)

10

42 12

Figure 13.2. (a) Typical weight-loss first-derivative curve from thermogravimetric analysis (TGA) and the three best-fit Gaussian peaks obtained from deconvolution (dotted lines). (b) Temperature of maximum of peak and area under the main deconvoluted curve (arrow) obtained from the TGA weight-loss first-derivative curve. SB, solid triangles and circles; and ASB, open triangles and circles.

Storage of Soy Bread with and without Almond

181

Figure 13.3. “Freezable” water in soy bread (SB) and almond-enriched soy bread (ASB) obtained from the differential scanning calorimetry endothermic peak at ∼0°C during 10 days of storage. “Freezable” water is expressed as percent weight of total water content in bread samples.

population (amount) evaporating at this temperature decreased (Lodi 2006), indicating a decrease in bound water during storage. Differential Scanning Calorimetry Percent FW levels in fresh SB and ASB (Figure 13.3) were not significantly different (P < 0.05). During the first 10 days of storage, the ASB FW content increased from ∼78% to ∼83% (expressed as percent weight of total moisture content in the sample), whereas %FW levels in fresh SB did not change significantly (P < 0.05). These results suggest that, during storage, bulk water progressively increases in ASB and more water becomes available for freezing. This is unlike wheat-based breads, where FW levels usually decrease during storage (Vodovotz and others 1996). However, Baik and Chinachoti (2000) observed that, upon storage of wheat bread without the crust (to minimize moisture-content gradients within the loaf section and thus prevent moisture migration from the crumb to the crust), FW levels remained constant during storage. As previously remarked, the addition of soy to bread formulations was found to promote a very homogeneous distribution of water molecules throughout the loaf (Figure 13.1). Therefore, lack of change in FW for SB can be attributed to the minimal moisture migration from the localized area tested. Furthermore, the addition of soy, which has high affinity for water molecules, may have an effect on FW, as well (Zhang 2004). In ASB, FW increased significantly during storage (in particular within the first

182

PART 1: Invited Speakers and Oral Presentations

day of storage), and this may be due to a phase separation of water and lipids (because of the higher amount of the latter in the ASB matrix), increasing the bulk-water fraction in the matrix as storage proceeds. A schematic summarizing some of the physicochemical differences between SB and ASB is presented in Figure 13.4. In particular, the specific loaf volume of ASB Fresh Soy Bread

Fresh Almond-Soy Bread

Soy Protein Wheat Protein Lipids Amylopectin Amylopectin (crystallized) Water

Air cells

Air cells

Stored Soy Bread

Air cells

Stored Almond-Soy Bread

Air cells Phase-separated water

Figure 13.4. Schematic of changes occurring in soy bread with and without almond during storage.

Storage of Soy Bread with and without Almond

183

was found to be larger than in regular SB, and the state and distribution of the water differed. Although water dynamics in SB during storage were found to differ considerably from those found in ASB, these changes dominated only in the first 3 days. Our previously published results (Lodi and others 2007b) indicate that the higher amount of lipids (∼2.5%), introduced in the almond-enriched soy product by the addition of 5% (wt/wt, dry ingredients) almond powder, might reinforce the plasticizing effect of water, as well as improve loaf quality (loaf specific volume in particular). Nonetheless, the differences encountered on addition of soy (and almond) may improve storage stability of traditional wheat bread significantly.

References Baik MY, Chinachoti P. 2000. Moisture redistribution and phase transitions during bread staling. Cereal Chem 77:484–8. Birt DF, Hendrich S, Wang W. 2001. Dietary agents in cancer prevention: flavonoids and isoflavonoids. Pharmacol Ther 90:157–77. Brouns F. 2002. Soya isoflavones: a new and promising ingredient for the health foods sector. Food Res Int 35:187–93. Chen PL, Long Z, Ruan R, Labuza TP. 1997. Nuclear magnetic resonance studies of water mobility in bread during storage. LWT Food Sci Technol 30:178–83. Fessas D, Schiraldi A. 2001. Water properties in wheat flour dough. I. Classical thermogravimetry approach. Food Chem 72:237–44. Fessas D, Schiraldi A. 2005. Water properties in wheat flour dough. II. Classical and Knudsen thermogravimetry approach. Food Chem 90:61–8. Food and Drug Administration. 1999. Food labeling: health claims; soy protein and coronary heart disease. Fed Reg 64:57699–733. Hasler CM. 2002. The cardiovascular effects of soy products. J Cardiovasc Nurs 16:50–63. Henkel J. 2000. Health claims for soy protein, questions about other components. FDA Consum 34:13–5, 8–20. Hyder MA, Hoseney RC, Finney KF, Shogren MD. 1974. Interactions of soy flour fractions with wheat components in breadmaking. Cereal Chem 51:666–75. Le Botlan D, Rugraff Y, Martin C, Colonna P. 1998. Quantitative determination of bound water in wheat starch by time domain NMR spectroscopy. Carbohydr Res 308:29–36. Lodi A. 2006. Physico-chemical and molecular characterization of soy bread containing almond [PhD diss]. Columbus: Ohio State University. Lodi A, Abduljalil A, Vodovotz Y. 2007a. Characterization of water distribution in bread using magnetic resonance imaging. Magn Reson Imaging 25:1449–58. Lodi A, Tiziani S, Vodovotz Y. 2007b. Molecular changes in soy and wheat breads during storage as probed by NMR. J Agric Food Chem 55:5850–7. Messina MJ, Loprinzi CL. 2001. Soy for breast cancer survivors: a critical review of the literature. J Nutr 131(Suppl 11):3095–108. Messina MJ, Persky V, Setchell KDR, Barnes S. 1994. Soy intake and cancer risk: a review of the in vitro and in vivo data [Review]. Nutr Cancer 21:113–31. Mizrahi S, Zimmermann G, Berk Z, Cogan U. 1967. The use of isolated soybean proteins in bread. Cereal Chem 44:193–203. Ofelt CW, Smith AK, Mills JM. 1954. Baking behavior and oxidation requirements of soy flour. II. Commercial defatted soy flours. Cereal Chem 31:23–8. Pomeranz Y, Shogren MD, Finney KF. 1969a. Improving breadmaking properties with glycolipids. I. Improving soy products with sucroesters. Cereal Chem 46:503–11.

184

PART 1: Invited Speakers and Oral Presentations

Pomeranz Y, Shogren MD, Finney KF. 1969b. Improving breadmaking properties with glycolipids. II. Improving various protein-enriched products. Cereal Chem 46:512–8. Roudaut G, van Dusschoten D, van As H, Hemminga MA, Le Meste M. 1998. Mobility of lipids in low moisture bread as studied by NMR. J Cereal Sci 28:147–55. Ruan RR, Wang XA, Chen PL, Fulcher RG, Pesheck P, Chakrabarti S. 1999. Study of water in dough using nuclear magnetic resonance. Cereal Chem 76:231–5. Scheiber MD, Liu JH, Subbiah M, Rebar RW, Setchell KDR. 2001. Dietary inclusion of whole soy foods results in significant reductions in clinical risk factors for osteoporosis and cardiovascular disease in normal postmenopausal women. Menopause 8:384–92. Schiraldi A, Piazza L, Riva M. 1996. Bread staling: a calorimetric approach. Cereal Chem 73:32–9. Setchell KD. 1998. Phytoestrogens: the biochemistry, physiology, and implications for human health of soy isoflavones. Am J Clin Nutr 68(Suppl):1333–46. Vodovotz Y, Ballard C, inventors. Calfee Halter & Griswold, assignee. 2002, 9 Oct. Formula and process for making soy-based bakery products. Patent application, docket no. 10-267845. Vodovotz Y, Hallberg L, Chinachoti P. 1996. Effect of aging and drying on thermomechanical properties of white bread as characterized by dynamic mechanical analysis (DMA) and differential scanning calorimetry (DSC). Cereal Chem 73:264–70. Vodovotz Y, Schwartz S, Zhang YC, inventors. Calfee Halter & Griswold, assignee. 2005, 29 Sept. Methods for enhancing soy-containing foods and foods made thereby. Patent application, docket no. 22727. Wang X, Choi SG, Kerr WL. 2004a. Effect of gluten content on recrystallisation kinetics and water mobility in wheat starch gels. J Sci Food Agric 84:371–9. Wang X, Choi SG, Kerr WL. 2004b. Water dynamics in white bread and starch gels as affected by water and gluten content. Lebensm Wiss Technol Food Sci Technol 37:377–84. Zhang YC. 2004. Physicochemical properties and isoflavone content of bread made with soy [PhD diss]. Columbus: Ohio State University.

14 Phase Separation of Ice Crystals in Starch-Based Systems During Freezing and Effects on Moisture Content and Starch Glass Transition T. Tran, K. Piyachomkwan, and K. Sriroth

Abstract To better understand the effects of freezing in starch-based food, this work focused on characterizing the decrease in moisture content of the starch fraction during freezing, for various total moisture contents (31.0%–245.6% dry-weight basis [dwb]) and under different freezing temperatures (−50° to −20°C), by using differential scanning calorimetry to measure the proportions of freezable and nonfreezable water. At low total moisture content (31.0%–71.6% dwb), most of the water was bound to the starch fraction, whereas, at high total moisture content (103.4%–245.6% dwb), the starch fraction became saturated with a constant moisture content in the range 33.9%–42.5% dwb, whereas the amount of freezable water increased from 60.9% up to 211.7% dwb. The proportion of nonfreezable water also decreased with freezing temperature, from 49.5% dwb at −20°C to 41.4% dwb at −50°C, in the case of a 71.6% dwb total moisture-content system. At low temperatures, the freezing equilibrium was reached more quickly, from 4 min at −50°C to 30 min at −30°C. At −20°C, the equilibrium was not reached after 1 h. By reporting the decrease in moisture content of the starch fraction on a starch state diagram, it became apparent that although all samples eventually reached the glassy state, samples frozen at −20°C remained in the rubbery state during most of the freezing process. These results suggest that real starch-based products frozen at −20°C may experience significant retrogradation before reaching the more stable glassy state.

Introduction The freezing of starch-based foods causes molecular changes detrimental to their texture. Understanding the interactions of water with starch during freezing is therefore important to improve product quality and shelf life. As illustrated on Figure 14.1, the freezing process results in a micro-scale phase separation, with the formation of ice crystals dispersed in the starch matrix, while the latter experiences a reduction of its moisture content, or freeze concentration (Jeong and Lim 2003). When freeze concentration reaches its maximum, the starch fraction is reported to retain 28% of unfrozen water on a wet-weight basis (wwb) (Hsu and others 2003). Maximum freeze concentration is achieved under slow freezing rates. However freezing of real food products is carried out under quick freezing conditions in order to reduce ice crystal 185

186

PART 1: Invited Speakers and Oral Presentations

Figure 14.1. Starch state diagram illustrating the freeze concentration of the starch fraction during freezing for a system initially at 50% moisture content.

size and limit changes to the product texture. The objectives of this work were to characterize the decrease in moisture content of the starch fraction during freezing under quick freezing conditions and to determine its moisture content at equilibrium, for various initial moisture contents and various freezing temperatures.

Materials and Methods Preparation of Pregelatinized Starch Cassava starch was fully gelatinized by heating to 95°C in excess water and stirring continuously for 20 min with a mechanical stirrer. Gelatinized starch was then transferred onto a plastic tray, dried in a hot-air oven at 85°C overnight, and ground into a powder by using a kitchen blender. The moisture content was equilibrated to 13.2% ± 0.1% wwb in a relative-humidity box containing a saturated solution of sodium chloride (NaCl) for 15 days. Moisture-content measurements were performed in triplicate following the Association of Analytical Chemists International method 950.46 (AOAC 1995). Measurement of Freezable and Bound Water by Differential Scanning Calorimetry Differential scanning calorimetry (DSC) was performed with a Perkin-Elmer DSC7 instrument (Wellesley, MA, USA) calibrated with indium. An empty stainless-steel pan was used as reference. The appropriate amounts of pregelatinized starch and distilled water were mixed directly in stainless-steel DSC pans so as to obtain starch

Freezing and Effects on Moisture Content and Starch Glass Transition

187

systems with 30%, 40%, and 50% (wwb) moisture contents. DSC pans contained typically 50–60 mg of sample. Following a homogenization pretreatment at 130°C for 15 min, each sample was placed in the DSC chamber and cooled to the required freezing temperature (−30°, −40°, or −50°C) at 100 K min−1. The sample was equilibrated for increasing times from 18 s to 30 min, and the ice-melting event was recorded at regular intervals by heating up to 30°C at 10 K min−1. The enthalpy of the ice-melting event was calculated from the peak area recorded by DSC and the weight of water in the sample. The frozen-water weight fraction in the sample was calculated as the ratio between the ice-melting event enthalpy and the melting enthalpy of the total amount of water in the sample, based on the latent heat of ice melting, 334 J/g (Ribotta and Le Bail 2007). The nonfreezable water fraction was expressed as the difference between the total water weight fraction and the frozenwater weight fraction. The proportions of freezable and nonfreezable water were further expressed as percentages of the solids’ weight and of the total sample weight.

Results and Discussion Effect of Initial Water Content and Freezing Temperature on the Starch-Fraction Moisture Content After Freezing The path followed by the starch samples (30%, 40%, and 50% wwb) on the state diagram during freezing by DSC is plotted in Figure 14.2. The initial sample cooling to the freezing temperature (−30°, −40°, or −50°C) is shown by the downward arrows.

Figure 14.2. Changes in moisture content of the starch fraction during freezing at various temperatures for samples with 30%, 40%, and 50% initial moisture contents (wet-weight basis).

188

PART 1: Invited Speakers and Oral Presentations

Table 14.1. Moisture content (%wwb) in the starch fraction after freezing at −30, −40, and −50°C for samples with 30%, 40%, and 50% overall moisture content Moisture content Freezing temperature

30%

40%

50%

−30°C

25

33

32

−40°C

27

32

31

−50°C

26

29

30

The moisture content of the starch fraction was assumed to remain constant during the cooling because of the high cooling rate (100°C · min−1) and small size of the sample. The following decrease in the moisture content of the starch fraction is shown by the horizontal arrows for each moisture content. The moisture contents in the starch fraction after equilibrium was reached are summarized in Table 14.1. The reduction in the starch-fraction moisture content was small in the 30% moisture-content sample (down to 25%–27%), whereas samples at 40% and 50% moisture content experienced a greater decrease, down to 29%–33%. This difference reflects the ability of starch to retain approximately 28% (wwb) bound water when maximum freeze concentration is reached, whereas any water above that level remains freezable. The starch fractions reached approximately the same moisture content regardless of the freezing temperature employed (−30°, −40°, −50°C), indicating the limited effect of freezing temperature on the ability of starch to retain bound water. Effect of Initial Water Content and Freezing Temperature on the Rate of Freeze Concentration The decrease in moisture content of the starch fraction in function of freezing time is shown on Figure 14.3 for starch systems at 30%, 40%, and 50% total moisture content and for three different freezing temperatures (−30°, −40°, and −50°C). The starch-fraction moisture contents after equilibrium is reached are summarized in Table 14.2, together with the time at which the equilibrium is attained. Table 14.2 also presents the times at which half the freezable water present in each sample became frozen. These results indicate that the rate of ice formation is influenced by both the freezing temperature and the initial moisture content of the system. As expected, ice formation was faster at −50°C than at −40° and −30°C. Ice formation was also faster in the systems at 40% and 50% moisture content (excess water conditions) than in the system at 30% moisture content (limited water conditions). Hence, the conditions for fastest ice formation were at −50°C in systems with 40%–50% moisture contents, for which the 29%–30% starch-fraction moisture content at equilibrium was reached in less than 3 min. When freezing at −30°C, the time to reach equilibrium was 15–30 min, notably longer than at −40° and −50°C; however, the time to freeze half of the freezable water remained short and comparable to freezing at −40° and −50°C (0.5–3.0 min).

Freezing and Effects on Moisture Content and Starch Glass Transition

189

Figure 14.3. Evolution of starch-fraction moisture content as a function of freezing time at various freezing temperatures.

Table 14.2. Effect of the freezing temperature on the time needed to freeze 50% and 100% of the available freezable water, and corresponding moisture content (MC) of the starch fraction Initial MC →

30% MC (%)

Freezing temp.

40%

Time (min)

MC (%)

50%

Time (min)

MC (%)

Time (min)

Time to freeze 50% of the available freezable water

−30°C

28

2.5

37

3

41

<0.5

−40°C

28

2

36

1

41

<0.3

−50°C

28

2.5

35

0.7

40

<0.3

−30°C

25

30

33

15

32

−40°C

27

10

32

6

31

3

−50°C

26

10

29

3

30

<3

Time to reach moisture-content equilibrium 15

Conclusion The effects of moisture content and temperature on the ice formation and freeze concentration of a starch system during freezing were investigated. At low moisture content (30% wwb), the starch fraction experienced a small reduction in moisture content, down to 25%–27% wwb, which corresponds to the maximum freeze concentration (Hsu and others 2003). At higher moisture contents (40% and 50% wwb), the moisture content of the starch fraction decreased to 29%–33% wwb, slightly above

190

PART 1: Invited Speakers and Oral Presentations

the maximum freeze concentration. The freezing temperatures investigated (−30°, −40°, and −50°C) had a limited effect on the final starch fraction moisture content, but influenced the rate of ice formation. When freezing at −40° and −50°C, equilibrium moisture content in the starch fraction was reached within 10 min, whereas, at −30°C, the equilibration time was at least 15 min and up to 30 min in the case of the 30% moisture-content sample (although half of the available freezable water was frozen in less than 3 min). The initial moisture content also influenced the rate of ice formation, with the 30% moisture-content sample markedly slower to reach equilibrium than the 40% and 50% moisture-content samples. For the manufacture of frozen foods weighing typically several grams instead of milligrams, as were the samples in this study, a practical implication of these results is that freezing at −30°C or above (e.g., −20°C) may delay reaching equilibrium compared to freezing at −40° or −50°C. This in turn suggests that samples frozen at temperatures above −30°C may spend longer time in the rubbery state before reaching the glass transition (Figure 14.2) and may undergo more molecular reorganizations, in particular starch retrogradation, which has consequences for the texture of the product. Longer equilibration times during freezing are also known to lead to the formation of larger ice crystals which are detrimental to the product texture.

Acknowledgments This work was supported by the National Center for Genetic Engineering and Biotechnology (BIOTEC) and the National Science and Technology Development Agency of Thailand.

References Association of Analytical Chemists International (AOAC). 1995. Official methods of analysis of AOAC International. 16th ed. Gaithersburg, MD: AOAC. Hsu C-L, Heldman DR, Taylor TA, Kramer HL. 2003. Influence of cooling rate on glass transition temperature of sucrose solutions and rice starch gel. J Food Sci 68:1970–5. Jeong H-Y, Lim S-T. 2003. Crystallinity and pasting properties of freeze-thawed high amylose maize starch. Starch/Stärke 55:511–7. Ribotta PD, Le Bail A. 2007. Thermo-physical assessment of bread during staling. LWT Food Sci Technol 40:879–84.

15 Carrot Fiber as a Carrier in Spray Drying of Fructose K. Cheuyglintase and K. R. Morison

Abstract Differential scanning calorimetry (DSC) at 0.1°C min−1 and dielectric analysis (DEA) in the range 200 Hz to 1 MHz, 10°–105°C, were applied to find the glass transition temperature (Tg) (onset) from freeze-dried mixtures of carrot fiber plus fructose. Spray-dried powders of fructose–carrot fiber in a laboratory-scale dryer were compared to those of fructose-maltodextrin (dextrose equivalent, 9.8 maximum). In DEA, the temperature of the peak derivative (assumed to be Tg) was found to be frequency dependent. When fructose–carrot fiber mixtures were spray dried, most of the powder stuck to the dryer walls. The powder swept from the walls was free flowing, with a moisture content of 2%–4%. There was no significant difference in stickiness and moisture content when the fiber content was higher. A DSC scan of 40%, 50%, 60%, and 70% carrot fiber–fructose showed transition temperatures were 107, 114.5°, 122.9°, and 130°C, respectively. These values indicated the wall buildup might be avoided in a larger-scale dryer. Solutions of at least 50% (wt/wt, dry) maltodextrinfructose were successfully dried at 175°/85°C inlet/outlet temperatures.

Introduction Food products in powder form obtained from spray drying are normally convenient, stable, concentrated, and natural. Spray drying is a continuous process, simple to operate, and has a low labor cost. However, the spray drying of fruit juice is difficult because fructose is a major component, and thus the spray-drying performance of fructose is an important indicator for the study of fruit-juice spray drying. Spray drying of fructose is difficult because it always exists as a very viscous, deformable plastic and seems never to form a perfect solid. It therefore needs a carrier to enable the production of free-flowing powder (Roos and Karel 1991; Bhandari and Howes 1999). Because of the lower Tg and plasticizing effect of fructose, the Tg of glucose-fructose and sucrose-fructose blends has been found to decrease with increasing fructose content (Roos 1995). The powder from spray drying is in an amorphous metastable form that is very sensitive to changes in temperature and moisture content (Roos and Karel 1991; Roos 1995; Cahn and others 2005). Spray drying a mixture of low molecular weight sugars could result in either a completely amorphous product or some microcrystalline regions dispersed in the amorphous mass (Bhandari and others 1997). 191

192

PART 1: Invited Speakers and Oral Presentations

The change in state from rubberlike to glasslike occurs as a second-order phase transition at what is known as the glass transition temperature (Tg) (McCrum and others 1991; Roos 1995; Cahn and others 2005). At temperatures near Tg, changes in specific volume, refractive index, density, heat content, thermal conductivity, and electric properties also occur (Newey and Weaver 1991; Lievonen and Roos 2003). However, transition temperatures vary with the method and the rate of the measurement (Grate and others 1992; Roos 1995; Cahn and others 2005). Differential scanning calorimetry (DSC) detects the typical change in specific heat at Tg. Usually, Tg values reported are either onset or midpoint temperatures of the Tg temperature range (Roos 1995). Because of the kinetic character of the transition, the Tg value should always be provided with the experimental conditions, such as the heating/cooling rates (Roos 1995; Park and Kojima 2000). Dielectric analysis (DEA) measures the dielectric properties of a material—permittivity and loss, as a function of temperature, time, and frequency—and is used for studying biopolymers (McCrum and others 1991; Bone and Zaba 1992; Lievonen and Roos 2003; Kilmartin and others 2004). At low moisture contents, the capacitance is directly proportional to the dielectric permittivity (Kilmartin and others 2004). Biopolymers are inherently microscopically heterogeneous and are composed of molecules that exhibit considerably different electrical properties. When an electric field is applied, the mobility of charge carriers, such as ions, migrating through the material can be significantly higher in some regions than in others. The heterogeneous system then exhibits frequency-dependent properties (McCrum and others 1991; Bone and Zaba 1992). A drying aid is added to the solution before spray drying to: improve the drying rate, prevent stickiness, reduce hygroscopicity, maintain flowability, encapsulate flavor and aroma, and maintain quality of the powder in storage (Roos 1995; Edwards and Langrish 2004). The most common carriers used are various maltodextrins whose Tg varies from 100° to 188°C, depending on their dextrose equivalent (DE) property (Bhandari and Howes 1999). As a practical matter, fruit or vegetable juice powder spray drying can be used when a sufficiently high concentration of natural fibers remains in the juice. The use of fibers, obtained from fruit processing, as carriers to produce free-flowing powder offers great potential for avoiding contamination with foreign substances and for further uses of the waste materials. The use of fiber as a carrier in spray drying could also be applied to other food products as a natural alternative to the carriers now used (Edwards and Langrish 2004). Fibers are composed of structural polysaccharides (cellulose and hemicellulose) bound as a fine latticework with gums, oligosaccharides (inulin and lignins), polymers based on phenylpropane units, and lipids (waxes and cutins) (Nyman and others 1993; Ramulu and Rao 2003). A study of dietary fiber in carrot pomace found 3.88% pectin, 12.3% hemicellulose, 51.6% cellulose, and 32.2% lignin (Nawirska and Kwasniewska 2005). A study of thermal transitions in carrot and its cell-wall components found that the Tg of the cell-wall-rich phase with low moisture content was about 62°C (Georget

Carrot Fiber as a Carrier in Spray Drying of Fructose

193

and others 1999). Researchers found that the Tg of cellulose, lignin, and hemicellulose was higher than 200°C (Salmen and others 1984; Georget and others 1999).

Materials and Methods Materials Fresh carrots were cleaned and, without any heat treatment, passed through a juice extractor. The fibers (retained by the extractor) were frozen at −30°C overnight before being freeze dried. A vacuum oven at 45°C was used to dry the fibers further before processing in an Ultra Centrifugal mill at 14,000 rpm with sieve no. 0.2. The carrot fiber powder was stored at −20°C. The particle size distribution was determined by a laser diffraction analysis machine (Microtrac X-100; Microtrac, Largo, FL, USA) with water as a fluid media. Dried D-fructose (Sigma-Aldrich, St. Louis, MO, USA) and deionized (DI) water were used to prepare the amorphous fructose and mixtures. A 90% fructose solution was mixed with dried carrot fiber powder at 14%, 21%, and 28% wt/wt (dry). DI water was added to the freeze-dried samples to equilibrate the moisture content to 2% and 5%. A vacuum oven (45°C until final weight stable) was used to determine moisture content before DSC and DEA and after DEA. Fieldose 10C AP maltodextrin (DE, 9.8 maximum) was obtained from Penford Australia (Lane Cove, Australia). Methods An HP (Hewlett-Packard) 1100 series HPLC (high-performance liquid chromatography) (Agilent Technologies, Santa Clara, CA, USA) with RID 1047A and a 35101 A Prevail carbohydrate ES column (Alltech Associates, Deerfield, IL, USA) was used at a flow rate of 1 mL/min with 75% acetonitrile in DI water as a solvent. D-fructose, D-glucose, and D-sucrose (Sigma-Aldrich) were used for sugar concentration standards. DI water was added to the carrot fiber to bring the total solid to its original value. A multicell differential scanning calorimeter (CSC 4100; Calorimetry Sciences, Linden, UT, USA) with MC-DSCRUN version 2.2.0 software (TA Instruments, New Castle, DE, USA) was used. The equipment was calibrated with sapphire and DI water. The samples were sealed in hastelloy-C ampoules with lids and were cooled to −15°C, annealed, and then heated at 0.1°C min−1 to produce the thermograms. A PM6306 Programmable Automatic RCL (resistance-capacitance-inductance) meter (Fluke, Everett, WA, USA) was used to measure the capacitance of the sample placed between two 50-mm-diameter aluminum disks of 1-mm clearance (Kilmartin and others 2004). The cell samples were placed in a vacuum oven at 45°C for 1 h to get rid of air; then the screw was immediately tightened, and all connecting parts were sealed with vacuum grease before being placed in a temperature-controlled chamber. A Visual Basic program (Microsoft) was used to record capacitance at frequencies from 200 Hz to 1 MHz every 30 s for the entire temperature range. The data were analyzed and filtered to reduce noise by using an Excel spreadsheet. The peak of the derivative of capacitance with respect to temperature was considered as the Tg of the

194

PART 1: Invited Speakers and Oral Presentations

samples. The Tg value from DSC for each sample was used as a guide to find a matching frequency for the DEA Tg. A laboratory scale spray dryer (Mobile Minor; GEA Niro, Søborg, Denmark) with a 0.6 × 1.8-m-high chamber and a rotary atomizer was used. An aqueous solution containing 40% total solids of a mixture of 40%, 50%, 60%, and 70% dry-weight carrot fiber–fructose was spray dried at inlet/outlet temperature 165°/75°C. The spraydrying temperatures for mixtures of maltodextrin-fructose solution were 175°/85°C. Immediately after spray drying, the moisture content was determined and DSC was performed.

Results and Discussion Carrot Fiber Powder Size and Sugar Content The particles of fiber were irregularly shaped, with most of the sizes distributed from 10 to 200 μm. This was small enough to be atomized by using a rotary disk. The carrot fiber was found to contain 0.3% fructose, 0.15% glucose, and 0.5% sucrose. DSC and DEA The Tg of freeze-dried amorphous fructose appeared at around 18°C (moisture content, 0.05%). Some researchers reported the Tg of melted amorphous fructose at 5°C (Roos 1995; Bhandari and Howes 1999). Figure 15.1 shows the derivatives of capacitances with respect to temperature versus temperature range at different frequencies for freeze-dried amorphous fructose. The derivative peak decreases with increasing frequency and appears at a higher temperature when frequency increases and matches Tg from DSC at 500 Hz.

Figure 15.1. Derivative of capacitance versus temperature for freeze-dried amorphous fructose. E, exponential (for example, 3.0E-11 = 3.0−11).

Carrot Fiber as a Carrier in Spray Drying of Fructose

195

Table 15.1. Tg from DSC and corresponding DEA frequencies for carrot fiber in freeze-dried amorphous fructose Carrot fiber/fructose ratio (dry basis) 0 : 100

Moisture content (% wet basis)

DSC Tg (°C)

DEA frequency for same Tg (Hz)

0.05

18.0

500

14 : 86

5.0

59.0

1,300

21 : 79

5.0

69.5

20,000

28 : 72

5.0

82.0

6,000

21 : 79

0.01

98.2

10,000

21 : 79

2.0

90.5

7,000

21 : 79

5.0

69.5

20,000

DEA, dielectric analysis; DSC, differential scanning calorimetry; and Tg, glass transition temperature.

Results from the freeze-dried mixtures are listed in Table 15.1. The Tg decreased when the moisture content increased, which conformed with the outcome reported by many researchers (Gupta 1978; Roos and Karel 1991; Roos 1995, 2002; Roos and others 1996). Increasing the ratio of carrot fiber to fructose increased the Tg, which was expected because an increase in a high molecular weight component normally increases the Tg (Roos 1995; Bhandari and Howes 1999). The DEA frequency at which the Tg value was the same as the DSC value showed no clear trend either with moisture content or the amount of carrot fiber. This casts doubt on the usefulness of DEA-based measurement of Tg. Spray Drying of Fructose Mixtures The moisture content of free-flowing spray-dried powder of all mixtures was found to be around 2%–4% wet basis. Values of Tg from the DSC of spray-dried powders are listed in Table 15.2. Higher fractions of fiber and maltodextrin in fructose resulted in a higher Tg. The 40% maltodextrin-fructose mixture was sticky in the dryer chamber and was not recovered as powder. The maltodextrin-fructose mixtures at 50%, 60%, and 70% produced a little free-flowing powder, some of which was deposited on the dryer wall as a sweepable powder. The fiber-fructose mixtures produced a little freeflowing powder, but most of the powder was deposited as a sweepable agglomerate on the dryer wall. The yield of swept powder was greater for the fiber-fructose mixture (∼62%) than for the maltodextrin-fructose mixture (∼20%). As Bhandari and Howes (1999) have reported, during the drying of a sugar-rich product the product might either remain as syrup or stick on the dryer chamber wall. It formed unwanted agglomerates on the walls of the dryer chamber and conveying system, which led to lower product yields and operating problems. The mechanism of stickiness and caking of amorphous sugars is through the phase change of the amorphous sugar from a glass to a rubber at temperatures above the Tg. The rate of cohesiveness development is thought to be proportional to the T − Tg, so, the higher temperature is, the faster the powder will develop liquid bridges, which may cause caking (Bhandari and Howes 1999).

196

PART 1: Invited Speakers and Oral Presentations

Table 15.2. Tg and yield from DSC of spray-dried carrot fiber–fructose and maltodextrinfructose at different ratios Carrot fiber/fructose ratio (dry basis)

DSC Tg (°C)

Yield (%)

Maltodextrin/fructose ratio (dry basis)

DSC Tg (°C)

40 : 60

107

51.5

40 : 60

50 : 50

115

63.1

50 : 50

119

12.9

60 : 40

123

65.7

60 : 40

124

19.7

70 : 30

130

68.0

70 : 30

135

27.0

100 : 0

187 ± 11

—

100 : 0

Not found

Yield (%)

183 ± 1

—

—

DSC, differential scanning calorimetry; and Tg, glass transition temperature.

In studies of the spray drying of citrus juice with maltodextrin (DE, 10) as a carrier, Gupta (1978) found the high temperature of the dryer wall melted the maltodextrin powder and formed a glassy coating. He also found that the powder ’s moisture content upon impacting the dryer wall resulted in the sticky coating. He further discussed that an incompletely dried product, with more than 4% moisture, would not be free flowing.

Conclusion Increasing carrot fiber in fructose solution increased the Tg of the mixtures. There was no clear relationship between peak frequency and Tg from DEA. Carrot fiber was a more effective carrier for spray drying than was maltodextrin when spray-drying yields were compared for powder with a similar Tg.

References Bhandari BR, Datta N, Howes T. 1997. Problems associated with spray drying of sugar rich foods. Drying Technol 15:671–84. Bhandari BR, Howes T. 1999. Implication of glass transition for the drying and stability of dried foods. J Food Eng 40:71–9. Bone S, Zaba B. 1992. Bioelectronics. London: John Wiley. Cahn RW, Haasen P, Kramer EJ. 2005. Materials science and technology, vol 9. Weinheim, Germany: Wiley-VCH. Edwards K, Langrish TAG. 2004. Characterisation of fruit fibre particles for use as carriers in spray drying. Paper presented at the Chemeca Conference, Sydney, September. Georget DMR, Smith AC, Waldron KW. 1999. Thermal transitions in freeze-dried carrot and its cell wall components. Thermochim Acta 332:203–10. Grate JW, Wenzel SW, White RM. 1992. Frequency-independent and frequency-dependent polymer transitions observed on flexural plate wave ultrasonic sensors. Anal Chem 60:413–23. Gupta AS. 1978. Spray drying of citrus juice. Paper presented at the seventh annual meeting of the American Institute of Chemical Engineers. Kilmartin PA, Reid DS, Samson I. 2004. Dielectric properties of frozen maltodextrin solutions with added NaCl across the glass transition. J Sci Food Agric 84:1277–84. Lievonen SM, Roos YH. 2003. Comparison of dielectric properties and non-enzymatic browning kinetics around glass transition. Innov Food Sci Emerg Technol 4:297–305.

Carrot Fiber as a Carrier in Spray Drying of Fructose

197

McCrum NG, Read BE, Williams G. 1991. Anelastic and dielectric effects in polymeric solids. New York: Dover. Nawirska A, Kwasniewska M. 2005. Dietary fibre fractions from fruit and vegetable processing waste. Food Chem 91:221–5. Newey C, Weaver G. 1991. Material principles and practice. London: Butterworth Scientific. Nyman M, Nylander T, Asp N-G. 1993. Degradation of water-soluble fibre polysaccharides in carrots after different types of processing. Food Chem 47:169–76. Park I-S, Kojima S. 2000. Glass transition and structural relaxation of intermediate liquids by MDSC and dielectric spectroscopy. Thermochim Acta 352:147–52. Ramulu P, Rao PU. 2003. Total, insoluble and soluble dietary fiber contents of Indian fruits. J Food Com Anal 16:677–85. Roos YH. 1995. Phase transitions in foods. San Diego, CA: Academic. Roos YH. 2002. Importance of glass transition and water activity to spray drying and stability of dairy powders. Lait 82:475–84. Roos YH, Karel M. 1991. Plasticizing effect of water on thermal behavior and crystallization of amorphous food models. J Food Sci 56:38–43. Roos YH, Karel M, Kokini JL. 1996. Glass transitions in low moisture and frozen foods: effects on shelf life and quality. Food Technol 50:95–108. Salmen L, Back E, Alwarsdotter Y. 1984. Effects of non-aqueous plasticizers on the thermal softening of paper. J Wood Chem Technol 4:347–56.

Session 3 Microstructured and Nanostructured Changes in Food

Invited Speakers

16 Taking the Measure of Water D. S. Reid

Abstract Water is ubiquitous in food systems. Its presence has a significant influence upon the properties of the systems. To understand the roles of water, it is necessary to have some metric to define the contribution from water. The challenge is to select the correct metric. The selection will depend upon the particular objective. The simplest metric is the gravimetric determination of actual water content. This is a necessary measure to allow for the description of system composition and is an important metric for many trade purposes. However, it provides little insight into the role of water. A metric of more relevance to the role of water is water activity, which describes the thermodynamic state of the water and can be correlated to the behavioral properties of many systems. However, this is an equilibrium measure and is of limited utility when attempting to understand many kinetic properties. For this, molecular mobilities, both of the water and of other system components, provide a clearer insight into system behavior. The issues inherent in the measurement of the influence of water are discussed from the aforementioned perspectives.

Introduction Water is the solvent of life. Not surprisingly, therefore, having some measurement that can quantify the influence of water within a system is important to the characterization of food systems. Since water has been the central theme of ISOPOW meetings since the first ISOPOW held in Glasgow, it seems appropriate for this, the Tenth International Symposium on the Properties of Water (ISOPOW 10), to discuss the issues that pertain to the measurement of the influence of water. My aim in this report is to provide a perspective of the many issues that arise when one decides to measure the water in a food system, and to discuss insights that might arise from the different metrics that we have available to us to quantify the influence of water upon a system. If we are to describe or quantify the role of water in a system, many factors have to be taken into consideration, and it is clear that no single metric will be a sufficient descriptor. Consider the kinds of factors we should expect to be relevant to the task. First we have the state of the water itself. Is it in the solid state, the liquid state, the gaseous state, or in some complex combination of states? Even within the three states, there will be further subclassifications to consider. Then the temperature of the system should be taken into account. In particular, we must be aware of the effects of 203

204

PART 1: Invited Speakers and Oral Presentations

temperature upon the states of the system, and also the effects of temperature upon the rates of processes, and the positions of equilibria. It is often convenient to consider separately frozen systems, where at least some of the water is in the solid state, and ambient systems, where the water is in more mobile states. Also, it can be helpful to consider the effects of temperature upon equilibria and thermodynamic properties separately from the effects of temperature upon reaction rates and molecular mobility. In this presentation, I consider the aforementioned factors identified in some more detail and their relevance to the task of providing useful metrics that will help quantify the contribution of water to the overall behavioral properties of food systems, and I attempt to provide a framework within which we can reach some understanding of the dynamics of systems.

States of Water Solid State When we describe a system as being in the solid state, we are describing a system in which molecular motions are very much constrained, primarily to vibrations and rotations about a defined central location, with a very limited potential for molecular translations. Only in the liquid and gaseous states does molecular translation become a major factor in physical properties. Three important scenarios place water in the solid state. The first, and most familiar, is water as ice, the crystalline solid phase of water in which the water molecules are arranged in a regular crystalline lattice. Second, we have the crystal hydrates; for example, the salt hydrates and organic hydrates. In these, the crystal lattice is usually that of the salt or the organic compound. An important subclass of the hydrates is that of the clathrate hydrates, in which the water lattice is the predominant crystal lattice. The difference between the ice lattice and the clathrate lattice, while spatially distinctive, results from a small change in the orientation of the water-water hydrogen bonds of the crystal lattices. In the ice crystal lattice shown in Figure 16.1 the hydrogen-bond geometry is staggered, as illustrated in Figure 16.3, whereas in the clathrate hydrate geometry, shown in Figure 16.2, the bond is eclipsed, as in Figure 16.3. The cagelike lattice of the clathrate hydrate is slightly less stable than that of regular ice, but stabilization can be achieved by the presence of the guest molecule in the cavities of the clathrate hydrate framework. With such stabilization, clathrate hydrates can be stable above 273 K. The third scenario is that of the aqueous glass, in which the structure is that of a liquid solution, but the molecular mobility has been constrained such that translational motions of the critical solute component have all but ceased. Though the water molecules in the system retain significant translational mobility (in contrast to the situation in ice or in clathrate hydrate), the lack of mobility of the solute results in the system having the overall behavior characteristic of a solid. This is discussed at more length in the section, “Molecular Mobility Approach”.

Taking the Measure of Water

205

Figure 16.1. Schematic of ice crystal lattice. Only water oxygen atoms are shown, linked by hydrogen bonds.

Figure 16.2. Schematic of a typical clathrate cage lattice. Only oxygen atoms are shown, linked by hydrogen bonds.

Liquid State In the liquid state, molecules are in general motion, but the overall volume of the system is defined. Vibration, rotation, and translation are all possible at significant rates. For aqueous systems, two categories are of particular interest: the dilute solution, where water is in considerable excess compared to the solute; and the concentrated solution, where there is much less water relative to solute. In the dilute solution, solute-solute

206

PART 1: Invited Speakers and Oral Presentations

Figure 16.3. Comparison of the orientation about the oxygen-oxygen hydrogen bond for ice (staggered) and clathrate (eclipsed) conformations.

interactions are small, and the properties of the solution can be largely predicted from theory. As the concentration increases, solute-solute interactions become more important, and adjusting factors have to be included into the dilute solution models. As concentrations increase still further, the dilute solution theories break down completely as solute-solute interactions become the primary factors determining system behavior, and in this state, empirical measurements, which may then be correlated to provide useful behavioral models, are required. As computational power increases, it is becoming feasible to model the behavior of such systems by ascribing appropriate interaction potentials to a molecular assembly. Vapor State In the vapor state, molecular motions are largely unconstrained, and the volume of the system is defined by its environment. In this state, intermolecular interactions are much reduced as compared to the condensed phases represented by the liquid and solid states. It can often be appropriate to characterize water by characterizing it in the vapor state.

Metrics for Water It is now appropriate to return to the initial question: How should we measure water? This requires consideration of a second question: Does the answer depend to any extent upon the state(s) of water present in the system?

Taking the Measure of Water

207

Two of the options available for providing a metric are (a) to describe the amount of water, or (b) to describe the effect of the water. Option b has two subsets to consider: (1) to determine and report thermodynamic parameters that describe the system, or (2) to determine and report the kinetic properties of the system, reporting on either the solvent or the solutes. We consider these options in turn. Amount of Water This is probably the oldest, and longest established measure of water. It is often a legal requirement. It has the easiest definition, though making measurements in accordance with the definition can be challenging. The simplest approach to the measurement would appear to be to make a direct determination by weight. To achieve this, one might determine the initial product weight and then apply drying conditions and determine the final product weight. The challenge comes in the drying step. First, how to be sure all of the water has been removed in the drying step, and second, how to be sure that only the water has been removed in the drying step? Oven drying is a technology commonly used in such procedures, introducing a further complication in that, due to heating, there may be chemical degradation of the material under study. Vacuum oven drying (a variant of regular oven drying), where the thermal degradation can be minimized, might reduce thermal degradation of the sample. Given these sources of uncertainty associated with direct weighing of the sample, as an alternative to direct weighing, water might be determined by chemical titration, using suitable chemistry by which water can be quantitatively reacted. Here a challenge is to ensure that all of the water within the sample has been extracted into the solvent system within which the reaction will be carried out. Another challenge is to ensure that no extraneous water is picked up from other sources. The most common chemistry used for this type of determination is that of Karl Fischer titration. A number of extant methodologies have been published for quantitation of water by using Karl Fischer titration. As an alternative to the use of chemical titration to quantify the amount of water extracted by the solvent, the water content of the extractant solvent can also be estimated using chromatographic techniques, where the size of a water peak can be related to sample water content. A third approach to determining the water content of a sample is to correlate an easily measured system property with system water content through standardized calibration curves. Here the primary sources of error are the quality of the calibration curve, and the assumption that only the change in water content is responsible for the property change being used in the correlation. An example of such a correlation is the linking of either refractive indices or specific gravities with the concentration of carbohydrate solutions (°Brix), where the assumption is that all carbohydrates have a similar effect on refractive index or specific gravity as does the standard solute, sucrose. It also assumes that no other solutes are contributing to the measured property. Isengard (1995, 2001) provides more detailed consideration of issues associated with the use of water content as a metric for water in foods, and a discussion of some analytic methodologies is presented by Reid and others (2005b).

208

PART 1: Invited Speakers and Oral Presentations

While water content is an important measure, the significance of any particular value for water content is system dependent. Consequently, measurement of water content provides little insight as to the sources of product stability or instability. To obtain such insight, we must approach the metric from a more physicochemical perspective. Two key themes in physical chemistry are thermodynamics, which describes the equilibrium states of systems and whose parameters are path independent; and kinetics, which describes the time course of the properties of systems and whose parameters are path dependent. Both thermodynamic and kinetic properties are influenced by temperature. It is important to note that, as a consequence of their pathindependent nature, thermodynamic properties can be predicted and calculated from other known properties, whereas kinetic properties, which are path dependent, can be determined only by experiment. Thermodynamic Approach The equilibrium thermodynamic properties of a system are time independent and a function of variables such as temperature and concentration. In a multiphase system at equilibrium, it is important to remember that thermodynamics require that the chemical potential of any component be the same in all phases. Thus, to determine the chemical potential, or any related property, such as the water activity (aw), it is sufficient to measure the chemical potential in the most readily accessed phase. This requirement that chemical potential is the same in all phases has consequences. Take butter as an example. Although the water contents of the aqueous phase and the oil phase are very different, the chemical potential of the water, and hence the aw, is the same in both phases! This is more easily understood by realizing that a water-saturated oil phase would have an aw close to 1, and reflects the ratio of the actual water content of the oil phase compared to the water content of the oil phase at saturation. The simplest approach to the measurement of aw is to determine ( pw pw0 )T in the vapor phase, where pw is the partial pressure of water in the system and pw0 is the partial pressure above pure water at the same temperature, with the isothermal constraint indicated by the subscript T. Though, conceptually, activity is defined as a ratio of fugacities under isothermal conditions, in real systems under normal conditions the partial vapor pressure of water is effectively the same as the fugacity of water. The equilibrium relative humidity (ERH) is the same ratio, expressed as a percentage, and is often used as an alternative means of expressing aw. The following constraints must be borne in mind. First, aw is a thermodynamic parameter. As such, the system must be at equilibrium. Also, the concept requires an isothermal comparison. Although conceptually simple, the practical accomplishment of these constraining conditions can be very difficult, especially in food systems, which are seldom at equilibrium. As a practical matter, however, steady-state conditions can often be achieved. In this case, only a measure of the relative vapor pressure (RVP) is achieved. Although the definition of aw is for an isothermal comparison of pw to a standard, this does not require that the entire measurement system be isothermal. It requires only that the sample itself be in an isothermal environment of known temperature. As

Taking the Measure of Water

209

Figure 16.4. Schematic of the general configuration of a cell for the determination of sample pw (the partial pressure of water in the system) at a known temperature. pa, pb, and pc, partial vapor pressure of water above the sample, at the sensor, and in the enclosed cell, respectively; and Ta and Tb, temperature of the sample and the sensor, respectively.

long as vapor transfer is facilitated, the sample vapor pressure can be measured in another, linked location where the conditions may be more conducive to measurement. Stated differently, the product must be in an isothermal environment of known temperature, and there must be easy vapor transfer, with no pressure loss, between the headspace above the sample (at sample temperature) and the headspace above the sensor, which is also in a defined environment. The sensor enables measurement of pw, and the temperature of the sample defines ( pw0 )T. This principle is illustrated in Figure 16.4, in which, with well-designed vapor paths, conditions should exist where pa = pb = pc. A question yet to be addressed is how RVP can be estimated. As with water content, it can be measured directly, but this is challenging. More often, an indirect measure is used. Some system property that correlates with pw, or with pw pw0, is determined, and it is assumed that the vapor has the appropriate pw. Perhaps the longest established correlation that enables the estimation of pw is that of the dew point temperature of the vapor volume of the system, since this is the temperature at which the system pw will be the same as pw0. The value of pw0 at the actual sample temperature can

210

PART 1: Invited Speakers and Oral Presentations

Figure 16.5. Schematic of the temperature dependence of pw and p w0. pw is the partial pressure of water in the system, and p w0 is the partial pressure above pure water at the same temperature. DF/DE and DH/DG describe the ratios of line lengths on the diagram. D, E, etc., are marked points.

be obtained from tables. Probes are available that have properties that correlate with pw pw0 of the atmosphere to which they are exposed. Most aw meters use one approach or the other. This topic is discussed in more detail elsewhere (Reid and others 2005a; Reid 2007). Figure 16.5 illustrates how vapor pressure depends upon temperature. The two curves represent pure water and a solution (or food sample). For a sample at T2, aw is given by DH/DG. At T1, the aw of the same sample would be DF/DE. In this example, T1 is the dew point for the sample held at T2. Thus, in an apparatus with easy vapor transfer, for a sample at T2 vapor condensation would be seen on a cooled surface at T1 in vapor-path contact with the sample. It is critical that, whatever method is used to determine pw, the actual temperature of the sample be accurately known. An error of 2 K in identifying the actual sample temperature results in an error in the estimation of pw0 of around 10%, with a consequent error in aw of 10%. In foods, establishment of equilibrium can be a very difficult task, especially since many foods are nonequilibrium systems. Since it is relatively easy to establish steadystate conditions, reporting data as RVP makes it clear that the result refers to steady

Taking the Measure of Water

211

Figure 16.6. Schematic representation of a moisture-sorption isotherm. RVP, relative vapor pressure.

state, but that equilibrium has not been confirmed. It must be recognized, however, that much of the literature uses aw, and providing it is realized that often this is RVP, the value of the data remains limited, absent any attempt to derive related thermodynamic parameters. Sorption Isotherms The two metrics discussed so far can be combined in the moisture-sorption isotherm (MSI), which plots water content against RVP. A schematic MSI is shown in Figure 16.6. These data can be separately obtained, or a special apparatus can be constructed in which the changing sample weight is monitored under different RVP equilibration conditions. Typical MSI shapes have been categorized. Once again, the issue as to whether the data relate to steady state or to true equilibrium is important. In particular, it is important to know how fast the underlying absorbent system is changing, and to understand whether there might be irreversible alterations in the physical state of the sorbent. By means of the sorption isotherm, both metrics can be related, and valuable behavioral correlations can be established linking both RVP and water content to the observed chemical and microbial stability of many classes of foods. Clearly, change itself is an important parameter, and measures that relate to the kinetics of change are of value. Bearing in mind that kinetic properties are temperature, time, and path dependent, how is this best approached? A simple and highly successful approach has been to establish correlations between reaction kinetics and the details of the MSI.

212

PART 1: Invited Speakers and Oral Presentations

There is an extensive body of literature relating food stability to RVP (e.g., see Barbosa-Cánovas and others 2007; Labuza and Altunakar 2007; and Sherwin and Labuza 2006). Such measurements and correlations have been important themes in ISOPOW meetings from ISOPOW 1. The use of RVP as a control factor has proved very successful in ambient temperature systems, with RVP ranges identified that provide particular characteristics. Only when the stability of frozen foods is considered does the approach fail to provide a useful metric. Since ice is one phase in a frozen food, the presence of ice within the system controls the aw of the food. All frozen foods held at the same temperature have the same aw. Also, the aw of frozen food reduces as the temperature is reduced. The initial freezing point of any ambient food is defined by its aw and is the temperature at which the aw of ice is the same as that of the ambient food. How do we explain the differing stabilities of frozen foods? An approach that moves beyond simply considering aw and considers other factors that might influence the rates and extent of processes is necessary for an improved understanding of the behavior of frozen systems. Molecular Mobility Approach In addressing the problem of understanding the relative stabilities of frozen systems, Levine and Slade (1986) and Slade and Levine (1991, 1995) introduced what they termed the food polymer science approach, based on the success of a similar approach in the understanding of the physical properties of polymer systems. The essential component of this approach was the consideration of the mobilities of the constituent molecules of the system. In its application to frozen foods, and later to low-moisture foods, two separate approaches have developed: (a) the determination of the phase behavior of the system, and that determination’s interpretation in terms of molecular mobilities, and (b) the more direct spectroscopic investigation of molecular mobility. Each approach leads to important insights, and the combination enables a clearer appreciation of the role of molecular mobilities in the behavioral dynamics of systems and also on the influence of water on the mobilities of all components of a system. Phase and State Diagrams A phase diagram is an entirely thermodynamic concept, tracking the concentrations of phases that can coexist at equilibrium as a function of temperature or pressure. The dependence of each phase concentration upon temperature (or pressure) is represented by a line on the diagram. In food systems, we are primarily interested in the role of temperature. Extending the equilibrium concept to take into consideration kinetic constraints and the behavior of systems at or close to steady-state conditions, in the same way as aw was extended to RVP, the general concept of a phase diagram can be extended to that of a state diagram. In a state diagram, the concentration of one or more of the coexisting phases at a given temperature or pressure, rather than being defined by thermodynamics, may lie within a range of possible concentrations, with the exact concentration dependent upon the pathway by which the phase has been formed and evolved as the system has separated into distinguishable phases. Since the separation into phases of concentrations

Taking the Measure of Water

213

Figure 16.7. A schematic temperature-composition phase diagram for a binary aqueous system. E, eutectic point; TE, eutectic temperature; and TmL and TmS , melting temperatures of the liquid and the solid lines, respectively.

different from the overall system concentration requires molecular translations, the kinetics of molecular translation will influence the rates of removal of one component from another. Whereas in a phase diagram the phase compositions present as a function of temperature and are represented by lines, in a state diagram some of the phase compositions are more properly seen as zones displaying the possible range of composition. Should such phase compositions be represented in the diagram by a line, this line usually represents the limit condition of the process that is responsible for the formation of the phase of interest. Figure 16.7 shows a typical simple phase diagram for a binary aqueous system, and Figure 16.8 shows a state diagram that could correspond to the same system should solute crystallization be inhibited by kinetic constraints. In these diagrams, the solid lines represent phase compositions as a function of temperature. The dotted lines (Figure 16.8) indicate the temperatures at which particular events can be observed in a system as a function of overall system composition. We must first consider how such diagrams can be constructed and then consider their interpretation. This discussion is limited to the construction of temperature-concentration diagrams. Since the conversion of a system from liquid state to solid state is accompanied by changes in the properties of the system, a phase diagram or state diagram can be

214

PART 1: Invited Speakers and Oral Presentations

Figure 16.8. The schematic temperature-composition state diagram for the system shown in Figure 16.7, where solute crystallization is inhibited. E, eutectic point; L, liquid; S, solid; TE, eutectic temperature; Tg, glass transition temperature; and TmL and TmS , melting temperatures of the liquid and the solid lines, respectively.

constructed by determining the patterns of property change as a function of both temperature and concentration. The most common approaches are either (a) to prepare systems of selected overall concentration and then monitor the selected property as a function of temperature for each concentration or (b) to study the change in the properties of a system at several constant temperatures as the concentration of the system is changed; for example, by removing water from an initially dilute system. In either case, examination of the property curve as a function of the changing variable (Figure 16.9) may reveal points in the curve where there is a clear change in the overall trend. The coordinates of these points are plotted to generate the state diagram. The assignment of meaning to the various lines and curves that can be identified in the diagrams relies upon an appreciation of the various processes that might be reflected within the diagram. One of the more common means of generating a state diagram is to follow the thermal behavior of the system, often by measuring the temperature dependence of the heat capacity or by employing differential scanning calorimetry. The temperature dependence of mechanical properties can also suggest the location of phase transformations. The challenges involved in interpreting instrumental data to construct a phase or state diagram are discussed in more detail elsewhere (Reid 2002).

Taking the Measure of Water

(a)

215

(b)

Figure 16.9. Schematic of the change expected in a system property in a system undergoing (a) a first-order phase transition and (b) a second-order phase transition.

In the Figure 16.8 state diagram, the thermodynamic phase diagram is represented down to the temperature TE: that of the eutectic point. The curve that shows concentration increasing with increase in temperature is the solubility curve of the solute. This curve is relevant only should the solute crystallize and the systems achieve equilibrium. The curve to the left of the eutectic point, where concentration reduces with increasing temperature is the melting temperature TmL curve tracking the melting point of the system, when ice is the only crystalline phase. Should the solute fail to crystallize, this curve extends below TE. In the two-phase region LS 1 bounded by the left axis and the TmL curve, the two phases are pure ice, whose temperature-composition profile is a vertical line down the left axis, and a solution of increasing concentration as temperature reduces, represented by the TmL curve. At a temperature close to that of Tg′ (or Tm′ ), the concentration of this phase can no longer be increased by further cooling and consequent formation of additional ice. This is considered to be a consequence of the reduced mobility of the solute molecules. For additional ice to form, the growth site has to reject the solute in order to accept an additional water molecule. If the solute mobility is such that this rejection is kinetically constrained, no more ice can form, and the phase composition becomes fixed within the timescale of the constraint. The loss of mobility of the solute is reflected in changes in the thermal and mechanical properties of the phase. It is generally accepted that, at this Tg′ temperature, the concentrated phase changes from a rubbery, viscous but mobile state to a state of high effective viscosity. It is the temperature at which, on warming, ice can begin to melt and dilute the concentrated amorphous phase.

216

PART 1: Invited Speakers and Oral Presentations

At lower temperatures in the diagram, a curve is depicted where, as the temperature rises, the concentration increases. This curve tracks the assumed location of the glass transition of the homogeneous system formed by cooling, absent any separation of an ice phase at higher temperatures. The Tm curve intersects this curve at Tg*, the temperature of the glass transition of the maximally freeze-concentrated matrix, concentration Cg*. This concentration will be very close to the matrix concentration Cg′ associated with Tg′. Note, however, that Tg* is associated with the maximally freezeconcentrated matrix, a defined concentration (though practically perhaps a conceptual ideal, unreachable in practice, as is “absolute zero”). Tg′ describes the temperature and concentration achieved in an actual system. Depending on the procedures used to induce the phase separation, and the number of cooling and heating cycles employed, the observed Tg′ and Cg′ tend toward a limit. At this limit, Cg′ will be the closest approximation to Cg*. In systems that cannot form ice (initial concentration greater than Cg*), Tg increases as Cg increases. In the low-moisture region, there is a range of compositions where Tg is within the range of possible storage temperatures. In these low-moisture systems, the changes in molecular mobility as the system transits from rubbery (above Tg) to glassy (below Tg) are important for storage stability and system characteristics. Thus, in considering the state diagram for a food-type system, the regions where restricted mobilities of molecular components may be important for storage stability and for other product characteristics are in the top right quadrant of Figure 16.8, which is relevant to low-moisture ambient products, and in the low-temperature region to left of Cg*, which helps characterize frozen products in low-temperature storage. This topic is discussed in more detail elsewhere (Reid and Fennema 2008). There is an extensive literature addressing studies in each region. Spectroscopic Approach In contrast to the phase diagram approach, which infers relationships involving molecular mobilities from the pattern of phase relationships, and the effect of varied timetemperature protocols on the identity of coexisting phases, spectroscopy attempts to make a more direct determination of molecular mobility. Nuclear magnetic resonance (NMR) techniques are sensitive to rotational and translational motions of water. Infrared and Raman spectroscopy can measure vibrational motions. Solute-solid mobility can also be observed. As with all other approaches to the characterization of a water metric, there is a divide between what might be conceptually possible and the limitations of the available methodologies. The measurement of frequencies and amplitudes of signals is relatively straightforward in spectroscopy. The unambiguous assignment of these signals is the challenge. In aqueous systems, a primary interest is to try to quantify the fraction of water that contributes to each component of the spectroscopic signal. Once again, a significant challenge is identifying the exact condition of the sample and identifying the processes that are contributing to the signal. Schmidt (2007) provides a comprehensive review and discussion of the capabilities and pitfalls of the use of NMR to characterize water mobility (see also Schmidt 2004). In many systems, it is clear from spectro-

Taking the Measure of Water

217

scopic studies that in what might be termed gels, and in the amorphous glasses of frozen systems and low-moisture glasses, the mobility of the water, although reduced compared to dilute aqueous systems, is still extensive. The reduced mobility of the bulk phase has, therefore, to be a consequence of reduced mobility in the other components of the phase. Spectroscopy enables the study of the molecular mobility of the other components, and it can be confirmed that their reduced mobility is the source of the observed bulk properties.

Conclusion To properly quantify the contribution of water to overall system properties, all metrics should be used. The historical development of our understanding of water in foods shows the use of available metrics and, as their limitations became recognized, the introduction of new metrics to be employed in concert with those already in use. In the 34-year history of ISOPOW, this progress can be tracked in the evolution of the discussions reported in the conference proceedings from each meeting. ISOPOW 10 confirms that the state of our art in understanding the role of water continues to advance.

References Barbosa-Cánovas GV, Fontana AJ, Schmidt SJ, Labuza TP. 2007. Water activity in foods: fundamentals and applications. Ames, IA: Wiley-Blackwell. Isengard HD. 1995. Rapid water determination in foodstuffs. Trends Food Sci Technol 6:155–62. Isengard HD. 2001. Water content, one of the more important properties of food. Food Control 12:395–400. Labuza TP, Altunakar B. 2007. Water activity prediction and moisture sorption isotherms. In: BarbosaCánovas GV, Fontana AJ, Schmidt SJ, Labuza TP, editors. Water activity in foods: fundamentals and applications. Ames, IA: Wiley-Blackwell. p 109–54. Levine H, Slade L. 1986. A polymer physico-chemical approach to the study of commercial starch hydrolysis products (SHPs). Carbohydr Polym 6:213–44. Reid DS. 2002. Use, misuse and abuse of experimental approaches to studies of amorphous aqueous systems. In: Levine H, editor. Amorphous food and pharmaceutical systems. Cambridge: Royal Society of Chemistry. p 325–38. Reid DS. 2007. Water activity: fundamentals and relationships. In: Barbosa-Cánovas GV, Fontana AJ, Schmidt SJ, Labuza TP, editors. Water activity in foods: fundamentals and applications. Ames, IA: Wiley-Blackwell. p 15–28. Reid DS, Fennema OR. 2008. Water and ice. In: Damodaran S, Parkin KL, Fennema OT, editors. Fennema’s food chemistry. 4th ed. Boca Raton, FL: CRC. p 17–82. Reid DS, Fontana AJ, Sabbiani SS, Rahman SS, Labuza TP, Guiznani N, Lewicki P. 2005a. Vapor pressure measurement of water. In: Wrolstad RE, Acree TE, Decker EA, Penner MH, Reid DS, Schwartz SJ, Shoemaker CF, Smith DM, Sporns P, editors. Handbook of food analytical chemistry. New York: Wiley. p 35–70. Reid DS, Ruiz RP, Chinachoti P, Krygsman PH. 2005b. Gravimetric measurement of water. In: Wrolstad RE, Acree TE, Decker EA, Penner MH, Reid DS, Schwartz SJ, Shoemaker CF, Smith DM, Sporns P, editors. Handbook of food analytical chemistry. New York: Wiley. p 5–34. Schmidt SJ. 2004. Water and solids mobility in foods. Adv Food Nutr Res 48:1–101. Schmidt SJ. 2007. Water mobility in foods. In: Barbosa-Cánovas GV, Fontana AJ, Schmidt SJ, Labuza TP, editors. Water activity in foods: fundamentals and applications. Ames, IA: Wiley-Blackwell. p 47–108.

218

PART 1: Invited Speakers and Oral Presentations

Sherwin CP, Labuza TP. 2006. Beyond water activity and glass transition: a broad perspective on the manner by which water can influence reaction rates in foods. In: Buera MP, Welti-Chanes J, Lillford PJ, Corti HR, editors. Water properties of food, pharmaceutical, and biological materials. Boca Raton, FL: CRC, Taylor and Francis. p 343–62. Slade L, Levine H. 1991. Beyond water activity: recent advances based on an alternative approach to the assessment of food quality and safety. Crit Rev Food Sci 30:115–360. Slade L, Levine H. 1995. Polymer science approach to water relationships in foods. In: Barbosa-Cánovas GV, Welti-Chanes J, editors. Food preservation by moisture control: fundamentals and applications. Lancaster, PA: Technomic. p 33–133.

17 Rehydration Modeling of Food Particulates by Using Principles of Water Transport in Porous Media I. S. Saguy, O. Troygot, A. Marabi, and R. Wallach

Abstract Water-imbibition theory has been shown to have a multidisciplinary validity and has applicability in modeling of the rehydration of dried porous food. Basic theory of flow in porous media has been widely used in several domains and promoted interdisciplinary research, and is also used as a bridge between food science and soil physics. Analogous to a food-sorption isotherm, the water-characteristic curve of a porous medium describes the functional relationship between water content and matric potential under equilibrium conditions. The use of porous-media theory for modeling of food rehydration requires employment of a known characteristic curve, the determination of which is time-consuming and cumbersome. Using watercharacteristic curves employed in soil physics to bridge between sorption isotherms commonly used in food science could furnish a novel and integrated approach necessary to overcome some of the immense complexity that has hampered previous modeling attempts and to open new avenues for studying and optimizing food rehydration.

Introduction In the last decade, there has been a continuous increase in the demand for convenience foods, including dehydrated products, mainly because of modern lifestyles. This trend has been accompanied by a decrease in the ability, desire, or time needed to prepare food and an increase in financial means, leading consumers to choose foods that are readily available and convenient and that require only minimal or no preparation before consumption (Tillotson 2003). The rehydration of dried foods is a fundamental unit operation in the food industry. The quality of the rehydrated and reconstituted products is affected by the drying conditions and rehydration processes used, ultimately influencing consumer acceptance. During the drying process, irreversible physicochemical changes occur—including textural and structural modifications, migration of solutes, and loss of volatiles and nutrients—that affect the quality of the final product. Therefore, the drying process needs to be understood and controlled to create a dried product with optimal nutritional, sensory, and rehydration characteristics. During rehydration, in addition to medium uptake, solids leach from the food product into the medium. Hence, the use of nondissolvable solids (NDS) was 219

220

PART 1: Invited Speakers and Oral Presentations

proposed (Marabi and others 2004a) and used in the determination of the rehydration ratio (RR): RR =

Wt − W0 We − W0

(17.1)

where W0, Wt, and We are g H2O per g NDS at times t0, t, and t∞ (s), respectively.

Mathematical Modeling In recent years, the food domain has experienced an encouraging transition from empirically based to physically based models. Nevertheless, most rehydration studies reported in the literature still include empirical and semiempirical models to describe the mechanisms of liquid uptake, solid leaching, and the kinetics of the processes (Marabi and Saguy 2007). Mathematical modeling facilitates understanding of process characteristics, provides insight into the governing mechanisms, and could be used to improve the process resulting in better products. Liquid uptake is typically modeled by applying a mechanistic approach or employing an empirical approach. Fick’s second law of diffusion is a typical example of a mechanistic approach, whereas the first-order model is merely a curve-fitting empirical model. The development of empirical models requires considerably less effort and therefore is used frequently. However, empirical models are limited and are nontransferable (Saguy and others 2005a).

Empirical and Semiempirical Models Five models that are often used are the exponential model (Equation 17.2) (Misra and Brooker 1980), Peleg’s model (Equation 17.3) (e.g., see Peleg 1988; GarciaPascual and others 2006; Giraldo and others 2006; Cunningham and others 2007), first-order kinetics (Equation 17.4) (Krokida and Philippopoulos 2005; Gowen and others 2007); the Weibull distribution function (Equation 17.5) (e.g., see Cunha and others 1998a, 1998b; Garcia-Pascual and others 2006; Cunningham and others 2007), and the normalized Weibull distribution function (Equation 17.6) (e.g., see Marabi and others 2003, 2004a, 2004b; Marabi and Saguy 2004, 2005) (Table 17.1). Among the empirical models, the Weibull distribution function is frequently used and has recently been improved to describe the rehydration of dried foods. It is important to note, however, that all of these empirical models provide an excellent basis for curve fitting and enable the process to be represented as a function of physical properties and conditions of both the rehydration medium and the rehydrating food particles. However, they offer rather limited insight into the fundamental principles involved, hindering an understanding of the transport mechanism(s) actually taking place (Marabi and Saguy 2007).

Using Principles of Water Transport in Porous Media

221

Table 17.1. Empirical models frequently used in curve fitting of rehydration data Exponential model

Mt − Me = exp ( − P1t P2 ) M 0 − Me

(Eq. 17.2)

Peleg’s model

Mt = M0 +

t ( P3 + P4 × t )

(Eq. 17.3)

First-order kinetics

Mt − Me = exp ( − kt ) M0 − Me

Weibull distribution function

⎡ t ⎤ Mt = 1 − exp ⎢ − ⎛ ⎞ ⎥ Me ⎣ ⎝ α⎠ ⎦

Normalized Weibull distribution function

⎡ t × Deff × Rg ⎞ ⎤ Wt − W0 = 1 − exp ⎢ − ⎛⎜ ⎠⎟ ⎥⎦ We − W0 L2 ⎣ ⎝

(Eq. 17.4) β

(Eq. 17.5) β

(Eq. 17.6)

Deff, effective diffusion coefficient (m2/s); L, half-slab thickness, or radius for spherical and cylindrical samples (m); M0, Mt, and Me, moisture content at times 0 and t and at equilibrium, respectively (kg H2O/kg dry solids); Ms, surface moisture content (kg H2O/kg DS); P1 and P2, parameters in the thin-layer rewetting equation; P3 and P4, constants related to the initial rate of sorption and to the equilibrium moisture content, respectively (kg D.S./kg H2O); k, rehydration rate (min−1); Rg, geometry factor (–); S, surface area (m2); t, time (s); V, volume (m3); W0, Wt, and We, moisture content at time 0 and t and at equilibrium, respectively (kg H2O/kg nondissolvable solids); α, scale parameter (s); and β, Weibull shape parameter (–). Adapted from Marabi and Saguy (2007).

The Diffusion Model The diffusion model is a combination of physical and empirical approaches founded on Fick’s first and second laws: J x = −D

dW dx

∂W ∂ 2W =D 2 ∂t ∂x

(17.7) (17.8)

where Jx is the flux in x direction (kg H2O/m2 s), W is the moisture content (kg H2O/ m3), and x is the spatial coordinate (m). To solve Fick’s second law, the following assumptions and simplifications are typically made: the process is controlled by moisture concentration gradient; no other transport mechanism is active; the diffusion coefficient is independent of moisture concentration; the matrix is uniform and isotropic; the distribution of initial moisture content is uniform; the surface attains saturation moisture instantly; the boundarylayer resistance is much smaller than the internal resistance; shrinkage or swelling during the process is negligible; the sample is generally considered as a sphere, cylinder, or slab; and heat generation/absorption and transfer are ignored. The effective diffusion coefficient, which is readily derived from experimental data, is an apparent value that encompasses all intrinsic hydraulic properties of the particles.

222

PART 1: Invited Speakers and Oral Presentations

Also, little effort has made to resolve microstructural aspects or study the mechanisms of flow in the porous particulates. Deff is defined (Gekas 1992) in Equation 17.9: Deff =

εD τ

(17.9)

where ε is the porosity (−) and τ is the tortuosity (−). Typical Deff values for moisture in foods range from 10−8 to 10−12 m2/s (Marabi and Saguy 2007), generally being closer to 10−10 m2/s (e.g., see Maroulis and others 2001), and is influenced by temperature, water content, pressure, physical properties, and the dried food’s structure. However, it is well documented that some or most of the aforementioned assumptions used in solving Fick’s laws are invalid. Moreover, mechanisms of mass transfer other than molecular or Fickian diffusion may also occur. Despite these drawbacks, the Fickian approach is frequently used to model a wide range of transport phenomena in foods (e.g., see Garcia-Pascual and others 2006; Cunningham and others 2007; Falade and Abbo 2007). These phenomena include both the transport of liquids in solids and the movement of solutes in the liquid phase. Modeling of the rehydration process by means of Fick’s laws of diffusion provides some insight into the rehydration process and enables the comparison of how different process parameters affect the overall transport phenomena. The use of Fick’s laws, despite many drawbacks, enables a practical assessment of the rehydration process. It also enables the characterization of different food particulates and drying methods. Therefore, their use will probably continue (a semimechanistic approach). However, since it is a simplified model, possible pitfalls should be carefully considered.

Paradigm Shift As most empirical models are primarily used to avoid more complex considerations and other complications, such as changes occurring within the product, the actual microstructure is not considered. The alternative is to apply a physically based approach that uses porous media in which the actual microstructure and void channels are important. This may sound straightforward but requires a paradigm shift and an interdisciplinary approach (Figure 17.1). This approach is based on integrating the know-how developed in other domains, such soil science, to study porous media and fluid transport in foods. Capillary Flow in Porous Media Rehydration studies of dried plant tissues reveal a very complex phenomenon involving different transport mechanisms, including molecular diffusion, convection, hydraulic flow, and capillary flow (Saravacos and Maroulis 2001). Several authors have also suggested that liquid enters a dried food sample by various mechanisms in tandem (Oliveira and Ilincanu 1999; Chiralt and Fito 2003; Marabi and others 2003; Marabi and Saguy 2005). The contribution of mechanisms involving mass flux due to a temperature gradient (Soret effect) is considered insignificant and is therefore often disregarded (Datta 2007a). Several studies using the Washburn equation to

Using Principles of Water Transport in Porous Media

Porous media Fluid transport

Food science

Food models & analyses

Soil science

Physically based models

Food, Earth, & mineral sciences

223

Microstructure

Figure 17.1. Interdisciplinary approaches to rehydration modeling.

represent the movement of liquids into porous food matrices are reviewed in the next paragraph. Some of the most common models based on the Washburn equation that have been applied in food science are described elsewhere (Marabi and Saguy 2007). To study the rehydration process further without having to rely on the common postulation that it is governed by Fickian diffusion, numerous studies have investigated the use of mathematical models based on capillary flow in porous media to describe the data obtained experimentally. These models are derived from the wellknown Lucas (1918) and Washburn (1921) equations (also referred to together as the Lucas-Washburn or Washburn-Rideal equation [Rideal 1922]). Porous media were represented by a bundle of capillaries with similar or different radii (simplest case), and water retention in the capillary is expressed by integrating the Laplace equation: P=

2γ cos θangle r

(17.10)

where P is the capillary pressure (Pa), γ is the surface tension (J m−2), θangle is the contact angle (°), and r is the capillary radius (m). With the Poiseuille equation for laminar flow: ΔP =

8QηL πr 4

(17.11)

where ΔP is the pressure loss due to friction (Pa), Q is the volumetric flow (m3), η is the dynamic viscosity (Pa s), and L is the length (m). The Lucas-Washburn equation (simplified assuming no gravity) can be expressed as shown in Equation 17.12: ⎛ r γ cos θangle ⎞ x2 = ⎜ ⎟⎠ t ⎝ 2η where x is the distance traveled by the liquid front (m).

(17.12)

224

PART 1: Invited Speakers and Oral Presentations

Capillary flow in porous media could be therefore expressed as shown in Equations 17.13 and 17.14 (Benavente and others 2002): W(t ) = Ctheor t S Ctheor = ερ

δγr cos θangle 2 τη

(17.13) (17.14)

where Ctheor is the uptake coefficient (kg/[m2 s0.5]), ρ is the density (kg/m3), S is the area (m2), W(t) is the weight gain (kg) at time t (s), and δ is the roundness (−). The value of Ctheor is derived from the model and the physical properties. The experimental coefficient, Cexp, is calculated from the initial slope obtained from plotting W(t)/S vs t . Following the imbibition of water by dried porous foods, it was shown that the process followed the aforementioned Washburn equation (Lee and others 2005; Saguy and others 2005b). However, discrepancies related to the use of a single “effective” cylindrical capillary and a constant contact angle were also reported (Saguy and others 2005b). An additional factor that may be responsible for inaccuracies in comparing experimental data with the Washburn equation may be related to the tortuosity of the pores within the food sample. The pore network is often regarded as a bundle of cylindrical and straight capillaries with a determined effective radius. Thus, the Washburn equation may be used in its original form (Aguilera and others 2004) or otherwise corrected with a tortuosity factor (e.g., Equation 17.14; see Carbonell and others 2004). A few other studies have arrived at similar conclusions with regard to the mechanisms of water movement during various common processes in the food industry. For instance, it was proposed that, during vacuum osmotic drying, the water transfer results from a combination of traditional Fickian diffusion and vacuum capillary flow, especially during the first few hours. The capillary-flow function was proposed to be closely related to the porosity of the apricot fruit being tested (Shi and Fito-Maupoey 1994). When the rehydration of pasta was studied by means of a radial nuclear magnetic resonance (NMR) microimaging technique, the researchers concluded that the rehydration was a non-Fickian process (Hills and others 1996). Transport mechanisms other than diffusion were reported during osmotic dehydration of apples. The proposed alternatives included capillary penetration or another fast-transport mechanism occurring near the interface of the samples (Salvatori and others 1999).

Flow in Unsaturated Porous Media Capillarity and Tension Head Water that has entered a porous medium but has not drained out of the sample will be retained within pores by capillary forces or will surround the surface of the medium particles by the molecular forces of adhesion and cohesion. Therefore, a simple measurement of the water content in the medium is insufficient to enumerate the complete

Using Principles of Water Transport in Porous Media

225

status of medium’s water. While the quantity of water present in the medium is very important, the potential or affinity with which the water is retained by the medium is perhaps more important, mainly if the dynamics of this water is considered. This potential may be defined as the amount of work done or potential energy stored, per unit volume, in moving a mass of water from the reference state (typically chosen as pure free water). In this manner, one may think of matric potential as potential energy per unit volume (J m−3), which is also expressed as Pascal (Pa, i.e., the force per unit area or pressure). This explains the use of the term pressure potential and is the reason why soil physicists refer to matric potential as soil pressure or if it is divided by the bulk density, as a pressure head (Wallach 2007). Capillarity refers to both the macroscopic and the statistical behavior of interfaces rather than their molecular structure. Capillarity is also referred to as surface tension (work per unit area expressed as J m−2). This phenomenon is extremely important in water retention in porous media. Surface tension occurs at the molecular level and involves two types of molecular forces: adhesive forces, which are the attractive forces of molecules of dissimilar substances, and cohesive forces, which are the attractions between molecules in similar substances. Cohesive forces decrease rapidly with distance and are strongest in solids, weaker in liquids, and weakest in gases. As water rises in a capillary, the meniscus grows spherical and concaves upward. By letting r equal the tube radius, the excess pressure above the meniscus compared to the pressure directly below it can be described under various assumptions by 2γ/r. As the pressure on the water surface outside the capillary tube is atmospheric, the pressure in the liquid below the meniscus will be less than the atmospheric pressure above the meniscus by 2γ/r. This will force the fluid up the tube until the hydrostatic pressure of the fluid column within the tube equals the excess pressure of 2γ/r. Because the circumference of the tube is 2πr, the total force (upward) on the fluid is given by Equation 17.10. This force supports the weight of the fluid column to the height (hc). The height of capillary rise can be expressed as in Equation 17.15 (Laplace equation): hc =

2γ cos θangle ρgr

(17.15)

where hc is the capillary rise (water tension head) (m), and g is the acceleration due to gravity (m/s2). Water-Retention Curve The water-characteristic curve of porous media is also known as the water-retention curve (RC) and describes the functional relationship between the volumetric water content (θv) and matric potential (ψ) under equilibrium conditions. The matric potential is usually replaced by the pressure potential head, h (m), which is the energy per unit weight of water. As the water in the unsaturated porous media is at subatmospheric pressure, the pressure potential head is commonly called the tension head. This curve is an important property related to the distribution of pore space (sizes,

226

PART 1: Invited Speakers and Oral Presentations

interconnectedness), which is strongly affected by texture and structure, as well as related factors, including organic matter content. The RC indicates the amount of water in a porous medium at a given tension head. It is also a primary hydraulic property required for modeling water flow in porous media. RCs are highly nonlinear functions and relatively difficult to determine accurately (Hillel 1998; Wallach 2007). The traditional method of determining the water retention involves establishing a series of equilibria between water in the porous medium sample and a body of water at known water tensions. The medium-water system is in hydraulic contact with the body of water via a water-wetted porous plate or membrane. At each equilibrium, the volumetric water content of the medium is determined and paired with a value of the tension head (h), determined from the pressure in the body of water and the gasphase pressure in the substrate. The data pair (θv, h) forms one point on the RC. A summary of methods used in soil science to determine the RC can be found (e.g., see Klute 1986). It is worth noting that RC quantification resembles sorption isotherm determination in foods. Measured values of water content and tension head (θv–h) are often fragmentary and usually based on relatively few measurements over the wetness range of interest. For modeling and analysis purposes and for the characterization and comparison of different substrates and scenarios, it is essential to represent the RC in continuous and parametric form. A parametric expression of the RC model should contain as few parameters as possible to simplify its estimation and describe the behavior at the limits (wet and dry ends) while closely fitting the nonlinear shape of the θv–h data. Many models were suggested to describe RC in soil science (Jury and Gardner 1991; Hillel 1998). The most frequently used are the Brooks and Corey (1966) (B-C) and van Genuchten (1980) (VG) models. The model parameters are typically determined by curve fitting of experimental data. The B-C model is as shown in Equation 17.16: λ

⎧ θV − θr ⎛ h⎞ ⎪⎪ θ − θ = Se = ⎜⎝ h ⎟⎠ s r b ⎨ θ θ − r ⎪ V = Se = 1 ⎪⎩ θs − θ r 2 + 2 ( mv + τ ) ⎧ K ( h ) = K s ⋅ Se ⎨ ⎩ K (h ) = Ks

h > hb h ≤ hb h > hb h ≤ hb

(17.16)

where Se is the effective saturation (−, which means dimensionless); hb is the air entry value (m); λ is the pore size index empirically determined parameter (−); θv, θr, and θs are the current, residual, and saturated volumetric water contents, respectively (m3/m3); Ks is the saturated hydraulic conductivity (m/s); and mv is an empirical parameter (−). The residual water content is somewhat arbitrarily defined as the water content at which the corresponding hydraulic conductivity is essentially zero, but very often it is used as an empirical constant when fitting hydraulic functions. As opposed to θs,

Using Principles of Water Transport in Porous Media

227

which has a clear physical significance, the meaning of θr and its estimation have not yet been resolved (Wallach 2007). When θr = 0, Se approaches S. Note that Se varies between 0 and 1. The other common parametric model for relating water content to the tension head was proposed by van Genuchten (1980): Se =

θV − θr ⎡ 1 ⎤ ⎢ n ⎥ θs − θ r ⎣ 1 + ( α h ⋅ h ) ⎦

mv

(17.17)

where αh (m−1) and mv (−) are empirical parameters (determining the shape of the RC and are derived by curve-fitting techniques). The value of the hydraulic conductivity of a saturated porous medium (Ks) depends on the properties of the medium and the flowing fluid. This dependence can be separated into two factors: fluidity f (defined as ρg/η) and intrinsic permeability k (m2): Ks = kf =

kρg η

(17.18)

The intrinsic permeability of a medium is a function of pore structure and geometry. Particles of smaller individual grains have a larger specific surface area, increasing the drag on water molecules that flow through the medium, which results in a reduced intrinsic permeability and Ks. Richards Equation, Boundary, and Initial Conditions The Darcy law describes the flow equation in saturated media: J = − Ks

ΔH Δs

(17.19)

where H is the hydraulic head (m), s is the distance along a stream line in the flow field (m), ΔH/Δs is the hydraulic-head gradient along the stream line (−), and J is the flux density or flow per unit area opposite to the direction defined by hydraulic-head gradient (m/s). The Darcy law can be coupled with the conservation of mass principles to derive a continuity equation. Using one horizontal dimension leads to Equation 17.20: ∂θ V ∂ ⎡ ∂h = K (h ) ⎤ ⎢ ∂t ∂x ⎣ ∂x ⎥⎦

(17.20)

Equation 17.20 contains two unknowns, namely, θv and h, and applying the chain rule for the left-hand side of Equation 17.20 provides the h-based continuity equation: C (h )

∂h ∂ ⎡ ∂h = K (h ) ⎤ ⎢ ∂t ∂x ⎣ ∂x ⎥⎦

(17.21)

228

PART 1: Invited Speakers and Oral Presentations

where C(h) = δθv/δh is the slope of the RC and is called the specific moisture capacity (1/m). Applying the chain rule for the derivative in the right-hand side of Equation 17.20 provides the θv-based continuity equation, which is a diffusion-type (nonlinear) equation: ∂ ⎛ ∂θ V ∂θ Dh ( θ V ) V ⎞ = ∂x ⎝ ∂t ∂x ⎠

(17.22)

where Dh = K(h) · δh/δθv (m2/s) is denoted as hydraulic diffusivity and is the ratio of the unsaturated hydraulic conductivity to the specific moisture capacity. The advantage of the θv-based form is that Dh does not vary with θv nearly as much as K varies with h. The last two equations, along with their various alternative formulations, combined are known as the Richards equation. Developing the Theory of Flow in Porous Media To use the theory of flow in porous media in rehydration of dry food particulates, the RC of these particulates is needed. The B-C, VG, and other functions used to model the RCs have significant limitations at low liquid saturations in that they use a parameter called residual saturation, which is the minimum liquid saturation calculated by these equations. At low water content, vapor diffusion is predominantly driven by vapor-pressure differences, which in turn are driven by local vapor-pressure lowering (related to capillary pressure through the Kelvin equation). Hence, vapor flow is significantly influenced by the RC at low water contents. The limitations of the two-phase characteristic curves in the low-water-content region leads to errors in liquid and vapor transport and in the prediction of the actual water content in this region. This value is important in the transition from liquid-solid sorption to vapor-solid sorption, which is typically at 4–6 monomolecular layers, or a liquid saturation of 0.11–0.16 (m3/m3) based on a soil surface area of 25 m2/g (Webb 2000). At higher liquid saturations, adsorption is primarily by liquid-solid processes. As the liquid saturation decreases, however, vapor-solid sorption contributes more and more to the adsorption process and may dominate at low levels of liquid saturation. Physically based modification of the RC has been suggested by a number of authors (e.g., see Ross and others 1991; Campbell and Shiozawa 1992; Rossi and Nimmo 1994; Morel-Seytoux and Nimmo 1999). Campbell and Shiozawa (1992) observed that the capillary pressure in the dry region is a linear function of liquid saturation on a semilog plot. They used this relationship and added a VG water-content relationship to derive the full capillary pressure curve, where the parameters in the VG equation were refit to the data, assuming zero liquid residual saturation. Subsequent models have all used the linear relationship observed by Campbell and Shiozawa (1992). The model proposed by Rossi and Nimmo (1994) is based on the B-C model, with zero residual saturation, which is equivalent to Campbell’s (1974) expression, with the dry-region function proposed by Campbell and Shiozawa (1992) in the low-watercontent region. Morel-Seytoux and Nimmo (1999) slightly modified the Rossi and

Using Principles of Water Transport in Porous Media

229

Nimmo (1994) model. Webb (2000) presented a similar approach to the dry-region modification of Morel-Seytoux and Nimmo (1999) in that it merged an existing RC with a dry-region expression. However, the approach by Morel-Seytoux and Nimmo (1999) requires the simultaneous solution of two nonlinear equations, whereas the Webb (2000) approach requires only a few iterations on a single equation. Following these approaches, we suggest a new method to predict the RCs of porous media (food particulates, in particular) from measured water-sorption isotherms. This method is based on a four-step procedure (Troygot and others 2007): 1. Water-Sorption Isotherm The suggested method adopts the fundamental water-isotherm approach to describe the relationship between water content and water activity (aw) for the dry zone. This selection is straightforward and takes into consideration the knowledge bank on the relationship between food stability and aw. Although the use of glass transition theory could also be considered, in this, however, the relationship between water and the gas phase is important for porous media. Water activity is converted to tension head by using the Kelvin equation: h=

RT ln ( aw ) ρw gM w

(17.23)

where R is the gas constant (J · K−1 · mol−1), Mw is the molecular water mass (0.018 kg/ mol), T is the absolute temperature (K), ρw is the water density (1000 kg/m3); and aw is the water activity (−). However, the theoretical aspects of Kelvin equation at low aw is still under development, and further studies are required. A typical water-sorption isotherm is depicted in Figure 17.2 for microcrystalline cellulose (MCC). Water content (g/100 g DS)

20 16

Experimental data GAB

12 8 4 0 0.0

0.2

0.4

0.6

0.8

1.0

Water activity

Figure 17.2. Typical water-sorption isotherm for microcrystalline cellulose at 25°C. DS, dry solids; and GAB, Guggenheim-Anderson-de Boer. Adapted from Troygot and others (2007).

230

PART 1: Invited Speakers and Oral Presentations

1.E+7 Prediction

Water tension (cm)

1.E+6

Sorption data Retention curve

1.E+5

Exponential sorption curve

1.E+4 1.E+3 1.E+2 1.E+1 1.E+0

0.0

0.2

0.4

0.6 3

0.8

1.0

3

qv (cm /cm )

Figure 17.3. Typical water-sorption and water-retention data for microcrystalline cellulose at 25°C. The retention curve was predicted by using the Brooks and Corey (1966) model adapted from Troygot and others (2007). E, exponential (for example, 1.E+7 = 1.07), and θv, volumetric water content.

2. Retention Curve Initially the sorption-isotherm data are converted to water data and plotted as tension head vs volumetric volume of water to comply with the RC presentation. Note that moisture at each data point of the sorption isotherm (Wc, aw) is converted to (h, θv) by using the Kelvin equation (Equation 17.23) and transforming the moisture content to a volumetric value: θV = M

ρb ρw

(17.24)

where M is moisture content (kg H20/kg dry solids), and ρb is product bulk density (kg/m3). Typical values are presented in Figure 17.3. 3. Prediction and Validation of Retention Curve from Water-Activity Data The sorption-isotherm data after their transformation into the adequate RC scale were used to derive the B-C model by using the two aforementioned requirements: continuity and smoothness at a point defined as θ*v ,h*. This yielded two equations from which θ*v and λ were derived. Once θ*v is known, it can be placed in one of the equations in order to derive λ and the B-C function can be drawn. Note that as a first approach, and for the simplicity of the treatment, it was assumed that the sorption isotherm is a linear function on a semilogarithmic scale (e.g., see Webb 2000). Another requirement for using the B-C function, when physical data are absent, is knowing the values of its parameters: hb, θs (air entry value, volumetric saturated water content), θr (volumetric residual water content), and λ (pore index parameter). The first two parameters can be experimentally

Using Principles of Water Transport in Porous Media

231

measured (although adaptation to food systems is required) and θr is an approximated parameter, whereas λ is evaluated through the model equations. To apply this approach in a VG model instead of the B-C model requires an additional parameter. This adds to the complexity of the solution without gaining a further conceptual insight into the overall approach and, as such, is not covered in this chapter. Validation of the RC by using the B-C model was then carried out. RC was measured by using the hanging-column method (Klute 1986), and the fit obtained is depicted on Figure 17.3. As the methods used for soils are probably difficult to implement in most food systems, measurements for few points are recommended, starting from the wet end of the curve. 4. Retention-Curve Validation Once the water RC is known, the unsaturated hydraulic conductivity function (Equation 17.16) is determined, and the Richards equation (Equation 17.22) can be solved. The saturated hydraulic conductivity is measured independently. Again, certain adaptations are necessary in evaluating Ks in food systems. Both measurement and prediction of 10-cm-column dry MCC rehydration are depicted on Figure 17.4. The simulations were made with Hydrus 1-D software (a code for simulating onedimensional movement based on the Richards equations for water flow [Šimůnek and others 2005]; Hydrus Software, Columbus, OH, USA) for the following boundary and initial conditions: constant upper-boundary flux and constant tension head at the bottom boundary set to zero. The initial moisture content was assumed as oven dry. The simulation predicts the cumulative bottom flux, multiplying this value with sample cross section is the amount (volume) of water gained by the sample: Rehydration of Foods by Using Porous Media Several research groups have applied the theory of capillary imbibition for modeling the rehydration of foods. A capillary-flow approach was used (Weerts and others

Cumulative bottom flux (cm)

8 7 6 5 4

Calculated

3

Experiment 1

2

Experiment 2

1 0

0

60

120

180

240

300

360

420

Time (s)

Figure 17.4. Typical rehydration data for microcrystalline cellulose at 25°C. Adapted from Troygot and others (2007).

232

PART 1: Invited Speakers and Oral Presentations

2003a, 2003b) to model the temperature and anisotropic effects during the rehydration of tea leaves. The predicted values agreed well with the experimental data derived from NMR measurements, leading to the conclusion that the physically based constitutive relationships of aw and hydraulic conductivity could be used to overcome the simplification of modeling water transport as a process governed by Fick’s laws. The approach could be extended to include gravity and osmotic pressure effects and also coupled with heat and solute transport in porous media for modeling heat, water, and chemical transport in general hydration and drying operations of porous food materials. In another study (Singh and others 2004), it was noted that one of the salient features of fluid transport through biological systems is the complex flow path presented by the biopolymeric matrix, thus expanding the capillary-flow approach. Similarly, it was recently proposed that use of an “effective” single cylindrical capillary radius is the simplest possible model to use for describing capillary penetration into a porous medium, and that this model may be insufficient in many cases (Saguy and others 2005b), especially when a significant distribution of pore sizes exists (Marmur and Cohen 1997), as previously shown for dried foods (Karathanos and others 1996). More specifically, and depending on the type of food and its processing history, the food may contain pores ranging in radius from 0.1 to 300 μm; those in the 10- to 300-μm range being most common (Bell and Labuza 2000). It is worth noting that, in the field of chemical engineering, the capillary model has also been widely studied and is being improved continuously. One such work (Marmur and Cohen 1997) used the kinetics of liquid penetration to characterize various porous media. It was shown that the effects of r and cos θangle can be evaluated independently, thus offering a possible solution for the aforementioned discrepancies. This was demonstrated in the case of a single vertical cylindrical capillary and also in the case of an assembly of vertical cylindrical capillaries in parallel (i.e., when a significant distribution of pore sizes exists). The model developed might be extended to evaluate the capillary mechanism taking place during the rehydration of food samples, leading to more accurate and representative results. Studies of vertical liquid penetration were recently presented based on Equation 17.14 (Saguy and others 2005a) and Equation 17.13 (Lee and others 2005); however, in neither case was the contact angle value resolved adequately. Although numerous models have been proposed and frequently applied, additional improvements are required before the process can be modeled based on the real mechanism(s). Nevertheless, recent studies toward more fundamental physical-based approaches have taken on a central role. Hence, significant progress is anticipated.

New Approaches and Other Advances The rehydration of dried foods involves different physical mechanisms, including water imbibition, internal diffusion (in the solid and its pores), convection and diffusion at the surface and within large open pores, hydraulic flow, capillary flow, and relaxation of the solid matrix. The foregoing sections describe the most relevant

Using Principles of Water Transport in Porous Media

233

factors affecting the rehydration of dry-food particles. In addition, the various approaches frequently used to model this process are highlighted. Two main approaches are often used to describe liquid transport into the dried solid matrix: empirical or semiempirical models (e.g., Peleg’s model and the Weibull distribution function) and the application of Fick’s laws of diffusion. Despite the significant and valuable insights these models have provided, several authors have recently proposed that the rehydration of dry-food particles cannot be explained and/or modeled solely by a Fickian mechanism. A growing body of data indicates that this simplification may not be fully warranted or, in some cases, might even result in misleading conclusions (Weerts and others 2003a, 2003b; Saguy and others 2005a; Datta 2007a). This realization is relatively new, and important insights are coming to light from the merging of different scientific disciplines, including food and soil sciences, biophysics, environmental sciences, and petroleum and chemical engineering. The ultimate goal of these interdisciplinary efforts is to provide more advanced and accurate physical models describing the mechanism(s) involved in the rehydration process. The models developed recently and applied successfully to the flow of liquids into dried foods are based on theories that are usually termed capillary-flow approach and/ or flow in porous media. The theories are already well developed in other fields, as are the methodologies needed to generate the experimental data required. However, there is still a considerable lack of data and, more importantly, standard methods enabling collection of the relevant physical properties of the food matrix and quantifying its interaction with the liquid. It is expected that, in the near future, collaborative studies will provide the information needed to apply these theories. The starting point for developing models that account for fluid flow in porous media is the Darcy equation and its Navier-Stokes analog (Weerts and others 2003a; Saguy and others 2005a; Datta 2007a). In addition, the main roles of the food particles’ physical properties (e.g., pore size distribution, heterogeneity, tortuosity), the embedding liquid (e.g., density, dynamic viscosity, temperature), and the interface (e.g., contact angle) are also considered in order to quantify the changes occurring during drying and to facilitate the modeling of the rehydration process. These studies on fluid flow in porous media have presented, in adequate detail, the development of the models, and the reader is referred to them for more detailed information. Some of the recent studies focusing on the various transport mechanisms in porous media include (a) molecular diffusion of gases, including water vapor; Darcy flow of gases due to pressure and Darcy flow of liquid due to gas and capillary pressures (Datta 2007a); (b) a variation of the Lucas-Washburn equation also used to model the rehydration of dried foods (Lee and others 2005; Saguy and others 2005b); and (c) the complexity of using the aforementioned mechanisms together with the input parameters needed to be derived from experimental data (Saguy and others 2005a; Datta 2007a, 2007b). This last category includes (1) density, porosity, thermal conductivity, and specific heat of the solid material; (2) molecular, capillary, and effective diffusivities; (3) moisture isotherms; (4) vapor and liquid permeability; and (5) waterpotential curves. Data in categories 4 and 5 are either nonexistent for foods or very difficult to obtain. Consequently, future research should address these topics.

234

PART 1: Invited Speakers and Oral Presentations

It is worth noting that, with new technologies and capabilities, such as X-ray microtomography (MCT), new possibilities are now feasible. X-ray MCT is a very powerful tool for detailed observations of the microstructure of porous foods and is expected to become a pertinent method in the near future for obtaining detailed information that can be further used in computational simulations of the imbibition of liquids into porous media. These data will open new avenues for ultimately quantifying and modeling internal changes during both drying and rehydration, thereby providing the necessary information for studying the mechanism(s) involved and enabling better control and quality improvements.

Research Needs Future research needs to focus on the accomplishment of four main goals: (a) expanding the aforementioned theory and approach to different aw, retention curve models, and foods; (b) applying theories of transport in porous media to model the fate of dissolved and retained substances in food particulates; (c) developing physical-based models, that is, multidisciplinary research integrating/assimilating of theories and knowledge from other fields; and (d) developing in situ quantification, that is, development of new methods/techniques for the in situ quantification of changes in food matrix microstructure, and water/vapor movement.

Conclusions The overall objective of this chapter was to highlight how porous-media physics could be implemented for modeling the rehydration of foods. The main conclusions are these: (a) Sorption-isotherm data was extended to a complete RC to be used in rehydration models. (b) Flow in porous-media theory was demonstrated to be applicable for food particulates. (c) The effectiveness of moving from empirical models to physically based models for foods was demonstrated. (d) Multidisciplinary collaboration was synergistic in opening new avenues for research and development. The field is entering a new era in which technology and scientific data can provide new insights and deeper knowledge. This will lead ultimately to much better food products that benefit from enhanced consumer appeal and acceptability.

References Aguilera JM, Michel M, Mayor G. 2004. Fat migration in chocolate: diffusion or capillary flow in a particulate solid? A hypothesis paper. J Food Sci 69:R167–74. Bell LN, Labuza TP. 2000. Moisture sorption: practical aspects of isotherm measurement and use. St Paul, MN: American Association of Cereal Chemists. Benavente D, Lock P, Del Cura MAG, Ordonez S. 2002. Predicting the capillary imbibition of porous rocks from microstructure. Transport Porous Media 49:59–76. Brooks RH, Corey AT. 1966. Properties of porous media affecting fluid flow. J Irrigation Drainage 92:61–8. Campbell GS. 1974. A simple method for determining unsaturated conductivity from moisture retention data. Soil Sci 117:311–4.

Using Principles of Water Transport in Porous Media

235

Campbell GS, Shiozawa S. 1992. Prediction of hydraulic properties of soils using particle-size distribution and bulk density data. In: van Genuchten MTh, Leij FJ, Lund L, editors. Proceedings of the international workshop on indirect methods for estimating the hydraulic properties of unsaturated soils. Riverside, CA: University of California. p 317–28. Carbonell S, Hey MJ, Mitchell JR, Roberts CJ, Hipkiss J, Vercauteren J. 2004. Capillary flow and rheology measurements on chocolate crum/sunflower oil mixtures. J Food Sci 69:E465–70. Chiralt A, Fito P. 2003. Transport mechanisms in osmotic dehydration: the role of the structure. Food Sci Technol Int 9:179–86. Cunha LM, Oliveira FAR, Ilincanu LA. 1998a. Application of the probabilistic Weibull distribution to rehydration kinetics: relationship between the model parameters and the underlying physical mechanisms. In: Oliveira JC, Oliveira FAR, editors. Proceedings of the third workshop of the Copernicus project. Porto, Portugal: Escola Superior de Biotecnologia. p 9–13. Cunha LM, Oliveira FAR, Oliveira JC. 1998b. Optimal experimental design for estimating the kinetic parameters of processes described by the Weibull probability distribution function. J Food Eng 37:175–91. Cunningham SE, McMinn WAM, Magee TRA, Richardson PS. 2007. Modelling water absorption of pasta during soaking. J Food Eng 82:600–7. Datta AK. 2007a. Porous media approaches to studying simultaneous heat and mass transfer in food processes. I. Problem formulations. J Food Eng 80:90–5. Datta AK. 2007b. Porous media approaches to studying simultaneous heat and mass transfer in food processes. II. Property data and representative results. J Food Eng 80:96–110. Falade KO, Abbo ES. 2007. Air-drying and rehydration characteristics of date palm (Phoenix dactylifera L.) fruits. J Food Eng 79:724–30. Garcia-Pascual P, Sanjuan N, Melis R, Mulet A. 2006. Morchella esculenta (morel) rehydration process modelling. J Food Eng 72:346–53. Gekas V. 1992. Transport phenomena of foods and biological materials. Boca Raton, FL: CRC. Giraldo G, Vazquez R, Martin-Esparza ME, Chiralt A. 2006. Rehydration kinetics and soluble solids lixiviation of candied mango fruit as affected by sucrose concentration. J Food Eng 77:825–34. Gowen A, Abu-Ghannam N, Frias J, Oliveira J. 2007. Modelling the water absorption process in chickpeas (Cicer arietinum L.): the effect of blanching pre-treatment on water intake and texture kinetics. J Food Eng 78:810–9. Hillel D. 1998. Environmental soil physics. San Diego, CA: Academic. Hills BP, Babonneau F, Quantin VM, Gaudet F, Belton PS. 1996. Radial NMR microimaging studies of the rehydration of extruded pasta. J Food Eng 27:71–86. Jury WA, Gardner WR. 1991. Soil physics. New York: John Wiley & Sons. Karathanos VT, Kanellopoulos NK, Belessiotis VG. 1996. Development of porous structure during air drying of agricultural plant products. J Food Eng 29:167–83. Klute A. 1986. Water retention: laboratory methods. In: Klute A, editor. Methods of soil analysis. Part I: Physical and mineralogical methods. Madison, WI: American Society of Agronomy. p 635–60. Krokida MK, Philippopoulos C. 2005. Rehydration of dehydrated foods. Drying Technol 23:799–830. Lee KT, Farid M, Nguang SK. 2005. The mathematical modelling of the rehydration characteristics of fruits. J Food Eng 72:16–23. Lucas R. 1918. Ueber das Zeitgesetz des kapillar Aufstiegs von Flussigkeiten. Kolloid-Zeitschrift 23:15–22. Marabi A, Dilak C, Shah J, Saguy IS. 2004b. Kinetics of solids leaching during rehydration of particulate dry vegetables. J Food Sci 69:FEP91–6. Marabi A, Jacobson M, Livings S, Saguy IS. 2004a. Effect of mixing and viscosity on rehydration of dry food particulates. Eur Food Res Technol 218:339–44. Marabi A, Livings S, Jacobson M, Saguy IS. 2003. Normalized Weibull distribution for modeling rehydration of food particulates. Eur Food Res Technol 217:311–8. Marabi A, Saguy IS. 2004. Effect of porosity on rehydration of dry food particulates. J Sci Food Agric 84:1105–10.

236

PART 1: Invited Speakers and Oral Presentations

Marabi A, Saguy IS. 2005. Viscosity and starch particle size effects on rehydration of freeze-dried carrots. J Sci Food Agric 85:700–6. Marabi A, Saguy IS. 2007. Rehydration and reconstitution of foods. In: Ratti C, editor. Advances in food dehydration. Contemporary Food Engineering Series. New York: Routledge. p 237–84. Marmur A, Cohen RD. 1997. Characterization of porous media by the kinetics of liquid penetration: the vertical capillaries model. J Colloid Interface Sci 189:299–304. Maroulis ZB, Saravacos GD, Panagiotou NM, Krokida MK. 2001. Moisture diffusivity data compilation for foodstuffs: effect of material moisture content and temperature. Int J Food Properties 4:225–37. Misra MK, Brooker DB. 1980. Thin-layer drying and rewetting equations for shelled yellow corn. Trans ASAE 23:1254–60. Morel-Seytoux HJ, Nimmo JR. 1999. Soil water retention and maximum capillary drive from saturation to oven dryness. Water Resour Res 35:2031–41. Oliveira FAR, Ilincanu L. 1999. Rehydration of dried plant tissues: basic concepts and mathematical modeling. In: Oliveira FAR, Oliveira JC, editors. Processing foods. Boca Raton, FL: CRC. p 201–27. Peleg M. 1988. An empirical model for the description of moisture sorption curves. J Food Sci 53:1216–9. Rideal EK. 1922. The flow of liquids under capillary flow. Philos Mag 44:1152–9. Ross PJ, Williams J, Bristow KL. 1991. Equation for extending water-retention curves to dryness. Soil Sci Soc Am J 55:923–7. Rossi C, Nimmo JR. 1994. Modeling of soil water retention from saturation to oven dryness. Water Resour Res 30:701–8. Saguy IS, Marabi A, Wallach R. 2005a. New approach to model rehydration of dry food particulates utilizing principles of liquid transport in porous media. Trends Food Sci Technol 16:495–506. Saguy IS, Marabi A, Wallach R. 2005b. Liquid imbibition during rehydration of dry porous foods. Innov Food Sci Emerg Technol 6:37–43. Salvatori D, Andres A, Chiralt A, Fito P. 1999. Osmotic dehydration progression in apple tissue II: generalized equations for concentration prediction. J Food Eng 42:133–8. Saravacos GD, Maroulis ZB. 2001. Transport of water in food materials. In: Transport properties of foods. New York: Marcel Dekker. p 105–62. Shi XQ, Fito-Maupoey P. 1994. Mass transfer in vacuum osmotic dehydration of fruits: a mathematical model approach. Lebensm Wiss Technol 27:67–72. Šimůnek J, van Genuchten MTh, Šejna M. 2005. The Hydrus-1D software package for simulating the movement of water, heat, and multiple solutes in variably-saturated media. Version 3.0, Hydrus Software Series 1. Riverside, CA: Department of Environmental Sciences, University of California. Singh PP, Maier DE, Cushman JH, Haghighi K, Corvalan C. 2004. Effect of viscoelastic relaxation on moisture transport in foods. Part I: Solution of general transport equation. J Math Biol 49:1–19. Tillotson JE. 2003. Convenience foods. In: Trugo L, Finglas PM, editors. Encyclopedia of food sciences and nutrition. Oxford: Academic. p 1616–22. Troygot O, Wallach R, Saguy IS. 2007. Utilization of water retention curves for rehydration modeling of foods. J Food Eng (in preparation). van Genuchten MTh. 1980. A closed-form equation for predicting the hydraulic conductivity of unsaturated soils. Soil Sci Soc Am J 44:892–98. Wallach R. 2007. Physical characteristics of soilless media. In: Raviv M, Lieth JH, editors. Soilless culture: theory and practice. New York: Elsevier. p 41–116. Washburn EW. 1921. The dynamics of capillary flow. Phys Rev 17:273–83. Webb SW. 2000. A simple extension of two-phase characteristic curves to include the dry region. Water Resour Res 36:1425–30. Weerts AH, Lian G, Martin DR. 2003a. Modeling the hydration of foodstuffs: temperature effects. AIChE J 49:1334–9. Weerts AH, Lian G, Martin D. 2003b. Modeling rehydration of porous biomaterials: anisotropy effects. J Food Sci 68:937–42.

18 Protein Hydration in Structure Creation P. J. Lillford and A.-M. Hermansson

Abstract Food processors have made structures from natural biopolymers for centuries, so why do we still not know the answers to all the problems of structure design, and its creation, in foods? Proteins are heteropolymers of amino acids, not designed for fabrication by processors, but evolved for particular biological functions at ambient temperature. Their performance derives from their distinctive native structure and that structure’s peculiar interaction with its aqueous environment. We are left to elucidate their behavior under more extreme conditions of pH, water activity (aw), temperature, pressure, etc. Adsorption isotherms of water on proteins at least show some similarities, and food processing operates across the entire aw or moisture-content spectrum. This report reviews what we know relative to the operating conditions and water contents used, with reference to general rules and the inevitable exceptions.

Introduction Proteins are ubiquitous in food systems, but their functional performance varies widely because their intrinsic structures vary enormously from soluble globular species (e.g., whey and soya), through disorganized swelling networks (e.g., gluten), to heavily cross-linked networks (e.g., collagen and elastin), and of course there is the unique structure of casein and its micelles. As food technologists, however, our job is to use them as interchangeably as possible to form gels, emulsions, foams, and, more recently, extruded snacks. So what is the polymer science of proteins and how does water affect behavior? This is a question asked not only by food technologists, but also in biotechnology, pharmaceuticals, and the biology of anhydrobiosis. The International Symposium on the Properties of Water (ISOPOW) has been particularly aware of this topic, and inspection of ISOPOW symposia 5–9 shows at least 13 major invited lectures on aspects of this topic. Here, frequent reference will be made to these contributions, and there is a wealth of other studies elsewhere. This presentation is an attempt to summarize what we do (and do not) know about the hydration of proteins and their role in structure formulation. As our measurement techniques have become more sophisticated, so has our realization that proteins are complex structures themselves, interacting quite specifically with water, small solutes, and other macromolecules. There are no simple predictive models available and probably never will be. At best, we can derive principles of behavior that are a qualitative guide to the choice of proteins as ingredients and their processability. 237

238

PART 1: Invited Speakers and Oral Presentations

h (g H2O/g) 0.3

0.2 1

0.1 2 0

0

0.2

0.4

0.6

0.8

1.0 P/P0

Figure 18.1. Typical adsorption-desorption cycles for biopolymers and foods. The lines marked 1 and 2 represent the first and second cycles; h, hydration (gram water/ gram protein); and P/P0, relative vapor pressure of the system.

Hydration Studies The early work examined the adsorption-desorption studies of water on proteins. Two phenomena are immediately observable: (1) Adsorption curves are complex, but most soluble proteins look similar; and (2) there is usually hysteresis between adsorption and desorption behavior (Figure 18.1). Adsorption is normally taken as the thermodynamic equilibrium line and can be fitted to a series of empirical isotherms. The Guggenheim-Anderson-de Boer (GAB) form is most usually applied, which gives us a set of characteristic parameters with which to compare protein but tells us nothing about the energetics or the molecular events taking place. Yang and Rupley (1979) have measured the energetics of adsorption and related them to the sequential hydration of specific groups. The overall conclusion is that all primary hydration sites are filled at a hydration of approx. 0.3 g/g protein. Finney and Poole (1984), updating this work, presented similar results showing that free water that could form ice was detected above approx. 0.35 g/g water, and that there was sequential loosening of the protein structure up to this hydration level. In dynamic uptake of heavy water (D2O), three regions were identified corresponding to surface amide exchange (0–0.5 g/g), a plateau (0.05–0.07 g/g), and finally a further rapid exchange. This was explained in terms of a change in the protein dynamic motion, and water acting as a plasticizer was proposed. Hills (1999) studied the motion of water itself by nuclear magnetic resonance (NMR) relaxation in gelatin samples. He came to the same conclusion (i.e., hydration leads to sequential loosening of the protein structure), and the even more remarkable observation that the transverse relaxation time of water correlates almost exactly with the water activity (aw) in the sample.

Protein Hydration in Structure Creation

239

Unfortunately, none of the foregoing workers related their observations to protein glass transitions, which is now known to move from high to ambient temperatures over these levels of hydration, though Hills (1999) reported a sudden increase in the slow-relaxation time component of water at greater than approx. 0.2 g/g protein. This we would now interpret as the point at which the glass transition meets the observation temperature (ambient). We return to the issue of hysteresis in adsorption-desorption later, when structuring of moist powders by extrusion is considered. Instead, we now examine the interaction of proteins and water at high aw, where hydration is completed and the formation of gels and interfacial protein structures is not limited by water availability.

Gelation Single-Protein Systems Heat Gelation That proteins can form gels was known long before their molecular structure was identified, and this forms a vital part of food technology. The molecular and microstructural events that take place have emerged much more recently. Studies have focused on the behavior of albumins and globulins that are easily solubilized in their native state but gel at higher temperature. We know that this involves denaturation, which is a cooperative phenomenon now routinely and easily measured by differential scanning calorimetry (DSC). We also know that this denaturation does not involve complete unfolding of the structure to produce a true random coil (Hermansson 1986b). Partial unfolding of the native structure is sufficient to cause insolubility and the formation of polymer aggregates that associate to form strings or chains of spherical units and higher-order aggregates. The application of microscopy has been preeminent in our understanding of these phenomena, and the techniques have been pioneered and reviewed by Hermansson (1998). Indeed, this work shows that aggregation begins before any denaturation event. Water plays its part insofar as it behaves as a good solvent for the native state, but a poor solvent for the denatured state, hence driving the aggregation into a space-filling gel. Whereas the protein forms the network, water is largely free in the interstices. The suggestion that water itself was in some way “structured” in gels resulted from NMR relaxation measurements, where a reduction in overall relaxation times was interpreted as “water binding.” Instead, this is now interpreted as the relaxation of water in bulk averaged with the smaller amounts (∼0.3 g/g) of water that are sampling the slower motional states of the polymer and its side chains (Lillford 1988). Gelation of proteins is determined by a competition between the kinetics of denaturation and aggregation. When the repulsive valance is sufficiently high, the onset of denaturation will initiate gelation and network formation. Denaturation is a cooperative process with a particular temperature for each protein, but it is not a sharp “melting point.” This means that the denaturation will occur at any temperature above the onset temperature, but its rate will vary throughout the phase transition. When denaturation predominates outside the isoelectric region and aggregation is suppressed, protein

240

PART 1: Invited Speakers and Oral Presentations

Figure 18.2. A particulate gel at pH 5.5 (left) and a transparent fine-stranded gel at pH 7.5 (right). Both gels are composed of 12% β-lactoglobulin (Hermansson 2008).

molecules can align into an ordered fine-stranded network structure. However, in the isoelectric region, when the net charge is low or at high ionic strength, aggregation can start prior to denaturation, and particulate gels are formed. This means that the overall morphology of the gel network is formed as dynamic arrest of an aggregated structure. The morphology is then strengthened by denaturation and subsequent cooling of the particulate network. Many dairy products are composed of particulate gels of whey proteins or casein (Hermansson 2008). Thus, the relative kinetics of aggregation and denaturation make it possible to form a whole set of variable network structures from the same protein. Figure 18.2 illustrates the enormous differences in gel structures that can be achieved by changes in the relative kinetics between aggregation and denaturation. Note that the scale bar in the micrograph of the particulate gel is 10 μm and that of the fine-stranded gel is 100 nm! There are many possibilities to manipulate the gel structure by processing when aggregation is predominant. This is illustrated by Figure 18.3 showing that a change in heating rate has a pronounced effect on aggregation and thus on the particulate network structure of whey-protein gels prepared under exactly the same conditions prior to heating (Stading and others 1993). The figure shows two-dimensional sections of gel networks formed at different heating rates. The slower the heating rate, the more time is available for aggregation and the coarser is the network structure formed. The change in heating rate affects pore size, as well as the size of the protein aggregates. The pore diameter increased from ∼20 to ∼100 μm when the heating rate was decreased from 12° to 1°C/min. Not only the pore size but also the degree of clustering and the particle size increased as the heating rate decreased. As shown in Figure 18.3, the polyelectrolyte nature of both the native and the denatured states means that both the denaturation and subsequent aggregation will depend on solvent conditions. As well as pH, ionic strength and Hofmeister ion effects

Protein Hydration in Structure Creation

241

Figure 18.3. Light-microscopic images of 10% β-lactoglobulin gels at pH 5.3 (Stading and others 1993).

are to be expected. Knowing this, we can predict the direction of the influence of these factors on gel properties but not their absolute effects on gel moduli or fracture properties. When in solution, the proteins of muscle foods (i.e., actomyosin, myosin, and the sarcoplasmic proteins) obey similar rules, though their intrinsic multidomain molecular structure also affects the properties of the gels formed. The gels themselves have been described as fractal structures, which is a convenient term for describing their architecture. However, in terms of gel mechanics they should be regarded as water-filled composites. Their small deformation (modulus) behavior will be dominated by the pore size and interconnectivity of the channels within them, since any deformation requires fluid flow, and properties will relate to the rate of deformation. Furthermore, deformation can be either (a) reversible, driven by equalization of osmotic forces and recoverable strain in the network chains; or (b) irreversible, if deformation induces new associative cross-links to form as chains approach

242

PART 1: Invited Speakers and Oral Presentations

each other. Progressive chain association may even be spontaneous at room temperature and lead to syneresis. Cold Gelation The insolubility of globulins and casein at their isoelectric point is also used to form coacervates or gels from soluble or heavily swollen hydrated proteins. In colloidal association terms, this is a simple process involving electrostatic interaction and hydrogen bonding. Its detailed manipulation by fermentation or direct acid addition results in the complete range of yoghurt, quark, and fermented sausage structures that are the basis of multimillion dollar businesses. Van Vliet and Walstra (1994) comprehensively reviewed the interactions of polymer and water in dairy systems. They reviewed the structures formed in terms of their fractal dimensions and showed that acid and rennet milk gels are initially of similar structure. They demonstrated the significance of network rearrangement in the syneresis of rennet casein gels and argued very systematically that the properties of the networks, rather than any special “binding” of water, lead to both syneresis and/or retention of water in gels. They showed by simple calculation that a pressure of 105 Pa (1 bar) reduces aw by 0.1, whereas an applied pressure of only 10–100 Pa can remove some water or lead to faster syneresis in cold-set gels. We must look to a larger scale (micromolar) to see why water holding varies. The pores in particulate protein gels can range from 1 to 100 μM. With complete wetting of the protein matrix, the negative pressure caused by capillary absorption is given by: ΔP = 2γ d ( where γ is the surface tension and d is the diameter of the capillaries ) . With a γ of 35 mNm−1 and 1-μM channels, ΔP ∼ 105 Pa (1 bar), a pressure under which most gels would fracture and flow, before releasing water. At 100 μM, ΔP = 103 Pa, allowing water to be released under much lower loads. Water Holding One of the key functional parameters often measured for proteins is their waterholding capacity; that is, their ability to retain water over time at equilibrium or under mechanical stress. It is obvious from the foregoing discussion that this is not an absolute property of a protein, but is a property of the architecture of the gels formed. In terms of capillary pressure in a gel, with capillaries of 1 μM, a pressure of 1 bar would be necessary to expel the water from a casein gel, and the pressure required is inversely related to capillary pore size. Not surprisingly, therefore, the results are highly dependent on the conditions of the gelation and the methods by which water retention is measured. The network of the fine-stranded gel formed at pH 7.5 shown in Figure 18.2 has such small pore dimensions that it is not possible to force any water from the gel by compression. When water is released from such transparent gels, it is most often because of syneresis caused by aggregation taking place after network formation and resulting in a denser network. The consequence is that there is simply less space

Protein Hydration in Structure Creation

= 5% Blood plasma protein = 12% Whey protein

50

Moisture loss, %

Moisture loss, %

40

30 20 10 0 0

= 5% Blood plasma protein, pH 9 = 12% Whey protein, pH 7

50

40

243

30 20 10

75

80 85 90 Temperature, °C

95

100

0 0

1

2

3 NaCl, %

4

5

Figure 18.4. Moisture loss as a function of temperature (left) and salt concentration (right) for blood plasma and whey-protein gels (Hermansson 1986a).

available for the water. On the other hand, for the particulate gels and gels with pore sizes above 200 nm, the capillary forces are low enough for water to be squeezed out by compression. The particulate gel structures formed with different heating rates shown in Figure 18.3 will have significantly different water-holding capacities. β-Lactoglobulin is responsible for the main gel characteristics of whey-protein mixtures. The effect of the kinetics of aggregation on the moisture loss is illustrated by results from particulate whey and blood plasma protein gels in Figure 18.4. The kinetics of aggregation and thereby the coarseness of the gel network have been increased by the heating rate and by the addition of salt. In both cases, the increase in aggregation results in a coarser network structure and substantial increase in the moisture loss (Hermansson 1986a). Permeability measurements have been made of coarsely aggregated dairy gels, such as casein gels (Mellema and others 2000). Roughly speaking, gels with pore sizes above 100 nm can release water without breaking when an external force is applied. However, the exact relationships between the pore size distribution and the waterholding properties of heterogeneous gel structures are far from being completely described, so much more work is needed in this area. The measurement of NMR relaxation time distributions can be related to the architecture of the non-deformed specimen since these can be related to porosity, but the response to deformation is not predictable without a measure of the viscoelastic behavior of the polymer chains themselves (Figure 18.5) (Ablett and others 1991). Mixed Systems Proteins in foods are rarely gelled as a single molecular species. Two vital pieces of information are required to understand phenomena in such systems: (1) Do the polymers mix intimately?; and (2) Does the gelation of one polymer influence the gelation of the second?

244

PART 1: Invited Speakers and Oral Presentations

% Compression 0 6 12 18 26 32 38

.001

.01

.1

1

10

T2 Relaxation time (s)

Figure 18.5. Change in the distribution of water-proton relaxation times during the compression of cooked beef (Ablett and others 1991).

Many proteins form intimate mixtures. Indeed, many raw materials are themselves mixtures. For example, whey contains a mixture of α-lactalbumin and β-lactoglobulin, and soya a mixture of glycinin (11S), conglycinin (7S) and a variety of smaller albumins (2S). The presence of a network formed from the first, restricts the formation of a network from the second. These effects are known, but the mechanical consequences are not predictable. Such studies are vitally important, however, since a shift in phase structure can completely change the functional performance of the mixed system. One example is given next. When a mixture of gelatin and whey protein is heated, the whey protein will gel whereas the gelatin remains liquid. The gelatin will then gel in the pores of the wheyprotein network when the system is cooled. Thus, a bicontinuous system is formed. On the other hand, at room temperature, the gelatin will form a gel and, when high pressure is applied, the whey protein will aggregate in the gelatin network. This second structure will be gelatin continuous. The size of the aggregates can be moderated by combining high temperature and high pressure treatments. These two gelled systems will have different behaviors because the gelatin-continuous systems will melt on heating. The bicontinuous systems will be firmer and will not melt on heating (Figure 18.6) (Walkenström and Hermansson 1997).

Protein Hydration in Structure Creation

Temperature

Temperature/high pressure

10 μm Bicontinuous

High pressure

10 μm Gelatin continuous

245

10 μm Gelatin continuous

Figure 18.6. Light-microscopic images of 12% whey protein + 3% gelatin, at pH 5.4, induced by temperature and high-pressure processing (Walkenström and Hermansson 1997).

In many cases, food polymers are not miscible and undergo phase separation. Protein/polysaccharide mixtures usually exhibit this phenomenon. The recognition and measurement of this phenomenon has been pioneered by Tolstoguzov (1999) and used by many other workers to create a particular mechanical and rheological performance (Blindt and others 2004). When phase separation occurs, the concentration of polymer in each phase is critical, as is the nature of the phase structure (water-in-water emulsion, bicontinuous, etc.). Unfortunately, whereas the resultant phase composition can be related to the relative compatibility of each solute with water (i.e., the Flory-Huggins parameters of each polymer and the polymer-polymer interaction), these parameters are not yet predictable, and therefore the phase diagram has to be measured for every binary system. The real world is even more complicated, since the phase diagrams will change with temperature and conformational changes prior to gel formation and phase separation. As well as spontaneous phase separation during gelation of mixed systems, the protein and polysaccharide components usually separate during cereal processing. Almost always, we see the separate phases of protein and polysaccharide in the finished microstructures. Figure 18.7 shows a high-moisture dough where starch is seen in a protein-continuous matrix. Emulsion Stabilization Proteins form two-dimensional structures at interfaces, a process that can be compared with gelation. This is too large a topic to be covered comprehensively in this review, but has been comprehensively treated by Dickenson (1994) and Morris (2006), where the competitive molecular dynamics between proteins and surfactants have been visualized.

246

PART 1: Invited Speakers and Oral Presentations

Figure 18.7. Pasta dough heated at 5°C/min (Thorvaldsson and others 1999).

Structuring by Extrusion Intermediate to High Water Content The formation of fibrous structures from proteins is essentially the formation of elongated gels by any thermal or chemical denaturation process combining with a subsequent aggregation, all while under the action of elongational shear. Few of these processes remain in commercial practice, but have been reviewed elsewhere (Lillford 2008). The development of the high-shear single or double screw extruder by the plastics industry presented a whole new technology to food restructuring. Here, dough is formed at low moisture contents (approx. 15–50 g water/g substrate), and the “melt” is forced through dies to produce specific shapes of final product. Proteins have been structured successfully by these processes, but moisture contents are low, and a dry powder (glass) is being transformed through hydration, temperature, and pressure treatment (rubber and melt states) back to a cooled extrudate (rubber or glass). The following two types of protein processing are currently practiced. Fibrous and Dense Composites These form when protein powders are hydrated in the extruder to levels of (approximately) 0.4–0.5 g/g water. At these moisture levels, hydration is complete, and some water can act as a lubricant. The doughs are sufficiently fluid that very high pressures and temperatures are not required, and expansion after the die is low and dense, but porous or aligned structures are formed (Figure 18.8).

Protein Hydration in Structure Creation

247

5 MM

Figure 18.8. Aligned flake extrudate.

The physicochemical changes imposed on the protein can be understood most easily by inspection of the DSC behavior of individual proteins. For example, gluten shows a glass transition and thermal relaxation as its proteins are initially hydrated. There is no cooperative denaturation, but a melt forms which, on cooling, reverts to a glass or rubber state, depending on the final moisture content (Figure 18.9). Non-denatured soya exhibits the same hydration process, but also denatures within the extruder barrel. This produces an almost instantaneous pressure rise and can frequently stop the flow, resulting in catastrophic damage. Only at higher moisture contents is this averted, so, surprisingly, native proteins are not ideal unless accompanied by diluting (lower viscosity) carbohydrate components in flours and concentrates. Under these conditions, protein is the discontinuous phase of the mix and instead of forming continuous matrices, it forms gelled inclusions in a carbohydrate matrix. Prior denaturation of the protein during raw material preparation (by thermal aggregation) has beneficial effects. A melt can be formed as hydrated gel particles fuse at higher temperatures. Expanded Solid Foams These are the low-density products used as components of breakfast cereals and snacks, normally extruded at moisture contents between 15% and 30%. The

248

PART 1: Invited Speakers and Oral Presentations

28.18 Gluten

Peak = 67.82 °C Onset = 62.13 °C End = 72.40 °C Peak Height = 0.4578 mW

9.3%mc

Area = 16.916 mJ Delta H = 1.1057 J/g

27 26

Heat flow endo up (mW)

Onset = 60.17 °C 25

Peak = 65.45 °C Peak Height = 1.4548 mW

24

Area = 61.057 mJ Delta H = 3.9907 J/g

23

End = 71.12 °C

11.4%mc Peak = 61.16 °C Peak Height = 1.5832 mW

22 21 12.5%mc

Onset = 56.30 °C End = 66.14 °C

Area = 59.983 mJ Delta H = 5.0406 J/g

20 19 18 17.54 7.504

15.7%mc Peak = 52.18 °C Peak Height = 0.3882 mW

20

30

40

50

End = 58.80 °C Area = 47.790 mJ Onset = 37.52 °C Delta H = 3.3655 J/g

70 60 Temperature (°C)

80

90

100

110

119.7

Figure 18.9. Scanning calorimetry of gluten protein as a function of moisture content (mc). Endo, endothermic heat flow direction is up; and mW, milli-Watts. Reproduced with permission of Food Futures Flagship.

materials science principles operating during the manufacture of these expanded solid foams have been reviewed (Lillford 2008). The continuous phase is usually starch, with protein as a heat-set inclusion in the matrix (Hermansson 1988). A light micrograph of such structures is shown in Figure 18.10. However, attempts to make “healthier” products by reducing carbohydrate/protein ratios are now in vogue to the point where proteins can become the continuous phase. Clearly, more details of the glass, rubber, and melt states of proteins at low hydration levels are required. For example, although the denaturation and aggregation of proteins at high aw have been studied extensively, completing this process requires much more detailed knowledge of the behavior of proteins as glassy and rubbery polymers. A simple model relating processing to melting temperature and glass transition temperature has been proposed (Strahm and others 2000). We know that plasticizers other than water (e.g., glycerol and low molecular weight sugars) influence their glass transitions (Gosline and Lillie 1993). However, since they are polyelectrolytes, there is every reason to believe that this fact will result in an influence of pH, ionic, and Hofmeister series on their glass/rubber transitions. There are almost no systematic data in the literature relating to the effect of

Protein Hydration in Structure Creation

249

Figure 18.10. Light micrograph of extruded wheat flour showing starch as the continuous structure with protein particles included.

these variables, so much remains to be done in the study of “proteins in the dry(ish) state.”

References Ablett S, Darke AH, Lillford PJ. 1991. The effect of mechanical deformation on the movement of water in foods. In: Levine H, Slade L, editors. Water relationships in foods. Advances in Experimental Medicine and Biology, vol 302. New York: Plenum. p 453–64. Blindt RA, Clark AH, Elliot B, Foster TJ, Norton IT, inventors. Unilever Home & Personal Care USA, Division of Conopco, Inc., assignee. 2004, Dec 2. Method for preparing a gelled food product. U.S. patent 20040241305. Dickenson E. 1994. Protein stabilised emulsions. In: Fito P, Mulet A, McKenna B, editors. Water in foods. Fifth International Symposium on the Properties of Water (ISOPOW 5). London: Elsevier Applied Science. p 59–74. Finney JL, Poole PL. 1984. Comments. Mol Cell Biophys 2:129. Gosline JM, Lillie MA. 1993. The effect of swelling solvents on the glass transition in elastin and other proteins. In: Blanshard JMV, Lillford PJ, editors. The glassy state in foods. Nottingham, UK: Nottingham University Press. p 133–56. Hermansson A-M. 1986a. Water- and fatholding. In: Mitchell JR, Ledward DA, editors. Functional properties of food macromolecules. London: Elsevier. p 273–314. Hermansson A-M. 1986b. Soy protein gelation. J Am Oil Chem Soc 63:658–66. Hermansson A-M. 1988. Gel structures in biopolymers. In: Blanshard JMV, Mitchell JR, editors. Food structure: its creation and evaluation. London: Butterworth’s. p25–40. Hermansson A-M. 1998. Supramolecular structures of biopolymer gels. In: Reid DS, editor. The properties of water in foods. Sixth International Symposium on the Properties of Water (ISOPOW 6). London: Blackie. p 3–29. Hermansson A-M. 2008. Structuring water by gelation. In: Aguilera JM, Lillford PJ, editors. Food materials science: principles and practice. New York: Springer. p 255–80. Hills BP. 1999. NMR studies of water mobility in foods. In: Roos YH, Leslie RB, Lillford PJ, editors. Water management in the design and distribution of quality foods. Seventh International Symposium on the Properties of Water (ISOPOW 7). Lancaster, PA: Technomic. p 107–31.

250

PART 1: Invited Speakers and Oral Presentations

Lillford PJ. 1988. The polymer/water relationship: its importance for food structure. In: Blanshard JMV, Mitchell JR, editors. Food structure: its creation and evaluation. London: Butterworth’s. p 75–92. Lillford PJ. 2008. Extrusion. In: Aguilera JM, Lillford PJ, editors. Food materials science: principles and practice. New York: Springer. p 415–35. Mellema M, Heesakkers JWM, van Opheusden JHJ, van Vliet T. 2000. Structure and scaling behavior of aging rennet-induced casein gels examined by confocal microscopy and permeametry. Langmuir 16:6847–54. Morris VJ. 2006. Studies on the molecular organisation at the water interface. In: Buera M, Welti-Chanes J, Lillford PJ, Corti HR, editors. Water properties of food, pharmaceutical, and biological materials. Food Preservation Series: ISOPOW 2004 conference proceedings. Boca Raton, FL: CRC. p 273–88. Stading M, Langton M, Hermansson A-M. 1993. Microstructure and rheological behaviour of particulate β-lactoglobulin gels. Food Hydrocolloids 7:195–212. Strahm B, Platner B, Huber G, Rokey G. 2000. Application of food polymer science and capillary rheometry in evaluating complex extruded products. Cereal Foods World 45:300–2. Thorvaldsson K, Stading M, Nilsson K, Kidman S, Langton M. 1999. Rheology and structure of heat-treated pasta dough: influence of water content and heating rate. Lebensm Wiss Technol 32:154–61. Tolstoguzov V. 1999. The role of water in intermolecular interactions in food. In: Roos YH, Leslie RB, Lillford PJ, editors. Water management and the distribution of quality foods. Lancaster, PA: Technomic p 199–233. van Vliet T, Walstra P. 1994. Water in casein gels. In: Fito P, Mulet A, McKenna B, editors. Water in foods. Fifth International Symposium on the Properties of Water (ISOPOW 5). Oxford: Elsevier. p 75–88. Walkenström P, Hermansson AM. 1997. Mixed gels of gelatin and whey protein formed by combining temperature and high pressure. Food Hydrocolloids 11:457–70. Yang P, Rupley JA. 1979. Protein-water interactions: heat capacity of the lysozyme-water system. Biochemistry 18:2654–61.

19 Water Partitioning in Colloidal Systems as Determined by Nuclear Magnetic Resonance P. Chinachoti and P. Chatakanonda

Abstract Food colloids are composed of complex molecular arrangements of biopolymers with various affinities for water. These molecules vary in chemical and physical properties. The physical history of the system, such as processing and storage conditions, can impact conformation and intermolecular or intramolecular interaction. Predicting behavior of food colloids can be complicated by these interrelated properties that result in the ability to interact with the surrounding water molecules. Water in food biopolymers may distribute or compartmentalize unevenly. The water molecules may inherently rotate freely and exchange rapidly with neighboring molecules but have more difficulty exchanging with those further away or physically separated. When trapped in a three-dimensional structure, water molecules may maintain some bulk water physical properties (e.g., freezing point and mobility), but their diffusive exchange ability is influenced or perturbed by physical barriers. The properties of physical barriers can influence the water exchange among domains either at fast or at slow rates with respect to the instrumental time frame or method. In the case of nuclear magnetic resonance (NMR), a fast-exchange regime on a microsecond scale results in seemingly simple relaxation (free-induction decay [FID]). Recent applications of a continuous model resulted in a way to decompose the FID to identify the contributions from multiple populations. These populations represent water rapidly exchanged among domains. Water behavior in starch influences the starch physical-chemical properties during thermal treatments (e.g., gelatinization and freeze-thaw cycling).

Introduction Water molecules serve as a universal solvent that supports all living systems. Not only does it exhibit unique physical and chemical properties as compared to other solvents (e.g., degree of hydrogen bonding, and melting and boiling points), its comparatively small size enables it to move and rotate dynamically. When molecularly associated, water molecules continue to rotate or diffuse dynamically at rates dictated by the local environment and the tumbling motions of the associated molecules. Some water molecules may be trapped in a three-dimensional structure such that channels for molecular exchange are long and/or restricted, so the water molecules cannot move freely but yet are not necessarily bound to a chemical surface. In the dynamic regime of instruments such as nuclear magnetic resonance spectrometers (NMRs), water in a 251

252

PART 1: Invited Speakers and Oral Presentations

solid state (ice) moves with correlation or relaxation times that are several orders of magnitude longer than those for free liquid water. Trapped or diffusively restricted water may also exhibit relaxation times that differ by a few orders of magnitude. Understanding the fundamental properties of water in colloidal systems can impact various areas of food science technologically in several ways. This chapter describes water in starch in different identifiable populations with relaxation times distinguishably different in the NMR time frame. We explore how starch may be hydrated in a starch granule and water be distributed in various amorphous/crystalline regions in ways that may influence the starch functional properties.

Starch Starch serves as a major source of energy in our diet. It can be converted to other food ingredients (e.g., syrups, maltodextrins, and alcohol). The hydration behavior of starch and starch-derived ingredients is investigated primarily because of the technological implications for applications to the food industry. Although the primary structure of starch is known (i.e., amylose and amylopectin), understanding of the secondary conformation of amylose and amylopectin within a starch granule has been limited by the complex architecture of starch granules and lack of instrumental capability of probing a granule. Starch genetics and biosynthesis are influenced by environmental factors, such as drought, and are currently under investigation (e.g., James and others 2003; Baguma 2004). In the case of potato and cassava starches, for instance, scientists have explored the impact of drought on starch synthesis (Munyikwa and others 1997; Geigenberger and others 1999; Baguma 2004). As a unique drought-tolerant crop, cassava-starch quality and its biosynthesis are of particular interest as future drought threat continues in various parts of the world (Sriroth and others 2001; Raemakers and others 2005). Linking starch structure to functionality in this context is discussed to clarify the nature of the water hydration and how this influences the gelatinization and freezethaw stability properties. A key question is whether changes in the internal arrangement of the starch granules occur in a way that influences the distribution of water, or whether the distribution of water influences the internal arrangement of the starch granules. Clarifying this could enable scientists to manipulate starch structure and understand how drought season impacts future crops. From the producer and manufacturer perspective, future grain and tuber crops are at risk from global warming and its associated more severe drought. Hopefully, this approach to the study of hydration will contribute to a future understanding of starch produced in adverse climates and of interventions by food manufacturers so that they may be able to find consistent raw materials and better predict starch performance of materials harvested in the different growth seasons.

Amylose, Amylopectin, and More Over the past 10–15 years, improved understanding of amylose and amylopectin has revealed new insights about the conformation and localized organization of their

Water Partitioning in Colloidal Systems

253

molecules (reviewed by Oates 1997; Jacobs and Delcour 1998; Zobel and Stephen 1995). In a 1997 study, Gallant and others proposed that amylopectin (crystalline) lamellae are formed in spherical blocklets located between amorphous growth rings (Figure 19.1). These blocklets were estimated to be in the range of 20–500 nm in diameter, depending on starch botanical type and on location in the starch granules. Presence of the blocklets also determines the local degree of crystallinity and hence susceptibility to hydrolysis, such as that in the semicrystalline shell (Oates 1997). Crystallinity may not be uniform across the granule. For instance, nonwaxy cereal-starch granules have been shown to exhibit higher crystallinity nearer the outer shell (Morrison 1995). One of the most challenging tasks in understanding starch is the determination of the architectural arrangement of amylose and amylopectin in a native intact granule. Study findings by Oostergetel and van Bruggen (1993) suggested that branch chains in amylopectin molecules are not distributed randomly; rather, they are clustered with portions forming double helices contributing to the starch crystalline structure. The localized, amorphous/gel-like chains may exhibit folding or multiple branching which prevents the molecule from forming ordered polymer structures in its amorphous regions. The region in crystalline lamellae of amylopectin side-chain clusters is on average 6 nm long, and amorphous lamellae of the branched zone are about 4 nm long (Gallant and others 1997). The starch growth ring consists of concentric shells or layers of alternating high and low refractive index, density, and crystallinity. Growth rings (potato and wheat) are visible under a light microscope but seldom are seen in smaller granules. Hydrolytic enzymes or acid treatment can erode the amorphous region, leaving these rings that are 1200–4000 Å thick (Jenkins and others 1993). Major crystalline amylopectin is located in 120- to 400-nm-thick, hard layers that are stacks of crystalline lamellae; namely, the backbone of the starch granule. Even though they are considered as highly crystalline, these shells consist of alternating amorphous and crystalline lamellae that are approximately 9–10 nm thick (Jenkins and others 1993; Oostergetel and van Bruggen 1993). The size of the amylopectin sidechain clusters within the crystalline lamellae varies, but on average is around 10 nm wide by 9–10 nm long; the majority of amylopectin is in these lamellar structures (Hizukuri 1986; Manners 1989; Oostergetel and van Bruggen 1993). Crystalline amylopectin lamellae are aligned within blocklets between the amorphous growth rings, as evidenced by scanning and transmission electron microscopy (SEM and TEM) and atomic force microscopy (Gallant and others 1997). The larger the blocklets are, the more resistant the starch tends to be (e.g., potato), and starch from different botanical types have blocklets of various size (20–500 nm). Other properties also influence starches’ resistance, including higher crystallinity nearer the outer shell in some nonwaxy starches (Morrison 1995). Very little is known about the organization of starch polymers in semicrystalline shells except for the presence of amylopectin in smaller blocklets (20–50 nm in diameter) with reduced crystallinity.

Semicrystalline lamellae Amorphous growth rings Pores Hilum

Granule surface

Whole granute Semicrystalline lamella Amorphous growth ring

Amorphous channels

Large blocklet

Small blocklet

Crystalline Amorphous Blocklet

Amylopectin clusters

Side view

Top view

Type A

Amylose

Lipid

Type B

Figure 19.1. Overview schematic diagram of starch granule structure redrawn from Gallant and others (1997). Within the box is the section area that was enlarged to picture the semicrystalline lamellae.

254

Water Partitioning in Colloidal Systems

255

Hydration Upon hydration, water molecules primarily occupy spaces in the amorphous regions of the starch granules. These are the zones between each lamella, between amylopectin clusters, and around each blocklet, both in the hard and the soft shells. The amorphous fraction, therefore, controls the mechanical elasticity and flexibility and granule volume by its ability to absorb and release the free water of the raw starch granules. The less flexible crystalline shells also limit granule expansion (Oostergetel and van Bruggen 1993) and hence granular enlargement upon melting during gelatinization. In starch gelatinization, granule swelling increases starch-chain mobility and swelling. At a more advanced stage of swelling, amylose leaches through 0.05- to 0.1-μmdiameter canals out of the starch granule to form a network structure (Figure 19.2).

(a)

(b)

(c)

(d)

Figure 19.2. Scanning electron (SEM) and transmission electron (TEM) micrographs of cassava and potato starches undergoing gelatinization. (a) Cassava starch SEM with surface initial swelling, (b) SEM and (c) TEM of cassava starch showing advanced swelling where amylose strands (arrows) are leaving the granules through channels, and (d) SEM of gelatinizing potato starch showing strands of amylose entanglement outside of the granule. From Notté (1993), Garcia (1996), and Gallant and others (1997).

256

PART 1: Invited Speakers and Oral Presentations

Nondiscriminating Properties It is important to pay attention to the methods used to evaluate water properties that can probe or differentiate diverse populations within a compound in a system. Classic approaches to determine average properties, such as water-sorption isotherm, average mobility of water (e.g., relaxation time and correlation time), and glass transition temperature, all use bulk properties that do not necessarily differentiate subpopulations. Absent a more robust and sophisticated dynamic instrumental approach, one should not expect classic averaged or bulk properties to be of much use when a subpopulation needs to be finely differentiated. For example, in a study of cassava starch from various crops (rainy season vs drought), water-sorption isotherms and average NMR relaxation time results showed no difference in properties, yet the functional properties are different. In the following discussion, we demonstrate the use of spinspin (transverse) relaxation time (T2) distribution analysis to differentiate subpopulations of water that are distributed in diverse amorphous regions in a starch granule. This is critical to further understanding how starch is gelatinized and how the amorphous network is formed. Water mobility in starch is primarily decoupled from starch-chain mobility (Chatakanonda and others 2003b) in that, even in the glassy state, though the starch chains are immobile, the water can be quite highly mobile, as observed by means of deuterium and proton NMR. Fortunately, this means that water in starch can be relatively easily evaluated by NMR even while it is trapped within the starch solids, and, upon gelatinization or melting of starch, it is possible to observe how water in different regions of starch domains changes. To determine the heterogeneous nature of starch hydration, time-domain NMR relaxation can be further analyzed to determine how water may be distributed.

Water Distribution NMR relaxation has been used for decades to describe the physical motion of water molecules. By using time-domain NMR, differentiation of water proton relaxation and solid protons has a long history of relying on the discrete-model assumption of components with distinctive relaxation times. The discrete model has been most widely used in the past to describe bound vs free water, which has been one of the most controversial issues in this area of water study in food and other complex systems. In simple homogeneous and isotropic systems, multiexponential relaxation theoretically means multiple relaxing components, and relaxation times can be estimated from the exponential decay curve (by assuming some sort of fast-exchange regimen). Unfortunately, most multicomponent systems with a complex structure (e.g., organelles, or cellular or physical compartmentalization such as in gels and emulsions) exhibit a rather complex proton relaxation due to a number of phenomena, including, for example, proton exchange at various rates depending on proximity and other dynamic and kinetic barriers (Belton and Hills 1987). Within an NMR experimental time frame, a sufficiently long diffusive exchange of protons may cause significant deviation in the exponential relaxing (decay) response, such as in the case of fast-

Water Partitioning in Colloidal Systems

257

frozen and slow-frozen food matrices (Belton and others 1993). In such cases, NMRrelaxometry data are complex and cannot be interpreted with simple model (e.g., two-sites, fast exchange). The distribution of T2 in heterogeneous systems has been introduced to determine water distribution in various biological systems (e.g., Lillford and others 1980; Kroeker and Henkelman 1986; Newcomb and others 1990; Menon and Allen 1991).

Discriminating NMR T2 Distribution A continuous model of proton relaxometry has been introduced and applied to several food systems (Ruan and Chen 1998; Tang and others 2001). A non-monoexponential free-induction decay (FID) curve can be evaluated by using a multiexponential model assuming a continuous distribution of relaxation time. Unlike the discrete model, where each population is described as having a single, uniform correlation time, the continuous model assumes a distribution of correlation or relaxation times (Figure 19.3). Relaxation of a heterogeneous system (mixture of surfaces of compatible and incompatible constituents) can be complicated by contributions from spin-lattice and spin-spin relaxation interaction. In such systems, the proximity of protons or other

z

x

Log Α

y

Discrete multiexponential decay of relaxation time: A = Σ Aie-t/T2i

Homogeneous

Continuous exponential decay of relaxation times:

gi = Σj=1mfje-ti/Tj

Heterogeneous

Figure 19.3. Discrete and continuous models of proton relaxation in interpretation of an exponential decay curve. Unlike the discrete model (left), where each population is described to have a single, uniform correlation time, the continuous model assumes a distribution of correlation or relaxation times.

258

PART 1: Invited Speakers and Oral Presentations

nuclei being used to observe the NMR signal may influence local environments, which in turn influence chemical and diffusive exchange affecting T2 (Provencher 1982a, 1982b; Tellier and others 1993). Since what we see is the sum of the contributions from all spins at different rates within the NMR timescale, it is reasonable that a continuum of different environments in which spins and different exchange rates exist resulting in a continuous T2 distribution (Provencher 1982a; Hills 1992; Tang and others 2000). The continuum-model approach has been pursued using the contin computer program (Tang and others 2001; Chatakanonda and others 2003a, 2003b) to process noisy NMR decay data. The T2 spectra obtained, for example, can be regarded as a probability distribution of spins at various relaxation times. Tang and his coworkers (2000) reported NMR T2 distribution of water in native starch granular microstructure after observing temperature and moisture dependence of the distribution of water proton transverse relaxation times. These water populations were assigned to (a) thin film of bulk water between the packed granules (∼50 ms), (b) water in the amorphous growth rings and the semicrystalline lamellae (∼8 ms), (c) fast-exchange (at 25°C) water and with a diffusive exchange lifetime of ∼1 ms, and (d) orientationally ordered, intragranular water of 1-kHz deuterium-NMR line width, assigned to water inside the hexagonal channels of B-type crystals (a polymorphic form of starch packed in hexagonal unit cells) in the semicrystalline lamellae (nonexchanging on the NMR timescale). Physical manipulation or treatment (e.g., freezing) helped confirm these assignments for potato, pea, and corn starches.

Cassava Starch Cassava starch is one of the most economically beneficial crops in Thailand because of its versatility (in terms of soil conditions, climate, and processing requirements). However, starch from different regions and climates has shown dramatically different functionality. Starch extracted from the rainy-season crops (at 6-month and 12-month harvest times) has relatively larger average granule sizes, lower gelatinization temperatures, and higher peak paste viscosities than the drought-seasoned starch samples (Sriroth and others 1999). There is also a significant difference in the amylose content between samples at a given harvest time. Amylose content is higher at the 6-month, than at the 12-month harvest time in the drought crops, but the same in the rainy crops. However, mean amylose chain length (DPn) (number-average degree of polymerization) in rainy-season samples is significantly shorter than that of dry-season samples (Sriroth and others 1999).

NMR Chatakanonda and others (2003b) reported a study (performed under the supervision of Brian Hills at the Institute of Food Research, Norwich, UK) that investigated starch by using water-saturated and deuterium oxide–saturated packed beds of starch granules whose NMR relaxation time was evaluated during both freezing and gelatinization. FID was determined with a single 90° pulse lasting 2 μs with a dwell time of 4 μs (Figure 19.4). T2 was measured by using the Carr-Purcell-Meiboom-Gill (CPMG)

Water Partitioning in Colloidal Systems

259

Free-Induction Decay (FID)

90o 2 μs

τ

τ = 150 μs

T2 CPMG

Recycle Delay 10-15 s ο

90

2 μs

τ

180 4 μs

o

τ

Figure 19.4. Nuclear magnetic resonance pulse sequences used in this study. CPMG, Carr-Purcell-Meiboom-Gill; and FID, free-induction decay.

pulse sequence (90°–180° pulse spacing of 150 μs and full phase cycling; 10- to 15-s recycle delay). A Bruker MSL100 NMR spectrometer (Bruker Spectrospin, Coventry, UK) operating at 100.13 MHz was used in this study. A continuous distribution evaluation of exponentials was performed (Tang and others 2001; Chatakanonda and others 2003a, 2003b) using a model to construct nonnegative least-squares (NNLS) continuous spectra with WinDXP software (Resonance Instruments, Oxfordshire, UK). A continuous distribution of exponentials for FID and CPMG experiments is based on the nonnegative constraints and optimal smoothing: m

gi = ∑ f j e − ti Tj

(19.1)

j =1

where gi is the unknown amplitude of the spectral component at time ti, fj is the preexponential multiplier of the distribution, and Tj is the exponential time constant (the T2 value). The Resonance Instruments WinDXP program solves this equation by minimizing the function: 2

m ⎛ m ⎞ gi = ⎜ ∑ f j e − ti Tj ⎟ + λ ∑ fx2 ⎝ j=1 ⎠ x =1

where λ is the smoothing parameter and

(19.2) m

∑f x=1

2 x

is the linear combination of functions,

added to overcome the ill-conditioning of the problem. By performing a zero-order regularization, λ is a weight factor (expand or narrow the features of the spectrum). More details for continuous models and their application to NMR relaxometry can be found in Ruan and Chen (1998).

260

PART 1: Invited Speakers and Oral Presentations

The contribution of water in different regions was defined by using the continuous model based on the different T2s of water and starch chains in crystalline and amorphous domains. On a large scale, water is localized in the amorphous growth rings and the semicrystalline lamellae. Within the semicrystalline lamellae, however, the amorphous channels between large and small blocklets, and the amorphous region inside the blocklets, also contain some water subpopulations. Inside the crystalline lattice of amylopectin clusters (inside the blocklets), the relaxation time of water is shorter than in the amorphous branch points separating the crystalline structure. Depending on local amylose and amylopectin content and relative chain mobility, water protons in each of these subgranular compartments are expected to have different relaxation times.

The Gelatinization Behavior of Water-Saturated Starch Granules Heating-stage microscopy was applied to capture granule swelling and gelatinization of drought- and rainy-crop cassava starch (Breuninger and others 2009). As shown in Figure 19.5, the rainy-crop starch granules were swollen and gelatinized at a lower temperature than the drought-crop starch of both 6-month and 12-month harvest times. This was also supported by viscoamylograph and differential scanning calorimetry (DSC). Further evaluation of test samples suggested that the growth-season difference impacts the starch synthesis process that produces a larger open amorphous region and less perfect crystalline lattice, as well as amylose chain length, etc. Such differences have a profound impact on the starch pasting property, which is difficult to adjust during food manufacturing. Therefore, understanding further the underlying properties with respect to intragranular local environment would help in identifying the fine structural difference.

Starch-Chain Mobility During Gelatinization Starch-backbone mobility can be observed through 1H-NMR FID relaxation (90° pulse) with sample exchangeable protons replaced by deuterons (a series of deuteration until saturation). Figure 19.6 shows drought and rainy starch (6-month and 12month harvest times) subject to stepwise heating for gelatinization. The FID for the heavy water (D2O)-packed bed of the starch granules pertains to the mobility of the nonexchanging starch CH chains. The 10-μs chain represents the rigid component in the semicrystalline lamellae, and the 800-μs chain represents the more mobile amylopectin of amorphous semicrystalline lamellae (mobile ASLs), which is observed also in waxy maize (Tang and others 2001). As temperature increases, the rigid component (10 μs) decreases proportionately to the increase in the amorphous component (800 μs). Simultaneously, the mobile 800-μs T2 peak emerges, suggesting the increasing mobile starch chain in the ASLs. This reflects a melting transition where the rainy-crop starch melted at temperatures lower than that of the drought crops. This suggested that limited water available led to more tightly packed amylopectin semicrystalline lamellae inside the granules.

Water Partitioning in Colloidal Systems

261

Figure 19.5. Granular swelling upon gelatinization of cassava starch grown in drought and rainy seasons and harvested at 6 and 12 months. Starch slurry (0.5% wt/wt) was heated in water on a heating-stage microscopy. From Breuninger and others (2009).

Three mobile starch-chain populations were identified, those with a T2 of 1-, 20-, and 50-ms (Chatakanonda 2003; Chinachoti and others 2006). CPMG 1H relaxometry was applied to the drought- and rainy-season cassava starch (54% water, dry basis) upon heating stepwise from 20° to 90°C (Chatakanonda 2003). The T2 distribution indirectly represents the dynamic states of the starch chains observed through the neighboring diluent. The more rigid the starch chains are, the shorter is the T2 of the exchangeable starch hydroxyl protons and, therefore, the shorter is the T2 of the water protons (Hills 1992). Figure 19.7 illustrates the T2 populations with corresponding peaks associated with (a) the water inside the granules (3 ms), (b)

262

PART 1: Invited Speakers and Oral Presentations

Figure 19.6. Starch-chain mobility T2 distribution as measured by proton 90° pulse free-induction decay of deuterated cassava starch saturated with heavy water (D2O) during gelatinization. From Chatakanonda (2003) and Chinachoti and others (2006). Cassava starch from drought and rainy seasons for 6- and 12-month harvest times: D6, drought season 6-month harvest; D12, drought season 12-month harvest; R6, rainy season 6-month harvest; and R12, rainy season 12-month harvest.

the water outside the granules (20 ms), and (c) the extragranular population trapped in the leached amylose network (100 ms). The latter began only after gelatinization was initiated, which had been shown earlier in cassava starch to include a significant amylose leaching through channel enlargement (Notté 1993; Garcia 1996; Gallant and others 1997). This was also confirmed by an observation that the amylose-iodine complex increased at gelatinization onset (Chatakanonda 2003). The amylose-iodine complex (observed as blue strands under the light microscope) was first seen on the outside edges of the granules at ≥65° and ≥60°C for drought- and rainy-cassava starches, respectively. Approximately 40% of the starch granules from rainy and drought crops showed amylose leaching at 65° and 70°C, respectively. At higher temperatures, the proportions of granules with leached amylose increased up to 100% at around 70° and 90°C for starch from rainy and drought seasons, respectively. This helps explain the appearance of the 100-ms peak in the T2 distribution assigned to the water in the leached amylose fraction (Figure 19.7).

Water Partitioning in Colloidal Systems

Gelatinization of Cassava Starch (D12)

Gelatinization of Cassava Starch (R12) Rainy 12 months

Drought 12 months

o

40 oC

40 C

50 oC

50 oC

o

60 oC

70 C

o

70 C

80 oC

80 C

90 oC

90 C

Intensity

60 C

102

103

104 Time (μs)

263

105

o

o

o

106 102

103

104 Time (μs)

105

106

Figure 19.7. Water mobility measured by Carr-Purcell-Meiboom-Gill (CPMG) pulse sequence. T2 proton relaxation time distribution of cassava starch (54% water) upon heating. From Chatakanonda (2003). Cassava starch from drought and rainy seasons for 6- and 12-month harvest times: D6, drought season 6-month harvest; D12, drought season 12-month harvest; R6, rainy season 6-month harvest; and R12, rainy season 12-month harvest.

There was more subtle observation that could be meaningful to the functionality difference. For instance, the water peaks for R12 (rainy 12-month cassava starch) showing relatively longer relaxation times could be related to the more open structure of the rainy crop (lower gelatinization temperature).

The Freezing Behavior of Water-Saturated Starch Granules Figure 19.8 depicts proton CPMG T2 distribution spectra for the water-saturated packed-bed cassava samples (54% moisture, dry basis) after rapid freezing and then heating were evaluated. At −62°C, only the short-relaxing (200 μs) component was observed. Note that ice cannot be detected at the millisecond timescale of the CPMG pulse sequence. At −43°C, the unfrozen or unfreezable water emerged showing a broad peak centered at 400 μs. Intragranular unfrozen water is likely to be closely interacting with the surface of amylose and amylopectin molecules. At −23° and −3°C,

264

PART 1: Invited Speakers and Oral Presentations

Figure 19.8. Water mobility measured by Carr-Purcell-Meiboom-Gill (CPMG) pulse sequence. The temperature dependence of proton transverse relaxation time distribution for cassava starches saturated with water (54% moisture, dry basis) during freezing. From Chatakanonda (2003) and Chinachoti and others (2006). Cassava starch from drought and rainy seasons for 6- and 12-month harvest times: D6, drought season 6-month harvest; D12, drought season 12-month harvest; R6, rainy season 6-month harvest; and R12, rainy season 12-month harvest.

the 200-μs component and the longer-relaxation (1 ms) component (possibly starch) coexisted. At above zero, a substantial appearance of liquid water was observed at 20 ms (probably melted extragranular water). The 3-ms and 200-μs components became negligible after thawing. In summary, NMR-relaxometry data showed that the starch chains from the rainy crop more readily swell and gelatinize. The shorter mean amylose chain length in the rainy 12-month crop facilitated the release of amylose upon heating more than that in the rainy 6-month crop. The presence of more closely packed amylopectin chains in the drought-season starches made it more difficult for the starches to swell and gelatinize (based on analysis of water at the amylose gel and rigid amylopectin peaks). The NMR T2 distribution results have shown that it can be used to follow phase changes, as well as other functional changes such as swelling and networking of the amorphous polymers inside and outside of starch granules. This provides information that cannot easily be determined on intact starch granules.

Acid Hydrolysis Starch acid-hydrolysis treatment is known to erode or destroy first the amorphous, and then the crystalline chains (Shi and Seib 1992; Morrison and others 1993; Chun and

Water Partitioning in Colloidal Systems

265

Table 19.1. Differential scanning calorimetry (DSC) data for acid-hydrolyzed cassava starch indicating double-helix destruction transition that broadens in temperature range upon acid hydrolysis as starch double-helix chains erode into various fragments Starch sample

Hydrolysis time (h)

6% HCl hydrolysis

Tp (°C)

Tc (°C)

ΔH (J/g)

Temperature range (Tc − To)

63.9 ± 0.2a

70.5 ± 0.1a

82.7 ± 1.0a

8.5 ± 0.1a,d,i

18.8

0

63.2 ± 0.4a

70.2 ± 0.1a

82.3 ± 0.4a

8.5 ± 0.2a,d,g,i

19.1

12

67.3 ± 0.2b

73.9 ± 0.1b

89.8 ± 1.2b

9.1 ± 0.0b,g,h

22.5

24

70.0 ± 0.0c

77.2 ± 0.1c

94.5 ± 0.8c,e 9.7 ± 0.1c,e,f,h

24.5 23.9

Native Control

To (°C)

48

72.0 ± 0.0d

79.3 ± 0.0d

95.9 ± 1.1c

8.8 ± 0.2a,b,g

96

71.5 ± 0.1d

81.4 ± 0.0e

98.6 ± 1.1d

9.4 ± 0.1b,c,h

27.1

192

61.5 ± 0.5e

74.2 ± 0.3b

92.9 ± 0.7e

8.5 ± 0.0a,d,g,i

31.4

384

52.3 ± 0.6f

73.9 ± 0.4b

101.8 ± 0.6f,i

8.2 ± 0.0d,i

49.5

768

53.3 ± 1.1g

73.6 ± 0.1b

103.3 ± 1.4f

8.0 ± 0.1d,i

50.0

Means with the same letter in each column are not significantly different (α = 0.05). From Atichokudomchai and others (2002).

others 1997; Jacobs and others 1998; Atichokudomchai and others 2002). DSC applied to follow the thermal transition of lintnerized tapioca starch in excess water found that the starch’s endotherm appears over a broader temperature and at higher onset and peak temperatures when compared with native starch endotherm (Chun and others 1997; Jacobs and others 1998). Since acid hydrolysis preferentially attacks the amorphous regions in granules, the crystallites are decoupled from the amorphous chains (i.e., no longer destabilized by the amorphous parts). Consequently, the starch crystallites of the acid-modified starches melt at a higher temperature, and the transition temperature range is broader due to increases in the relative short-range double-helix contents and long-range crystallinity (Jacobs and others 1998; Atichokudomchai and others 2004). Gelatinization involves both changes in the amorphous regions that may impact melting of the crystalline regions (Jane and others 1999). This coupling process is disturbed when acid hydrolysis destroys the amorphous chains. The loss of the amorphous domains and corresponding swelling before gelatinization could result in melting endotherm broadening. When extensive hydrolysis occurs, the broader melting transition results from the shorter and more heterogeneous chain length of the double helices of amylopectin. The unchanged enthalpy indicates that the degree of molecular order (i.e., relative hydrogen bonds broken) remains unchanged with hydrolysis. But shorter amylopectin chain length can lead to a decrease in the onset temperature, whereas the degree of structural order may not be affected (Donovan and others 1983; Moates and others 1997; Jane and others 1999; Waigh and others 2000; Atichokudomchai and others 2002, 2004). The CPMG T2 distribution of water is shown in Figure 19.9 for acid-hydrolyzed starch and the control. Predominant in the two populations was the extragranular water population at T2 around the 100-ms range. When heated above the gelatinization

Figure 19.9. Water mobility measured by Carr-Purcell-Meiboom-Gill (CPMG) pulse sequence. The temperature dependence of proton transverse relaxation time distribution for acid-modified cassava starches diluted with water (80% moisture) as measured by a DPX Maran NMR spectrometer (Resonance Instruments, Whitney, UK) during heating. Data show spectra for 0, 12, and 24 h of acid hydrolysis. e, exponential (for example, 1e+1 = 1+1). From Atichokudomchai (unpublished data).

266

Water Partitioning in Colloidal Systems

267

temperature, the intragranular component in the control increased at >65°C. Simultaneously, the extragranular water peak broadened, skewing toward the higher temperature. This was associated with the leaching of the amylose-forming network and opening of granular channels and melted starch chains that allowed water to distribute rapidly. But broader water T2 distribution resulted from the great degree of intermolecular networking of amylose and partly of amylopectin linear chains. On the other hand, acid-hydrolyzed starches (12 and 24 h) that had lost amylose and amorphous amylopectin through acid erosion showed a different behavior. A smaller (very little) contribution by the intragranular water was observed at all temperatures, suggesting that this population was associated with the presence of intact amorphous chains. In addition, the extragranular water population showed a very slight broadening and shifting, suggesting that the absence of amylose networking beyond the granules and the shorter amylopectin chains resulted in much less heterogeneity after the crystalline starch melting. This was likely because the average molecular weight is greatly reduced. The presence of heterogeneous amorphous starch chains in the control sample helps demonstrate the importance of the intact amorphous amylose and amylopectin needed in granular swelling, which opens channels for the amylose chains to leave the granule and form the extragranular network. The hydrated amorphous regions in the semicrystalline lamellae and regions between clusters and blocklets could play a role in maintenance of a certain degree of “imperfection” by imposing certain torsion strains on the helical structure and reducing the melting temperature onset of the starch. Removing the amorphous components resulted in a higher degree of short-range (13C NMR) and long-range (X-ray) molecular order increasing the melting temperature even though the total enthalpy (number of hydrogen bonds) remained relatively unchanged.

Syneresis and Freeze-Thaw Stability The starch and water T2 distribution offers a way to evaluate key differences in functionality and molecular origin of gelatinization. For example, a comparison of gelatinized wheat flour, cassava-starch rice flour, rice starch, and waxy rice starch was studied (Waraho 2003). T2 distribution broadening can be used as a marker of the degree of cross-linking or networking observed during the five freeze-thaw cycles in some starch samples with significant syneresis, such as native rice and cassava starches (Figure 19.10). In the case of freeze-thaw stable starch (waxy rice), the T2 distribution remained in a relatively narrow range (with only a very small decrease in average T2). This indicates little or no networking or cross-linking. Native rice starch, on the other hand, showed the widest T2 broadening, whereas cassava starch exhibited the largest drop in average T2 after the same freeze-thaw period. Another syneresis study of pectin gel found that calcium-induced cross-linking broadened the T2 distribution. This was also seen with the aged, refrigerated gel when it was aged (syneresis) with some downward averaged T2 shift. The 1H T2 distribution approach is quite useful in understanding freeze-thaw stability and in detection of networking of hydrocolloid gel systems.

268

PART 1: Invited Speakers and Oral Presentations

T2 distribution

Intensity

wheat starch cassava starch rice flour rice starch waxy rice starch

1e+2

1e+3

1e+4

1e+5

1e+6

1e+7

1e+8

T2 (μs)

Figure 19.10. T2 distribution of protons in gelatinized starch or flour suspensions that had undergone five freeze-thaw cycles. The solid lines depict before and the dashed lines after the freeze-thaw treatments. The extent of syneresis can be observed from the water redistribution as water is squeezed out from starch gel network. Exponential, e (for example, 1e+2 = 1+2). From Waraho (2003).

Conclusions Water can distribute in starch granules at locations dependent on the local starch domain hydration sites and the conforming spatial flexibility. Upon heating, drying, or freezing, changes in thermal energy induce water- and starch-molecule phase changes that can reversibly and irreversibly modify the properties of the starch. Presented here was an example of an investigative approach for describing starch functionality that can be tied to specific molecular events. The heterogeneous nature of water in various starch domains is an area that links the missing gaps between basic physical properties and the way in which the starch molecules are put together (their fine structure). This enables us to better understand the nature of starch granules and the seasonal change or modification that affects their properties.

References Atichokudomchai N, Varavinit S, Chinachoti P. 2002. Gelatinization transitions of acid-modified tapioca starches by differential scanning calorimetry (DSC). Starch 54:296–302. Atichokudomchai N, Varavinit S, Chinachoti P. 2004. A study of ordered structure in acid-modified tapioca starch by 13C CP/MAS solid-state NMR. Carbohydr Polym 58:383–9. Baguma Y. 2004. Regulation of starch synthesis in cassava [PhD diss]. Uppsala: Department of Plant Biology and Forest Genetics, Swedish University of Agricultural Sciences. Belton PS, Colquhoun IJ, Hills BP. 1993. Applications of NMR to food sciences. Annu Rep NMR Spectrosc 26:1–53.

Water Partitioning in Colloidal Systems

269

Belton PS, Hills BP. 1987. The effect of diffusive exchange in heterogeneous systems on NMR line shapes and relaxation processes. Mol Phys 61:999–1018. Breuninger WF, Sriroth K, Piyachomkwan K. 2009. Tapioca/cassava starch: production and use. In: Whistler RL, BeMiller JN, editors. Starch: chemistry and technology. 3rd ed. New York: Elsevier. p 541–68. Chatakanonda P. 2003. Water and starch chain mobility in cassava starch as monitored by NMR: effects of heat-moisture treatments, growth conditions and harvest time [PhD diss]. Amherst: Department of Food Science, University of Massachusetts. Chatakanonda P, Chinachoti P, Sriroth K, Piyachomkwan K, Chotineeranat S, Tang HR, Hills B. 2003b. The influence of time and conditions of harvest on structure-function properties of cassava starch: a proton NMR relaxation study. Carbohydr Polymer 53:233–40. Chatakanonda P, Dickinson LC, Chinachoti P. 2003a. Mobility and distribution of water in cassava and potato starches by 1H and 2H NMR. J Agric Food Chem 51:7445–9. Chinachoti P, Vittadini E, Chatakanonda P, Vodovotz Y. 2006. Characterization of molecular mobility in carbohydrate food systems by NMR. In: Webb GA, editor. Modern magnetic resonance. Part 3: Application in materials science and food science. Dordrecht, The Netherlands: Springer. p 1675–9. Chun J, Lim S, Takeda Y, Shoki M. 1997. Properties of high-crystalline rice amylodextrins prepared in acid-alcohol media as fat replacers. Cereal Foods World 42:813–9. Donovan JW, Lorenz K, Kulp K. 1983. Differential scanning calorimetry of heat-moisture treated wheat and potato starches. Cereal Chem 60:381–7. Gallant DJ, Bouchet B, Baldwin PM. 1997. Microscopy of starch: evidence of a new level of granule organization. Carbohydr Polym 32:177–91. Garcia V. 1996. Transitions thermiques de l’amidons de manioc en mililieux peu hydratés [Thermal transitions of cassava starch in low-hydrated milieu] [PhD thesis]. Paris-Grinon: Institut National Agronomique [National Institute of Agronomy]. Geigenberger P, Reimholz R, Deiting U, Sonnewald U, Stitt M. 1999. Decreased expression of sucrose phosphate synthase strongly inhibits the water stress-induced synthesis of sucrose in growing potato tubers. Plant J 19:119–29. Hills BP. 1992. The proton exchange cross-relaxation model of water relaxation in biopolymer systems. Mol Phys 76:489–508. Hizukuri S. 1986. Polymodal distribution of the chain lengths of amylopectins and its significance. Carbohydr Res 147:342–7. Jacobs H, Delcour JA. 1998. Hydrothermal modifications of granular starch, with retention of the granular structure: a review. J Agric Food Chem 46:2895–905. Jacobs H, Eerlingen CR, Rouseu N, Colona P, Delcour AJ. 1998. Acid hydrolysis of native and annealed wheat, potato and pea starches: DSC melting features and chain length distributions of lintnerised starches. Carbohydr Res 308:359–71. James M, Denyer K, Myers A. 2003. Starch synthesis in the cereal endosperm. Curr Opin Plant Biol 6:215–22. Jane J, Chen YY, Lee FL, McPherson EA, Wong KS, Radosavlievic M, Kasemsuwan T. 1999. Effects of amylopectin branch chain length and amylose content on the gelatinization and pasting properties of starch. Cereal Chem 76:629–37. Jenkins PJ, Cameron RE, Donald AM. 1993. A universal feature in the structure of starch granules from different botanical sources. Starch/Stärke 45:417–20. Kroeker RM, Henkelman RM. 1986. Analysis of biological NMR relaxation data with continuous distribution of relaxation times. J Magn Reson 69:218–35. Lillford PJ, Clark AH, Jones DV. 1980. Distribution of water in heterogeneous foods and model systems. In: Rowland SP, editor. Water in polymers. Washington, DC: American Chemical Society. p 177–95. Manners DJ. 1989. Recent developments in our understanding of amylopectin structure. Carbohydr Polym 11:87–112. Menon RS, Allen PS. 1991. Application of continuous relaxation time distribution to the fitting of data from model systems and excised tissue. J Magn Reson 86:214–27.

270

PART 1: Invited Speakers and Oral Presentations

Moates GK, Noel TR, Parker R, Ring SG. 1997. The effect of chain length and solvent interactions on the dissolution of the B-type crystalline polymorph of amylose in water. Carbohydr Res 298:327–33. Morrison WR. 1995. Starch lipids and how they relate to starch granule structure and functionality. Cereal Foods World 40:437–46. Morrison WR, Tester RF, Gidley MJ, Karkalas J. 1993. Resistance to acid hydrolysis of lipid-complexed amylose and lipid-free amylose in lintnerised waxy and non-waxy barley starches. Carbohydr Res 245:289–302. Munyikwa TRI, Langeveld S, Salehuzzaman SNIM, Jacobsen E, Visser RGF. 1997. Cassava starch biosynthesis: new avenues for modifying starch quantity and quality. Euphytica 96:65–75. Newcomb CH, Graham SJ, Bronskill MJ. 1990. Effects of nonlinear signal detection on NMR relaxation time analysis. J Magn Reson 90:279–89. Notté C. 1993. Structure et fonctionnalité des amidons modifiés chimiquement [Functionality and structure of chemically modified starches]. Nantes, France: Faculty of Sciences and Technology, University of Nantes. Oates CG. 1997. Towards an understanding of starch granule structure and hydrolysis. Trends Food Sci Technol 81:275–82. Oostergetel GT, van Bruggen EFJ. 1993. The crystalline domains in potato starch granules are arranged in a helical fashion. Carbohydr Polym 21:7–12. Provencher SW. 1982a. A constrained regularization method for inverting data represented by linear algebraic or integral equations. Comput Phys Commun 27:213–27. Provencher SW. 1982b. Contin: a general purpose constrained regularization program for inverting noisy linear algebraic and integral equations. Comput Phys Commun 27:229–42. Raemakers K, Schreuder M, Suurs L, Furrer-Verhorst H, Vinken JP, de Vetten N, Jacobsen E, Visser RGF. 2005. Improved cassava starch by antisense inhibition of granule-bound starch synthase I. Mol Breeding 16:163–72. Ruan RR, Chen PL. 1998. Water in foods and biological materials: a nuclear magnetic resonance approach. Lancaster, PA: Technomic. Shi YC, Seib PA. 1992. The structure of four waxy starches related to gelatinization and retrogradation. Carbohydr Res 227:131–45. Sriroth K, Piyachomkwan K, Santisopasri V, Oates CG. 2001. Environmental conditions during root development: drought constraint on cassava starch quality. Euphytica 120:95–102. Sriroth K, Santisopasri V, Petchalanuwat C, Kurotjanawong K, Piyachomkwan K, Oates CG. 1999. Cassava starch granule structure-function properties: influence of time and conditions at harvest on four cultivars of cassava starch. Carbohydr Polym 38:161–70. Tang HR, Brun A, Hills BP. 2001. A proton NMR relaxation study of the gelatinization and acid hydrolysis of native potato starch. Carbohydr Polym 46:7–18. Tang HR, Godward J, Hills BP. 2000. The distribution of water in native starch granules: a multinuclear NMR study. Carbohydr Polym 43:375–87. Tellier C, Mariette F, Guillement J, Marchal P. 1993. Evolution of water proton nuclear magnetic relaxation during milk coagulation and syneresis: structural implications. J Agric Food Chem 41:2259–66. Waigh TA, Gidley MJ, Komanshek BU, Donald AM 2000. The phase transformations in starch during gelatinization: a liquid crystalline approach. Carbohydr Res 328:165–76. Waraho T. 2003. Comparative study of freeze-thaw stability of starch gels [master ’s thesis]. Amherst: Department of Food Science, University of Massachusetts. Zobel HF, Stephen AM. 1995. Starch: structure, analysis and application. In: Stephens AM, editor. Food polysaccharides and their applications. New York: Marcel Dekker. p 19–66.

20 Physical Changes in Confectionery Products Caused by the Availability of Water, with a Special Focus on Lactitol Crystallization M. H. Lim, B. Lampen, L. F. Siow, and T. Rades

Abstract Crystallization of sugars produces significant effects on the sensory and physical properties of food. Whereas crystallization of sucrose has received much attention, crystallization of polyols has not been well researched. Lactitol ([+]-4-O-β-Dgalactopyranosyl-D-glucitol), a polyol, is being used increasingly as a sweetener in foods. This project studied the crystallization behavior of lactitol in order to better control the processing and quality of lactitol products. It was studied using two main approaches: (a) the melting enthalpy of the crystallizing lactitol samples and (b) their moisture sorption. The degree of crystallinity of lactitol during storage was determined by the melting enthalpy during storage. An iterative modeling equation was developed to describe the crystallization behavior of lactitol based on its moisture-sorption pattern. This model was based on the moisture adsorption of the amorphous lactitol, the moisture desorption of the crystalline lactitol, and the crystallizing lactitol, which was modeled using the Avrami equation. The rate of crystallization increased with increased temperature and relative humidity (RH). Lactitol crystallized down to 30% RH at 20°C and down to 22% RH at 32°C. The glass transition temperature of lactitol was successfully used to predict whether crystallization would occur at different storage temperatures and relative humidities.

Introduction The major component of confectionery products (candies) is sugars. Sugars may exist in different physical states, including crystalline, glassy, rubbery, melt, and solution. These different physical states enable a wide variety of confectionery products to be produced that differ in their textural attributes, even though the composition of the confectionery products remains the same. This can be illustrated in comparing hard candies to fondants and tablets for all of which the main ingredient is sucrose. Whereas sucrose crystals are considered a defect in hard candies, they are the source of the desired structure and texture of fondant and tablets. To maintain the quality of confectionery products, their water activity should be kept as low as possible (within their optimal range). The stability of confectionery products also depends on their storage temperatures. Below the glass transition temperature (Tg), some hard candies, such as lollipops and hard caramels, are generally 271

272

PART 1: Invited Speakers and Oral Presentations

stable. However, above the Tg beyond which the products are in the rubbery state, sugars may undergo crystallization caused by the increase in molecular mobility and diffusion, as the viscosity of the product matrix decreases above Tg after the absorption of moisture from the surrounding environment (Slade and Levine 1991). At a constant temperature, crystallization of lactose depends on relative humidity (RH) and water content during storage (Roos and Karel 1992). The two steps that limit crystal growth are mass transfer and surface integration (Hartel 2001). Mass transfer involves the movement of molecules from a bulk solution to the surface of a crystal lattice, and surface integration involves the orientation of molecules with the right sizes and shapes to form part of an existing crystal lattice (Hartel 2001). The degree and mode of crystallization affect the viscosity, hygroscopicity, grain size, and consistency of the confectionery products. Whereas crystallization of simple sugars has received much attention, crystallization of polyols, which are the alternative sweeteners, has not been well researched. Polyols are sugar alcohols or polyhydric alcohols with less sweetness than sucrose but also many fewer calories. The market demand for sugar-free products is increasing. Sugar alcohols can be manufactured into sugar-free confectionery products in the forms of hard-boiled sweets, chocolate, chewing gum, directly compressed products, chewy sweets, jellies, hard gums, soft gums, marshmallows, nougat, and cream fillings (Sicard and Le Bot 1994). In the current study, the crystallization behavior of lactitol, a polyol, was studied because lactitol has been increasingly used as a food sweetener. The objectives of the current study were to investigate the effect of RH, storage temperature, and moisture content on the crystallization behavior of lactitol and to relate that behavior to its glass transition temperatures.

Materials and Methods Preparation of Amorphous Lactitol Lactitol (Cultor Food Science, Kotka, Finland) and Milli-Q grade water were mixed for at least 1 h at room temperature to produce a 10% (wt/wt) solution. The solution was placed into shallow plastic trays to a depth of ∼2 cm and kept at −80°C for at least 24 h. The trays were then covered with aluminum foil with pinholes and freezedried for at least 72 h using a Sentry vacuum freeze dryer (VirTis, Gardiner, NY, USA). The freeze dryer has an initial chamber temperature of −50°C and an ultimate vacuum below 10.7 kPa. After freeze drying, the samples were transferred to desiccators that contained phosphorus pentoxide (P2O5) and stored at room temperature for at least 2 weeks before further examination. Preparation of Saturated Salt Solutions Saturated salt solutions were prepared using Milli-Q grade water and salts to produce equilibrium RH ranging from 7% to 97% as shown in Table 20.1. Airtight plastic vacuum containers with salt solutions were left for at least 48 h at 20°C (room temperature) and 32°C (in an incubator) for equilibration before the samples were placed in the containers. The RH and temperature of the containers were measured by placing

Physical Changes in Confectionery Products

273

Table 20.1. Saturated salt solutions of varying relative humidity (RH) measured at 20° and 32°C Salt

Measured %RH 20°C

LiBr

32°C

7

LiCl

11

13

CHCOOK

21

22

MgCl2

30

33

K2CO3

39

41*

Ca(NO3)2

49*

52

NH3NO3

58*

KI

63

64*

(NH4)2SO4

81*

81*

KCl

85

BaCl2

90

K2SO4

97

* Relative humidities used in this study to measure the melting enthalpy, melting temperature, and glass transition temperature of samples.

an HM34 humidity and moisture meter (Vaisala Sensor Systems, Helsinki, Finland) into the containers. Measurements were taken after the samples were left for at least 12 h in the containers. Determination of Moisture Content A glove bag, which is a thick plastic film bag with base dimensions of approximately 1 × 1 m and a height of ∼0.5 m with attached gloves, was produced to reduce moisture sorption during the handling of the samples. The glove bag was flushed with dry air for at least 5 min before samples were introduced. In the glove bag, ∼0.5 g of sample was transferred into 2-mL glass vials for moisture determination. Moisture content was determined by using a 736 GP Titrino KarlFisher titrator (Metrohm, Zofingen, Switzerland) calibrated at least three times using the Milli-Q standard water. Moisture content was taken in triplicate, and the readings were averaged for the final moisture content. Modeling of Moisture Content of Crystallizing Amorphous Lactitol Moisture content of the crystallizing amorphous lactitol over time was modeled by using both Equation 20.1 (the modified Avrami equation) (Avrami 1939) and Equation 20.2. Several parameters, including the degree of crystallization (α); time (t); moisture sorption of the amorphous material (Meq,A = equilibrium moisture content and kA = rate constant); crystallization behavior (i = induction time, k = crystallization rate constant, and n = Avrami index); and moisture sorption of the crystalline material (Meq,X = equilibrium moisture content and kX = rate constant) were determined (Lampen 2000).

274

PART 1: Invited Speakers and Oral Presentations

α = 1 − exp ( − k ( t − i )

n

)

where t ≥ i, else α = 0

M t + dt = [αk x ( M eq , X − M t ) + (1 − α ) k A ( M eq , A − M t )] dt

(20.1) (20.2)

Sample Preparation for Differential Scanning Calorimetry Amorphous lactitol that had been dried over P2O5 was hygroscopic; therefore, samples were prepared in a glove bag flushed with dry air. Samples (6–9 mg) were placed in aluminum pans, sealed, and then placed in desiccators containing P2O5 for at least 72 h. Samples were then weighed to ±0.001 mg by using a Sartorius Electronic Ultra Microbalance (model 4431 MP8; Sartorius Instruments, Epsom, UK). These amorphous samples were used for glass transition and melting characterization. Determination of Melting Enthalpy, Melting Temperature, and Glass Transition Temperature Lactitol was placed into containers of 49%, 58%, and 81% RH at 20°C, and 41%, 64%, and 81% RH at 32°C. The samples were removed at intervals of no less than 90 min. A differential scanning calorimeter Pyris 1 (Perkin Elmer, Shelton, CT, USA) was used to determine the glass transition temperatures, melting enthalpies, and melting temperatures of lactitol samples by scanning the samples over a temperature range of between 50° and −130°C at 10°C/min. The calorimeter was equipped with an Intracooler II (FIS, New York, NY, USA) and was calibrated and checked for temperature and heat flow by using Milli-Q grade water (melting temperature [Tm], 0°C; enthalpy of fusion [ΔHm], 333.8 J/g) and indium (Tm, 156.6°C; ΔHm, 28.45 J/g). The melting enthalpy of the crystallizing lactitol increased at a high rate over time to an enthalpy that was specific to the crystallization conditions (RH and temperature). For the purpose of this study, this value was determined to be the enthalpy of maximum crystallization (Hm,max). Modeling of Glass Transition Temperature The Gordon-Taylor equation (Equation 20.3), where Tg1 and Tg2 are the glass transition temperatures of component compounds, k is a constant, and w1 and w2 are the weight fractions of the components, was modified as shown in Equation 20.4 to model the effect of moisture content on the glass transition temperature. Tg,1 and Tg,w are the glass transition temperatures of anhydrous lactitol and water, respectively, k is a constant, and M is the moisture content on a wet-weight basis. The value of −135°C was used for the Tg for water (Roos 1995). Tg =

w1Tg1 + kw2 Tg 2 w1 + kw2

(20.3)

Tg =

(1 − M ) Tg,l + kMTg,w (1 − M ) + kM

(20.4)

Physical Changes in Confectionery Products

275

Glass transition temperature was modeled using Equation 20.4. The squares of the difference between modeled and experimental values for Tg were calculated (i.e., the residuals). The square root of the sum of these values was then calculated to judge how well the model fit the experimental values. The unknown parameters (k and Tg,l) were then optimized using the Solver add-in software (Frontline Systems, Incline Village, NV, USA) in Microsoft Excel to minimize the root sum of the residuals.

Results and Discussion Melting Temperatures of Lactitol Crystallites The thermograms from the melting of lactitol at 49% RH and 20°C show two overlapping endothermic peaks with onset temperatures at 69°–70°C and 90°–91°C (thermogram b in Figure 20.1). The first peak (short arrow) was initially large and diminished over time as the second peak (long arrow) became larger. The temperatures of these endotherms were similar to the temperatures of endotherms of ground lactitol observed by Koichi and others (1997). After 20 h of storage, only a single peak was observed with onset temperature at around 97.2° ± 0.6°C (thermograms c and d in Figure 20.1), which was similar to the reported values of 94°–97°C for the melting temperature of lactitol monohydrate (Van Velthuijsen 1979). In addition, a small endothermic peak was observed between −25°

Heat flow endo up (mW)

50 40 d (192 hours)

30 20 10

c (20 hours) * b (11 hours) a (5 hours)

0 –50

0 50 Temperature (°C)

100

Figure 20.1. Thermograms of lactitol crystallization at 20°C and 49% RH. (a) After 5 h of storage, a glass transition (ˆ) occurred at around 10°C and a small peak around 70°C (short arrow). (b) After 11 h of storage, there appeared to be two peaks between 60° and 100°C and a thermal event above 100°C. (c) After 20 h of storage, a small endothermic peak (endo) occurred below 0°C (*), the first peak at around 70°C had diminished, and the peak at 90°C (long arrow) was predominant. (d) After 120 h of storage, only one peak (long arrow) existed. mW, milliwatts.

276

PART 1: Invited Speakers and Oral Presentations

and 0°C (thermograms c and d in Figure 20.1). This endothermic peak occurred only in scans where no glass transition was observed and when the degree of crystallization was apparent. This peak was believed to be due to the melting of water that was liberated during crystallization, as several previous studies reported that lactose crystallization resulted in a loss of sorbed water at the corresponding RH (Lai and Schmidt 1990; Karmas and others 1992; Labrousse and others 1992; Ozimek and others 1992). Rate and Extent of Crystallization Lactitol stored either at 20°C at 49%, 58%, and 81% RH; or 32°C at 41%, 64%, and 81% RH, was measured for its melting temperature at a fixed time interval (Figure 20.2). Lactitol crystallization occurred in a time-dependent manner. Onset of melting

(a)

Melting temperature (°C)

100 95 90 85 80 75 49%

70

58% 81%

65 60 0

5

10

15 Time (h)

20

25

30

(b)

Melting temperature (°C)

90 85 80 75 70

41%

65

64% 81%

60 0

5

10

15 Time (h)

20

25

30

Figure 20.2. Melting temperatures of lactitol stored at (a) 20°C at 49%, 58%, and 81% RH and (b) 32°C at 41%, 64%, and 81% RH.

Physical Changes in Confectionery Products

277

temperatures increased from ∼70° to ∼90°C with increasing storage time for all the storage conditions, with an exception for storage of lactitol at 32°C and 41% RH, which remained around 70°C up to 10 h (Figure 20.2). The leveling off of melting temperatures at around 90°C suggested that the extent of lactitol crystallization had reached a maximum. At higher storage temperatures, the time needed for the melting temperatures to level off was shorter than at the lower storage temperatures (Figure 20.2). This had been previously reported in the leveling off of the melting enthalpies of corn starch (Jouppila and Roos 1997) and potato starch (Nakazawa and others 1985), which indicated the maximum extent of crystallization. The leveling off of melting temperatures of lactitol crystallites was achieved more rapidly at the storage conditions with higher RH. A similar observation of the increase in crystallization rate to a maximum as the RH increased during storage had also been reported in an X-ray diffraction study of lactose crystallites (Jouppila and others 1997). To determine the extent of lactitol crystallization, the melting enthalpy of lactitol crystallites stored at 20° or 32°C at the various RHs was measured, and a schematic diagram was plotted as shown in Figure 20.3. After the initial induction period, the melting enthalpy of the lactitol crystallites appeared to reach a plateau. The plateau in melting enthalpies was higher for lactitol stored at 32°C compared with 20°C (Table 20.2), which could be related to the T − Tg conditions as discussed in the next section. At each temperature, the plateau in melting enthalpies was higher at the medium RH compared with both higher and lower RHs (Table 20.2). At low RH, the lower extent of crystallization could probably be due to the lower molecular mobility and diffusion,

slow enthalpy change final melting enthalpy

plateau melting enthalpy

Moisture content

induction period

Melting enthalpy

rapid enthalpy change

Time

Figure 20.3. Schematic diagram showing melting enthalpy measured as a function of time. The induction period includes moisture sorption of the samples and the time of nuclei formation. When crystallization occurs, there is a rapid enthalpy change to a plateau melting enthalpy. Then a slow enthalpy change occurs where the migration of water and maturation of crystals occur and melting enthalpy increases to a final melting enthalpy.

278

PART 1: Invited Speakers and Oral Presentations

Table 20.2. Plateau melting enthalpies (ΔHm) for lactitol crystallization at 20° and 32°C storage temperatures and relative humidities, measured after the initial period of crystallization Temperature (°C) % Relative humidity

20 49

20 58

20 81

32 39

32 64

32 81

ΔHm at plateau (J/g)

95

115

105

120

130

120

as was suggested for crystallization of amorphous lactose and starch (Jouppila and Roos 1997; Jouppila and others 1997). Furthermore, at low moisture content that occurred at low RH, the difference between the storage temperature and the glass transition is small, and a high nucleation rate relative to crystal growth produces many small crystals (Slade and Levine 1991). The many small crystals may also inhibit further crystallization because all the small crystals may not be able to orient themselves into a crystal lattice (Jouppila and others 1997). On the other hand, at high RH, the extent of crystallization decreased, probably because of the higher moisture content and increased solubilization of the lactitol in sorbed water. In parallel to the initial increase in melting enthalpy, the moisture content of lactitol also increased (Figure 20.3). However, once the melting enthalpy leveled off, which indicated that the lactitol had maximally crystallized, the moisture content decreased to the equilibrium moisture content. At storage temperatures of 20° and 32°C, the moisture content of amorphous lactitol increased to a maximum and then decreased to reach equilibrium with increasing storage time, as shown in Figure 20.4. At storage conditions where the lactitol crystallized (29% RH and above), the moisture content of the amorphous samples was modeled using Equations 20.1 and 20.2. The extent of crystallization could also be indicated by the melting enthalpy of water because water was liberated during crystallization at a critical RH, as reported for lactose crystallization (Lai and Schmidt 1990; Karmas and others 1992; Labrousse and others 1992; Ozimek and others 1992). The melting enthalpy of water increased with increasing moisture content of the lactitol (Figure 20.5), which suggests that the absorption of water by the glassy lactitol resulted in plasticization and hence caused massive crystallization. The melting enthalpy and moisture content of lactitol stored at 20°C were generally greater than at 32°C. At higher RH, the water melting enthalpy of lactitol crystallites was higher than those stored at lower RH. This could be related to the higher moisture content of lactitol stored at the higher RH compared with the lower RH.

Glass Transition Temperature Glass transition temperatures (Tg) during storage at 20°C at 49%, 58%, and 81% RH; and at 32°C at 39%, 64%, and 81% RH, are shown in Figure 20.6. The Tg decreased as the moisture content of the samples increased as a result of water plasticization (Jouppila and others 1997). Gordon-Taylor fit predicted that lactitol with a moisture content of 5.8% has a Tg of 20°C and that lactitol with a moisture content of 4.4% has a Tg of 32°C (Figure 20.6). Therefore, a transition from the glassy to the

(a) 12% 90% 85% 81% 63% 58% 49% 39% 30% 21% 7%

11% 10%

Moisture content

9% 8%

90% fit 85% fit 81% fit 63% fit 58% fit 49% fit 39% fit 30% fit 11%

7% 6% 5% 4% 3% 2% 0

50

100

150

200 Time (h)

250

300

350

400

(b) 12% 11% 10%

Moisture content

9%

13%

13%

33% fit

33%

41% fit

41%

53%

64%

81%

8% 7% 6% 5% 4% 3% 2% 0

50

100

150

200 Time (h)

250

300

350

400

Figure 20.4. Moisture sorption of lactitol stored at (a) 7%, 11%, 21%, 30%, 39%, 49%, 58%, 63%, 81%, 85%, and 90% RH and 20°C; and at (b) 13%, 22%, 33%, 41%, 53%, 64%, and 81% RH and 32°C. Solid lines indicate modeled moisture content. RH, relative humidity.

279

280

PART 1: Invited Speakers and Oral Presentations

14 12 DH (J/g)

10

81%, 20°C

8

58%, 20°C

6 4 49%, 20°C 63%, 32°C

2 0 5%

81% 32°C

41%, 32°C

6%

7%

8% 9% 10% Moisture content

11%

12%

13%

Figure 20.5. Sample water peak melting enthalpy (ΔH) and moisture content of lactitol samples were grouped according to their storage conditions: 20°C at 49% RH (solid diamonds), 58% RH (plus signs), and 81% RH (open squares); and 32°C at 41% RH (asterisks), 63% RH (solid circles), and 81% RH (minus signs). The shaded and unshaded areas represent the two storage temperatures. RH, relative humidity.

Tg Gordon-Taylor fit

Temperature (°C)

60 40

32°C 20°C

20 0 0% –20 –40

5%

10%

15%

20%

Moisture content

Figure 20.6. Gordon-Taylor fit of the effect of moisture content on the glass transition temperatures (Tg) of lactitol. The broken lines at 20° and 32°C represent the storage temperatures of lactitol in this study.

rubbery state occurs when the moisture content increases above 5.8% and 4.4% for storage at 20° and 32°C, respectively. Lactitol that was stored at 20°C and 21% RH showed no evidence of crystal melting after 16 days of storage (results not shown). Crystal melting, however, was observed for samples stored at 20°C and 30% RH. Therefore, during storage at 20°C, there was a critical RH for crystallization to occur between 21% and 30% RH. At 21% RH, lactitol had an average final moisture content of 5.6%, which was below the critical moisture content of 5.8% (Figure 20.6). The measured Tg was 25° ± 1°C, which was 5° ± 1°C above the storage temperature (T − Tg = −5° ± 1°C). Lactitol was therefore predicted to be in the metastable glassy state. At 30% RH, lactitol crystallized and

Physical Changes in Confectionery Products

281

attained a maximum moisture content of 7.1%. For this moisture content, the GordonTaylor equation predicted a Tg of 11°C, which was 9°C below the storage temperature (T − Tg = 9°C) (Figure 20.6). This result indicated that the lactitol stored at 30% RH was in the rubbery state and thus underwent crystallization. During storage at 32°C for 11 days, lactitol stored at 13% RH did not crystallize, but samples stored at 33% RH crystallized, and samples stored at 22% RH at least partially crystallized (results not shown). The maximum moisture content of lactitol stored at 22% RH was 5.8%. The Gordon-Taylor equation predicted a Tg of 20°C (Figure 20.6), which was 12°C below the storage temperature of 32°C (T − Tg = 12°C). Hence, the amorphous lactitol was predicted to be in the rubbery state and potentially able to crystallize. The amorphous lactitol stored at 13% RH had an equilibrium moisture content of 3.7%. The Tg of lactitol at this moisture content was predicted to be 38°C (Figure 20.6), which was 8°C above the storage temperature of 32°C (T − Tg = −8°C). This result indicated that the amorphous lactitol at its equilibrium moisture content was in a metastable glassy state.

Conclusions Lactitol crystallized at a rate that increased with increasing RH and storage temperature. The crystallization occurred during storage at 20°C at RH down to 30%, whereas, at 32°C, crystallization occurred at RH down to 22%. Lactitol crystallization was enhanced by water plasticization that depressed the Tg to below the storage temperature. Therefore, moisture content and Tg are useful tools in predicting the stability of amorphous lactitol or sugar-free confectionery products that consist of lactitol.

References Avrami M. 1939. Kinetics of phase change I: general theory. J Chem Phys 7:1103–12. Hartel RW. 2001. Crystal growth. In: Crystallization in foods. Frederick, MD: Aspen. p 192–232. Jouppila K, Kansikas J, Roos YH. 1997. Glass transition, water plasticization, and lactose crystallization in skim milk powder. J Dairy Sci 80:3152–60. Jouppila K, Roos H. 1997. The physical state of amorphous corn starch and its impact on crystallization. Carbohydr Polym 32:95–104. Karmas R, Buera MP, Karel M. 1992. Effect of glass transition on rates of nonenzymatic browning in food systems. J Agric Food Chem 40:873–9. Koichi Y, Okahira, A, Hoshino M. 1997. Transformation of lactitol crystals and dehydration with grinding. Chem Pharm Bull 45:1677–82. Labrousse S, Roos YH, Karel M. 1992. Collapse and crystallization in amorphous matrices with encapsulated compounds. Sci Aliments 12:757–69. Lai HM, Schmidt SJ. 1990. Lactose crystallization in skim milk powder observed by hydrodynamic equilibria, scanning electron microscopy and H2 nuclear magnetic resonance. J Food Sci 55:994–9. Lampen BJ. 2000. Crystallisation of water-plasticised lactitol [master ’s thesis]. Dunedin, New Zealand: University of Otago, 93 p. p 34–5. Nakazawa F, Noguchi S, Takahashi J, Takada M. 1985. Retrogradation of gelatinized potato starch studied by differential scanning calorimetry. Agric Biol Chem 49:953–7. Ozimek L, Switka J, Wolfe F. 1992. Water sorption properties of ultrafiltration retentate skim milk powders. Milchwiss Milk Sci Int 47:751–4. Roos YH. 1995. Phase transitions in food. San Diego, CA: Academic.

282

PART 1: Invited Speakers and Oral Presentations

Roos YH, Karel M. 1992. Crystallization of amorphous lactose. J Food Sci 57:775–7. Sicard PJ, Le Bot Y. 1994. Manufacturing opportunities with non-sugar sweeteners. In: Rugg-Gunn AJ, editor. Sugarless: towards the year 2000. Cambridge, UK: Royal Society of Chemistry. p 112–35. Slade L, Levine H. 1991. Beyond water activity: recent advances based on an alternative approach to the assessment of food quality and safety. Crit Rev Food Sci Nutr 30:115–360. Van Velthuijsen JA. 1979. Food additives derived from lactose: lactitol and lactitol palmitate. J Agric Food Chem 27:680–6.

Oral Presentations

21 Entrapment of Probiotic Bacteria in Frozen Cryoprotectants and Viability in Freeze Drying Y. H. Roos and K. S. Pehkonen

Abstract The effects of freezing, freeze drying, and storage relative water-vapor pressure (water activity) on the viability of Lactobacillus rhamnosus GG (LGG) were investigated. Lactose and trehalose and their 1 : 1 mixtures with very similar glass transition behavior and thermal properties were used as cryoprotectants. LGG suspensions (109–1010 colony forming units/mL) were frozen (−22° or −43°C), freeze-dried, and stored under controlled water-vapor pressure (0%, 11%, 23%, and 33% relative vapor pressure) conditions. Glass transitions were determined by differential scanning calorimetry, and freezing properties of dehydrated materials were observed by microscopy. LGGs were observed to become entrapped in a maximally freeze-concentrated cryoprotectant phase in freezing and retained encapsulated in freeze-dried cryoprotectant matrices. Glassy frozen and dehydrated matrices seemed to retain viability better than matrices in the vicinity of, or above glass transition.

Introduction Freeze drying is often used to preserve bacteria for both food and pharmaceutical applications. Recovery rates of probiotic cultures after freeze drying in the absence of protective agents have been reported to be as low as 0.2% (Abadias and others 2001). Protective systems, such as monosaccharides, disaccharides, skimmed milk, bovine albumin, peptone, glycerol, and methanol, are often mixed with cultures before freezing to prevent damage from intracellular ice formation, to prevent membrane damage, and to facilitate dehydration and improve storage stability (Palmfeldt and others 2003; Patist and Zoerb 2005). Freezing temperature is a significant factor affecting cell viability in a freeze-drying process (Abadias and others 2001; Zhao and Zhang 2005). Although state diagrams and sorption isotherms have been determined for bacterial suspensions (Fonseca and others 2001), the fundamental aspect of the glass transition of the protective matrix as the main parameter controlling ice formation in frozen cell suspensions, as well as that of dehydrated systems subsequent to dehydration, have not been studied thoroughly. The present study considered the glass transition behavior of protective agents, freezing temperature, and water plasticization in storage as the critical factors affecting the viability of probiotic bacteria in frozen and freeze-dried cultures. The objective of 285

286

PART 1: Invited Speakers and Oral Presentations

the present study was to design cryoprotectant systems with similar glass transition behavior and observe differences in viability of Lactobacillus rhamnosus GG (LGG) in such matrices.

Materials and Methods LGG was incubated in MRS broth for 17 h at 37°C (optimal conditions to reach a cell density of 109–1010 colony forming units/mL). Following fermentation, the cells were harvested under aseptic conditions by centrifugation (25 min at 5000 rpm, ∼4000 g), and the resultant pellet was then suspended in the protective medium. The number of colonies was counted after 24 h of incubation in anaerobic conditions. Anaerocult (Merck, Darmstadt, Germany) was used to establish anaerobic conditions. All experiments were performed in triplicate. Solutions (20%, wt/wt) of α-lactose monohydrate (Sigma-Aldrich, Steinheim, Germany), trehalose dihydrate (Hayashibara, Okayama, Japan), and lactose-trehalose 1 : 1 were autoclaved (15 min at 121°C) and used to protect cells during freezing and freeze drying. Suspensions with LGG in glass vials (10 mL; 2.5-mL aliquots) (Thermo Life Sciences, Lund, Sweden) were frozen for 21 h at −22° or −43°C, followed by 5 h at −80°C, and freeze-dried (Lyovac GT 2 freeze dryer; Steris, Tuusula, Finland) for 48 h (P < 0.1 mbar; ice T less than −40°C). Vials were closed under vacuum by using a stoppering device at the end of the drying process prior to release of vacuum with ambient air. Freeze-dried materials in vials were stored at relative vapor pressure (RVP) of 0 (closed vials stored in an evacuated desiccator) and over saturated solutions of potassium acetate (23% RVP) and magnesium chloride (33% RVP). Vials with trehalose were also stored over 11% RVP (saturated lithium chloride solution). Colonies were counted after 24 h at 37°C under anaerobic conditions. All analyses were made in triplicate. SPSS for Windows was used for analysis of variance, and Tukey’s HSD (honestly significant difference) test was used to identify statistically significant differences in viability after freezing and freeze drying. Glass transitions of frozen, freeze-dried, and rehumidified protectant materials with and without LGG were determined by differential scanning calorimetry (model 821e DSC; Mettler-Toledo, Columbus, OH, USA). Freeze-dried materials (3.5–10.0 mg) were placed in preweighed pans (40 μL) (model 27331; Mettler-Toledo) and stored at room temperature in vacuum desiccators over 11%, 23%, and 33% RVP for at least 6 days. Onset temperatures of glass transition (Tg) for each material were determined at least twice, and the result was given as an average value. Also measured were the glass transition temperatures of maximally freeze-concentrated solutes (Tg′ ) and the onset temperature of ice-melting (Tm′ ) of 20% (wt/wt) lactose, trehalose, and lactose-trehalose solutions in the presence and absence of LGG as described by Roos and Karel (1991). Analyses were made in triplicate. Freezing of 20% (wt/wt) lactose-trehalose solution in the presence of LGG was observed microscopically by using an optical microscope (Olympus BX51; B & B Microscopes, Pittsburgh, PA, USA) equipped with a cold stage (Linkam LTS 350;

Entrapment of Probiotic Bacteria in Frozen Cryoprotectants

287

Linkham Scientific Instruments, Waterfield, UK). Micrographs were taken of the culture solution at −43° and at −22°C. A micrograph of the solution was taken as the ice crystals were almost melted at −2.9°C to observe the location of the cells in freezeconcentrated systems. The electron micrographs of dry lactose, trehalose, and lactose-trehalose cultures were obtained by using a scanning electron microscope (JSM-5600; JEOL, Tokyo, Japan).

Results and Discussion Glass transition temperatures (onset temperatures of glass transition) (Tg) determined for dry and rehumidified lactose, trehalose, and lactose-trehalose with and without LGG; as well as the glass transition temperatures of maximally freeze-concentrated solutes (Tg′); and onset temperatures of ice melting (Tm′ ) of 20% (wt/wt) lactose, trehalose, and lactose-trehalose solutions in presence and absence of LGG are listed in Table 21.1. The results confirmed that freezing of solutions at −43°C with bacteria enabled maximum ice formation to take place prior to freeze drying. Freezing at −22°C did not enable maximum freeze concentration, and the systems were only partially frozen prior to freeze drying. Water was an effective plasticizer, and the Tg decreased with increasing water content. As for freeze-concentrated systems, the Tg values determined for lactose, trehalose, and the lactose-trehalose mixture with LGG were found not to be affected by the presence of bacteria. The Tg′ and Tm′ determinations provided data for observations of frozen-state properties of probiotic suspensions. Ice melting was observed at around −30°C in accordance with the Tm′ . The maximum freeze concentration at −43°C enabled formation of solid, glassy nonfrozen solute systems with entrapped microbial cells. Such dispersion of the probiotic bacteria cells is illustrated in Figure 21.1. Bacterial cells were

Table 21.1. Glass transition of maximally freeze-concentrated solutes (Tg′ ) and onset temperature of ice melting (Tm′ ) of lactose, trehalose, and the lactose-trehalose solutions; as well as glass transition temperature (Tg) at various relative humidities of the freeze-dried systems in the presence and absence of Lactobacillus rhamnosus GG (LGG) RVP (%)

Glass transition (°C) Lactose

Lactose LGG

Trehalose

Trehalose LGG

Lactose trehalose

Lactose trehalose LGG

0

104

106

110

108

107

107

11

54

56

53

54

54

54

23

40

40

40

40

38

39

33 Tg′

29

30

30

30

30

30

−40

−40

−41

−41

−39

−41

Tm′

−29

−29

−30

−31

−30

−31

288

PART 1: Invited Speakers and Oral Presentations

Freeze-concentrated unfrozen solute phase

Entrapped cells

Ice

Freeze-dried, glassy solute membranes

Pores

Entrapped cells

Figure 21.1. A schematic representation of entrapment of bacteria in frozen and freeze-dried carbohydrate (cryoprotectant) matrices.

located in the continuous, nonfrozen phase. As freeze drying enabled removal of dispersed ice, the SEM micrographs also illustrated the location of the dispersed cells in the maximally freeze-concentrated nonfrozen solute phase. After freeze drying, the material retained its solid amorphous form and the structure looked like broken glass (Figure 21.2). The structure of freeze-dried trehalose was similar to that of lactose in its freeze-dried, glassy state. The smallest particles entrapped in solid glassy structure were identified as the dehydrated LGG cells, and their size was the typical cell size (∼2 μm). Freezing temperature before freeze drying had significant effects on cell viabilities after dehydration (Figure 21.2). The loss of viability was greater when the cells were frozen at −22°C than at −43°C. The LGG viability was between 30% and 70% when frozen at −43°C and <38% when frozen at −22°C. These results suggest that freezing to a maximally freeze-concentrated state and formation of a solid, glassy structure of an nonfrozen continuous phase with entrapment of LGG cells enhances retention of viability during freezing and freeze drying. It seems that the level of freeze concentration and formation of the maximally freeze-concentrated nonfrozen continuous solute phase controlled the cell viability during the freezing process. Retention of viability decreased with decreasing glass transition temperature. Highest viabilities were maintained when trehalose and lactose-trehalose were used to protect the entrapped cells at anhydrous conditions (0% RVP). When the materials were stored at 23% and 33% RVP, the glass transitions of the protective matrices were further depressed, as a result of water plasticization, to approximately 40° and 30°C, respectively, and considerable loss in cell viability was observed. When the materials were stored at 33% RVP, Tg decreased to ∼30°C (Table 21.1). Cells freeze-dried with lactose and trehalose and stored at 33% RVP lost their viability during storage, which was likely a result of water plasticization and associated changes in physicochemical properties of the continuous dehydrated solid phase and also the

Entrapment of Probiotic Bacteria in Frozen Cryoprotectants

289

log10 (cfu/ml)

10 9 RVP (%)

8 7

0 11 23 33

6 5 4

Frozen at -43°C Frozen at -22°C

3 2 1

LGG in trehalose

0 0

log10 (cfu/ml)

10 9 8

10

20 30 Storage time (days)

40

Storage time (day) RVP (%)

7 6 5 4 3

0 23 33

2 1 0

LGG in lactose + trehalose (1:1) 0

10

20 30 Storage time (days)

40

Figure 21.2. Scanning electron microscopy (SEM) images of Lactobacillus rhamnosus GG (LGG) encapsulated in trehalose (top) and lactose-trehalose (bottom) glass, with data for colony forming units (CFU) depending on prefreezing temperature (−43°C, open symbols; −22°C, solid symbols) and storage relative vapor pressure (RVP).

physical state of the dehydrated cells, which could be more sensitive to water than to the continuous extracellular protective matrix surrounding entrapped bacteria. When stored at 23% and 33% RVP, the cells frozen at the lower freezing temperature (−43°C) retained a higher viability. The results of this study indicate that disaccharides enhance the retention of viability of LGG during freezing, freeze drying, and storage. LGG viability seems to depend on freezing temperature, cryoprotectant, and storage conditions. It appears that materials encapsulating frozen and dehydrated bacteria should be stored below their Tg to avoid loss of viability.

References Abadias M, Benabarre A, Teixidò J, Viñas UI. 2001. Effect of freeze drying and protectants on viability of the biocontrol yeast Candida sake. Int J Food Microbiol 65:173–82. Fonseca F, Obert JP, Béal C, Marin M. 2001. State diagrams and sorption isotherms of bacterial suspensions and fermented medium. Thermochim Acta 366:167–82.

290

PART 1: Invited Speakers and Oral Presentations

Palmfeldt J, Rådström P, Hahn-Hägerdal B. 2003. Optimisation of initial cell concentration enhances freeze-drying tolerance of Pseudomonas chlororaphis. Cryobiology 47:21–9. Patist A, Zoerb H. 2005. Preservation mechanisms of trehalose in food and biosystems. Colloids Surf [B] 40:107–13 Roos YH, Karel M. 1991. Amorphous state and delayed ice formation on sucrose solutions. Int Food Sci Technol 26:553–66. Zhao G, Zhang G. 2005. Effect of protective agents, freezing temperature, rehydration media on viability of malolactic bacteria subjected to freeze-drying. J Appl Microbiol 99:333–8.

22 Fracture Behavior of Biopolymer Films Prepared from Aqueous Solutions I. Yakimets, S. S. Paes, N. Wellner, and J. R. Mitchell

Abstract To investigate the effect of water on the fracture mechanism of solid biopolymer systems, the mechanical response to the tensile loading of cassava starch, gelatin, and hydroxypropyl cellulose notched films was studied and compared. The biopolymer films were prepared by casting from aqueous solutions. To visualize the crack propagation, single-edge notched tension (SENT) specimens were submitted to tensile testing while being filmed with a high-speed video system. To evaluate the effect of water on the brittle-ductile transition of biopolymer films, the tensile tests were performed on both double-edge notched tension (DENT) and SENT specimens under different humidity conditions. The brittle-ductile transition was identified by plotting the proportion of stable crack propagation relative to total crack length as a function of the water content in the biopolymer films.

Introduction An important disadvantage of biopolymers is that they interact naturally with water, which leads to water-induced structural transformations (e.g., amorphous-crystalline transition) that have a strong impact on their molecular mobility and functional properties (Marzec and others 2007; Yakimets and others 2007a, 2007b). The aim of this work was to compare the water-dependent fracture behavior of different biopolymer films and to visualize the crack propagation by using in situ visualization with a highspeed camera. This study also included the comparison, by polarized Fourier transform infrared spectroscopy, of the water ’s effect on the size and shape of the residual plastic deformation zone formed during the stretching of biopolymer films.

Materials and Methods Preparation of Specimens and Conditioning The gelatin used in this study was type B from bovine skin (G9382 gelatin; SigmaAldrich, Poole, UK), bloom 225 and molecular weight of 142,000 g/mol. Hydroxypropyl cellulose (Klucel; Hercules, Wilmington, DE, USA), was LF type with molar substitution of 3.8 and molecular weight of 95000 g/mol. Native cassava starch (National Starch, Bridgewater, NJ, USA) contained 18.0% ± 0.9% amylose. Aqueous solutions (4 : 100 wt/wt) were prepared, by heating in a water bath, gelatin (70°C, 30 min), HPC (22°C, 5 h), and cassava (90°C, 1 h) powders. The solutions were cast in 291

292

PART 1: Invited Speakers and Oral Presentations

polystyrene Petri dishes and dried under controlled conditions (20°C and 45%–55% relative humidity [RH]) to obtain thick films (∼100 μm). Single-edge notch tension specimens (10 mm wide and 100 mm long) were prepared by introducing a 3-mm-long notch. The specimens were predried over phosphorus pentoxide and subsequently equilibrated at specified RHs for at least 3 days (in chambers containing saturated salt solution). Mechanical Tests The mechanical tests were conducted on a texture analyzer (TA.XT Plus; Stable Micro Systems, Godalming, UK) with a 30-kg load cell by using a controlled RH chamber (by mixed dry and wet airflow) to maintain the RH of the samples during stretching. Tensile tests were performed at high speed equal to 30 mm/min by using a grip separation of 50 mm. This speed was chosen in order to match the acquisition rate of the mechanical data and images recorded by the high-speed camera. Visualization Techniques Two monitoring methods were used in this study: (a) in situ visualization by highspeed camera during the stretching of the films and (b) post-deformation plastic zone by using polarized Fourier transform infrared spectroscopy (FTIR). For in situ visualization, a monochrome high-speed video system (Photron Ultima APX; Photron, Marlow, UK), which can record at up to 120,000 fps, was used. For this work, the system was operated at a frame rate of 6000 fps and maximum resolution of 512 × 512 pixels. The samples were marked with black points to better visualize the development of the plastic zone deformation during stretching. The postdeformation plastic zones close to the fractured edges of thin film fragments (∼10 μm) were imaged under the FTIR microscope. The RH inside the FTIR chamber was maintained by a flow of mixture of dry and wet air. Polarized infrared maps were recorded with a 128 × 128-element focal point array detector (Stingray; BioRad, Hercules, CA, USA). For each map, 128 scans were averaged at 8 cm−1 resolution. The backgrounds of the empty windows were measured with the same parameters. The spatial resolution is about 7–10 μm (corresponding to the diffraction limit of the infrared light).

Results and Discussion Effect of Water Content on Physicochemical and Structural Properties of Biopolymer Films The effect of water content on the physicochemical and structural properties of gelatin (G), cassava starch (CS), and hydroxypropyl cellulose (HPC) films was discussed extensively in our previous work (Yakimets and others 2007a), which demonstrated that the increase in water content within these three biopolymer systems leads to the well-known plasticizing effect, which decreases the glass transition temperature associated with the decrease in elastic properties (storage modulus measured by dynamic mechanical thermal analysis). Water played an important role in

Fracture Behavior of Biopolymer Films Prepared from Aqueous Solutions

293

stabilization of the triple-helical structural order of G films. In contrast, the elastic properties of the G films were not strongly affected by the increase in water content (below 25%, dry basis) (Yakimets and others 2005). In contrast, the HPC films demonstrated a low capability of changing their conformation during hydration, associated with a large decrease in elastic properties due to water uptake. The CS films revealed the most distinctive structural organization, which is strongly related to their method of preparation (Paes and others 2007). Also demonstrated was that the shear rate and temperature of the gelatinization process has a significant effect on the structural and mechanical properties of CS films. Films prepared at milder conditions (low shear, T < 90°C) exhibited better mechanical properties compared with the films prepared under severe conditions (high temperature and high shear rate) because of the more homogeneous amylose-amylopectin distribution and presence of the granule remnants within the structure. Because of that, in the present work CS films were prepared at low gelatinization temperature (90°C) and low shear (150 rpm). In situ Visualization of Crack Propagation To visualize the fracture mechanism and evaluate the effect of water content on the fracture properties of the biopolymer films studied, a high-speed camera was used to record the images of the stretching of notched biopolymer films conditioned at different RHs. At a water content of up to 25% (dry basis; ambient temperature, 25°C), the G films in the glassy state exhibited only unstable crack propagation. Unstable crack propagation in G films did not usually involve crack branching or a multicrack mechanism (data not shown). In contrast, in CS films, the unstable crack propagation exhibited at a low water content occurred mostly under multicrack conditions. It was postulated that, when the crack propagation velocity exceeds a critical value, brittle systems develop a new mode of energy dissipation, which results in the formation of additional branches (Sharon and others 1995). It was shown that this critical velocity for the onset of branching is equal to Vc = 0.36 VR, where VR is the Rayleigh wave speed in the material. Consequently, the multicrack fracture mechanism of CS films shows that (a) this system is the most unstable of the three systems studied, and (b) its brittle fracture is associated with the highest velocity of crack propagation. We also previously showed that CS films—because of the presence of remaining starch granules, which also contribute to their relatively high instability in the dry state—presented the most heterogeneous structure among the three biopolymer systems studied. The HPC films did not exhibit pure unstable crack propagation. The stable propagation of the crack tip always preceded the unstable propagation within the dry HPC system. This indicates that HPC films are the most thermodynamically stable system in the dry state. Only HPC and CS films exhibited stable crack propagation at high water content (≥44% RH). In our previous work, we showed that the method of preparation of starch films affects their functional properties and should be considered in order to optimize the final structure of these films (Paes and others 2007). In this work, we showed that, as with HPC films, the optimized CS films, which do not contain any additional

294

PART 1: Invited Speakers and Oral Presentations

(a)

0 sec

7.515 sec

12.024 sec

13.632 sec

(b)

0 sec

3.322 sec

3.775 sec

3.93 sec

Figure 22.1. In situ visualization of the stable crack propagation for (a) hydroxypropyl cellulose film and (b) cassava starch film at 69% relative humidity by using the high-speed camera.

plasticizers, can develop plastic deformation and stable crack propagation at high water content. The mechanism of stable crack propagation in CS films was compared to that of HPC films (Figure 22.1). Figure 22.1 shows that HPC films develop a large, homogeneous plastic deformation (stress-whitening zones and spot deformations). Two different mechanisms of fracture are involved during the stretching of these two biopolymer films: the crack propagated by void formation in CS films and by formation of the extensive plastic deformation zone in HPC films. It is well known that, when an external force is applied, two main mechanisms are activated to absorb the energy by a polymer system: shear yielding and crazing. The crazing results in limited energy absorption compared with the shear yield, which is a characteristic deformation mode for ductile polymers. The stress whitening of HPC films during tensile deformation clearly shows that this biopolymer deforms by a shear-yield mechanism. In contrast, the crack propagation by void formation in CS films can be compared with the crazing mechanism observed in synthetic polymers such as polymethyl methacrylate (PMMA), polystyrene (PS), and polyvinyl chloride (PVC). Despite the similarity of these two fracture mechanisms, the crazing mode will absorb more energy because of the high molecular orientation of fibrils formed under crazing conditions, which can support higher stress than the bridges formed between the voids in starch films.

Fracture Behavior of Biopolymer Films Prepared from Aqueous Solutions

Fractured edge

(a)

Fractured edge

(b)

295

Fractured edge

Potential zone of voids formation (dark grey zones indicated by arrows)

(c)

Figure 22.2. Infrared maps for (a) gelatin film, (b) hydroxypropyl cellulose film, and (c) cassava starch film in the hydrated state (60% relative humidity).

The Post-Deformation Plastic Zone of Biopolymer Films To further characterize the deformation zone around the crack tip of stretched thin films (Marzec and others 2007), FTIR maps of the fracture edges were acquired with a Digilab UMA 600 microscope attached to a FTS6000 FTIR spectrometer (Varian, Palo Alto, CA, USA) by using polarized infrared light (Figure 22.2). To construct these maps, the intensity of the amide III band at 1242 cm−1 (0–50%, black; 50%–80%, red; 80%–90%, green; >90%, blue), intensity of the 1377-cm−1 band (0–57%, black; 57%–85%, red; 85%–93%, green; >93%, blue), and the intensity of the 1356-cm−1 band (0–50%, black; 50%–86%, red; 86%–93%, green; 93%–100%, blue; >100%, yellow) were used for G, HPC, and CS films, respectively. These maps show the changes within the thickness of deformed films, which reflect the local distribution of the deformation. An additional level was necessary because, due to their microstructure (i.e., remnants of starch granules), the thin CS films were not entirely flat. The thickness of the non-deformed film was taken as 100%. It was observed that, in the dry state, the three biopolymer films (G, HPC, and CS) did not exhibit any plastic deformation near the fracture edge (data not shown). However, significant differences were observed for the hydrated biopolymer films (Figure 22.2). As previously discussed, the G films did not develop plastic deformation during the crack propagation under 60% RH conditions. In contrast, the HPC films developed a large plastic zone under 60% RH, with even distribution of its thickness, which indicated the classic shear-yield plastic deformation. The CS films showed a highly irregular distribution of thickness within the hydrated films and particularly near the crack edge. In addition, FTIR microscopy demonstrated isolated thin areas of spots in Figure 22.2c that can develop into voids when additional force is applied. This mechanism of fracture by void formation is clearly the heterogeneity of the structure, which maintains the starch granule remnants. Prediction of Dependence of Fracture Characteristics on Water Content for Biopolymer Films: Brittle-Ductile Transition The parameter chosen to model and predict the fracture mechanism was the ratio between stable and unstable crack propagation, which was associated with brittleductile transition. The G films always showed unstable crack propagation even at 75%

296

PART 1: Invited Speakers and Oral Presentations

RH and did not exhibit brittle-ductile transition within the range of RH conditions studied. The fracture characteristics of CS films were random because of the relatively heterogeneous structural order given the highly scattered mechanical data. The data obtained enabled us to conclude that the capability of the CS films to develop the stable crack propagation generally increased with increased water content. Also, the high scattering of the mechanical data masked the brittle-ductile transition of this biopolymer. On the other hand, a clear quasi–brittle-ductile transition was observed for HPC. In our previous work, we showed that this brittle-ductile transition can be associated with the appearance of the multilayer water within the structure of HPC films, which adds mobility to the biopolymer system (Yakimets and others 2007a).

Acknowledgments This work was supported by the Biotechnology and Biological Sciences Research Council (BBSRC), with grants BBSB 12962/12733 and CAPES (Ministry of Education of Brazil). We are also grateful to the European Polysaccharide Network of Excellence (EPNOE) for their scientific contribution.

References Marzec A, Lewicki PP, Ranachowski Z. 2007. Influence of water activity on acoustic emission of flat extruded bread. J Food Eng 79:410–22. Paes SS, Yakimets I, Mitchell JR. 2007. Influence of gelatinization process on functional properties of cassava starch films. Food Hydrocolloids 22:788–97. Sharon E, Gross SP, Fineberg J. 1995. Local crack branching as a mechanism for instability in dynamic fracture. Phys Rev Lett 74:5096–9. Yakimets I, Paes SS, Wellner N, Smith AC, Wilson RH, Mitchell JR. 2007a. Effect of water content on the structural reorganization and elastic properties of biopolymer films: a comparative study. Biomacromolecules 8:1710–22. Yakimets I, Wellner N, Smith AC, Wilson RH, Farhat I, Mitchell J. 2005. Mechanical properties with respect to water content of gelatin films in glassy state. Polymer 46:12577–85. Yakimets I, Wellner N, Smith AC, Wilson RH, Farhat I, Mitchell J. 2007b. Effect of water content on the fracture behaviour of hydroxypropyl cellulose films studied by the essential work of fracture method. Mechanics Mater 39:500–12.

Session 4 Biomaterial Sciences: Water in Stability and Delivery of Active Biomolecules

Invited Speakers

23 The Plasticization-Antiplasticization Threshold of Water in Microcrystalline Cellulose: A Perspective Based on Bulk Free Volume S. P. Chamarthy, F. X. Diringer, and R. Pinal

Abstract The influence of water on the compaction properties of microcrystalline cellulose (MCC) was investigated. The solid fraction and tensile strength of MCC compacts were studied as a function of compression pressure and relative humidity. The interplay between water content and porosity determines whether the solvent plasticizes (softens) or antiplasticizes (hardens) the polymer. At low concentrations, water increases the mechanical strength of the compacts. At higher concentrations, water has the effect of a typical plasticizer, reducing the tensile strength. The antiplasticizing effect of water was more pronounced for tablets with higher porosity. The crystallinity of MCC also changes as a function of water content. It is shown however, that moisture-induced changes in crystallinity are not the cause of antiplasticization. The results strongly suggest that when the plasticizer concentration is low enough to occupy the accessible void volume in the polymer matrix, antiplasticization occurs. Conversely, when the volume occupied by the solvent exceeds the accessible void volume, the plasticizer disrupts polymer-polymer interactions, resulting in plasticization.

Introduction The importance of polymers in various disciplines cannot be overemphasized. Polymers have unrivaled versatility in their uses. The main reason is arguably that, when using polymers, scientists can control the mechanical properties of such materials almost at will, through the use of plasticizers. In the pharmaceutical and food sciences, carbohydrate polymers play a central role and are a major component in many products from each respective industries. Microcrystalline cellulose (MCC) occupies a unique place in pharmaceutical applications and is the focus of this report. An estimated 84% of all pharmaceutical products are designed to be ingested (Abrahamsson and Lennernäs 2003). Moreover, most of these products are made in the form of tablets; compacts made by compressing powdered blends of the ingredients. The mechanical strength of tablets makes them the most convenient type of dosage form, and tablets are, whenever possible, the first choice for delivering an active pharmaceutical ingredient to patients. Many active pharmaceutical ingredients, however, do not have compaction properties that make them amenable by themselves 301

302

PART 1: Invited Speakers and Oral Presentations

to producing compacts of desirable strength and mechanical properties. For this reason, pharmaceutical formulations call for the use of excipients (pharmacologically inactive ingredients) in order to attain the desired properties in the final dosage form. In addition to being innocuous, MCC has compaction properties that make it ideally suited for use as a compaction aid in pharmaceuticals. Water, on the other hand, is the plasticizer of choice whenever possible because it provides a safe, inexpensive means of controlling the mechanical properties of materials used in the preparation of compacts. The addition of a plasticizer has the effect of softening polymers; that is, increasing their ductility while decreasing their mechanical strength, thus contributing to the formation of compacts. Another important aspect of water, however, is that since it is ubiquitous in the environment, it is an unavoidable factor in the preparation of compacts. Environmental moisture represents an unintentional and/or uncontrolled means of incorporating water into pharmaceutical tablets. Environmental moisture contributes rather low levels of water into the raw materials used for making tablets. This is important because, more often than not, low levels of plasticizer actually produce the exact opposite of the expected softening effect on the polymer. Such a phenomenon, termed antiplasticization, is the subject of this study. Below a critical plasticizer concentration, the plasticizer actually decreases the ductility, while increasing the mechanical strength of polymers (Sears and Darby 1982). This is true for just about any type of polymer, and pharmaceutical systems are no exception (Chamarthy and Pinal 2007). Here we present the results of a detailed study of the effect of water on the compaction properties of MCC. The results from this study, although directly associated with numerous existing and developing products, should also be useful in terms of the possible effect of water on other carbohydrate polymers.

Materials and Methods Microcrystalline Cellulose MCC (Avicel PH-200) was a kind gift from FMC (Newark, DE, USA). Preparation of Compacts MCC powder was equilibrated under different equilibrium relative humidities (ERHs), including 0%, 6%, 11%, 23%, 33%, 43%, 59%, 75%, 85%, and 100%; over saturated salt solutions in desiccators for a 2-week period. The controlled-humidity chambers were preequilibrated at these different moisture levels for 12 weeks by using silicone gel to seal the chambers. The salts used to generate the different ERH levels are listed in Table 23.1, based on the ASTM method (ASTM International, 2002). After moisture equilibration for a period of 2 weeks, MCC powder samples were compacted in a single-station automated press (Carver Laboratory Press, Hydraulic Unit Model; Carver, Wabash, IN, USA) under different compaction pressures (23, 50, 67, 134, 201, 268, 335, and 402 MPa) and a 30-s dwell time. The compacts were made using a 13-mm, flat-faced, round punch-and-die set (Natoli Engineering,

Plasticization-Antiplasticization Threshold of Water in MCC

303

Table 23.1. Saturated salt solutions used to maintain fixed relative humidity levels in desiccators % Relative humidity

Salt

0

P2O5

6

LiBr

11

LiCl

23

CH3COOK

33

MgCl2

43

K2CO3

59

NaBr

75

NaCl

85

KCl

100

KNO3

Charles, MO, USA). Care was taken to minimize the time each sample spent outside the ERH-controlled chamber during compaction. The freshly formed tablets were immediately returned to the corresponding controlled ERH desiccators for 48 h prior to any testing. This step allowed reequilibration of the compacts under the same ERH conditions used for the powder before compaction. The weight of the tablets and their cylindrical dimensions were used to determine the apparent density of the compacts. Mechanical Properties After storage for 48 h under controlled humidity, the mechanical properties of the tablets were measured. Force-vs-displacement profiles were obtained in triplicate by using the three-point beam-bending method on a TA.XT2i Texture Analyzer (Stable Micro Systems, Godalming, UK). The data were collected, by using Texture Expert Exceed (version 2.5; Stable Micro Systems) software supplied with the instrument, at a rate of 200 points per second and a test speed of 0.2 mm/s. The instrument was calibrated prior to data collection by using a 5-kg weight. The tensile strength (σ) and Young’s modulus of elasticity (E), were obtained according to Equations 23.1 and 23.2: σ=

3 Fl 2 w th2

(23.1)

E=

F l3 4 d w th3

(23.2)

where F is the applied force at the point of fracture, l is the distance between the lower supports, w is the width (or diameter) of the compact, d is the displacement of the center beam when the compact breaks, and th is the thickness of the compact. The mechanical properties were tested in a room where temperature and humidity levels were maintained at 25°C and 22% ERH, respectively.

304

PART 1: Invited Speakers and Oral Presentations

True Density The true density of the MCC powder was measured by using a gas displacement pycnometer, AccuPyc 1330 (Micromeritics Instruments, Norcross, GA, USA). Helium was used as the gas for the measurements. The accuracy of the instrument was checked by using the AccuPyc 1330 calibration standard (Micromeritics Instruments) with a known volume of 6.3723 mL. Before measurements, the powder samples were stored at 0% ERH (over phosphorus pentoxide) in a desiccator for a 2-week period in order to remove any residual moisture. During measurement, powder samples were purged with dry helium and returned to vacuum conditions ten times in the instrument test chamber. This was done to effectively remove any residual surface moisture prior to final data collection. The results reported are the average of 15 consecutive measurements. Samples were weighed immediately after the last measurement. The weight change was used to correct the true density reported by the instrument. Solid Fraction and Porosity A tablet can be considered as a special type of dispersion. The solids and the air trapped in pores constitute two phases of the system. The solid fraction of a tablet is defined as the portion of the total tablet volume occupied by solid material. The solid fraction in a compact corresponds to the ratio of the apparent density of the tablet (ρapp) and the true density (ρtrue) of the solid. Conversely, porosity (ε), given by Equation 23.3, represents the fraction of the total volume of the compact made up by void space (air), including interparticulate and intraparticulate voids (Sun 2004). The compaction pressure directly influences these parameters, and the porosity usually decreases as the pressure increases. ε = 1−

ρapp ρtrue

(23.3)

For a compact with simple geometry (a flat-faced cylindrical tablet), density can be calculated from its weight (wtab) and volume (Vtab). The true density is measured by using established methods like helium pycnometry (Sun 2004). When a plasticizing diluent is added to the system, the volume and weight contributed by the plasticizer have to be taken into account. The corrected porosity of the tablet is given by ε = 1−

( wtab − w plas ) ( Vtab − Vplas ) ρtrue

(23.4)

where wplas and Vplas are the weight and the volume of the plasticizer, respectively. In this report, water is the plasticizer and the definition of porosity used represents the combined fraction of the volume of the compact occupied by air and by water.

Results and Discussion The hygroscopicity of the same lot of MCC (Avicel PH-200) used in this study was measured and reported previously (Chamarthy and Pinal 2007) and is shown in Figure 23.1. MCC picks up about 12% (wt/wt) of water in the ERH range of 0–90%.

Plasticization-Antiplasticization Threshold of Water in MCC

305

Water Uptake / % (wt/wt)

12 10 8 6 4 Sorption 2

Desorption

0 0

20

40

60

80

100

%RH

Figure 23.1. Moisture-sorption isotherm of microcrystalline cellulose (Avicel PH-200) measured at 25°C. From Chamarthy and Pinal (2007). RH, relative humidity.

When a powder bed is compressed, its total volume is reduced through a decrease in the space occupied by voids between the particles; that is, through a decrease in porosity. The effect of varying levels of water on the compression behavior of MCC is shown in Figure 23.2. At all ERH values investigated, there is an initial rapid decrease in porosity upon compression. As the compression pressure is further increased, the reduction in porosity exhibits a plateau. This behavior reflects the packing ability of MCC powder. When introduced into the die, the powder bed contains high levels of voids. At low pressures, particles rearrange easily to fill these voids and readily reduce the total volume. However, when the porosity is substantially decreased, slippage and rearrangement of particles are no longer sufficient volume-yielding mechanisms. A further increase in pressure causes the larger particles to fracture into smaller ones in order to fill smaller voids, a process that involves substantially more work. Consequently, the compressibility profiles level off. Figure 23.2 shows that, for a given compression pressure, increasing the water content shifts the curve down. This means that the presence of water enhances the volumeyielding properties of the material. As stated, porosity levels in Figure 23.2 correspond to the fractional volume occupied by air and water combined. Therefore, under the same compression pressure, increasing water levels compact the higher solid fraction. In addition to particle fracture, low porosities (or high compression pressures) lead to the formation of cohesive bonds among particles, mostly through Lifschitz–van der Waals interactions (Rietema and others 1993), thus resulting in the formation of the compact. For any given compressed material, the lower the porosity, the greater the number of cohesive interactions among particles and, consequently, the stronger

306

PART 1: Invited Speakers and Oral Presentations

0.5 0% RH 11% RH 33% RH 59% RH 85% RH

0.4

Porosity

0.3

0.2

0.1

0.0 0

100

200

300

400

Compaction Pressure (MPa)

Figure 23.2. Plots of porosity vs compaction pressure showing the compressibility of microcrystalline cellulose under different conditions of relative humidity (RH).

the resulting compact. In a plot of porosity vs compression pressure like Figure 23.2, a shift down is expected to result in a compacted material with greater mechanical strength. By this notion alone, we could expect that, for the samples shown in Figure 23.2, at any given compression pressure, higher water content should result in a stronger compact. Water however, is a widely used plasticizer that, as such, has the effect of reducing the mechanical strength of polymers. Then, by this second notion alone, we could expect that, at any given compression pressure, higher water content should result in compacts with lower tensile strength and lower Young’s modulus. The question is, therefore, which one of these two contradictory notions is really at play in this type of situation. To explore this question, it is more convenient to use porosity (instead of compression pressure) as the control variable when comparing different samples. As per the definition of porosity used here, regardless of their water content, in any two samples of equal porosity, the fractional volume occupied by MCC is the same. This way, any differences in mechanical properties between the two samples cannot be attributed to geometric factors such as the polymer contribution to the crosssectional area of the compact. Hence, the differences can be assessed in the context of the interactions between the polymer and water. Figure 23.3 shows the effect of increasing water content on the tensile strength of MCC compacts at constant porosity. At the higher porosity (ε = 0.3), addition of water clearly shows antiplasticization, where the tensile strength of the compact first increases before higher concentrations of the plasticizer produce the typically expected decrease. At lower porosity (ε = 0.1), increasing water concentration first shows a plateau in tensile strength before higher water concentrations begin to reduce the

Plasticization-Antiplasticization Threshold of Water in MCC

(a)

(b)

Tensile Strength / MPa

Tensile Strength / MPa

6 5 4 3

16 12 8 4

Porosity = 0.3 2

307

0

20

Porosity = 0.1 40

60 %RH

80

100

0

20

40 60 %RH

80

100

Figure 23.3. Effect of relative humidity (RH) on the tensile strength of compacts at constant porosity of (a) 0.3 and (b) 0.1.

strength of the compact. It is noteworthy that, in either case, water does not act as a plasticizer until the ERH exceeds ~40% (when the solid line falls below the horizontal dotted line), which corresponds to about 4%–5% water content. Notice that lower porosity resulted in stronger compacts throughout, as would be expected from the general notions of compaction. It is the effect of water that does not follow the traditional notions of plasticizing effects. Figure 23.3 suggests that, as long as sorbed water can fill the available void volume without saturating it, the water will strengthen the compact. To explore this possibility, the tensile strength at zero porosity was obtained based on the Ryshkewitch equation. This is one of the most commonly used relationships for analyzing compaction data (Ryshkewitch 1953; Sun 2005). The approach is based on an exponential relationship between the porosity (ε) of a material and its tensile strength (σ). The equation provides the mechanical strength of the material at the hypothetical zero porosity: σ = σ 0 e − bε

(23.5)

where σ0 is the tensile strength at zero porosity and b is a material specific constant. Figure 23.4 shows a plot of σ0 as function of relative humidity obtained by fitting the compaction data to Equation 23.5. There are two noteworthy features in the figure. First, the effect of water is exclusively that of a plasticizer: the water solely decreases tensile strength, and the effect is more pronounced with increasing water concentrations. Second, the tensile strength is highest for the lowest ERH (or water content). These results strongly suggest that the presence of available volume for the plasticizer to fill is critical for antiplasticization to take place. The preceding discussion is presented in terms of what can be considered the most fundamental pair of variables: porosity and tensile strength. In practice however, formulation and processing decisions are made on the basis of more readily accessible

308

PART 1: Invited Speakers and Oral Presentations

Tensile Strength / MPa

30

20

10

Porosity = 0 0 0

20

40

60

80

100

%RH

Figure 23.4. Tensile strength as a function of relative humidity (RH) at a porosity of zero obtained from Ryshkewitch analysis.

variable sets, such as compression pressure and tensile strength. Although subtle, the distinction is important: cause-and-effect considerations are critical for elucidating mechanisms. Increasing compression pressure results in stronger compacts. However, greater compression pressure, per se, does not strengthen the compacts. It does so only indirectly. Greater pressure leads to lower porosity, which in turn results in greater mechanical strength. Compaction is a balance between (1) the material’s ability to yield volume upon compression and (b) its tendency to form cohesive bonds between particles when the volume is reduced. Figures 23.2–23.4 show that water alters the behavior of MCC on both accounts. Water (a) augments the volume-yielding capacity of MCC and (b) increases at first (i.e., at low concentrations) the bond-forming tendency of MCC, but has the opposite effect at higher water content levels. The effect of water on point (b) above seems to depend on the ability of water to replace air in the compressed powder without forcing any other geometric change. In other words, antiplasticization is observed as long as water can occupy the accessible volume in the matrix. The foregoing considerations notwithstanding, the effect of compression pressure on the mechanical strength of compacts is important because the relationship involves those variables that are accessible during the processing of compaction-based products. Figure 23.5a shows the effect of compression pressure on tensile strength in the absence of water. The profile is typical in the sense that it shows an increase in compact strength with compression pressure, followed by a plateau. For purposes of the present discussion, we selected three compression pressures: 23, 134, and 268 MPa. These pressures correspond to the low, middle, and high (plateau) portions of the

Plasticization-Antiplasticization Threshold of Water in MCC

(b)

Tensile Strength / MPa

15

10 0% RH 5

0

0

200 300 400 100 Compaction Pressure / MPa

Normalized Tensile Strength / %

(a)

309

2.0 1.6 1.2 0.8 23 MPa 134 MPa 268 MPa

0.4 0

20

40 60 %RH

80

100

Figure 23.5. Relationship between tensile strength and compaction pressure of microcrystalline cellulose (MCC). (a) Dry conditions used as reference. The vertical dotted lines mark the selected compression pressures. (b) Effect of relative humidity (RH) on the tensile strength of MCC compacts prepared at the selected compression pressures of 13, 134, and 268 MPa.

profile, as indicated by the dotted vertical lines in Figure 23.5a. The effect of water (ERH) on the (normalized) tensile strength at those compression pressures is shown in Figure 23.5b. In all cases, the presence of water initially increases the strength of the compacts. The strengthening effect of water reaches a maximum; thereafter, additional water reduces the tensile strength. It should be noted that, even though there is a maximum in each curve, we can identify two critical points in the profiles. One corresponds to the maximum in tensile strength. The other corresponds to the point where the tensile strength is the same as that of the compact devoid of any water; that is, the point where the tensile-strength profile and the horizontal dotted line intersect in Figure 23.5b. When the porosity is high (low compression pressure), the latter critical point is never attained; water-containing compacts are always stronger than those produced in the absence of water. As long as the profiles are above the horizontal dotted line, water is acting as an antiplasticizer since its effect is to increase the tensile strength in relation to a compact free of any water. The range where water acts as an antiplasticizer decreases with reduced porosity (with increasing compression pressure). In the case of MCC, high ERH levels (about 60% and 80% for 134 and 268 MPa, respectively) are needed for water to begin to act as a plasticizer. The results obtained support the notion that the accessible volume is a key factor in the antiplasticization phenomenon. More specifically, accessible volume seems to play the role of a pseudo–free volume, or a bulk free volume (BFV). A clarification is in order here, since free volume has an unambiguous definition. Free volume is the difference in specific volume of the material at the temperature of interest and the specific volume at absolute zero. In other words, free volume is the volume of the intramolecular interstices at the temperature of interest. In the case of

310

PART 1: Invited Speakers and Oral Presentations

antiplasticization, to the extent that the true free volume can be occupied by the plasticizer molecules, it is part of the accessible volume (or BFV). Strictly speaking, true free volume is a subset of BFV and may or may not be accessible, depending on the particular polymer-plasticizer system. Our data suggest that antiplasticization is the result of BFV and the ability of the plasticizer to occupy it. There is, however, an alternative explanation that needs to be addressed. MCC is a semicrystalline compound; it consists of ordered crystalline regions (crystallites) and less organized amorphous regions. It has been proposed in the literature that for semicrystalline systems, crystallinity changes induced by the plasticizer cause antiplasticization (Lourdin and others 1997). By this account, small amounts of plasticizer increase the two-dimensional ordered regions in a semicrystalline polymer, and such an increase in crystallinity has been proposed to be a direct cause of antiplasticization (Sears and Darby 1982). We determined the crystallinity indices of MCC as a function of ERH by two methods: X-ray powder diffraction (XRD) and Fourier transform infrared spectroscopy (FTIR). Crystallinity indices from XRD were determined according to the method reported by Nelson and O’Connor (1964) for cellulose. Briefly, this method compares the peak intensities at the 2θ angles of 22.6° (crystalline material) and 19.0° (amorphous background). Crystallinity indices from FTIR were obtained according to the method reported by Ek and others (1995) for MCC. Briefly, this method compares the absorbance at 1372 cm−1 and 2900 cm−1. The crystallinity indices obtained by the two methods are shown as a function of ERH in Figure 23.6. The profiles support the notion that the initial increase in tablet strength at low amounts of water is linked to an increase in crystallinity. By such an

100

XRD FTIR

95

95

90

90

85

85

80

80

75 0

20

40

60

80

Crystallinity Index (FTIR)

Crystallinity Index (XRD)

100

75 100

%RH

Figure 23.6. Apparent crystallinity index of microcrystalline cellulose obtained from two different methods—X-ray diffraction (XRD) and Fourier transform infrared spectroscopy (FTIR)—as function of equilibrium relative humidity (RH).

Plasticization-Antiplasticization Threshold of Water in MCC

311

account, the greater molecular mobility afforded by the plasticizer facilitates the reorganization of the material, inducing a growth of the ordered (crystalline) regions. However, this notion does not hold at higher water content, where the degree of crystallinity of the polymer plateaus but the tensile strength of its compacts begins to decrease as a function of water concentration. An even more dramatic challenge to the notion of changes in crystallinity as the cause of antiplasticization is posed by other systems, such as soluble starch. Soluble starch is a semicrystalline polymer whose crystallinity continuously increases with increasing water content. The antiplasticizing effect of water in this case results in a profile where the tensile strength of the compacts shows a sharp drop while the increase in crystallinity continues with increasing water content (Chamarthy 2007). The case against crystallinity as the cause of antiplasticization becomes overwhelming if we consider that increases in tensile strength by addition of small amounts of plasticizer are also observed in noncrystalline polymers. In other words, antiplasticization takes place even in the total absence of crystallinity in the polymer (Chamarthy 2007). The foregoing considerations suggest that, rather than acting as the cause, changes in crystallinity in semicrystalline polymers are actually a manifestation of antiplasticization. It follows that an alternative explanation for antiplasticization is necessary, which brings our discussion back to the BFV. The proposed role of BFV is depicted in Figure 23.7.

Figure 23.7. Depiction of a feasible explanation for the association between antiplasticization and plasticization, based on bulk free volume (BFV). As long as the plasticizer can occupy the BFV without changing the total volume, antiplasticization can be expected. When the volume occupied by the plasticizer exceeds that available as BFV, plasticization occurs. RH, relative humidity; and TS, tensile strength.

312

PART 1: Invited Speakers and Oral Presentations

Consider the case of a compressed polymer matrix of high porosity (Figure 23.7, top row). At low levels of plasticizer, the (comparatively smaller) plasticizer molecules can easily penetrate the void spaces in the polymer matrix. New bonds are created between polymer chains and plasticizer molecules, in addition to the already present polymer-polymer interactions. At a bulk level, this increase in cohesive interactions increases the rigidity of the matrix, thus antiplasticizing the polymer. During this stage, the ordered regions of the polymer chains can also grow in semicrystalline systems. A further increase in plasticizer content will result in more sites of plasticizerpolymer interactions until no more BFV is available. At that point, additional plasticizer molecules cannot be incorporated into the polymer matrix without disrupting the interactions between polymer chains, so additional plasticizer starts acting as a diluent. This situation has the effect of adding ductility to the system (i.e., plasticization [Figure 23.7, top row, right]), so the tensile strength of the compact begins to drop. In a plot of tensile strength vs ERH as the depiction in the top left corner of Figure 23.7, the tensile-strength profile will lie above the horizontal dotted line (the value for a compact free of any water) over a wide ERH range. Let us now consider the case of a compressed polymer matrix of low porosity (Figure 23.7, bottom row). The situation is similar in the sense that water molecules penetrate the BFV in the matrix. However, since the porosity is lower, disruption of polymer-polymer interactions by the incorporation of water occurs at lower water concentrations. This means that water will begin to decrease the tensile strength of the tablet at lower water content than would be the case for higher porosity. In the low-porosity case, a plot of tensile strength vs ERH, depicted in the bottom left corner of Figure 23.7, results in a profile where the curve lies above the horizontal dotted line within a considerably narrower ERH range. In addition (see Figure 23.5), since the tensile strength of the water-free compact of lower porosity is greater, the relative effect of water is smaller, such that the distance between the maximum tensile strength and the horizontal reference line is smaller at low porosity. According to the interpretation presented here, the plasticization threshold (the point where the solid and dotted lines cross) should depend on two main factors: the size of the plasticizer molecule and the total BFV available. For larger BFV and a smaller plasticizer molecule, a greater amount of plasticizer can go into the BFV. This increases the total number of adhesion interactions in the matrix, making it mechanically stronger. Such a situation corresponds to the antiplasticizing effect. In the converse case (smaller BFV and larger plasticizer molecule), smaller amounts of plasticizer fill the BFV, and further incorporation of the plasticizer is impossible without disrupting the already present cohesive interactions between polymer chains. In the hypothetical limit of zero porosity, every molecule of plasticizer added to the system would disrupt polymer-polymer interactions in the matrix. The only effect expected in this case would be the weakening of the polymer matrix, leading to the plasticization of the material. This notion is consistent with the profiles established by the Ryshkewitch analysis shown in Figure 23.4. The analysis shows that no antiplasticization occurs at zero porosity.

Plasticization-Antiplasticization Threshold of Water in MCC

313

Conclusions The results presented here show that, in low and intermediate concentrations, water acts as an antiplasticizer of MCC before exerting its expected plasticizer properties at higher concentrations. The antiplasticizing effect of water is strongly dependent on the porosity of the compacts, which is set by the compression pressure used to create them. These results are significant because MCC is arguably the most commonly used compression aid in pharmaceuticals, and water is present, to some degree, in every pharmaceutical product. Particular attention needs to be paid to the interplay between the porosity of the polymer matrix and the compaction pressure applied during the manufacturing process. This is especially relevant in situations where low amounts of moisture are inadvertently (or unavoidably) present in the MCC. The crystallinity of MCC also changes as function of water content. The changes in crystallinity, however, cannot explain the antiplasticizing effect of water on MCC. It is concluded that changes in crystallinity are a result, rather than the cause, of the antiplasticizing effect of water on the polymer. An alternative explanation, based on BFV, is proposed here. The proposed explanation is based on the notion that antiplasticization occurs when the plasticizer molecules can be incorporated into the polymer matrix without breaking the polymer-polymer interactions already present. The incorporation of low concentrations of the plasticizer leads to polymer-plasticizer interactions that contribute to the overall rigidity and strengthening of the polymer matrix. When the porosity of the polymer matrix is very low or when the voids in the matrix have been filled with plasticizer, further incorporation of plasticizer molecules comes at the cost of disrupting the existing interactions in the system. The disruption of the polymer-polymer interactions leads to a plasticized (softened) polymer material. Further investigations focused on molecular mobility—that is, the ability of the polymer ’s chains to move freely within the matrix as a function of plasticizer concentration—should add significant insight in assessing the validity of the proposed account.

Acknowledgments We thank the National Science Foundation IUCRC 000364-EEC, Dane O. Kildsig Center for Pharmaceutical Processing Research. The French-American Fund for Academic Partnerships program is acknowledged for its support of F.-X.D. We also thank Niraj Trasi for his help in running some of the experiments.

References Abrahamsson B, Lennernäs H. 2003. Application of the biopharmaceutic classification system now and in the future. In: van de Waterbeemd H, Lennernäs H, Artursson P, editors. Drug bioavailability: estimation of solubility, permeability, absorption and bioavailability. Weinheim, Germany: Wiley-VCH. ASTM International. 2002. ASTM E104–02: standard practice for maintaining constant relative humidity by means of aqueous solutions. West Conshohocken, PA: ASTM International. Chamarthy SP. 2007. The different roles of surface and bulk effects on the functionality of pharmaceutical materials [PhD diss]. West Lafayette, IN: Purdue University.

314

PART 1: Invited Speakers and Oral Presentations

Chamarthy SP, Pinal R. 2007. Moisture-induced antiplasticization in microcrystalline cellulose compacts. Tablets Capsules 5:22–33. Ek R, Wormald P, Ostelius J, Iversen T, Nystrom C. 1995. Crystallinity index of microcrystalline cellulose particles compressed into tablets. Int J Pharm 125:257–64. Lourdin D, Bizot H, Colonna P. 1997. “Antiplasticization” in starch-glycerol films? J Appl Polym Sci 63:1047–53. Nelson ML, O’Connor RT. 1964. Relation of certain infrared bands to cellulose crystallinity and crystal lattice type. Part II: A new infrared ratio for estimation of crystallinity in celluloses I and II. J Appl Polym Sci 8:1325–41. Rietema K, Cottaar EJE, Piepers HW. 1993. The effect of interparticle forces on the stability of gas-fluidized beds. II. Theoretical derivation of bed elasticity on the basis of van der Waals forces between powder particles. Chem Eng Sci 48:1687–97. Ryshkewitch E. 1953. Compression strength of porous sintered alumina and zirconia. 9. To ceramography. J Am Ceram Soc 36:65–8. Sears JK, Darby JR. 1982. The technology of plasticizers. New York: John Wiley & Sons. Sun CQ. 2004. A novel method for deriving true density of pharmaceutical solids including hydrates and water-containing powders. J Pharm Sci 93:646–53. Sun CQ. 2005. True density of microcrystalline cellulose. J Pharm Sci 94:2132–4.

24 Understanding the Role of Water in Nonaqueous Pharmaceutical Systems B. D. Anderson, S. S. Rane, and T.-X. Xiang

Abstract This chapter focuses initially on the dynamics and organization of water molecules and factors governing the extent of water uptake in nonaqueous pharmaceutical systems such as lipid-based drug-delivery vehicles and amorphous and crystalline solid-dosage forms. In systems that have been investigated in this laboratory, including triglyceride-monoglyceride lipid mixtures and amorphous polymers (e.g., polyvinylpyrrolidone [PVP]), water distribution is heterogeneous, with molecules tending to form strands or small clusters at relatively low moisture content (<10%) driven by hydrogen-bonding interactions with polar functional groups in the excipients and other water molecules. In lipid mixtures, water is a key contributor to the formation of organized local domains, a characteristic of microemulsions. Over short time scales, water and other solute displacements in both lipid vehicles and amorphous glasses at low water content are dominated by entrapment and jump motions, whereas the formation of water clusters at higher water content tends to alter solute diffusion such that it more closely resembles that in bulk water. The heterogeneity in water distribution, its effects of solute diffusivity and plasticization of amorphous matrices, and the higher diffusivity of water due to its relatively small molecular size become critical factors governing physical and chemical stability in solid formulations. Whereas diffusion-controlled decomposition reactions seldom need to be considered in solution, molecular mobility plays a critical role in amorphous solid-state reaction kinetics. Spatial heterogeneity in dynamic relaxation processes and their time dependence, as well as possible heterogeneity in the distribution of drug molecules, water, and other formulation components, may also be important.

Introduction Water plays a central role in determining the physicochemical properties and performance of virtually every pharmaceutical material and dosage form. Most obvious is the importance of water when it is the solvent for the active pharmaceutical ingredient (API) in a formulation, as in the case of sterile products for injection, oral syrups and suspensions, and other solution or suspension products. For aqueous formulations, the solubility and stability of the API, both of which are primary concerns during product development, are largely determined by the chemistry of the API as influenced by water. 315

316

PART 1: Invited Speakers and Oral Presentations

Considering first drug solubility, the ability of water to participate as a hydrogenbond donor and hydrogen acceptor, the increased polarization of the oxygen and hydrogen atoms that occurs when a water molecule participates in hydrogen-bond formation, the small size of the water molecule, and the tendency of water to selfassociate to form a three-dimensional hydrogen-bonded network are properties that make water unique in several respects that influence API solubility. The ability of water to participate in hydrogen-bonding interactions with drug molecules containing polar functional groups can lead to strong intermolecular (solvation) interactions between water and the drug molecule that enhance solubility. On the other hand, the water-water H-bonding interactions that lead to water ’s structure are a key determinant of the hydrophobic effect, a factor that reduces solubility of APIs having significant nonpolar surface area. The dipolar nature of the water molecule and the electron redistribution that occurs on hydrogen-bond formation leads to cooperativity in hydrogen-bond formation that is further induced by the solvation of ions. The ability of water to solvate ions is a key factor in determining the solution pH and the state of ionization of the drug molecule, both of which influence solubility. In considering the factors contributing to the aqueous solubility of a given compound, solvation in the aqueous phase is only part of the story, however, because the aqueous solubility of a given compound is also governed by its escaping tendency from the solid, which in turn reflects the crystal-lattice forces in the solid phase (for a compound that is crystalline). Even this aspect may be influenced by water, as both the strength of water ’s hydrogen-bond formation with polar functional groups on the drug molecule and the small size of the water molecule can lead to the incorporation of water molecules into the crystal lattice of many APIs to form stoichiometric hydrates as the most stable crystal form in equilibrium with the aqueous solution. The chemical stability (or instability) of the API in aqueous solution or suspension formulations is often a direct consequence of the ability of water to act as a nucleophile, as in hydrolysis reactions, but water plays a critical role even when it is not itself a reactant. Solvation of the transition state by water in relation to its solvation of the ground state is likely a critical factor for nearly any type of chemical reaction in aqueous systems. Also the concentration of catalytic species such as the hydronium ion or other general acids or bases, as well as the state of ionization of the drug molecule itself, which combine to determine the drug’s reactivity all depend on the unique properties of water. Physical instability such as protein aggregation in an injectable formulation or conversion of an API solid in a suspension to a new crystal hydrate also involves water in a central role. If we restrict our inquiry to so-called nonaqueous pharmaceutical systems and therefore consider the properties of either solids such as crystalline or amorphous bulk APIs or pharmaceutical excipients such as various sugars and polymers; nonaqueous liquids that are used as formulation vehicles such as lipids that might be employed in lipid-based oral delivery systems, controlled-release injectables, etc.; or mixtures of these materials, we quickly realize that water continues to play a central role in defining the properties of these materials. Regardless of the chemical composi-

Understanding the Role of Water in Nonaqueous Pharmaceutical Systems

317

tion of the material, water is likely to be present either as an impurity adsorbed on the surface or absorbed (i.e., dissolved) within the material matrix. Although the water concentration may be so low that it is considered only a trace impurity on a weight basis, its chemical potential may be quite high if the system is at equilibrium with atmospheric humidity; sufficient to induce conversion of a pure crystalline solid to a hydrate, for example. Moreover, the small molecular size and therefore low molecular weight of water means that even trace levels of water on a weight basis may be sufficient on a molar basis to account for a relatively high percentage of the reactive nucleophiles present in the solid matrix. This review focuses on the role of water in nonaqueous liquid and amorphous solid pharmaceutical formulations. Several questions are addressed that require better answers before the properties of nonaqueous pharmaceutical materials and the stability, equilibrium state, and performance of the API in these systems can be predicted. Our laboratory, among others, has been approaching the aforementioned problems by using experimental methods and molecular dynamics (MD) simulations. The emphasis herein is on the insights gleaned from MD simulations, so the general procedure for conducting such simulations is presented first. Molecular Dynamics Simulations MD computer simulations involve integration of Newton’s equation of motion for each atom in an ensemble as a function of time in order to generate information on the position, velocity, and acceleration of each particle. For details, the reader may want to refer to a tutorial on the topic (Allen 2004). The force acting on each atom is obtained from a potential function, U, such as that depicted in Equation 24.1 (Tieleman and others 1997). U=

θ kijb kijk 2 eq 2 r − r + ( ) ∑ 2 ij ij ∑ 2 (θijk − θijkeq ) + ∑ k φ [1 + cos ( n (φ − φ eq ))] dihedrals bonds angles qi q j ⎤ ⎡ Aij Bij + ∑ ⎢ 12 − 6 + rij 4π ε o rij ⎥⎦ i < j ⎣ rij

(24.1)

The various terms refer to interactions involving covalent bonds and nonbonded interactions (e.g., van der Waals and electrostatic). A variety of force fields (i.e., the parameter values in the foregoing equation) have been used in simulations, all of which are empirically derived from comparisons of computed properties to experiments. Thus, the suitability of a given force field for a particular application must be established by comparing the results of various properties to experimental data. The computed properties reflect averages over all molecular species, conformations, and solvent structures sampled in the ensemble. MD simulations may provide insights that help answer some of the following questions: (a) What molecular interactions account for the water affinity of a given amorphous solid or nonaqueous pharmaceutical excipient, and can water uptake be predicted? (b) Where is the water localized in the material, and does it affect the

318

PART 1: Invited Speakers and Oral Presentations

organization of the material? (c) How do water molecules diffuse in the matrix? (d) How does the presence of water alter the translational and local mobility of surrounding molecules? Water Uptake and Its Implications in an Amorphous Glass (PVP) Water Uptake and Its Distribution in PVP By probing water uptake and its distribution in amorphous polyvinylpyrrolidone (PVP) glasses, we hoped that MD simulations would provide another perspective and thereby complement experimental studies aimed at addressing the following questions:(a) What molecular interactions account for the water affinity of a given amorphous solid or nonaqueous pharmaceutical excipient, and can water uptake be predicted? (b) Where is the water localized in the material, and does it affect the organization of the material? Two common types of moisture uptake in pharmaceutical solids as relative humidity is varied are stoichiometric, and nonstoichiometric uptake (Byrn and others 1999). As illustrated in Figure 24.1 (top) crystalline hydrate formation typically exhibits stepwise, stoichiometric water uptake with increasing relative humidity (Higgins and others 2003) while amorphous solids and lipid excipients display water uptake profiles that are continuous functions of percent relative humidity (i.e., nonstoichiometric), as illustrated for PVP, polyvinylacetate (PVAc), and their copolymers (Figure 24.1, bottom) (Taylor and others 2001). As shown in Figure 24.1, PVP can spontaneously absorb a substantial amount of water even at a low humidity. The sorbed water can significantly alter the glass transition temperature, polymer structure, and stability of dissolved small molecules or proteins (Hancock and Zografi 1994; Lai and others 1999). Evidently, the wateruptake profiles for PVP and PVAc indicate an important role for the functional group composition of the polymer as a key determinant of water affinity. Taylor and others (2001) concluded from the carbonyl peak shifts in the Raman spectra for the PVPPVAc used as an example in the bottom of Figure 24.1 that at all water contents the PVAc carbonyl interacted with water to a lesser extent than did the PVP carbonyl. Moreover, they demonstrated that the water uptake of the PVP-PVAc copolymer was very close to the predicted profile, assuming additivity of the contributions of PVP and PVAc (i.e., the dashed line in Figure 24.1, bottom), with PVAc sorption of water contributing only to a minor extent. Xiang and Anderson (2004, 2005) conducted MD simulations to explore the distribution and plasticization effects of water, among other phenomena, in PVP polymer assemblies containing six PVP chains (each having 40 monomers) with 8–167 water molecules, corresponding to approximately 0.5% and 10% wt/wt water, respectively. The molecular weight for each PVP chain was 4456 Da, close to the average molecular weight of 4000, which is the reported molecular weight for Kollidon K12 (BASF, Ludwigshafen, Germany), a commercial PVP polymer. Once constructed, the PVP assemblies were energy minimized (500 iterations each of steepest descent and conjugate gradient) to eliminate possible bad contacts and then equilibrated at 1 bar and 700–900 K and 298 K, respectively, by dynamic simulations (1–2 ns) subjected to

Understanding the Role of Water in Nonaqueous Pharmaceutical Systems

319

Moles of water/mole API

5 4 tetrahydrate

3 2

dihydrate

1 0

hemihydrate

anhydrate 0

10

20

30

40

50

60

70

80

90

100

% Relative humidity

Water uptake (wt. fraction)

0.45 0.40 0.35

O

N

0.30

CH

0.25

CH 3

0.20

O

0.15

C H2

n

PVP

O

CH 3

HC C H2

0.10

n

HC

HC

O C H2

C H2

PVAc

0.05

N

O O

n

PVP-PVAc

0.00 0

20

40

60

80

100

% Relative humidity

Figure 24.1. Stoichiometric and nonstoichiometric water uptake. (Top) Stoichiometric water uptake of a bulk drug active pharmaceutical ingredient (API) as a function of percent relative humidity at 25°C (modified from Higgins and others 2003). (Bottom) Nonstoichiometric water uptake in polyvinylpyrrolidone (PVP K12), polyvinylacetate (PVAc), and a PVP-PVAc. copolymer as a function of percent relative humidity at 22°C. The dashed line is the predicted profile, assuming additivity of the contributions of PVP and PVAc. Modified from Taylor and others (2001).

periodic boundary conditions. The equilibrated PVP systems were then subjected to dynamic runs during which they were cooled to 200 K at 0.1 K/ps. After cooling, a selected PVP-water assembly corresponding to a certain temperature (e.g., 298 K) from the acquired trajectory file was used as a restarting configuration for a prolonged dynamic run (0.1 μs) at the same temperature and pressure to acquire system trajectories at intervals of 2–4 ps.

320

PART 1: Invited Speakers and Oral Presentations

At experimental heating rates of 10 K/min, the temperature at which PVP K12 undergoes a transition from an amorphous glass to a rubbery state is approximately 373 K (Khougaz and Clas 2000). The extremely rapid cooling rate (0.1 K/ps) in an MD simulation (necessary because of the limited computational times possible) results in a rapid increase in the polymer viscosity such that the polymer chains cannot undergo the relaxation motions necessary to maintain equilibrium even at a relatively high temperature. Consequently, the system falls out of equilibrium to form a glass at a much higher apparent glass transition temperature (Tg). Apparent Tg values obtained for the simulated PVP glasses were ca. 200 K higher than the experimental value. The densities of the simulated systems at 298 K and density and enthalpy versus temperature profiles indicated that these simulated metastable glasses continued to relax over the time frame of the simulations. Even after a 100-ns simulation, however, the simulated glass densities were 3%–6% higher than the experimentally observed densities of PVPs of similar molecular weight (Khougaz and Clas 2000), indicating that present-day simulations do not produce the glasses that are accessible experimentally. Snapshots of the spatial distribution of water in the simulated PVP (PVP chains are present but not illustrated) are shown in Figure 24.2. At 0.5% wt/wt water (left), the water molecules are distributed uniformly throughout the polymer and appear to be mostly monomeric, while, at 10% wt/wt water (right), clusters or strands of water molecules occupying channels between the polymer chains predominate. At both the low and high water contents, the majority of the water molecules were hydrogen bonded to a PVP carbonyl. Figure 24.3 displays a distinct peak slightly beyond the close-contact distance (>2.6 Å) for the radial distribution function, g(r), between the oxygen atoms in water and the carbonyl oxygen at 0.5% wt/wt water,

Figure 24.2. Spatial distributions of water molecules obtained in molecular dynamics simulations from two instantaneous structures of the polyvinylpyrrolidone (PVP) glasses containing 0.5% (left) and 10% (right) water by weight. Reproduced from Xiang and Anderson (2005), with permission.

Understanding the Role of Water in Nonaqueous Pharmaceutical Systems

321

8

O C

6

H

g (r)

O

H

4

2

0 0

2

4

6

8

10

r (Å)

Figure 24.3. Radial distribution functions, g(r), between the oxygen atoms in water and carbonyl oxygen atoms in polyvinylpyrrolidone (PVP) (solid diamonds) or the hydrogen atoms in all PVP methylene groups (open diamonds) at 0.5% water content. Adapted from Xiang and Anderson (2004), with permission.

indicating a preference for water molecules to reside near PVP carbonyl groups (Xiang and Anderson 2004). Figure 24.3 also shows that there is no such preferred location for water molecules relative to the hydrogen atoms in the methylene groups in PVP. Similar results were obtained at 10% wt/wt water (Xiang and Anderson 2005). The state of water in PVP could be characterized more quantitatively by determining the number of water molecules (nw) within a distance of 3.4 Å of another PVP carbonyl oxygen, a distance consistent with hydrogen-bond formation, as shown in the left of Figure 24.4 at 10% wt/wt water (Xiang and Anderson 2005). Nearly twothirds of the carbonyl oxygen atoms in PVP are within 3.4 Å of at least one water molecule and the average number of water molecule neighbors per carbonyl oxygen at 10% wt/wt water was 0.86, with a rather wide standard deviation. Since the number of PVP carbonyl oxygen atoms (240) exceeds the number of water molecules (167) in the simulation even at 10% wt/wt water, it appears that the majority of water molecules in PVP are hydrogen bonded to PVP carbonyls. Fourier transform infrared (FTIR) spectra of sorbed water in PVP obtained by Lebedeva and others (1999) were consistent with these simulation results. Lebedeva and others (1999) concluded from the absence of a band in the spectra for unbound -OH groups that all the water molecules were either hydrogen bonded to other absorbed water molecules or to PVP. They concluded that at low water content the water molecules are predominantly hydrogen bonded to the PVP carbonyl groups. The tendency for water to self-associate at higher concentrations in PVP has also been observed experimentally. PVP carbonyl and water Fourier transform Raman peak shifts at various relative humidities generated by Taylor and others (2001) were “consistent with a model in which the first one or two water molecules hydrogen bond directly to the hydrophilic group of the polymer side chain, whereas additional water

322

PART 1: Invited Speakers and Oral Presentations

Figure 24.4. Probability distributions for the number of water molecules (nw) surrounding a given polyvinylpyrrolidone (PVP) carbonyl oxygen atom (left) or a given water molecule (right) in a simulated PVP glass with 10% wt/wt water. Adapted from Xiang and Anderson (2005), with permission.

molecules associate and hydrogen bond with other water molecules” (p 899). The concave-upward curvature in the PVP water uptake versus percent relative-humidity profile shown previously (Figure 24.1) may reflect, at least in part, the tendency for water to self-associate with increasing concentration in PVP although plasticization of polymer chains at higher water content also occurs. Water Mobility in Simulated PVP Glasses The translational mobility of water molecules in pharmaceutical glasses may be important in determining both the rates of degradation of drug molecules and the products formed. How do water molecules diffuse in a PVP glass, and how does the presence of water alter the translational and local mobility of surrounding molecules? MD simulations enable one to monitor the relative displacement trajectories of individual molecules (|r(t) – r(0)|), which can then be combined to obtain mean squared displacements (<|r(t) – r(0)|2> or ). Shown in Figure 24.5 are representative displacement profiles versus simulation time for two water molecules in PVP at 0.5% wt/wt water and 298 K. In the inset of Figure 24.5 are shown similar displacement profiles for two water molecules diffusing in bulk water at the same temperature. In the PVP glass, solute displacement is distinctly heterogeneous, as illustrated by the disparate movements of the two water molecules shown in Figure 24.5. Some water molecules reside in the same location

Understanding the Role of Water in Nonaqueous Pharmaceutical Systems

60

323

400 300 200

50

100

|r(t)-r(0)| (Å)

0 0

40

5000

10000 15000 20000

30 20 10 0 0

20000

40000

60000

80000

100000

Time (ps)

Figure 24.5. Representative displacement (|r(t) – r(0)|) profiles versus simulation time for two water molecules in polyvinylpyrrolidone (PVP) at 0.5% wt/wt water and 298 K. (Inset) Displacement profiles for two water molecules diffusing in bulk water at the same temperature. Reproduced from Xiang and Anderson (2004) with permission.

for a relatively long period before hopping to a new location. At 0.5% wt/wt water, hopping of water molecules back and forth between interconnected microdomains within the virtually frozen polymer glass matrix occurred many times (about 82% of the jumps for water were followed by subsequent hops back to the original location). A small fraction of such jumps led to a water molecule’s escape from its original location into a new microdomain, only to again find itself entrapped. Displacement of individual water molecules in bulk water is more continuous, as shown by the curves in the inset of Figure 24.5. Apparent diffusion coefficients can be calculated from the slopes of linear fits of versus time ( = 6 Dt). The diffusion coefficient for water was estimated in simulations to be ca. 1 × 10−7 cm2/s in PVP containing 0.5% wt/wt water and 3.4 × 10−5 cm2/s in bulk water (Xiang and Anderson 2004). The latter value compares favorably with the experimental value for water of 2.2 × 10−5 cm2/s, but the value in simulated PVP appears to be higher than the experimental diffusivity of water in PVP (Oksanen and Zografi 1993; Chang and others 1997; Rodriguez and others 2003). Oksanen and others (1993), for example, reported a value of 6.5 × 10−9 cm2/s in PVP K90 at 25°C at 1.5% wt/wt water, significantly lower than the simulated value. This disparity may be due to the lower density (i.e., higher free volume) of the simulated glass compared to the PVP glass used in experiments, or it may reflect the short time during which diffusion coefficients are calculated in simulations and the relatively higher apparent (non-Einsteinian) diffusion observed over short time frames due to the relative ease of solute hopping between interconnected microdomains.

324

PART 1: Invited Speakers and Oral Presentations

Regardless of these differences, both experiments and simulations indicate that water is highly mobile in PVP even in the glassy state, in sharp contrast to the immobilized nature of the PVP molecules themselves. A single tripeptide molecule (PheAsn-Gly, MW = 336) was also included in the MD simulations in PVP at both 0.5% and 10% wt/wt water (Xiang and Anderson 2004, 2005). This larger molecule remained virtually frozen in place over the time frame of the simulations, demonstrating the significant mobility advantage of the much smaller water molecule relative to a molecule comparable in size to a typical drug molecule. The diffusion coefficient of water increased significantly in simulations performed at a water content of 10% wt/wt, to ca. 5 × 10−7 cm2/s. Similarly, marked increases in translational mobility of water were observed experimentally with increasing water content, independent of transitions of the polymer from a glassy to a rubbery state (Oksanen and Zografi 1993). Simulations indicated that a large fraction of the water molecules in PVP exist in strands or clusters at a water content of 10% wt/wt. Was this effect of higher water concentration on water mobility due to a global effect on the polymer structure (e.g., plasticization of polymer chains), or was it more closely related to the effect of neighboring water molecules on diffusivity of a given water molecule? To explore the latter possibility, all water molecules in the simulated PVP glasses at either 0.5% or 10% wt/wt water were tracked to determine their diffusion coefficients and sorted according to the number of surrounding (first shell) water molecule neighbors each possessed (averaged over the entire simulation). The results, shown in Figure 24.6, indicate that water molecules surrounded by other water molecules tend to have a higher diffusivity, indicative of a local plasticization of water movement by other water molecules. A least-squares regression analysis gave a posi-

D (cm2/s)

1.E-06

5.E-07

0.E+00 0.0

0.5

1.0

1.5 nw

2.0

2.5

3.0

Figure 24.6. Relationship between the diffusion coefficient (D) for water in a simulated polyvinylpyrrolidone (PVP) glass containing 0.5% or 10% wt/wt water and the number of first-shell water molecules. Modified from Xiang and Anderson (2005). (Inset) A tagged (white arrow) water molecule H-bonded to another water molecule on the right within a free-volume cavity in the simulated PVP glass. E, exponential (for example, 1.E-06 = 1 × 10−6); and nw, number of water molecules.

Understanding the Role of Water in Nonaqueous Pharmaceutical Systems

325

tive slope of 2.6 × 10−7 cm2/s for the increase in water diffusivity per nearest neighbor water molecule. Implications of Water Uptake and Mobility on Pharmaceutical Stability in Amorphous Solids Generally, it is expected that chemical reactions in amorphous solids will be coupled to structural relaxation of the matrix since both require molecular mobility to occur (Shamblin and others 2006; Abdul-Fattah and others 2007). Thus, the well-known Tg lowering in amorphous materials with increasing water content leads to the expectation that chemical reactions will be facilitated at a higher water content (Oksanen and Zografi 1990, 1993). While such generalizations are conceptually appealing, there are many relaxation processes to consider that produce a distribution of relaxation times (Shamblin and others 1999, 2000). The problem of which of the relaxation processes are coupled to reactivity in a given reaction remains an open question (Shamblin and others 2006). Some chemical reactions in amorphous solids exhibit little or no sensitivity to Tg or other indicators of molecular mobility. Yoshioka and others (2000) found, for example, that the hydrolysis of aspirin and cephalothin in lyophilized formulations as a function of temperature did not show distinct breaks either at Tg or at a proton nuclear magnetic resonance (NMR) based critical mobility temperature. The mobility of water itself also does not appear to be sensitive to the underlying transformations of PVP in going from a glassy to a rubbery state (Oksanen and Zografi 1993). A more recent example of a lack of coupling of reactivity to other measures of structural relaxation is presented in Table 24.1, which lists the times for 10% degradation of cefoxitin sodium either as the amorphous drug alone or in mixtures with sucrose or trehalose in relation to molecular mobility as reflected in Tg or structural relaxation as measured by enthalpy relaxation times (τβ) by using a thermal activity monitor (TAM) (Shamblin and others 2006). The expectation if reactivity were coupled to structural relaxation is that sucrose, having a lower Tg, would increase the rate of chemical degradation of cefoxitin, whereas trehalose would have only a small effect compared to the reactivity of the drug alone. Instead, sucrose provided significant stabilization. The authors concluded that there was “no correlation whatsoever between the rate of chemical degradation and molecular mobility” (p 2262) when the τβ values and t10% values were compared. An important complicating factor, that has not received adequate attention in studies probing the relationships between chemical reactivity and other indicators of structural relaxation, is that most reactions of pharmaceutical importance, such as acyl-transfer reactions, the Maillard reaction, and various oxidation processes, are multistep reactions involving one or more reactive intermediates. The rate-determining step in such reactions may involve the formation or breakdown of a given intermediate. Also, the products formed in such reactions may be determined by reactions occurring after the rate-determining step for the overall reaction and reflect competition among several potential reactants in the formulation for a reactive intermediate. An illustration of this complexity is shown in Figure 24.7, where an acyl-transfer reaction occurring at

326

PART 1: Invited Speakers and Oral Presentations

Table 24.1. Times for 10% degradation (t10%) of cefoxitin sodium compared with measurements of structural relaxation, including enthalpy relaxation times (τβ), by using isothermal microcalorimetry and glass transition temperatures (Tg) in amorphous cefoxitin sodium formulations at 40°C H 2N

O O

Cefoxitin sodium CH3

S

O O

N S

NH O

+

O

Hydrolytic & nonhydrolytic degradation

Na-O

t10% (h)

τβ (h)

Tg (°C)

Drug only

110

39

127

Drug-trehalose (1 : 10)

150

97

114

Drug-trehalose (1 : 3)

120

105

114

Formulation

Drug-sucrose (1 : 10)

380

28

75

Drug-sucrose (1 : 3)

250

22

79

From Shamblin and others (2006).

the A-21 position in the insulin molecule was found to result primarily in deamidation products or covalent dimer formation, depending on water content (Strickley and Anderson 1996, 1997, 2001). The overall rate of insulin degradation was governed by the formation of an anhydride intermediate in an intramolecular reaction step, whereas the products formed were determined by a bimolecular reaction of either water or another insulin molecule with this intermediate. Figure 24.7 shows that while the rate constant for the overall reaction was independent of Tg, the products formed were highly coupled to Tg. This was attributed to the intramolecular nature of the rate-determining first reaction step, which would likely depend on local segment mobility, whereas the fraction of dimer formation depends on translational and/or reorientational motion of the total insulin molecule. Water has a relative mobility advantage over insulin at temperatures well below Tg, because of its small molecular size, and therefore deamidation predominates in the glass. Water Uptake, Distribution, and Effects on Drug Solubility in Lipid Vehicles Composed of Triglycerides and Monoglycerides Interest in lipid-based drug-delivery systems has grown recently because of the high percentage of new chemical entities entering drug development that have inadequate solubility and therefore limited oral absorption (Gursoy and Benita 2004; Rane and Anderson 2008a). The potential advantage in using hard or soft gel capsules contain-

Understanding the Role of Water in Nonaqueous Pharmaceutical Systems

327

O H2O (MW = 18)

A-21 Terminus O R

NH 2 r.d.s. OH

R

O Deamidation

O

NH 3

OH OH

O R

O

O Insulin (MW = 6000)

1st step – intramolecular

Dimerization

2nd step -bimolecular

0.5

3

Fraction insulin dimer

2

0.3 0.2 = fraction dimer 0.1

= overall rate const.

Tg

1

0 (collapsed phase)

(glassy phase) 0

5

Overall rate const. (h–1 × 104)

0.4

10

15

20

15

% Water

Figure 24.7. Differential coupling of the overall rate constant for insulin degradation and the fraction dimerization to Tg in amorphous lactose formulations at 25°C and varying percentages of water (Strickley and Anderson 2001).

ing the drug dissolved in a semisolid or liquid lipid vehicle for oral administration is that the drug is introduced in its dissolved state, thus bypassing the slow dissolution step that is often rate limiting for absorption of highly lipophilic and poorly watersoluble drug candidates. Typical lipid-based delivery vehicles are composed of oils or oil mixtures (e.g., triglycerides of varying chain length) along with ionic or nonionic surfactants or emulsifiers (e.g., long-chain carboxylic acids, various pegylated surfactants, monoglycerides and diglycerides, or phospholipids) and possibly water-soluble organic solvents (e.g., polyethylene glycol or glycerol). Water is typically present at varying levels in natural oils and other components of lipid vehicles. In vegetable oils consisting largely of long-chain triglycerides, water contents in the range of 300–1000 have been reported (Kurashige and others 1991;

328

PART 1: Invited Speakers and Oral Presentations

Land and others 2005) but can be much higher in certain oils. Castor oil, for example, has a reported water content of 9000 ppm at saturation (Takaoka and others 1999). Monoglycerides can contain substantially larger amounts of water (e.g., 10%–15% wt/wt [Friberg and Mandell 1970]). The water present in lipid-based drug-delivery vehicles may have important consequences on the properties of the formulation in terms of (a) influencing the solvation of drug molecules by altering the hydrogen-bond donating/accepting nature of the vehicle; (b) inducing or promoting organizational structure within the vehicle, which in turn may affect drug solubility by altering the microenvironment in which the solute resides or creating interfaces that may be important sites for drug solubilization; (c) providing the driving force for hydrate formation; or (d) facilitating chemical reactions either directly (e.g., as a nucleophile in hydrolysis reactions) or indirectly through solvation effects. Water uptake in lipid mixtures tends to correlate with the molar concentration of polar, hydrogen-bonding functional groups in the vehicle, as demonstrated in Figure 24.8, where the percent wt/wt of water in tricaprylin-monocaprylin and tricaprylinmonocaprin mixtures at 37°C and 100% relative humidity are plotted as a function of the molar concentration of the monoglyceride component (Rane and others 2008). Profiles of water uptake tend to fall on the same “master curve” for different monoglycerides in the same triglyceride when the concentration of the monoglyceride is expressed as a molar quantity, indicating the importance of the concentration of the polar functional groups in the monoglyceride (rather than acyl chain length) in solvating water. 10 OH O

8 % wt/wt water

OH

CH3 O 1-Monocaprylin (C8)

6

4 1-Monocaprin (C10)

OH

2

O OH

0 0

0.5

CH3 O

1 mol/L monoglyceride

1.5

2

Figure 24.8. Water uptake in tricaprylin-monocaprylin (solid circles) and tricaprylinmonocaprin (open circles) mixtures at 37°C and 100% relative humidity as a function of the molar concentration of the monoglyceride component. Modified from Rane and others (2008).

Understanding the Role of Water in Nonaqueous Pharmaceutical Systems

329

The organization of the lipid molecules in triglyceride-monoglyceride-water or monoglyceride-water mixtures has been studied experimentally by freeze-fracture electron microscopy (Gulik-Krzywicki and Larsson 1984) and X-ray scattering (Larsson 1979), leading to the view that microemulsions formed by these combinations are highly structured, with layers of hydrocarbon chains being separated by undulating water layers lined by the polar groups of the lipid molecules. Drug solubility in such interfacial systems may be quite different than would be expected if the vehicles were homogeneous. For example, Spernath and others (2002) demonstrated that the solubility of lycopene in a complex microemulsion changed with the microstructure. Upon the addition of water, the microstructure varied from a water in oil (w/o) microemulsion, to a bicontinuous microemulsion, to an oil in water (o/w) microemulsion. The changes in lycopene solubility were related to the preferred location of the solute in these heterogeneous systems. MD computer simulations may be quite useful for exploring the molecular organization of the lipid components, water, and solute in lipid-based drug-delivery vehicles and thereby contribute to a more fundamental understanding of drug solubility in these systems. This insight may ultimately enable the prediction of drug solubility as a function of lipid-based drug-delivery system composition. To mimic experimental systems that were also investigated in these laboratories, we conducted MD simulations in a 60% tricaprylin–40% 1-monocaprylin mixture saturated with water. The cubic simulation box contained 111 molecules of 1-monocaprylin and 77 molecules of tricaprylin with 219 molecules of TIP3P water (transferable intermolecular potential 3P water model) (Rane and Anderson 2008b). A single solute molecule (benzamide) was also included in the simulation box. The simulations were performed using AMBER 8 (molecular dynamics simulation program packages; University of California, San Francisco, USA), and charges were assigned using the AM1-BCC charge model (AM1 atomic charge–bond charge-correction model for condensed-phase simulations; Merck Frosst Canada, Quebec, Canada). Shown in Figure 24.9 is a snapshot of the distribution of water molecules in the aforementioned simulation. At the water content present at equilibrium at 100% relative humidity, the water molecules are clearly not uniformly distributed. Most are present in clusters or interconnecting strands. Over the time course of a simulation, the individual water molecules are highly mobile, moving from one cluster to another. Water aggregates may disappear as new ones form, or a given cluster may expand while another contracts. Figure 24.10 shows the radial distribution function between water molecules in the water-saturated lipid mixture and C = O oxygen atoms in monocaprylin and tricaprylin (solid line), -OH oxygen groups in monocaprylin (long dashes), and alkyl carbon atoms in monocaprylin (dotted line) or tricaprylin (dash-dot line) along with an inset illustrating a small water cluster, its proximity to neighboring polar hydrogen-bonding functional groups, and the hydrogen bonds that have formed as indicated by light-gray bands. Both the radial distribution function and the inset snapshot indicate that the microenvironment surrounding the H-bonded water cluster is rich in polar -OH and C = O groups and depleted in alkyl groups, consistent with suggestions from the results

330

PART 1: Invited Speakers and Oral Presentations

Figure 24.9. Snapshot from a molecular dynamics simulation of a 60% tricaprylin– 40% 1-monocaprylin lipid mixture saturated with water at 37°C. The distribution of water molecules is shown with the lipid chains in the background. A single solute molecule (benzamide) was also present in the simulations.

of experimental studies. The peak heights in the radial distribution function at a distance of ca. 3 Å suggest that the region surrounding water is slightly richer in hydroxyl groups than in carbonyl ester groups. Further, more detailed analyses of the solvation energies obtainable in MD simulations will be necessary to ascertain the potential consequences of water and waterinduced structure on drug solubility in lipid-based drug-delivery systems. One very important factor to keep in mind when considering the effects of water content on drug solubility in lipid vehicles is possible hydrate formation. A seemingly small water content in a lipid mixture on a percent wt/wt basis at 100% relative humidity may lead to the conversion of an anhydrate to a hydrate crystal form with lower solubility due to the high chemical potential for water in such systems. Land and others (2005) examined the solubility of several poorly water-soluble model steroids, including progesterone, estradiol, and testosterone, in various triglycerides in the desiccated (<0.0025% wt/wt water) or fully hydrated (0.08%–0.24% wt/wt water) states to assess the influence of water content on solubility. They found a 30%–40% lower solubility for estradiol (Figure 24.11) and testosterone in the hydrated oils compared to the

8 7 6

RDF

5 4 3 2 1 0 0

5

10 Distance, Å

15

20

Figure 24.10. Radial distribution function (RDF) for water oxygen with monocaprylin and tricaprylin C=O oxygen atoms (solid line), monocaprylin -OH oxygen groups (long dashes), and alkyl carbon atoms in monocaprylin (dotted line) or tricaprylin (dash-dot line). (Inset) Snapshot of a small water cluster showing the various hydrogen bonds (light-gray bands, white arrows) between water molecules and neighboring hydroxyl and ester carbonyl groups.

Figure 24.11. Equilibrium solubility of testosterone in triglyceride oils that were either desiccated (dry) or water saturated (Land and others 2005). The starting material for the attainment of the equilibrium solubilities was either testosterone monohydrate or anhydrous testosterone (anhydrate).

331

332

PART 1: Invited Speakers and Oral Presentations

desiccated oils at equilibrium. In hydrated oils, the solid phase at equilibrium was a hydrate regardless of whether the compound initially added to the lipid vehicle was a hydrate or an anhydrate. Based on these results, it seems that the water content in lipid excipients may need to be carefully controlled in some cases.

Conclusions Moisture uptake in amorphous pharmaceutical matrices and lipid vehicles may have profound effects on both physical stability (e.g., solubility and hydrate formation) and chemical stability of the API (e.g., hydrolysis and deamidation). The results of equilibrium uptake experiments as well as MD simulations indicate that the uptake of moisture in nonaqueous pharmaceutical systems is governed largely by hydrogen bonding of water to polar functional groups in the excipient at low water concentrations and to other water molecules at higher moisture levels. This may give rise to nearly superimposable master curves for moisture uptake versus molarity of the more polar excipient in excipient mixtures wherein the components differ only in chain length. Self-association of water at higher moisture levels has been observed in amorphous glasses (e.g., PVP), as well as in lipid-based drug-delivery vehicles (e.g., triglyceride-monoglyceride-water mixtures) to produce water clusters or strands. Microdomains of high water content may induce local structure in lipid vehicles that alter drug solubility. In amorphous glass formulations, the diffusion coefficient of water is enhanced in water clusters. Enhanced local mobility in amorphous glasses may be important in the chemical stability of API in such formulations. Because translational mobility in amorphous solid formulations appears to be highly sensitive to the molecule size of the diffusing molecule, chemical reactions involving smaller, more mobile reactants such as water may occur preferentially in highly viscous solid matrices.

References Abdul-Fattah AM, Dellerman KM, Bogner RH, Pikal MJ. 2007. The effect of annealing on the stability of amorphous solids: chemical stability of freeze-dried moxalactam. J Pharm Sci 96:1237–50. Allen MP. 2004. Introduction to molecular dynamics simulation. In: Attig N, Binder K, Grubmuller H, Kremer K, editors. Computational soft matter: from synthetic polymers to proteins, lecture notes. Juelich, Germany: John von Neumann Institute for Computing. p 1–28. Byrn SR, Pfeiffer RR, Stowell JG. 1999. Solid-state chemistry of drugs. 2nd ed. West Lafayette: SSCI. Chang MJ, Myerson AS, Kwei TK. 1997. The effect of hydrogen bonding on vapor diffusion in watersoluble polymers. J Appl Polym Sci 66:279–91. Friberg S, Mandell L. 1970. Phase equilibria and their influence on the properties of emulsions. J Am Oil Chem Soc 47:149–52. Gulik-Krzywicki T, Larsson K. 1984. An electron microscopy study of the L2-phase (microemulsion) in a ternary system: triglyceride/monoglyceride/water. Chem Phys Lipids 35:127–32. Gursoy RN, Benita S. 2004. Self-emulsifying drug delivery systems (SEDDS) for the improved oral delivery of lipophilic drugs. Biomed Pharmacother 58:173–82. Hancock BC, Zografi G. 1994. The relationship between the glass transition temperature and the water content of amorphous pharmaceutical solids. Pharm Res 11:471–7.

Understanding the Role of Water in Nonaqueous Pharmaceutical Systems

333

Higgins JP, Arrivo SM, Reed RA. 2003. Approach to the determination of hydrate form conversions of drug compounds and solid dosage forms by near-infrared spectroscopy. J Pharm Sci 92:2303–16. Khougaz K, Clas S-D. 2000. Crystallization inhibition in solid dispersion of MK-0591 and poly(vinylpyrrolidone) polymers. J Pharm Sci 89:1325–34. Kurashige J, Takaoka K, Takasago M, Taru Y, Kobayashi K. 1991. State of dissolved water in triglycerides as determined by Fourier transform infrared and near infrared spectroscopy. Yukagaku Zasshi 40:549–53. Lai M, Hageman MJ, Schowen RL, Topp EM. 1999. Chemical stability of peptides in polymers. 1. Effect of water on peptide deamidation in poly(vinyl alcohol) and poly(vinylpyrrolidone) matrices. J Pharm Sci 88:1073–80. Land LM, Li P, Bummer PM. 2005. The influence of water content of triglyceride oils on the solubility of steroids. Pharm Res 22:784–8. Larsson K. 1979. An X-ray scattering study of the L2-phase in monoglyceride-water system. J Colloid Interface Sci 72:152–3. Lebedeva TLF, Mikhail M, Kuptsov SA, Plate NA. 1999. A.V. spectroscopy of biological molecules. In: Greve J, Puppels GJ, Otto C. editors. Spectrocopy of biological molecules: new directions. Eighth European Conference on the Spectroscopy of Biological Molecules, Aug 29–Sept, Enschede, The Netherlands. Dordrecht, The Netherlands: Kluwer Academic. p 581–2. Oksanen CA, Zografi G. 1990. The relationship between the glass transition temperature and water vapor absorption by poly(vinylpyrrolidone). Pharm Res 7:654–7. Oksanen CA, Zografi G. 1993. Molecular mobility in mixtures of absorbed water and solid poly(vinylpyrrolidone). Pharm Res 10:791–9. Rane SS, Anderson BD. 2008a. Molecular dynamics simulations of functional group effects on solvation thermodynamics of model solutes in decane and tricaprylin. Mol Pharmaceutics 5:1023–36. Rane SS, Anderson BD. 2008b. What determines drug solubility in lipid vehicles: is it predictable? Adv Drug Deliv Rev 60:638–56. Erratum in Adv Drug Deliv Rev 60:1674, 2008. Rane S, Cao Y, Anderson BD. 2008. Quantitative solubility relationships and the effect of water uptake in triglyceride/monoglyceride microemulsions. Pharm Res 25:1158–74. Rodriguez O, Fornasiero F, Arce A, Radke CJ, Prausnitz JM. 2003. Solubilities and diffusivities of water vapor in poly(methylmethacrylate), poly(2-hydroxyethylmethacrylate), poly(N-vinyl-2-pyrrolidone) and poly(acrylonitrile). Polymer 44:6323–33. Shamblin SL, Hancock BC, Dupuis Y, Pikal MJ. 2000. Interpretation of relaxation time constants for amorphous pharmaceutical systems. J Pharm Sci 89:417–27. Shamblin SL, Hancock BC, Pikal MJ. 2006. Coupling between chemical reactivity and structural relaxation in pharmaceutical glasses. Pharm Res 23:2254–68. Shamblin SL, Tang X, Chang L, Hancock BC, Pikal MJ. 1999. Characterization of the time scales of molecular motion in pharmaceutically important glasses. J Phys Chem [B] 103:4113–21. Spernath A, Yaghmur A, Aserin A, Hoffman RE, Garti N. 2002. Food-grade microemulsions based on nonionic emulsifiers: media to enhance lycopene solubilization. J Agric Food Chem 50:6917–22. Strickley RG, Anderson BD. 1996. Solid-state stability of human insulin. I. Mechanism and the effect of water on the kinetics of degradation in lyophiles from pH 2–5 solutions. Pharm Res 13:1142–53. Strickley RG, Anderson BD. 1997. Solid-state stability of human insulin. II. Effect of water on reactive intermediate partitioning in lyophiles from pH 2–5 solutions: stabilization against covalent dimer formation. J Pharm Sci 86:645–53. Strickley RG, Anderson BD. 2001. Solid-state stability of human insulin: the effect of water content and glass transition on the chemical kinetics of the Maillard reaction, covalent dimerization, and deamidation. Paper presented at the 15th Annual Meeting of the American Association of Pharmaceutical Scientists, Denver, CO. Takaoka K, Kobayashi K, Takahashi M, Sone M. 1999. Studies on interaction between castor oil and dissolved water by hot wire method. Nippon Kagaku Kaishi 10:649–54. Taylor LS, Langkilde FW, Zografi G. 2001. Fourier transform Raman spectroscopic study of the interaction of water vapor with amorphous polymers. J Pharm Sci 90:888–901.

334

PART 1: Invited Speakers and Oral Presentations

Tieleman DP, Marrink SJ, Berendsen HJ. 1997. A computer perspective of membranes: molecular dynamics studies of lipid bilayer systems. Biochim Biophys Acta 1331:235–70. Xiang T-X, Anderson BD. 2004. A molecular dynamics simulation of reactant mobility in an amorphous formulation of a peptide in poly(vinylpyrrolidone). J Pharm Sci 93:855–76. Xiang T-X, Anderson BD. 2005. Distribution and effect of water content on molecular mobility in poly(vinylpyrrolidone) glasses: a molecular dynamics simulation. Pharm Res 22:1205–14. Yoshioka S, Aso Y, Kojima S. 2000. Temperature dependence of bimolecular reactions associated with molecular mobility in lyophilized formulations. Pharm Res 17:925–9.

25 Crystallization, Collapse, and Glass Transition in Low-Water Food Systems Y. H. Roos

Abstract Collapse phenomena in food systems are important because they affect food dehydration characteristics and deterioration during the storage of foods at intermediate and low water contents. These phenomena often include stickiness and flow properties of particles in dehydration and manufacturing of powders; collapse in freeze drying; and stickiness, caking, and loss of structure during the storage of dehydrated materials. Collapse phenomena result from a change in stiffness and flow properties as a result of the glass transition of food solids. Diffusion, reaction rates, and component crystallization can be affected significantly by the glass transition and the consequent changes in food structure. The glass transition and structural properties of low-water food systems can be described by state diagrams that include information on water plasticization and frozen-state properties, as well as saturation of dissolved components. Food composition can be manipulated to control collapse phenomena in dehydration and storage of low-water food systems.

Introduction At low water content, food systems exist as solid, noncrystalline materials or as highly concentrated, viscoelastic fluids. The physicochemical and flow properties of lowwater foods are often controlled by the physical state of their hydrophilic continuous phase with dissolved and dispersed components. These systems are typically in a thermodynamically metastable, nonequilibrium state. The importance of the physical state of low-water systems, including frozen and other foods, and its sensitivity to temperature and water content have been well recognized and discussed in numerous studies and reviews (e.g., see Roos 1995, 2008; Slade and Levine 1995). Collapse phenomena in food systems are important because they affect food dehydration characteristics and deterioration during the storage of foods at intermediate and low water contents (Tsourouflis and others 1976; Slade and Levine 1995). These phenomena often include the stickiness and flow properties of particles in dehydration and manufacturing of powders; collapse in freeze drying; and stickiness, caking, and loss of structure during the storage of dehydrated materials (Roos 1995). Collapse phenomena result from a change in stiffness and flow properties as a result of the glass transition of food solids. Diffusion, reaction rates, and component crystallization can be affected significantly by the glass transition and the consequent changes in food 335

336

PART 1: Invited Speakers and Oral Presentations

structure. Molecules of materials above their glass transition temperature exhibit translational mobility, and the glass transition causes rapid changes in flow properties. This is observed from the study of interparticle cohesion and surface adhesion of particles as liquid properties appear at particle surfaces; structural collapse as a result of time-dependent, viscous flow; and crystallization of food components. For example, stickiness, caking, collapse, and crystallization of amorphous lactose are caused by lactose glass transition in dehydrated dairy foods (Jouppila and Roos 1994a, 1994b; Roos 2002). These collapse phenomena are time dependent and significantly enhanced by increasing temperature or water content above those characterizing the glass transition. Structural collapse in low-water foods often affects the stability of encapsulated components, the rates of deteriorative changes, and contributes to flavor loss since diffusion and reaction rates are enhanced by structural transformations (Roos 2008). Crystallization is a first-order phase transition that may occur during the storage of a number of amorphous foods containing sugars, polysaccharides, and proteins. Crystallization, which is the most dramatic structural change, leads to instant changes in the rates of other physical and chemical changes and reactions (Roos 1995). Food composition can be manipulated to control collapse phenomena during the dehydration and storage of low-water food systems. This often requires modification of miscible components responsible for bulk glass transition and water plasticization properties (Hartel 2001). Furthermore, protective packaging, as well as water-activity control and temperature control, are used to maintain glassy structures in low-water foods. Collapse and crystallization phenomena in this review are considered as glass transition–related time-dependent changes that may occur and govern quality changes during numerous food processes and the storage of low-water and frozen foods.

Flow in Glassy and Liquid States Glass Transition Glass transition, which is a property of amorphous, supercooled materials, appears as a solid-to-liquid transformation with no latent heat. Glass transition exhibits a change in heat capacity indicating a change in thermodynamic properties and molecular mobility. Supercooled liquids undergo a glass transition at about 100°–150°C below the equilibrium melting temperature of their crystalline states (Sperling 1992). The amorphous state, however, is a nonequilibrium state whose properties cannot be well defined in thermodynamic terms. Furthermore, no exact glass transition temperature (Tg) can be defined because the transition is time dependent and occurs over a temperature range. The observation of Tg also depends on the methodology used to detect changes in thermodynamic properties, molecular mobility, or flow properties (Roos 1995, 2008). In foods, miscible components may show composition-dependent glass transition behavior. Several studies have reported Tg-composition relationships for food systems and various approaches for predicting Tg based on component glass transition behavior (Roos and Karel 1991a, 1991b; Kalichevsky and Blanshard 1993). Often, the glass

Crystallization, Collapse, and Glass Transition in Low-Water Food Systems

337

transition of food solids is measured at various water contents, and their water plasticization and accompanying reduction of their Tg with increasing water content are predicted by using the Gordon-Taylor relationship (Roos 1995). Foods, however, may contain miscible, partially miscible, and immiscible components that may show overlapping glass transitions of separate phases and one or more glass transitions, depending on their composition. Furthermore, other transitions, such as fat crystallization and melting, often overlap glass transitions at some levels of water plasticization. Overlapping transitions cannot be separated from each other without individual studies of melting and glass transition properties of food components. Foods can also exhibit high heterogeneity within their microstructure, and numerous glass transitions of miscible, partially miscible, and immiscible components may occur locally within a food system. Glass transition temperatures of anhydrous food components may also be below room temperature, or other typical storage temperatures. Storage at a temperature above the Tg often limits processing and handling properties, as well as storage stability (Roos 1993, 2008). The glass transition is a characteristic property of noncrystalline solids and of foods with low water content. The amorphous state of food components is often a result of processes where solvent water (e.g., dehydration) is rapidly removed or concentrated solids (e.g., extrusion) are cooled to a nonequilibrium state without sufficient time for dissolved solids to attain an equilibrium state (Figure 25.1). This requires that the rate of solvent removal or cooling is faster than the rate of nucleation, thus inhibiting crystallization and enabling molecules to remain in their original amorphous arrangement typical of the dissolved state, or of the liquidlike arrangement of a melt or concentrated fluid. Nonequilibrium amorphous structures have a higher enthalpy and volume than do the equilibrium crystalline structures under the same conditions (Figure 25.2).

Equilibrium Liquid

D

eh

n

n

io

io

at liz

at

ra

tio

n

tic

lu

as Pl

n

io

So

at (Pressure)

Cooling

RUBBER

Nonequilibrium Solid GLASS

Crystallization

oo C id ap R

Sl

g

ow

Co

Cooling

g in ol

H in

ea t

g

in

Heating

at

He

lin

g

Heating

CRYSTAL

yd

iz

bi

Sa tu r

Equilibrium Solid

Flow and Collapse Phenomena

SOLUTION

Equilibrium Liquid

MELT

Figure 25.1. Equilibrium and nonequilibrium states. Collapse phenomena result from time-dependent flow in nonequilibrium states.

338

PART 1: Invited Speakers and Oral Presentations

V H

Liquid Anomalous changes in thermodynamic properties depending on glass characteristics

S

Equilibrium states

Supercooled liquid

Glass

Nonequilibrium states

Crystal

≈ 100-150°C

Tg

Tm

T

Figure 25.2. Enthalpy (H), entropy (S), and volume (V) at various states of materials. Supercooled liquid and glassy states are nonequilibrium states. Temperatures are shown for glass transition (Tg) and equilibrium melting (Tm).

In freeze drying and the dehydration of liquids to powders and in extrusion processes, dissolved molecules must be “frozen” in the solid, glassy state in order to maintain product stability. An infinite number of glassy states of the same material may be formed in food processes. The physical state of the glass produced may be affected by the timedependent characteristics of the “freezing” of molecules to the solid, glassy state. Cooling rate affects the free volume of molecules within the amorphous matrix. This may be observed from various relaxation processes observed when rapidly cooled materials are heated to above the Tg. The glass transition itself is a time-dependent property of the amorphous state, and the relative rate of glass formation may contribute to the properties of food systems. Food freezing also requires the removal of solvent water. Freezing has similarities to dehydration processes. However, in freezing, water is removed from the system into an ice phase, which is dispersed within a freeze-concentrated, unfrozen solute phase. The extent of freeze concentration depends on temperature and solute properties following the melting temperature depression of water caused by the dissolved solids. Maximum freeze concentration may occur at temperatures slightly below the onset temperature of ice melting, where ice formation becomes kinetically inhibited as the unfrozen phase approaches its glassy state (Roos and Karel 1991c, 1991d). The glassy state is a solidlike state with a viscosity exceeding 1012 Pa · s (Sperling 1992). Molecular mobility changes at near glass transition produce dramatic changes in dielectric, mechanical, and thermal properties, and a solid-to-liquid state transition takes place over the Tg range. Relaxation times of molecular, mechanical, dielectric,

Crystallization, Collapse, and Glass Transition in Low-Water Food Systems

Temperature

Tm

T – Tg

Relative

(°C)

relaxation time 10–8

40

2.4 ×

20

1.3 × 10–5

0

1.0 × 100

339

Supercooled liquid

Tg′

s Glas 0

ion ansit

sy las

id

sol

G

tr

Thermal plasticization

Tm′

Water plasticization

Weight fraction of solids

1.0

Figure 25.3. Schematic representation of relaxation times above glass transition temperature (Tg) and the effect of water plasticization on glass transition and relaxation times. The equilibrium ice-melting temperature (Tm) is shown by the Tm curve. At temperatures Tm′ < T < Tm , ice forms until ice and the freeze-concentrated unfrozen solution exist at equilibrium at T = Tm. At Tm′ , ice ceases to form because of increasing viscosity of the unfrozen phase, which shows a glass transition onset temperature at Tg′ . Both Tg′ and Tm′ are independent of the initial concentration of the system prior to ice formation. Tg′ , glass transition temperature at maximally freeze-concentrated solution; and Tm′ , melting temperature at maximally freeze-concentrated solution.

and thermal responses of amorphous materials change rapidly around and above the glass transition (Figure 25.3). Below glass transition, an Arrhenius-type, slow decrease in relaxation times with increasing temperature can be expected (Sperling 1992). Relaxation times above glass transition are often modeled by using the WilliamsLandel-Ferry (WLF), power law, Vogel-Tamman-Fulcher (VTF), Fermi, and other models (Roos and Karel 1991a, 1991b, 1992; Slade and Levine 1991; Peleg 1992, 1993). The exponential decrease in viscosity around and above glass transition is fundamental for understanding changes in flow properties and understanding collapse behavior of food materials at low water content and in the frozen state (Figure 25.4). Relaxation Phenomena of Amorphous Foods Amorphous materials show various relaxation phenomena, depending on the physical state of the material and time allowed for the material to respond to changes in temperature, pressure, or the amount of a plasticizer. Relaxation phenomena are often observed on heating near the glass transition.

PART 1: Invited Speakers and Oral Presentations

Flow

Fermi’s Model (M. Peleg) COLLAPSING LIQUID Increasing Diffusion

SOLID Structural Transformations

Relaxation time

Days Hours Minutes Seconds

Glass Transition

LIQUID

Stability Zone

Critical Zone

Mobility Zone

‘Solid’

‘Highly time-dependent’

‘Instant changes’

Extent of change in property

Months

Glassy State

Crispness

Years

Hardening, Cracking

340

Temperature, water activity or water content

Figure 25.4. Effect of thermal and water plasticization on mechanical, flow, and collapse phenomena on relaxation times and time-dependent changes in food solids.

Enthalpy Relaxations Enthalpy relaxations are observed in calorimetric measurements of supercooled liquids around the glass transition. Enthalpy relaxations may be either endothermic or exothermic, depending on the thermal history and glass formation conditions (Figure 25.5). Several authors have reviewed and discussed enthalpy relaxations in synthetic polymers (Tant and Wilkes 1981; Wunderlich 1981; Roos 1995, 2008). Typically, glass formation by rapid cooling freezes component molecules into a glassy state with a large free volume. A subsequent slow reheating of the material to above glass transition allows sufficient time around the glass transition for molecular rearrangement. This is enhanced by transitional mobility, which appears around glass transition. Therefore, an exotherm and a decrease in volume and entropy can be observed when rapidly cooled amorphous materials are heated to above glass transition. This behavior also suggests that flow properties of the material change toward a liquidlike state, which results in time-dependent collapse phenomena above the glass transition (Roos 1995, 2008). An endotherm around glass transition is often observed when a slowly cooled material (“dense glass”) is heated to above glass transition. This heat input to the material is needed because the increasing mobility of component molecules enables them to attain a higher-mobility state corresponding to expansion and higher heat content of the material at and above the glass transition. Enthalpy relaxations around glass transition result from changes in diffusion, and the enhanced molecular mobility with increasing temperature explains observed increases in enthalpy, volume, and entropy around the glass transition (Roos 2008).

Crystallization, Collapse, and Glass Transition in Low-Water Food Systems

341

Exothermal heat flow

Glassy State A Rate a > Rate b

Exotherm

Glassy State B

Cp Time-dependent changes

Endotherm T

Figure 25.5. Schematic representation of changes in volume associated with endothermic and exothermic enthalpy relaxations around glass transition. Cp, specific heat capacity (at constant pressure).

Structural Relaxations Mechanical and dielectric relaxation times of amorphous materials change dramatically around the glass transition. These relaxations indicate changes in molecular mobility and flow, and can be used as flow or collapse indicators. Observed relaxation times give a measure of the time-dependent response of amorphous materials to mechanical or dielectric disturbance (Figure 25.3). This time-dependent behavior is a result of the nonequilibrium state, and measured responses are also temperature dependent and frequency dependent (Figure 25.6). This behavior indicates clearly the nonequilibrium and time-dependent characteristics of supercooled liquids. Measured structural relaxations often appear around the glass transition measured by calorimetric techniques, and relaxation times decrease rapidly with increasing temperature difference T – Tg (Sperling 1992; Roos 2008). Viscosity changes often follow structural relaxations around and above glass transition, which was a general observation for relaxation times above glass transition (Williams and others 1955; Sperling 1992; Slade and Levine 1995).

Collapse Phenomena Solid materials can withstand deformation, whereas liquid systems flow under an external force or pressure. Flow properties of amorphous materials govern their stability under gravity. Collapse refers to structure loss resulting from the viscous flow of

342

PART 1: Invited Speakers and Oral Presentations

Mechanical or dielectric property

Storage modulus or dielectric constant

Increasing frequency

Mechanical and dielectric relaxations

Loss modulus or dielectric loss

β relaxation

Tg

α relaxation Temperature

Figure 25.6. Schematic representation of mechanical and dielectric relaxations in amorphous materials below and around their glass transition.

components within food structure. Solid, glassy structures can support their structure under gravity, whereas around and above glass transition a rapidly increasing flow produces various collapse phenomena in food processing and storage. These phenomena have been addressed by numerous authors (To and Flink 1978a–c; Downton and others 1982; Karel and others 1994; Levine and Slade 1988). Collapse phenomena occur as molecular mobility increases and relaxation times decrease around and above glass transition. Collapse phenomena are typical of food dehydration and low-water food systems, and include shrinkage and collapse in drying processes, stickiness and caking of powders, and loss of crispness of low-water foods (Figure 25.7). An essential requirement in many dehydration processes is that food solids are transformed into solid, stable structures. This occurs in proper freeze-drying and spray-drying processes (Figures 25.7 and 25.8). The typical solid structures formed in these processes are glassy, noncrystalline solid forms of food solids. Hence, dissolved solids appear as highly saturated components, and their crystallization may be observed in storage (e.g., lactose in powders and in frozen dairy foods). Transformations toward equilibrium states are affected by kinetic limitations of low molecular mobility, diffusion, and high viscosity of the concentrated low-water hydrophilic phases. One of the main requirements in food dehydration and for stability of dehydrated foods is that food solids form a solid structure that can support its own weight under gravity. In dehydration, the temperature and the temperature gradients in the materials undergoing dehydration may be controlled to follow glass transition to minimize flow. Reduced or increased surface plasticization is used to avoid or enable stickiness and caking of particles. A typical example is spray drying, where a rapid formation of a

Crystallization, Collapse, and Glass Transition in Low-Water Food Systems

343

Temperature

Glass Transition Region - Reduced Stickiness, Shrinkage and Collapse

Stickiness Region - Increased Stickiness, Shrinkage and Collapse

Flow Region - Rapid Collapse Dry State

Initial State

Glassy State Dehydration Path

Water content Figure 25.7. Glass transition, stickiness, and collapse properties of amorphous food solids in dehydration at various water contents.

Freeze-concentrated unfrozen solute phase

Freeze-dried glass solute membranes

Pores

Ice

id

Sol

Flo

w

Collapsed liquid

Figure 25.8. Collapse in freeze drying. Freeze-drying conditions maintaining solid, glassy structures of freeze-concentrated solids enable the retention of volume while flow in nonmaximally freeze-concentrated solids results in collapse and loss of quality.

sufficiently solid surface in drying particles is required to avoid particle adhesion and stickiness in equipment. Surface plasticization and dehydration are the basis of, for example, powder agglomeration, manufacturing of instant dairy powders, and granulation processes. In production of air-dried systems, shrinkage may be reduced by control of temperaturewater-time relationships, which minimizes the flow of structure-forming components (Karel and others 1994; Roos 2002). Glass formation as a result of freeze concentration in the freezing and dehydration of glassy, freeze-concentrated solids by sublimation of the dispersed ice phase is

344

PART 1: Invited Speakers and Oral Presentations

required for successful freeze-drying processes. In freeze drying, water must freeze to a maximum extent to form a solid, freeze-concentrated, unfrozen matrix. This structure needs to be formed in food freezing and retained during sublimation of ice to resist flow of the unfrozen solids phase, structural changes, and collapse during freeze drying (Figure 25.8) (Roos 1995; Pehkonen and others 2008). During freezing, solutes in an unfrozen phase must vitrify at temperatures approaching the glass transition of the maximally freeze-concentrated unfrozen phase, which provides a freeze-drying system with required stability under gravity. The glassy, freeze-concentrated solids formed in freezing are dehydrated in the freeze-drying process to a porous, solid matrix that retains the original volume of the frozen material. However, the solid, glassy structures must be maintained in the storage of freeze-dried materials to support and retain their physical structure and volume. A flowing plasticized mass of food solids is often produced by thermal and water plasticization in high temperature–short time (HTST) extrusion processes. Viscous flow is desirable for the mixing and extrusion stages. Expansion and structure formation can be based on the flow and collapse or expanding characteristics of plasticized solids. For example, the solids can be expanded and “frozen” to a glassy structure by using simultaneous dehydration and cooling at the extruder die. Stickiness and Caking Stickiness and caking, which may occur when amorphous food materials are heated or exposed to high-humidity conditions (Figures 25.3 and 25.4), are typical of amorphous food powders. Caking may be considered as a collapse phenomenon that occurs when permanent aggregates form and a powder hardens into a solid mass. This causes an obvious loss of free-flowing properties in powders. The dramatic decrease in viscosity above Tg reduces contact time between particle surfaces needed for particle cohesion and formation of interparticle liquid bridges. Downton and others (1982) estimated that cohesion-produced stickiness occurred with a contact time of 1–10 s. This suggested that a surface viscosity lower than 106–108 Pa · s was sufficient for stickiness and particle cohesion. Soesanto and Williams (1981) found that viscosity changes above the glass transition of a fructose-sucrose system followed the WLF temperature dependence. It may be assumed that sticky-point temperatures measured by the well-known method of Lazar and others (1956) correspond to an isoviscosity state above glass transition and a constant T – Tg (Slade and Levine 1991). Roos and Karel (1990, 1991a, 1991b) used differential scanning calorimetry (DSC) in determining, as a function of water content, the Tg of the fructose-sucrose mixture. They found that sticky points were at a T – Tg of ∼20°C and that the sticky-point temperature corresponded to the endset of the glass transition measured by DSC. The WLF equation with the universal constants C1 = –17.44 and C2 = 51.6 (Williams and others 1955) predicted an isoviscosity state of 107 Pa · s at ∼20°C above Tg (Figure 25.3). This viscosity agreed with the experimental results and predicted critical viscosity values for stickiness found by Downton and others (1982).

Crystallization, Collapse, and Glass Transition in Low-Water Food Systems

345

The Tg and its dependence on water content are essential properties of powders that can be used in control of stickiness problems, especially in the production of spraydried powders. The main importance of the relationship between the sticky point and Tg is that the Tg of amorphous food powders can be used in powder manufacturing and storage as a critical parameter for free-flowing properties, as well as a stability indicator. Powders retain their free-flowing properties at temperatures lower than the Tg, but stickiness may occur around the glass transition. Caking of amorphous powders results from the change in the material from the glassy state to the less viscous liquidlike state, which allows liquid flow and formation of interparticle liquid bridges (Peleg 1977, 1983). Peleg (1983) pointed out that “humidity caking” is the most common mechanism of caking in food powders. Humidity caking is a consequence of increasing water content, water plasticization, and depression of the Tg to below ambient temperature (Figure 25.4) (e.g., see Slade and Levine 1991; Roos 1995, 2002). Collapse Collapse can be related to viscosity and glass transition of amorphous foods. Tsourouflis and others (1976) found that the decrease in viscosity resulting in flow and structure or collapse could occur with various combinations of water content and temperature. Collapse temperatures (Tc) of freeze-dried materials were high at low water contents and decreased with increasing water content. These authors suggested that collapse required a sufficient flow above a critical viscosity, which was a function of temperature and water content. The relationships between Tc and water content were confirmed in a comprehensive study of collapse in dehydrated materials by To and Flink (1978a– c). To and Flink (1978b) also found significant similarities between the collapse temperature and glass transition behavior of synthetic polymers. Roos and Karel (1991b) reported that the Tg values determined with DSC were lower than the corresponding collapse temperatures but showed a very similar decrease with increasing water plasticization. The temperature difference between collapse and Tg is a measure of the difference between the experimental time scale of the Tg observation with DSC and the collapse (flow) determination methods, including the determination of the sticky-point temperature (Roos 1995).

Crystallization Phenomena Amorphous materials, including amorphous food components, exist in a thermodynamically metastable, nonequilibrium state. Translational diffusion of molecules within amorphous glassy solids cannot take place to enable diffusion and molecular arrangement to form the highly ordered crystalline equilibrium state. Thermal or water plasticization of amorphous solids to above glass transition transforms the glassy state into the supercooled liquidlike state. Translational mobility of molecules is characteristic of the resultant liquidlike states. Additional plasticization enhances molecular mobility, which often leads to time-dependent crystallization with rates increasing with increasing T – Tg.

346

PART 1: Invited Speakers and Oral Presentations

Crystallization of Amorphous Sugars Makower and Dye (1956) studied the crystallization properties of amorphous sucrose. Their results were based on dehydration of sorbed water from amorphous sugars at constant relative humidity (RH) as a result of crystallization. Their study suggested that amorphous sucrose crystallized at an RH-dependent rate and that crystallization was a time-dependent phenomenon. For amorphous sucrose, Karel (1973) established a sorption isotherm based on the results reported by Makower and Dye (1956). He showed that sorbed water content of amorphous sugars at a constant temperature increased with increasing storage RH, but that time-dependent crystallization occurred only above a critical storage RH. The sorption isotherm suggested that amorphous sucrose was unstable at intermediate RHs as a result of instant crystallization. DSC studies of amorphous sugars show that glass transition is often followed by a crystallization exotherm. This behavior is similar to that of many synthetic polymers and suggests instant crystallization at a critical T – Tg value enabling the rapid molecular mobility required by instant molecular arrangements to form the crystalline state. According to Roos and Karel (1990), amorphous sugars crystallize above Tg. The time required for crystallization was found to depend on temperature and water content. The crystallization rates of amorphous lactose suggested that Tg was the main factor that controlled crystallization at various temperatures and water contents (Roos and Karel 1992). Crystallization at constant RH conditions enabled crystallization to occur at a constant T – Tg. The experimental time to crystallization at low T – Tg values was longer than was predicted by Arrhenius-type temperature dependence. Roos and Karel (1992) used the universal WLF constants to predict the temperature dependence of time to crystallization and showed that this time increased dramatically at low T – Tg conditions. During water sorption, lactose may crystallize either into the anhydrous α form or into β-lactose monohydrate (Figure 25.9). The crystalline form produced depends on the RH and temperature. According to Vuataz (1988), lactose crystallizes into the anhydrous β form at relatively low water activities and into α-lactose monohydrate at water activity above 0.57 at room temperature. At higher temperatures, crystallization behavior may change following the stability of the crystalline form at the crystallization temperature. Furthermore, recrystallization of anhydrous α-lactose and β-lactose to α-lactose monohydrate at intermediate water activities has been reported (Haque and Roos 2005). α = 1 − e − kt

n

ln [ − ln (1 − α )] = ln k + n ln t

(25.1) (25.2)

In Equations 25.1 and 25.2, α is crystallinity, t is time and k and n are constants. The rate of crystallization of polymers is often modeled using the Avrami equation (Equation 25.1). The rate increases as crystallization proceeds, it decreases again as crystallinity approaches unity, and follows the sigmoid relationship of Equation

Crystallization, Collapse, and Glass Transition in Low-Water Food Systems

347

50

Water content (g/100 g of solids)

Lactose 40

30

20

(and Dairy Powders) Extrapolated water sorption isotherm for non-crystalline lactose

Time-dependent crystallization at constant T α-lactose monohydrate crystals

10

0 0.0

Anhydrous Recrystallization α/β mixed crystals

0.2

0.4

0.6

0.8

1.0

Water activity

Figure 25.9. Water-sorption and time-dependent crystallization and recrystallization of amorphous lactose. Crystallization above a critical water activity or water content results from depression of the glass transition to below ambient temperature.

25.2 against time. Roos and Karel (1992) reported that the rate of lactose crystallization at various T – Tg conditions did not fully follow the Avrami equation. The kinetics of crystallization at a constant temperature above Tg can be related to water content and water activity, which define T – Tg at various levels of water plasticization (Jouppila and others 1997). Sugars exist in the amorphous state in a number of low-water and frozen foods. Crystallization of amorphous sugars during storage is often uncontrolled. Crystallization of amorphous sugars in low-water foods is detrimental to quality and may significantly decrease shelf life. Water-sorption properties and the stability of food powders are known to be affected by temperature. Foods that contain mixtures of sugars show more complex crystallization behavior. For example, milk powders, which contain lactose hydrolyzed into galactose and glucose, show no break in their sorption isotherms (Jouppila and Roos 1994a). Crystallization of sugar components in the glucosegalactose mixture is probably delayed in comparison to crystallization of sugar components in lactose. In frozen foods, sugar may crystallize as a result of freeze concentration of the solutes and supersaturation of the freeze-concentrated solute matrix. At sufficiently low temperatures, solutes may vitrify into the glassy state, and crystallization ceases. Sugar in frozen foods may crystallize above the Tg of the maximally freeze-concentrated solute matrix (Tg′ ) (Slade and Levine 1991, 1995). Lactose is one of the least soluble sugars and a typical component of frozen desserts containing milk. Crystallization of amorphous lactose in ice cream is the main cause of “sandiness”

348

PART 1: Invited Speakers and Oral Presentations

(White and Cakebread 1966). Ice recrystallization is another phenomenon that frequently reduces the quality of frozen desserts and other foods (Roos 1995). Ice Formation and Recrystallization Ice formation and recrystallization may affect the stability of food materials significantly during frozen storage. Ice formation in freeze-concentrated systems is time dependent and temperature dependent. In simple solutions and in foods, ice forms as the temperature of the material is reduced to below the initial concentration-dependent equilibrium freezing temperature. The relationships between temperature and the physical state of freeze-concentrated food materials can be determined from state diagrams (Roos and others 1996). The time-dependent ice formation is observed in most supercooled binary sugar solutions. Ice formation at low solute concentrations is not significantly delayed, but increasing the solute concentration to more than 50% (wt/wt) is often sufficient to delay and reduce ice formation during rapid cooling. Heating of such solutions in DSC may produce a low-temperature glass transition, which is followed by an exothermal event that is often referred to as devitrification (e.g., see Roos and Karel 1991c, 1991d; Roos 1995). Devitrification indicates ice formation during rewarming of rapidly cooled solutions. The devitrification temperature is often located slightly above the observed Tg. The time dependence of ice formation, which is associated with the heating of rapidly cooled solutions, can also be observed in annealing treatments. Such annealing may be accomplished by heating a rapidly cooled solution to a predetermined temperature above Tg, which is followed by isothermal holding at that temperature and recooling. Roos and Karel (1991c, 1991d) studied time-dependent ice formation in fructose, glucose, and sucrose solutions. Annealing at a temperature slightly below Tm′ showed that the observed Tg of the annealed material increased with an increase in annealing time. This suggested that plasticizing water was removed from the partially freeze-concentrated solute matrix during annealing. The increase in ice formation was also evidenced by the size of the ice-melting endotherm, the size of which increased with an increase in annealing time.

State Diagrams State diagrams, which can be used to describe the glass transition and structural properties of low-water food systems (Figure 25.10), include information on water plasticization and frozen-state properties, as well as the saturation of dissolved components (Roos and others 1996). State diagrams are effective tools in establishing relationships between the physical state of food materials, temperature, and water content. State diagrams show the Tg as a function of water content and the effect of ice formation on Tg and on the equilibrium ice-melting temperature (Tm). State diagrams may also show solubility as a function of temperature and provide information on various changes that may occur due to the metastable state of amorphous food solids and their approach toward the equilibrium state. In food formulation and design, state diagrams enable the evaluation of the effects of food composition and water content on the

Crystallization, Collapse, and Glass Transition in Low-Water Food Systems

100

Solubility (equilibrium mixture

Glass transition range

of α- and β-lactose)

Supercooled

Tg

liquid

0

-50

T'm

Equilibrium freezing zone

T'g

Temperature range for maximum ice formation

Gla ss

Temperature (°C)

50

Tg

-100

Glass

-150 0.0

349

0.2

0.4 0.6 Weight fraction of lactose

C'g 0.8

1.0

Figure 25.10. State diagram of amorphous lactose. Glass transition at various water contents is shown by the glass transition temperature (Tg) range. Maximally freezeconcentrated solutions show the onset of glass transition at Tg′ and onset of ice melting at Tm′ . C g′ , lactose weight fraction at Tg′ ; Tg′ , glass transition temperature at maximally freeze-concentrated solution; and Tm′ , melting temperature at maximally freezeconcentrated solution.

physical state and physicochemical properties during processing and storage. State diagrams may be used in food formulation as “maps” that show the physical state at various water contents and temperatures (Roos and others 1996).

Conclusions Food solids in low-water systems form amorphous structures with time-dependent properties. Foods with high sugar content often show clear glass transitions and changes in flow properties as temperature and water content exceed values, resulting in the glass transition. Decreasing viscosity and flow above the glass transition causes stickiness and caking of particles, collapse in dehydration and of dehydrated structures, and crystallization of amorphous food components. These changes are time dependent, and their rates increase at increasing temperature and water content. In future studies, the relaxation times of structural changes in amorphous food systems should be determined and related to their dielectric, mechanical, and thermal properties. It is also important to manipulate food composition to improve the stability of low-water food systems for improved protection from deterioration resulting from collapse phenomena and crystallization of amorphous components.

350

PART 1: Invited Speakers and Oral Presentations

References Downton GE, Flores-Luna JL, King CJ. 1982. Mechanism of stickiness in hygroscopic, amorphous powders. Ind Eng Chem Fundam 21:447–51. Haque MK, Roos YH. 2005. Crystallization and X-ray diffraction of spray-dried and freeze-dried amorphous lactose. Carbohydr Res 340:293–301. Hartel RW. 2001. Crystallization in foods. Gaithersburg, MD: Aspen. Jouppila K, Kansikas J, Roos YH. 1997. Glass transition, water plasticization, and lactose crystallization in skim milk powder. J Dairy Sci 80:3152–60. Jouppila K, Roos YH. 1994a. Water sorption and time-dependent phenomena of milk powders. J Dairy Sci 77:1798–808. Jouppila K, Roos YH. 1994b. Glass transitions and crystallization in milk powders. J Dairy Sci 77:2907–15. Kalichevsky MT, Blanshard JMV. 1993. The effect of fructose and water on the glass transition of amylopectin. Carbohydr Polym 20:107–13. Karel M. 1973. Recent research and development in the field of low-moisture and intermediate-moisture foods. CRC Crit Rev Food Technol 3:329–73. Karel M, Anglea S, Buera P, Karmas R, Levi G, Roos Y. 1994. Stability-related transitions of amorphous foods. Thermochim Acta 246:249–69. Lazar ME, Brown AH, Smith GS, Wong FF, Linquist FE. 1956. Experimental production of tomato powder by spray drying. Food Technol 10:129–34. Levine H, Slade L. 1988. “Collapse” phenomena: a unifying concept for interpreting the behaviour of low moisture foods. In: Blanshard JMV, Mitchell JR, editors. Food structure: its creation and evaluation. London: Butterworth’s. p 149–80. Makower B, Dye WB. 1956. Equilibrium moisture content and crystallization of amorphous sucrose and glucose. J Agric Food Chem 4:72–7. Pehkonen KS, Roos YH, Song M, Ross RP, Stanton C. 2008. State transitions and physicochemical aspects of cryoprotection and stabilization in freeze-drying of Lactobacillus rhamnosus GG (LGG). J Appl Microbiol 104(6):1732–43. Peleg M. 1977. Flowability of food powders and methods for its evaluation. J Food Proc Eng 1:303–28. Peleg M. 1983. Physical characteristics of food powders. In: Peleg M, Bagley EB, editors. Physical properties of foods. Westport, CT: AVI. p 293–323. Peleg M. 1992. On the use of the WLF model in polymers and foods. Crit Rev Food Sci Nutr 32:59–66. Peleg M. 1993. Mapping the stiffness-temperature-moisture relationship of solid biomaterials at and around their glass transition. Rheol Acta 32:575–80. Roos Y. 1993. Melting and glass transitions of low molecular weight carbohydrates. Carbohydr Res 238:39–48. Roos YH. 1995. Phase transitions in foods. San Diego: Academic. Roos YH. 2002. Importance of glass transition and water activity to spray drying and stability of dairy powders. Lait 82:475–84. Roos YH. 2008. The glassy state. In: Lillford PJ, Aguilera JM, editors. Food materials science. New York: Springer. p 67–81. Roos YH, Karel M. 1990. Differential scanning calorimetry study of phase transitions affecting the quality of dehydrated materials. Biotechnol Prog 6:159–63. Roos YH, Karel M. 1991a. Plasticizing effect of water on thermal behavior and crystallization of amorphous food models. J Food Sci 56:38–43. Roos YH, Karel M. 1991b. Phase transitions of mixtures of amorphous polysaccharides and sugars. Biotechnol Prog 7:49–53. Roos YH, Karel M. 1991c. Amorphous state and delayed ice formation in sucrose solutions. Int J Food Sci Technol 26:553–66. Roos YH, Karel M. 1991d. Nonequilibrium ice formation in carbohydrate solutions. Cryo Lett 12:367–76.

Crystallization, Collapse, and Glass Transition in Low-Water Food Systems

351

Roos YH, Karel M. 1992. Crystallization of amorphous lactose. J Food Sci 57:775–7. Roos YH, Karel M, Kokini JL. 1996. Glass transitions in low moisture and frozen foods: effects on shelf life and quality. Food Technol 50:95–108. Slade L, Levine H. 1991. Beyond water activity: recent advances based on an alternative approach to the assessment of food quality and safety. Crit Rev Food Sci Nutr 30:115–360. Slade L, Levine H. 1995. Glass transitions and water-food structure interactions. Adv Food Nutr Res 38:103–269. Soesanto T, Williams MC. 1981. Volumetric interpretation of viscosity for concentrated and dilute sugar solutions. J Phys Chem 85:3338–41. Sperling LH. 1992. Introduction to physical polymer science. 2nd ed. New York: John Wiley & Sons. Tant MR, Wilkes GL. 1981. An overview of the nonequilibrium behavior of polymer glasses. Polym Eng Sci 21:874–95. To ET, Flink JM. 1978a. “Collapse,” a structural transition in freeze dried carbohydrates. I. Evaluation of analytical methods. J Food Technol 13:551–65. To ET, Flink JM. 1978b. “Collapse,” a structural transition in freeze dried carbohydrates. II. Effect of solute composition. J Food Technol 13:567–81. To ET, Flink JM. 1978c. “Collapse,” a structural transition in freeze dried carbohydrates. III. Prerequisite of recrystallization. J Food Technol 13:583–94. Tsourouflis S, Flink M, Karel M. 1976. Loss of structure in freeze-dried carbohydrates solutions: effect of temperature, moisture content and composition. J Sci Food Agric 27:509–19. Vuataz G. 1988. Preservation of skim-milk powders: role of water activity and temperature in lactose crystallization and lysine loss. In: Seow CC, editor. Food preservation by water activity control. Amsterdam: Elsevier. p 73–101. White GW, Cakebread SH. 1966. The glassy state in certain sugar-containing food products. J Food Technol 1:73–82. Williams ML, Landel RF, Ferry JD. 1955. The temperature dependence of relaxation mechanisms in amorphous polymers and other glass-forming liquids. J Am Chem Soc 77:3701–7. Wunderlich B. 1981. The basis of thermal analysis. In: Turi EA, editor. Thermal characterization of polymeric materials. New York: Academic. p 91–234.

26 Carbohydrates in Amorphous States: Molecular Packing, Nanostructure, and Interaction with Water J. Ubbink

Abstract The last few decades have witnessed a considerable impact of the study of phase transitions in carbohydrate-based systems on the prediction and optimization of product stability and shelf life in low-moisture foods and pharmaceutics. Such studies have in particular been aimed at the analysis of the glass transition and the role of water as a plasticizer. There is a growing awareness, however, that a number of issues cannot be resolved satisfactorily without an understanding of the molecular physics of carbohydrate matrices. This includes the properties of carbohydrate glasses, the phase behavior and molecular structure of carbohydrate-water systems, and the understanding of the role of carbohydrates as bioprotectants. In this chapter, an overview is provided of recent advances in the molecular physics of carbohydrate matrices, including our own work on the elucidation of the carbohydrate nanostructure by positron annihilation lifetime spectroscopy (PALS) and molecular dynamics (MD) simulations. The implications of these findings for the development of encapsulation systems and bioprotectants are discussed.

Glassy Carbohydrates in Food and Pharmaceutical Stability During the last few decades, significant progress has been made in the understanding of the physical properties of carbohydrate-water systems in terms of their temperatureand water-dependent phase transitions (Figure 26.1). This has resulted in the widespread use of phase and state diagrams to predict the stability of systems based on carbohydrate-based foods (Roos 1995) and pharmaceutics (Roberts and Debenedetti 2002). Of particular importance in this context is the glass transition temperature (Tg) of amorphous carbohydrates, its depression by water, and relation to the rheology of the carbohydrate system (Slade and Levine 1995). The concept of the glass transition has been used to such effect that nowadays the glassy state of a food or pharmaceutical product is assumed to be almost synonymous with stability (Table 26.1). This is obviously correct when viscosity-dependent processes are considered (Angell 1995). For example, a powder consisting of a matrix of amorphous carbohydrates, such as milk powder (Vuataz 2002), or a glass-encapsulation system (Ubbink and Schoonman 2007), may be said to be physically stable below its Tg; large-scale molecular rearrangements and macroscopic flow of the matrix are absent, and the powder will keep its physical state virtually indefinitely. In addition, 353

354

PART 1: Invited Speakers and Oral Presentations

Glassy state

Glass transition

Rubbery Supersaturated Solubility Dilute limit solution state solution

Crystalline state Increasing water content

Figure 26.1. Outline of phase transitions and physical states of carbohydrate-water systems in the nonfrozen state at constant (e.g., ambient) temperature. The glass transition, here indicated to occur at a distinct water content of the carbohydrate-water system, will occur at higher temperatures for lower water contents and vice versa. The carbohydrates that have propensity to do so will crystallize between the glass transition and the solubility limit of the carbohydrate, but with kinetics that depend strongly on the viscosity of the matrix and thus may vary by many orders of magnitude. Complex mixtures of carbohydrates often do not crystallize or crystallize only with significant delay.

Table 26.1. Dual role of glassy carbohydrates in the stabilization and delivery of active ingredients Encapsulation

Biostabilization

Principle

Macroscopic Entrapment by formation of vitreous barrier

Molecular Stabilization by hydrogen bonding between vitreous scaffold and active ingredient “Water replacement” hypothesis

Bioactive ingredient

Flavors Drugs PUFAs Various nutrients

Drugs Vesicles and membranes Proteins and peptides Nucleic acids

Technologies

Spray drying Melt extrusion Vacuum drying Freeze drying

Mainly freeze drying In nature, exploited by microorganisms (yeasts) and plants

PUFAs, polyunsaturated fatty acids.

many of the systems will have an increased chemical stability because the rate of chemical reactions comprising bimolecular reaction steps of higher molecular weight species (such as the carbohydrate molecules and proteins in the Maillard reaction [Eichner and Karel 1972]) will be much lower below Tg. Several dynamic processes, however, are known to proceed with appreciable rates below the Tg. These processes include the diffusion of water (Tromp and others 1997) and other small molecules such as gases (Schoonman and others 2002), and local

Carbohydrates in Amorphous States

355

rearrangements of the carbohydrate molecules resulting in glassy-state aging of the matrix (Noel and others 2005). In addition, in the glassy state, variations are observed in the protective effects of various carbohydrates for complex and sensitive biological systems such as membrane and lipid systems (Crowe and others 1997) and proteins (Cicerone and Soles 2004). Increasingly often, it is observed that important aspects of the stability of such systems do not correlate with the matrix Tg. Alternative mechanisms have been proposed to explain the stabilizing effects of carbohydrates below the Tg. These include first and foremost the so-called waterreplacement hypothesis (Crowe and others 1992), which states that those carbohydrates that can replace the hydrogen bonding between water and lipid membranes will be particularly efficacious in stabilizing encapsulated lipid complexes (Table 26.1). Although partially corroborated by various experimental studies using Fourier transform infrared spectroscopy (FTIR), which indeed show a positive correlation between increases in the interaction strength between carbohydrates and the lipid complexes, and the stability of the encapsulated complexes (Crowe and others 1997), the hypothesis does not fully explain why certain carbohydrates show these increased interactions, nor does it provide detailed physical insight into the properties of the carbohydrate matrix itself. Apart from FTIR, several other techniques have proven to be useful in probing the properties of carbohydrates in the glassy state. These include dielectric spectroscopy (Noel and others 2000), dilatometry (Benczédi and others 1998), nuclear magnetic resonance (NMR) (Derbyshire and others 2004), electron spin resonance (ESR) (van den Dries and others 2000), and neutron scattering (Cicerone and Soles 2004). These techniques cover α and β relaxations (dielectric spectroscopy); matrix and solute dynamics (NMR, ESR, and neutron scattering); and some structural aspects (NMR and neutron scattering). One important aspect of glassy carbohydrates that has not yet been covered in significant detail in the scientific literature is their local structure at the molecular level. In accordance with the extensive literature that exists on, for example, synthetic polymers, I expect this local structure to significantly influence the properties in the glassy state. Apart from dilatometry (Benczédi and others 1998), which provides quantitative information on the specific volume of the matrix, and from which some structural conclusions can be drawn provided a statistical mechanical model is assumed, none of these techniques provides significant insight into the structure of the carbohydrate matrices, in particular at the molecular level and in the glassy state. For this reason, we have introduced positron annihilation lifetime spectroscopy (PALS) to explore the molecular structure of amorphous carbohydrates in the glassy state (Kilburn and others 2004). In this chapter, I review our recent investigations of the glassy-state structure using PALS and show that useful insight into the physics of glassy carbohydrate matrices is obtained from this technique. Our PALS studies are then related with the experimental results reported by other groups, and with molecular dynamics (MD) simulations performed both by Limbach and Ubbink (2008) and by others (Roberts and

356

PART 1: Invited Speakers and Oral Presentations

Debenedetti 1999; Molinero and Goddard 2005), to arrive at an interpretation of the structure, interactions, and dynamics of carbohydrate-water systems in the glassy state and just above the Tg. In addition, I present an extended framework that will facilitate a rational use of glassy carbohydrates in food and pharmaceutical applications. In this framework, I make a distinction between processes influencing the product stability that depend on the viscosity of the system, and those that do not. For the latter class of processes, the concept of molecular packing (Kilburn and others 2005; Townrow and others 2007) is demonstrated to be useful to interpret and predict stability-related properties such as the modulation of moisture sorption in the glassy state (Ubbink and others 2007), the stability of encapsulated oxidation-sensitive compounds (Anandaraman and Reineccius 1986), and the stabilization of biological systems and surfactant complexes in glassy carbohydrate matrices (Crowe and others 1997).

Effects of Water on the Structure of Carbohydrate Glasses In Figure 26.2a, the hole size of thermally annealed maltodextrin DE12 matrices prepared by solvent casting is shown as a function of the temperature for various water contents. These water contents are prepared by equilibrating the matrices at various water activities at T = 25°C. The hole size as probed by PALS is shown to expand linearly with increasing temperature. Two regimes can be discerned: a glassy state at low temperatures and a rubbery state at high temperatures. The temperature at which the crossover between the two linear branches of the hole volume as a function of temperature occurs is identified as the Tg (Figure 26.2b). It is of interest to note the very good agreement between the Tg determination from the thermal expansion of the molecular holes and by differential scanning calorimetry (DSC), because PALS is essentially a static technique and DSC provides a dynamic measure of the Tg. An important additional observation from Figure 26.1a is that, at constant temperature, the hole size increases with increasing water content. This is the case in both the glassy and the rubbery states. This expansion of the average hole size with increasing water content is most likely a general feature of amorphous carbohydrates, because it is observed both for maltodextrins (Figure 26.1a), which are highly polydisperse and branched, and for simple disaccharides such as trehalose (Figure 26.3a). In a previous report (Kilburn and others 2005), we interpreted the effect of water on the expansion of the hole size in glassy carbohydrate matrices as resulting from the interference of water with the hydrogen bonding between the carbohydrate molecules (Figure 26.4a). As long as no direct experimental evidence on water content– dependent changes in the hydrogen bonding pattern in glassy carbohydrates is available, this remains a tentative explanation. In a subsequent report (Townrow and others 2007), we again addressed this issue by considering not only the average hole size as determined by PALS, but also the width of the hole size distribution. In combination with a number of distinct models for the effect of water on the average hole size and the width of the hole size distribution in glassy carbohydrate matrices (Figure 26.4), we have strengthened our conclusion that the principal mechanism by which water acts on the structure and dynamics of glassy carbohydrate matrices is by influencing

Carbohydrates in Amorphous States

357

(a) 90 80

3

vh [Å ]

70 60 50 40 30 0

20

40

60

80

100

120

140

160

T [°C ] (b)

Tg from PALS [°C]

100 80 60 40 20 0 0

20

40 nd

Tg(onset) at 2

60

80

100

heating from DSC [°C]

Figure 26.2. (a) The hole volume of annealed maltodextrin matrices (DE12) as a function of the temperature for various water activities (aw) equilibrated at 25°C: , aw = 0.11; •, aw = 0.22; , aw = 0.33; , aw = 0.54; , aw = 0.54; and , aw = 0.75. (b) Comparison of the glass transition temperatures from positron annihilation lifetime spectroscopy (PALS) and differential scanning calorimetry (DSC) for maltodextrin DE12 matrices at various water concentrations: dashed line, slope = 1; and solid line, linear regression. Parts a and b reproduced from Kilburn and others (2005) with permission from the American Chemical Society. Vh, hole volume of annealed maltodextrin matrices.

the hydrogen bonding between the carbohydrate molecules. Further proof using additional techniques, such as FTIR and NMR, will be needed to resolve this issue fully. When considering the structural information provided by PALS, it should be realized that this technique probes only the holes between the carbohydrate and water molecules making up the matrix. It thus provides only a coarse impression of the

2.05

r [Å]

2.00

1.95

1.90

1.85 0.00

0.02

0.04

0.06

0.08

Qw

Figure 26.3. Mean radius of holes in amorphous trehalose as a function of the water content at 25°C. The solid line is the linear regression of the experimental data (R2 = 0.87). Reproduced from Kilburn and others (2006). r[Å], mean radius in Angstroms; Qw, water content; and circles, actual data points.

(a)

water uptake

vh

vh

(b)

water uptake

radius

radius

(c)

water uptake

Figure 26.4. Suggested mechanisms leading to carbohydrate hole sizes increasing with water content in the glassy state. (a) Hole expansion because of interference of water with the hydrogen bonding between the carbohydrate molecules in the glassy state. (b) Preferential filling of small holes in a glassy matrix with a polydisperse hole size distribution. (c) The increase in water content leads to an effective two-phase system with distinct hole size distributions, with the water holes being larger than the holes in the carbohydrate system. Vh, average hole volume of amorphous carbohydrate matrices.

358

Carbohydrates in Amorphous States

(a)

359

(b)

Figure 26.5. (a) Snapshot from a molecular dynamics (MD) simulation of a carbohydrate-water system (80 wt% maltotriose in water) (T = 25°C). All molecules (carbohydrates and water) are shown in a non–space-filling stick-and-ball representation; one water molecule is shown with its actual size as given by the van der Waals radii of the atoms. (b) Free-volume holes in an anhydrous carbohydrate matrix as determined from MD simulations. The matrix consists of maltopentose, and the temperature is 310 K. Part b reproduced from Limbach and Ubbink (2008), with permission from the Royal Society of Chemistry.

structure of the system, in some sense similar to the information contained in a silhouette determined from the shadow cast by an object placed in a light beam. Full insight into the structure of a carbohydrate-water system can consequently be obtained only by combining PALS with other techniques, such as FTIR, NMR, and electron paramagnetic resonance (EPR); and scattering techniques such as neutron and X-ray scattering. The complexity of the structure of an amorphous carbohydrate-water system may be gauged from snapshot images plotted from the positions of atoms in MD simulations (Figure 26.5a). From such MD simulations, the actual holes in a carbohydratewater system can be derived by filling the spaces between the atoms (whose sizes are defined by their van der Waals radii) with virtual spheres that are smaller than the smallest detail of the molecular holes one aims to determine (Limbach and Ubbink 2008) (Figure 26.5b). Such representations of the holes between the molecules enables one to assess the complexity of the structures for which only simplified information is obtained from PALS. In particular, from a first inspection of Figure 26.5b, it is obvious that the shape of the molecular holes is far from spherical, as is conventionally assumed in PALS, and, furthermore, the hole size distribution is far from monodisperse, as is also often assumed.

Molecular Packing in Glassy Carbohydrates Apart from water content and temperature, we may anticipate that the carbohydrate composition influences the structure of the carbohydrate matrix. In Figure 26.6a, the

(a)

90 80

vh [Å]

70 60 50 40 30 20 0

20

40

60

80

100

120

140

160

100

120

140

160

T [°C] (b) 0.680

0.670

3

-1

Vsp [cm ·g ]

0.675

0.665 0.660 0.655 0.650 0.645 0

20

40

60

80

T [°C] (c) 0.675

Various molecular weights

0.670

aw = 0.22

0.660

3

-1

Vsp [cm ·g ]

0.665

0.655 0.650

aw = 0.54

0.645 0.640 0.635 30

40

50

60

vh [Å3]

360

70

80

Carbohydrates in Amorphous States

361

average hole size from PALS is plotted as a function of temperature for maltodextrins of various degrees of hydrolysis (Kilburn and others 2005). It is immediately observed that, in the glassy state, significant differences in hole size occur, whereas, in the rubbery state, no significant differences can be detected. In the glassy state, the hole size decreases with increasing degree of maltodextrin hydrolysis and decreases thus with decreasing molecular weight. In effect, by lowering the (average) molecular weight, the local packing of the carbohydrates becomes more dense. It is of interest to note that the dependence of the specific volume on temperature and carbohydrate composition is very similar to the behavior of the hole volume (Figure 26.6b), with the specific volume decreasing with decreasing molecular weight in the glassy state, whereas it is independent of the carbohydrate molecular weight in the rubbery state. These similarities extend even so far that a direct correspondence between the hole volume (a molecular property) and the specific volume (a macroscopic property) is observed (Kilburn and others 2005; Townrow and others 2007) (Figure 26.6c). MD simulations may be used to verify the observations from the PALS experiments. For instance, when a series of carbohydrate systems of varying degree of polymerization are cooled down in computer simulations, the free volume of the matrix is seen to be independent of the degree of polymerization in the rubbery state, whereas the free volume decreases with decreasing degree of polymerization in the glassy state, in line with the observations from PALS (Figure 26.7). It should be noted here that MD simulations unfortunately cannot be performed on time scales that enable a precise determination of a Tg (Limbach and Ubbink 2008). In a recent PALS study, the effects of variations in carbohydrate composition and water content have been investigated systematically for blends of a fairly monodisperse maltopolymer (Mw/Mn = 2.2, where Mw and Mn are the weight and number average molecular weight, respectively) and the disaccharide maltose (Townrow and others 2007). When the molecular hole size is plotted as a function of the water content for a series of such maltopolymer-maltose blends, a rather complex picture of the structural behavior of the carbohydrate matrices is obtained (Figure 26.8a). Two of the previous observations on maltodextrins are confirmed: (a) in the rubbery state, the hole size tends to be independent of the carbohydrate composition, and (b) the  Figure 26.6. Hole volume (a) and specific volume (b) of annealed maltodextrin matrices equilibrated at water activity (aw) = 0.22 (at T = 25°C) as a function of the temperature for various molecular weight distributions: , DE6; •, DE12; and , DE33. (c) The specific volume vs hole volume of annealed maltodextrin matrices as a function of the temperature for various molecular weight distributions. Open symbols, aw = 0.22 (at T = 25°C); solid symbols, aw = 0.54; circles, DE6; downward-pointing triangles, DE12; diamonds, DE21; upward-pointing triangles, DE33; Vh, average hole volume of amorphous carbohydrate matrices; and Vsp, specific volume of annealed maltodextrin matrices. Parts a and b reproduced from Kilburn and others (2005), with permission from the American Chemical Society.

362

PART 1: Invited Speakers and Oral Presentations

18

3

Free volume fraction x10 [-]

20

16 14

glucose (DP-1) maltose (DP-2) maltotriose (DP-3) maltotetraose (DP-4) maltopentose (DP-5)

12 10 8 6 4 2 250

300

350

400

450

500

T [K] Figure 26.7. Temperature dependence of the free-volume fraction of maltooligomerwater mixtures at a carbohydrate concentration of 90 wt%. Each data point is the average of 20 independent quench runs. DP, degree of polymerization (e.g., DP-2 means two polymer molecules). The [-] indicates that the physical quantity plotted on the y-axis is dimensionless. Reproduced from Limbach and Ubbink (2008), with permission from the Royal Society of Chemistry.

hole size in the glassy state decreases with decreases in carbohydrate molecular weight (see also Figure 26.8b). Several additional observations can be made. First, at very low water content, hole filling is observed not only for low maltose but also for water (Townrow and others 2007) (Figure 26.8a). This hole-filling mechanism can possibly be related to the mechanism of antiplasticization observed at very low water contents (Benczédi and others 1998). Secondly, in the glassy state at nonzero water content, the hole size of a matrix consisting of carbohydrate polymers decreases rapidly with the first ∼30% of low molecular weight carbohydrates that are added (Figure 26.8b). The disparate effect of water and low molecular weight sugars as plasticizers for carbohydrate polymers strongly suggests that water acts as a plasticizer mainly by interacting with the carbohydrate polymers and interfering with their hydrogen bonding. Low molecular weight sugars, conversely, do not modify the interactions between the carbohydrate polymers as much, but, due to their small size, they reduce the average number of entanglements experienced by a polymer chain and thus enable molecular reorganizations under conditions where the polymer chains themselves would already be frozen into a glassy state. The differences in glassy-state properties observed for the various molecular weight distributions relating to the temperature dependence of both hole volume and specific volume are thus essentially determined by value of the Tg for the specific combination of molecular weight distribution and water content. As I discuss next, this has important technological implications.

(a) 60 55 50

3

vh [Å ]

45 40 35 30 25 20 15 0.00

0.04

0.08

0.12

0.16

Qw (b) 50 45

3

vh [Å ]

40 35 30 25 20 0

20

40

60

80

100

Maltose content on total carbohydrate [wt.%]

Figure 26.8. (a) Hole volume as a function of the weight fraction of water for a range of maltopolymer-maltose matrices equilibrated at water activities between 0 and 0.75 (T = 25°C). The various maltopolymer-maltose blends are indicated by the following symbols: , 100% maltopolymer; •, 95% maltopolymer–5% maltose; , 90% maltopolymer–10% maltose; , 80% maltopolymer–20% maltose; , 60% maltopolymer–40% maltose; , 30% maltopolymer–70% maltose; and , 100% maltose. (b) Hole size at 25°C as a function of the weight fraction of maltose on a total carbohydrate basis for the anhydrous matrices () and for the matrices at a water content of 5 wt% (•) (T = 25°C). The line is a fit of a simple overlap model as proposed by Townrow and others (2007). The maltopolymer used for the experiments reported in the figures is a chromatographically fractioned maltodextrin (Mw/Mm = 2.2). Reproduced from Townrow and others (2007), with permission from the American Chemical Society.

363

364

PART 1: Invited Speakers and Oral Presentations

Dynamic Properties Close to the Glass Transition It is important to investigate whether the structural variations as discussed in the foregoing sections affect the dynamics of the carbohydrate-water systems. Several techniques, such as NMR and EPR; and scattering techniques, such as neutron and X-ray scattering, can in principle be used to obtain quantitative information on such structure-property relations. Unfortunately, whereas NMR is perfectly feasible in aqueous solutions of carbohydrates, and even in the rubbery state, it becomes very difficult to obtain structural and dynamic information from NMR in the glassy state. Difficulties are also encountered in the application of neutron and X-ray scattering. A major issue with neutron scattering on carbohydrate matrices is the large incoherent scattering length of hydrogen, and from X-ray scattering only limited information may be obtained on the amorphous state of carbohydrates, because generally only a broad amorphous peak is observed. ESR and neutron scattering, however, have delivered some important information on the dynamics of glassy carbohydrates (van den Dries and others 2000; Cicerone and Soles 2004). MD simulations have been used with increasing success to predict the structure and dynamics of concentrated carbohydrate-water systems (Roberts and Debenedetti 1999; Molinero and Goddard 2005; Limbach and Ubbink 2008). MD simulations are also extremely useful for investigating structure and dynamics of complex food systems at the molecular level, because they enable molecular structure, physical properties, and interactions to be incorporated in a highly detailed manner unattainable through analytic statistical mechanics (Limbach and Kremer 2006). Such MD simulations can either be based on all-atom models or can use coarse-graining procedures to reduce the level of molecular detail and thus the computer power required (Limbach and Kremer 2006). Recent advances in computing performance, force fields, and simulation methods now enable the simulation of systems of a complexity that is sufficiently high to enable direct application for quite complex carbohydrate-water systems. Detailed mechanisms for the diffusion of water in such systems and for the hydrogen bonding between the carbohydrate and water molecules have been derived from such simulations (Roberts and Debenedetti 1999; Molinero and Goddard 2005; Limbach and Ubbink 2008). In Figure 26.9, the Arrhenius plot of the diffusion coefficient of water in maltooligomers is shown for three carbohydrate concentrations. Note that, in solution and in the rubbery state, the diffusion coefficient of water is independent of the degree of polymerization of the carbohydrate and solely is determined by the carbohydrate concentration. In Figure 26.10, the ratio τ gluc τ H2 O of the rotational correlation times of water (τ H2 O) and glucose (τgluc) are shown for different carbohydrate concentrations. For low concentrations, the ratio τ gluc τ H2 O is relatively constant when the temperature is changed. The situation changes when the glucose concentration is increased. Above ∼50 wt% glucose, the ratio τ gluc τ H2 O rather suddenly increases by a factor of 3–4 in a temperature window of ∼100 K. With increasing concentration and lower temperatures, the errors in the ratio τ gluc τ H2 O become larger because the error in determining τgluc is increasing. Qualitatively similar behavior is observed for the ratio of the

Figure 26.9. Arrhenius plot for the temperature dependence of the water diffusion coefficient (D) for different carbohydrate concentrations and different molecular weights. From top to bottom: 20, 50, and 70 wt% maltooligomer content. , glucose; , maltose; , maltotriose; , maltotetraose; and , maltopentose. 1e, exponential (for example, 1e-08 = 10−8). Reproduced from Limbach and Ubbink (2008), with permission from the Royal Society of Chemistry.

Figure 26.10. Temperature dependence of the ratio of the rotational time of glucose, τgluc and the rotational time for water, τH2O for various glucose concentrations in glucose-water mixtures. The decoupling of the mobilities at higher glucose concentrations with decreasing temperature can be seen. Reproduced from Limbach and Ubbink (2008), with permission from the Royal Society of Chemistry.

365

366

PART 1: Invited Speakers and Oral Presentations

diffusion constants of the two molecules (Limbach and Ubbink 2008), but a quantitative analysis is impossible because of the rapidly increasing uncertainty in the value of the diffusion coefficients when the glass transition approaches. This decoupling of mobilities in the approach to the glass transition, which has also been observed in experiments (van den Dries and others 2000), explains that the diffusion of water shows an Arrhenius-type behavior although the mobility of the maltooligomer molecules is almost frozen. In a recent article (Limbach and Ubbink 2008), we have also investigated whether differences in carbohydrate molecular weight result in differences in structural packing at the molecular level and, hence, in differences in water mobility. Unfortunately, significant molecular weight differences in the molecular packing of maltooligomers occur only below the Tg, and in this regime, MD simulations do not yet provide sufficiently quantitative results. As I discuss in the next section, the decoupling of the mobility of water, and most likely other small molecules such as gases, from the mobility of the carbohydrate molecules in the approach to the glass transition has important implications in the food and pharmaceutical industries. This is because the mobility of these compounds is, for example, related to the shelf-life stability of products containing such compounds or to the barrier properties of membranes.

Technological Implications The current findings on the structure of carbohydrates in the glassy state and the dependence of the structure on, in particular, composition may help to define rational strategies for optimizing the stability of products in the food and pharmaceutical fields. In these applications, the first distinction that needs to be made is between viscositycontrolled stability issues and stability issues that are related to the structure and dynamics of the carbohydrate glassy state (Table 26.2). The first category comprises stability issues such the collapse of powder structure, the undesired crystallization of the carbohydrates constituting the product matrix (such as lactose in milk powder), and to some extent, chemical reactions (namely, those between larger molecular weight species whose mobility is controlled by the mobility of the carbohydrate matrix). For this category of stability issues, the highest stability is obtained by increasing the Tg

Table 26.2. Criteria for physical product stability Compromise between density and Tg

Only Tg requirementa

Encapsulation/biostabilization

Physical stability of powders

Modulation of moisture level

Caking

Matrix expansion below Tg

Agglomeration/compaction

Protection against oxygen

(Avoidance of) crystallization

Mechanical properties

Flowability

Tg, glass transition temperature. To quantitatively predict collapse, caking, and agglomeration, the analysis of the water content–dependent glass transition is usually combined with the rheology of the system in the rubbery state.

a

Carbohydrates in Amorphous States

367

Magnitude

Matrix density Glass transition

Carbohydrate molecular weight

Figure 26.11. Diagram schematically showing the effect of carbohydrate molecular weight on the matrix density and on the glass transition temperature.

to the highest practical value by either increasing the carbohydrate molecular weight or by lowering the water content of the matrix. The second category encompasses rather diverse phenomena such as those in which the mobility of low molecular weight compounds or the local structure and dynamics of the carbohydrate matrix (in particular hydrogen bonding) play the central role and includes situations in which the moisture uptake (Ubbink and others 2007), the diffusion of oxygen, and irreversible transitions in biomolecular complexes are limiting the stability of the system. In these situations, a subtle compromise between two important factors needs to be made in order to arrive at the optimal stability of the product (Figure 26.11). On the one hand, it is required that the product matrix remains in the glassy state during the shelf life of the product. As for the viscosity-related stability issues, the compromise is best achieved by increasing the carbohydrate molecular weight and by lowering the water content. On the other hand, the molecular packing of the carbohydrate matrix, as reflected, for example, by its density or the molecular hole size, is controlling the product stability. If, for example, the barrier properties against oxygen are limiting the stability of the product (e.g., it contains an encapsulated oxidation-sensitive compound such as a polyunsaturated fatty acid [PUFA]), the oxygen uptake may be minimized by maximizing the matrix density (at constant water content) and therefore by decreasing the molecular weight of the carbohydrates constituting the encapsulation matrix (Figure 26.11). This leads to a situation where a compromise needs to be made between the Tg requirement and the matrix density. Consequently, the best strategy for maximizing the stability in cases where it is limited by processes occurring in the glassy state is to strive for the carbohydrate matrix’s lowest average molecular weight that still satisfies the requirement that the product remains in the glassy state during its shelf life; at this molecular weight the barrier properties for oxygen are the highest. Several examples from the literature illustrate our approach. First, I consider the study on the oxygen uptake by carbohydrate-encapsulated citrus oils as published by Anandaraman and Reineccius (1986). By spray drying, those authors prepared a series of maltodextrin-based capsules with a variation in the degree of hydrolysis of the

PART 1: Invited Speakers and Oral Presentations

Oxygen uptake (mL O2/g· oil) × 1000

368

9

6

3

200

600 400 Storage time (h)

Figure 26.12. Oxidation of citrus oil encapsulated in maltodextrins of varying molecular weight distribution at T = 45°C: , DE4; , DE10; , DE20; , DE25; and •, DE37. Reproduced from Anandaraman and Reineccius (1986), with permission of the Institute of Food Technologists.

maltodextrins. They observed that, during storage at T = 45°C, the rate of oxygen uptake varied by more than 1 order of magnitude (Figure 26.12). The rate of oxygen uptake decreased with increasing degree of maltodextrin hydrolysis and thus decreased with decreasing carbohydrate molecular weight. This was in agreement with our results on the carbohydrate molecular packing in the glassy state. As data on the water content of the capsules are lacking in this study, and as no density measurements were performed, further, better-controlled studies are warranted. A second example relates to the stabilization of lipid vesicles and related biological lipid membrane systems, such as red blood cells, by glassy carbohydrates. Crowe and others (1997) have observed that, by varying the molecular weight of the carbohydrate matrix, the membrane phase-transition temperature (Tm) may be influenced. Specifically, they observed that, by lowering the average molecular weight of the carbohydrate matrix by increasing the fraction of glucose in hydroxyethyl starch (HES)–glucose mixtures, Tm is significantly depressed. This results in an increased membrane fluidity and consequently an improved liposome integrity during storage, and thus in an increased retention of active ingredients encapsulated within the liposomes (in this case, a fluorescent dye) (Figure 26.13). At the same time, they observed that the interactions between the carbohydrate matrix and the phospholipids were increasingly modified with increasing glucose levels (as observed from FTIR analysis of the phosphate asymmetrical stretch vibrations). Again, this is in line with our results on

Carbohydrates in Amorphous States

369

60 40 20 0

% of transition > 20°C

% retention of CF after drying

80

Tm Diameter Retention

100

250

80 200

60 40

150

20

100

Average diameter after drying (nm)

300 100

0 50 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 Weight fraction of HES

100% HES

Fewer H-bonds

100% Glucose

Lower density Figure 26.13. Effect of the weight fraction of hydroxethyl starch (HES) on total carbohydrate (HES and glucose) and on the retention of carboxyfluorescein (CF) in liposomes, membrane phase-transition temperature, and average diameter following rehydration. Reproduced from Crowe and others (1997), with permission from Academic Press. The annotations below the figure have been added for the present chapter. H-bonds, hydrogen bonds.

the molecular packing of carbohydrates in the glassy state. By reducing the average molecular weight of the carbohydrate matrix, one increases the local density of packing of the carbohydrate molecules, thereby allowing for a higher density of hydroxyl groups in the vicinity of the phospholipid head group and thus increased levels of hydrogen bonding between the phospholipids and the carbohydrate encapsulation matrix. This in turn influences the phase behavior of the liposomes and increases their stability during storage under anhydrobiotic conditions.

Concluding Remarks In this chapter, recent developments in the understanding of the properties of amorphous carbohydrates have been discussed, with emphasis on their properties in the glassy state. The impact of the glass transition concept on food and pharmaceutical technology is discussed briefly, and it is argued that those aspects of the performance of amorphous carbohydrates that are exclusively related to the glassy state are only indirectly related to the Tg of the matrix. Our investigations using PALS reveal a molecular weight–dependent structural signature of carbohydrates in the glassy state, and a relation with structural and dynamic properties as determined from MD simulations is established. The molecular packing of carbohydrates in the glassy state is discussed as a new concept to rationalize several hitherto unexplained phenomena related to the performance of glassy carbohydrates in food and pharmaceutical applications. It turns out that, in particular, the molecular weight distribution of the carbo-

370

PART 1: Invited Speakers and Oral Presentations

hydrates influences the molecular packing and consequently the properties in the glassy state. The molecular packing of carbohydrates in the glassy state affects important applications, such as the modulation of moisture sorption, the encapsulation of oxidation-sensitive compounds, and the stabilization of biological systems and surfactant complexes.

Acknowledgments I would like to take the opportunity to thank Prof. Ashraf Alam, Dr. Duncan Kilburn, Sam Townrow, and Mina Roussenova from the University of Bristol (UK) for their stimulating collaboration. I also thank my colleagues at the Nestlé Research Center who have contributed to the experimental and theoretical program on the molecular physics of carbohydrate systems at low water content. Special thanks are due to Dr. Hans-Joerg Limbach, Maria-Isabelle Giardiello, Dr. Vincent Meunier, Jean-Pierre Marquet, and Johanna Claude.

References Anandaraman S, Reineccius GA. 1986. Stability of encapsulated orange peel oil. Food Technol 40:88–93. Angell CA. 1995. Formation of glasses from liquids and biopolymers. Science 267:1924–35. Benczédi D, Tomka I, Escher F. 1998. Thermodynamics of amorphous starch-water systems. 1. Volume fluctuations. Macromolecules 31:3055–61. Cicerone MT, Soles CL. 2004. Fast dynamics and stabilization of proteins: binary glasses of trehalose and glycerol. Biophys J 86:3836–45. Crowe JH, Hoekstra FA, Crowe LM. 1992. Anhydrobiosis. Annu Rev Physiol 54:579–99. Crowe JH, Oliver AE, Hoekstra FA, Crowe LM. 1997. Stabilization of dry membranes by mixtures of hydroxyethyl starch and glucose: the role of vitrification. Cryobiology 35:20–30. Derbyshire W, Van den Bosch M, Van Dusschoten D, MacNaughtan W, Farhat IA, Hemminga MA, Mitchell JR. 2004. Fitting of the beat pattern observed in NMR free-induction decay signals of concentrated carbohydrate-water solutions. J Magn Reson 168:278–83. Eichner K, Karel M. 1972. The influence of water content and water activity on the sugar-amino browning reaction in model systems under various conditions. J Agric Food Chem 20:218–23. Kilburn D, Claude J, Mezzenga R, Dlubek G, Alam A, Ubbink J. 2004. Water in glassy carbohydrates: opening it up at the nanolevel. J Phys Chem [B] 108:12436–41. Kilburn D, Claude J, Schweizer T, Alam A, Ubbink J. 2005. Carbohydrate polymers in amorphous states: an integrated thermodynamic and nanostructural investigation. Biomacromolecules 6:864–79. Kilburn D, Townrow S, Meunier V, Richardson R, Alam A, Ubbink J. 2006. Organization and mobility of water in amorphous and crystalline trehalose. Nat Mater 5:632–5. Limbach HJ, Kremer K. 2006. Multi-scale modelling of polymers: perspectives for food materials. Trends Food Sci Technol 17:215–9. Limbach HJ, Ubbink J. 2008. Structure and dynamics of maltooligomer-water solutions and glasses. Soft Matter 4:1887–9. Molinero V, Goddard WA. 2005. Microscopic mechanism of water diffusion in glucose glasses. Phys Rev Lett 95:045701. Noel TR, Parker R, Brownsey GJ, Farhat IA, MacNaughtan W, Ring SG. 2005. Physical aging of starch, maltodextrin, and maltose. J Agric Food Chem 53:8580–5. Noel TR, Parker R, Ring SG. 2000. Effect of molecular structure and water content on the dielectric relaxation behaviour of amorphous low molecular weight carbohydrates above and below their glass transition. Carbohydr Res 329:839–45.

Carbohydrates in Amorphous States

371

Roberts CJ, Debenedetti PG. 1999. Structure and dynamics in concentrated, amorphous carbohydrate-water systems by molecular dynamics simulation. J Phys Chem [B] 103:7308–18. Roberts CJ, Debenedetti PG. 2002. Engineering pharmaceutical stability with amorphous solids. AIChE J 48:1140–4. Roos YH. 1995. Phase transitions in foods. New York: Academic. Schoonman A, Ubbink J, Bisperink C, Le Meste M, Karel M. 2002. Solubility and diffusion of nitrogen in maltodextrin/protein tablets. Biotechnol Prog 18:139–54. Slade L, Levine H. 1995. Glass transition and water-food structure interactions. Adv Food Nutr Res 38:103–269. Townrow S, Kilburn D, Alam A, Ubbink J. 2007. Molecular packing in amorphous carbohydrate matrices. J Phys Chem [B] 111:12643–8. Tromp RH, Parker R, Ring SG. 1997. Water diffusion in glasses of carbohydrates. Carbohydr Res 303:199–205. Ubbink J, Giardiello M-I, Limbach H-J. 2007. Sorption of water by bidisperse mixtures of carbohydrates in glassy and rubbery states. Biomacromolecules 8:2862–73. Ubbink J, Schoonman A. 2007. Flavor delivery systems. In: Hui YH, et al., editors. Kirk-Othmer concise encyclopedia of chemical technology. 5th ed. New York: Wiley. p 940–4. van den Dries IJ, van Dusschoten D, Hemminga MA, van der Linden E. 2000. Effects of water content and molecular weight on spin probe and water mobility in malto-oligomer glasses. J Phys Chem [B] 104:10126–32. Vuataz G. 2002. The phase diagram of milk: a new tool for optimizing the drying process. Lait 82:485–500.

27 Ice Crystallization in Gels and Foods Manipulated by the Polymer Network N. Murase, S. Yamada, and N. Ijima

Abstract Freezing behavior of polymer gels was investigated. Studies using differential scanning calorimetry found that water in polymer gels sometimes remains partially unfrozen during cooling and freezes during rewarming. Ice crystallization during rewarming, observed with cross-linked dextran gels, depended on density of cross-links, as well as water content and cooling rate. When examined, the density of cross-links was considered to be interrelated with the flexibility of the polymer network and water diffusivity in gels. Water in the gels was found to transform into a vitreous state at the time of freezing. Ice crystals formed in the gel are probably small because the diffusion of water molecules is more or less obstructed by the polymer network. In the polymer-water systems containing a small amount of water, no ice crystallization was observed when cooled. Instead, water behaves as a plasticizer, decreasing the glass transition temperature of the polymers.

Introduction Many biological systems and foods include a gelling state under normal conditions. Thus, it is of practical importance to investigate the obstructive effect of the polymer network in gels on water-molecule mobility and ice crystallization for the implementation of cryopreservation of biological systems, food refrigeration, and freeze drying. Differential scanning calorimetry (DSC) studies conducted from this standpoint found that water in polymer gels sometimes remains partially unfrozen during cooling below −50°C and freezes during subsequent heating (rewarming). Ice crystallization during rewarming, especially observed with cross-linked dextran (MacKenzie 1977; Murase and others 1982, 1986) and cross-linked polyacrylamide gels (Murase and others 1983), depends on the density of cross-links, as well as on water content and cooling rate. The density of cross-links is interrelated with the flexibility of the polymer network and water diffusivity through the polymer network in gels (Murase and Watanabe 1992; Murase and others 1997). In this connection, a concept of compartmentalized water was introduced for water in polymer gels that differed from water in pores. Ice crystallization during rewarming is also observed with food biopolymer gels, such as those of agarose (Nishinari and others 1991) and starch (Murase and Watanabe 1992). However, the mechanism of ice crystallization during the rewarming of these polymer gels still remains unclear after intensive studies using 373

374

PART 1: Invited Speakers and Oral Presentations

various physical techniques such as DSC, electron spin resonance (ESR), nuclear magnetic resonance (NMR), simultaneous X-ray diffraction (XRD)-DSC measurements, Raman scattering (Murase and others 2002), and synchrotron radiation XRD (Murase and others 2004). The physical state of unfrozen water in polymer gels is another matter of concern in this chapter. Since the glass transition temperature (Tg) of pure water is reported to be around 136 K (Johali and others 1987), unfrozen water in polymer gels might turn into ice or a glassy state by further cooling. However, it might also be vitrified at the point of partial freezing at higher subzero temperatures (Murase and others 2002). The polymer-water systems discussed thus far contain a large amount of water, where water forms a phase separated from hydrated polymers. In polymer-water systems with a small amount of water forming a single phase, ice crystallization does not occur during cooling, but water in the systems is known to behave as a plasticizer that decreases the Tg of polymers. In this case, the glass transition associated with the vitrified water, assuming that it is present in the polymer gels at subzero temperatures, should be differentiated from that of the hydrated polymers observed at higher temperatures. Based on experimental results, the freezing behavior of polymer gels and the glass transition of the systems are discussed with a focus on ice crystallization during rewarming. Ice crystallization during rewarming in food biopolymer gels is also considered.

The Ice Crystallization Exotherm During Rewarming Observed with Cross-Linked Dextran Gels The materials used in this study were cross-linked dextrans (Sephadex; GE Healthcare UK, Giles, England). Sephadex, used as a matrix of gel chromatography, is in a bead form that is ∼100 μm in diameter. The various Sephadexes differ in the density of their cross-links. The order of cross-link densities is G10 > G15 > G25 > G100, and their pore sizes expressed in molecular exclusion limit when swollen with water is ∼700, ∼1500, ∼5000, and ∼10,000 Da, respectively. Typical DSC rewarming traces obtained with Sephadex gels are shown in Figure 27.1. When a Sephadex G25 gel containing ∼50 wt% water is cooled, a freezing exotherm appears at around −18°C. During rewarming, the heating trace begins to drift in an endothermic direction at around −20°C, followed by an exotherm resulting from ice crystallization at around −12°C and a large endotherm resulting from ice melting. Anomalous freezing behavior during rewarming is observed only with a G25 gel among Sephadex gels of different cross-link densities. The amount of unfrozen water in a G25 gel containing 50 wt% water, estimated to be ∼37% of total water at maximum, is more than that in both G15 and G100 gels, where the amount of unfrozen water is ∼29%. By comparing the excess amount of unfrozen water with that in both G15 and G100 gels, ice crystallizes when the exotherm occurs as a G25 gel rewarms (Murase and others 1986). The result suggests that this amount of unfrozen water was vitrified during cooling, and that the anomalous freezing behavior during rewarm-

Ice Crystallization in Gels and Foods Manipulated by the Polymer Network

375

Heat Flow Endo Up/mW

80

60

40

20

0 -50

-40

-30

-20

-10

0

10

Temperature/°C Figure 27.1. Differential scanning calorimetry rewarming traces obtained with Sephadex G25 gels. Cooling and heating rates, 5°C min−1; cooling temperature, −50°C; and sample weight, ∼10 mg. Water content (%) is wet basis.

ing is a characteristic of water compartmentalized by the polymer network, the concept of which is different from water in pores. In the case of a G25 gel, the exotherm during the rewarming becomes remarkable with the increase in water content until 50– 55 wt%, as shown in Figure 27.1, but a further increase causes the exotherm to decrease. Although the exotherm during rewarming disappears above a water content of 70 wt% when cooled slowly, it is still observed when the sample is quenched in liquid nitrogen. To observe the exotherm during rewarming, it is sufficient to cool the system below the temperature at which the freezing exotherm is completed in the cooling step prior to the heating scan (i.e., at around −22°C). Moreover, the extent of the exotherm observed with a G25 gel containing 50 wt% water scarcely depends on the rate of the prior cooling within the range of 0.5° to ∼103°C min−1. DSC rewarming traces dependent on the prior freezing condition are shown in Figure 27.2. As mentioned, when cooled from ambient temperature, where nonequilibrium freezing occurs in the gels after substantial undercooling, an ice crystallization exotherm is observed during rewarming of a G25 gel. However, the exotherm disappears after equilibrium freezing, where the gel freezes without any additional ice being capable of formation due to seeding at lower temperatures. Besides disappearance of the exotherm, the endothermic trend in the DSC rewarming trace shifts to higher temperatures after equilibrium freezing (Murase 1993; Murase and others 1997). The extent of undercooling in the gel was reduced by the addition of ice nucleation protein (Snomax; York Snow, Centennial, CO, USA) and controlled by the amount of the protein. When this addition raises the freezing temperature higher than around −15°C, the exotherm disappears in the subsequent rewarming trace. Therefore, appearance of the exotherm has a kinetic origin. Ice crystallization during rewarming is a

G – 10

0.00

–5.00

Endo.

Heat Flow

mW

–10.00 –40.00

–20.00 0.00 Temp [°C]

G – 25

mW 5.00 Heat Flow

20.00

0.00 –5.00

–10.00 –40.00

–20.00 0.00 Temp [°C]

20.00 G – 100

mW

Heat Flow

0.00

–5.00

–40.00

–20.00 0.00 Temp [°C]

20.00

Figure 27.2. Differential scanning calorimetry rewarming traces obtained with Sephadex gels dependent on the prior freezing. Samples contained 6 mg of Sephadex and 6 mg of water. Cooling and heating rates, 5°C min−1; and cooling temperature, −50°C. Solid curves represent the traces obtained during rewarming after nonequilibrium freezing; the dotted curves represent the traces obtained during rewarming after equilibrium freezing in which the first rewarming was interrupted before complete melting, followed by a second cooling. mW, milliwatts. See the text and Murase and others (1997).

376

Ice Crystallization in Gels and Foods Manipulated by the Polymer Network

377

characteristic not only of a cross-linked dextran gel but also of other cross-linked polymer gels, such as polyacrylamide gels (e.g., Bio-Gel; BioRad, Hercules, CA, USA). Among the Bio-Gels, the exotherm during rewarming is observed with the P6 gel, which is known to have almost the same pore size as Sephadex G25. From these results, factors influencing ice crystallization of compartmentalized water in polymer gels are summarized. Compartment size must be appropriate for the ice to crystallize during rewarming. The continuity of water between adjacent compartments and the flexibility of polymer chains also must be appropriate. Note that these three factors are interrelated via the density of their cross-links. Thus, the density of cross-links must be appropriate for the ice to crystallize during rewarming.

Origin of an Endothermic Trend Observed Prior to the Exotherm During Rewarming The endothermic trend can be considered to be due to the melting of small ice crystals, the melting temperature of which is depressed by the Kelvin effect. This theory is supported by the trend becoming larger after equilibrium freezing, during which ice crystals can grow. However, the trend might be related to glass transition and/or enthalpy relaxation. Assuming that the endothermic trend is a result of the glass transition, the heat capacity (ΔCp) was tentatively calculated from the DSC rewarming trace. The value of ∼100 J/(g °C) for ΔCp is too large for the glass transition compared with that of glucose or pure water because the values of 0.52–0.65 J/(g °C) and <0.1 J/(g °C) are reported for glucose (Finegold and others 1989; Roos and Karel 1991) and pure water (Johari and others 1987), respectively. The annealing treatment was conducted just below the initiation temperature of the endothermic trend for 2 h. The expectation is that the trend would become more remarkable if enthalpy relaxation is involved with the event. But the trend did not become remarkable. Thus, measurements using oscillation DSC were tried. Irrespective of the oscillation condition (i.e., amplitude and frequency of the oscillation), rewarming traces scarcely changed, except for the oscillation behavior. These results suggest that the endothermic trend is not of a kinetic origin. Assuming the endothermic trend is a result of the melting of small ice crystals, the size of the ice crystals was estimated by using Laplace-Young and Gibbs-Duhem equations: ΔTf =

2Vm Tf0 σ il cos θ ΔHf r

(27.1)

where ΔTf is the depression of the freezing point (= melting point); Vm is the molecular volume of ice; Tf0 is the freezing point of bulk water; σil is the ice-water interfacial energy (∼30 × 10−7 J cm−2); θ is the contact angle (assumed to be zero); ΔHf is the enthalpy of fusion of ice; and r is pore radius. Depression of melting temperature (ΔTf) is proportional to the reciprocal of ice crystal radius (r), and the dimension of nanometer was estimated for the melting temperature of −12°C. However, neither the presence of such small ice crystals nor their traces was

378

PART 1: Invited Speakers and Oral Presentations

indicated by the scanning electron microscopic observation of freeze-dried samples. Instead, the ice crystals observed were around a micrometer in dimension. Contradiction between this result and Equation 27.1 might have come, mainly, from the wrong application of σil measured for the ice-water interface (Kerr and others 1985). Instead, the interfacial energy between water hydrating to polymer and ice should be used. Thus, the value for the energy might be much larger, and the existence of larger ice crystals (i.e., micrometer size) can be explained. To verify the assumption that the endothermic trend is due to ice melting, simultaneous XRD-DSC measurements were performed. Five diffraction peaks for 2θ (θ is the diffraction angle) were observed between 20° and 40° with the frozen G25 gels. The peaks are ascribed to those of hexagonal ice, but the existence of cubic ice cannot be excluded. Diffraction intensities increased coincidentally with the appearance of the exotherm during rewarming in the DSC trace, which confirmed that the exotherm is a result of ice crystallization. However, a decrease in the intensities within the temperature range of the endothermic trend, which had been expected assuming the trend is due to the ice melting, was not confirmed because of the fluctuation of intensity. The result indicates that some mode of water-molecule motion had already been initiated in the gels. To obtain more precise information, twodimensional XRD-DSC measurements were carried out by using synchrotron radiation. Three-ring images at −40°C are diffractions produced by hexagonal ice corresponding to those obtained by one-dimensional XRD measurements (Figure 27.3). Continuous, but dim ring images observed with a frozen G25 gel are characteristic of diffraction by powder crystals whose axes are randomly oriented. The size of ice crystals, estimated by the diffraction pattern, was about 10 μm or smaller. Ring images became gradually discontinuous, with the temperature rise starting at around −24°C, and spotty images appeared. The spots indicate either the growth of ice crystals at

G 10

G 25

G 100

Figure 27.3. Two-dimensional X-ray diffraction images obtained with frozen Sephadex gels containing 50% water. Temperature, −50°C. Reproduced from Murase and others (2004), with permission.

Ice Crystallization in Gels and Foods Manipulated by the Polymer Network

379

certain places in the sample or recrystallization. Moreover, the result indicates that the ice crystals melt during the endothermic trend, which exceeds the exothermic heat of ice crystallization and/or recrystallization. At the initiation temperature of the exotherm during rewarming, the number of spots suddenly increased corresponding to the crystallization of larger ice. Two-dimensional XRD images of frozen gels were different, depending on the density of their cross-links. Many spots were observed on rings of a frozen G10 gel. Spotty images, indicating the formation of large ice crystals, can be explained by the rigidity of the polymer network and resultant crystal growth. It means that the network structure with a higher-density of cross-links changes little during freezing. Pairs of arcs facing each other were seen with a frozen G100 gel. The result indicates the preferential growth of ice crystals in a certain direction so that lengthy ice crystals are formed.

Vitrified Water in a Frozen G25 Gel The temperature dependence of ESR spectra was investigated using tempone, a nitroxide radical, as a spin probe (Murase and others 1986). ESR-derivative spectra produced three sharp lines due to the radical in liquid water down to −18°C. Freezing the motion of the radical starts below this temperature during cooling, and a broad signal due to immobilized radicals began to appear. At about −50°C, the signal from the mobile radicals in liquid water disappeared. During rewarming, the signal from the mobile radicals, once appearing, decreased simultaneously with appearance of the exotherm in the DSC trace. These results indicate that unfrozen water below −50°C is already vitrified. Raman spectroscopic studies were conducted with Sephadex gels. The spectrum around 3400 cm−1 is from the OH stretching band and can be used as an indicator of the strength of hydrogen bond. The increase in Raman intensity in the region of wave numbers lower than 3400 cm−1 when a G100 gel froze indicates that the hydrogenbond network was strengthened by ice crystallization. With a G25 gel, however, the intensity remained low when partially frozen at −24°C, compared with a frozen G100 gel, and was similar to that obtained at ambient temperature. It increased at −11°C when ice crystallized during rewarming. These results are evidence of the presence of vitrified water in a frozen G25 gel.

Freezing Scheme for Compartmentalized Water in Polymer Gels Taking the aforementioned results into consideration, we now discuss the freezing scheme of water in a G25 gel. Assuming that ice crystals are formed in a particular compartment, ice crystal growth proceeds either by inoculating other ice in an adjacent compartment or by dehydrating water from the adjacent compartment. When cooled from ambient temperature, ice crystals grow quickly after substantial undercooling because the rate of ice crystal growth depends on the extent of undercooling. Fast freezing induces an instantaneous change in the polymer network, and the diffusion

380

PART 1: Invited Speakers and Oral Presentations

Table 27.1. A scheme for ice crystallization of compartmentalized water in polymer gels Nonequilibrium freezing

Equilibrium freezing

↓

↓

Instantaneous change in the polymer network structure

Gradual change in the polymer network structure

↓

↓

G25

Vitrified water

→

Disappearance of vitrified water

Small ice crystals

→

Growth of ice crystals or recrystallization

G100

Lengthy ice crystals grown in a certain direction

→

Growth of ice crystals

G10

Small ice crystals Large ice crystals

→

Hard to grow because the network structure is hard to change

of water molecules between adjacent compartments is prevented. Then, some of the compartmentalized water in a G25 gel is trapped by the polymer network and either remains unfrozen to eventually vitrify, or small ice crystals form. During equilibrium freezing on the other hand, where cooling is in the presence of ice or an ice nucleator, ice crystals grow slowly without undercooling, and the polymer network changes more gradually compared with the case of nonequilibrium freezing. Then, most of the compartmentalized water is barely trapped by the polymer network at the time freezing is initiated and forms larger ice crystals, which explains the shift of the endothermic trend toward a higher temperature observed through DSC, as well as the disappearance of the exotherm during rewarming. The freezing scheme is summarized in Table 27.1. In the case of gels with a lower density of cross-links, such as a G100 gel, ice crystal growth in a certain direction is not significantly prevented because water between adjacent compartments is highly continuous, and lengthy crystals can be formed even under nonequilibrium freezing. In the case of gels with a higher density of cross-links, such as a G10 gel, small ice crystals may form in small compartments. As the polymer network is difficult to change at the time of freezing, however, ice crystal growth between adjacent compartments is possible, and large ice crystals can also be formed. This is confirmed by the results of two-dimensional XRD experiments. The appearance of freeze-dried Sephadex beads detected through environmental scanning electron microscopy (ESEM) clearly showed their flexibility, dependent on the density of their cross-links (Ruike and others 1999b). Flexibility of the polymer network was also indicated by the preparation of water-sorption isotherms plotted against relative humidity. Water-desorption isotherms lay on the upper side than sorption isotherms, and the extent of hysteresis was more remarkable with Sephadex gels of lower density of cross-links. The result suggests that absorbed water is readily trapped by the polymer network when the relative humidity is lowered (Ruike and others 1999a).

Ice Crystallization in Gels and Foods Manipulated by the Polymer Network

(b)

250

8

200 6 4

Tg/°C

Heat-Flow Endo. Up/mW endo

(a)

381

21% 18%

2

150 100 50 0

0 –10 0 10 20 30 40 50 60 70 80 Temperature (°C)

–50

0

0.1

0.2 Wc/g · g

0.3

–1

Figure 27.4. (a) Differential scanning calorimetry rewarming traces obtained with Sephadex G25 gels containing a small amount of water. Endo, endotherm; and mW, milliwatts. (b) Glass transition temperature (Tg) of the system dependent on water content.

Glass Transition of Cross-Linked Dextrans Containing a Small Amount of Water There can be two types of glass transitions in polymer gels which should be clearly differentiated: one is associated with water and the other is associated with polymer matrices. The glass transition observed with a G25 gel containing a small amount of water, first reported by Cojazzi and Pizzoli (1999), was confirmed and is shown in Figure 27.4. Tg decreases with increasing water content. Therefore, water behaves as a plasticizer for the cross-linked dextran. Above the water content of ∼0.34 (g water/g dry matter), two phases of water and the hydrated polymer became separated. Data are then analyzed by means of the Gordon-Taylor equation. The result confirmed that the Sephadex-water system forms a single phase within the water content of 0.34, and that the hydrated polymer indicates a glass transition plasticized by water. Tg decreased to lower than 0°C above the water content of ∼0.25. When the system with the water content of ∼0.34 is cooled, the hydrated polymer might turn first into a glassy state, followed by the glass transition of water at subzero temperatures after partial freezing. However, two glass transitions within the same system have not yet been identified. The glass transition behavior of the system around the water content where phase separation occurs still remains unclear.

Ice Crystallization During Rewarming Observed with Various Food Biopolymer Gels Some food biopolymer gels, such as those of agarose, gelatin, or starch, display ice crystallization during rewarming although the exotherm is not so remarkable

382

PART 1: Invited Speakers and Oral Presentations

20 min

Endo

50 min

80 min 110 min 140 min 170 min

-30

-10 10 Temperature (°C)

30

Figure 27.5. Differential scanning calorimetry rewarming traces obtained with gelatinized potato-starch gels. Gels were quenched by liquid nitrogen, followed by heating. Water content, 49.5 wt%; sample weight, 20 mg; temperature of gelatinization, 120°C, and heating rate, 5°C min−1. Retention time at ambient temperature before quenching by liquid nitrogen is indicated in the figure. Endo, endotherm.

compared with that shown by a Sephadex G25 gel. DSC rewarming traces obtained with a potato-starch gel after being quenched in liquid nitrogen are shown in Figure 27.5. The extent of the exotherm depended on the retention time at ambient temperature after gelatinization. The exotherm initially increased with time but decreased after a while. Both water content and gelatinization temperature also influence the extent of the exotherm and its time dependency. During the retention time at ambient temperature, the polymers once dissociated by heating become entangled to form cross-links. Considering the case of a Sephadex G25 gel, it seems that gelatinized starch gels display an exotherm due to ice crystallization during rewarming when the density of cross-links becomes sufficiently high in the process of retrogradation. The fact that the exotherm seen in the potato-starch gel was not as remarkable as that of a Sephadex G25 gel means that the preparation of the gel with a certain density of cross-links is difficult. Apart from the difficulty, it should be pointed out that the exotherm during rewarming can be used as an indicator of the rheological properties of food biopolymer gels.

Ice Crystallization in Gels and Foods Manipulated by the Polymer Network

383

Conclusions Molecular diffusion of water in polymer gels is more or less obstructed by the polymer network. As a result, ice crystals of various size and dimension are formed dependent on the characteristics of the polymer network and especially on the density of its cross-links. With gels of a certain density of cross-links, water is partially vitrified when frozen; trapped by the polymer network. Two types of glass transition are observed in gels: one is associated with vitrified water and the other is associated with polymer matrices plasticized by water. The presence of the network structure in biopolymer gels should be considered for successfully cryopreserving biological systems and refrigerating or freeze drying foods.

References Cojazzi G, Pizzoli M. 1999. Thermal behavior of water in crosslinked dextran. Macromol Chem Phys 200:2356–64. Finegold L, Franks F, Hatley RHM. 1989. Glass/rubber transitions and heat capacities of binary sugar blends. J Chem Soc Faraday Trans I 85:2945–51. Johari GP, Hallbrucker A, Mayer E, Hofer K. 1987. The glass-liquid transition of hyperquenched water. Nature 330:552–3. Kerr KL, Feeney RE, Osuga DT, Reid DS. 1985. Interfacial energies between ice and solutions of antifreeze glycoproteins. Cryo Lett 6:371–8. MacKenzie AP. 1977. Non-equilibrium freezing behaviour of aqueous systems. Philos Trans R Soc Lond [B] 278:167–89. Murase N. 1993. Origin of an endothermic trend observed prior to ice crystallization exotherm in the DSC rewarming trace for polymer gels. Cryo Lett 14:365–74. Murase N, Abe S, Takahashi H, Katagiri C, Kikegawa T. 2004. Two-dimensional diffraction study of ice crystallization in polymer gels. Cryo Lett 25:227–34. Murase N, Fujita T, Gonda K. 1983. Low-temperature calorimetric studies of compartmentalized water in hydrogel systems (II). Cryo Lett 4:19–22. Murase N, Gonda K, Watanabe T. 1986. Unfrozen compartmentalized water in gels and its anomalous crystallization during warming. J Phys Chem 90:5420–6. Murase N, Inoue T, Ruike M. 1997. Equilibrium and nonequilibrium freezing of water in crosslinked dextran gels. Cryo Lett 18:157–64. Murase N, Ruike M, Yoshioka S, Katagiri C, Takahashi H. 2002. Glass transition and ice crystallisation of water in polymer gels, studied by oscillation DSC, XRD-DSC simultaneous measurements, and Raman spectroscopy. In: Levine H, editor. Amorphous food and pharmaceutical systems. Cambridge: Royal Society of Chemistry. p 339–46. Murase N, Shiraishi M, Koga S, Gonda K. 1982. Low-temperature calorimetric studies of compartmentalized water in hydrogel systems (I). Cryo Lett 3:251–4. Murase N, Watanabe T. 1992. Ice crystallization during rewarming of polymer gels. In: Maeno N, Hondoh T, editors. Physics and chemistry of ice. Sapporo: Hokkaido University Press. p 249–53. Nishinari K, Watase M, Williams Peter A. 1991. Effect of sugars and polyols on water in agarose gels. In: Levine H, Slade L, editors. Water relationships in food. New York: Plenum. p 235–49. Roos YH, Karel M. 1991. Non-equilibrium ice formation in carbohydrate solutions. Cryo Lett 12:367–76. Ruike M, Inoue T, Takada T, Horie K, Murase N. 1999a. Water sorption and drying behavior of crosslinked dextrans. Biosci Biotechnol Biochem 63:271–5. Ruike M, Takada S, Murase N, Watanabe T. 1999b. Changes in the bead structure of crosslinked polymer gels during drying and freezing. Cryo Lett 20:61–8.

28 Marine-Inspired Water-Structured Biomaterials A.-M. Hermansson, P. Olofsson, S. Ekstedt, M. Pihl, and P. Gatenholm

Abstract Biomimetics can provide new insights for the design of the next generation of biomaterials based on hierarchical structures. In an inspirational project for research on structured materials with tailored mass transport properties, we have used the jellyfish as a model system because of its remarkable water-holding capacity. The microstructure of the jellyfish, Aurelia aurita, has been characterized by confocal laser scanning microscopy (CLSM) and by transmission electron microscopy (TEM). The characterization was focused on the mesoglea in order to understand the water management of jellyfish and related rheological properties. The results revealed a structure spanning a wide range of length scales. The mesoglea consists of long fibers and a network spanning the space between them. It is a relatively coarse and open network consisting of proteins, as well as polysaccharides, which suggests a complex supermolecular proteoglucan structure. Jellyfish are sensitive to variations in salt content, and change in volume, shrinking with increasing salt content. The hierarchical structural levels are necessary for jellyfish to maintain their overall structure in water with varying salt contents. The coarse fibers stabilize the jellyfish structure, and the network structure between the fibers is most likely responsible for water management. By understanding the hierarchy and structures at different length scales, we can gain important information for the development of a new generation of hierarchically structured biomaterials with tailor-made mass transport, water management, and molecular mobilities.

Introduction Biomimetics can provide insights for the design of new biomaterials with tailored water-holding properties. Within a new Center of Excellence on Supramolecular Biomaterials: Structure Dynamics and Properties, we are establishing a 9-year program for research on structured heterogeneous biomaterials with tailored mass transport properties and related functionalities. Structures over a range of length scales and heterogeneity will be designed to administer mass transport by a combination of diffusion, capillary, and hydrodynamic flow with applications for the hygiene, pharmaceutical, cellulose, and food sectors. In an inspirational project for research on superabsorbents, we have studied jellyfish because of their remarkable water-holding properties. Superabsorbents are key components in modern hygiene products (e.g., diapers, feminine products, and 385

386

PART 1: Invited Speakers and Oral Presentations

incontinence products), where they help to maintain high liquid absorption capacity. Development is needed to increase absorption capacity even further and to understand different types of absorption kinetics, especially under pressure, in salt-rich environments, and at varying pH. The most commonly used superabsorbent today is cross-linked polyacrylic acid because it is efficient and inexpensive. Superabsorbent polymers (SAPs) based on acrylic monomers are used in about 90% of the 19 × 109 disposable baby diapers produced annually in the United States alone. However, the main drawbacks of using acrylate-based superabsorbents are that they are produced from nonrenewable sources and are not biodegradable, thus preventing their disposal through composting. Therefore, research is focused on developing a new generation of “green” superabsorbents. Jellyfish are well known for holding a large amount of salt water in relation to the amount of their bodily biological material. There is hardly any information about the mechanism they use, their microstructure, or chemical composition. In this report, we present new findings about the microstructure and the water-holding properties of jellyfish to illustrate length scales that are important for their water-holding properties in different ionic environments.

Jellyfish (Aurelia aurita) Jellyfish contain approximately 99% salt water and only 1% organic material. The salt water contains approximately 2%–3% salt, which means that jellyfish consist of around 96%–97% pure water. Structural analysis has shown higher concentrations of organic material on the jellyfish surface that seems to cover jellyfish quite well, serving as a “skin.” Consequently, not much material can be responsible for water retention. The nuclear magnetic resonance (NMR) spectra in Figure 28.1 provide a general view of the chemical composition of dialyzed and freeze-dried jellyfish before

Water Manet, dialysed and lyophilized HR MAS NMR, 600 MHz T2 filter 5000 Hz

Aromatic amino acids

8.0 7.5

Fatty acid

Fatty acids Amino acids Amino acids Oligosaccharides

Fatty acids

7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5

1.0

ppm

Figure 28.1. High-resolution magic angle spinning nuclear magnetic resonance images of dialyzed and freeze-dried jellyfish.

Marine-Inspired Water-Structured Biomaterials

387

medusa

planula larva ephyra budding polyp polyp

Figure 28.2. Jellyfish life cycle. Redrawn from Whitaker and others.

hydrolysis and fractionation. The peak at 1 ppm can also originate from aliphatic amino acids.

Life Cycle and Accessibility The life cycle of jellyfish (A. aurita) is 1 year (Figure 28.2). In the late summer, the parent jellyfish drops small larvae that attach to the sea floor and become polyps. From each polyp, small ephyras will bud off that are laciniated with eight “wings” each. The ephyras are only 2–3 mm in diameter when they leave the polyp in the early spring, around February–March in Sweden. Around May, the first jellyfish, only a few centimeters in diameter, are visible along the Swedish coast. During June–August, fully grown jellyfish can be found in the ocean, and these are the type used in our experiments. Mainly male, smaller jellyfish were caught; the capture of the normally larger females carrying eggs was avoided.

Water-Holding and Textural Properties of Jellyfish Superabsorbents (SAPs) can absorb enormous amounts of pure water. However, as illustrated in Figure 28.3, the water uptake is drastically reduced by the addition of salt corresponding to physiological conditions in humans. Thus, there is a need to develop new biomaterials in which the water uptake can be efficient also in the presence of salts and other constituents in body fluids. Data from Hirst and Lucas (1998) show that the water holding of jellyfish does not change with the salinity of water to the same extent as does the water-holding ability of SAP. This schematic example clearly illustrates why jellyfish can serve as an inspiration for the development of new superabsorbents for the hygiene industry. This does not mean that we want to develop a superabsorbent with the same structure as jellyfish, but that we want insight into and understanding of the length scales of the structural constituents responsible for jellyfish texture and water-holding properties of structured materials.

388

PART 1: Invited Speakers and Oral Presentations

Figure 28.3. Water retention by superabsorbent polymer (SAP) and jellyfish (grams salt water/grams dry material) as a function of salt concentration.

Figure 28.4. A sample was taken from the mesoglea for a puncture test where stress and strain were recorded in an Instron universal testing machine (Instron, Norwood, MA, USA).

To see if and how jellyfish change their water content and texture depending on changes in the environment, water baths with different salt types and concentrations were prepared. Initial experiments showed that the jellyfish stopped moving when salinity changed suddenly. Salt concentration was therefore changed gradually over 2 days. The size, water content, and texture of living jellyfish was determined. One issue was to clarify whether jellyfish water-holding properties depend on the living organism or are determined by its structure. In addition to intact jellyfish, pieces of the mesoglea were cut off and placed directly in saline water prior to texture measurements (Figure 28.4). The mesoglea is the transparent “structure-less” part of the jellyfish. Similar results were obtained for pieces of mesoglea and intact jellyfish, which shows that the structure is responsible for the textural properties. The effects of added sodium chloride (NaCl), potassium chloride (KCl), magnesium chloride (MgCl2), and calcium chloride (CaCl2) to the sea water were evaluated at 0.04, 0.08, 0.12, and 0.16 M chloride. The additions of NaCl and MgCl2 at the lower con-

0.60

control + 0.04 M NaCl + 0.08 M NaCl + 0.12 M NaCl + 0.16 M NaCl

0.50 0.40 0.30 0.20 0.10 0.00

Water content in jellyfish [%]

Maximum load [N]

Marine-Inspired Water-Structured Biomaterials

389

98.00 97.80 97.60

MgCl2 NaCl

97.40 97.20 97.00 96.80 96.60 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 Added salt concentration in M [mol Cl–/Liter]

Figure 28.5 Left: The maximum load of a 2.4-mm probe penetrating the jellyfish at a speed of 0.5 mm/s at increasing additions of sodium chloride (NaCl). Right: The water content of jellyfish as a function of added salt concentration. MgCl2, magnesium chloride.

centrations did not affect the jellyfish to a great extent. At the highest levels of NaCl, the jellyfish shrank but were still quite active. At higher concentrations of MgCl2, the organisms survived but were gradually deformed with time. At the two higher concentrations of CaCl2, none survived. Most surprisingly, the jellyfish in KCl were completely paralyzed after a few hours of salt addition. However, if a specimen was taken from a 0.04 M KCl bath and placed back into the control with only sea water, it started to move again. This phenomenon was not observed at higher concentrations of KCl. The results from the puncture test showed that the jellyfish modulus increased with increasing addition of salt to the surrounding water. This is most likely due to a decrease in water content, as illustrated in Figure 28.5. The water content decreased with salt concentration independently of the type of salt. In the case of NaCl, the decrease was linear, whereas the decrease in MgCl2 was nonlinear and the jellyfish contained more water at higher salt concentrations. One possibility is that the jellyfish could not adapt their water-regulation system at higher concentrations of MgCl2. The results clearly show that the jellyfish shrink and grow firmer with increasing salt concentrations. This can be a mechanism by which the organism regulates its density, preventing sinking or floating to the surface, depending on variations in environmental conditions. However, it is not possible to conclude exactly how the jellyfish regulates its water content.

Jellyfish Microstructure The structure of A. aurita was characterized to obtain an understanding of the microstructure and how the different constituents relate to water-holding properties. The clear, jellylike body the mesoglea was given the greatest attention, since it is composed of the most water-holding structures. The Figure 28.6 images of different parts of the jellyfish were taken with a confocal laser scanning microscope (CLSM). Both proteins and polysaccharides were labeled, but only polysaccharide-labeled images are shown.

390

PART 1: Invited Speakers and Oral Presentations

400 mm

200 mm

40 mm

80 mm

Figure 28.6. The grainy surface of the jellyfish (top right) and the transport canals (top left). Bottom: The filamentous structures in the mesoglea with interdispersed cells, close to the surface (right).

The epidermis, the outer skin of the jellyfish, has a grainy surface, with the grains consisting of groups of cellular structures that are visible even without the aid of the microscope. These groups of cells consist mostly of polysaccharides; little protein can be found in this structure. By the naked eye, one can see canals running from the center out to the periphery. Half of the canals transport nutrition in, while the other half transports waste products out. Specific labeling revealed that the canals are composed of proteins and polysaccharides. Large pores, which are up to about 50 μm in diameter, are visible in the structure of the canals. The radial canals have an opening in the center, which is easily visualized when the canals are cross-sectioned. Moving into mesoglea of the jellyfish, there is a coarse network of fibers, with interspersed cells. Just below the surface, the fibers are arranged more or less parallel to the upper surface, as shown in Figure 28.6. Further down, the fibers seem to be arranged more randomly. Specific labeling showed that the fibers are composed of polysaccharides and proteins. These fibers could be a kind of skeleton that is important

Marine-Inspired Water-Structured Biomaterials

391

8 mm

20 mm

Figure 28.7. Confocal laser scanning microscopic (CLSM) images of cells and fibers in the mesoglea.

Sik

Sik

1000 nm

Sik

1000 nm

Figure 28.8. A longitudinal cut (left) and a cross section (right) of a fiber in the mesoglea. The arrow indicates the area that was enlarged in the top square.

for the overall structure and shape of the jellyfish. The cells are seen as bright spots in the structure. The cells seem to be associated with the fibers in the mesoglea. One hypothesis is that cells interspersed in the mesoglea produce the fibers. As shown in Figure 28.7, some cells have a string attached, like spider-web strands. The fibers seen in CLSM can also be visualized at high magnification in transmission electron microscopy (TEM). What seems to be one fiber in CLSM turns out to be bundles of thin threads. The crosscuts and the longitudinally cut sections in Figure 28.8 show the thinner threads in the fiber. The most interesting issue is to determine the structure between the fibers in the mesoglea, because this structure is responsible for the water-holding properties of

392

PART 1: Invited Speakers and Oral Presentations

1000 nm

1000 nm

8 mm 100 nm Proteins

100 nm Mesoglea

Polysaccharides

Figure 28.9. Transmission electron microscopic (TEM) images and a confocal laser scanning microscopic (CLSM) image demonstrating that the area in the middle box of the mesoglea consists of an open network structure composed of polysaccharides and proteins.

jellyfish. Nothing can be seen from the CLSM images. For these length scales, TEM provides important information. Figure 28.9 shows CLSM and TEM images revealing details of the fine structure of the mesoglea. The empty spaces seen in CLSM contain a network structure composed of proteins and polysaccharides. It is very open, with pore sizes in the micrometer range. In addition to microscopic analysis at different length scales, a fractionation scheme has been developed and applied to differentiate between the polysaccharide and the protein fractions by degrading the constituents in the last fraction. An extensive NMR study combined with chemical analysis provided information about the chemical structure of polysaccharides in this very complex structure. The polysaccharides present in A. aurita were mainly composed of galactose, N-acetylgalactosamine and N-acetylglucosamine. Chemical analysis showed that jellyfish contain sulfated polysaccharides. We have also studied the different cation contents (Mg2+ and Ca2+) of fresh mesoglea, seawater, and dialyzed jellyfish, to obtain insights about jellyfish structure. We found that Mg2+ but also Ca2+ are strongly bonded by the biopolymeric structure. We subjected fresh and dialyzed jellyfish, as well as alginate gel, to EDTA, a substance that chelates metal cations. All samples were rapidly dissolved by adding EDTA; the effect was less distinct, however, for the dialyzed sample because of lower magnesium and calcium ion content. We suggest that magnesium and calcium induces crosslinking of polysaccharides and proteins in the mesoglea. The hypothesis is that the structure responsible for the water-holding capacity of jellyfish is a highly negatively charged proteoglycan. This makes sense because

Marine-Inspired Water-Structured Biomaterials

393

proteoglycans are common biopolymers in other tissues, such as cartilage. These proteoglycans often consist of a protein backbone decorated with long, repeating, linear polysaccharides composed of repeating units of uronic acids and amino sugars, many of which are sulfated. This creates a highly negatively charged, giant macromolecule.

Gel Structures, Hierarchy, and Mass Transport The results show that the mesoglea is composed of a transparent proteoglucan network and fibrous structure of completely different length scale. Transparent gels have nanoscale networks that require electron microscopy for their visualization. Even, so there are large variations in the length scales, as illustrated in Figure 28.10. The SAP gel has a very fine network structure with pores that are ∼15 nm in diameter. This is very different from the open network structure in the mesoglea. The images shown in Figure 28.10 are freeze-etched replicas, whereas the images in Figure 28.8 are thin sections (thickness, ∼80 nm), which is why the structures have a different appearance (Hermansson 2007). Both techniques show that the network is very open, with voids in the micrometer range. In addition to the capillary forces holding the water, osmotic forces can be involved because of the negative charges of the biopolymer. An interesting question is whether the proteoglucan gel structure alone or whether the hierarchy of fibrils and proteoglucans of the mesoglea determine the characteristic texture and water-holding properties of the mesoglea. We can regard the structure as a gel network reinforced by the filamentous structure. We can also regard the structure as composed of a highly viscoelastic solution or weak gel with a few calcium crosslinks that keep the filaments apart (Figure 28.11). They then act as the jellyfish skeleton. Both approaches can give rise to the characteristic texture and water-holding properties of the mesoglea.

100 nm SAP H2O

200 nm Gelatin

2000 nm Jellyfish

Figure 28.10. Transmission electron microscopic (TEM) images of a freeze-etched replica of a superabsorbent polymer (SAP) superabsorbent gel (left), a gelatin gel (middle), and a jellyfish network (right).

394

PART 1: Invited Speakers and Oral Presentations

Gel

or

Viscoelastic solution

Figure 28.11. Schematic illustration of possible structural roles of the proteoglucans in the mesoglea: one where they themselves form a gel structure reinforced by the fibers, and one where they form a viscoelastic solution or weak gel that, together with the fiber network, gives rise to the gel characteristics of the mesoglea.

The mass transport in a hierarchical structure is dependent on the length scale. The typical length scales determine water uptake and release, water holding, diffusion, and molecular mobility of soluble compounds (Hermansson and others 2006). In many hierarchical structured products, water management includes several mass transport mechanisms such as hydrodynamic flow, capillary flow, and molecular self-diffusion, depending on the length scale. Hydrodynamic flow, which is active in large and open structures, is driven by external forces such as gravity, or by differences in the chemical potential; that is, differences in concentrations at different locations in the structure. Capillary flow also depends on surface tension and occurs in channels and pores on shorter length scales than in hydrodynamic flow. A capillary gel structure will hold water, and external pressures equivalent to the capillary pressure are needed to remove the water. If we take the jellyfish structure as an example, we can assume that hydrodynamic flow dominates in the coarse transport canals shown in Figure 28.5, whereas capillary flow and diffusion prevail in the mesoglea. Figure 28.12 illustrates a simple simulation consisting of three connected compartments. The top one (section I) represents hydrodynamic flow in the transport canals, where the structure has limited influence on the flow. Section II illustrates capillary flow at smaller length scales, where surface tension and the diameter determine the mass transport; and section III illustrates diffusion at the smaller length scales. For these three situations, we have scaled the diffusion coefficients to 45, 15, and 1, respectively. Molecules enter the left side of Section I, and the compartments increase in brightness with mass transport. The mass transport is very fast in section I, with hydrodynamic flow relative to the equivalent to the capillary flow in section II and the slow diffusion in section III. All three mass transport mechanisms are crucial for all living hierarchically built structures like jellyfish, trees, and humans. There are numerous ways to design new biomaterials with a variety of mass transport and water-binding properties based on

Marine-Inspired Water-Structured Biomaterials

395

1 (a)

(b)

(c)

0.5

0 (d)

(e)

(f)

Figure 28.12. Schematic illustration of mass transport by hydrodynamic flow (I), capillary flow (II), and diffusion (III) at time units (a) 25, (b) 200, (c) 500, (d) 1000, (e) 2500, and (f) 10,000. Relative rate: I = 45, II = 15, and III = 1.

combinations of length scales. In a sophisticated product like a diaper, a wound-care product, or a drug with controlled-release properties, different parts of the structure can be designed for different functions, which requires water management over different length scales, as well as time scales. The water may need to be quickly transported, stored, and released predictably. Here, biomimetics can play an important role as an inspiration for the vast possibilities of creating new functions based on hierarchy and heterogeneity.

References Hermansson AM. 2007. Structuring water by gelation. In: Aguilera J, Lillford PJ, editors. Food materials science: principles and practice. Food Engineering Series. New York: Springer. p 255–80. Hermansson AM, Lorén N, Nydén M. 2006. The importance of microstructure for water diffusion on micro and nano length scales. In: Buera P, Welti-Chanes J, Lillford P, Corti H, editors. Water properties of food, pharmaceuticals and biological materials. Boca Raton, FL: CRC Taylor and Francis. p 79–100. Hirst AG, Lucas CH. 1998. Salinity influences body weight quantification in the scyphomedusa Aurelia aurita: important implications for body weight determination in gelatinous zooplankton. Mar Ecol Prog Ser 165:259–69. Whitaker DJ, King R, Knott D. Jellyfish. Columbia: Sea Science, South Carolina Department of Natural Resources. Available from: http://www.dnr.sc.gov/marine/pub/seascience/pdf/Jellyfish.pdf.

PART 2 Poster Presentations

Session 5 Role of Water Mobility/Dynamics in Food and Pharmaceutical Systems

29 Another Unusual Property of Water: It Increases the Glass Transition Temperature of a Glassy Polymer S. P. Chamarthy and R. Pinal

Abstract We report on the effect of water on the glass transition temperature (Tg) of three amorphous systems: an acrylic-based polymer (Eudragit E100), a pharmaceutically relevant, low molecular weight compound (indomethacin), and a 50 : 50 mixture of the two. The Tg of the polymer increased with water content up to RH ∼0.06, followed by a decrease at higher water levels. Consistent with the Tg results, the permeability of water in the polymer decreased with water content, up to RH ∼0.06, and began to increase with higher water concentration. Conversely, indomethacin showed exclusively a decrease in Tg when water was present. For the drug-polymer mixture, the Tg exhibited a plateau in the RH range where the pure polymer Tg increased, followed by a sharp decrease at higher water content, where the two pure components exhibited decreasing Tg. These results are consistent with previous observations of antiplasticization in two ways: (a) low concentrations of plasticizer have been found to increase, not decrease, the rigidity of polymeric materials, and (b) in the case of low molecular weight materials, plasticizers exclusively produce softening effects. However, there has been no previous report of a plasticizer producing an increase in the Tg of any material. The results presented here demonstrate that water can increase the Tg of a polymer. These results provide a new perspective on antiplasticization and negate any theoretical argument precluding the increase in Tg by addition of a small amount of plasticizer. Results on the Tg of the drug-polymer mixture show that the antiplasticizing effect of water is significant and can have important implications when polymers are used to stabilize amorphous systems.

Introduction Water is arguably the most common plasticizer in the pharmaceutical and food science fields. Its innocuous character makes water the plasticizer of choice when plasticization is desired. Its ubiquitous nature, on the other hand, also makes water a frequent unintentional plasticizer. The intentional (or not) alteration of the physical properties of polymeric materials is a subject of interest in numerous industrial applications. The effect produced by mixing a plasticizing solvent, such as water, with a polymer can be quantified through changes in the physicomechanical properties of the material. Plasticizers have a softening effect in polymers, and such an effect is observed either as an increase or a decrease in certain bulk properties. Plasticization is observed as a 401

402

PART 2: Poster Presentations

decrease in properties such as tensile strength, Young’s modulus, viscosity, and the glass transition temperature (Tg) in the case of glassy polymers. Properties that increase with the softening of a polymer include elongation and permeability. One interesting aspect of polymer-plasticizer interactions is that, in the domain of low plasticizer concentrations, the added solvent very often has the effect of increasing the rigidity of the polymer, before higher concentrations produce the softening effect typically expected. The rigidifying effect of small amounts of plasticizer, termed antiplasticization, has been observed in a significant number of polymer-plasticizer systems. The phenomenon is observable in the form of increases in mechanical strength parameters, as well as in terms of a decrease in transport and diffusion properties. However, an increase in the Tg of a polymer by the addition of small amounts of plasticizer has not been reported to date. Consequently, the prevailing view is that addition of a plasticizer will reduce the Tg of the polymer even if other (mechanical or transport) properties show antiplasticization. In this report, we present an investigation showing that the unique properties of water at low concentrations actually make it increase the Tg of a polymer.

Materials and Methods We selected Eudragit® E100 (Evonik Industries, Essen, Germany) and water as the polymer and plasticizer, respectively. Eudragit E100 is an aminoalkyl methacrylate copolymer (butyl methacrylate, methyl methacrylate, and dimethylaminoethyl methacrylate copolymer [1 : 1 : 2]). Indomethacin, also included in the study, is a small molecule that has been extensively studied in the glassy state, thus serving as a comparative reference regarding the reliability of the Tg values obtained by the analytical method chosen here. Low levels of the plasticizer (water) were introduced into the polymeric matrix by equilibration with environments of controlled relative humidity (RH). The change in Tg of the polymer as a function of the environmental RH was determined by inverse gas chromatography (IGC) by using a reported method (Thielmann and Williams 2000). A commercially available IGC instrument (SMS iGC; Surface Measurements Systems, London, UK) was used to determine the Tg under different conditions of RH. This particular analysis method was chosen because it is the only one that enables continuous and accurate control of RH of the environment in equilibrium with the sample throughout the glass transition measurements, even as the temperature is varied. About 800 mg of Eudragit E100 was packed in a column (340-mm length × 4-mm internal diameter). The polymer was initially conditioned under dry helium (5 mL · min−1) at 30°C for 40 h and further conditioned for 2 h at 60°C. This treatment ensured removal of residual water, as well as the thermal history of the sample. The polymer was then subjected to a temperature program ranging from 24° to 60°C. This temperature range encompasses the Tg of the polymer. Methanol (p/po = 0.03) was used as the vapor probe, and methane (p/po = 0.03) was used as reference. Probe and internal reference were both injected at various temperatures within the

Another Unusual Property of Water

403

selected temperature range. The outflow from the column was analyzed using a flame ionization detector. Preselected conditions of partial pressure of water vapor within the column were achieved with the humidity controls of the instrument. For measuring the Tg as a function of RH, the sample was first equilibrated at the desired RH level for 40 h at 30°C and 5 mL · min−1 flow rate. The attainment of equilibrium at each RH, including the column conditioning phase, was monitored by using a thermal conductivity detector also attached to the instrument. The Tg measurements were conducted under low-humidity conditions, with RH ranging from 0 to 0.15. The application of IGC to glass transition measurements is based on a change in the heat of interaction (chromatographic retention) between the probe and the polymer. Such a change is readily observable by plotting the retention data in the form: lnVn = −

ΔH +c RT

(29.1)

where Vn is the net retention volume, T is the absolute temperature, R is the universal gas constant, ΔH is the heat of interaction, and c is a constant. The slope in the retention volume (ln[Vn/T]) of the vapor probe as a function of reciprocal temperature will change as the polymer in the stationary phase changes from the glassy to the rubbery state. Figure 29.1 shows an example of a Tg determination by IGC. The point where the two straight lines cross corresponds to the Tg of the material. 0.2 Tg

In (Vn/T)

0.0

–0.2

–0.4

–0.6 3.00

3.05

3.10

3.15 T

3.20

3.25

3.30

· T –1/K

Figure 29.1. Determination of the glass transition temperature (Tg) by inverse gas chromatography. The slope of the retention volume data as a function of temperature shows two distinct profiles, corresponding to the glassy and rubbery regions. The glass transition temperature is the point where the two lines cross.

404

PART 2: Poster Presentations

(a)

(b)

Figure 29.2. Effect of low concentrations of water on the glass transition temperature (Tg) of indomethacin. (a) Water consistently decreases the Tg of indomethacin. (b) Water uptake of amorphous indomethacin in the low-water-activity region. RH, relative humidity.

Results and Discussion Figure 29.2a shows the effect of low water concentrations on the glass transition temperature of amorphous indomethacin produced by quench cooling of the melt. The corresponding water uptake (VTI, Hialeah, FL, USA) is shown in Figure 29.2b. The Tg values for indomethacin obtained by IGC are consistent with the values reported in the literature (Andronis and others 1997). It should be pointed out that the nature of the IGC method makes it such that the Tg values obtained correspond very closely to values obtained by differential scanning calorimetry at slow heating rates. Even when present at very low levels (less than ∼0.3%), water exclusively reduces the Tg of indomethacin. This result is consistent with the well-known effect of water on the Tg of many molecular organic glasses. The effect of low concentrations of water on the Tg of Eudragit E100 is shown in Figure 29.3a. In this case, as the level of water increases, an initial increase in Tg is observed, followed by a decrease in Tg at higher water contents. Figure 29.3b shows that the water content of Eudragit E100 and indomethacin (Figure 29.2b) are similar in the low-RH range investigated. Similar thermodynamic water activity results in a drastically different effect on the Tg of amorphous indomethacin and Eudragit E-100. The increase in Tg as a function of water content in Figure 29.3 is the first reported case where small amounts of a plasticizer (a liquid whose Tg is substantially lower than that of the polymer) results in a Tg that is greater than that of the pure polymer. The data show that, in small concentrations, water “antiplasticizes” Eudragit E100. It was mentioned that antiplasticization is a phenomenon commonly observed in the transport properties of polymers (Lachman and Drubulis 1964; Maeda and Paul 1987; Pant and others 1994). We can use the water uptake data of Figure 29.3b to investigate whether water has an antiplasticizing effect on permeability. From Fick’s second law of diffusion, it is possible to obtain the permeability of water through the polymer by

Another Unusual Property of Water

(a)

405

(b)

Figure 29.3. Effect of low concentrations of water on the glass transition temperature (Tg) of Eudragit E100. (a) Antiplasticization: water initially increases the Tg of the polymer before higher water levels produce the typically expected depression on Tg. (b) Water uptake of amorphous Eudragit E100 in the low-water-activity region. RH, relative humidity.

fitting the weight gain as a function of time under different water-activity conditions; namely, 2 6 M ⎛ π Dp ln ⎛⎜ 1 − t ⎞⎟ = ln ⎛ 2 ⎞ − ⎜ 2 ⎝π ⎠ ⎝ r ⎝ Me ⎠

⎞ ⎟⎠ t

(29.2)

where t is time; Mt and Me are the moisture uptakes at time t and at equilibrium, respectively; r is the mean particle radius; and Dp is the permeability of water through the polymer matrix. From the above expression, the permeability of water through the polymer can be obtained from the slope of a plot of ln(1 − Mt/Me) against t. In Figure 29.4, permeability values of water through Eudragit E100 are shown as a function of water content. The antiplasticizing effect of water is clearly observed. Incorporation of water into the polymer initially decreases permeability. Higher water levels result in the typically expected increase in permeability. The effect of water on the Tg and the permeability of Eudragit E100 are independent indications that small amounts of water increase the rigidity of the polymer. Decreases in permeability induced by small concentrations of plasticizer have been reported in the antiplasticization literature (Lachman and Drubulis 1964; Maeda and Paul 1987; Pant and others 1994). In contrast, the effect on Tg shown in Figure 29.3 has not been reported. The conventional view on antiplasticization (Sears and Darby 1982) is that “although the Tg value is lowered by the small amount of plasticizer, the modulus and tensile strength may increase appreciably before they begin to decrease” (p 52). It should be pointed out, however, that there is no first-principles argument precluding an increase in Tg by the addition of a small amount of plasticizer. It is rather the lack of reported instances to the contrary, that has led to the expectation that a plasticizer can only and should always decrease the Tg of a polymer.

406

PART 2: Poster Presentations

Permeability of water / cm2 s–1

2.4 × 10–5

2.0 × 10–5

1.6 × 10–5

1.2 × 10–5

0

5

10

15 RH

20

25

30

Figure 29.4. Permeability of water in Eudragit E100 as a function of relative humidity (RH). The antiplasticizing effect of water is observable as an initial decrease in permeability. Values were obtained from the kinetics of water uptake according to Equation 29.2.

The effect of water on the Tg of Eudragit E100 demonstrates that an increase in Tg upon the addition of small amounts of plasticizer is not a physical impossibility but rather a rare occurrence. However, Figure 29.3 does not explain what makes this particular polymer-plasticizer system behave this way. A likely explanation is that the Tg behavior shown in Figure 29.3 results from a combination of two factors: the specific interactions between the polymer and water, and the temperature range involved in the Tg measurements. We propose that antiplasticization on the Tg is more likely to be observed when (a) there are strong interactions between the polymer and plasticizer and (b) when the polymer has a low Tg. Sorption of small amounts of water by a substrate (polymer in this case) will result in substrate-water interactions at the most energetically favored sites (Carvajal and Staniforth 2006). This implies that, when sorbed water is scarce, the molecules will be more strongly bound than when water is more abundant. The strong interactions between water and polymer result in restrictions in molecular mobility that are manifested as the observed decrease in permeability. The same restrictions in molecular mobility are also responsible for the observed increase in Tg. However, from an experimental perspective, there is a significant difference between permeability and Tg measurements. Whereas permeability can be determined at any arbitrarily low temperature (e.g., room temperature or below), measurement of the Tg requires heating the material to the highest temperature where the glassy state can exist before transforming into the liquid. By heating a glassy sample to reach its Tg, we are unavoidably weakening any polymer-plasticizer interac-

Another Unusual Property of Water

407

tions present. It follows that, if the Tg is sufficiently high, it will take place when the polymer-plasticizer interactions have significantly obliterated by the thermal motion at that temperature. From these considerations, we can expect, in the case of water, that a Tg well below 100°C (the boiling point of water) would be necessary in order to see the manifestation of water-polymer interactions on the Tg. The system chosen for this study has a Tg in the range of 34°–42°C, not far above room temperature, where the effect on permeability is clearly observable. This is a likely reason why the increase in Tg is clearly observable, as well. In searching for evidence of an antiplasticization effect on the Tg, we necessarily obliterate, to some extent, the very effect we are trying to observe. The reduction in molecular mobility induced by small amounts of plasticizer is unavoidably countered by the heating involved in the Tg measurement itself. Such a countereffect becomes predominant to the point of expunging restricted mobility when the Tg is sufficiently high. This, however, does not mean that the modes of molecular motion responsible for the glass transition event are unaffected by the plasticizer when the temperature is well below the Tg. Both the Tg and permeability results in this study show that molecular mobility is reduced in the presence of small amounts of the plasticizer. We can conclude that, when a glassy polymer undergoes antiplasticization, the modes of motion responsible for the glass transition will also be restricted even if the effect is not observable at higher temperatures, such as those required for the glass transition event to be observed. Small amounts of water increase the Tg of the polymer (Figure 29.3) but reduce the Tg of indomethacin, a low molecular weight compound. These results are consistent with what is known about antiplasticization; namely, antiplasticization is a phenomenon observed in polymers, not in small molecules. To explore this further, we prepared a 1 : 1 (wt/wt) amorphous mixture of Eudragit E100 and indomethacin, and investigated the effect of low levels of water on the Tg of the mixture. The results, shown in Figure 29.5, exhibit all the features of a combination of the effects observed on the individual components. The dotted line, which corresponds to the linear combination of the individual glass transition temperatures of the two components (Figures 29.2 and 29.3), serves as an indication of the degree to which the behavior of the mixture is a combined effect. The Tg of the mixture shows a plateau in the region where water increases the Tg of the polymer but decreases the Tg of the molecular glass. The Tg of the mixture begins to decrease with water content roughly at the point where water depresses the Tg for the two individual components. These results strongly support the notion that water affects molecular mobility differently in polymeric and small-molecule glasses. The molecular origins of the different effects of water are unclear, but the difference in properties of the free volume in small molecular weight compounds and polymers may hold part of the answer. Reports on the effect of water on glassy carbohydrates have shown that water increases free volume in low molecular weight materials (Kilburn and others 2006) and also in polymers (Kilburn and others 2004). We propose that, for glasses of low molecular weight compounds, water solely increases free volume, thus acting as a plasticizer at all concentrations. For macromolecules, on the other hand, water at very low concentrations can also decrease free

408

PART 2: Poster Presentations

(a)

(b)

Figure 29.5. Effect of low concentrations of water on the glass transition temperature (Tg) of a 1 : 1 (wt/wt) mixture of Eudragit E100 and indomethacin. (a) Effect of water on the Tg of the mixture. The dotted line is the calculated value from the linear combination of the data in Figures 22.2 and 22.3. (b) Water uptake of the amorphous mixture in the low-water-activity region. RH, relative humidity.

volume in some instances, and the effect should be clearly observable below the Tg. The same type of argument can be made for direct measurements of molecular mobility. At very low RHs, water uptake by polymers has the effect of blocking free volume because, at low water levels, the strongly bound sorbed molecules dominate. At high RH, the more numerous and less strongly interacting molecules predominate (Cava and others 2006).

Conclusions The results presented here provide a new perspective on antiplasticization. Even though the antiplasticizing effect of small amounts of plasticizer has been documented extensively with regard to the mechanical and transport properties of polymers (Seow and others 1999), it has been widely believed that plasticizers can only decrease the Tg (Anderson and others 1995). The data presented here demonstrate that an increase in Tg produced by small amounts of plasticizer is a rare occurrence rather than a physical or fundamental impossibility. The results presented here negate any theoretical argument precluding the possibility of Tg undergoing an increase upon the addition of a plasticizer. We believe that the fact that water is the first plasticizer shown to increase the Tg of a polymer is no coincidence. Water is arguably a most unusual solvent: its remarkably strong interactive capacity, especially in relation to its small molecular dimensions, make it the plasticizer most likely to exert a significant decrease in molecular mobility; strong enough to be observed as an increase in the Tg. At low concentrations in a polymer, water can be expected to adopt the character of confined water: an intermediate state having solidlike and fluidlike properties. The relaxation time of confined water is slower, that is, of a different order of magnitude than that of bulk

Another Unusual Property of Water

409

water (Mashl and others 2003). The restricted mobility of confined water seems to “wedge in” the mobility of the polymer chains, giving rise to the observed increase in Tg. There is no way of foretelling how many other systems showing an increase in Tg by addition of a plasticizer are likely to be found in the future. We can anticipate that the number of such systems will remain small and that the unique properties of water make it likely to be again the plasticizer involved. The significance of the effect on Tg shown in Figure 29.3, however, does not depend on the abundance of such systems. The results of this study indicate that, below a critical concentration, water does decrease the molecular mobility of a glassy polymer. Even though such an effect may not be observable under the relatively high temperatures necessary to observe many glass transitions, water has the ability to restrict the molecular mobility of the polymer in the glassy state (i.e., below the Tg).

Acknowledgment National Science Foundation IUCRC 000364-EEC, Dane O. Kildsig Center for Pharmaceutical Processing Research.

References Anderson SL, Grulke EA, DeLassus PT, Smith PB, Kocher CW, Landes BG. 1995. A model for antiplasticization in polystyrene. Macromolecules 28:2944–54. Andronis V, Yoshioka M, Zografi G. 1997. Effects of sorbed water on the crystallization of indomethacin from the amorphous state. J Pharm Sci 86:346–51. Carvajal MT, Staniforth JN. 2006. Interactions of water with the surfaces of crystal polymorphs. Int J Pharm 307:216–24. Cava D, Cabedo L, Gimenez E, Gavara R, Lagaron JM. 2006. The effect of ethylene content on the interaction between ethylene–vinyl alcohol copolymers and water: (I) application of FT-IR spectroscopy to determine transport properties and interactions in food packaging films. Polym Test 25:254–61. Kilburn D, Claude J, Mezzenga R, Dlubek G, Alam A, Ubbink J. 2004. Water in glassy carbohydrates: opening it up at the nanolevel. J Phys Chem [B] 108:12436–41. Kilburn D, Townrow S, Meunier V, Richardson R, Alam A, Ubbink J. 2006. Organization and mobility of water in amorphous and crystalline trehalose. Nat Mater 5:632–5. Lachman L, Drubulis A. 1964. Factors influencing the properties of films used for tablet coating. I. Effects of plasticizers on the water vapor transmission of cellulose acetate phthalate films. J Pharm Sci 53:639–43. Maeda Y, Paul DR. 1987. Effect of antiplasticization on gas sorption and transport. 1. Polysulfone. J Polym Sci [B] 25:957–80. Mashl RJ, Joseph S, Aluru NR, Jakobsson E. 2003. Anomalously immobilized water: a new water phase induced by confinement in nanotubes [Letter]. Nano Lett 3:589–92. Pant BG, Kulkarni SS, Panse DG, Joshi SG. 1994. Modification of polystyrene barrier properties. Polymer 35:2549–53. Sears JK, Darby JR. 1982. The technology of plasticizers. New York: John Wiley & Sons. Seow CC, Cheah PB, Chang YP. 1999. Antiplasticization by water in reduced-moisture food systems. J Food Sci 64:576–81. Thielmann F, Williams D. 2000. Determination of the glass transition temperature of maltose and its dependence on relative humidity by inverse gas chromatography. Dtsch Lebensm Rundsch 96:255–7.

30 Molecular Mobility Interpretation of Water-Sorption Isotherms of Food Materials by Means of Gravimetric Nuclear Magnetic Resonance W. P. Weglarz, M. Witek, C. Inoue, H. Van As, and J. van Duynhoven

Abstract The use of moisture-sorption isotherms as a common method to characterize hydration properties and behavior of materials is well established. Moisture-sorption isotherms enable reliable predictions of shelf life and have become essential for designing stable multicomponent food materials. Several formalisms have been proposed to describe sorption isotherms analytically, but none of these approaches adequately reflects the molecular reality of the events taking place during hydration (e.g., phase transitions). Nuclear magnetic resonance (NMR) relaxometry is a well-recognized method for studying hydration effects at the level of molecular mobility. From the acquired transversal relaxation decays, the populations of protons in glassy/crystalline, semisolid (rubbery), and liquid phases can be obtained, based on differences in spin-spin relaxation time (T2). However, according to our knowledge, so far no general procedure has been described for linking relaxometric parameters with the hydration phenomena of food materials. In this chapter, an approach to combine classic moisture-sorption isotherms with time domain (td) proton nuclear magnetic resonance (1H NMR) (i.e., tdNMR) relaxometry is presented. We propose a general formalism for analytic description of a correlated gravimetric and tdNMR data set in terms of a population of different phases as a function of water uptake. To test usefulness of gravimetric NMR methodology, a range of food materials (vegetable powders and cereal-based bakery products) was equilibrated at different water activities. Thus, gravimetric measurements were conducted in conjunction with a 1H NMR relaxometry experiment. This enabled features to be assigned in the water-sorption isotherms, as well as in the NMR relaxation decays. The populations of protons in different phases varied with hydration state, and this could be modeled in terms of plasticization of biopolymers and/or dissolution of low molecular weight species in the different food materials. This approach can be applied to a broad range of food materials to gain insight into hydration behavior at the molecular level.

Introduction Foods can be considered as complex mixtures of macromolecules, solutes, solvents, and plasticizers. Water is a constituent of virtually every food product and can act 411

412

PART 2: Poster Presentations

both as a solvent and as a plasticizer. To describe the wide range of chemical, microbial, and physical processes that can occur in foods, researchers have relied on two general concepts. For predicting chemical and microbial stability (Roos and others 2007), one commonly employs the concept of water activity (aw), which is considered a good measure for the chemical potential of water (Schmidt 2004). For predicting physical phenomena, such as stickiness, sogginess, and crispness, the glass to rubber transition temperature (Tg) has been proven useful. Determination of the Tg considers water-induced mobility of the solid matrix on structural and macromolecular levels (Rahman 2006), whereas aw is related to the molecular mobility of water itself (Vittadini and Chinachoti 2003). To describe and predict the interaction between foods and water (Chirife and Buera 1995), the use of both aw and Tg has been recommended (Roos 1995a, 1995b). In many cases, the use of molecular mobility as defined by nuclear magnetic resonance (NMR) relaxometry shows better correlation with physical stability and microbial stability than do the classic measures. The use of “NMR state diagrams” has been proposed (Schmidt 2004; Lin and others 2006). In this work, we propose the combined application of time-domain NMR (tdNMR) relaxometry and gravimetric measurements to resolve the relations among aw, Tg, and the underlying mobility of water and the solid matrix. The experimental methodology is outlined, together with a formalism for quantifying and interpreting results for food samples of low and moderate levels of hydration. The approach is demonstrated on cereal crackers and vegetable powders.

Theory The total 1H NMR signal amplitude obtained from hydrated food material can be described as a sum of signals from the matrix (SM) and water (SW), which are proportional to the product of their masses (mM and mW) and proton mass fractions (ρM and ρW), respectively: S = SM + SW = ρM mM + ρW mW

(30.1)

Under the simplifying assumptions that water added during hydration is (a) not immobilized into a crystalline or glassy structure, (b) is not dissolving or plasticizing matrix material, and (c) is not involved in a chemical reaction with the matrix, then the relative water-to-matrix signal should be linearly dependent on added water content: Δm SW ρW ( m − mdry ) = = ρWM ρM mdry SM mdry

(30.2)

where ρWM = ρW/ρM is the relative water/matrix proton mass fraction, mdry is the true dry mass of the sample, and Δm = m − mdry is the mass of water adsorbed in the sample.

Molecular Mobility Interpretation of Water-Sorption Isotherms of Food Materials

413

SW/SM

complete dissolution

nondissolved matrix

dissolution Δm/mref

Figure 30.1. Schematic illustration of typical behavior of the SW/SM as a function of water content. Excess mobile signal leading to nonlinear dependence appears in case of the presence of a solvable (plasticizable) fraction. The dotted line corresponds to complete dissolution (plasticization) of the matrix material.

With increased hydration levels, the linearity expected from Equation 30.2 may no longer hold should a soluble fraction be present in the solid matrix (Harańczyk and others 1999) or should matrix material be plasticized by interaction with water. If the soluble part of the matrix is actually dissolved (or becomes more mobile) during hydration, the mobile signal shows excess over that expected from a mass increase of the water added alone, as illustrated in Figure 30.1. If the amount of the soluble fraction is limited, the dependence of the SW/SM on water content will become linear again at higher hydration level when all soluble material has dissolved. If the entire matrix is soluble, the relative water-to-matrix signal will reach infinity when all matrix material is dissolved. The same approach can be applied for the matrix fraction, which undergoes a glass-to-rubber transition (plasticization) induced by water uptake. In that case, the mobility of macromolecular chains is increased, which is reflected in the less solidlike appearance of the NMR decay (i.e., a change toward exponential behavior and increasing T2). Assuming that mS/mdry is a fraction of the matrix that is in solution or in a rubbery state, the relative tdNMR signal can be expressed as SW ρW Δ m + ρS mS , = SM ρNS ( mdry − mS )

(30.3)

where ρS and ρNS are proton mass fractions for solute (rubbery) and nondissolved (nonplasticized) matrix material, respectively. After rearrangement, it can be expressed in terms of the mobile (rubbery) fraction:

414

PART 2: Poster Presentations

SW ρ W 1 Δ m ρS = ⋅ ⋅ + SM ρNS ⎛ mS ⎞ mdry ρNS ⎜⎝ 1 − m ⎟⎠ dry

⋅

m 1 ⋅ S . mS ⎞ mdry ⎛ ⎜⎝ 1 − m ⎟⎠ dry

(30.4)

In a hydration range where the solute (rubbery) fraction is constant, this formula describes the linear dependence of relative NMR signal on relative mass of water added (Figure 30.1).

Methods Hydration Experiments Wheat flour–based crackers were produced locally on a pilot bakery scale. Samples were equilibrated at 15% relative humidity (RH) for 2 weeks, which was defined as the reference mass. Then, the samples were placed under 100% RH, and tdNMR decays were monitored over a 4-week period. Tomato and corn powders of commercial origin were predried in a vacuum oven at 60°C for 6 h. Eight NMR samples were prepared from each powder. The reference mass and NMR signal were established after a 1-month equilibration in 0% RH. Samples were then distributed to desiccators with controlled RH of 0%, 11%, 33%, 45%, 58%, 61%, 70%, and 85%, and equilibrated for another month prior to tdNMR measurements. Before measurements, the mass of the equilibrated samples was obtained. The hydration kinetics of cracker samples were measured by a 30-MHz Maran spectrometer (Resonance Instruments, Oxford, UK) at 20°C. T2 decays consisted of a free-induction decay (FID) and a subsequent Carr-Purcell-Meiboom-Gill (CPMG) echo train. The FID was sampled at 100 points every 0.8 μs, starting from 13 μs. Subsequently, 1800 CPMG echo amplitudes were collected every 0.2 ms. The π/2 pulse length was 4.9 μs. A recycle delay of 2 s was used and signal was accumulated during 2048 acquisitions. tdNMR measurements for equilibrated vegetable powders were done at 20°C, using a 20-MHz Bruker MiniSpec spectrometer (Bruker Optik, Ettlingen, Germany). The FID was sampled at 230 points every 0.4 μs starting from 8 μs, after which 1800 CPMG echo amplitudes were collected every 0.2 ms. The π/2 pulse length was 2.5 μs, the recycle delay was set at 10 s, and 256 signals were accumulated. T2 Decay Analysis T2 decay curves were fitted with a model comprised of functions describing the signals of solid (SS) and intermediate mobility and/or liquid (SL) components: 2 ⎛ ⎛ t ⎞ ⎞ ⎛ ⎛ t ⎞ ⎞ sin ( wt ) SS = Ag exp ⎜ − ⎜ A exp + − s ⎜ ⎟ ⎟ ⎟ ⎜⎝ ⎝ T ⎠ ⎟⎠ ⋅ wt ⎝ ⎝ T2 g ⎠ ⎠ 2s 2

⎛ t ⎞ SL = ∑ Aei exp ⎜ − ⎝ T2 ei ⎟⎠ i S = SS + SL

(30.5)

Molecular Mobility Interpretation of Water-Sorption Isotherms of Food Materials

415

The solid part of the signal is described by the sum of the Gaussian and damped sinc functions (Derbyshire and others 2004). For the present work, the amplitude of total solid signal (i.e., Ag + As) is of importance. Intermediate mobility and liquid signals are described by a discrete number of exponential functions, which are then grouped into hydration-dependent (i.e., water) and hydration-independent (i.e., liquid-fat) fractions. The liquid-fat fraction was then considered as part of the matrix, whereas the water fraction was separated from the liquid component to allow for correlation with gravimetric data. Equation 30.5 was used to fit the experimental data by a nonlinear least-squares procedure in Origin software (Origin Lab, Northampton, MA, USA).

Results In Figure 30.2a, the water-sorption isotherms for a porous cracker sample are presented for the aw range of ca. 0.2–0.95 (moisture content [MC] dry basis, ca. 0.05– 0.48). In Figure 30.2b, the results of the discrete component fit according to Equation 30.5 are presented. Upon hydration, the number of water components increases from one (T2, ∼400 μs) for MC below ∼0.2, to three (with T2 between ∼1 and ∼30 ms) for MC above ∼0.6. Also, a gradual increase in T2 values was observed for these waterrelated components. This is in marked contrast to the oil and solid matrix fractions, for which T2 is more or less constant. In Figure 30.2c, the dependence of the relative amplitude of the water NMR signal on the relative mass (SW/SM) increase during water uptake is shown. Three hydration ranges can be distinguished from the graph that corresponds to the theoretical prediction from Figure 30.1. For MC up to ∼0.2 and above ∼0.4, a linear dependence is present. The slopes for both linear regions have values characteristic of carbohydratebased materials. These are expected to lie between 1.5 and 1.8, based on known proton mass fractions of carbohydrate material. In the intermediate MC region, a nonlinear dependence showing an excess of liquid NMR signal over that expected from the mass of water uptake is detected. It should be noted that the onset of nonlinearity is correlated with the appearance of a second water-associated component with a T2 on the order of 1 ms (see Figure 30.2b). Based on Equation 30.4, and under rough assumption that ρNS and ρS are similar, it can be estimated from the results that ∼10% of the cracker dry mass is transferred to solution. Figure 30.3a shows the water-sorption isotherms of two vegetable powders, which display distinct differences. The tomato powder exhibited a significantly higher water uptake for the same aw than did corn powder. Figure 30.3b presents a correlation between the SW/SM and water content (Δm/mref) for these two vegetable powders. An approximately linear dependence is observed for water content below 5%–7%, whereas an excess amount of mobile fraction is present at higher hydration. The slope of the linear dependence is within the theoretical range for carbohydrate matrix with some oil and crystalline water fractions (1.5–1.6). Significant differences between samples are visible with increased hydration, with tomato showing much greater hydrationinduced excess of mobile fraction than does corn. The matrix fraction mobilized during

416

PART 2: Poster Presentations

(a)

(b)

0.6

1000

0.5

100

0.4 MC

T2 (ms)

10

0.3 0.2

0

0

0.2

0.4

aw

0.6

0.8

1

water

1

0.1

0.1

lipids

solid

0.01 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Δm/m15%

(c)

1.0 SW/SM

1.50

0.5 1.65 0.0 0.0

0.1

0.2

0.3 0.4 Δm/m15%

0.5

0.6

0.7

Figure 30.2. Water-sorption isotherms (a), dependence of spin-spin relaxation time (T2) on hydration (b), and dependence of relative mobile tdNMR signal on hydration (c) for a cereal cracker sample. Moisture content (MC) is dry-basis water content, shown in a; values of calculated slopes are shown in c. The open and closed symbols in b correspond to different components in solid, water, and lipid fractions of a tdNMR signal.

hydration, calculated from Equation 30.4 at the highest aw achieved is shown in the last column of Table 30.1.

Discussion The results obtained for crackers show that a relatively limited amount of material in the carbohydrate matrix is transferred to the mobile phase during hydration. The correlation of the onset of nonlinearity with the appearance of the relatively low-mobility mobile component (i.e., T2, ∼1 ms or less) strongly suggests that plasticization, rather than dissolution of the matrix material is responsible for this effect.

Molecular Mobility Interpretation of Water-Sorption Isotherms of Food Materials

(b) 40

12 0.5

35

10

30

tomato corn

25

SW/SM

Δm/mref [%]

(a)

417

20 15

6 0.0 4

10

0

5

10 15

tomato corn

2

5 0 0.0

8

0.4 0.3 0.2 0.1

0.2

0.4

aw

0.6

0.8

1.0

0

0

5

10

15 20 25 Δm/mref

30

35

Figure 30.3. Water-sorption isotherms (a) and dependence of relative mobile tdNMR signal on hydration (b) for two different vegetable powders. Expanded view of the low-hydrated region is presented in inset in b. The dotted line corresponds to the theoretical dependence in case of the absence of the soluble fraction (b).

Table 30.1. Main (dry-mass based) components of the vegetable powders used for the hydration experiments, compared with the fraction of the matrix component mobilized during hydration Powder

% Carbohydrates

% Sugar

% Protein

% Fat

% Mobile fraction

Tomato

74.6

43.9

12.9

0.40

86

Corn-flour whole-grain yellow

76.8

0.6

3.80

37

6.90

For vegetable powders, the amount of excess mobile signal corresponding to matrix material mobilized during hydration correlates positively with sugar content (see Table 30.1). However, calculation based on Equation 30.4 shows that water-soluble sugar can be only partially responsible for the presence of the mobilized fraction. Assuming that sugar is the predominant solvable fraction for both powders, ∼40% wt/ wt matrix fraction can undergo plasticization during hydration. While beyond the scope of this report, exact characterization of the different mobile fractions (i.e., solvable, plasticized) is possible by careful assessment of the mobility of the corresponding T2 components. Here, only the total mobile fraction is considered.

Conclusions Quantitative information about the amount and dynamics of water, as well as the possibility of assessing the presence of a solvable fraction and a plasticizable fraction within the solid matrix, make our proposed methodology a useful tool for studying hydration and dehydration behavior of food materials. The presented formalism for describing gravimetric NMR experiments enables a better understanding of classic

418

PART 2: Poster Presentations

water-sorption isotherms, typically used for characterization of food hydration behavior.

Acknowledgments We gratefully acknowledge support by the European Community within the framework of a Marie Curie Intra-European Fellowship (MEIF-CT-2005-009475). Part of this work was supported by the Dutch IS (Innovative Studies) Programme (Dutch Ministry of Economic Affairs, project IS042042). Nic Franciosi is thanked for skillfully preparing the cracker sample.

References Chirife J, Buera MP. 1995. A critical review of some non-equilibrium situations and glass transitions on water activity values of foods in the microbiological growth range 3. J Food Eng 25:531–52. Derbyshire W, van den Bosch M, van Dusschoten D, MacNaughtan W, Farhat IA, Hemminga MA, Mitchell JR. 2004. Fitting of the beat pattern observed in NMR free-induction decay signal of concentrated carbohydrate-water solutions. J Magn Reson 168:278–83. Harańczyk H, Weglarz WP, Sojka Z. 1999. The investigation of hydration processes in horse chestnut (Aesculus hippocastanum, L.) and pine (Pinus silvestris, L.) bark and bast using proton magnetic relaxation. Holzforschung 53:299–310. Lin XY, Ruan R, Chen P, Chung MS, Ye XF, Yang T, Doona C, Wagner T. 2006. NMR state diagram concept. J Food Sci 71:R136–45. Rahman MS. 2006. State diagram of foods: its potential use in food processing and product stability. Trends Food Sci Technol 17:129–41. Roos YH. 1995a. Characterization of food polymers using state diagrams. J Food Eng 24:339–60. Roos YH. 1995b. Glass transition-related physicochemical changes in foods. Food Technol 49:97–102. Roos YH, Leslie RB, Lillford PJ, editors. 2007. Water management in the design and distribution of quality food. Lancaster, PA: Technomic. Schmidt SJ. 2004. Water and solids mobility in foods. Adv Food Nutr Res 48:1–101. Vittadini E, Chinachoti P. 2003. Effect of physico-chemical and molecular mobility parameters on Staphylococcus aureus growth. Int J Food Sci Technol 38:841–7.

31 Kinetics of Enthalpy Relaxation in Corn Syrup–Sucrose Mixtures B. R. Bhandari and R. W. Hartel

Abstract Amorphous glasses of corn-syrup solids–sucrose mixtures at 30 : 70, 50 : 50, 70 : 30, and 100 : 0 proportions were prepared by boiling at 150°C, followed by cooling to ambient temperature. The enthalpy relaxation at various aging intervals (up to 24 h) was studied. The relaxation was rapid at higher water activity (aw), although the maximum enthalpy relaxed (ΔHmax) was lower. An increased proportion of corn-syrup solids in the mixture decreased the average relaxation time (τeff), even though the maximum relaxation enthalpy values (ΔHmax) remained similar for all the mixtures.

Introduction The amorphous glassy state is essentially a thermodynamically nonequilibrium state. When a glass is held at temperatures near the transition region or below, an excess thermodynamic quantity relaxes toward thermodynamic equilibrium. When the glass is held at a given condition, it is called annealing or physical aging. The macroscopic or molecular level changes (thermal or mechanical) taking place in the glass during aging or annealing phenomenon are known as enthalpy relaxation, volume relaxation, structure relaxation, or β relaxation (the glass formation process is termed cooperative α relaxation). This reduces the configuration energy (Yoshida 1995; Yamamuro and others 1998; Sun and others 1999; Bhandari and Hartel 2005) equivalent to a supercooled liquid. The β relaxation indicates local diffusion; both rotational diffusion and translational diffusion occur in loosely packed, isolated regions. This is seen as “islands of mobility” or “defects” in a glassy structure (Johari 2002, p 317), with the relaxation attributed to a small number of molecules or molecular groups in the confined region. These regions of low density or high density in the disordered structure of a glass are produced during the freezing-in process. There is also an assumption that this phenomenon occurs due to temperature-independent small-angle reorientation of all the molecules in the glassy structure (Vogel and others 2000). However, this theory regarding the latter mechanism is not widely supported (Johari 2002). Because of relaxation, the enthalpy of the glassy state decreases with aging time. Therefore, during the glass transition from solid to liquid state, an overshoot in the enthalpy is observed in the differential scanning calorimetry (DSC) thermogram. This overshoot is the recovery of the enthalpy lost during structural relaxation. Physical aging can be reversed by annealing the sample at temperatures 20°–40°C above the 419

420

PART 2: Poster Presentations

glass transition temperature (Tg) and then melting and cooling the sample at a suitable (high) cooling rate (Urbani and others 1997). The enthalpy relaxation is time and temperature dependent and is also influenced by additives (Shamblin and Zografi 1998; Sun and others 1999; Kim and others 2001; Bhandari and Roos 2003). The objective of this research was to extend the investigations to complex carbohydrate systems such as those found in candy formulations. The kinetics of relaxation were investigated as a function of water activity (aw), temperature, and composition.

Materials and Methods Sample Preparation in Glassy State at Various Water Activities The materials used to study the enthalpy relaxation behavior were mixtures of sucrose (Domino, Tate & Lyle American Sugars, Baltimore, MD, USA) and corn-syrup solids (DE 42; ADM, Decatur, IL, USA) at corn-syrup solids–sucrose ratios of 30 : 70, 50 : 50, 70 : 30, and 100 : 0. A 100% sucrose (corn syrup–sucrose = 0 : 100) system was not studied because of the difficulty in preparing the glass without crystallization during boiling and cooling. The method used to prepare the glass was similar to the method used to make hard candy. The sucrose (500 g), water (200 g), and required amount of corn-syrup solids were mixed in a Teflon-coated sauce pan and boiled to 150°C. The total cooking time varied between 14 and 18 min for various mixtures. The boiled mixture was poured onto an aluminum tray in the form of a thin, flat sheet and allowed to cool quickly to room temperature. The cooled sheet was broken and ground into a fine powder in a mortar. The ground mass was then sieved through a 100-μm screen, and the powder particles passing through the sieve (<100-μm size) were collected in glass bottles. These samples were then equilibrated at three water activities (0.11, 0.22, and 0.32) in a humidity-controlled chamber by using saturated salt solutions in the room maintained at 22° ± 1°C. It was previously determined by means of DSC that all mixtures, except that with a 30 : 70 ratio at an aw of 0.32, remained in the glassy state within this aw range. Determination of Enthalpy Relaxation Kinetics The glassy samples equilibrated at each aw had a previous thermal history and would have already relaxed prior to analysis; therefore, it was necessary to reverse these effects. Triplicate samples (close to 10 mg each) were quickly drawn and weighed in 50-mL DSC pans in a dry environment (relative humidity, <20%), and the pans were hermetically sealed. The samples were scanned by DSC at a rate of 50°C/min to 20°C above their predetermined Tg. The scanned samples were then immediately cooled to 22°C at the rate of 50°C/min and aged for up to 24 h at this temperature in a temperature-controlled room to determine the kinetics of enthalpy relaxation. A differential scanning calorimeter (DSC-7; Perkin Elmer, Wilton, CT, USA) connected to a low-temperature refrigeration system (Intracooler II, Perkin Elmer) was used. Dry nitrogen was used to purge the sample head and the dry box. The instrument was calibrated using mercury (−38.87°C) and indium (156.60°C) at the same temper-

Kinetics of Enthalpy Relaxation in Corn Syrup–Sucrose Mixtures

421

ature-scanning rate as the sample. Incubated samples in pans at various time intervals were scanned in the DSC at a rate of 10°–20°C/min above the Tg. The enthalpy relaxation was calculated by integrating the excess enthalpy area in the DSC thermogram. This area of excess enthalpy was determined by subtracting the relaxed sample thermogram from that of the fresh (zero relaxed time) sample in the glass transition region. The same scanned samples were then cooled to 22°C and used in subsequent aging study. Since the same samples were used for all the kinetic studies, it was necessary to confirm that no degradation had occurred in the sample. Since the Tg of any of the samples was <65°C, it was assumed that no thermal degradation would have occurred. The unchanged color of the repeatedly scanned sample and lack of change in its Tg (after reversing the relaxation effect) confirmed this assumption. The temperature effect on enthalpy relaxation was determined by incubating the samples at 10°, 20°, 30°, 35°, and 40°C in the DSC oven itself.

Results and Discussion Glass Transition Temperature of the Mixture The Tg and ΔCp (heat capacity) values of samples at various water activities and corn syrup–sucrose ratios are presented in Table 31.1. The increased Tg with the increased proportions of corn-syrup solids and the decreased Tg with an increase in aw are evident. Analysis of Enthalpy Relaxation Data Increase in the enthalpy relaxation values as a function of time can be observed in representative DSC scans in Figure 31.1. The ΔH is obtained from these scans by Table 31.1. Glass transition temperature (Tg) and heat capacity (ΔCp) value of corn syrup– sucrose (C-S) mixtures C-S ratio 30 : 70

50 : 50

70 : 30

100 : 0

Water activity

Tgon (°C)

SD (°C)

Tgmid (°C)

SD (°C)

Tgend (°C)

SD (°C)

ΔCp (J/g · °C)

SD (J/g · °C)

0.11

33.7

1.6

39.4

1.0

48.2

0.5

0.66

0.05

0.22

26.5

1.7

33.7

2.0

39.5

0.9

0.57

0.01

0.32

17.1

2.0

23.0

1.4

29.7

2.6

0.64

0.06

0.11

34.6

0.6

40.7

0.6

49.7

0.9

0.63

0.05

0.22

26.7

1.3

36.7

1.6

45.9

3.1

0.51

0.06

0.32

20.2

0.2

25.4

0.2

29.7

0.2

0.60

0.02

0.11

40.5

1.5

47.5

2.0

54.7

1.0

0.55

0.01

0.22

37.1

0.2

42.0

0.9

48.6

1.4

0.48

0.06

0.32

27.5

0.6

33.4

1.1

36.6

0.7

0.51

0.02

0.11

47.6

1.44

53.3

2.1

61.8

1.5

0.40

0.05

0.22

43.5

1.5

48.9

2.2

55.5

0.4

0.34

0.02

0.32

34.9

3.7

40.4

1.9

50.1

5.0

0.32

0.05

SD, standard deviation; Tgon, onset glass transition temperature; Tgmid, midpoint glass transition temperature; and Tgend, end-set glass transition temperature.

422

PART 2: Poster Presentations

(a) 22

∆H (mJ)- Endo up

16.5h 5h 3h

21

1h 0h

20

25

30

35

40

45

50

55

60

65

Temperature (°C)

(b) 16 ∆H(mJ) Endo up

5h 3h

15 1h 0h

14

13 10

20

30

40

50

60

Temperature (°C)

Figure 31.1. Representative scanning calorimetric thermograms of corn-syrup solids (a) and corn syrup–sucrose (30 : 70) (b) indicating increased enthalpy relaxation overshoots at longer annealing (aging) time (water activity, 0.22; and aging temperature, 22°C). Endo, endothermic.

integrating the ΔCp difference between aged and nonaged samples. At a given temperature, enthalpy relaxation values ΔH (ta, Ta) as a function of aging time and maximum extent of enthalpy relaxation ΔHmax (t∞, Ta) are described by the KohlrauschWilliams-Watts (KWW) equation (Equation 31.1) (Schmidt and Lammert 1996). Φ ( ta ) = 1 −

Δ H ( ta , Ta ) Δ H max ( t∞ , Ta )

= exp

t − ⎛⎜ a ⎞⎟ ⎝ τ eff ⎠

β

(31.1)

Here, Φ(ta) is a decay function, and τeff is some effective or average relaxation time of the system, which depends on the aging temperature (Ta) and on the departure of the system from the final equilibrium state. ΔH is the enthalpy relaxation value measured by DSC analysis (Figure 31.1) found at any time ta at a given Ta. The β describes the extent to which the relaxation process is nonexponential and thus the width of the

Kinetics of Enthalpy Relaxation in Corn Syrup–Sucrose Mixtures

423

relaxation time distribution. Since the relaxation processes in amorphous systems are nonexponential, β should be <1. The ΔHmax (t∞, Ta) is the enthalpy relaxation of a sample annealed for infinite time (t∞) at Ta. This value is calculated by using Equation 31.2 (Urbani and others 1997). Δ H = ΔCp (Tg − Ta )

(31.2)

The β and τeff are obtained from the linearized form (Equation 31.3) of the KWW equation. ln ( − ln Φ ( ta )) = β ln ( ta ) − β ln ( τ eff )

(31.3)

From this equation, the values of β and τeff are obtained by linear least-squares fitting of the ln(−ln Φ[ta]) vs ln(ta), assuming that τeff is a time-independent parameter (Moon and others 2001). The KWW relaxation kinetics and relaxation parameters studied in this work are summarized in Table 31.2. The time required for 50% (t50) of the maximum relaxation (ΔHmax) to occur is determined by Equation 31.4. Note that Φ = 0.5 when there is 50% of the maximum relaxation. The same equation was used to calculate the time required for 99% (t99) of the maximum relaxation to occur (in this case, Φ will equal 0.01). 0.5 = exp

β t − ⎛⎜ 50 ⎞⎟ ⎝ τ eff ⎠

(31.4)

Table 31.2. Kohlrausch-Williams-Watts parameters and other kinetic data determined for corn syrup–sucrose (C-S) mixtures C-S ratio 30 : 70

50 : 50

70 : 30

100 : 0

a

Water activity

Tg − Ta

Hmax (J/g)

β

τeff (h)

t50% (days)a

t99 (days)a

MSE%

0.11

16.4

10.8

0.33

151

2.1

5,015

0.22

10.7

6.1

0.29

16

0.2

1,366

4.3

0.32

0.0

Sample

Crystallized

—

—

—

—

1.2

0.11

17.7

11.1

0.36

79

1.2

1,514

10.2

0.22

13.7

7.0

0.41

36

0.6

329

6.6

0.32

2.7

1.6

0.38

0.11

24.5

13.5

0.41

462

7.8

4,580

10.7

0.22

19.0

9.1

0.25

695

6.6

215

3.7

0.32

10.4

5.3

0.42

4

0.1

37

4.7

0.11

30.3

12.1

0.25

18,720

174.7

6,737,581

3.2

0.22

25.9

8.8

0.32

361

4.7

16,430

4.4

0.32

17.4

5.6

0.41

24

0.4

252

13.6

0.4

0.01

5.3

t50 and t99 are the times required for 50% and 99% relaxation; and MSE, mean square error.

1.1

424

PART 2: Poster Presentations

22

∆H(mJ) Endo up

a w=0.32

a w=0.2

21 a w=0.11

20 15

25

35

45

55

65

Temperature (°C)

Figure 31.2. Representative differential scanning calorimetric thermograms of cornsyrup solids (annealing for 5 h at 22°C) indicating the increased enthalpy relaxation overshoot at higher water activity (aw). Endo, endothermic.

Enthalpy Relaxation as a Function of Water Activity The kinetics of enthalpy relaxation was analyzed at three different water activities. The aw values were chosen in such a way that the glasses prepared generally did not crystallize during storage. This experiment was conducted only within a narrow range of water activities where crystallization did not occur. One exception was the sample at the corn syrup–sucrose ratio of 30 : 70, which crystallized at aw 0.32. Representative thermograms illustrating the effect of aw are presented in Figure 31.2. Table 31.2 shows that the enthalpy relaxation parameter β varied from 0.25 to 0.42. This is in the same range as the value (0.32) reported by Urbani and others (1997) for sucrose. Values of β much less than 1 indicate a complex exponential nature (breadth of the distribution) of relaxation. The maximum enthalpy relaxation value (ΔHmax) was high for systems with low aw, although the average relaxation time (τeff) was short. This means that the rate of relaxation is high (glass approaches the equilibrium state faster) at higher aw, and that the total relaxation value necessary to reach the equilibrium is also low at higher aw. Since the Tg of a glass at low aw is high, the differential temperature (Tg − Ta) will be high, and the total relaxation enthalpy to reach to equilibrium is large. In previous publications, it has been stated that the water acts as an antiplasticizer within the glassy state (Seow and others 1999; Lourdin and others 2003; Braga and Cunha 2004). This means that, instead of softening, the product toughens at higher moisture levels. This phenomenon is evident from these results, since the relaxation is faster at higher aw, and therefore the product will increase in density much earlier than it will at lower aw levels, because of the facilitation of localized molecular mobility by water. This will certainly increase the hardness of the material within the timescale of observations; however, our results demonstrate that the total structural relaxation achieved by the glass at lower aw is low because of the low rate of relaxation. Therefore, if the measurement is done after longer aging, the product with lower aw will be found to be denser, such as if we compare the time to relax 99% of the

Kinetics of Enthalpy Relaxation in Corn Syrup–Sucrose Mixtures

425

Table 31.3. Temperature effect on relaxation behavior of a candy formulation (Tg = 37.9; corn syrup–sucrose, 30 : 70)a Tg − Ta (°C)

Hmax (J/g)

Hmax

10

27.9

18.41

0.349

2,814

20

17.9

11.81

0.512

39

30

7.9

5.21

0.335

35

2.9

1.91

0.359

40

−2.1

0.35*

—

Temp. (°C)

a b

β (h)

t50 (h)b 983 19.2

4.2 0.09 —

1.41 0.03 —

t99 (days)b 68,443

MSE% 0.3

126

3.4

131

11.9

1.9 —

8.9 —

Observed enthalpy relaxation value. t50 and t99 are the times required for 50% and 99% relaxation; and MSE, mean square error.

22

∆H(mJ) Endo up

30:70 50:50 70:30

21 100:0

20 15

25

35

45

55

65

Temperature (°C)

Figure 31.3. Representative differential scanning calorimetric thermograms of corn syrup–sucrose mixtures at various proportions (water activity, 0.11; aging time, 5 h; and aging temperature, 22°C). Endo, endothermic.

maximum (t99%) value (Tables 31.2 and 31.3). This is contradictory to the results usually reported in previous publications. Probably other results were based on the measurements done within a limited aging timescale (days but not in years, as suggested by our maximum aging times needed to reach to equilibrium). This can be one of the areas for further investigation. The predictability of the relaxation time by the KWW model at three different water activities was reasonably good because most of the mean square error (MSE) values were <10%. Figure 31.3 demonstrates the observed values and those predicted by using the KWW function. There is generally excellent agreement between experimental data and the model; however, in some cases, the observed values were found to be somewhat scattered. These types of variations have also been observed in other published reports. The relaxation values might be strongly influenced by the way the glass is formed, and any small variation in process conditions may affect these relaxation values, even though a constant scanning rate was used in DSC. The microscale effect could not be explained.

426

PART 2: Poster Presentations

Enthalpy Relaxation as a Function of Corn Syrup–Sucrose Ratio A small amount of additive can influence the enthalpy relaxation of a principal component through molecular interaction, due to coupling of molecular motions. In our case, both ingredients (corn syrup and sucrose) can be considered as additives, depending on which is in the lower proportion. In general, smaller molecules relax more rapidly and to higher degree than do large molecules. An additive with larger molecules slows the relaxation and vice versa. Table 31.2 demonstrates that the average relaxation time (τeff) or the t50 was longer as the proportion of corn syrup was increased in the mixture, with the 100% corn-syrup sample (corn syrup– sucrose = 100 : 0) having the highest value. However, the maximum relaxation enthalpy varied within a narrow range for all mixtures. Both observations were also reported by Shamblin and Zografi (1998) for binary mixtures of sucrose and high molecular weight additives. The maximum enthalpy relaxed by mixtures of sucrose and additives reported by Shamblin and Zografi (1998) is similar to that found in our experiments (10–14 J/g at aw 0.11). The Tg of the mixture at a given aw increases with an increase in proportion of high average molecular weight corn syrup (Table 31.1). It is normally expected that, when Tg − Ta increases, relaxation enthalpy should also increase. This was not the case in these mixtures, which indicates that molecular interactions may take place between sucrose and corn-syrup components. The observed and predicted enthalpy relaxation values for a mixture at aw 0.11 are presented in Figure 31.4. The KWW function was found to predict relaxation enthalpies reasonably well (mostly MSE < 10%) for the mixtures. An MSE higher than 10% might be due to experimental error, as described previously.

5.00

Enthalpy recovery (J/g)

4.50 4.00 3.50 3.00 2.50 2.00

Predicted

1.50 1.00 0.50 0.00 0

5

10

15

20

Aging time (h)

Figure 31.4. Observed and predicted (Kohlrausch-Williams-Watts model) enthalpy relaxation values for corn syrup–sucrose mixture (30 : 70) at an aging temperature of 22°C (water activity, 0.22).

Kinetics of Enthalpy Relaxation in Corn Syrup–Sucrose Mixtures

427

Enthalpy Relaxation as a Function of Temperature The 30 : 70 corn syrup–sucrose mixture was analyzed to investigate the influence of temperature on the enthalpy relaxation of glasses. Since the enthalpy relaxation depends on how far the glass has been cooled below the Tg, a higher temperature difference (Tg − Ta) is expected to need a longer time to reach equilibrium (the rate in reaching equilibrium will be slowed). The effect of temperature on relaxation behavior is exemplified in Table 31.3. An additional aging experiment at 40°C (slightly above the Tg) was conducted to determine whether any relaxation was observed when the matrix was held above its Tg. A low level of relaxation (∼0.35 J/g) was observed that held fairly constant until 2 h of aging time (Table 31.3). This relaxation might result from some microregions of glass in the rubbery liquid held just above the Tg. From the data in Table 31.3, it appears that β is not temperature dependent; however, the average relaxation time (τeff) decreased with an increase in temperature. The maximum enthalpy relaxation value also increased nearly linearly with Tg − Ta. An Arrheniustype analysis of 1/(Tg − Ta) against enthalpy relaxation of the sample did not show a linear relationship. Urbani and others (1997) proposed calculating apparent activation energy based on the logarithm of average relaxation time (1/τeff) and difference in relaxation enthalpy (ΔHmax − ΔH) at a given aging time. Since our data did not cover a wide enough range of temperature from the Tg, further experiments are needed to find the activation energy.

Conclusion Enthalpy relaxation was found to exist in corn syrup–sucrose glasses. The relaxation rate and extent were dependent on the time, temperature, and aw of storage. Increased corn syrup in sucrose generally reduced the average rate of relaxation. From the enthalpy relaxation data, it is clear that corn-syrup solids–sucrose will both increase in density with age, affecting their mechanical properties. We were unable to analyze the changes in the mechanical properties by using dynamic mechanical thermal analysis (DMTA), because of either the inability to prepare uniform samples or the insensitivity of the instrument. This area warrants further investigation.

References Bhandari BR, Hartel RW. 2005. Phase transitions during food powder production and powder stability. In: Onwulata C, editor. Encapsulated and powdered foods. New York: Taylor and Francis. p 261–91. Bhandari BR, Roos Y. 2003. Dissolution of sucrose crystals in the anhydrous sorbitol melt. Carbohydr Res 338:361–7. Braga ALM, Cunha RL. 2004. Plasticization and antiplasticization by small molecules in brittle cellular food: TMDSC and mechanical properties. Int J Food Properties 7:105–20. Johari GP. 2002. Localized molecular motions of β-relaxation and its energy landscape. J Non-Cryst Solids 307–10:317–25. Kim YJ, Suzuki T, Hagiwara T, Yamaji I, Takai R. 2001. Enthalpy relaxation and glass to rubber transition of amorphous potato starch formed by ball milling. Carbohydr Polym 46:1–6. Lourdin D, Colonna P, Brownsey GJ, Noel TR, Ring SG. 2003. Structural relaxation and physical ageing of starchy materials. Carbohydr Res 337:827–33.

428

PART 2: Poster Presentations

Moon IK, Jeong YH, Furukawa T. 2001. Enthalpy and dielectric relaxation in the glass transition region of polypropylene glycol. Thermochim Acta 377:97–104. Schmidt SJ, Lammert AM. 1996. Physical aging of maltose glasses. J Food Sci 61:870–5. Seow CC, Cheah PB, Chang YP. 1999. Antiplasticization by water in reduced-moisture food systems. J Food Sci 64:577–81. Shamblin SL, Zografi G. 1998. Enthalpy relaxation in binary amorphous mixtures containing sucrose. Pharm Res 15:1828–34. Sun N, Yang J, Shen D. 1999. The effect of water absorption on the physical ageing of amorphous poly(ethylene terephthalate) film. Polymer 40:6619–22. Urbani R, Sussich F, Sandra P, Cesaro A. 1997. Enthalpy relaxation and glass transition behaviour of sucrose by static and dynamic DSC. Thermochim Acta 304–5:359–67. Vogel M, Medick P, Rossler E. 2000. Slow molecular dynamics in binary organic glass formers. J Mol Liquids 86:103–8. Yamamuro O, Oishi Y, Nishizawa M, Matsuo T. 1998. Enthalpy relaxation of glassy glycerol prepared by rapid liquid quenching. J Non-Cryst Solids 235–7:517–21. Yoshida H. 1995. Relationship between enthalpy relaxation and dynamic mechanical relaxation of engineering plastics. Thermochim Acta 266:119–27.

32 Development of a Novel Phase Transition Measurement Device for Solid Food Materials: Thermal Mechanical Compression Test (TMCT) Y. Liu, P. Intipunya, T. T. Truong, W. Zhou, and B. R. Bhandari

Abstract A thermal mechanical compression test (TMCT) device was developed for the measurement of glass-rubber transition and surface stickiness properties of food particulate materials. This device can be used to measure changes in mechanical properties during a glass-rubber transition of solid food materials. The glass-rubber transition temperature (Tg) of various food materials such as sugars, skim-milk powder, fruitjuice powders, whey powders, starch, candy, rice, and nonfood polymers were tested by using this technique, and the data were compared with those from the standard technique. In this TMCT method, about 1 g of powder samples or a reasonably sized multiple or single particle grain is subjected to compression force (ca. 10–40 Newtons) and scanned under compression at 30°C/min heating rate. The inflection point is observed at the time of probe displacement when the sample changes from the glassy state to the rubbery state. The results obtained from this technique were comparable with Tg analyzed by a standard differential scanning calorimetry technique.

Introduction Solid food materials commonly exist in either a crystalline or an amorphous state (Liu and others 2006). Through heating, crystalline food undergoes first-order phase transition (i.e., melting), whereas amorphous food undergoes second-order phase transition (i.e., glass-liquid transition) (Liu and others 2006). These transition temperatures can be determined by using instruments that analyze the changes in physical properties, such as thermal mechanical analysis (TMA) or dynamic thermal mechanical analysis (DMTA) for rheological properties (viscosity and modulus); and differential scanning calorimetry (DSC) for thermal properties (enthalpy and heat capacity). However, application of those conventional methods have limitations. DSC is insensitive in determining glass transition temperature (Tg) in food polymers (e.g., proteins, starches) because of the small change in heat capacity (Boonyai and others 2006), and in real food because of the complexity and heterogeneity in its composition (Liu and others 2007). In such cases, TMA or DMTA may be more sensitive because of the large mechanical property change, but each requires a sample with proper shape 429

430

PART 2: Poster Presentations

Heating elements Thermocouple for temperature recording

Thermocouple for temperature control

Heating elements Sample

Sample cell

Figure 32.1. Schematic diagram of the sample cell.

(e.g., a sheet or strip). In addition, the powder sample needs to be compressed into a tablet before the test. Through this preparation, the sample properties may be altered by moisture loss, pressure effects, and other possible causes. Meanwhile, slow heating rate, small force, and high cost also limit the applications of TMA (Boonyai and others 2006). A new measuring device—the thermal mechanical compression test (TMCT), as shown in Figure 32.1, measuring the probe displacement resulting from mechanical property change (i.e., viscosity) during phase transition—has been developed by Bhandari and coworkers at the University of Queensland to determine the Tg for solid food materials. The studies by Boonyai and others (2005, 2006, 2007), found the TMCT to be simple, fast, accurate, robust, reproducible, and economical. The sample can be in powder, granular, flat-plate, or any other form. In many cases, there is no need for preparation of the sample.

Methodology Design and Setup of TMCT As shown in Figure 32.2, an aluminum sample cell is constructed with an embedded heating element and with a thermocouple for heating control. The round cell in the center of the metal block works as a sample holder. The heating elements and thermocouple for heating control are linked to the heater controller (Figure 32.1), which uses proportional integral derivative (PID) principles to control the heating rate and final temperature. The performance of sample cell and the heater controller is presented in Figure 32.3. A texture analyzer is used to facilitate the test (Figure 32.1). A 35-mm-diameter aluminum probe is used to compress the sample powder in the sample cell with a constant force. The force of the probe and its speed are controlled by the texture analyzer software, which also records data on compression force, probe position, and sample cell temperature. When the solid sample in the sample cell is compressed under the probe with constant pressure at constant temperature at the equilibrium state, the sample does not flow because of its high viscosity in the range of more 1012 Pa · s (Liu and others 2006). Once the melting or glass-liquid phase transition occurs, the sample’s viscosity

Thermal Mechanical Compression Test

37 mm

431

25 mm

50 mm

5 mm

Thermocouple for heating control Thermocouple for temperature recording

Heating elements

Sample cell

Figure 32.2. Design of the sample cell (Boonyai 2005). 240 Temperature measured

Temperature (°C)

210 180 150 Temperature set: 120

Heat from 30 to 250°C at 30°C/min

90 60 30 0

50

100

150

200 250 Time (second)

300

350

400

450

Figure 32.3. Performance of the heater controller. The dashed line is the temperature set (heat from 30° to 250°C at 30°C/min), and the solid line is the temperature measured during the experiment.

abruptly drops to the order of 104–107 Pa · s or even lower (Liu and others 2006). Under pressure, the supercooled melt (rubber) or the melt will flow and thus abruptly displace the probe, as shown in Figure 32.4c. Operation Protocol of TMCT The TMCT operation protocol has been studied and optimized in the research by Boonyai and others (2005, 2006, 2007) and has been improved in this study, as shown in Figure 32.4. The probe position at 30°C is normalized as the reference level, where

432

PART 2: Poster Presentations

Probe Probe

Sample Cell

Sample Cell

Displacement (mm)

(a) Blank Curve with Thermal Insulation Board

Probe

Sample Cell

Sample Cell

Sample curve subtracted by blank curve

Probe Probe

Before phase transition

After phase transition

Blank 50

70

90 110 130 Temperature (°C)

150

170

0.3 0.2 Glucose

0.1 0 –0.1 –0.2 30

Displacement (mm)

(c)

Probe

Displacement (mm)

(b) Sample curve

0.4 0.3 0.2 0.1 0 –0.1 –0.2 –0.3 30 0.4

0.4 0.3 0.2 0.1 0 –0.1 –0.2 –0.3 30

50

70

90 110 130 Temperature (°C)

150

170

Glucose - Blank Glucose

Blank 50

70

90 110 130 Temperature (°C)

150

170

Figure 32.4. Standard operation protocols for the thermal mechanical compression test (TMCT).

the probe displacement is 0. When the probe moves downward, the probe displacement is positive; when the probe moves upward, its displacement is negative. First, a blank curve indicating the expansion of the aluminum sample cell is determined by heating the sample cell alone with a thermal insulation board (36-mm diameter) between the probe and sample cell to avoid overheating the probe (Figure 32.4a). A negative displacement curve suggests the probe’s upward movement caused by the thermal expansion of sample cell, which could be modeled by using the following second-order mathematical equation: y = ax2 + bx + c (Boonyai and others 2005, 2006, 2007). When solid sample powder (0.5–1.0 g) is placed in the sample cell, 5 kg of force is applied through the 35-mm-diameter probe, where the pressure is approximately 5.09 × 104 Pa. Within 5 min, the probe displacement reaches a constant value because the sample cannot be compressed further in this experimental condition. During heating (30°C/min starting at room temperature), the sample surface contacting the sample cell undergoes the phase transition and then subsequently flows and causes the positive probe displacement. The signal measured is the composition of expansion of both the sample cell and the sample flow (Figure 32.4b). After subtraction of the blank curve, the probe displacement attributable to the sample cell thermal expansion and sample phase transition is isolated (Figure 32.5). Probe displacement due to sample thermal expansion is normally insignificant in the

Thermal Mechanical Compression Test

433

0.4 0.35 Displacement (mm)

0.3

Glucose - Blank

0.25 0.2 0.15 0.1 0.05 0 Tm-TMCT

-0.05 -0.1 30

50

70

90

110

130

150

170

Temperature (°C)

Figure 32.5. Determination of the onset of glucose melting temperature (Tm–TMCT), based on the mechanical behavior measured by the thermal mechanical compression test (TMCT).

Table 32.1. Onset of the melting temperature (Tm) of sugar crystals and some plastic polymer beads measured by thermal mechanical compression test (TMCT) and differential scanning calorimetry (DSC) Crystalline sample

Tm–TMCT (°C)

Tm–DSC (°C)

Sucrose

187.99 ± 2.74

189.97 ± 0.83

Maltose

128.18 ± 2.81

127.28 ± 1.83

Glucose

161.47 ± 0.49

158.06 ± 0.46

Fructose

132.89 ± 0.43

121.44 ± 0.61

Lactose

214.44 ± 0.68

214.04 ± 0.25

Polyethylene

121.01 ± 1.15

123.68 ± 0.72

Polypropylene

153.06 ± 2.51

155.69 ± 0.31

Low-density polyethylene (LDPE)

106.16 ± 0.89

106.29 ± 0.16

Linear LDPE

112.29 ± 1.05

114.58 ± 2.49

experimental temperature range. The corresponding temperature of the intercept between the displacement lines before and within the phase transition is the onset of the phase transition temperature (Figure 32.5).

Applications TMCT is ideal for measuring the phase transition of solid food materials. The firstorder phase transition (melting) of crystalline food solids using common sugar crystals is presented in Table 32.1. The displacement curve of glucose melting measured by TMCT is presented in Figures 32.4 and 32.5. Table 32.1 also presents the results obtained from some synthetic polymers (crystalline) that are commonly used in food packaging. For comparison purpose, the values from conventional analytic techniques, such as DSC and TMA, reported in the literature are presented along with

434

PART 2: Poster Presentations

Figure 32.6. Comparison of glass transition temperatures (Tg) measured by thermal the mechanical compression test (TMCT), dynamic thermal mechanical analysis (DMTA)/thermal mechanical analysis (TMA), and differential scanning calorimetry (DSC) for whey powder.

Table 32.2. Onset of the glass transition temperature (Tg) of food powders measured by thermal mechanical compression test (TMCT), differential scanning calorimetry (DSC), and thermal mechanical analysis (TMA) Sample Whey powder

Moisture content (% dry basis)

Tg–TMCT(°C)

2.77 ± 0.12

69.4 ± 0.8

5.84 ± 1.22

61.8 ± 0.1

Tg–DSC (°C)

Tg–TMA (°C)

See Figure 32.1

9.03 ± 1.41

37.7 ± 0.4

Honey powder

2.68 ± 0.36

58.5 ± 0.6

53.9 ± 1

Apple-juice powder

3.38 ± 0.18

55.3 ± 1.4

50.3 ± 2

Skim-milk powder

2.01 ± 0.07

59.13 ± 1.64

68.76 ± 0.60

93.73 ± 0.10

3.54 ± 0.23

52.96 ± 1.34

56.91 ± 0.39

89.25 ± 2.76

4.83 ± 0.05

38.66 ± 0.48

39.24 ± 1.26

83.02 ± 1.93

88.7 ± 2 81.7 ± 4

the TMCT results. The scanning rate for DSC is 10°C/min. The melting temperatures measured by TMCT are very close to those measured by DSC. This suggests the reliability of TMCT for the measurement of melting. For most solid food materials, Tg is the most interesting physical property because of its significance for physical and chemical stability during storage (Liu and others 2006). However, its measurement by conventional methods is not easy because of several reasons suggested earlier. TMCT was originally developed for measuring the Tg of food powders. Some spray-dried food powders were analyzed for their Tg, as shown in Table 32.2. Their glass transition temperatures were also measured by DSC

Displacement (mm)

Native Cornstarch Increase Moisture Content 0.3 0.25 0.2 0.15 0.1 0.05

0 30 Candy 1.1

50

70

90 110 130 Temperature (°C)

150

170

Displacement (mm)

0.9 0.7 0.5 0.3 0.1 –0.1 30

35

Pasta

40 45 50 Temperature (°C)

55

60

130

150

Displacement (mm)

0.05 0.04 0.03 0.02 0.01 0

–0.01 30

70 90 110 Temperature (°C) Individual Rice Kernels 0.2 19.50% MC Displacement (mm)

50

0.15

14.46% MC

0.1 7.81% MC

0.05 0

–0.05 30

50

70 90 110 Temperature (°C)

130

150

Figure 32.7. Glass-rubber transitions measured by TMCT for pasta, candy, native cornstarch, and individual rice kernels. MC, moisture content.

435

436

PART 2: Poster Presentations

and TMA. The scanning rate for DSC was 10°C/min, and the scanning rate for TMA or DMTA was 5°C/min with 9.81 Newtons of compression force. Their comparison is shown in Table 32.2 and Figure 32.6. Due to the complexity of glass transition measured by different methods (Liu and others 2006), including the influence of the kinetic nature of the measurement and thermal history of the sample itself, the glass transition temperatures measured do not agree closely with each other. But, based on the effect of moisture content, the consistency of the glass transition temperatures measured is shown in Table 32.2. In addition to measuring the Tg of food powders, TMCT is extremely useful for measuring the Tg of solid food itself, despite its internal heterogeneity and small change in heat capacity during the transition. Figure 32.7 shows the results of Tg measurements for native cornstarch, candy, pasta, and individual rice kernels.

Conclusion TMCT is a quick and easy method for measuring the phase transitions of solid food materials. It is accurate and robust, with minimum sample preparation required. Together with its low operation cost, it is ideal for the routine analysis of solid food samples and has a great potential for wide application in research institutes and industries.

References Boonyai P. 2005. Development of new instrumental techniques for measurement of stickiness of solid particulate food powders [PhD diss]. Brisbane, Australia: University of Queensland. Boonyai P, Bhandari B, Howes T. 2005. Measurement of glass-rubber transition temperature of skim milk powder by static mechanical test. Drying Technol 23:1499–514. Boonyai P, Bhandari B, Howes T. 2006. Applications of thermal mechanical compression tests in food powder analysis. Int J Food Properties 9:127–34. Boonyai P, Howes T, Bhandari B. 2007. Instrumentation and testing of a thermal mechanical compression test for glass-rubber transition analysis of food powders. J Food Eng 78:1333–42. Liu YT, Bhandari B, Zhou WB. 2006. Glass transition and enthalpy relaxation of amorphous food saccharides: a review. J Agric Food Chem 54:5701–17. Liu YT, Bhandari B, Zhou WB. 2007. Study of glass transition and enthalpy relaxation of mixtures of amorphous sucrose and amorphous tapioca starch syrup solid by differential scanning calorimetry (DSC). J Food Eng 81:599–610.

33 Proton Nuclear Magnetic Resonance Studies of Molecular Mobility in Potato Systems in Relation to Nonenzymatic Browning N. C. Acevedo, C. Schebor, and M. P. Buera Abstract Water-solid interactions were analyzed in potato samples in a wide range of relative humidities (RHs) by means of proton nuclear magnetic resonance (1H NMR), differential scanning calorimetry (DSC), and water-sorption isotherms. Nonenzymatic browning development was studied at 70°C and in relation to the physical properties of the samples. Proton relaxation was analyzed by 1H-NMR measurements. The spinspin relaxation time (T2) attributed to the fast-relaxing protons (from solids and most tightly associated water molecules) was measured using a free-induction decay (FID) analysis. The T2 associated with the slow-relaxing protons of water molecules was measured by using the Hahn spin-echo pulse sequence. The FID analysis showed a single T2 component at around 7–20 μs, which was associated with the mobility of the polymer chains. Samples between 11% and 52% RH showed a slight increase in T2 values as the RH increased and did not change along the temperature scale significantly. The systems at 75%, 84%, and 93% RH presented much higher T2 values because of the plasticizing effect of water and showed a marked increase in T2 along the temperature scale close to the glass transition. The Hahn analysis showed two components. A short component on the order of 30 μs, present at all the RHs analyzed, which corresponds to solid polysaccharide protons as the T2 measured by FID analysis. Above 22% RH, a long component in the range of 400–950 μs, which reflects the relaxation time of water protons, was also observed the value of which increased with the increase in water content. Nonenzymatic browning was detected in the glassy state; however, the maximum color development was detected at 84% RH, well above the glass transition temperature (T − Tg = 40°C), and then, above this RH, the reaction rate decreased, coinciding with the appearance of freezable water (determined by DSC) and with the higher T2 values. The integrated analysis of the water sorption, water, and solids mobility and of thermal transitions aids in the interpretation of the mechanisms and kinetics of chemical reactions.

Introduction Nonenzymatic browning (NEB) is one of the most important chemical phenomena that may affect food quality in processing and storage (Namiki 1988). The control of the mechanisms that determine this reaction rate has been given much attention. Some studies have evaluated the impact of water activity (aw) and the glass transition on 437

438

PART 2: Poster Presentations

chemical reactivity (Karmas and others 1992; Bell and Hageman 1994; Buera and Karel 1995). Diffusion-controlled chemical and enzymatic reactions are particularly dependent on translational diffusivity of the reactants (or on the viscosity of the matrix material) and are thus susceptible to the physical state of the system (Slade and Levine 1991). Rather than water affecting chemical reactions via aw or by plasticizing amorphous systems, and considering the inhibitory effect of water as a reaction product of Maillard condensations, water mobility itself may have a direct impact on chemical reactivity in low-moisture and intermediate-moisture solids. Nuclear magnetic resonance (1H NMR) is a powerful technique that allows water and food solids mobility to be studied independently. Although there have been a number of NMR studies on proton relaxation in food systems, its relationship with chemical reactivity has not been analyzed before, and information regarding water mobility and solids mobility in dried potato products is still limited. The present work analyzed the influence of water-solids interactions and water mobility on the kinetics of NEB in freeze-dried potato systems.

Materials and Methods Peeled potatoes were cut into discs, frozen, and freeze-dried for 48 h by using an Alpha 1-4 LD/2-4 LD-2 freeze dryer (Martin Christ Gefriertrocknungsanlagen, Osterode am Harz, Germany). Several potato pieces were powdered and distributed into vials for the determination of water content, thermal transitions, and molecular mobility. The remaining samples were used to determine NEB. Dehydrated samples were equilibrated over saturated salt solutions (in a range of 11%–93% RH) for 13 days at 23°C. Some mixtures were prepared by the addition of distilled water to the potato powder (aw = 0.93–0.98). The water content was determined gravimetrically and the water activity (aw), by using an electronic aw meter. After equilibration, potato discs were placed inside rubber O-rings that in turn were sandwiched between two glass plates held hermetically with metal clamps. The glass sample holder was then placed in an air-convection oven operated at 70° ± 1°C. At suitable intervals, samples were removed from the oven, color was determined, and the samples were placed back in the oven to continue the heat treatment. Glass transitions were determined by differential scanning calorimetry (DSC) (onset values) using a DSC (822e DSC; Mettler Toledo, Schwerzenbach, Switzerland). A Bruker PC 120 Minispec pulsed nuclear magnetic resonance instrument (Bruker AXS, Madison, WI, USA), with a 0.47 T magnetic field operating at resonance frequency of 20 MHz, was used to measure molecular mobility. The spin-spin relaxation time (T2) associated with the fast-relaxing protons was measured by using freeinduction decay (FID) analysis after a single 90° pulse. The T2 associated with the slow-relaxing protons was measured by using the Hahn spin-echo pulse sequence (90°-τ-180°) with an interpulse (τ) range of 0.001–2.0 ms. The degree of Maillard reaction was determined by reflectance measurements of the color-attribute luminosity (L*) with a white background of reflectance (L0*). A

Potato Systems in Relation to Nonenzymatic Browning

439

Figure 33.1. Color development at 70°C in potato discs at different water activity (aw). L*0 − L*, changes in luminosity value.

handheld tristimulus reflectance spectrocolorimeter (CM-508-d; Minolta, Ramsey, NJ, USA) was employed. Color functions were calculated for illuminant D65 at 2° standard observer and in the CIELab uniform color space. An average of six replicates was reported.

Results and Discussion Browning development in freeze-dried potato discs increased as a function of storage time at 70°C (Figure 33.1), and the browning rate increased along the aw scale, showing a maximum value at aw = 0.84. Above aw = 0.84, the browning rate decreased, most likely because of the presence of enough water to inhibit this chemical reaction. Figure 33.2 shows the experimental points of water-sorption isotherm at 23°C for potato powder together with the results of the fitting to the Guggenheim-Anderson-de Boer (GAB) model (van den Berg and Bruin 1981). The isotherm shows a characteristic sigmoid shape. The GAB fit gave a monolayer water content of 6.0% dry basis, which corresponds to about aw = 0.21. According to Timmermann and Chirife (1991), three different stages can be identified in the water-sorption isotherm: a first stage that corresponds to water contents up to the monolayer; a second stage in which the state of water is different from that of the monolayer and of bulk water; and a third stage corresponding to high water activities in which the systems show a water sorption larger than that predicted by the GAB model. The third stage can be identified by means of a linearized form of the GAB equation, which shows a linear trend up to a certain aw value, after which the points deviate downward, showing the onset of the third stage. The linearized form of the GAB equation has been applied to the experimental points of the potato

440

PART 2: Poster Presentations

Figure 33.2. Water-sorption isotherms at 23°C fitted by the Guggenheim-Andersonde Boer (GAB) equation. aw, water activity; and db, dry basis.

powder–sorption isotherm, and additional points at the high aw region were included for better observation (not shown). A deviation from the linear behavior was located at water activities higher than 0.84, suggesting a change in the physical state of adsorbed water; that is, the water molecules would be more liquidlike than in the preceding layers, coinciding with the appearance of freezable water as detected by DSC (not shown), which is also reflected in a decreased browning rate (Figure 33.1). Figure 33.3a shows the T2 relaxation times determined by FID analysis after a single 90° pulse, as a function of temperature. This fast-decaying component, called T2-FID, was attributed to protons from solid polysaccharides and from water molecules that interact strongly with the solid matrix by hydrogen bonding (i.e., monolayer water) (Kalichevsky and others 1992; Le Botlan and Helie-Fourel 1995; Ruan and others 1999). Between 0.11 and 0.52 aw, T2-FID values slightly increased as RH and temperature increased. The potato powder equilibrated at aw = 0.75, 0.84, and 0.93 presented much higher T2-FID values than those obtained at lower aw. Also above aw = 0.52, a marked effect of temperature on T2-FID increase was observed (Figure 33.3a). The increase of T2-FID with water content can be attributed to the plasticizing effect of water, which provides greater mobility to solid protons. Chatakanonda and others (2003) studied the molecular mobility in potato starch and wheat starch, respectively, by 1H NMR and also found that T2 increased as aw increased (aw range, 0–0.97). The break observed in the T2-FID vs temperature curves at aw = 0.75 and 0.84 (Figure 33.3a, arrows) could be related to the glass transition. This break was not observed at lower aw values because the systems were in the glassy state below aw = 0.75 at all the temperatures analyzed. In the samples at aw = 0.93, the break was not observed because the glass transition temperature was lower than 25°C (the lower temperature analyzed).

Potato Systems in Relation to Nonenzymatic Browning

441

(a) 20.5

T2-FID, ms

18.0 15.5

Tg= 36°C 13.0

Tg= 49°C 10.5 8.0 20

30

40

50

60

Temperature, ºC (b) 950

TH2-2, ms

800

650

500

350 0

10

20

30

40

50

Water content, db% Figure 33.3. (a) Spin-spin relaxation time (T2-FID) obtained by a single 90° pulse by proton nuclear magnetic resonance for potato powder as a function of temperature. Samples were equilibrated at different water activities: 0.11 (solid squares), 0.33 (solid triangles), 0.52 (x’s), 0.75 (open squares), 0.84 (asterisks), and 0.93 (open diamonds). Glass transition temperature (Tg) values obtained by differential scanning calorimetry are indicated. (b) Spin-spin relaxation time (T2H− 2 ) obtained by spin-echo Hahn sequence for potato powder as a function of water content; db, dry basis.

The analysis of the Hahn spin-echo data of the protons associated with the water present in the potato powder samples at different RHs showed two T2 components. A short component ( T2H−1) on the order of 30 μs was present at all the water activities analyzed. This faster-decaying component corresponds to protons of solid polysaccharide and of water strongly interacting with solids, as the T2-FID measured by a single

442

PART 2: Poster Presentations

90° pulse. Above aw = 0.22, a long component ( T2H− 2) in the range of 400–950 μs was also observed whose value increased with the increase in water content (Figure 33.3b). The onset of the third sorption stage can be identified by means of the T2H− 2, as the beginning of this stage corresponds to the appearance of a plateau in the T2 values along the water-content scale. Above a water content of 26% (aw = 0.84), T2H− 2 values tended to reach a plateau that corresponded to more “free” water, as detected by DSC, as freezable water (not shown).

Conclusions The results of this work enable the rate of NEB to be related to the degree of interactions between water and solid molecules and with the physical state of the food matrix. At low aw, the water molecules strongly interact with the solid matrix and a water increase enhances the reaction rate. At increasing aw, the relaxation times T2H− 2 increase. Upon the appearance of freezable water and highly mobile water (determined by the third sorption stage and T2H− 2 value), the behavior changes markedly. At this point, the NEB rate decreases because the water molecules can inhibit the reaction and/or dilute the reactants. The T2-FID values displayed a marked increase around the glass transition temperature (close to the values determined by DSC), and 1H NMR proved to be an effective tool for the study of glass transition in potato powders. The influence of water on NEB kinetics (i.e., aw and water plasticization), and the molecular mobility aspects determined by 1H NMR, enabled us to conclude that the reaction rate decreases as the third sorption stage of water appears in the system. An integrated approach leads to conclude that the browning rate increases from the monolayer value, a further increase occurs well above the Tg value, and then a decrease is observed above 0.84, coinciding with the appearance of freezable water and changes in T2H− 2 dependence with aw. The complementary analysis of 1H-NMR relaxations enables one to evaluate each of those stages and to predict the dependence of NEB on aw.

Acknowledgments The authors acknowledge the financial support of Agencia Nacional de Promoción Científica y Tecnológica (PICT 32916), Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET, PIP 5977), and the Universidad de Buenos Aires (EX226).

References Bell LN, Hageman MJ. 1994. Differentiating between the effects of water activity and glass transition dependent mobility on a solid state chemical reaction: aspartame degradation. J Agric Food Chem 42:2398–401. Buera MP, Karel M. 1995. Effect of physical changes on the rates of nonenzymic browning and related reactions. Food Chem 52:167–73.

Potato Systems in Relation to Nonenzymatic Browning

443

Chatakanonda P, Chinachoti P, Sriroth K, Piyachomkwan K, Chotineeranat S, Tang H, Hills B. 2003. The influence of time and conditions of harvest on the functional behavior of cassava starch: a proton NMR relaxation study. Carbohydr Polym 53:233–40. Kalichevsky MT, Jaroszkiewicz EM, Ablett S, Blanshard JMV, Lillford PJ. 1992. The glass transition of amylopectin measured by DSC, DMTA and NMR. Carbohydr Polym 18:77–88. Karmas R, Buera MP, Karel M. 1992. Effect of glass transition on rates of nonenzymatic browning in food systems. J Agric Food Chem 40:873–9. Le Botlan D, Helie-Fourel I. 1995. Assessment of the intermediate phase in milk fat by low-resolution nuclear magnetic resonance. Anal Chim Acta 311:217–23. Namiki M. 1988. Chemistry of the Maillard reaction: recent studies on the browning reaction mechanism and the development of antioxidants and mutagens. Adv Food Res 32:115–84. Ruan R, Wang X, Chen PL, Fulcher RG, Pescheck P, Chakrabarti S. 1999. Study of water in dough using nuclear magnetic resonance. Cereal Chem 76:231–5. Slade L, Levine H. 1991. Beyond water activity: recent advances based on an alternative approach to the assessment of food quality and safety. Crit Rev Food Sci Nutr 30:115–60. Timmermann EO, Chirife J. 1991. The physical state of water sorbed at high activities in starch in terms of the GAB sorption equation. J Food Eng 13:171–9. Van den Berg C, Bruin S. 1981. Water activity and its estimation in food systems. In: Rockland LB, Stewart F, editors. Water activity: influence on food quality. New York: Academic. p 147–77.

34 Nonenzymatic Browning Reaction and Enthalpy Relaxation of Glassy Foods K. Tsuji, K. Kawai, M. Watanabe, and T. Suzuki

Abstract The progress of the nonenzymatic browning (NEB) reaction is observed significantly in glassy-state foods; NEB causes various quality changes (in flavor, color, and nutritional value). In addition, it is known that enthalpy relaxation progresses gradually during storage; causing various physical changes (excess volume and water-sorption ability). Although it has been assumed that degradation rate depends on the molecular mobility of the glassy foods, there have been few kinetic studies of this degradation. Model glassy foods were prepared by freeze drying 10% (wt/wt) aqueous solutions of glucose-lysine-trehalose (1 : 1 : 98 dry-weight basis). Moisture content and glassrubber transition temperature (Tg) of the samples were determined by Karl Fischer titration and by differential scanning calorimetry (DSC), respectively. The samples were stored at temperatures below Tg (35°–70°C). After the storage, the relaxed enthalpy (ΔHrelax) of the samples was evaluated by DSC. The samples were rehydrated with distilled water, and then the extent of NEB was evaluated spectrophotometrically by absorbance at 280 nm (ABS280). The samples stored at 70°C showed a clear enthalpy relaxation, and the value of ΔHrelax increased with increases in storage temperature and time. It was found that the rate of enthalpy relaxation was much lower than that of NEB.

Introduction The nonenzymatic browning (NEB) reaction is one of the most important chemical reactions in food processing and preservation because the progress of NEB causes various quality changes (in flavor, color, and nutritional value) in food products. NEB is a major factor in the stability of stored dry foods. It is known that most dry foods are in an amorphous state, and that their chemical and physical stabilities are affected by glass-rubber transition. It is expected that degradation rate is restricted in the glassy state because of its extremely low molecular mobility. In fact, previous studies demonstrated a significant effect of glass transition on the NEB rate of model amorphous food products; the dependence of temperature on the NEB rate drastically changed at glass transition temperature (Tg) (Karmas and others 1992; Roos and Himberg 1994; Buera and Karel 1995; Lievonen and others 1998; Maltini and others 2003). NEB progress, however, has been observed even 445

446

PART 2: Poster Presentations

at temperatures below Tg (Schebor and others 1999). It is, therefore, of practical, as well as fundamental interest to prevent the progress of NEB in glassy food products. Lately there has been considerable interest in the enthalpy relaxation process in glassy materials in the biological, pharmaceutical, and food industries because enthalpy relaxation time relates to the macroscopic molecular mobility of glass (Hancock and others 1995; Duddu and others 1997; Kawai and others 2005a). Understanding the enthalpy relaxation time is expected to be useful in predicting its long-term stability. For example, the aggregation rate of protein embedded in the glassy matrices was suggested to be predicted by their enthalpy relaxation time (Duddu and others 1997). However, very few studies have been reported on the relationship between the NEB rate and the enthalpy relaxation time of food products in the glassy state. Therefore, this study investigated the progress of the NEB reaction and the enthalpy relaxation of a model glassy-state food sample during longterm storage.

Materials and Methods Sample Preparation A 10% (wt/wt) aqueous solution of glucose-lysine-trehalose (1 : 1 : 98, dry-weight basis) was prepared, and a 2-mL fraction of each solution was placed into a 10-mL glass vial. The solution was frozen for 24 h at −40°C and transferred to a precooled freeze dryer (RLE-52; Kyowa Shinku Gijyutsu [Kyowa Vacuum Technology], Tokyo, Japan). The frozen formulation was freeze-dried at 3.0 × 10−2 Torr with warming from −40° to 25°C over a 3-day period. The residual moisture of the freeze-dried solid was removed by storage over phosphorus pentoxide in a vacuum desiccator for 7 days at room temperature. The moisture content of the obtained samples was confirmed to be below 0.5% (wt/wt). Nonenzymatic Browning Reaction The samples were stored at 50°, 60°, and 70°C over a 14-day period. After the storage, the samples were rehydrated to 0.04 g sample/mL with distilled water, and the extent of NEB was investigated spectrophotometrically by absorbance at 280 nm (ABS280) (U-1100; Hitachi, Tokyo, Japan). The measurement was performed in triplicate and averaged. Glass Transition and Enthalpy Relaxation Glass transition and enthalpy relaxation properties were investigated by differential scanning calorimetry (DSC) (DSC-50; Shimadzu, Kyoto, Japan). Alumina powder was used as a reference material. The temperature and heat flow by DSC were calibrated with indium and distilled water. Each sample (∼10 mg) was placed into an aluminum pan, and then measured by DSC at 5°C/min in the temperature range of 25°–150°C.

Nonenzymatic Browning Reaction and Enthalpy Relaxation of Glassy Foods

447

6

Relaxed enthalpy (J/g)

5

τΗ = 4520 h β = 0.75

4 glass transition

3

shoulder 0h 43 h

2

50mwt/g

969 h

enthalpy relaxation

1 40

60 80 100 120 Temperature (ºC)

0 0

200

400 600 800 Storage time (h)

1000

1200

Figure 34.1. Time courses of ΔHrelax for the samples stored at 70°C. The solid curve was obtained by fitting Equation 34.1 to the data. The insert shows typical differential scanning calorimetric thermograms with or without storage at 70°C.

Results Glass Transition and Enthalpy Relaxation Typical DSC thermograms for the trehalose-glucose-lysine system are shown in the insert in Figure 34.1. The nonstored sample showed an endothermic shift due to glass transition at 92°C. The Tg of the sample was slightly lower than that of pure trehalose (Tg = 114°C) because of the plasticizing effect of glucose and lysine. The sample stored for up to 120 h at 70°C initially showed an endothermic shoulder before the glass transition; this shoulder is often observed at an initial stage in the enthalpy relaxation (Pappin and others 1994). With the progress of enthalpy relaxation, the shoulder evolves into an endothermic peak located in the glass transition region, and relaxed enthalpy (ΔHrelax) was evaluated from the area of this peak. The samples stored at 50° and 60°C showed only an endothermic shoulder over the storage period because the molecular mobility of the samples would be too low to permit an enthalpy relaxation. Enthalpy Relaxation Time of Glassy Trehalose-Glucose-Lysine System The time course of development of ΔHrelax for the sample stored at 70°C is shown in Figure 34.1. It is known that the enthalpy relaxation process in glassy materials does not follow a simple exponential decay. To analyze the enthalpy relaxation process,

448

PART 2: Poster Presentations

two different approaches have been used successfully: the Kohlrausch-WilliamsWatts (KWW) approach (Hancock and others 1995; Kawai and others 2005a) and the Adam-Gibbs (AG) approach (Yamamuro and others 1998; Kawai and others 2005a). The KWW method expresses stretched exponential decay by assuming the distribution of the enthalpy relaxation time (Equation 34.1). β

1 − Δ H relax Δ H ∞ = exp ⎡⎣ − ( t τ H ) ⎤⎦

(34.1)

where t, τH, and β are storage time, mean average enthalpy relaxation time, and nonexponential parameter, respectively. It is understood that β relates to the distribution of the enthalpy relaxation time (Hancock and others 1995; Kawai and others 2005a), and its value changes between 0 (untidy) and 1 (unity). ΔH∞ is the maximum relaxed enthalpy, which can be expressed by Equation 34.2: Δ H ∞ = ΔCp (Tg − T )

(34.2)

where ΔCp and T are heat capacity change at Tg and storage temperature, respectively. On the other hand, the AG method assumes that the molecular numbers involved in the region undergoing cooperative rearrangement increase with the enthalpy relaxation, and thus the enthalpy relaxation time increases with an increase in ΔHrelax. The AG approach, however, could not be used in this study because more ΔHrelax data between nonequilibrium and equilibrium were required in order to apply the approach successfully. So, only the KWW approach was used, and the values of τH and β were evaluated as fitting parameters as listed in Figure 34.1. Progress of NEB in Glassy Trehalose Matrix Figure 34.2 shows the time courses of ABS280 for trehalose-glucose-lysine systems stored at each temperature. The ABS280 increased with increases in storage temperature and time. In previous studies, the initial NEB rate was evaluated as a pseudo-zeroorder reaction rate employing the initial time course of ABS280 (Roos and Himberg 1994). The time course of ABS280, deviated from the zero-order formula at longer times. To evaluate NEB rates and compare the rates of the enthalpy relaxation and the NEB, the KWW formula (Equation 34.3) was developed: β

1 − ABS280 ABS∞ = exp ⎡⎣ − ( t τNEB) ⎤⎦

(34.3)

where ABS∞, τNEB, and β are assumed to be a maximum value of ABS280, the reciprocal of the NEB rate, and the distribution of τNEB, respectively. Equation 34.3 was fitted to the time course of ABS280, and ABS∞, τNEB, and β were evaluated as fitting parameters as listed in Figure 34.2. The β of τNEB increased from 0.79 to 1.0 with storage temperature decrease. Based on the KWW approach, the distribution of the NEB rate decreases with a storage temperature decrease.

Nonenzymatic Browning Reaction and Enthalpy Relaxation of Glassy Foods

449

9 at 70ºC τNEB = 90 h β = 0.79

8 7

ABS280

6

at 60 ºC τNEB = 450 h β = 0.98

5 4

at 50 ºC τNEB = 2000 h β = 1.0

3 2 1 0 0

100

200

300

400

500

Storage time (h)

Figure 34.2. Time courses of absorbance at 280 nm (ABS280). The solid curves were obtained by fitting Equation 34.3 to the data. The maximum value of ABS280 (ABS∞) at each temperature was determined to be 8.5 from the fitting result of the sample stored at 70°C.

Discussion When τH and τNEB of the sample stored at 70°C were compared, τH (4520 h) was much longer than τNEB (90 h). In addition, although the progress of the enthalpy relaxation could not be observed at 50° or 60°C over the storage period, the values of τNEB were low still (450 h at 60°C and 2000 h at 50°C). On the other hand, when the β of the enthalpy relaxation and the β of the NEB were compared, the β of τNEB (0.79–1.00) tended to be greater than that of τH (0.75). Based on these results, the progress of the NEB was much greater and more homogeneous than that of the enthalpy relaxation. This means that the progress of NEB is at least partially independent of the molecular mobility of the glassy matrix. The results of our previous study (Kawai and others 2004, 2005b) suggested that the formation of intermolecular hydrogen bonds among amino acids, reducing sugars, and the glassy matrix would cause a “local reaction” promoting the progress of the NEB reaction in the matrix. When intermolecular hydrogen bonds are formed between the glassy matrix and each NEB reactant (amino acid and reducing sugar), the NEB reactant is forced to break the hydrogen bonds in the molecular rearrangement process. Thus, the NEB progress is prevented efficiently by the glassy matrix. Some NEB reactants, however, form intermolecular hydrogen bonds between amino acids and reducing sugars (Figure 34.3a). NEB reactants in these states can react smoothly without requiring molecular rearrangement in the glassy matrix. This is understood as a local reaction because the NEB progress is independent from the entrapment of the glassy matrix.

450

PART 2: Poster Presentations

(a) Before storage (at 25 °C)

(b) Storage at 70 °C

hydrogen bonds between lysine and glucose

(c) Storage at 50 °C and 60 °C Hydrogen bonds between trehalose and lysine or glucose glucose lysine trehalose matrix hydrogen bond NEB product

Figure 34.3. Model of local mobility promoting the progress of the nonenzymatic browning reaction in the glassy matrix: (a) Glucose and lysine are embedded in the glassy trehalose matrix. (b) After storage at 70°C. (c) After storage at 50° and 60°C. NEB, nonenzymatic browning.

Taking the suggestion into account, the results in this study are interpreted as follows. Since localized and/or hydrogen-bonded glucose-lysine reacted smoothly without molecular rearrangement of the glassy trehalose matrix, the NEB progress was much greater compared with the molecular mobility of the glassy trehalose matrix. Furthermore, since the glassy trehalose matrix had relative high molecular mobility at 70°C (the progress of the enthalpy relaxation was significant during the storage period), glucose and lysine entrapped by the matrix also reacted slowly (Figure 34.3b). As a result, the sample stored at 70°C showed a relatively wide distribution of NEB rate (β of τNEB = 0.79). On the other hand, since the glassy trehalose matrix had low molecular mobility at 50° and 60°C (the enthalpy relaxation was not seen during the storage period), localized and/or hydrogen-bonded glucose-lysine reacted mainly in the matrix (Figure 34.3c). Therefore, the samples stored at 50° and 60°C showed a narrow distribution of NEB rate (β of τNEB = 0.98–1.00). In conclusion, it was found that the progress of the NEB was much higher and more homogeneous than that of the enthalpy relaxation. A possible extension of this research would involve clarification of the local reaction caused by the intermolecular hydrogen bonds in a glassy matrix.

References Buera MP, Karel M. 1995. Effect of physical changes on the rates of nonenzymic browning and related reactions. Food Chem 52:167–73.

Nonenzymatic Browning Reaction and Enthalpy Relaxation of Glassy Foods

451

Duddu SP, Zhang G, Dal Monte PR. 1997. The relationship between protein aggregation and molecular mobility below the glass transition temperature of lyophilized formulations containing a monoclonal antibody. Pharm Res 14:596–600. Hancock BC, Shamblin SL, Zografi G. 1995. Molecular mobility of amorphous pharmaceutical solids below their glass transition temperature. Pharm Res 12:799–806. Karmas R, Buera MP, Karel M. 1992. Effect of glass transition on rate of nonenzymatic browning in food system. J Agric Food Chem 40:873–9. Kawai K, Hagiwara T, Takai R, Suzuki T. 2004. Maillard reaction rate in various glassy matrices. Biosci Biotechnol Biochem 68:338–42. Kawai K, Hagiwara T, Takai R, Suzuki T. 2005a. Comparative investigation by two analytical approaches of enthalpy relaxation for glassy glucose, sucrose, maltose and trehalose. Pharm Res 22:490–5. Kawai K, Hagiwara T, Takai R, Suzuki T. 2005b. The rate of non-enzymatic browning reaction in model freeze-dried food system under the glassy state. Innov Food Sci Emerg Technol 6:346–50. Lievonen SM, Laaksonen TJ, Roos YH. 1998. Glass transition and reaction rates: nonenzymatic browning in glassy and liquid system. J Agric Food Chem 46:2778–84. Maltini E, Torreggiani D, Venir E, Bertolo G. 2003. Water activity and the preservation of plant food. Food Chem 82:79–86. Pappin AJ, Hutchinson JM, Ingram MD. 1994. The appearance of annealing pre-peaks in inorganic glasses: new experimental results and theoretical interpretation. J Non-Cryst Solids 172–174:584–91. Roos YH, Himberg J. 1994. Nonenzymatic browning behavior, as related to glass transition of a food model at chilling temperatures. J Agric Food Chem 42:893–8. Schebor C, Buera MP, Karel M, Chirife J. 1999. Color formation due to non-enzymatic browning in amorphous, glassy, anhydrous, model system. Food Chem 65:427–32. Yamamuro O, Oishi Y, Nishizawa M, Matsuo T. 1998. Enthalpy relaxation of glassy glycerol prepared by rapid liquid quenching. J Non-Cryst Solids 235–237:517–21.

35 Film-Forming Ability of Duck Egg White and Its Water-Vapor Barrier Property W. Garnjanagoonchorn, A. Yimjaroenpornsakul, N. Poovarodom, and S. Praditdoung

Abstract Duck egg white is a Thai dessert manufacturing by-product that is mostly used in animal feed. We studied the film-forming ability and water-vapor barrier property of duck egg-white protein. We investigated the effects of egg freshness, egg-white solution pH, and concentration of plasticizers on edible film properties. The results showed that egg white from fresh duck eggs kept at 28° ± 3°C for no longer than 4 days formed edible film when pH was adjusted to 10.5 and 50% (wt/wt of egg protein) sorbitol was added as a plasticizer. The film was clear, transparent, and smooth, and proved to be a good water-vapor barrier, with a water-vapor permeability (WVP) of 0.098 (g · mm)/(m2 · h · mmHg). Increasing the pH to 11.25 increased the WVP of the film.

Introduction Edible films and coatings are the answer for future food packaging. Research on edible films for over 40 years has developed films with various physical and gas barrier properties from biopolymers. Protein films have been developed from wheat gluten, soy protein, corn zein, meat protein, collagen, and hen egg white. Studies of eggalbumen films have been reported by Gennadios and others (1996) and Handa and others (1999). Desugared spray-dried egg-white solids from hen eggs were used in these experiments. The effects of plasticizers, the partial substitution of egg-yolk solids for egg white, the pH of aqueous egg-white solutions, and heating of egg-white solutions prior to casting on the physical properties of egg-white films were determined. Egg-albumen films showed water-vapor permeability (WVP) values in the same range as films of other proteins (Gennadios and others 1996). Duck egg white is a manufacturing by-product from the production of some Thai desserts. Differences in the ratios of amino acid composition in hen, and duck egg white have been reported (DuckEggs.com). Since Gennadios and Weller (1990) have pointed out that amino acids play an important role in protein film–forming properties, it is appropriate to develop edible films from fresh duck egg white to determine whether the amino acid composition might influence the film properties. We also investigated the effects of egg freshness, the pH of egg-white solutions, and the plasticizer content on egg-white film properties. 453

454

PART 2: Poster Presentations

Materials and Methods Egg Freshness Fresh duck shell eggs brought from a farm were stored at 28° ± 3°C for 4, 11, and 18 days before being prepared for film casting. Egg freshness was determined from 3 eggs by using a Haugh gauge to measure Haugh units. Viscosity The viscosity of egg-white solutions was determined by using a Brookfield model RVDV III viscometer (Brookfield Engineering Laboratories, Middleboro, MA, USA) with a UL adapter, spindle number 00, at speed 200 rpm, spindle time 1 min. Film Preparation Egg white was separated from shell eggs and strained through cheesecloth. The pH of the egg-white solution was determined. Film-forming solutions were prepared by adding silicone oil at 0.1% wt/wt of egg-white protein, stirred for 5 min, and then de-aerated by ultrasonic bath for 20 min. Sorbitol was added as a plasticizer at 50% and 60% wt/wt of protein. After stirring, the solution pH was adjusted to 10.5–11.25 with 5 N sodium hydroxide. Egg-white solutions were heated in a water bath at 45°C for 20 min and then cooled to 28° ± 3°C. Solutions were cast on flat glass plates that had been covered with linear low-density polyethylene. Film thickness was controlled by casting the same amount of solution (25 mL) on each plate (20 × 30 cm). Castings were kept at 28° ± 3°C for 8 h to solidify and then placed in a chamber (25°C, 50% relative humidity) for 24 h to complete the drying. Films were peeled from plates, cut to appropriate sizes, and incubated at 28° ± 3°C and 65% ± 2% relative humidity prior to testing. Physical Properties Film thickness was determined by using a Mitutoyo Digimatic Thickness Gauge (Mitutoyo America, Aurora, IL, USA). Tensile strength (TS) and elongation at break (% E) were determined according to ASTM D882-91 (ASTM 1992). Color values of prepared films were measured as Hunter L, a, b with a HunterLab MiniScan XE (HunterLab, Reston, VA, USA) equipped with a D65, 10° angle. Barrier Properties The WVP of films was measured according to ASTM E96-85 (ASTM 1996). Oxygen permeability of films was determined according to ASTM D1434-82 (ASTM 1988). Statistical Analysis A 2 × 2 × 2 factorial in randomized complete block experimental design was used. The independent variables were egg freshness, pH of the film-forming solution, and plasticizer content. Two replications were carried out. Analysis of variance and the Duncan multiple-range test were performed, setting a significant level of 95%.

Film-Forming Ability of Duck Egg White and Its Water-Vapor Barrier Property

455

Table 35.1. Haugh unit, viscosity, and pH of duck egg white stored at room temperature (28° ± 3°C) for 18 days Time (days)

Haugh unit

Viscosity (cP)

pH

4

73.41 ± 7.49

12.87 ± 0.29

9.46 ± 0.03

11

40.70 ± 12.06

12.02 ± 0.46

9.97 ± 0.08

18

28.39 ± 12.06

7.82 ± 0.67

10.05 ± 0.09

Results and Discussion Egg Freshness Duck eggs stored at 28° ± 3°C for 4 days showed a Haugh unit of 73.14, a viscosity of 12.87 cP, and pH of 9.46 (Table 35.1). The Haugh unit of duck eggs decreased sharply during storage, indicating the rapid loss of egg freshness when stored at room temperature. The viscosity of egg white remained ∼12 cP after 11 days of storage but had decreased almost 50% after 18 days at 28° ± 3°C. The separation of ovomucinlysozyme complexes at a pH close to 10 (lysozyme pI = 10) is considered responsible for the thinning of egg albumen during storage (Stadellman and Cotterill 1995). The pH of egg white increased with increasing storage time and reached pH 10.05 after 18 days because of a loss of carbon dioxide through the shell pores (Powrie 1995). According to the US Department of Agriculture (USDA) egg-grading manual, the higher the Haugh unit is, the better is the albumen quality of the egg. The good-quality hen eggs showed Haugh units of 73–100 (USDA 2000). In Thailand, it is common to keep shell eggs at room temperature, where the quality remains good over 4 days of storage, as indicated in this study. Film Formation Viscosities of egg-white solutions were determined after the addition of two levels of sorbitol (50% or 60% wt/wt) and on pH adjustment to either pH 10.5 or pH 11.25. Storage time of shell eggs, pH, and sorbitol content showed significant effects (P < 0.05) on the viscosity of egg-white solutions. The viscosity decreased with the increasing storage time of duck eggs (Table 35.2). Egg-white films made from 4-dayold eggs appeared smooth and homogeneous. Increasing the storage time of duck eggs caused egg-white solutions to form brittle film at both pH 10.5 and 11.25 with 50% and 60% sorbitol. This is due to the denaturation of egg-white protein during storage as alkaline conditions reduce disulfide bonds, allowing complete protein dispersion, while disulfide bonds re-form through air oxidation as the film dries (Gennadios and others 1996). Although sorbitol helps in softening the film, which reduces its brittleness, the addition of 50%–60% sorbitol is not enough when forming films from 11- or 18-day-old egg white. Color Properties Color values of duck egg-white films from 4- to 18-day-old eggs showed slight changes in lightness (L: 88–89). The greenness and yellowness increased in films prepared from 11- and 18-day-old eggs (Table 35.3).

456

PART 2: Poster Presentations

Table 35.2. Viscosity and film-forming ability of duck egg white separated from shell eggs that were kept at 28° ± 3°C for 4, 11, and 18 days Time (days)

% Sorbitol

Solution pH

Viscosity (cP)

50

10.50

20.63 ± 0.71d

Smooth clear film

50

11.25

25.40 ± 0.77f

Smooth clear film

60

10.50

24.40 ± 1.35e

Smooth clear film

60

11.25

24.68 ± 0.82ef

Smooth clear film

50

10.50

9.45 ± 0.31b

Partly brittle film

50

11.25

13.15 ± 0.52c

60

10.50

60

11.25

50

10.50

8.77 ± 0.47ab

Very brittle film

50

11.25

8.80 ± 0.57ab

Very brittle film

60

10.50

8.72 ± 0.51ab

Brittle film

60

11.25

8.39 ± 0.48ab

Clear film difficult to peel from plate

4

11

18

8.77 ± 0.44ab 13.58 ± 0.31c

Appearance

Brittle film Smooth clear film Brittle film

a, b, c, etc.: The difference in any two means in same column that are followed by same letter was not significant (P < 0.05). Table 35.3. Color values of duck egg-white film made from shell eggs that were kept at 28° ± 3°C for 4, 11, and 18 days Time (days) 4

11

18

Sorbitol (% wt/wt)

pH

Color value

50

10.50

89.36 ± 0.22d

−0.92 ± 0.04c

1.71 ± 0.34a

50

11.25

89.34 ± 0.32d

−0.91 ± 0.10bc

1.83 ± 0.28ab

60

10.50

89.10 ± 0.23bcd

−0.88 ± 0.06c

1.95 ± 0.15abc

60

11.25

89.26 ± 0.15cd

−0.90 ± 0.02c

2.08 ± 0.18abcd

50

10.50

88.66 ± 0.30a

−0.93 ± 0.04bc

2.35 ± 0.39d

50

11.25

88.90 ± 0.34ab

−0.91 ± 0.04c

2.32 ± 0.19cd

60

10.50

89.03 ± 0.30bcd

−0.92 ± 0.07bc

2.33 ± 0.35cd

60

11.25

88.91 ± 0.24ab

−0.90 ± 0.03c

2.13 ± 0.20bcd

50

10.50

89.11 ± 0.21bcd

−0.99 ± 0.07ab

2.32 ± 0.04cd

50

11.25

88.90 ± 0.26ab

−1.03 ± 0.08a

2.15 ± 0.40bcd

60

10.50

88.96 ± 0.18abc

−0.94 ± 0.06bc

2.16 ± 0.19bcd

60

11.25

88.94 ± 0.30abc

−0.96 ± 0.09abc

2.30 ± 0.22cd

L

a

b

a, b, c, etc., in the table body: The difference in any two means in same column that are followed by same letter was not significant (P < 0.05). In the table column heads, L, lightness; a, greenness; and b, yellowness.

Mechanical Properties The TS (tensile strength) is the maximum tensile stress a material can sustain, whereas E (elongation) indicates film flexibility and stretchability (Gennadios and others 1996). Values of TS and E for egg-white films plasticized with 50% and 60% sorbitol with pH adjusted to 10.5 and 11.25, prepared from 4- to 18-day-old duck eggs, were compared (Table 35.4). The storage time of eggs, pH of solution, and sorbitol content

Film-Forming Ability of Duck Egg White and Its Water-Vapor Barrier Property

457

Table 35.4. Mechanical properties of duck egg-white film made from shell eggs that were kept at 28° ± 3°C for 4, 11, and 18 days Time (days) 4

11

18

Sorbitol (% wt/wt)

pH

Film thickness (mm)

Tensile strength (kg/mm2)

50 50 60 60

Elongation at break (%)

10.50

0.113

0.57 ± 0.01d

6.92 ± 0.47d

11.25

0.121

0.70 ± 0.04f

12.81 ± 0.47f

10.50

0.116

0.40 ± 0.01b

19.08 ± 0.43g

11.25

0.118

0.58 ± 0.01d

22.14 ± 0.95h

50

10.50

0.121

0.62 ± 0.04e

3.41 ± 0.39b

50

11.25

0.123

0.71 ± 0.04f

2.71 ± 0.26a

60

10.50

0.124

0.46 ± 0.04c

5.42 ± 0.57c

60

11.25

0.125

0.56 ± 0.01d

9.28 ± 0.15e

60

10.50

0.122

0.28 ± 0.01a

5.85 ± 0.35c

60

11.25

0.122

0.28 ± 0.09a

9.59 ± 0.29e

a, b, c, etc.: The difference in any two means in same column that are followed by same letter was not significant (P < 0.05).

showed significant effects (P < 0.05) on TS and E. The lowest tensile strengths were obtained from film of 18-day-old eggs. Since water vapor is lost through shell pores during storage (Stadellman and Cotterill 1995), the egg-white solution from 18-dayold eggs contained more concentrated egg-white protein. Thus, during formation, film can form more links between protein molecules, resulting in increased brittleness, which reduces TS. The TS decreased and E increased with increases in sorbitol content, which agrees with results reported by Gennadios and others (1996). The TS and E increased with a pH increase from 10.5 to 11.25. Egg-white film prepared from 4-day-old duck eggs showed a TS in the range of 0.4–0.7 kg/mm2 and an E of 6.92–22.14, depending on the pH of the egg-white solution and the sorbitol content (Table 35.4). Barrier Properties Storage time of shell eggs, the pH of the egg-white solution, and the sorbitol content had significant (P < 0.05) effects on the WVP values of duck egg-white films (Table 35.5). The WVP values increased with increasing storage time and sorbitol content. Increasing of film brittleness may cause the WVP values to increase. Film prepared from 4-day-old egg white plasticized with 50% wt/wt sorbitol and pH adjusted to 10.5 showed the lowest WVP (0.098 (g · mm)/(m2 · h · mmHg) and the lowest oxygen permeability (7.04 (mL · mm)/(m2 · h · atm). WVP values of duck egg-white film in this study are slightly lower than hen egg-white film prepared by Gennadios and others (1996).

Conclusion Films can be prepared from fresh duck egg white that has been stored for not longer than 4 days at 28° ± 3°C. Mechanical and gas barrier properties of such films changed with differing levels of plasticizer and pH. Films were smooth and transparent with

458

PART 2: Poster Presentations

Table 35.5. Gas barrier properties of duck egg-white film made from shell eggs that were kept at 28° ± 3 °C for 4 and 11 days Time (days) 4

11

Sorbitol (% wt/wt)

pH

Thickness (mm)

WVP

P(O2)

50

10.50

0.113

0.098 ± 0.002a

7.04 ± 2.66a

50

11.25

0.121

0.115 ± 0.002b

13.90 ± 4.93b

60

10.50

0.116

0.121 ± 0.002c

13.61 ± 2.48b

60

11.25

0.118

0.118 ± 0.002b

22.41 ± 2.52c

50

10.50

0.121

0.118 ± 0.003b

—

50

11.25

0.123

0.141 ± 0.002d

—

60

10.50

0.124

0.159 ± 0.003e

—

60

11.25

0.125

0.143 ± 0.002d

—

a, b, c, etc.: The difference in any two means in same column that are followed by same letter was not significant (P < 0.05). —, film was broken and could not be used for P(O2) measurement; WVP, water-vapor permeability (mL · mm)/(m2 · h · atm); and P(O2), oxygen permeability (mL · mm)/(m2 · h · atm).

WVP values slightly lower than for hen egg-white protein films prepared from spraydried egg-white solids.

Acknowledgment This study was supported financially by the Kasetsart University Research and Development Institute Fund.

References ASTM. 1988. D 1434-82: standard test method for determining gas permeability characteristic of plastic film and sheeting. Annual Philadelphia: ASTM. ASTM. 1992. ASTM D 882-91: standard test method for tensile properties of thin plastic sheeting. Philadelphia: ASTM. ASTM. 1996. ASTM E96-85: standard test method for water vapor transmission of materials. Philadelphia: ASTM. DuckEggs.com. Duck egg–chicken egg nutrient comparison per 100 grams of edible portion. Available from: http://www.duckeggs.com/duck-egg-nutrition-compare.html. Gennadios A, Weller CL. 1990. Edible films and coatings from wheat and corn protein. Food Technol 44:63–9. Gennadios A, Weller CL, Hanna MA, Froning GW. 1996. Mechanical and barrier properties of egg albumen films. J Food Sci 61:585–9. Handa A, Gennadios A, Froning GW, Kuroda N, Hanna MA. 1999. Tensile, solubility, electrophoretic properties of egg white films as affected by surface sulfhydryl groups. J Food Sci 64:82–5. Powrie WD. 1995. Chemistry of eggs and egg products. In: Stadellman WJ, Cotterill OJ, editors. Egg science and technology. 4th ed. New York: Food Produce. p 72–3. Stadellman WJ, Cotterill OJ. 1995. Egg science and technology. 4th ed. New York: Food Produce. US Department of Agriculture (USDA). 2000. Egg-grading manual. Agricultural handbook 75. Washington, DC: USDA. Revised July 2000. Available from: http://www.ams.usda.gov/AMSv1.0/getfile?dDocNam e=STELDEV3004502.

36 Water-Vapor Permeability of Chitosan and Methoxy Poly(ethylene glycol)-bpoly(ε-caprolactone) Blend Homogeneous Films N. Niamsa, N. Morakot, and Y. Baimark

Abstract Water-soluble methoxy-poly(ethylene glycol)-b-poly(ε-caprolactone) diblock copolymer (MPEG-b-PCL) was blended for increasing water-vapor sensitivity of chitosan film by solution blending before film casting. The 1% (vol/vol) acetic acid aqueous solution was used as a blend solvent. The interaction between chitosan and MPEGb-PCL was investigated with Fourier transform infrared spectroscopy. Results of scanning electron microscopy suggested that the blend films were homogeneous. Blending with MPEG-b-PCL can enhance the water-vapor resistance of the chitosan blend film compared to the chitosan-only film. Water-vapor permeation of blend films decreased with increasing MPEG-b-PCL ratios.

Introduction Chitosan, which is produced by N-deacetylation of chitin, is a biopolymer whose biodegradability and biocompatibility have led to great attention regarding its use in a variety of applications (Ravi Kumar and others 2004; Muzzarelli and Muzzarelli 2005). Several studies on the potential of chitosan for biodegradable food-packaging applications especially as edible films and coatings have been published, for example, regarding the storage of fruits (Ghaouth and others 1991; Kittur and others 2002) and seafood products (Jeon and others 2002), because chitosan has good film-forming properties (Kittur and others 1998; Balua and others 2004). A property of chitosan film that makes it unpopular for packaging applications, however, is its high sensitivity to water, since it has a large number of hydrogen bonds (Olabarrieta and others 2001). To try to improve the water-resistant properties of chitosan films, a blending method was used, adding some hydrophobic biodegradable polyesters, such as poly(3-hydroxybutyric acid) (Ikejima and Inoue 2000), poly(εcaprolactone) (PCL) (Olabarrieta and others 2001; Senda and others 2002), and polylactide (PL) (Suyatma and others 2004; Sebastien and others 2006). However, these films are microcomposites whose properties, such as mechanical properties, may be inconsistent. In this study, water-soluble methoxy-poly(ethylene glycol)-b-poly(ε-caprolactone) (MPEG-b-PCL) was synthesized. A series of blend homogeneous films with different chitosan/MPEG-b-PCL blend ratios was prepared by a solution-blending method. The 459

460

PART 2: Poster Presentations

intermolecular bonds between chitosan and MPEG-b-PCL and the water-vapor permeability (WVP) of the blend films were investigated as a function of blend ratios.

Materials and Methods Materials Chitosan (90% deacetylation; molecular weight, 80,000 g/mol) was purchased from Seafresh Chitosan (Bangkok, Thailand) and used as received. The MPEG-b-PCL (molecular weight, 5800 g/mol) was synthesized by ring-opening polymerization of caprolactone (CL) monomer (99%) (Acro Organics, Morris Plains, NJ, USA) at 130°C for 24 h. The MPEG (Fluka, Buch, Germany) and stannous octoate (95%) (Sigma, St. Louis, MO, USA) were used as the initiating agents. A stannous octoate concentration of 0.02 mol% was chosen. The MPEG-b-PCL obtained was purified by melting at 120°C in vacuo and dissolved in distilled water before filtering. The water-soluble MPEG-b-PCL solution was freeze-dried. A yield of ∼80% purified water-soluble MPEG-b-PCL was obtained. The MPEG/CL molar ratio of 1:7 was calculated from proton nuclear magnetic resonance (1H NMR) spectrum. The number-average molecular weight of MPEG-bPCL, calculated from the MPEG/CL molar ratio from 1H-NMR analysis on the basis of MPEG molecular weight, was found to be 5800 g/mol. Blend Films Preparation The 1% (wt/vol) chitosan solution was prepared by using 1% (vol/vol) acetic acid aqueous solution as the solvent. The chitosan/MPEG-b-PCL blend solutions with different blend ratios were prepared by dissolving respectively 20, 40, and 60 mg of MPEG-b-PCL in 18, 16, and 14 mL of 1% (wt/vol) chitosan solutions to obtain the chitosan/MPEG-b-PCL blend ratios of 90 : 10, 80 : 20, and 70 : 30 (wt/wt). The blend solutions were adjusted to 20 mL with distilled water and vigorous stirring for 6 h before film casting on Petri dishes. The blend solutions were dried at 40°C for 48 h before being dried in vacuo at room temperature for a week. Chitosan and Blend Films Characterization Fourier transform infrared spectroscopy (FTIR) spectra of the samples were collected by using a PerkinElmer Spectrum GX Series Fourier transform infrared spectrophotometer (Waltham, MA, USA). A resolution of 2 cm−1 was used for 32 scans. Film morphology was studied with a JSM-6460LV scanning electron microscope (JEOL, Tokyo, Japan). The film was cut with a paper cutter and sputter-coated with gold prior to examination. The WVP of the films was investigated by using the method described by Rutnakornpituk and Ngamdee (2006). The sample films were tightly adhered to the top rim of glass vials with an approximate volume 24 cm3 that had been filled either with a preweigh amount of anhydrous calcium chloride, or as the control, an amount of small glass beads of approximately the same weight as that of the anhydrous calcium chloride. The vials were kept in a desiccator with 90% ± 5% relative humidity

Water-Vapor Permeability of Chitosan Blend Homogeneous Films

461

maintained with a saturated sodium chloride solution at 30° ± 2°C. The vials were weighed again after 14 days in the desiccator. The WVP rate was calculated as follows: Rate of water-vapor permeability ( g day liter ) = [(Tf − Ti ) − ( Cf − Ci )] × 1000 (14v )

(36.1)

where Ti and Tf are the initial and final weights (g) of the sample vials, respectively; Ci and Cf are the initial and final weights (g) of the control vials, respectively; and v is the volume (cm3) of each vial. The reported values are the average of three different measurements.

Results and Discussion FTIR Analysis FTIR spectra of chitosan film, blend films, and MPEG-b-PCL powder are shown in Figure 36.1. The FTIR spectrum of chitosan film shows absorption bands at 1654 and 1587 cm−1, which represent the carbonyl in the amide group (amide I) and free amino group, respectively. The FTIR spectrum of MPEG-b-PCL powder shows absorption bands at 1720 cm−1, which represent the carbonyl group. The FTIR spectra of blend films show absorption bands with characteristics of both chitosan and MPEGb-PCL. The intensities of the carbonyl bands (at ∼1720 cm−1) increased with increasing

Figure 36.1. Fourier transform infrared spectroscopy spectra of (a) chitosan film, (b) 90 :1 0 blend film, (c) 80 : 20 blend film, (d) 70 : 30 blend film, and (e) MPEG-bPCL powder. %T, percent transmittance.

462

PART 2: Poster Presentations

(a)

(b)

(c)

(d)

Figure 36.2. Scanning electron micrographs of surfaces and fracture surfaces of (a) chitosan film, (b) 90 : 10 blend film, (c) 80 : 20 blend film, and (d) 70 : 30 blend film.

MPEG-b-PCL ratios. The band positions of amide I and free amino groups of the chitosan were slightly shifted to a lower wave number when MPEG-b-PCL was blended. This result indicates that interaction bonds existed between chitosan and MPEG-b-PCL (Pawlak and Mucha 2003). Morphology Scanning electron micrographs of the chitosan and blend films are presented in Figure 36.2. The film surface and fractures of the chitosan film were smooth (Figure 36.2a). The surfaces and fractures of the films were roughened during blending (Figure 36.2b–d). The film roughness increased when the MPEG-b-PCL ratios increased. However, consistent roughness throughout the blend films suggested that the chitosan/ MPEG-b-PCL blend films are homogeneous. Water-Vapor Permeability Hydrophobic biodegradable PL and PCL have been blended to improve the WVP of chitosan film (Olabarrieta and others 2001; Suyatma and others 2004). In this

Water-Vapor Permeability of Chitosan Blend Homogeneous Films

463

Table 36.1. Rate of water-vapor permeability (WVP) of chitosan and blend films Chitosan/MPEG-b-PCL blend ratio (wt/wt)

WVP rate (g/day/liter)

100/0

6.6 ± 0.5

90/10

5.2 ± 0.6

80/20

4.8 ± 0.4

70/30

4.5 ± 0.4

MPEG-b-PCL, methoxy-poly(ethylene glycol)-b-poly(ε-caprolactone).

work, the presence of the hydrophilic MPEG block is thought to enhance water solubility of the hydrophobic PCL block. Table 36.1 illustrates influence of the MPEG-bPCL with different blend ratios on the rate of WVP of chitosan film. The WVP rates of the blend films were lower than that of the chitosan film. This suggested that the hydrophobic characteristics of PCL blocks dominated the hydrophilicity of the MPEG and, therefore, exhibited water vapor–resistant properties. The WVP rates of blend films decreased significantly when the MPEG-b-PCL ratios increased.

Conclusions The chitosan/MPEG-b-PCL blend homogeneous films were successfully prepared by film casting of the solution blends. The intermolecular bonds between chitosan and MPEG-b-PCL were determined by FTIR spectra. The film roughness increased, whereas the WVP of the chitosan films decreased as the proportion of MPEG-b-PCL in the blend increased. The blend films have potential for food-packaging applications because they have lower WVP than the chitosan film.

Acknowledgment The authors gratefully acknowledge the financial support from Mahasarakham University (fiscal year 2007, grant 5001024).

References Balua L, Lisa G, Popa MI, Tura V, Melnig V. 2004. Physico-chemical properties of chitosan films. Cent Eur J Chem 2:638–47. Ghaouth A, Arul J, Ponnampalam R, Boulet M. 1991. Chitosan coating effect on storability and quality of fresh strawberries. J Food Sci 56:1618–20. Ikejima T, Inoue Y. 2000. Crystallization behavior and environmental biodegradability of the blend films of poly(3-hydroxy butyric acid) with chitin and chitosan. Carbohydr Polym 41:351–6. Jeon Y-J, Kamil JYVA, Shahidi F. 2002. Chitosan as an edible invisible film for quality preservation of herring and Atlantic cod. J Agric Food Chem 50:5167–78. Kittur FS, Kumar KR, Tharanathan RN. 1998. Functional packaging properties of chitosan films. Z Lebensm Unters Forsch [A] 206:44–7. Kittur FS, Saroja N, Tharanathan RN. 2002. Storage studies of mango packed using biodegradable chitosan film. Eur Food Res Technol 215:504–8. Muzzarelli RAA, Muzzarelli C. 2005. Chitosan chemistry: relevance to the biomedical sciences. Adv Polym Sci 186:151–209. Olabarrieta I, Forsström D, Gedde UW, Hedengvist MS. 2001. Transport properties of chitosan and whey blended with poly(ε-caprolactone) assessed by standard permeability measurements and microcalorimetry. Polymer 42:4401–8.

464

PART 2: Poster Presentations

Pawlak A, Mucha M. 2003. Thermogravimetric and FTIR studies of chitosan blends. Thermochem Acta 396:153–66. Ravi Kumar MNV, Muzzarelli RAA, Muzzarelli C, Sashiwa H, Domb AJ. 2004. Chitosan chemistry and pharmaceutical perspectives. Chem Rev 104:6017–84. Rutnakornpituk M, Ngamdee P. 2006. Surface and mechanical properties of microporous membranes of poly(ethylene glycol)-polydimethysiloxane copolymer/chitosan. Polymer 47:7909–17. Sebastien F, Stephane G, Copinet A, Coma V. 2006. Novel biodegradable films made from chitosan and poly(lactic acid) with antifungal properties against mycotoxinogen strains. Carbohydr Polym 65:185–93. Senda T, He Y, Inoue Y. 2002. Biodegradable blends of poly(ε-caprolactone) with α-chitin and chitosan: specific interaction, thermal properties and crystallization behavior. Polym Int 51:33–9. Suyatma NE, Copinet A, Tighzert L, Coma V. 2004. Mechanical and barrier properties of biodegradable films made form chitosan and poly(lactic acid) blends. J Polym Environ 12:1–6.

37 Ice Formation in Concentrated Aqueous Glucose Solutions P. Thanatuksorn, K. Kajiwara, N. Murase, and F. Franks

Abstract This study used differential scanning calorimetry (DSC), X-ray diffraction (XRD), and low-temperature light microscopy to clarify the crystallization of the concentrated glucose-water system during cooling and to clarify the kinetics of recrystallization during rewarming, with a view to improving the understanding the phase transitions that occur in the freeze-thaw process. Ice formation in glucose solutions within the concentration range of 10%–55% wt/wt was studied by using DSC from 25° to −120°C with a scanning rate of 5°C/min. During cooling, the single freezing exotherm was clearly observed in 10%–50% glucose solutions. For 51%–55% glucose solutions, the double-crystallization exotherms were observed. The ice crystallization in 51%– 55% glucose solutions was observed by using a low-temperature light microscope. Between −50° and −60°C, the growth of spherical ice crystals was observed. Below −60°C, smaller crystals formed between the already formed ice crystals in the supercooled liquid. This coincided with the second exothermic peak that appeared at the lower temperature. The XRD pattern suggested that the double-crystallization exotherms in these solutions are both associated with growth of cubic ice (Ic) crystals. The stability of Ic was investigated by annealing as a function of time at −35°, −40°, −45°, and −50°C. The results revealed that Ic suddenly transformed into hexagonal ice (Ih) when the temperature reached −35°C. At −40° and −45°C, the Ic transformed into Ih within 30 min. Ic held at −50°C for 3 h was still stable. These results coincide with the very weak endothermic signals in the neighborhood of −43°C in the DSC heating scans, which might indicate a polymorphic transition from Ic to Ih.

Introduction Thermal, mechanical, and dynamic properties of concentrated sugar-water mixtures have become subjects of interest because they can govern the storage stability and processing characteristics of a wide variety of foods and pharmaceutical products. A better understanding of physical and chemical transformations, as well as the factors that govern the kinetics of such processes, is required. Highly concentrated glucose aqueous solution serves as a useful generic model because the significant role of glucose in metabolism has received much attention and because glucose is widely used as a pharmaceutical excipient. Hexagonal ice (Ih) is the stable phase of ice upon freezing at or below ambient pressure. The cubic form of ice (Ic), in contrast, is a metastable phase that transforms 465

466

PART 2: Poster Presentations

irreversibly into Ih. The cooling process, as generally performed, leads only to the crystallization of the stable Ih polymorph. On the other hand, Ic cannot be produced directly by cooling liquid water in bulk, nor can it be formed by cooling Ih, as is currently believed. Nevertheless, it has been suggested that small droplets of water in the upper atmosphere may often freeze first into cubic ice. Experimental studies on Ic production have reported that Ic was formed by freezing water in nanoporous silica (Dunn and others 1988) and by the devitrification of amorphous water or quenching of a water-glycerol mixture (Mayer and Hallbrucker 1987; Vigier and others 1987). Murray and others (2005) stated that cubic ice can be formed by freezing ammonia– sulfuric acid–water micrometer-sized droplets. Ice formation in highly concentrated aqueous solutions with high viscosity is still in question because initiating nucleation and growth of ice in the complex, viscous matrix under unstable conditions can be quite difficult. When ice crystal morphologies were observed by Luyet and his associates, the pioneer researcher relied on a microscopic method (Franks 1982). Crystallization in pure-water ice is generally observed in the hexagonal dendrite, whereas irregular crystal forms are considered to be a transitional stage between hexagonal and spherulitic crystals in higher concentrated aqueous solutions. However, there have been no reports about the formation of ice polymorphs and their stability in highly concentrated aqueous solutions in bulk. The present study was initiated to provide information about any measurable physical changes in maximally freeze-concentrated aqueous solutions of polyhydroxy compounds at sub–glass transition temperature (sub-Tg). Therefore, the object of this study using differential scanning calorimetry (DSC), X-ray diffraction (XRD), and lowtemperature light microscopy was to clarify the ice crystallization in concentrated glucose-water systems during cooling and during rewarming, and to elucidate the phase transitions that occur in the freeze-thaw process.

Materials and Methods Glucose was obtained from Wako (Osaka, Japan) and used without further purification. The freeze-thaw behavior of aqueous glucose solutions within the concentration range of 10%–60% wt/wt was studied by DSC, cryomicroscopy, and XRD. Calorimetric measurements were carried out in the range of 25° to −120°C with the aid of a differential scanning calorimeter 822e, connected to the cooling unit (Mettler Toledo, Schwerzenbach, Switzerland). The instrument had previously been calibrated with indium, n-hexane, and n-octane. Aqueous glucose solutions (10 mg each) were hermetically sealed in the 40-μL standard aluminum pans. Cooling and warming rates used were usually 5°C min−1. Ice crystallization was observed by using a cryomicroscope equipped with a Olympus DP7 CCD (charged-coupled device) video camera. The cryomicroscope consisted of an Olympus BX51 microscope and a Linkam THMS600 cold stage. A small volume of the solution was placed between two coverslips. The ice crystallization during cooling and its melting during warming were observed with a rate of temperature change of 5°C min−1 from 25° to −120°C.

Ice Formation in Concentrated Aqueous Glucose Solutions

467

Simultaneous X-ray diffractory and differential scanning calorimetry (XRD-DSC) were used with the aid of a RINT-UltimaIII unit connected to a ThermoPlus2 DSC (Rigaku, Tokyo, Japan). These measurements were performed with glucose solution in open aluminum pans covered with a polypropylene sheet. The DSC component of this instrument was calibrated at a scanning rate of 5°C min−1 with indium and water. The samples were cooled and warmed at a rate of 5°C min−1. However, with the limits of the DSC component, the lowest temperature that could be reached was −60°C. An independent X-ray diffractometer equipped with the medium- to low-temperature attachment (Rigaku) was used to observe the lowest temperatures. XRD was carried out with 40-kV × 40-mA CuKα wavelength = 1.54 Å and scanned at a rate of 10°min−1 (2θ min−1) with a 20°–50° diffraction-angle scanning range. The stability of Ic was investigated as a function of annealing time at several temperatures. The 53% glucose solution was cooled at a rate of 5°C min−1 to −60°C and then annealed at −50°, −45°, −40°, and −35°C. While the solution was held at the annealing temperature, the XRD pattern was observed.

Results and Discussion The single freezing exotherm of 10%–50% glucose solutions was clearly observed during cooling. The crystallization temperature decreased with increasing glucose concentration. The double-freezing exotherms were observed for 51%–54% solutions during the cooling period. Because of the detection of two exothermic peaks in 51%–54% glucose solutions, the crystallization in 53% glucose solution was selected for special study by cryomicroscopy. A representative DSC thermogram of 53% glucose solution is shown in Figure 37.1, together with ice crystals at the indicated temperature. Between −50° and −70°C, the spherical ice crystals increase in size rapidly. This coincides with the first DSC exothermic peak. At temperatures below −70°C, new, smaller crystals began to appear between the ice crystals in the supercooled liquid. This corresponds to the second exothermic peak. For the corresponding warming scan of the 53% glucose solution, the ice crystals begin to change at −70°C. At −60°C, many small dots appear on the larger, spherical ice crystals. Between −45° and −40°C, the ice crystals gradually change in appearance, and eventually uniform crystals are observed at −28°C. The type and structure of crystals that give rise to the first and second exothermic peaks were studied by XRD. Figure 37.2 shows a heat-flow curve for a cooling-heating cycle of the 53% glucose solution, together with representative XRD patterns between 25° and −60°C obtained from simultaneous XRD-DSC; because of the limits of the DSC component, only the first exothermic peak was observed. The XRD pattern in the cooling scan is consistent with Ic. Because of the limits of the XRD-DSC, the independent X-ray diffractometer equipped with the low-temperature attachment was used to observe crystals in the second exothermic peak. XRD of the second exothermic peak (at −80°C) also was consistent with Ic (data were not shown). This might be because water in these concentrations becomes partly ice crystals in the highertemperature exothermic peak. However, available water may form the ice at the lower

Figure 37.1. Thermogram and image from cryomicroscope at the temperatures indicated.

Temperature °C

20.0

120000

–5.0

–25.0

–45.0

–65.0

100000

32.0

90000

28.0

Time/min

24.0

70000

20.0

60000

16.0

50000 40000

12.0

–40 °C

30000

8.0

Cooling

Intensity [cps]

80000

Warming

36.0

110000

20000

26.0

32.0

38.0

2Theta [deg]

44.0

50.0 7.5

DSC

0 20.0

4.0

TEMP

10000

2.0

–2.0

–6.0

–10.0

0.0 –14.3

Heat flow/mW

Figure 37.2. Simultaneous X-ray diffractory and differential scanning calorimetry of 53% glucose solution.

468

Ice Formation in Concentrated Aqueous Glucose Solutions

469

temperature. Ice nucleation does not occur easily in concentrated solutions, so the metastable phase of Ic occurred in these solutions. In the case of warming, the XRD pattern begins to change from Ic to Ih at about −40°C (Figure 37.2). This result coincides with the weak endothermic signals in the DSC heating scans in the region of −40°C, which might indicate the polymorphic transition from Ic to Ih. We suggest, therefore, that the subendotherm represents the scenario in which the mobility of Ic has reached the point where it transitions to the stable polymorph within the period of observation. Therefore, the transformation of Ic to Ih in 53% glucose solution as a function of annealing time was investigated. To investigate the stability of Ic, the intensity ratio of the XRD peak at 2θ ≈ 40° and 44° (I44/I40) was used as the index indicating conversion to Ih (Mayer and Hallbrucker 1987; Murray and others 2005; Murray and Bertram 2006). The intensity at 44° is an indicator for the growth of the 103 reflection exclusive to Ih, while the reflection at 40° is common to both Ic and Ih. The I44/I40 ratio determined during the isothermal transformation measurements is plotted in Figure 37.3. According to the report by Murray and others (2005), the value of I44/I40 for stacking faulty Ic is 0, indicating that no bulk Ih formed in the droplets and therefore that the dominant product was Ic, whereas, for pure Ih, the value is 0.79 ± 0.03 in micrometer-sized droplets of water. The increase in the I44/I40 ratio corresponds to a decrease in the amount of Ic. In this study, the value of I44/I40 of ice formed in 53% glucose solution is nearly 0.3 when cooled to −60°C. The XRD pattern of Ic did not change when ice was warmed to −50°C and then annealed at that temperature for 3 h (I44/I40 = 0.31 ± 0.02). The I44/I40 ratio at this temperature seems to be unchanged when held for more than 3 h. However, this part of study could not last longer than 3 h because of the limits of the cooling unit. When the ice was held at −45°C, XRD peaks at 26°, 33° and 44°C, an ordinary indication of Ih, were observed after 30 min (the XRD pattern is not 0.9

–35 °C –40 °C –45 °C –50 °C

0.8

0.6 0.5 0.4 0.3 0.2

Before annealing

I44/I40

0.7

0

50 100 Annealing time (min)

150

Figure 37.3. The intensity ratio I44/I40 of the isothermal transformation measurement at varying temperatures.

470

PART 2: Poster Presentations

shown). The I44/I40 of the isothermal transformation measurement at this temperature increased significantly. At −40°C, the diffraction pattern began to convert to that of Ih after 10 min with an I44/I40 of 0.44 ± 0.02. Conversely, the XRD pattern of Ic suddenly converted to Ih at the beginning of the annealing experiment when ice was held at −35°C (I44/I40 = 0.64 ± 0.08). Transformation of Ic to Ih has been reported. Ic, which was made from the devitrification of a quenching water-glycerol mixture, transformed to Ih after 1 h when annealed at −70°C (Vigier and others 1987), whereas about 70% of Ic transformed in amorphous water at −43°C became completely Ih at −33°C after 30 min (Mayer and Hallbrucker 1987). Murray and Bertram (2006) reported that the Ic from emulsified water droplets transformed to perfect Ih after 178 min at −35°C. The difference in thermal stability between cubic ice made from the droplets and that made from the liquid is probably the result of differences in surface areas (Mayer and Hallbrucker 1987; Vigier and others 1987). Cubic ice crystals made from water droplets in the upper atmosphere may be even more stable because they are separated from one another. Similarly, the water droplets suspended in an oil matrix were separated from one another, and hence mass transfer and possibly surface nucleation, were blocked, which provides an explanation for their lasting longer at higher temperatures. As already mentioned, the stability of Ic is possibly correlated with the subendotherm in the warming scan. This subendotherm coincides closely with the thermal transition generally observed in the DSC warming scan, referred to as the glass transition of the maximally freeze-concentrated amorphous matrix ( Tg′) by Levine and Slade (1988). The Tg′ of the glucose solution at this concentration was observed at around −43°C. Ic started to transform into Ih after 10 min when ice was held at −40°C and suddenly transformed into Ih at −35°C without annealing (Figure 37.3). In addition, the two sizes of spherical ice crystals observed by cryomicroscopy began to disappear at −40°C and subsequently transformed into uniform crystals at −28°C (Figure 37.1). These results suggest that ice polymorphs transform when reaching Tg′. It has been believed that the Tg′ governs ice recrystallization and that stability is maintained during storage below Tg′. Below Tg′, the matrix exists as a kinetically metastable amorphous solid at temperature (Levine and Slade 1988; Blond 1989). The metastable Ic is still in this polymorph when ice is kept below Tg′. When the ice is warmed to above Tg′, Ic can thermodynamically orient to Ih (Takahashi and Kobayashi 1983) because the matrix surrounding the ice crystals behaves like a viscous liquid. These results can be used in understanding ice information in highly concentrated glucose aqueous solutions and the transition in the DSC thermogram during the warming scan. More experiments on the kinetics of ice transformation, including the relationship between ice polymorphic transformation and Tg′, are required.

References Blond G. 1989. Water-galactose system: supplemented state diagram and unfrozen water. Cryo-Lett 10:299–308. Dunn M, Dore JC, Chieux P. 1988. Structural studies of ice formation in porous silicas by neutron diffraction. J Cryst Growth 92:233–8.

Ice Formation in Concentrated Aqueous Glucose Solutions

471

Franks F. 1982. The properties of aqueous solutions at subzero temperatures. In: Felix F, editor. Water: a comprehensive treatise. New York: Plenum. p 215–338. Levine H, Slade L. 1988. Principles of “cryostabilization” technology from structure/property relationships of carbohydrate/water systems: a review. Cryo-Lett 9:21–63. Mayer E, Hallbrucker A. 1987. Cubic ice from liquid water. Nature 325:601–2. Murray BJ, Bertram AK. 2006. Formation and stability of cubic ice in water droplets. Phys Chem Chem Phys 8:186–92. Murray BJ, Knopf DA, Bertram AK. 2005. The formation of cubic ice under conditions relevant to Earth’s atmosphere. Nature 434:202–5. Takahashi T, Kobayashi T. 1983. Role of cubic structure in freezing of a supercooled water droplet on an ice substrate. J Cryst Growth 64:593–603. Vigier G, Thollet G, Vassoille R. 1987. Cubic and hexagonal ice formation in water-glycerol mixture (50% w/w). J Cryst Growth 84:309–15.

38 Effects of Sodium and Potassium Ions on the Viscosities in the Sodium/ Potassium-Glucose-Water Ternary System M. Soga, K. Kurosaki, and K. Kajiwara Abstract Sodium ions and potassium ions, are important in the activity of living bodies. Despite the chemical similarities of these ions, their cellular distributions are different and asymmetric; the external concentration of sodium ions exceeds their intracellular concentration and the reverse is true for potassium ions. To clarify the role of each ion in living bodies, the physicochemical properties of each ion has been studied in the alkali halide–saccharide–water ternary system. The viscosity of sodium (NaCl)- or potassium (KCl)-glucose-water with varying concentrations was investigated in this study. The density of the NaCl- or KCl-glucose-water ternary system was measured with a Gay-Lussac–type pycnometer. The flow time of the ternary system was measured with an Ubbelohde viscometer. The viscosity of the solutions was calculated from the density and the flow time. All measurements were conducted at 25° ± 0.01°C. The concentration of the NaCl-KCl-glucose solution used in this study was expressed in molar ratios (nalkali-metal salt/nwater and nglucose/nwater, where n is the number of moles). The concentrations of NaCl, KCl, and glucose in water were nNaCl/nwater = 0.00018, 0.00090, and 0.00181; nKCl/nwater = 0.00018, 0.00090, and 0.00181; and nglucose/nwater = 0.0009, 0.0018, 0.0099, 0.016, 0.022, and 0.030. For the NaCl-glucose-water system, the relative viscosity increases with increasing NaCl concentration. The results for KCl are more complex. When the glucose concentration is less than 0.0018, the relative viscosity increases with increasing KCl concentration, whereas when the glucose concentration exceeds 0.0099, the relative viscosity decreases with increasing concentration. A similar change in the dependence of relative viscosity on KCl concentration is seen in the binary KCl-water system as a function of temperature.

Introduction Many electrolytes, including sodium and potassium chloride, exist in a mammal’s living body. Despite the chemical similarities of the sodium and potassium ion, their cellular distributions are different and asymmetric; the extracellular concentration of sodium ions exceeds their intracellular concentration, the reverse being the case for potassium ions, suggesting that the ions have differing functions in the cell. The saccharides are also important tissue components that influence tissue fluid behavior. To clarify the roles of sodium ions and potassium ions in living 473

474

PART 2: Poster Presentations

bodies, in our previous research we examined the physicochemical characteristics of each ion in the alkali halide-saccharide-water ternary system by measuring its solubility (Kajiwara and Franks 2004) and freeze-thawing behavior (Kajiwara and others 2001). In the experiment on solubility, sodium chloride in the alkali halide–saccharide–water ternary system has both salting-out and salting-in effects on the solubility of saccharides, whereas potassium chloride has only salting effects on the solubility of saccharides. The freeze-thaw experiment showed that when the concentration of glucose is low in the sodium (NaCl)-glucose-water ternary system the eutectic elements vitrify. When the glucose concentration is high, only the water vitrifies. The results are comparable in the potassium (KCl)-glucose-water ternary system. The viscosity has been employed previously as a physicochemical index, often related to potential solution structural changes. In this study of NaCl/KCl-glucose-water ternary systems, the influence of sodium ions and potassium ions on solution viscosities was studied as a potential structural probe.

Materials and Methods Glucose, NaCl, and KCl were purchased from Wako Pure Chemical Industries (Osaka, Japan) and used without further purification. Pure water was used in all experiments. The density of the NaCl/KCl-glucose-water ternary system was measured with a GayLussac–type pycnometer. Density measurements were carried out at least three times. The flow time of the ternary system was measured with a Ubbelohde viscometer. All measurements were carried out in a temperature-controlled water bath (25° ± 0.01°C). The viscosity of the solutions was calculated by using the density and the flow-time results. The viscosity of each solution was divided by the viscosity of 25°C water and used as the relative viscosity. The concentrations of the NaCl-KCl-glucose in solution were expressed in molar ratios with water (nalkali-metal salt/nwater and nglucose/ nwater, where n is the number of moles). The NaCl, KCl, and glucose concentrations in water were nNaCl/nwater = 0.00018, 0.00090, and 0.00181; nKCl/nwater = 0.00018, 0.00090, and 0.00181; and nglucose/nwater = 0.00090, 0.0018, 0.0099, 0.016, 0.022, and 0.030.

Results and Discussion Figures 38.1 and 38.2 show, for different glucose concentrations, how the relative viscosity depends upon the salt concentration. In the NaCl-glucose-water system, relative viscosity increased with increasing NaCl concentration. This result agrees well with the tendency reported previously (Moreira and others 2003). In KCl-glucosewater, the viscosity behavior is more complex. When the concentration of glucose is less than 0.0018, the relative viscosity increases with increasing KCl concentration. In contrast, when the concentration of the glucose exceeds 0.0099, the relative viscosity decreases with increasing KCl concentration. A similar change in the dependence of viscosity on KCl concentration has been seen in the binary KCl-water system as a function of temperature. In this case, at lower temperatures, the relative viscosity

The Sodium/Potassium-Glucose-Water Ternary System

475

0.02 Difference of relative viscosity

in pure water 0.0009(nGlu/nH2O) 0.0018(nGlu/nH2O)

0.015

0.0099(nGlu/nH2O) 0.0160(nGlu/nH2O)

0.01

0.0220(nGlu/nH2O)

0.005

0 0

0.0004

0.0008 0.0012 nNaCl/nH2O

0.0016

0.002

Figure 38.1. Differences in relative viscosities of ternary solutions with sodium chloride–glucose–water at 25°C. nGlu/nH2O, molar ratio of glucose to water.

Difference of relative viscosity

0.01

0

in pure water 0.0009 (nGlu/nH2O) 0.0018 (nGlu/nH2O) 0.0099 (nGlu/nH2O) 0.0220 (nGlu/nH2O) 0.0300 (nGlu/nH2O)

-0.01

-0.02 0

0.0004

0.0008 0.0012 nKCl/nH2O

0.0016

0.002

Figure 38.2. Differences in relative viscosities of ternary solutions with potassium chloride–glucose–water at 25°C. nGlu/nH2O, molar ratio of glucose to water.

decreases with increasing KCl concentration, whereas, at higher temperatures, the relative viscosity increases with increasing KCl concentration. Next we plan to measure the viscosity of the ternary systems at higher salt concentrations.

References Kajiwara K, Franks F. 2004. Solubilities in the ternary system alkali metal halide–glucose–water. J Solution Chem 33:157–67. Kajiwara K, Motegi A, Murase N. 2001. Freeze-thawing behaviour of highly concentrated aqueous alkali chloride-glucose systems. Cryo Lett 22:311–20. Moreira R, Chenlo F, Pereira G. 2003. Viscosities of ternary aqueous solutions with glucose and sodium chloride employed in osmotic dehydration operation. J Food Eng 57:173–7.

39 Comparison of Water Sorption and Crystallization Behaviors of Freeze-Dried Lactose, Lactitol, Maltose, and Maltitol K. Jouppila, M. Lähdesmäki, P. Laine, M. Savolainen, and R. A. Talja

Abstract Molecular structures of sugars and sugar alcohols are quite similar, but their properties and behavior under various processing and storage conditions may vary. In this study, water sorption and crystallization behaviors of various freeze-dried sugars and sugar alcohols were compared. The water-sorption properties of samples were determined gravimetrically when stored at different relative vapor pressures (RVPs) ranging from 11% to 86% at 25°C. Crystallization was observed as the gradual loss of sorbed water during storage. The glass transition temperature (Tg) of samples at various water contents was determined by using differential scanning calorimetry (DSC). The solid state of samples, which were freeze-dried after water-sorption study, was analyzed by using X-ray diffraction. The water-sorption isotherms of lactose, maltose, and lactitol were quite similar, and their water sorption was slightly higher than that of maltitol. The Tg values of anhydrous sugars were about 50°C higher than those of corresponding anhydrous sugar alcohols. Sugar alcohols with lower Tg values crystallized at lower RVP than did sugars. The lowest RVP at which both lactitol and maltitol were observed to crystallize was 44%, whereas the lowest RVP was 54% and 66% for lactose and maltose crystallization observed, respectively, in the time frame of the experiment. When these sugars and sugar alcohols are used as encapsulating matrices in low-moisture food or pharmaceutical products, crystallization during storage may result in the loss of protection ability, affecting the quality of encapsulated compounds.

Introduction Various sugars and sugar alcohols have quite similar molecular structures but differences in their properties, such as sweetness, solubility, dissolution rate, water sorption, tendency to crystallize, and melting temperature. Also, their behavior varies under various processing and storage conditions. Thus, it is important to know the properties and behavior of various sugars and sugar alcohols when seeking alternatives to sugars currently used in foods and pharmaceutical products. For example, the incidence of lactose intolerance and the contribution of sucrose to the development of dental caries have brought about the need for substitutive sugars and/or sugar alcohols with suitable properties. 477

478

PART 2: Poster Presentations

The present study compared the water-sorption properties and crystallization behavior of two sugars (the disaccharides lactose and maltose) and two sugar alcohols derived from those disaccharides (lactitol and maltitol) as a function of time.

Materials and Methods Lactose, maltose, and maltitol were purchased from Sigma-Aldrich (St. Louis, MO, USA), and lactitol was donated by Danisco Sweeteners (Copenhagen, Denmark). Solutions (5 mL) containing 10% (wt/wt) either sugar (lactose or maltose) or sugar alcohol (lactitol or maltitol) and 90% (wt/wt) distilled water were prepared and transferred into 20-mL glass vials. Samples were frozen at −20°C and stored at −80°C for at least 24 h prior to freeze drying (Lyovac GT2; Amsco Finn-Aqua, Hürth, Germany). Freeze-dried samples were further dehydrated in vacuum desiccators containing phosphorus pentoxide. The water-sorption properties of samples were determined gravimetrically by storing them in vacuum desiccators over saturated salt solutions giving relative vapor pressures (RVPs) ranging from 11% to 85% at 25°C (Labuza and others 1985). The water content of samples when stored at an RVP of 11%–44% for 94 h (lactose and maltose) or 98 h (lactitol and maltitol) was taken as steady-state water content, which was used in the calculation of the Brunauer-Emmett-Teller (BET) water-sorption isotherms. The crystallization of sugars and sugar alcohols was estimated based on the gradual loss of sorbed water during storage. The glass transition temperature (Tg) of samples at various water contents was determined by using differential scanning calorimetry (DSC) (DSC30; Mettler Toledo, Greifensee, Switzerland). Samples (∼15 mg) in standard aluminum 40-μL sample pans were scanned from temperatures that were 30°C lower than the Tg to temperatures that were 30°C higher than the Tg at a heating rate of 5°C min−1. The Tg was determined as the onset temperature of glass transition; that is, the onset temperature of a change in heat capacity in the DSC thermogram. The decrease in the Tg as a result of water plasticization was modeled by using the Gordon-Taylor equation. The critical values for water content and water activity that depress the Tg of material to storage temperature (Roos 1993a) were calculated. The solid state of samples, which were freeze-dried after water-sorption study, was analyzed by using an X-ray powder diffractometer (D8 Advance; Bruker AXS, Karlsruhe, Germany). The experimental X-ray powder diffraction (XRPD) patterns were compared with theoretical XRPD patterns, based on data obtained from Cambridge Crystallographic Data Centre (CCDC).

Results and Discussion The water sorption of freeze-dried sugars and sugar alcohols was modeled by using the BET model (Figure 39.1). The BET water-sorption isotherms of lactose, maltose, and lactitol were quite similar, and monolayer water content obtained from BET modeling for lactose, maltose, and lactitol was also quite close to one another (6.6, 6.2, and 6.9 g/100 g of solids, respectively). The water sorption of maltitol, as well as

Comparison of Water Sorption, Crystallization of Freeze-Dried Sugars and Sugar Alcohols

(b) 100

25

60

20

40 20

15

25

0 -20

7.6

10

-40 -60

5

-80

60

0

0.2

0.4

0.6

20

15

25

0 -20

10

-40

5.0

-60

5

-80

0.28

-100

0 1.0

0.8

20

40

0.37

-100

25

80

Glass transition temperature (°C)

80

Water content (g/100 g of solids)

0

0.2

(c)

0.6

0 1.0

0.8

(d) 100

100

80 60

20

40 20

15

25

0 -20

10

-40 -60

5

1.9

0.12

-80 -100 0

0.2

0.4

0.6

Water activity

0.8

Glass transition temperature (°C)

25

Water content (g/100 g of solids)

Glass transition temperature (°C)

0.4

Water activity

Water activity

25

80 60

20

40

25

20

15

0 -20

10

-40 -60

5

0.07

-80 0 1.0

0.8

-100 0

0.2

0.4

0.6

0.8

Water content (g/100 g of solids)

Glass transition temperature (°C)

100

Water content (g/100 g of solids)

(a)

479

0 1.0

Water activity

Figure 39.1. Experimental glass transition temperature (Tg) (open circles) and water contents (solid circles) after storage of 94 h for freeze-dried lactose (a) and maltose (b) and after storage of 98 h for freeze-dried lactitol (c) and maltitol (d) as a function of water activity at 25°C. Critical water activities and critical water contents depressing Tg to 25°C are shown.

its monolayer water content (5.6 g/100 g of solids), was slightly lower than those of lactose, maltose, and lactitol. The Tg values obtained for anhydrous sugars were about 50°C higher than those obtained for corresponding anhydrous sugar alcohols (Figure 39.1). The Tg values of anhydrous freeze-dried lactose and maltose were almost identical with those reported by Roos (1993b). However, the Tg value determined for anhydrous freeze-dried maltitol was slightly lower than the value determined for quench-cooled anhydrous maltitol melt (Roos 1993b). The Tg values decreased with increasing water content, and this relationship was modeled by using the Gordon-Taylor equation. The k values obtained for lactose, maltose, lactitol, and maltitol were 6.0, 7.6, 8.1, and 7.1, respectively. The k values of maltose and maltitol were higher than those estimated by Roos (1993b), but the k value of lactose was quite similar. The critical water activity and critical water content of lactose (Figure 39.1) were identical to the values reported by Jouppila and Roos (1994). The critical water content of maltose was lower than the value reported by Roos and Jouppila (2003) of 6.3 g/100 g

480

PART 2: Poster Presentations

(a)

(b) 30

Water content (g/100 g of solids)

Water content (g/100 g of solids)

15

10

5

20

15

10

5

0

0 0

50

100

150

200

250

300

Time (h)

(c)

0

50

100

150

200

250

300

Time (h)

(d) 25

Water content (g/100 g of solids)

25

Water content (g/100 g of solids)

25

20

15

10

5

0

20

15

10

5

0 0

100

200

300

Time (h)

400

500

0

100

200

300

400

500

Time (h)

Figure 39.2. Water content of lactose (a), maltose (b), lactitol (c), and maltitol (d) during storage at 25°C at various relative vapor pressures: 44% (circles), 54% (upwardpointing triangles), 66% (squares), 76% (downward-pointing triangles), and 85% (diamonds). Crystallization can be observed as the loss of sorbed water as a function of time.

of solids, which was obtained by using the calculated k value. Critical values obtained for lactitol and maltitol were very low. They were even lower than critical values obtained for freeze-dried skim milk with hydrolyzed lactose (Jouppila and Roos 1994) containing monosaccharides with very low Tg values (∼30°C [Roos 1993b] for anhydrous monosaccharides). The critical water content of maltitol was lower than the value (1.8 g/100 g of solids) estimated by Roos and Jouppila (2003). Sugar alcohols with lower Tg values crystallized at a lower RVP than did sugars (Figure 39.2). The lowest RVP at which both lactitol and maltitol were observed to crystallize was 44%, whereas the lowest RVPs at which lactose and maltose were observed to crystallize were 54% and 66%, respectively, in the time frame of the experiment. The rate of crystallization of sugars generally increases with increasing RVP (Jouppila and others 1998). This phenomenon was observed also in the present

Comparison of Water Sorption, Crystallization of Freeze-Dried Sugars and Sugar Alcohols

(a)

481

(b)

400

300

350

10000

250

250 200 5000 150

Intensity (counts)

Intensity (counts)

10000 300

200

150 5000 100

100 50 50 0 5

10

15

20

25

30

35

0

0 40

5

10

Diffraction angle, 2θ (°)

20

25

30

35

0 40

Diffraction angle, 2θ (°)

(c)

(d) 400

250

10000

150

5000

100

Intensity (counts)

10000

200

Intensity (counts)

15

300

200 5000

100 50

0 5

10

15

20

25

30

Diffraction angle, 2θ (°)

35

0 40

0 5

10

15

20

25

30

35

0 40

Diffraction angle, 2θ (°)

Figure 39.3. X-ray powder diffraction (XRPD) patterns of lactose (a), maltose (b), lactitol (c), and maltitol (d) at the end of storage at 25°C at relative vapor pressures of 54%, 76%, 66%, and 54%, respectively. Theoretical XRPD patterns of prevailing crystal forms (gray patterns) are included: α-lactose monohydrate (LACTOS11) (Smith and others 2005) (a), β-maltose monohydrate (MALTOS11) (Gres and Jeffrey 1977) (b), lactitol monohydrate (JEXZER) (Kanters and others 1990) (c), and anhydrous maltitol (BIZHIP01) (Park and others 1989) (d).

study. However, at high RVPs, water adsorption, water desorption due to crystallization, and solubilization occurred simultaneously, which may explain the slower loss of sorbed water or the lower water content during water-sorption study than was expected. All of the samples in which the loss of sorbed water was observed were crystalline. Lactose crystallized mainly as α-lactose monohydrate, but samples stored at an RVP of 54% contained also anhydrous β-lactose (Figure 39.3a). Maltose and maltitol seemed to crystallize into one crystal form: maltose crystallized as β-maltose monohydrate (Figure 39.3b) and maltitol as anhydrous maltitol (Figure 39.3d). Lactitol was observed to crystallize as a mixture of several crystal forms. Lactitol monohydrate was shown to be one of the crystal forms present (Figure 39.3c); it is probably the prevailing crystal form in crystallized lactitol samples.

482

PART 2: Poster Presentations

Conclusions There were differences in Tg values and crystallization behavior of sugars and sugar alcohols studied. However, differences in water-sorption behavior were not so notable. The rate of crystallization in sugars and sugar alcohols was affected by the RVP, and thus, the degree of water plasticization. An increase in water plasticization increases diffusion and molecular mobility, resulting in the acceleration of crystallization. Knowledge of the water plasticization and crystallization behavior of various sugars and sugar alcohols is important when they are used in low-moisture food or pharmaceutical products; for example, as encapsulating materials for the protection of functional compounds against oxidation. Increased molecular mobility as a result of water plasticization results in various changes, such as crystallization, and thus, loss of protection ability which can affect the quality of encapsulated compounds.

References Gres SME, Jeffrey GA. 1977. A neutron diffraction refinement of the crystal structure of β-maltose monohydrate. Acta Crystallogr [B] 33:2490–5. Jouppila K, Kansikas J, Roos YH. 1998. Crystallization and X-ray diffraction of crystals formed in waterplasticized amorphous lactose. Biotechnol Prog 14:347–50. Jouppila K, Roos YH. 1994. Glass transitions and crystallization in milk powders. J Dairy Sci 77:2907–15. Kanters JA, Schouten A, van Bommel M. 1990. Structure of lactitol (4-O-β-D-galactopyranosyl-D-glucitol) monohydrate: an artificial sweetener. Acta Crystallogr [C] 46:2408–11. Labuza TP, Kaanane A, Chen JY. 1985. Effect of temperature on the moisture sorption isotherms and water activity shift of two dehydrated foods. J Food Sci 50:385–91. Park YJ, Shin JM, Shin W, Suh I-H. 1989. The crystal and molecular structure of maltitol. Bull Korean Chem Soc 10:352–6. Roos YH. 1993a. Water activity and physical state effects on amorphous food stability. J Food Process Preserv 16:433–47. Roos YH. 1993b. Melting and glass transitions of low molecular weight carbohydrates. Carbohydr Res 238:39–48. Roos YH, Jouppila K. 2003. Plasticization effect of water on carbohydrates in relation to crystallization. In: Kaletunç G, Breslauer KJ, editors. Characterization of cereals and flours. New York: Marcel Dekker. p 117–49. Smith JH, Dann SE, Elsegood MRJ, Dale SH, Blatchford CG. 2005. α-Lactose monohydrate: a redetermination at 150 K. Acta Crystallogr [E] 61:o2499–501.

40 Sorption Behavior of Extruded Rice Starch in the Presence of Glycerol J. Enrione, S. Hill, J. R. Mitchell, and F. Pedreschi

Abstract The modeling approach to predict the sorption behavior of nonelectrolytic mixtures has been related to moisture content at a specific equilibrium relative humidity (ERH) and to the component weight fraction in the system. The work presented here attempts to identify deviations of the predicted sorption profile of a model system based on rice starch and glycerol. Accurate sorption isotherms (25°C) were obtained by using dynamic vapor sorption (DVS) for the ERH range of 0%–90%. The glass transition temperature (Tg) of the samples was obtained by differential scanning calorimetry (DSC). Sorption studies show a reduction in moisture content for an ERH of <60% in the presence of glycerol. This behavior was represented by a reduction in Guggenheim-Anderson-de Boer (GAB) modeling parameters, the monolayer value (from 8.9% to 6.6%); a reduction in the constant related to the net heat of sorption (from 6.4 to 3.5); and an increase in the constant related to the heat of sorption at the multilayer (from 0.72 to 0.94). The sorption profiles of the starch-glycerol mixtures were also modeled using sorption isotherms and moisture contents from each component. The estimated interaction factor of <1 indicated a reduction in water uptake specifically for an ERH of <60%. The 10% and 20% glycerol systems showed a significant increase in independent interaction factors (IIFs) from ∼0.6 to ∼0.9, suggesting a relationship between water uptake and the matrix-polyol interaction. This behavior was related to the molecular mobility of the system, as indicated by plotting IIF versus T − Tg (T = 25°C).

Introduction It is well known that small changes in moisture can cause major changes in the textural properties of food materials (Suwonsichon and Peleg 1998). Water and other plasticizers (e.g., glycerol and glycol) are small and ubiquitous molecules whose effect on texture can be related to the increase in the free volume of the polymer forming the food matrix. Polyols are commonly used in the food industry to reduce the overall water-vapor pressure or water activity (aw) in the mixture, helping to control microbial growth while retaining the desirable texture attributes of the product. The relationship between total moisture content and aw at constant temperature yields the sorption isotherm (Brunauer and others 1940). Several mathematical models have been developed to represent the sorption behavior of food materials, with the GuggenheimAnderson-de Boer (GAB) equation being the most widely used (Equation 40.1). The 483

484

PART 2: Poster Presentations

GAB model distinguishes physical chemistry between the bulk water and multilayer water (Anderson 1946; Van den Berg 1985; Bader and Göritz 1994). M=

mo CKa w

(1 − Ka w ) (1 − Ka w + CKa w )

(40.1)

Here, M is the equilibrium moisture content, aw is the water activity (relative humidity at thermodynamic equilibrium), mo is the monolayer value, C is the constant related to the heat of sorption at the monolayer, and K is the constant related to the heat of sorption at the multilayer. To predict the sorption profile of multicomponent systems, sorption isotherms and their component weight fraction can be combined (Lang and Steinberg 1980; Leiras and Iglesias 1991; Kaminski and Malik Al-Bezweni 1994; Rahman 1995). In practice, it is assumed that individual ingredients do not interact with each other, and that the aw is the same for all the components in the mixture. A simple equation describing this approach has been proposed by Kaminski and AlBezweni (1994), who introduced the interaction coefficient ξi, representing the physicochemical interaction between particular components. If an interaction ξ occurs, the moisture content of a component ith (Xei) becomes ξi Xei. For an n-component mixture, the mean moisture content Xem in the mixture is defined as n

X em = ∑ WSi ξi X ei

(40.2)

i =1

where WSi is the mass fraction of the ith component in the mixture. The relationship between moisture content of an ith component and its aw can be determined from an equation describing the sorption isotherm (e.g., the GAB model). An aspect that has not been considered here, is the likely variation in interactions if the composition of the mixture is varied. It is therefore clear that moisture uptake of complex food systems in the presence of a combination of plasticizers is not a straightforward prediction. To establish how glycerol affects water sorption, to determine possible matrixplasticizer interactions, and to determine whether this behavior can be predicted, a model system based on rice starch and glycerol was prepared by hot extrusion. Wideangle X-ray diffraction (WAXD) was used to assess the starch conversion during sample preparation. The sorption behavior was followed by dynamic vapor sorption (DVS). The glass transition temperature (Tg) was assessed by differential scanning calorimetry (DSC). To improve the understanding of the sorption profiles obtained, mathematical models were applied to the experimental data.

Materials and Methods Native rice starch (Sigma-Aldrich UK, Poole, England) and glycerol (Merck UK, Feltham, England) were processed in a twin-screw co-rotating extruder (Clextral BC21; Clextral, Firminy, France) operating at barrel temperatures of 25°, 85°, 85°, and 65°C. Glycerol and water were individually pumped directly into the extruder barrel to obtain 40% moisture (dry-weight basis [db]) and 5%, 10%, 15%, and 20% of the

Sorption Behavior of Extruded Rice Starch in the Presence of Glycerol

485

polyol (db) samples. The processing specific mechanical energy was ∼63 Watt h/kg. After extrusion, the ribbons were hermetically packed, quenched under liquid nitrogen, freeze-dried, ground (particle diameter, <212 μm), and stored at −80°C. WAXD An X-ray diffractometer, CuK → 0.154 nm (model D5005; Bruker, Karlsruhe, Germany), was used. Experimental settings were the following: scanning, from 2θ = 4°–38°; angular step, 0.05°/3 s; and holder rotational speed, 60 rpm. All the samples were run in duplicate. DVS A DVS-1 (Surface Measurements Systems, London, England) was used. Experimental settings were ∼4 mg of ground sample equilibrated at an ERH from 0% to 90% in 10% increments (9 points). Sorption was “at equilibrium” if dm/dt (slope of the changing mass [m] with time [t]) is <0.002 (mass%/min). All samples were analyzed in triplicate. DSC A Pyris Diamond differential scanning calorimeter (PerkinElmer, Waltham, MA, USA) was used. The DSC was calibrated by using indium and cyclohexane. The experimental settings were the following: 50 mg of sample, equilibrated to different moisture contents (over saturated salt solutions), were loaded into stainless-steel pans and heated from −30° to 140°C (run and rerun) at 10°C/min. The Tg was determined as the temperature at half-value for the step change in heat capacity. All samples were analyzed in triplicate.

Results and Discussion WAXD Figure 40.1 indicates a semiamorphous-like pattern with small peaks at ∼13° and ∼20°, both associated with amylose-lipid complex (Buléon and others 1998; Becker and others 2001; Gelders and others 2004). The peaks at ∼18° and ∼23° seem to be related to a fraction of native starch not fully converted during extrusion. It is possible that a high glycerol concentration could have a lubricating effect protecting the starch’s native structure from the shear effect on extrusion. DVS Figure 40.2 shows the obtained rice-glycerol extrudate sorption data. Due to experimental difficulties, the sorption isotherm for pure glycerol was obtained from the literature (Bell and Labuza 2000). Glycerol contributes to an increase in moisture uptake at an equilibrium relative humidity (ERH) of >70%. In the highly plasticized starchy matrix, a glycerol fraction seems to be available that attracts water molecules to the system. On the contrary, at an ERH of <60%, higher concentrations of the polyol reduced water uptake.

PART 2: Poster Presentations

20% Glycerol

Arbitrary units (A.U.)

486

10% Glycerol 5% Glycerol 0% Glycerol

4°

38°

20° Diffractive angle (2θ)

Figure 40.1. X-ray diffraction patterns for the extruded rice starch with 0%, 5%, 10%, and 20% glycerol. A.U., arbitrary units. 50.0% 45.0%

0 % Glycerol 5 % Glycerol

Moisture content (db)

40.0% 35.0%

10 % Glycerol 20 % Glycerol Pure Glycerol

30.0% 25.0% 20.0% 15.0% 10.0% 5.0% 0.0% 0.0%

20.0%

40.0%

60.0%

80.0%

100.0%

ERH (%)

Figure 40.2. Sorption isotherms (25°C) for the extruded rice starch with 0%, 5%, 10%, and 20% glycerol. db, dry-weight basis; and ERH, equilibrium relative humidity.

Myllärinen and others (2002) suggested that, at a low ERH, water is displaced from the starch active polar groups by glycerol. The trihydric alcohol (3 -OH) structure of this polyol and the solid acting as an adsorbent at these low ERHs may explain these results (Van den Berg and Bruin 1981), suggesting a predominance of glycerol over water.

Sorption Behavior of Extruded Rice Starch in the Presence of Glycerol

487

Table 40.1. Guggenheim-Anderson-de Boer (GAB) parameters obtained for the rice-glycerol extrudates % Glycerol (db)

% Monolayer (mo)

C

K

0

8.9

6.4

0.72

% EF 0.9

5

6.0

9.7

0.88

5.2

10

5.6

8.1

0.93

3.9

20

6.6

3.5

0.94

1.9

db, dry-weight basis; mo, monolayer value; C, constant related to the net heat of sorption; K, constant related to the heat of sorption at the multilayer; and EF, error function. Table 40.2. Interaction factor ξi and error function (EF) for the rice starch–glycerol mixture % Glycerol (db)

ξi

% EF

5

0.89

2.5

10

0.84

11.8

20

0.81

22.4

db, dry-weight basis.

Modeling Sorption Isotherms The accuracy of the modeling was assessed following the approach suggested by McMinn and others (2004). An error function of <10% was considered an accurate fit. Table 40.1 shows the GAB parameters obtained for the rice-glycerol extrudates. The mo obtained for 0% glycerol was ∼8.9% db. Similar values have been reported in the literature for starch-based materials. Bader and Göritz (1994) obtained an mo of ∼9.1% (db) for high-amylose cornstarch. The mo decreased to ∼6.6% for the 20% glycerol samples. It seems that, at low ERH (e.g., moisture contents less than the monolayer coverage), glycerol decreases the active sorption sites, hence reducing the surface area available for water molecules. The parameter C also varied in the presence of glycerol, giving a C of ∼3.5 for the 20% glycerol. C is related to the adsorption energies at the monolayer (Brunauer and others 1938). Glycerol’s hydroxyl groups would occupy these sites, reducing the enthalpy of the water-matrix interaction. In terms of K, an increase in this parameter was observed when the polyol concentration increased (K, ∼0.94). An increase in K toward a value of 1 would indicate a reduction in sorption energy at the multilayer, which suggests destructuring behavior from a multilayer to a bulk liquidlike domain. Prediction of Sorption Isotherms Equation 40.2 was tested to predict the sorption isotherms of the rice starch–glycerol mixtures. The ξi values were obtained by the same approach used for the calculation of the GAB parameters. Table 40.2 summaries the ξi values obtained and the error function for each fit.

488

PART 2: Poster Presentations

(a)

(b) 1.0

Independent interaction factors

Independent interaction factors

1.0 0.9 0.8 0.7 0.6 0.5

5 % Glycerol 10 % Glycerol

0.4 0.3 0%

20 % Glycerol

0.9 0.8 0.7 0.6 0.5 0.4

5 % Glycerol 10 % Glycerol 20 % Glycerol

0.3

10%

20%

30%

MC(%) (db)

40%

50%

-150

-100

-50

0

50

T – Tg (°C, T = 25°C)

Figure 40.3. Independent interaction factors (IIF) versus moisture content (MC) (a) and T − Tg (b).

The ξi values were <1, indicating possible matrix-glycerol interactions by a reduction of the availability of sorption sites (Leiras and Iglesias 1991). This can be correlated with the variations observed on the GAB parameters mo, C, and K. Independent interaction factors (IIFs) were estimated as a function of moisture content (Figure 40.3a). The IIF for the rice starch–5% glycerol sample did not change significantly when the moisture content increased from 2.9% to 18.7% (db), and the IIF maintained its value up to ∼0.90. Only at 28% (db) did moisture content IIF increase to ∼0.95. The 10% and 20% glycerol extrudates showed an increase in IIF from 0.66 to 0.91 and 0.52 to 0.85, respectively. IIF values of <0.70 at low moisture contents (20% db) indicated stronger starch-glycerol interactions. This behavior seems to be correlated with glycerol concentration, with the 20% glycerol extrudates having lower IIF values. Another interesting finding is related to the overlapping of IIF values at moisture contents that are >20% (db). This behavior could be explained in terms of the system molecular mobility, as indicated by plotting IIF values versus T − Tg (Figure 40.3b). The overlapping seems to occur at T − Tg of ∼0°C, the point at which an increase in the polyol concentration will not affect the plasticizer-matrix interaction, resulting in IIF values of ∼1.

Conclusions Two different sorption behaviors were observed, depending on ERH when glycerol was added to the system. At an ERH of >70%, glycerol contributed to an increase in the moisture content of the samples, but, at an ERH of <60%, the polyol decreased the water sorption. The GAB modeling suggests that, in low-vapor-pressure environments, glycerol molecules may fill the sorption sites of the matrix normally occupied

Sorption Behavior of Extruded Rice Starch in the Presence of Glycerol

489

by water. This behavior was represented by a reduction in mo and C and an increase in K when the polyol concentration increased. These findings were supported by IIF values, which decreased (greater interaction) to a minimum for the 20% glycerol extrudates. The overlapping of IIF as function of moisture content was attributed to the system molecular mobility, which was related to the Tg of each mixture. Further studies, including those investigating sorption isotherms at various temperatures, are required to improve our understanding of the thermodynamics describing the sorption behavior of these systems in the presence of different combinations of plasticizers.

References Anderson RB. 1946. Modifications of the Brunauer, Emmett and Teller equation. J Am Chem Soc 68:686–91. Bader H, Göritz D. 1994. Investigations on high amylose corn starch films. Part 2: Water vapor sorption. Starch/Stärke 46:249–52. Becker A, Hill SE, Mitchell JR. 2001. Relevance of amylose-lipid complexes to the behaviour of thermally processed starches. Starch 53:121–30. Bell LN, Labuza TP. 2000. Moisture sorption: practical aspects of isotherm measurement and use. 2nd ed. St Paul, MN: American Association of Cereal Chemists. Brunauer S, Deming LS, Deming WE, Teller E. 1940. On the theory of the van der Waals adsorption of gases. J Am Chem Soc 62:1723–32. Brunauer S, Emmett PH, Teller E. 1938. Adsorption of gases in multimolecular layers. J Am Chem Soc 60:309–19. Buléon A, Colonna P, Planchot V, Ball S. 1998. Starch granules: structure and biosynthesis. Int J Biol Macromol 23:85–112. Gelders GG, Vanderstukken TC, Goesaert H, Delcour JA. 2004. Amylose-lipid complexation: a new fractionation method. Carbohydr Polym 56:447–58. Kaminski W, Al-Bezweni M. 1994. Calculation of water sorption for multicomponent protein-containing mixtures. Int J Food Sci Technol 29:129–36. Lang KW, Steinberg MP. 1980. Calculation of moisture content of a formulated food system to any given water activity. J Food Sci 45:1228–30. Leiras MC, Iglesias H. 1991. Water vapour sorption of two cake mixes and their components. Int J Food Sci Technol 26:91–7. McMinn W, Al-Muhtaseb A, Magee TRA. 2004. Assessment of two- and three-parameter Lewicki models for description of sorption phenomena of starch materials. J Sci Food Agric 84:1695–700. Myllärinen P, Partanen R, Jukka S, Forssell P. 2002. Effect of glycerol on behaviour of amylose and amylopectin films. Carbohydr Polym 50:355–61. Rahman S, ed. 1995. Food properties handbook. 1st ed. Boca Raton, FL: CRC. Suwonsichon T, Peleg M. 1998. Instrumental and sensory detection of simultaneous brittleness loss and moisture toughening in three puffed cereals. J Texture Stud 29:255–74. Van den Berg C. 1985. Development of BET-like models for sorption of water on foods: theory and relevance. In: Simatos D, Multon JL, editors. Properties of water in foods in relation to quality and stability. NATO Science Series E. New York: Springer. Van den Berg C, Bruin S. 1981. Water activity and its estimation in food systems: theoretical aspects. In: Rockland LB, Stewart GF, editors. Water activity: influences on food quality. London: Academic. p 1–60.

41 Water State and Mobility Affect the Mechanical Properties of Coffee Beans P. Pittia, G. Sacchetti, P. Rocculi, L. Venturi, M. Cremonini, and M. Dalla Rosa

Abstract The plasticization and antiplasticization effects of water in raw and roasted coffee beans were investigated as described by sorption isotherms, as well as water “freedom” and mobility. Sorption isotherms were obtained in the 0.24–0.94 water activity (aw) range and fitted by the Guggenheim-Anderson-de Boer (GAB) equation. Mechanical properties of samples equilibrated at different water activities were studied by uniaxial compression. Differential scanning calorimetry (DSC) and low-resolution nuclear magnetic resonance (LR-NMR) results showed that the rehydration process in coffee beans can be divided into two different stages. The first stage (aw from 0.24 ± 0.01 up to 0.76–0.78 and to 0.86 for raw and roasted samples, respectively), where sorbed water acts as an antiplasticizer, with an increase in fracture force at increasing moisture levels, represents the solid-matrix hydration. In the second stage, at increasing hydration levels (above aw 0.78 and 0.86 for raw and roasted samples, respectively), after the completion of the monolayer hydration, water behaves as a bulklike agent, as confirmed by the appearance of an endothermic melting peak and by a significant increase in the transverse relaxation time (T2). Only in this stage does water act as a plasticizer. These results suggest that bulklike water accounts for the plasticization effect as described by mechanical analysis.

Introduction Water is the most effective plasticizer in food matrices, decreasing glass transition temperature (Tg) and mechanical resistance. In general, a progressive softening occurs with the increase in its concentration. However, upon water sorption, amorphous porous matrices can undergo a toughening effect up to certain moisture content, corresponding to a critical water activity (aw) value, above which plasticization takes place. This behavior is referred to as antiplasticization (Seow and others 1999; Pittia and Sacchetti 2008). Hypotheses on the causes of the antiplasticization effect of water and the interest of scientists in this phenomenon are increasing, as documented by literature on different food matrices (Moraru and others 2002; Gondek and Lewicki 2006; Marzec and Lewicki 2006). However, there lack studies aimed at better understanding the relationship between the state of the water; microstructural, macrostructural, and compositional properties; and the antiplasticization effect of water in foods (Pittia and Sacchetti 2008). 491

492

PART 2: Poster Presentations

In the case of coffee beans, textural changes were studied as a function of aw and evidenced an antiplasticization effect of water above the Brunauer-Emmett-Teller (BET) monolayer value (Pittia and others 2007). The plasticization threshold was calculated to occur at different aw values in raw products (aw, 0.75 ± 0.02) and roasted products (aw, 0.86 ± 0.02). Since it was first introduced, the concept of aw has been regarded as an index of water availability in food systems. The aw measurement is based on the assumption that, in thermal equilibrium, the partial vapor pressure above the food system is the same as that of water within it. Unfortunately, equilibrium can be considered satisfied only in very dilute food systems (Slade and Levine 1991). Moreover, the limits of the use of this physicochemical parameter describing some structural aspects and phenomena of metastable amorphous systems are recognized. Such systems are better explained in terms of water mobility (Slade and Levine 1991). The approach of combining different techniques has been used to better investigate the mobility of water in several systems (Cornillon 2000; Capitani and others 2003). A calorimetric method, such as differential scanning calorimetry (DSC), and nuclear magnetic resonance (NMR) can offer a different, but complementary perspective in the study of the dynamics of the water-binding process, as well as the water state in food matrices. By DSC, bound water has been traditionally determined as the amount of unfreezable water within a sample after it has been cooled to a low temperature (e.g., −70°C) (Simatos and others 1975). Low-resolution NMR (LR-NMR) has also been proposed by some authors as an alternative technique to use in investigating the state of water in food (Hills 2006). The measurement of the transverse relaxation times (T2) reveals a multicomponent behavior of T2 which reflects the existence of different fractions or populations of water. In the present research, the water state in coffee was studied by the joint use of sorption isotherms, DSC, and LR-NMR. The results obtained by this multianalytic approach were combined to better understand the role of the state of water in the antiplasticization effect.

Materials and Methods Materials Raw coffee beans from a blend of Coffea arabica and Coffea canephora variety Robusta were provided by Saquella Caffè (Pescara, Italy). Coffee was roasted in a rotary laboratory roaster (Probat, Emmerich am Rhein, Germany) with 1-kg capacity at 170°C for 10 min. The texture analysis was performed on single intact coffee beans while sorption isotherms, DSC, and NMR analyses required samples ground in a mill (model Super Junior S; Moulinex, Paris, France). Water-Sorption Isotherm Coffee beans were ground and transferred into glass desiccators, containing phosphorus pentoxide, to dry samples completely. Moisture was equilibrated inside sterilized,

Water State and Mobility Affect the Mechanical Properties of Coffee Beans

493

hermetically sealed hygrostats (glass jars) containing different solutions with percent relative humidity (RH) from 24 to 94 at 22° ± 2°C. Dried samples (1 g) were introduced into 10-mL glass capsules that were left half-open and placed in the hygrostat, above the solution during the experiment. The reaching of equilibrium was determined according to the method of Spiess and Wolf (1983). Samples equilibrated to different aw values (0.24, 0.47, 0.58, 0.73, 0.84, 0.89, and 0.94) were obtained. The aw of the equilibrated samples was checked by an Aqualab (Decagon Devices, Pullman, WA, USA) dew-point hygrometer. The dry-matter content was determined by drying the samples in a vacuum oven (AOAC 1990). Water-sorption isotherms were fitted by the Guggenheim-Anderson-de Boer (GAB) model (Equation 41.1): we =

w0 CKa w

(1 − Ka w ) (1 + (C − 1) Ka w )

(41.1)

where we is the water content (g water/g solids), aw is the water activity, w0 is the monolayer value (g water/g solids), C is the constant related to monolayer sorption heat, and K is the constant related to multilayer sorption heat. Textural Properties Uniaxial compression of single beans was carried out at 22° ± 2°C according to the method of Pittia and others (2007) by using an Instron UTM (universal testing machine) dynamometer model 4301 (High Wycombe, UK) equipped with a 1-kN load cell at a rate of 0.83 cm/s until the bean failed. Fracture force (Newtons [N]) corresponding to the force at the major failure event on the force-displacement curve was considered a measure of strength. To highlight the effect of water on textural properties, data are reported as normalized fracture force (adimensional) obtained by dividing the fracture force of the sample rehydrated at a specific aw value (Fawi) by that of the dried samples (Faw0). DSC Measurements The amount of unfreezable water was evaluated on samples at different water activities by using a Pyris 6 DSC (Perkin Elmer, Wellesley, MA, USA). The differential scanning calorimeter was equipped with a low-temperature cooling unit (Intracooler II, Perkin Elmer). Temperature was calibrated with ion-exchanged distilled water, indium, and zinc while the heat flow was calibrated by using the heat of fusion of indium. In both cases, the same heating rate was adopted for sample measurements and a dry nitrogen gas flux of 20 mL/min was used. Samples (∼20 mg) were weighed in 50-μL aluminum pans, hermetically sealed, and then loaded onto the calorimeter at room temperature. An empty pan of the same type was used as a reference. Samples were then cooled at 5°C/min to −60°C, held for 1 h, and then scanned at 5°C/min to 20°C. Unfreezable water was evaluated as the maximum water content for which no enthalpic peak was detected that was obtained from the intercept at ΔH = 0 of a linear fit of melting enthalpies versus water-content percentages (Capitani and others 2003).

494

PART 2: Poster Presentations

Table 41.1. Guggenheim-Anderson-de Boer (GAB) parameters computed by fitting of sorption isotherms of raw and roasted coffee w0

C

K

R2

Raw

5.28 ± 2.43

0.70 ± 0.64

0.94 ± 0.04

0.992

Roasted

4.23 ± 0.52

1.13 ± 0.43

0.95 ± 0.01

0.999

Coffee sample

w0, monolayer value (g water/g solids); C, constant related to monolayer sorption heat; and K, constant related to multilayer sorption heat.

NMR Relaxation Measurements The transverse relaxation time (T2) of water protons was measured in triplicate for each moisture level. Samples were analyzed at 24°C with the Carr-Purcell-MeiboomGill (CPMG) pulse sequence by using a Minispec PC/20 spectrometer (Bruker, Karlsruhe, Germany) working at 20 MHz. The exponential decay comprised 5000 echoes, an echo time of 80 μs, and a recycle delay of 3.5 s. To maximize the signal/ noise ratio and to prevent signal clipping, the number of scans, as well as the amplification factor, were adjusted according to the moisture content of the samples analyzed. The raw CPMG decays were normalized to the dry-matter weight of the samples before being inverted by the uniform-penalty inversion (UPEN) program (Borgia and others 1998). The application of the UPEN led to a continuous T2 distribution interpretable in terms of different proton pools. The intensity of an NMR pool spanning a certain range of T2s on the relaxogram was obtained from the fraction of the cumulative signal percentage provided by the UPEN in that range, multiplied by the UPEN total extrapolated NMR signal (XSig).

Results and Discussion Water-Sorption Isotherms The water-sorption isotherms were fitted by the GAB equation. The calculated parameters are reported in Table 41.1 for raw and roasted coffee, together with the R2 values. Textural Properties Textural changes of raw (green) and roasted coffee were studied as a function of aw. A similar hardening effect caused by water sorption above the BET monolayer value was noticed, as shown in Figure 41.1, where data on normalized fracture force are reported. Softening caused by water plasticization occurs at high aw values, which were different for raw and roasted coffee. In the case of roasted coffee, the increase in the fracture force, as well as of the fracture energy (Pittia and others 2007), at an increasing degree of hydration occurred in the amorphous state. This hardening effect in the glassy state at a temperature lower than the glass transition temperature (Tg) is referred to as the antiplasticization effect. Upon further hydration, plasticization starts at a moisture content of about 18%–20% (Pittia and others 2007), where the glass transition of roasted coffee occurs at ambient temperature (Geiger and others 2002).

Water State and Mobility Affect the Mechanical Properties of Coffee Beans

495

Normalized fracture force

3 2.5 2 Green

1.5

Roasted

1 0.5 0

0

0.2

0.4

aw

0.6

0.8

1

Figure 41.1. Normalized fracture force of rehydrated green and roasted coffee as a function of water activity (aw).

Figure 41.2. Normalized differential scanning calorimetry (DSC) thermograms of raw coffee, rehydrated at water activity (aw) 0.11 and 0.94. Endo, endotherm; and mW, milliwatts.

On the basis of sorption isotherm data, this moisture content corresponds to the aw of the upward concavity (0.86) in the sorption isotherm when the monolayer is completely hydrated and the state of water changes from hydration water to physically entrapped water. In the case of green coffee, however, plasticization occurs at a lower moisture content corresponding to aw = 0.75, which was lower than that of the upward concavity in the sorption isotherm. DSC Results In Figure 41.2, normalized thermograms of roasted coffee samples rehydrated at aw 0.11 and 0.94 are reported. From DSC analysis, it was noticed that the endothermic

496

PART 2: Poster Presentations

Figure 41.3. Uniform-penalty inversion (UPEN) T2 relaxograms of raw coffee samples rehydrated at different water contents. aw, water activity; a.u., absorbance units; and db, dry basis.

peak of freezable water, which did not appear at low aw values, is otherwise evident at high aw values. The values of freezable water per gram of sample were plotted as a function of moisture content. The amount of unfreezable water, defined as the maximum amount of water present in the samples that is not associated with any endothermic peak, was calculated as the intercept of a linear fit on the x-axis as proposed by Capitani and others (2003). Results indicate that the water adsorbed by the samples can be considered as unfreezable until an aw of 0.79 ± 0.01 (R2 = 0.988) and 0.86 ± 0.01 (R2 = 0.999) is reached for green coffee and roasted coffee, respectively. NMR Results In the continuous T2 relaxograms of green and roasted coffee samples, three different proton populations (hereafter named T2a, T2b, and T2c), characterized by an increased level of mobility, can be observed (Figure 41.3). The major population (T2b) has an average T2, starting from ∼0.1 ms at low hydration levels, that gradually reaches the final value of 1.5 and 1.7 for raw and roasted samples, respectively. The extrapolated signal per gram dry basis (db) of this pool shows a linear increase with moisture content. This was to demonstrate that it can be assigned to the water progressively added to the matrix. By plotting the T2b peak relaxation rate (R2 = T2 − 1) versus moisture (expressed in g H2O/g db), it was observed that a sudden jump toward longer relaxation rates occurred after a certain level of

Water State and Mobility Affect the Mechanical Properties of Coffee Beans

497

hydration (data not shown). The slope break point has been calculated to occur at a different hydration degree for raw and roasted coffee corresponding to the aw values of 0.61 and 0.75. Furthermore, this peculiar shift in the R2 rate occurred approximately along with a width narrowing of the T2b population. This revealed a shoulder (T2a) at about T2 = 0.15 ms, indicating the presence of a new fast-relaxing proton pool. The slowest relaxing component, T2c, can be assigned to the apolar phase in coffee, as previously reported by Mateus and others (2007). In fact, the T2c population seemed not to be influenced by hydration. Its signal per gram dry basis and its T2 value were constant along all the hydration process. Overall, our results suggest the presence of a relatively low-mobility water population (T2b), probably located inside the cell interacting with the cell wall biopolymers, at low- and medium-hydration degrees of coffee beans. This water, which can be thought of as being tightly associated with macromolecules, acted as an antiplasticizing agent, based on the semisolid nature of its signal. The sudden jump of T2b toward longer relaxation rates occurring at aw 0.61 and 0.75 for raw and roasted coffee, respectively, might indicate the beginning of a different water-matrix interaction and water state affecting the mechanical properties. This interpretation was confirmed by the appearance of the T2a population that we ascribe to the solid matrix (of which NMR signal decayed too fast to be detected by CPMG experiments under our conditions). The signal was gradually more and more detectable in the low-T2 region after being hydrated by water. Its mobility was enhanced at these levels of hydration. The aw values at which a water state change occurs as identified by NMR are lower than those determined by DSC analysis (0.79 and 0.86 for raw coffee and roasted coffee, respectively), based on the appearance of the melting peak of freezable water. However, this difference could be attributed to a different sensibility of the two techniques in evaluating the water state.

Conclusions Mechanical properties of coffee beans at different hydration degrees seemed to be related to the state and mobility of water. The rehydration process of dry coffee beans was divided into two different stages. The first one, which ranged from the monolayer value to aw value of 0.76–0.78 for green coffee beans and to 0.86 for roasted coffee beans, could represent the solid-matrix hydration where water was completing the monolayer and interacted with the solid matrix (unfreezable water). This hypothesis is supported by the very short T2 values obtained by NMR, as well as by the concomitant absence of any endothermic peak in the DSC pattern. Under these conditions, water induced hardening, and an antiplasticization effect occurred. The second stage, which took place at higher aw values, corresponded to the end of the solid-matrix hydration when water acted as a plasticizing agent.

References AOAC (Association of Official Analytical Chemists). 1990. Official methods of analysis. n. 950.46. 15th ed. Arlington, VA: AOAC.

498

PART 2: Poster Presentations

Borgia GC, Brown RJS, Fantazzini P. 1998. Uniform-penalty inversion of multiexponential decay data. J Magn Reson 132:65–77. Capitani D, Mensitieri G, Porro F, Proietti N, Segre AL. 2003. NMR and calorimetric investigation of water in a superabsorbing crosslinked network based on cellulose derivatives. Polymer 44:6589–98. Cornillon P. 2000. Characterization of osmotic dehydrated apple by NMR and DSC. LWT Food Sci Technol 33:261–7. Geiger R, Perren R, Schenker S, Escher F. 2002. Mechanism of volume expansion in coffee beans during roasting. In: Proceedings of the 19th International Conference on Coffee Science, 14–19 May 2001. Paris: ASIC. Gondek E, Lewicki PP. 2006. Antiplasticization of cereal-based products by water. Part II: Breakfast cereals. J Food Eng 77:644–52. Hills BP. 2006. Applications of low-field NMR to food science. Annu Rep NMR Spectrosc 58:177–230. Marzec A, Lewicki PP. 2006. Antiplasticization of cereal-based products by water. Part I: Extruded flat bread. J Food Eng 73:1–8. Mateus ML, Champion D, Liardon R, Voilley A. 2007. Characterization of water mobility in dry and wetted roasted coffee using low field proton nuclear magnetic resonance. J Food Eng 81:572–9. Moraru CI, Lee T-C, Karwe MV, Kokini JL. 2002. Plasticizing and antiplasticizing effects of water and polyols on a meat-starch extruded matrix. J Food Sci 67:3396–401. Pittia P, Nicoli MC, Sacchetti G. 2007. Effect of moisture and water activity on textural properties of raw and roasted coffee beans. J Texture Stud 38:116–34. Pittia P, Sacchetti G. 2008. Antiplasticization effect of water in amorphous foods: a review. Food Chem 106:1417–27. Seow CC, Cheah PB, Chang YP. 1999. Antiplasticization by water in reduced-moisture food systems. J Food Sci 64:576–81. Simatos D, Faure M, Bonjour E, Couach M. 1975. Differential thermal analysis and differential scanning calorimetry in the study of water foods. In: Duckworth RD, editor. Water relations of foods. London: Academic. p 193–209. Slade L, Levine H. 1991. Beyond water activity: recent advances based on an alternative approach to the assessment of food quality and safety. CRC Crit Rev Food Sci Nutr 30:115–360. Spiess WEL, Wolf WR. 1983. The results of the COST 90 Project on water activity. In: Jowitt R, Escher F, Hallström B, Meffert HF, Spiess WEL, Vos G, editors. Physical properties of foods. London: Applied Science. p 65–91.

42 Effect of Water Activity on the Release Characteristics of Encapsulated Flavor A. Soottitantawat, H. Yoshii, and T. Furuta

Abstract D-limonene and L-menthol as model flavors were encapsulated in gum arabic or modified starch wall materials by spray drying. The water activity (aw) played an important role in the release rate of the encapsulated flavor. In the case of the waterinsoluble flavor, D-limonene, the release rate increased with the aw up to around 0.50 and then decreased to a minimum release rate at the aw of around 0.70–0.80. The release rate increased again with high aw. That could be explained by the change of structure of the capsule wall materials. On the other hand, for the partially soluble flavor (L-menthol), the release rate increased monotonously with the increasing aw. This might be because of the solubility of L-menthol. For partially water-soluble flavors, dissolved flavors can increasingly diffuse through the wall matrices with increasing aw. The release rate was controlled by the water adsorption by wall matrices and by the flavor solubility.

Introduction Flavor microencapsulation is a technology of enclosing flavor compounds in a carrier matrix to provide dry and free-flowing powders. Two advantages of this technology are the following: it protects against degradative reactions and prevents the loss of flavors during storage. In addition to the primary means of flavor stabilization and protection, a controlled or extended flavor release at an appropriate time may also be required. The release characteristics of encapsulated flavors in powder matrices are also important for estimating the shelf life of products, as well as for controlled-release applications in foods. Several mechanisms have been proposed for the controlled release of flavor. Whorton (1995) provided an overview of the basic release mechanisms in various capsule matrices. Flavor release from the spray-dried powder may be considered to be associated with the diffusion mechanisms of both flavor and water, because the solubilization of the wall matrices by water would be followed by subsequent release of the encapsulated flavor.

A version of this article has been published as Soottitantawat A, Yoshii H, Furuta T, Ohgawara M, Forssell P, Partanen R, Poutanen K, Linko P. 2004. Effect of water activity on the release characteristics and oxidative stability of D-limonene encapsulated by spray drying. J Agric Food Chem 52:1269–76.

499

500

PART 2: Poster Presentations

Rosenberg and others (1990), who observed that the release of an encapsulated flavor ester during storage increases with an increase in relative humidity (RH), suggested that water uptake at a high RH destroys the structure of the encapsulation matrix. Whorton and Reineccius (1995) evaluated the mechanism associated with the controlled release of the flavor encapsulated in maltodextrins of different dextrose equivalents. The flavor release increased with the increase in water activity (aw) up to the point where collapse of the matrix occurred. Yoshii and others (2001) also showed the higher release rate of ethyl-n-butyrate with higher RH. Soottitantawat and others (2004) showed the contrasting results of the effect of aw on the release of encapsulated D-limonene. They proposed that the mechanisms of the higher release rate of flavor occurred near the glass transition temperature of the carrier matrix. Since these reported results are not in agreement with one another, a better understanding of the effect of the aw on the release of encapsulated flavor would be useful in the quality control and applications of these powders. The present study focused on the effect of RH on the release characteristics of encapsulated insoluble flavor D-limonene and partly soluble L-menthol, which were prepared by spray drying. The Avrami equation was used to fit the course of the release time of encapsulated flavor in order to explain the mechanism and determine the release-rate constant.

Materials and Methods Materials D-limonene and gum arabic were purchased from Nacalai Tesque (Kyoto, Japan). L-menthol was purchased from Soda Aromatic (Tokyo, Japan). Maltodextrin with ∼20 dextrose equivalent (Amycol no. 1) and modified starches (Capsul and Hi-Cap 100) were gifts from Nippon Starch Chemicals (Osaka, Japan) and Nippon NSC (Tokyo, Japan), respectively. The organic chemicals used in the analyses were of analytic grade. Encapsulated Flavor Powder Preparation Wall materials (gum arabic, Capsul, Hi-Cap 100, or maltodextrin) were added to distilled water to obtain a 10%–40% wt/wt mixture, depending on the solubility of each material, and allowed to hydrate overnight. The model flavor of D-limonene or L-menthol was added to the 25% wt/wt (dry basis) solution. Then, an emulsion was prepared by homogenizing the mixture of the model flavor and the carrier solution by using a Polytron homogenizer (PT-10; Kinematica, Littau, Switzerland), followed by a microfluidizer (110T; Microfluidics, Newton, MA, USA) at 12,000 psi (pound-force per square inch [82.8 MPa]). For L-menthol, the mixture was heated to 60°C before homogenization. The emulsified flavor liquid was dried in a spray dryer (L8; Okawara, Yokohama, Japan) equipped with a centrifugal atomizer. The operational conditions of the spray drying were as follows: inlet temperature of air, 200°C; outlet temperature of air, 110° ± 10°C; feed rate, 45 mL/min; airflow rate, 110 kg/h (at outlet temperature); and atomizer rotational speed, 30,000 rpm.

Effect of Water Activity on the Release Characteristics of Encapsulated Flavor

501

Release of Encapsulated Flavor from the Spray-Dried Powder One-tenth of a gram of each dried powder was weighed and spread in a thin layer in a 15-mL glass bottle (20 φ × 48 mm), which was placed in a desiccator with saturated salt solution to maintain a constant RH at 23% ± 5%, 33% ± 5%, 51% ± 5%, 75% ± 5%, 83% ± 5%, and 96% ± 5% (Rockland 1960). Fifteen sample bottles were placed in a desiccator to study the release and oxidation kinetics for 25 or 30 days. Furthermore, humid air of the same RH was blown into the desiccator to purge the gas for 10 min at 6-h intervals. At prescribed time intervals, bottles were picked out of the desiccators and then, using solvent flavor extraction and gas chromatographic analysis, the residual amounts of D-limonene were measured (Soottitantawat and others 2004).

Results and Discussion Effect of Relative Humidity on the Release of D-Limonene from the Powder The release time courses of D-limonene from the spray-dried powder were measured at 50°C and at 23%, 51%, 75%, and 96% RH, as shown in Figure 42.1. The RH greatly affected the release rate of D-limonene. The dependence, however, was not simple. Considering only 23% and 96% RH, the release of D-limonene increased with increasing RH. However, when comparing the 51% and 75% RH, the release of D-limonene at 51% RH was higher than that observed at 75% RH. These results suggest that the release of D-limonene is closely related at least to the aw of the powder, which is in agreement with many earlier observations.

(a)

(b)

(c)

D-limonene retention

1.0

0.8

0.6

0.4

0.2

0

0

5

10 15 Time (day)

20

0

5

10 15 20 Time (day)

0

5

10 15 20 Time (day)

Figure 42.1. Release time courses of encapsulated D-limonene in spray-dried powders stored at various relative humidities (RHs) and 50°C: (a) blend of gum arabic and maltodextrin, (b) blend of Hi-Cap 100 and maltodextrin, and (c) Hi-Cap 100. , 23% RH; , 51% RH; , 75% RH; and , 96% RH.

502

PART 2: Poster Presentations

(a)

(b)

(c)

L-menthol retention

1.0

0.8

0.6

0.4

0.2

0

0

5

10

15

Time (day)

20

0

5

10

15

Time (day)

20

0

5

10

15

20

Time (day)

Figure 42.2. Release time courses of encapsulated L-menthol in spray-dried powders stored at various relative humidities (RHs) and 50°C: (a) gum arabic, (b) Capsul, and (c) Hi-Cap 100. , 33% RH; , 51% RH; , 75% RH; and , 83% RH.

Effect of Relative Humidity on the Release of L-Menthol from the Powder The release time course of L-menthol in spray-dried powder was measured at 45°C and at 33%, 51%, 75%, and 83% RH, as shown in Figure 42.2. The RH showed a pronounced effect on the release rate of L-menthol, especially in gum arabic and Hi-Cap 100 wall matrices. On the other hand, the effect of RH on the release of Lmenthol from Capsul matrix could not be clearly observed. Analysis of the Release Rate and Release Mechanism by the Avrami Equation To study the release characteristics of encapsulated flavor in response to various influence factors, the release-rate constants were evaluated. According to Yoshii and others (2001), the time courses of flavor release were correlated with the Avrami equation (Equation 42.1) as shown by the solid lines in Figures 42.1 and 42.2. R = exp ⎡⎣ − ( kt ) ⎤⎦ n

(42.1)

Here R is the retention of flavor, t is the storage time, k is the release-rate constant, and n is a parameter representing the release mechanism. The relation of the release-rate constant (k) and n values compared with aw (aw = RH/100) are shown in Figures 42.3 and 42.4, respectively. For the insoluble flavor of D-limonene in Figure 42.3a, the release-rate constants first increased with the increase in aw and then decreased at aw of ∼0.70. At a higher aw range, the release rate tended to increase again because the powder matrices were destroyed. These results were discussed clearly by Soottitantawat and others (2004). At a low aw (≈0.23), the

Effect of Water Activity on the Release Characteristics of Encapsulated Flavor

(b) 350

Release-rate constant ¥ 103 (day-1)

Release-rate constant ¥ 103 (day-1)

(a)

503

Blend of GA-MD Blend of HI-Cap 100–MD HI-Cap 100

300 80 60 40 20 0 0

0.2

0.4

0.6

0.8

350

GA Capsul HI-Cap 100

300 80 60 40 20 0

1.0

0

0.2

Water activity

0.4

0.6

0.8

1.0

Water activity

Figure 42.3. Effect of water activity on the release-rate constant of encapsulated Dlimonene (a) and L-menthol (b) at 50°C. The error bars indicate 95% confidence. GA, gum arabic; and MD, maltodextrin.

(b)

(a)

1.0

Blend of GA-MD Blend of Hi-Cap 100-MD Hi-Cap 100

0.8

Release-rate constant ¥ 103 (day-1)

Release-rate constant ¥ 103 (day-1)

1.0

0.6

0.4

0.2

0

0

0.2

0.4

0.6

Water activity

0.8

1.0

GA Capsul Hi-Cap 100

0.8 0.6 0.4 0.2 0

0

0.2

0.4 0.6 Water activity

0.8

1.0

Figure 42.4. Effect of water activity on the n values for the release of encapsulated D-limonene (a) and L-menthol (b) from spray-dried powder at 50°C storage. The error bars indicate 95% confidence. GA, gum arabic; and MD, maltodextrin.

spray-dried powders were still in the glassy state, resulting in low mobility of Dlimonene molecules in the glassy state of the capsule matrices. The higher rate was observed at aw of ∼0.5 because of the higher mobility of flavor molecules near the glass transition temperature of the carrier. When aw increased to ∼0.7, the powders began to be rehydrated. At this stage, it may be assumed that the effective surface-area

504

PART 2: Poster Presentations

decrease resulted in a decrease in D-limonene evaporation from the surface of the powder particles. Most particles were detected to be clumped and adhered together in a pastelike mass, or the rubbery form of the capsule matrices. At a very high aw, the release rate increased because the flavor droplet was crushed when the capsule matrix was destroyed. On the other hand, for the partly soluble flavor of L-menthol, at a low aw, only a small amount of encapsulated L-menthol was released from the spray-dried powder. Then, the release-rate constant dramatically increased with increasing aw, especially with gum arabic and Hi-Cap 100, when aw was higher than 0.5. Similar results were also reported in our previous work (Yoshii and others 2001) for the increase in release rate of encapsulated ethyl-n-butyrate from wall matrices with an increase of RH. That might be explained by the change in matrix structure, as mentioned by Rosenberg and others (1990) and Soottitantawat and others (2004). As long as the individual structure of the capsule remains intact, high retention of volatile is maintained. At low aw, the capsule matrix is still in the glassy state. The slower release in the low-aw region is most likely due to the lower mobility of flavor molecules in the glassy state of the capsule matrix, as already mentioned. Once the capsule structure is damaged by water uptake, the release rates increase, possibly because of the higher mobility of the flavor molecules (i.e., the capsule matrix starts to be plasticized). However, contrasting results were reported for the water-insoluble flavor, D-limonene. For the soluble flavor, although the pastelike mass of encapsulated powder was also observed at aw of ∼0.75, a reduction of the release-rate constant was not observed, as shown in Figure 42.3b. This might be because of the water solubility of L-menthol. For partially watersoluble flavors, dissolved flavors can increasingly diffuse through the wall matrices with increasing aw, resulting in a higher release of flavors. The Avrami parameter n relates to the release mechanism. As indicated in previous work (Whorton 1995; Whorton and Reineccius 1995), n = 1 represents the first-order release mechanism, and n < 1 (theoretically, n = 0.54) means that the molecular diffusion of D-limonene is rate limiting. For n > 1, the release is rapid, with an induction period. The influence of aw on n is shown in Figure 42.4. The release of encapsulated insoluble D-limonene is shown in Figure 42.4a. The n values were between around 0.30 and 0.80 over a wide range of aw. This implies that the release of encapsulated D-limonene was generally controlled by the diffusion mechanism. However, for the partly soluble L-menthol shown in Figure 42.4b, the values of n for all wall materials are in the range of 0.10–1.00. The n values increased with increasing aw. This indicates that the release of encapsulated, partly soluble L-menthol is controlled by the diffusion mechanism through the wall of the particles and becomes the first-order release mechanism at high aw.

Conclusion For the insoluble flavor, D-limonene, the release rate increased with increasing aw in the low-aw range (0.10–0.50), then decreased around the 0.70–0.80 aw range, and dramatically increased again in the high aw (0.96). This is explained by the change in

Effect of Water Activity on the Release Characteristics of Encapsulated Flavor

505

structure from the glassy state to the rubbery state in the range of 0.10–0.50 aw. The decrease of the release rate was explained by the formation of a pastelike mass of powder in the rubbery state, resulting in the decrease in effective surface area for release. On the other hand, with the partially water-soluble flavor, L-menthol, the release-rate constant increased with increasing water. This might be because of the solubility of L-menthol. For partially water-soluble flavors, dissolved flavors can increasingly diffuse though the wall matrices with increasing aw. For insoluble flavors, the release rate was controlled by the change in matrix structures. On the other hand, for soluble flavors, the release rate was controlled by the water adsorption by wall matrices and by the solubility of flavors.

Acknowledgments We acknowledge Nippon Starch Chemicals (Osaka, Japan) and Nippon NSC (Tokyo, Japan) for their kind gift of maltodextrin and modified starches (Capsul and Hi-Cap 100), respectively.

References Rockland LB. 1960. Saturated salt solution for static control of relative humidity between 5°C and 50°C. Anal Chem 32:1375–6. Rosenberg M, Kopelman IJ, Talmon Y. 1990. Factors affecting retention in spray-drying microencapsulation of volatile materials. J Agric Food Chem 38:1288–94. Soottitantawat A, Yoshii H, Furuta T, Ohgawara M, Forssell P, Partanen R, Poutanen K, Linko P. 2004. Effect of water activity on the release characteristics and oxidative stability of d-limonene encapsulated by spray drying. J Agric Food Chem 52:1269–76. Whorton C. 1995. Factors influencing volatile release from encapsulation matrices. In: Risch SJ, Reineccius GA, editors. Encapsulation and controlled release of food ingredients. American Chemical Society (ACS) Symposium Series 590. Washington, DC: ACS. p 134–42. Whorton C, Reineccius GA. 1995. Evaluation of the mechanisms associated with the release of encapsulated flavor materials from maltodextrin matrices. In: Risch SJ, Reineccius GA, editors. Encapsulation and controlled release of food ingredients. American Chemical Society (ACS) Symposium Series 590. Washington: ACS. p 143–60. Yoshii H, Soottitantawat A, Liu XD, Atarashi T, Furuta T, Aishima S, Ohgawara M, Linko P. 2001. Flavor release from spray-dried maltodextrin/gum arabic or soy matrices as a function of storage relative humidity. Innov Food Sci Emerg Technol 2:55–61.

43 Water and Protein Modifier Effects on the Phase Transitions and Microstructure of Mung-Bean Starch Granules P. Hongsprabhas and K. Israkarn

Abstract This study investigated the roles of protein modifiers (i.e., 25 mM cysteine and 25 mM cysteine plus 50 mM calcium lactate) in the thermal characteristics of mung-bean (MB) starch. Although proteins were present as only 0.16% of the starch granule, the alterations of the reactive groups in protein molecules by cysteine as reducing agent, and by calcium lactate as cross-linking agent, increased the onset temperature (To), peak temperature (Tp), and end temperature (Te) of the first endotherm when the water content was above 50% (wt/wt). The increase in water content from 40% to 60% lowered To, Tp, and Te of the first endotherm only when MB starch was heated in systems containing 25 mM cysteine plus 50 mM calcium lactate (P < 0.05). At 40% and 50% (wt/wt) water content, the second endotherm of MB starch occurred within the range of 147°–173°C. Its phase-transition temperatures were unaffected by water content (P ≥ 0.05). They were lowered below those of MB starch only in the systems containing 25 mM cysteine. The findings from confocal laser scanning microscopy suggested at least three separate microphases dispersed in the MB starch network: the closely packed swollen granule; the protein phase at the inter-phase between adjacent swollen granules; and the randomly distributed, separate starch-rich phase.

Introduction Generally, gelatinized starch paste and gel are polydispersed in nature. They contain swollen granules, granule ghosts, leached amylose, randomly oriented and open-chain amylose, amylose-rich starch remnants, granular amylopectin in the remnants, and disrupted remnants (Hermansson and Svegmark 1996). In addition, the proteincontaining envelope of some plants (e.g., mung bean and cassava) generated during gelatinization plays an important role in determining the granule’s ability to retain starch content after heat treatment (Israkarn and others 2007). Mung bean (Vigna radiata [L.] Wilczec) (MB) is a major starch crop in Thailand. Its starch contains high amylose content, and the gelatinized starch is prone to retrogradation and can form resistant starch (Biliaderis and others 1980; Kasemsuwan and others 1998). The limited swelling capability of legume starches, such as MB, was reported to be influenced by the existence of starch-granule peptide bridges that maintain the structure of the granule ghosts (Oates 1990). 507

508

PART 2: Poster Presentations

By definition, starch granule–associated proteins are defined as the proteins naturally positioned in and on starch granules (Baldwin 2001). They are different from storage proteins and are bound tightly to the surface and/or are integrated constituents within the granule structure. These proteins are mainly starch biosynthetic enzymes whose molecular weights are around 5–149 kDa (Baldwin 2001). Although present in minute quantities, starch granule–associated proteins influence the rheological properties of starch paste, as shown in maize starch (Han and others 2002). This study explored the changes in thermal and architectural characteristics of starch that are induced by protein fractions when water is limited. The molecular interactions of protein fractions were altered by the addition of cysteine and calcium lactate at a pH of around 5–6 by using a method described by Israkarn and others (2007). The insights may help in designing the physicochemical properties of starch granules used in the fabrication of thermoplastic bio-based materials.

Materials and Methods Food-grade MB starch (Pine Brand; SithiNan, Bangkok, Thailand) was obtained from a local supermarket. Rhodamine B (Invitrogen, Carlsbad, CA, USA) was used to stain protein. The MB contained 11.35% moisture, 0.16% protein, 0.08% ash, and 0.08% crude fiber, with no detectable lipid (AOAC 2000). Thermal Properties Differential scanning calorimetry (DSC) was used to determine the thermal properties of food-grade starches. Briefly, a differential scanning calorimeter (DSC 822e/400 W; Mettler Toledo, Schwerzenbach, Switzerland) was used to characterize the thermal properties of 40%–60% (wt/wt) water content in different dispersants. The starches were dispersed in distilled water, 25 mM cysteine solution, or 25 mM cysteine plus 50 mM calcium lactate solution prepared as described by Israkarn and others (2007). The suspension was incubated at 25°C for 24 h in a hermetically sealed stainless-steel pan prior to the measurements. The samples were heated at a rate of 5°C/min from 25° to 210°C to determine the transition temperature and enthalpy of gelatinization. The transition temperatures reported were the onset (To), peak (Tp), and end (Te) temperatures of the gelatinization endotherm. The enthalpy of the gelatinization (ΔH) was estimated by integrating the area between the thermogram and a baseline connecting the points of onset and end temperature and is expressed in J/g starch (dry basis). Gel Microstructure MB starch slurries (1 mL; 40%–60% [wt/wt] water content) were prepared by dispersing the starch in the aforementioned dispersants in 1.5-mL Eppendorf tubes, incubated at 27°C for 24 h, and autoclaved at 121°C for 15 min. The starch gels were sectioned (∼2 mm thick) by razor blade. A solution of rhodamine B (0.01% in 95% ethanol) was added to the MB starch gel section. The excess dye was washed off with distilled water after incubation for

Water and Protein Modifier Effects on Mung-Bean Starch Granules

509

5 min. Each sample was observed for the location of fluorescent-labeled protein by confocal laser scanning microscopy (CLSM) (Axio Imager MI; Carl Zeiss, Göttingen, Germany). A helium-neon (HeNe) laser with an excitation wavelength of 543 nm was used. CLSM digital images were acquired by using an LSM (laser scanning microscope) 5 Pascal program (Carl Zeiss), whereas the digital images from the same location observed under cross-polarized light were acquired by using the Image-Pro Plus program (MediaCybernetics, Bethesda, MD, USA). Statistical Analysis The experiments were conducted in two separate trials, each run in duplicate. The data were subjected to analysis of variance (ANOVA) with significance at P < 0.05. Significant differences among mean values were determined by the Duncan multiplerange test. All statistical analyses were performed using SPSS software version 12 (SPSS, Chicago, IL, USA).

Results and Discussion The presence of a protein modifier altered the transition temperatures of MB starch significantly at intermediate water content (i.e., 40%–60% wt/wt) (P < 0.05) but had no effect on gelatinization enthalpy. The addition of 25 mM cysteine plus 50 mM calcium lactate increased the phase-transition temperature of MB starch at 60% water content (Table 43.1). However, at the water content of 40% (wt/wt), the influences

Table 43.1. Effect of water content and protein modifiers on phase-transition characteristics of mung-bean starch in the first endotherm Water content (% wt/wt)

Protein modifier No modifier

+25 mM cysteine

+25 mM cysteine + 50 mM calcium lactate

40

69.02aA

70.31aA

72.09aA

50

69.24

bA

60

68.55

bA

40 50

To abA

71.86aA

bA

69.14

71.21aB

73.04aA

73.59aA

75.51aA

72.70

aA

aA

75.05aB

60

71.76

bA

bA

72.43

74.39aC

40

80.34aA

80.27aA

82.35aA

50

78.35

aA

aA

80.33aB

77.68

bA

abA

80.05aB

69.92

Tp 73.70

Te

60

79.59

78.30

To, onset temperature; Tp, peak temperature; and Te, end temperature. Means in the same column that are followed by a different uppercase superscript in each To, Tp, or Te set are significantly different(P < 0.05). Means in the same row that are followed by a different lowercase superscript in each To, Tp, or Te set are significantly different (P < 0.05).

PART 2: Poster Presentations

(b)

200

MB: no modifier MB: 25 mM cysteine MB: 25 mM cysteine+ 50 mM calcium lactate

100

0

no peak

40 50 60 Total water content (% wt/wt)

2nd Endotherm: Tp (°C)

2nd Endotherm: To (°C)

(a)

(b)

200

MB: no modifier MB: 25 mM cysteine MB: 25 mM cysteine+ 50 mM calcium lactate

100

no peak

0

40 50 60 Total water content (% wt/wt)

2nd Endotherm: Te (°C)

510

200

MB: no modifier MB: 25 mM cysteine MB: 25 mM cysteine+ 50 mM calcium lactate

100

0

no peak

40 50 60 Total water content (% wt/wt)

Figure 43.1. Effect of modifier on transition temperatures (To, Tp, and Te) in the second endotherm at different levels of water content. MB, mung bean; To, onset temperature; Tp, peak temperature; and Te, end temperature.

(a)

(b)

50 μm

50 μm

Figure 43.2. Light micrograph of mung-bean starch granules under (a) visible light and (b) cross-polarized light. Bar = 50 μm.

of protein modifiers ceased. In addition, the second endotherm (within the range of 147°–173°C) was observed only when the water content was less than 60% (wt/ wt) (Figure 43.1). The presence of 25 mM cysteine lowered the To, Tp, and Te of the second endotherm (P < 0.05) but not the enthalpy (P ≥ 0.5), whereas the addition of 25 mM cysteine plus 50 mM calcium lactate to MB starch slurries had no effect (P ≥ 0.05). It should be noted that the temperature range of the second endotherm was higher than that of the amylose-lipid complex and was apparently affected by protein modifiers. MB starch granules are round to oval, their size ranging from 7 to 28 μm (Figure 43.2). The unheated MB starch showed birefringence under cross-polarized light. CLSM illustrated the closely packed structure of the entire network, which was composed of swollen MB starch granules (Figure 43.3). The protein fractions, fluoresced in white under CLSM, were observed in all samples prepared in different dispersants (Figure 43.3). The proteins were retained in the swollen starch granules and existed at the envelope observed as a strong signal surrounding the granules.

Water and Protein Modifier Effects on Mung-Bean Starch Granules

(a)

(b)

50 μm

511

(c)

50 μm

50 μm

Figure 43.3. Confocal laser scanning micrographs of mung-bean starch gel containing 60% water content (wt/wt) prepared by autoclaving at 121°C for 15 min: (a) mung-bean starch, (b) mung-bean starch in the presence of 25 mM cysteine; and (c) mung-bean starch in the presence of 25 mM cysteine and 50 mM calcium lactate. Bar = 50 μm.

The presence of proteins, which were locally concentrated on the envelope, is likely to alter the surface properties of swollen starch granules in terms of charge characteristics and the number of thiol groups. The addition of protein modifiers slightly altered the appearance of the entire network. The dark area (nonprotein phase) was randomly distributed in the matrix with different degrees of continuity. This periodic shape was likely a starch-rich phase, possibly leached amylose, which was separated from the closely packed swollen granules via segregative phase separation or thermodynamic incompatibility with proteins at the inter-phase. In the presence of 25 mM cysteine plus 50 mM calcium lactate, the periodic shape of the dark area was larger and had more continuity than the others. This suggests that the presence of 25 mM cysteine plus 50 mM calcium lactate increased the permeability of the granular envelope and leaching of starch content to a greater extent than in the presence of cysteine alone, although the mixture of 25 mM cysteine plus 50 mM calcium lactate did not alter phase transition temperatures significantly (Table 43.1). The light micrographs (Figure 43.4a and c) illustrate that the densely packed swollen granules in the network had lost their birefringence characteristics because of the high temperature (i.e., 121°C for 15 min) during gel preparation. The CLSM also revealed the continuity of the protein network fluoresced in white under the same field of observation (Figure 43.4b). In the area in which the MB starch-rich fraction showed a short rodlike structure (Figure 43.4a), the fluorescent signal (Figure 43.4b) was very low, suggesting that the structure was a nonprotein entity. This structure could be the amylose-rich phase observed under lower magnification (Figure 43.3). This study demonstrates the influences of protein modifiers on the starch network architecture at intermediate water content. Cysteine is a reducing agent that reduces the existing disulfide bond to sulfhydryl groups in the protein fractions. The lower phase transition of the whole starch granule at sufficient water content (i.e., up to 60% [wt/wt] for MB starch) may result from the alteration of the granule-associated

512

PART 2: Poster Presentations

(a)

(b)

50 μm

(c)

50 μm

50 μm

Figure 43.4. Micrographs of mung-bean starch gel containing 50% water content (wt/wt) prepared by autoclaving at 121°C for 15 min under the same observation field: (a) a light micrograph, (b) a confocal laser scanning micrograph, and (c) a light micrograph under the polarized light. Bar = 50 μm.

proteins located within the granule, in addition to the loss of the crystalline structure of the starch fraction, alone. Calcium lactate, on the other hand, reacts with the carboxylic groups of proteins and cross-links the adjacent negatively charged molecules. The presence of calcium lactate in addition to cysteine could modify the permeability of the envelope further by altering the nature of pores in the envelope chemically and physically. It is likely that both starch and protein molecules in the pores, and possibly the channels (Huber and BeMiller 2000; Han and Hamaker 2002; Fannon and others 2004; Han and others 2005), facilitate the permeation of starch content from the granule interior to the environment. Overall, the CLSM suggested at least three separate microphases dispersed in the MB starch network: the closely packed swollen granule, the protein phase at the inter-phase between adjacent swollen granules, and the randomly distributed, separate starch-rich phase. In the presence of both cysteine and calcium lactate, the separate starch-rich microphase was more continuous than the others, both of which showed the periodic short rodlike entities scattered randomly in the network.

Water and Protein Modifier Effects on Mung-Bean Starch Granules

513

Conclusions The molecular alterations of granule-associated proteins can be used to alter the thermal and microstructural properties of MB starch granules and the MB starch network at water content above 50% (wt/wt). At 40% water content, the role of protein modifiers on MB starch was less significant than the role of water. Nevertheless, the protein-protein and protein-starch interactions at the inter-phase of the swollen granules still need further clarification, particularly those in the presence of 25 mM cysteine plus 50 mM calcium lactate, for the fabrication of thermoplastic bio-based materials.

Acknowledgment The financial support of the National Center for Genetic Engineering and Biotechnology (project BT-B-01-CG-11-5002), Thailand, is gratefully acknowledged.

References AOAC (Association of Official Analytical Chemists). 2000. Official methods of analysis. 17th ed. Arlington, VA: AOAC. Baldwin PM. 2001. Starch granule-associated proteins and polypeptides: a review. Starch/Stärke 53:475–503. Biliaderis CG, Maurice TJ, Vose JR. 1980. Starch gelatinization phenomena studied by differential scanning calorimetry. J Food Sci 45:1669–74. Fannon JE, Gray JA, Gunnawan N, Huber KC, BeMiller JN. 2004. Heterogeneity of starch granules and the effect of granule channelization on starch modification. Cellulose 11:247–54. Han X-Z, Benmoussa M, Gray JA, BeMiller JN, Hamaker BR. 2005. Detection of proteins in starch granule channels. Cereal Chem 82:351–5. Han X-Z, Campanella OH, Guan H, Keeling PL, Hamaker BR. 2002. Influence of maize starch granuleassociated protein on the rheological properties of starch paste. Part II: Dynamic measurements of viscoelastic properties of starch pastes. Carbohydr Polym 49:323–30. Han X-Z, Hamaker BR. 2002. Association of starch granule proteins with starch ghosts and remnants revealed by confocal laser scanning microscopy. Cereal Chem 79:892–6. Hermansson A-M, Svegmark K. 1996. Developments in understanding of starch functionality. Trends Food Sci Technol 7:345–53. Huber KA, BeMiller JN. 2000. Channels of maize and sorghum starch granules. Carbohydr Polym 41:269–76. Israkarn K, Hongsprabhas P, Hongsprabhas P. 2007. Influences of granule-associated proteins on physicochemical properties of mungbean and cassava starch granules. Carbohydr Polym 68:314–22. Kasemsuwan T, Bailey T, Jane J. 1998. Preparation of clear noodles with mixtures of tapioca and high amylose starches. Carbohydr Polym 32:301–12. Oates CG. 1990. Evidence for protein crosslinks in mung-bean starch. In: Phillips GO, Wedlock DJ, Williams PA, editors. Gums and stabilisers for the food industry. Oxford: IRL. p 203–6.

44 Evaluation of the Disintegration and Diffusion of Pharmaceutical Solid Matrices by Image Processing and Nonlinear Dynamics D. I. Téllez-Medina, A. Ortíz-Moreno, J. J. Chanona-Pérez, L. Alamilla-Beltrán, and G. F. Gutiérrez-López Abstract The objective of this work, using pharmaceutical and maltodextrin matrices as experimental models, was to study the disintegration and diffusion of solid matrices into water, and the involved interphases generated during these phenomena by means of image and fractal analysis and by the application of nonlinear dynamics. Special diffusion cells were used to observe the disintegration and diffusion of aspirin and of maltodextrin compressed matrices in degasified distilled water without agitation. Digital images were captured at different time intervals during the disintegration phenomena. Images from different perspectives were processed using the appropriate software measuring the fractal dimension (df) values of the perimeter of the matrices. The fractal dimension of the lateral face of the matrices, the number of layers, and the velocity and behavior of turbulences and interphases were observed, and the data were statistically analyzed.

Introduction Nonlinear dynamics has helped in the description of the kinetic behavior of several phenomena, such as crystallization and diffusion, contributing to our understanding of, and improvements in processes and new technologies (Alligood and others 2000). Fractal geometry is a valuable tool that enables us to treat complicated questions in a simpler manner; for example, the quantitative characterization of the outline, surface, and structure of irregular materials (Peleg 1993; Chanona and others 2003). This has been useful for evaluating the complexity of biological systems. Here we study the disintegration and diffusion of solid matrices into a liquid medium by means of image analysis, determination of fractal dimension of outlines and surfaces, and nonlinear dynamics.

Methods In a specially constructed diffusion cell that eliminates vibration, air movement, and reflections, the disintegration and diffusion of aspirin and maltodextrin matrices with a strength of 4.5 kgF prepared by direct compression in degasified distilled water and 515

516

PART 2: Poster Presentations

df

1.8 1.7 1.6 1.5 1.4 1.3 1.2 1.1 Aspirin® Maltodextrin 1.0 0.9 0 2 4 6 8 10 12 14 16 18 202224262830 Time (s)

Figure 44.1. Average fractal dimension (df) of the edge of aspirin and maltodextrin matrixes within the first 30 s of their disintegration and diffusion. These images were analyzed to calculate df.

without agitation, and the interphases generated to obtain matrices, were observed. With a computerized video capture system, digital images were captured at different times during the diffusion process. Using Adobe PhotoShop 7.0, Sigma Scan 5.0, and ImageJ 1.37 software, images were processed, and the fractal dimension (df) values of the matrices’ perimeter were calculated. The images were obtained from different perspectives and with cycling illumination orientations (top, tangential, and bottom), which allowed for evaluation of the fractal dimension of the side face of the matrices, the number of formed layers, and the velocity and behavior of observed turbulence and generated interphases. The data were statistically analyzed with Sigma Stat 2.0 and Minitab 13 software.

Results and Discussion In the case of aspirin matrices, the edge irregularity rapidly increases, as indicated by an increase in df. Maltodextrin matrices do not increase markedly in edge irregularity during their disintegration. In the first 12 s of the process, df grows continuously for the edges of both aspirin and maltodextrin matrices (see Figure 44.1), but the increment is greater for aspirin than for maltodextrin matrices. After the first 12 s, the trend changes: irregularity of matrix edges shows a period of stabilization; that is, df starts to stabilize in the range of 1.55–1.62. After 20 s, df follows statistically identical kinetics for both types of matrix.

Image Processing and Nonlinear Dynamics

(a)

1.30

517

(b)

1.25 1.20

1st

1.15

df

1.10

1.05 1.00

Aspirin® Maltodextrin

0.95

2nd

0.90 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 Time (s)

Figure 44.2. (a) Average fractal dimension (df) of the side face of aspirin and maltodextrin matrices within the first 30 s of their disintegration and diffusion. (b) Tangential perspective of a maltodextrin matrix in which three layers can be distinguished.

Differences in df kinetics could be attributed primarily to the chemical composition of the two types of matrix, although other factors, such as variations in compression pressure (during manufacture) and moisture content of the matrix, could also contribute to this difference (Parrot 1970; Banker and Anderson 1986; Pratt 1986; Abdou and others 2001). The df corresponds to the lowest layer of the matrix, in a tangential perspective. Moreover, this layer is affected by the superior ones and shows a rapid increase in the irregularity that is followed by compensation and persistence in the df after the first 16 s. In all cases, three layers were quantified, and collectively they provide a value up to 1.2 for the df of the side face of the matrices (Figure 44.2). By means of top perspective and bottom illumination, and using the diffusion cell, the disintegration process was visualized as a group of concentric fractures (like rings) that begin to form and separate from one another (like flower petals) when water diffuses into the matrix. Then concentric structures commence to fall from the matrix center to the periphery (like dominoes). This mechanism results in the quantified layers observed (see Figure 44.3). For maltodextrin and aspirin, 4 s and 8 s after the matrices were placed in the cell, respectively, vortices were observed around them. In the captured images (Figures 44.3 and 44.4), a halo can be identified between the matrix edge and the zone of vortices. This halo is probably a boundary layer as has been described for a cylindrical shape within a fluid stream (Fischer and others 1979; Treybal 1988; McCabe and others 2002). The halo exhibits a width of 3.18 ± 0.37 mm for aspirin and 3.47 ± 0.41 mm for maltodextrin matrices, but these values are valid for only the particular cell employed. With 30-cm-diameter recipients, the zone between the vortices region and matrix edge has a width of 11.26 ± 0.51 mm (aspirin) and 8.33 ± 0.50 mm (maltodextrin).

Figure 44.3. Images of an aspirin matrix in degasified distilled water and without agitation, captured from a top perspective and illumination on the bottom of the cell. There is a series of concentric fractures, like rings, at 16 and 20 s. Also visible are the halo and vortices around the matrix.

Vortices with opposite-sense movement

Interphase solid-water

Boundary layer

Figure 44.4. Scheme of the vortices arrangement in opposite-sense movement around the halo formed (boundary layer). The image is of an aspirin matrix at 16 s after being placed in degasified distilled water and without agitation.

518

Image Processing and Nonlinear Dynamics

519

These vortices show an alternating succession or opposite-sense; that is, a vortex with clockwise movement is between two vortices with counterclockwise movement. Particles leave the matrix edge with an average velocity of 0.06 m/s and traverse the halo to enter the zone of vortices. In the interior of these vortices, most of the particles describe a negative spiral (toward vortex center), and the remainder move to adjacent fluid further from the matrix. This perhaps reflects the water flows generated around the matrix as a consequence of the entrance and exit of water. Vortex movement could be characterized as exhibiting an average frequency of 432 ± 58 rpm, an average tangential velocity on the order of 0.058 ± 0.001 m/s, and average centripetal acceleration of 2.69 ± 0.43 m/s2 (the intervals refer to respective standard errors). These values appear somewhat high, noting that particle sizes and mass are very low. Considering the observed periodical movement in two directions (x- and y-axes) of the solid particles inside the vortices, a nonlinear mathematical simulation could be made of the particle distribution within an observed vortex during the diffusive process of an aspirin matrix (Figure 44.5). The reported simulation is the result of

(b)

(c) Position in “y” axis (pixel)

(a)

f(x,y) = [(1.1 + 0.91 (xi cos Q + yi sen Q)), (0.91 (–xi sen Q + yi cos Q))]

30 2 1.5 1 0.5 0

–0.5 –1

–1.5 –1 –0.5 0 0.5 1 1.5 2 Position in “x” axis (pixel)

(e) Position in “y” axis (pixel)

Position in “y” axis (pixel)

(d)

Q = (–5.6) / (1 + xi2 + yi2)

40 30 20 10 0 –10 –20 –20 –10 0 10 20 30 40 50 60 Position in “x” axis (pixel)

30 20 10 0 –10 –20 –30 –20 –10 0 10 20 30 40 50 60 Position in “x” axis (pixel)

Figure 44.5. Proposed mathematical simulation for one of the vortices during diffusion of an aspirin matrix. (a) The image at 10 s. (b) Amplification of the modeled vortex. (c) A digitalized vortex. (d) Graph for the proposed iterative (1000 times) model. (e) A 20° clockwise-turned digitalized vortex.

520

PART 2: Poster Presentations

iterating the written equation 1000 times. The correspondence of the simulation to the digitalized vortex is 93% as calculated with the image-analysis software. The proposed mathematical model could be considered a fractal model because (a) it is derived from an iterative process, and (b) it shows self-similarity and a high level of detail at different observation scales (Mandelbrot 1977). It has a fractional dimension on the order of df = 1.55. Such a df indicates that this mathematical model consists of a series of trajectories or lines that trend asymptotically to a fully flat surface, so the proposed mathematical construction has a dimension between 1 and 2 (a Euclidean line has a dimension equal to 1, and a Euclidean surface has a dimension equal to 2) (Mandelbrot 1977; Peleg 1993). Therefore, the proposed mathematical model could be cataloged as a strange attractor whose intersection with the plane z = 0 (section of Poincaré) gives the distribution of points presented in the graphic of Figure 44.5d (Alligood and others 2000; Nagle and others 2001). If the observed particle distribution can be modeled by means of a strange attractor, this suggests that a nonlinear mechanism could be operating on the system. Thus, solid matrix diffusion might be studied in the frame of deterministic chaos (Redondo and others 2004).

Conclusions The edge of both types of matrix shows a fractal nature, and the df values have statistically identical kinetics after the first 20 s. The measured df corresponds to the lowest layer of the matrix; this layer is affected by the superior ones and shows a rapid increase in irregularity that is followed by compensation and persistence in the df after the first 16 s. Three layers were quantified, and these collectively provide a value ranging to 1.2 for the df of the side face of the matrices. Aspirin and maltodextrin solid matrices, in the experimental conditions, show a disintegration process comparable to the opening of the petals of a flower, with fractures during such opening. A mathematical model was developed, applying nonlinear dynamics, for one of the turbulences observed during the diffusion process of an aspirin matrix. A strange attractor with a df of 1.55 was obtained, whose structure approximates 93% of the particle distribution inside the vortex. Apparently, a molecular diffusion mechanism is not the only, nor the main, mechanism operating in solid-matrix diffusion processes under the experimental conditions. Probably, deterministic chaos is also present in the diffusion of solid matrices.

Acknowledgments The authors thank the Mexican National Council of Science and Technology (CONACYT), Committee on Operation and Development of Academic Activites (COFAA), and National Polytechnic Institute (IPN) for economic support by means of projects SIP-IPN 20070631 and 20071011, as well as CONACYT 48061-Z and 59730.

Image Processing and Nonlinear Dynamics

521

References Abdou HM, Hanna S, Muhammad N. 2001. Disolución. In: Gennaro AR, editor. Remington Farmacia. 20th ed. Buenos Aires: Editorial Médica Panamericana. p 764–9. Alligood KT, Sauer TD, Yorke JA. 2000. Chaos: an introduction to dynamical systems. 1st ed. New York: Springer-Verlag. p 105–267, 359–97. Banker GS, Anderson NR. 1986. Tablets. In: Lachman L, Lieberman HA, Kanig JL, editors. The theory and practice of industrial pharmacy. 3rd ed. Philadelphia: Lea & Febiger. p 293–343. Chanona PJJ, Alamilla BL, Farrera RRR, Quevedo R, Aguilera JM, Gutiérrez LGF. 2003. Description of the convective air-drying of a food model by means of the fractal theory. Food Sci Technol Int 7:207–13. Fischer HB, List EJ, Koh RCY, Imberger J, Brooks NH. 1979. Mixing in inland and coastal waters. San Diego, CA: Academic. p 116–21, 145–72. Mandelbrot BB. 1977. The fractal geometry of nature. 1st ed. New York: WH Freeman. p 5–19, 25–46. McCabe WL, Smith JC, Harriott P. 2002. Operaciones unitarias en ingeniería química. 6th ed. Mexico City: McGraw-Hill. p 545–83. Nagle RK, Saff EB, Snider AD. 2001. Ecuaciones diferenciales y problemas con valores en la frontera. 3rd ed. Mexico City: Pearson Education. p 261–323. Parrot EL. 1970. Pharmaceutical technology. 1st ed. Minneapolis, MN: Alpha. p 73–92, 158–69. Peleg M. 1993. Fractals and foods. Crit Rev Food Sci Nutr 33:149–65. Pratt W. 1986. The entry, distribution, and elimination of drugs. In: Pratt WB, Taylor P, editors. Principles of drug action: the basis of pharmacology. 3rd ed. New York: Wiley & Sons. p 222–7. Redondo JM, Platonov AK, Grau J. 2004. Application of multifractal feature analysis to the sea surface. In: Chashechkin YuD, Baydulov VK, editors. Proceedings of International Conference on Fluxes and Structures in Fluids, 2003. Moscow: Russian Academy of Sciences (RAS). p 66–8. Treybal RE. 1988. Operaciones de transferencia de masa. 2nd ed. Mexico City: McGraw-Hill. p 21–31, 60–75, 100–9.

Session 6 Properties and Stability of Food and Biological Systems

45 Effect of Water Content on Physical Properties of Potato Chips F. Pedreschi and P. Moyano

Abstract The objective of this research was to study the effect of violent drying during immersion frying on the texture, color, oil content and distribution, and porosity of potato chips. Some potato slices were blanched in hot water at 85°C for 3.5 min and were deep fried in sunflower oil at 120°, 150°, and 180°C. A model based on a variable diffusion coefficient during the frying process was used to model water loss. The effective moisture diffusion coefficient increased with frying time and temperature. Oil uptake was high even for short frying times (high moisture content), suggesting that oil wetting is an important mechanism of oil uptake during frying. The normalized maximum force (MF*) parameter was used in modeling textural changes in the potato slices during frying in both stages: (a) the initial tissue-softening stage and (b) the later crust-development process. Higher temperatures accelerated both processes; however, neither the temperature nor the pretreatment had a significant effect (P > 0.05) on the final texture of the fried potato chips. The potato chip color-difference parameter (ΔE) tends to increase sharply as the moisture loss increases with frying time. Porosity increased abruptly as result of violent drying during frying of blanched potato slices at 180°C.

Introduction Frying in hot oil at temperatures between 160° and 180°C is characterized by very high drying velocities, which are critical to improve not only the mechanical, but also the structural properties of the potato chips (Baumann and Escher 1995). Potato chips are thin slices whose moisture content decreases from around 80% to almost 2% when they are fried. However, the drying in oil inevitably leads to a considerable oil uptake of ∼35%, which is mostly located in chip surface (there is almost no penetration during frying), and it adheres to the surface at the end of the frying. Therefore, a high proportion of oil penetrates the food microstructure during the postfrying cooling stage (Ufheil and Escher 1996; Aguilera and Gloria-Hernández 2000; Bouchon and others 2003). Most recently, consumers’ preference for low-fat and fat-free products has been the driving force of this food industry to produce lower oil–content fried potatoes that still retain their desirable texture and flavor. Bouchon and others (2003) defined three different oil fractions that can be identified as a consequence of the different absorption mechanisms in fried potato cylinders: (a) structural oil (StO), which 525

526

PART 2: Poster Presentations

represents the oil absorbed during frying; (b) penetrated surface oil (PSO), which represents the oil suctioned into the food during cooling after removal from the fryer; and (c) surface oil (SO), which is the oil that remains on the surface and does not penetrate the microstructure. These authors showed that only a small amount of oil penetrates potato cylinders during frying because most of the oil is picked up at the end of the process, suggesting that oil uptake and water removal are not synchronous. Crispness is an important parameter to be controlled during processing of potato chips (Scanlon and others 1994). The crispy structure of potato chips is the result of changes at the cellular and subcellular levels in the outermost layers of the product. These chemical and physical changes include physical damage produced when the product is cut and a rough surface is formed with release of intracellular material; starch gelatinization and consequent dehydration; protein denaturation; breakdown of adhesive forces between cells; water evaporation; rapid dehydration, expansion, and browning of the tissue; and finally oil uptake itself (Bouchon and others 2001). Among the different classes of physical properties of foods and foodstuffs, color is considered the most important visual attribute in the perception of product quality. Fried potato color is the result of the Maillard reaction, which depends on the content of reducing sugars and amino acids or proteins at the surface, and the temperature and time of frying (Márquez and Añón 1986). Porosity is a very important structural parameter that changes during frying. This research studied the effect of violent drying during frying on the texture, color, oil content and distribution, and porosity of potato chips.

Materials and Methods Materials Potatoes (variety Panda) and sunflower oil (Chef; Coprona, Santiago, Chile) were the raw materials. Slices (2.5 mm thick) were cut, and a circular cutting mold was used to provide chips with a diameter of 37 mm. Slices were rinsed immediately after cutting for 1 min in distilled water to eliminate some starch material adhering to the surface prior to frying (control). Slices immersed in hot water at 85°C for 3.5 min were called blanched slices. Frying Conditions Ten slices per sample time of each pretreatment were deep fried in 8 L of hot oil contained in an electrical fryer (Beckers model F1-C, Treviglio, Italy) at each of the three temperatures tested: 120°, 150°, and 180°C. Slices were fried at different time intervals until reaching a final moisture content of ∼1.8% (wet basis). Analysis The mean moisture content of potato chips was measured by drying the samples in a convection oven until constant weight at 105°C. The oil content was determined by Bligh and Dyer methodology (1959). Fractions of absorbed oil were determined

Effect of Water Content on Physical Properties of Potato Chips

527

according the method of Bouchon and others (2003). Color was determined by a computer vision system according to the method of Pedreschi and others (2006). Porosity was determined according to the method of Moreira and others (1999). Modeling Water Loss The model used to describe water loss during frying considers that physical properties of the potato slices vary with the moisture content during the process, making the effective moisture diffusion coefficient a function of time (Alvarez and Legues 1986). In the case of potato chips, Equation 45.1 was adequately used for slice geometry to model water loss in an axial direction during frying: mt − me 8 = m0 − me π 2

⎧ 1 ∑ ⎨ (2n + 1)2 n =1 ⎩ ∞

⎡ ⎛ ( 2 n + 1) π 2 ⎢exp ⎜⎝ − 4 (1 + b ) ⎣ 2

⎡⎛ D0 t ⎞ ⎢⎝ 1 + l 2 ⎠ ⎣

1+ b

⎤⎞ ⎤ ⎫ ⎥⎟⎠ ⎥ ⎬ ⎦ ⎦⎭

(45.1)

where l is the half-thickness of the slice; mt is the moisture content at time t (dry basis); m0 is the initial moisture content (dry basis); me is the equilibrium moisture content (dry basis); t is the frying time, D0 is the effective moisture diffusion coefficient at time 0, and b is a dimensionless parameter. It is reasonable to assume that the moisture content is negligible when equilibrium is reached in the frying process, so me = 0 in Equation 45.1. Effective diffusivity (Deff) is postulated in Equation 45.2 as a function of time as D Deff = D0 ⎛ 1 + 20 t ⎞ ⎝ l ⎠

b

(45.2)

Results and Discussion Figure 45.1 shows that water loss increases with the frying temperature and time for control potato chips as expected. Figure 45.2 shows that the Deff increases with frying time and temperature for control slices. For the time interval studied at the three frying temperatures, effective diffusivity (Deff) always increases, suggesting that the water leaves potato tissue easily without a pronounced resistance of the external crust progressively formed (which could eventually make Deff diminish with time). Figure 45.3 shows kinetics of total oil (TO) uptake and its different fractions (PSO, StO, and SO) in the structure of control potato chips fried at 180°C. At very short frying times (between 1 and 4 min) or high moisture content, almost 75% of the TO content of final control potato chips (chips with ∼1.8% of moisture content, total basis) is absorbed. After that time interval, the TO content of potato chip is reached and remains almost constant until 1.8% of moisture content (total basis) is reached. As the frying temperature decreases, the frying time required to reach that final moisture content and the TO increases. Similar results have been found in pretreated potato chips, fried potato cylinders, and tortilla chips (Moreira and others 1997; Bouchon and others 2003; Durán and others 2007). This result suggests that the TO

Figure 45.1. Observed and predicted water loss during the frying of control slices at various temperatures. Solid lines represent the values predicted by the variable diffusivity model (Equation 45.1).

Figure 45.2. Effective moisture diffusivity (Deff) during frying of control potato slices at 120°, 150°, and 180°C (Equation 45.2).

Oil content (g/g dry basis)

0.50 SO STO PSO TO

0.40 0.30 0.20 0.10

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

Moisture content (g/g dry basis)

Figure 45.3. Kinetics of oil-uptake fractions and total oil (TO) in control potato slices during frying at 180°C. PSO, penetrated surface oil; SO, surface oil; and STO, structural total oil.

528

Effect of Water Content on Physical Properties of Potato Chips

529

6.0

MF (N)

4.5

3.0

180 °C

1.5

150 °C 120 °C

0.0 0

0.8

1.6 2.4 3.2 Moisture content (g water/g dry solid)

4

Figure 45.4. Maximum force (MF) vs moisture content for control potato slices fried at 120°, 150°, and 180°C. N, Newton.

in potato chips is absorbed primarily in the initial stage of frying once the potato slices are placed inside the hot oil. Most of the oil penetrates the potato microstructure after the potato chip is removed from the fryer, during the cooling period. Almost no oil penetrates the potato tissue while the potato slices are within the fryer oil. Maximum force (MF) for control samples increases as the moisture content of the slices decreases, which corresponds to an increase in the crispness of the chips (Figure 45.4). There is no apparent effect of the frying temperature on MF when slices with the same moisture content are compared. The frying temperature did not have a significant effect (P > 0.05) on the final texture represented by the MF of the fried potato slices (moisture content of ∼1.8%, wet basis). Total color change was calculated by the total color-difference parameter ΔE = ([L*0 − L*]2 + [a*0 − a*]2 + [b*0 − b*]2 )1 2. Color changes in the potato slices during frying were followed by ΔE, since this color parameter showed a notorious increase during frying as the potato slices dried (Figure 45.5). Potato slices tend to darken during frying (as a result of the surface nonenzymatic browning reaction), as indicated by the progressively increasing ΔE values with frying time. Finally, Figure 45.6 shows the effect of frying time and temperature on the porosity changes in blanched potato slices during frying. At longer frying times and lower moisture content, the porosity increased considerably.

Conclusions The model with a variable coefficient of diffusion was suitable for modeling moisture loss during the frying of potato slices at three temperatures. At higher oil temperatures, moisture was lost faster. Most of the oil in a potato chip is formed by SO that penetrates the structure during cooling. Almost no oil penetrates the potato chip

530

PART 2: Poster Presentations

60 50

ΔE

40 30 20 10 0 0

1

2 3 4 Moisture content (g/g dry solid)

5

6

Figure 45.5. Color evolutions of blanched potato slices fried at 180°C. ΔE, colordifference parameter.

0.60

Porosity

0.45

0.30

0.15

0.00 0

1

2

3

4

5

Moisture content (g/g dry solid)

Figure 45.6. Changes in porosity in blanched potato slices fried at 180°C.

microstructure during frying. As the potato slices dry violently during frying, their porosity increases and their color darkens. At low moisture content, potato chips fried at lower temperature seemed to be crispier.

Acknowledgment The authors acknowledge financial support from Fondo Nacional de Desarrollo Científico y Tecnológico (FONDECYT) (National Fund for Scientific and Technological Development) projects 1070031 and 1030411.

References Aguilera JM, Gloria-Hernández H. 2000. Oil absorption during frying of frozen parfried potatoes. J Food Sci 65:446–79. Alvarez I, Legues PA. 1986. Semi-theoretical model for the drying of Thompson seedless grapes. Drying Technol 4:1–17.

Effect of Water Content on Physical Properties of Potato Chips

531

Baumann B, Escher E. 1995. Mass and heat transfer during deep fat frying of potato slices: rate of drying and oil uptake. LWT Food Sci Technol 28:395–403. Bligh EG, Dyer WA. 1959. A rapid method of total lipid extraction and purification. Can J Biochem Physiol 37:911–7. Bouchon P, Aguilera JM, Pyle DL. 2003. Structure oil-absorption relationships during deep-fat frying. J Food Sci 68:2711–6. Bouchon P, Hollins P, Pearson M, Pyle DL, Tobin MJ. 2001. Oil distribution in fried potatoes monitored by infrared microspectroscopy. J Food Sci 66:918–23. Durán M, Pedreschi F, Moyano P, Troncoso E. 2007. Oil partition in pre-treated potato slices during frying and cooling. J Food Eng 81:257–65. Márquez G, Añón MC. 1986. Influence of reducing sugars and amino acids in the color development of fried potatoes. J Food Sci 51:157–60. Moreira RG, Castell-Perez ME, Barrufet MA. 1999. Deep-fat frying: fundamentals and applications. Gaithersburg, MD: Aspen. Moreira RG, Sun X, Chen Y. 1997. Factors affecting oil uptake in tortilla chips in deep-fat frying. J Food Eng 31:485–98. Pedreschi F, León J, Mery D, Moyano P. 2006. Implementation of a computer vision system to measure the color of potato chips. Food Res Int 39:1092–8. Scanlon MG, Roller R, Mazza G, Pritchard MK. 1994. Computerized video image analysis to quantify colour of potato chips. Am Potato J 71:717–33. Ufheil G, Escher F. 1996. Dynamics of oil uptake during deep-fat frying of potato slices. LWT Food Sci Technol 29:640–4.

46 Predicting Water Migration in Starchy Food During Cooking S. Thammathongchat, M. Fukuoka, T. Hagiwara, T. Sakiyama, and H. Watanabe

Abstract A starchy food that is initially a single-phase body turns into a multiphase body during cooking because of starch gelatinization. The relative water content (RWC) model has been proposed to describe water migration in multiphase food systems. In the RWC model, water migration is driven by the gradient of water content (m) divided by the water-holding capacity (WHC, m*), m/m*. In this work, a WHC profile as a function of water content at boiling temperature was assumed based on information concerning starch gelatinization. Several WHC profiles in different shapes were applied to the RWC model by which the change in water-content profile in a slab of wheat-flour dough during boiling was simulated. Some correlation between the shape of the WHC profile and the shape of the calculated transient water-content profile was revealed. Assisted by this correlation, a WHC profile was proposed and found to be applicable in conjunction with the RWC model in describing some distinguishing features (e.g., bending points) of the reported transient water-content profile in the slab of wheatflour dough during boiling.

Introduction Water distribution in food is one of the main factors that affects the texture and shelf life of food. Texture of cooked rice, pasta, and noodles depends greatly on watercontent distribution inside the food body after boiling (Gonzalez and others 2000; Kojima and others 2000; Irie and others 2004). Therefore, an understanding of water migration inside these starchy foods during boiling is necessary. With magnetic resonance imaging, water distribution inside the food body can be monitored. It was found that the water-content profile inside the starchy food that had been soaked in water at a temperature higher than the gelatinization temperature (boiling) showed an irregular water-content profile that could not be explained by Fick’s second law of diffusion (Stapley and others 1997; Takeuchi and others 1997; Fukuoka and others 2000). This irregular water-content profile in the food body may be attributed to a heterogeneous structure that is caused by starch gelatinization. Since Fick’s law is applicable to homogeneous systems only, it clearly cannot be applied in this case. In this work, the relative water content (RWC) model was applied to explain water migration inside the model starchy food (a slab of wheat-flour dough) during boiling. In the RWC model (Equation 46.1), water migration is driven by the 533

534

PART 2: Poster Presentations

gradient of water content (m) divided by water-holding capacity (WHC, m*) (Watanabe and others 2006, 2007). j = −ρsolid Dm ∗

∂ ⎛ m⎞ ⎜ ⎟ ∂ x ⎝ m* ⎠

(46.1)

When the diffusion flux equation of the RWC model is combined with the equation of continuity, we have a one-dimensional diffusion equation.

{

∂m ∂ ∂ m⎞ (ρsolid m ) = ρsolid Dm ∗ ⎛ ∂t ∂x ∂x ⎝ m* ⎠

}

(46.2)

In a system where expansion or shrinkage is negligibly small, we have a simple form of diffusion equation. ∂m ∂ ⎛ ∂ m ⎞⎞ = Dm ∗ ⎛ ∂t ∂x ⎝ ∂x ⎝ m* ⎠ ⎠

(46.3)

The water-holding capacity, defined as the maximum water content that a food body can adsorb at equilibrium, plays an important role in the RWC model. In this work, the effect of the WHC profile on the transient water-content profile was studied.

Method Modeling In this work, the RWC model was applied to calculate water migration in a slab of wheat-flour dough during boiling at 100°C. The size (3.8-mm thickness) and the initial water content of the dough slab (0.72 kg water/kg solid) selected was the same as those used in the experiment reported by Fukuoka and others (2000). The watercontent profile calculated was compared with the profile measured by those authors. Since the water-diffusion process is several tens of times slower than the heat-transfer process, the water-content profile inside the slab may be governed by the waterdiffusion process at constant temperature (100°C). For the sake of simplicity, swelling of the dough slab was disregarded, so Equation 46.3 was solved by an explicit method. A constant diffusivity value (5 × 10−10 m2/s) was used in this work so that the effect of WHC on the water-content profile simulation was emphasized. The water-content values, obtained by using magnetic resonance imaging, at a position ∼0.1 mm below the slab surface was used as the boundary condition in the calculation. The water content at this position was found to increase linearly as the time increased (Equation 46.4). m = 0.000555t + 0.7333

m ≤ 1.7

(46.4)

Predicting Water Migration in Starchy Food During Cooking

535

Water-Holding Capacity Profile The WHC of starchy food was reported to depend on the extent of starch gelatinization. For example, the WHC of fully gelatinized rice grain was several times larger than that of the nongelatinized grain (Watanabe and others 2001). Starch granules were gelatinized rapidly to reach a specified extent of gelatinization within 1 or 2 min (Lund and Wirankartakusumah 1984; Gomi and others 1998). This upper limit in the extent of gelatinization is termed the terminal extent of gelatinization (TEG), which depends on both temperature and water content (Fukuoka and others 2002). The TEG increases from 0 (nongelatinized) to 1 (fully gelatinized) as water content increases from 0.4 to 1.7 kg water/kg solid at 100°C. Therefore, the WHC should increase from that of nongelatinized dough to that of fully gelatinized dough when water content increases from 0.4 to 1.7 kg water/kg solid. The WHC of nongelatinized dough measured by soaking dough in pure water is 0.95 kg water/kg solid (Yahata and others 2006). The maximum WHC value of the model dough selected was 1.7 kg water/kg solid, which is the maximum water content in the water-content profile of the dough slab measured during boiling (Fukuoka and others 2000). In this work, three schematics of WHC profiles—linear (A), convex (B), and concave (C)—were used to calculate a transient water-content profile in the model food (a slab of wheat-flour dough) during boiling (Figure 46.1).

Results and Discussion The transient water-content profiles calculated by using the WHC profile in Figure 46.1 is shown in Figure 46.2. The shape of water-content profile that was calculated

Figure 46.1. Water-holding capacity (WHC) profiles used for the calculation of the transient water-content profile in a slab of wheat-flour dough during boiling: line A, a linearly varying WHC profile; line B, a WHC profile with a breaking point at m = 1.25 (db), WHC = 1.6 (db); and line C, a WHC profile with a breaking point at m = 1.25 (db), WHC = 1.3 (db). db, dry basis (kg water/kg solid).

536

PART 2: Poster Presentations

Figure 46.2. A water-content profile in a slab of wheat-flour dough calculated using water-holding capacity (WHC) profiles in Figure 46.1. (a) A linear WHC profile (line A in Figure 46.1). (b) A bending WHC profile (m = 1.25 (db), WHC = 1.6 (db); line B in Figure 46.1). (c) A bending WHC profile (m = 1.25 (db), WHC = 1.3 (db); line C in Figure 46.1). The numbers indicate boiling time (min). db, dry basis (kg water/kg solid).

by using line A as the WHC profile was apparently similar to that of water-content profiles calculated by the Fickian diffusion equation (Figure 46.2a). On the other hand, the convex (line B) or concave (line C) shape of the WHC profile was reflected in the shape of water-content profile calculated. Moreover, the water content at which the break points occurred in each water-content profile (Figure 46.2b and c) coincided

Predicting Water Migration in Starchy Food During Cooking

537

Figure 46.3. (a) A water-holding capacity (WHC) profile with three bending points. (b) A water-content profile in a slab of wheat-flour dough calculated using the WHC profile given in part a is compared with the profile measured. The numbers indicate boiling time (min). db, dry basis (kg water/kg solid); solid lines, calculated profile; and broken lines, measured profile.

with the break-point water content in the WHC profile (1.25 kg water/kg solid) (Figure 46.1, lines B and C). The water-content profile inside a slab of wheat-flour dough measured by Fukuoka and others (2000) is shown by the broken lines in Figure 46.3b. The characteristic features of the water-content profile measured can be summarized as follows: 1. Water migration in the wheat-flour dough was very slow. Even after 30 min of boiling, the water content at the center of the slab remained at the initial level while water content at the surface reached an equilibrium value of 1.7 kg water/kg solid. The water content inside the slab leveled off to an equilibrium of 1.7 kg water/kg solid when the slab was boiled for 120 min. 2. At 60 min of boiling, a nearly flat water-content profile emerged in three regions: at the center, near the surface, and in the intermediate part of these two regions. These flat regions were tied with a sharp gradient curve in the water-content profile. To mimic the water-content profile measured, a WHC profile with three bending points was proposed. Referring to the relationship between the bending point in the WHC profile and that in the transient water-content profile, three levels of water content (0.85, 1.1, and 1.6 kg water/kg solid) were selected as the water content of the bending points in the proposed WHC profile (Figure 46.3a). The transient watercontent profiles calculated are presented in Figure 46.3b, which shows that the outline of the transient water-content profile in the dough slab during boiling is well described by using the RWC model with a water diffusivity of 5 × 10−10 m2/s and the WHC profile proposed. The bending points that emerged at the three levels of water content (m = 0.85, 1.1, and 1.6 kg water/kg solid) in the water-content profile measured at 30 min and 60 min of boiling were depicted successfully.

538

PART 2: Poster Presentations

Conclusion The effect of the WHC profile (WHC plotted against water content with which starchy food is heat treated) was examined to describe the transient water-content profile in a wheat-flour dough slab during boiling. A WHC profile with break points was found useful for describing certain characteristic features of transient water content.

Nomenclature D j M m* t x ρsolid

diffusivity (m2/s) water-diffusion flux (kg water [m2/s]−1) water content (kg water/kg solid) water-holding capacity (kg water/kg solid) time (s) position (m) solid density (kg solid [m3]−1)

References Fukuoka M, Mihori H, Watanabe T. 2000. MRI observation and mathematical model simulation of water migration in wheat flour dough during boiling. J Food Sci 65:1343–8. Fukuoka M, Ohta K, Watanabe H. 2002. Determination of the terminal extent of starch gelatinization in a limited water system by DSC. J Food Eng 53:39–42. Gomi Y, Fukuoka M, Mihori T, Watanabe H. 1998. The rate of starch gelatinization as observed by PFGNMR measurement of water diffusivity in rice starch/water mixtures. J Food Eng 36:359–69. Gonzalez JJ, McCarthy KL, McCarthy MJ. 2000. Textural and structural changes in lasagna after cooking. J Texture Stud 31:93–108. Irie K, Horigane AK, Naito S, Motoi H, Yoshida M. 2004. Moisture distribution and texture of various types of cooked spaghetti. Cereal Chem 81:350–5. Kojima T, Sekine M, Suzuki T, Horigane A, Nagata T. 2000. Effect of moisture distribution on texture of boiled Japanese noodles (udon). Nippon Shokuhin Kagaku Kogaku Kaishi 47:142–7. Lund DB, Wirankartakusumah M. 1984. A model for starch gelatinization phenomena. In: McKenna BM, editor. Engineering and food, vol. 1. New York: Elsevier. p 425–32. Stapley AGF, Hyde TM, Gladden LF, Fryer PJ. 1997. NMR imaging of the wheat grain cooking process. Int J Food Sci Technol 32:355–75. Takeuchi S, Fukuoka M, Gomi Y, Maeda M, Watanabe H. 1997. An application of magnetic resonance imaging to the real time measurement of the change of moisture profile in a rice grain during boiling. J Food Eng 33:181–92. Watanabe H, Fukuoka M, Tomiya A, Mihori T. 2001. A new non-Fickian diffusion model for water migration in starchy food. J Food Eng 49:1–6. Watanabe H, Yahata Y, Fukuoka M, Sakiyama S, Mihori T. 2006. Relative water content model that describes non-Fickian diffusion in multiphase inhomogeneous systems [in Japanese]. Jpn J Food Eng 7:129–39. Watanabe H, Yahata Y, Fukuoka M, Sakiyama S, Mihori T. 2007. The thermodynamic basis for the relative water demand model that describes non-Fickian water diffusion in starchy foods. J Food Eng 83:130–5. Yahata Y, Fukuoka M, Sakiyama T, Watanabe H. 2006. Water sorption in pre–heated-treated wheat flour dough. Jpn J Food Eng 7:163–72.

47 Nonenzymatic Browning May Be Inhibited or Accelerated by Magnesium Chloride According to the Level of Water Availability and SaccharideSpecific Interactions P. R. Santagapita, S. B. Matiacevich, and M. P. Buera

Abstract The effect of magnesium chloride (MgCl2) on the kinetics of nonenzymatic browning (NEB) was evaluated and related to water and/or sugar interactions with the salt. Liquid systems incubated at 70°C consisted of 50%–70% wt/vol saccharide (glucose or trehalose) and 0.5% wt/vol L-glycine in phosphate buffer pH 5 with and without 0.14 M MgCl2. The progress of NEB was followed by absorbance at 445 nm, by fluorescence (excitation/emission, 340/492 nm), and by pH changes. Proton nuclear magnetic resonance (1H NMR) transversal relaxation time (T2) was measured by CarrPurcell-Meiboom-Gill (CPMG) pulse sequence. The presence of MgCl2 could greatly inhibit NEB in glucose systems. The reaction was accelerated in trehalose systems, but the type of fluorophore obtained in both sugars was not affected by the salt. Whereas the T2 values were slightly affected in systems containing glucose (the waterglucose interaction was unmodified by salt), in trehalose systems T2 values were 6%–14% lower with the salt. These lower values reflect that the salt imposed a certain local order to water molecules, reducing their mobility. Consequently, the relative effect of MgCl2 on NEB rate depends on the rate of disaccharide hydrolysis and on the inhibitory effect of water, which over time depends on the interactions of the cation with water and the complexation of the cation with sugar. MgCl2 effects must be taken into account, considering the high technological interest in finding strategies either to inhibit or to enhance NEB, depending on the application.

Introduction The Maillard reaction involves a complex set of steps, and its interpretation represents a challenge in basic and applied aspects of food science. During storage of foods and biological systems, amino acids or proteins react with reducing sugars, and brown pigments and fluorescent products are produced through the Maillard reaction. The rate of the Maillard reaction and the nature of its products are governed by the immediate chemical environment of the reactants as defined by the chemical composition of the system: water content, pH, presence and type of buffer salts, temperature, and exposure to light (Baisier and Labuza 1992). The Maillard reaction is one of the causes of nonenzymatic browning (NEB), which affects the quality of food and pharmaceutical products. 539

540

PART 2: Poster Presentations

The addition of salts may influence the kinetics of NEB and hydrolysis reactions (Mazzobre and Buera 1999; Matiacevich and Buera 2006). Both reactions lead to food-quality loss. The presence of magnesium chloride (MgCl2) in freeze-dried trehalose or sucrose systems (22% and 43% relative humidity) increases water retention and delays sugar crystallization. Whereas an inhibition of the NEB reaction was observed with the presence of MgCl2 in freeze-dried trehalose systems, the effect was the opposite in sucrose systems (Santagapita and Buera 2006). The presence of MgCl2 in glucose-glycine solutions delayed the formation of fluorescent compounds and brown pigments without affecting their spectral characteristics (Matiacevich and Buera 2006). In addition, sugar-metal complexes can be formed in solution (Carpenter and others 1987). Thus, the use of MgCl2 may help to avoid the addition of sulfites for NEB inhibition. Sulfites are used in food products to control microbial growth, browning, or bleaching. However, sulfite sensitivity, which occurs most often in asthmatic adults, can be problematic (Lester 1995). Therefore, the search for innocuous compounds (such as alkali metal chloride) is of high technological interest. Pulsed nuclear magnetic resonance (NMR) is a technique for investigating the mobility of water in foods (Schmidt 1990). The spin-spin or transverse relaxation time constant of protons (T2) measures the molecular mobility of protons in the system. This relaxation process is affected by the chemical exchange process taking place between different sites of different mobilities. Protons of water molecules that interact closely with solids and are highly immobilized show reduced T2, whereas protons that are readily mobile have a relatively long T2 (Kuo and others 2001). Thus, the present work analyzed the effect of MgCl2 on NEB as related to water interactions, measured by NMR.

Materials and Methods Preparation of Model Systems Liquid systems consisted of 50%–70% wt/vol saccharide (trehalose [Havashibara, Leatherhead, UK] or glucose [Merck, Rahway, NY, USA]) and 0.5% wt/vol L-glycine (Gly; Merck) with 0.14 M MgCl2 · 6H2O (Mallinckrodt, St. Louis, MO, USA). The systems were prepared in phosphate buffer 0.1 M, pH 5 (Merck); controls without MgCl2 and without glycine were also prepared. All reactants were analytic grade. Aliquots of 2.5 mL of each model were placed in 5-mL vials. The samples were hermetically sealed in the vials and stored at 70° ± 1°C in a forced-air convection oven to accelerate the NEB reaction. Since trehalose is a non-reducing sugar, its participation in NEB reaction occurred after hydrolysis. Methods The progress of NEB was followed by a browning index (BI) defined as absorbance readings (UV-Vis 1620; Shimadzu, Columbia, MD, USA) at 445 nm multiplied by the dilution factor and by fluorescence excitation at 340 nm/emission at 492 nm (spectrofluorometer; Ocean Optics, Dunedin, FL, USA). The samples used for the

Nonenzymatic Browning May Be Inhibited or Accelerated by Magnesium Chloride

541

determination of fluorescence were diluted in order to have systems with absorbance values lower than 0.1 at excitation wavelength (340 nm) to avoid inner filter effects. Changes in pH during NEB reactions were measured with a pH meter (MettlerToledo, Columbus, OH, USA). Sugar hydrolysis was analyzed by measuring the amount of glucose released during the incubation time by means of an enzymatic method. Time-resolved proton nuclear magnetic resonance (1H NMR) was used to determine water mobility at 25°C. After standard shaking, samples of exactly 3 mL of each sugar system at adequate incubation times at 70°C were used to measure the T2 in a Minispec mq20 (20 MHz; Bruker AXS, Madison, WI, USA) spectrometer using the spin echo Carr-Purcell-Meiboom-Gill (CPMG) method (Meiboom and Gill 1958). Data were averaged over four acquisitions of 256 points, with phase cycling, using a gain of 68 and τ of 0.5. Sufficient echoes were recorded to give a zero baseline, and the T2s were determined for the echo-decay envelope by using an exponential unpeeling program.

Results and Discussion The rate of fluorescence and brown pigment development during storage at 70°C was dependent on the composition of the liquid systems. Figure 47.1 shows the browning index (BI) obtained for the systems at 50% wt/vol concentration of solids, but the same behavior was obtained for the 60% and 70% wt/vol systems. Since trehalose is a non-reducing sugar and its participation in NEB reaction occurred only after hydrolysis, trehalose systems had developed much less browning than did glucose systems (Figure 47.1). It is also well known that trehalose is very stable in hydrolysis (Schebor and others 1999). However, trehalose hydrolysis occurred during heat treatment at 70°C, and was two times faster in the presence of MgCl2, thus accelerating the NEB reaction. Moreover, the MgCl2 displayed a great ability to inhibit NEB in glucose systems (50%–70% wt/vol), but in trehalose systems (50%–70% wt/vol) the reaction was accelerated, as shown in Figure 47.1. It is interesting to note that in both systems the same reactant (glucose) participates in the browning reaction, and it can be hypothesized that the interactions of trehalose with MgCl2 and/or with water may play a role in the browning inhibition observed. Fluorescence intensity was negligible at incubation time zero, it increased with increasing incubation time, and paralleled browning data in all systems. The fluorescence characteristics were studied to analyze the effects associated with sugar type and their interactions with MgCl2. Fluorescence emission spectra were similar in both sugar samples with and without the salt, as shown in Figure 47.2. This result indicates that MgCl2 does not affect the type of fluorophore obtained. A Raman peak, produced by water dispersion close to 420 nm, was also observed and did not interfere with the measurements of the fluorescent products. Moreover, samples without amino acids did not develop fluorescence. Thus, caramelization involving the sugar reactions in the systems did not contribute to fluorescence nor to browning development. Therefore, according to previous studies (Matiacevich and Buera 2006), the composition of the

542

PART 2: Poster Presentations

2.8 G

2.4

BI

2.0

G+MgCl2

1.6 1.2

T+MgCl2

0.8 0.4 0.0

T 0

25 50 75 100 125 150 175 200 225 250 275 300 Time (h)

Figure 47.1. Browning index (BI) of 50% wt/vol systems during storage at 70°C. BI, absorbance at 445 nm × dilution factor; G, glucose; MgCl2, magnesium chloride; and T, trehalose. 90

50 T 50 G 50 TM 50 GM

Fluorescence intensity (au)

80 70 60 50 40 30 20 10 0 400

t=0h G or T 425

450

475

500

525

550

575

600

Wavelength (nm)

Figure 47.2. Fluorescence emission spectra of 50% wt/vol sugar systems at 70°C for 168 h/240 h (glucose and trehalose systems, respectively). Incubation time zero for both sugar systems (t = 0 h, G or T). au, arbitrary units; G, glucose; M, magnesium chloride; and T, trehalose.

systems affected the kinetics of the reaction without affecting their chromatic or fluorescence characteristics. Despite the 0.1 M buffer used, the development of the NEB reaction promoted a reduction in pH values in both the trehalose and the glucose systems. The decreased NEB reaction rate could have been a consequence of a pH drop. However, for both sugar systems, the change in pH as a function of time was greater in the absence of salt. Thus, the attenuation of pH change while NEB progressed caused by the presence

Nonenzymatic Browning May Be Inhibited or Accelerated by Magnesium Chloride

543

Table 47.1. Nuclear magnetic resonance spin-spin transverse relaxation data (T2) and their analysis as a function of the presence of magnesium chloride (MgCl2) and/or brown pigment T2L valuesa (ms)

f (MgCl2)b f (brown pigment)c (%) (%) Without MgCl2 With MgCl2

System

% wt/vol

Glucose

50

840 ± 4

859 ± 4

1.02

1.21

60

693 ± 4

728 ± 4

1.05

1.15

1.10

70

589 ± 3

569 ± 3

0.97d

1.16

1.09

50

1191 ± 2

1035 ± 5

0.87

0.99

0.82

60

1009 ± 1

901 ± 4

0.89

1.04

0.94

70

895 ± 1

840 ± 3

0.94

1.09

0.84

Without MgCl2 With MgCl2

Trehalose

1.08

a

T2L, free-water relaxation. f (MgCl2) = T2L initially with MgCl2/T2L initially without MgCl2. c f (brown pigment, BP) = T2L in the presence of BP/T2L in the absence of BP. d Values of <1 indicate the major order of molecules and minor water mobility. Values of >1 indicate the opposite. b

of salt was independent of the brown pigment concentration achieved by the systems. Therefore, pH change was not the path by which MgCl2 inhibited the NEB reaction. The mobility of water protons was studied through the 1H-NMR spin-spin transverse relaxation times (T2) in the different systems. As shown in Table 47.1, two sets of T2 values were obtained. One of them (T2S) was between 90 and 270 ms (comprising 5%–25% of the total population) and the other (T2L) was between 500 and 1200 ms (representing 75%–95% of the total population). The presence of T2S values was attributed to the presence of a water molecule population displaying stronger interactions with solids than bulk water (assigned to T2L). As a consequence of a higher ratio of sugar molecules to water in the glucose system, trehalose systems had higher T2 values (T2S and T2L) than did glucose systems. As the sugar concentration increased, both T2 values diminished in the sugar systems as a consequence of reduced mobility. Whereas in systems containing glucose the T2 values were slightly affected by the salt (the water-glucose interaction was unmodified), in trehalose systems they were between 6% and 14% lower in the presence of salt. The relationship between the T2L values with and without MgCl2 is shown in Table 47.1. These lower values reflect that the salt imposed a certain local order to water molecules, reducing their mobility. According to Miller and de Pablo (2000), the local environment of the ions contained in a trehalose system has more water molecules with respect to those corresponding to a uniform distribution of water, which can explain the lower T2 obtained. The combined effect of pigment and salt was analyzed at fixed pigment concentration (absorbance values = 0.18), and the relative T2 values are shown in Table 47.1, column f (brown pigments). The presence of brown pigments did not modify T2 values in trehalose systems without salt (the relative values were close to 1), while in presence of salt T2 values decreased. On the other hand, T2 values increased when brown pigments were present (with or without Mg) in all glucose systems. This effect could be associated to a pigment-MgCl2 interaction, which could reduce the pigment-water interaction. The complexation of Maillard products with MgCl2 has been previously reported by O’Brien and Morrissey (1997).

544

PART 2: Poster Presentations

Conclusion It can be proposed that the different interactions with water according to the type of sugar (Suggett and others 1976; Birch and others 1989) and the interaction/ complexation between MgCl2 and brown pigments generated during the Maillard reaction in glucose systems could be responsible for inhibiting the NEB in these systems. In trehalose systems, NEB is accelerated by the presence of MgCl2 due to the enhanced sugar hydrolysis rate and to the reduction in water mobility caused by the salt (lower T2 values), which reduces the effect of water inhibition by the product.

Acknowledgments The authors acknowledge the financial support of Agencia Nacional de Promoción Científica y Tecnológica (National Agency for the Promotion of Science and Technology) (PICT 20545), Consejo Nacional de Investigaciones Científicas y Técnicas (Argentine National Scientific and Technical Research Council) (CONICET, PIP 5799), and the University of Buenos Aires (X024).

References Baisier WM, Labuza TP. 1992. Maillard browning kinetics in a liquid model system. J Agric Food Chem 40:707–13. Birch GG, Grigor J, Derbyshire W. 1989. Identification of proton type in concentrated sweet solutions by pulsed NMR analysis. J Solution Chem 18:795–801. Carpenter JF, Crowe LM, Crowe JH. 1987. Stabilization of phosphofructokinase with sugars during freezedrying: characterization of enhanced protection in the presence of divalent cations. Biochim Biophys Acta 923:109–15. Kuo MI, Gunasekaran S, Johnson M, Chen C. 2001. Nuclear magnetic resonance study of water mobility in pasta filata and nonpasta filata mozzarella. J Dairy Sci 84:1950–8. Lester MR. 1995. Sulfite sensitivity: significance in human health. J Am Coll Nutr 14:229–32. Matiacevich SB, Buera MP. 2006. A critical evaluation of fluorescence as a potential marker for the Maillard reaction. Food Chem 95:423–30. Mazzobre MF, Buera MP. 1999. Combined effects of trehalose and cations on the thermal resistance of β-galactosidase in freeze-dried systems. Biochim Biophys Acta 1473:337–44. Meiboom S, Gill D. 1958. Modified spin-echo method for measuring nuclear magnetic relaxation times. Rev Sci Instrum 29:688–91. Miller DP, de Pablo JJ. 2000. Calorimetric solution properties of simple saccharides and their significance for the stabilization of biological structure and function. J Phys Chem [B] 104:8876–83. O’Brien J, Morrissey PA. 1997. Metal ion complexation by products of the Maillard reaction. Food Chem 58:17–27. Santagapita PR, Buera MP. 2006. Chemical and physical stability of disaccharides as affected by the presence of MgCl2. In: Buera MP, Welti-Chanes J, Lillford P, Corti H, editors. Water properties of food, pharmaceutical, and biological materials. Boca Raton, FL: CRC, Taylor and Francis. p 663–9. Schebor C, Burin L, Buera MP, Chirife J. 1999. Stability to hydrolysis and browning of trehalose, sucrose and raffinose in low-moisture systems in relation to their use as protectants of dry biomaterials. LWT Food Sci Technol 32:481–5. Schmidt SJ. 1990. Characterization of water in foods by NMR. In: Finley JW, Schmidt SJ, Serianui AS, editors. NMR application in biopolymers. New York: Plenum. p 415–59. Suggett A, Ablett S, Lillford PJ. 1976. Molecular motion and interactions in aqueous carbohydrate solutions. II. Nuclear-magnetic-relaxation studies. J Solution Chem 5:17–31.

48 Combined Effect of Cinnamon Essential Oil and Water Activity on Growth Inhibition of Rhizopus stolonifer and Aspergillus flavus and Possible Application in Extending the Shelf Life of Bread S. Nanasombat, N. Piumnoppakun, D. Atikanbodee, and M. Rattanasuwan Abstract Essential oils of cinnamon (Cinnamomum verum) barks, pummelo (Citrus maxima) peels, and kaffir lime (Citrus hytrix) peels were preliminarily examined for their antifungal activities by an agar dilution method. Cinnamon oil was most inhibitory to Rhizopus stolonifer and Aspergillus flavus, with lowest minimum inhibitory concentrations (MICs) of 20 and 80 μg/mL, respectively. Cinnamon oil in combination with reduced aw had a synergistic effect on the inhibition of these fungal growths on bread model agar (BMA). Cinnamon oil (≥10 μg/mL) inhibited R. stolonifer at aw 0.90, but 20 μg/mL of this oil was inhibitory at aw 0.93 and 0.97. Water activity reduced to 0.87 alone inhibited this mold completely. Cinnamon oil (≥20 μg/mL) in BMA (aw 0.87) inhibited A. flavus completely, whereas ≥40 μg/mL cinnamon oil at aw 0.90–0.97 was inhibitory. Then, the effect of ground cinnamon (0.16%–1.30%) on the inhibition of these fungal species on bread stored at 30°C in 75% and 90% relative humidity (RH) was studied. Storage of bread at 75% RH resulted in slower fungal growth compared to the growth at 90% RH. Overall, 1.3% cinnamon showed the greatest inhibitory effect.

Introduction Bread spoilage is caused mainly by mold growth. Rhizopus stolonifer and Aspergillus flavus are molds that can grow at the characteristic water activity (aw) of bread (0.96–0.98). This causes a limited shelf life for breads. Moreover, some molds may secrete mycotoxins, which is a consumer safety concern (Smith and others 2003). Cinnamon and some citrus fruits have been reported to possess antifungal activity (Bullerman and others 1977; Caccioni and others 1998). Employing different inhibitants, such as cinnamon and citrus oils, in combination with reduced aw may effectively control the growth of R. stolonifer and A. flavus on bread. Thus, the first objective of this study was to examine the antifungal activity of cinnamon and citrus oils. Then, the combined effect of selected essential oils and water activities on the inhibition of these molds on bread model agar (BMA) were determined, along with their application in extending the shelf life of bread. 545

546

PART 2: Poster Presentations

Materials and Methods Fungal Strains and Preparation of Conidial Suspension Conidia of 7-day-old A. flavus TISTR 3041 and R. stolonifer TISTR 3199 on potato dextrose agar (PDA) slant were harvested by adding 5 mL of 0.01% Tween 80 in tubes and scraping the surface to release the conidia. Conidial concentration was adjusted to 106 conidia/mL by using a hemocytometer. Antifungal Activity of Essential Oils Antifungal activities of cinnamon (Cinnamomum verum J.S. Presl) bark oil, pummelo (Citrus maxima Merr.) peel oil, and kaffir lime (Citrus hytrix) peel oil, extracted by steam distillation, were examined against R. stolonifer and A. flavus by an agar dilution method (Collins and others 2001). Each oil was solubilized with 10% dimethyl sulfoxide (DMSO) solution, serially diluted twofold with sterile distilled water, and mixed with molten BMA (PDA with 2% wheat flour) to obtain final concentrations of 0.6–5000 μg/mL. After surface drying, 5-μL conidial suspension was inoculated at the center of each BMA plate and incubated at 30°C. The radial growth of mycelium was recorded at 3- to 4-day intervals for 14 days. Distilled water and potassium sorbate were used for negative and positive controls, respectively. The lowest concentration of essential oils that inhibited visible mold growth completely was recorded as the minimum inhibitory concentration (MIC). Effect of Cinnamon Oil and Water Activity on Inhibition of Rhizopus stolonifer and Aspergillus flavus on Bread Model Agar The inhibitory effects of cinnamon oil at 0–20 and 0–80 μg/mL against R. stolonifer and A. flavus, respectively, on reduced aw BMA (aw 0.87–0.97) were tested. Similarly, the appropriate volumes of the DMSO-solubilized oil and water were mixed with BMA adjusted to aw 0.87, 0.90, 0.93, or 0.97 by using glycerol in a sterile Petri dish to obtain each concentration of oil. All plates were incubated at 30°C. The radial growth diameters were measured at 3-day intervals for 21 days. Application of Cinnamon and Controlled Relative Humidity in Extending the Shelf Life of Bread Bread Preparation Breads with varying amounts of ground cinnamon were prepared using a straightdough method (McWilliams 2005). The bread ingredients were 100 g of wheat flour, 62 g of water, 5 g of sucrose, 1.5 g of salt, 5 g of shortening, 1 g of yeast, and ground cinnamon (0.16%, 0.32%, 0.65%, or 1.30% of dough weight). After baking, breads were aseptically cut into 1.3-cm-thick slices. The initial aw of bread was measured using a Thermoconstanter TH200 (Novasina, Zurich, Switzerland). Inoculation, Incubation, and Growth Measurement Conidial suspension (5 μL) was inoculated into the center of each bread slice. Each piece of bread was then packed in a sterile polyethylene bag and stored at 30°C in

Effect of Cinnamon Oil and Water Activity on Growth Inhibition of Mold

547

Table 48.1. Minimum inhibitory concentrations (MICs) of essential oils and potassium sorbate against Rhizopus stolonifer and Aspergillus flavus MICs (μg/mL)

Samples tested

R. stolonifer

A. flavus

Cinnamon bark oil

20

80

Kaffir lime peel oil

5000

>5000

Pummelo peel oil

5000

2560

Potassium sorbate

640

2560

75% and 90% relative humidity (RH) (using saturated sodium chloride and barium chloride, respectively). The aw, pH, and mycelial growth on the surface of bread slices were measured at 3-day intervals for 21 days. Statistical Analysis of the Data The data of three replications were analyzed by using analysis of variance and the Duncan multiple-range test.

Results and Discussion Antifungal Activity of Cinnamon and Citrus Oils Cinnamon oil had the highest antifungal activity against R. stolonifer and A. flavus, showing the lowest MICs of 20 and 80 μg/mL, respectively (Table 48.1). Compared to R. stolonifer, A. flavus was more resistant to almost all samples tested, except for pummelo peel oil. This was in agreement with previous findings (Juglal and others 2002), as cinnamon bark oil contains the major active components cinnamic aldehyde (76.0%) and eugenol (4.0%) (Wright 1991). Bullerman and others (1977) reported that these compounds exhibited antifungal activity. Some active compounds in pummelo peel and kaffir lime peel oils may contribute to their antifungal activity. The main compounds in pummelo peel oil are limonene, myrcene, β-pinene, α-pinene, sabinene, and linalool (Thavanapong 2006), whereas the major compounds in kaffir lime peel oil are citronella (17.0%), linalool (7.1%), terpineol (17.9%), terpine-4-ol (26.5%), and geranyl acetate (8.4%) (Insuan 2003). Effect of Cinnamon Oil and Water Activity on Inhibition of Rhizopus stolonifer and Aspergillus flavus on Bread Model Agar Increasing the concentration of cinnamon oil and decreasing the aw levels in BMA significantly decreased the diameter of mycelial growth in R. stolonifer and A. flavus (P < 0.05) (Figure 48.1). Cinnamon oil (20 μg/mL) inhibited growth of R. stolonifer on BMA completely at aw 0.93 and 0.97 (Figure 48.1a and b). No mycelial growth was observed on BMA (aw 0.90) with ≥10 μg/mL cinnamon oil or on BMA (aw 0.87) without cinnamon oil (Figure 48.1c and d). Cinnamon oil (≥40 μg/mL) inhibited growth of A. flavus at aw 0.90–0.97, but ≥20 μg/mL of this oil was inhibitory at aw 0.87 (Figure 48.1e–h). Aspergillus flavus

(a)

aw 0.93

(b)

aw 0.97

10 8 6 4 2 0

Mycelial growth (cm)

Mycelial growth (cm)

10 8 6 4 2 0 0

3

6

9

12

15

18

0

21

3

6

(c)

9

12

15

18

21

15

18

21

15

18

21

15

18

21

Time (days)

Time (days) (d)

aw 0.90

aw 0.87

Mycelial growth (cm)

10 8 6 4 2 0

Mycelial growth (cm)

10 8 6 4 2 0 0

3

6

9

12

15

18

21

0

3

6

Time (days) (e)

9

12

Time (days) (f)

aw 0.97

aw 0.93

Mycelial growth (cm)

10 8 6 4 2 0

Mycelial growth (cm)

10 8 6 4 2 0 0

3

6

9

12

15

18

21

0

3

6

Time (days) (g)

9

12

Time (days) (h)

aw 0.90

aw 0.87

Mycelial growth (cm)

10 8 6 4 2 0

Mycelial growth (cm)

10 8 6 4 2 0 0

3

6

9

12

Time (days)

15

18

21

0

3

6

9

12

Time (days)

Figure 48.1. Effect of cinnamon oil and water activity (aw) on growth inhibition of Rhizopus stolonifer (a–d) and Aspergillus flavus (e–h) on bread model agar at 30°C. Concentrations of cinnamon oil for R. stolonifer: , 0 μg/mL; , 5 μg/mL; , 10 μg/mL; and , 20 μg/mL. Concentrations of cinnamon oil for A. flavus: , 0 μg/mL; , 20 μg/mL; , 40 μg/mL; and , 80 μg/mL.

548

Effect of Cinnamon Oil and Water Activity on Growth Inhibition of Mold

(a)

(b)

75% RH 4 2 0

4 2 0

0

3

6

(c)

9 12 Time (days)

15

18

21

75% RH

0

3

6

9 12 Time (days)

15

18

21

15

18

21

90% RH

(d)

8

8 Mycelial growth (cm)

Mycelial growth (cm)

90% RH 6

Mycelial growth (cm)

Mycelial growth (cm)

6

549

6 4 2 0

6 4 2 0

0

3

6

9 12 Time (days)

15

18

21

0

3

6

9 12 Time (days)

Figure 48.2. Effect of cinnamon and relative humidity (RH) on growth inhibition of Rhizopus stolonifer (a and b) and Aspergillus flavus (c and d) on bread during storage at 30°C. Concentrations of ground cinnamon: , 0.16%; , 0.32%; , 0.65%; and , 1.30%.

was more resistant to cinnamon oil and low aw, compared to R. stolonifer. Aspergillus flavus had a lower minimum aw (0.78) for growth (Ayerst 1969) than R. stolonifer (minimum aw, 0.93) (Jay and others 2005). Application of Cinnamon and Controlled Relative Humidity in Extending the Shelf Life of Bread Concentrations of ground cinnamon tested were estimated based on the oil content of cinnamon bark (0.61%) and the MIC of each fungal species. The highest concentration of cinnamon (1.3%) effected the slowest growth of both fungal species on bread. At 75% RH, they grew significantly slower than at 90% RH at most storage times (P < 0.05). Storage of bread with 1.3% cinnamon at 75% RH could delay growth of R. stolonifer for 10 days and A. flavus for 6 days (Figure 48.2). The aw of breads changed as storage time increased. After a 10-day storage at 75% RH, the bread aw reduced from 0.925–0.951 to 0.905–0.908, but the aw at 90% RH increased from 0.921–0.951 to 0.935–0.949. Cinnamon concentration in breads did not affect change in aw during storage. However, the pH of breads slightly changed during storage. Bread pH values were 5.06–5.19. At 90% RH, A. flavus grew faster than at 75% RH because of its high optimum aw (0.98) (Ayerst 1969). Reduction of bread aw at 75% RH may slow the growth of these molds. Seiler (1976) reported similar findings. Storage of cake at low RH (82%) caused a 2% loss of total weight, thereby extending its shelf life by 50%. Therefore,

550

PART 2: Poster Presentations

the addition of cinnamon to bread and proper storage in suitable RH may prevent mold growth on bread.

References Ayerst G. 1969. The effects of moisture and temperature on growth and spore germination in some fungi. J Stored Prod Res 5:127–41. Bullerman LB, Lieu FY, Seier SA. 1977. Inhibition of growth and aflatoxin production by cinnamon and clove oils, cinnamic aldehyde and eugenol. J Food Sci 42:1107–9, 1116. Caccioni DRL, Guizzardi M, Biondi DM, Renda A, Ruberto G. 1998. Relationship between volatile components of citrus fruit essential oils and antimicrobial action on Penicillium digitatum and Penicillium italicum. Int J Food Microbiol 43:73–9. Collins CH, Lyne PM, Grange JM. 2001. Collins and Lyne’s microbiological methods. 7th ed. London: Arnold. Insuan W. 2003. Extraction of aromatic compounds from citrus fruit peel by using superheated water [MS thesis]. Bangkok: Kasetsart University. Jay JM, Loessner MJ, Golden DA. 2005. Modern food microbiology. 7th ed. New York: Springer Science & Business Media. Juglal S, Govinden R, Odhav B. 2002. Spice oils for the control of co-occurring mycotoxin-producing fungi. J Food Protect 65:683–7. McWilliams M. 2005. Food: experimental perspectives. 5th ed. Upper Saddle River, NJ: Pearson Prentice Hall. Seiler DAL. 1976. The stability of intermediate moisture foods with respect to mould growth. In: Davis R, Birch GG, Parker KJ, editors. Intermediate moisture foods. London: Applied Science. p 166–81. Smith JP, Daifas DP, El-Khoury W, Austin JW. 2003. Microbial safety of bakery products. In: Ashurst PR, editor. Microbial safety of minimally processed foods. Boca Raton, FL: CRC. p 3–33. Thavanapong N. 2006. The essential oil from peel and flower of Citrus maxima [MS thesis]. Bangkok: Silpakorn University. Wright J. 1991. Essential oils. In: Food flavourings. Glasgow: Blackie and Son. p 24–53.

49 From Water to Ice: Investigation of the Effect of Ice Crystal Reduction on the Stability of Frozen Large Unilamellar Vesicles L. F. Siow, T. Rades, and M. H. Lim

Abstract The solid phase of water, ice, is generally detrimental to cells because it causes freeze injury of cells during cryopreservation. Nonpermeable cryoprotective agents (CPAs) such as sucrose, trehalose, and glucose and/or permeable CPAs such as dimethyl sulfoxide (DMSO) and ethylene glycol (EG) are often added to freezing solutions to reduce such effects. The effect of ice formation on the stability of phospholipid bilayers was investigated using a model membrane: 1,2-dipalmitoyl-rac-glycero-3phosphocholine (DPPC) large unilamellar vesicles (LUVs) encapsulated with carboxyfluorescein (CF) solution. The LUV dispersion was added with a CPA, cooled to −40°C, and heated to 20°C at 10°C/min. LUV stability was described by the degree of CF leakage. In the presence of sugars, DMSO, or EG, ice formation was reduced. Simultaneous to the ice reduction, an increase of unfrozen fraction prevented LUV aggregation and leakage. The leakage was observed to decrease above a defined sugar concentration and as the concentration of DMSO or EG increased. Below the defined sugar concentration, LUV leakage increased with increasing sugar concentrations. Nonpermeable CPAs were more effective for the frozen LUVs compared to permeable CPAs. The current result suggests that LUVs were spaced out from one another in the presence of nonpermeable CPAs. On the other hand, LUVs in 10% (wt/wt) of permeable CPAs could probably undergo structural destabilization.

Introduction Cryopreservation is a technique that is widely used in the biopreservation field to preserve organisms at low temperature. Low-temperature preservation is only part of the cryopreservation procedure. A successful cryopreservation also includes the survival of the frozen organisms, sperms, oocytes, cells, or organs at room temperatures after warming and thawing of ice. The main goal of cryopreservation is to preserve organisms to prevent their extinction. Cryopreserved gametes are used in breeding programs to improve the breed of organisms, which ultimately leads to a collection of high-quality cells and tissues. Today, most cryopreservation procedures are developed empirically because the optimal procedure varies among oocytes, sperm, cells, tissues, and organs of various organisms, and each procedure is species dependent. Cryopreservation of different organisms involves various cooling and 551

552

PART 2: Poster Presentations

heating procedures, as well as the use of different types of cryoprotective agents (CPAs) at various stages of the procedures. During cooling, ice formation (Sterling 1968; Rubinsky and others 1980; Sei and Gonda 2006; Siow and others 2007a, 2007b) and the freeze-concentration effect of the unfrozen matrix (Morris and McGrath 1981; Nei 1981; Mazur and Cole 1985) have been reported as the major causes of freeze injury to various organisms, which occurred at the plasma membrane level (Morris and McGrath 1981). Similar freeze injury of plasma membranes as caused by ice formation has been demonstrated in model membrane studies of large unilamellar vesicles (LUVs) (Siow and others 2007a, 2007b). At slow cooling rates, formation of extracellular ice promotes an osmotic gradient across membrane bilayers, which could lead to cell dehydration and shrinkage (Mazur and others 1972; Pegg and Diaper 1989). At fast cooling rates, there is a limited time for cells to completely dehydrate, especially cells that have low water permeability. As a result, at fast cooling rates, intracellular freezing is likely (Mazur 1984; Pegg 2002). The intracellular ice could recrystallize during warming and disrupt cells (Mazur 1984). To improve cell survival during cryopreservation, nonpermeable and permeable CPAs are added during cryopreservation. Nonpermeable CPAs such as saccharides are usually larger than permeable CPAs, and they remain in the extracellular space during cryopreservation. In the extracellular space, nonpermeable CPAs provide osmotic gradients that dehydrate cells and prevent any intracellular ice formation which is lethal to cells. In addition, nonpermeable CPAs depress the freezing point of water (Miles and others 1997; Bhatnagar and others 2005; Sei and Gonda 2006), reduce the ice formation, and thus minimizing the increase of solute concentration by increasing the phase volume of the unfrozen matrix. The permeable CPAs such as dimethyl sulfoxide (DMSO) and ethylene glycol (EG) are smaller molecules that can permeate cell membranes and depress the freezing temperature of intracellular water, which offers the advantage of suppressing intracellular ice formation (Lovelock and Bishop 1959; Muldrew and others 2004), as well as their effect on extracellular ice reduction. This study investigated the effect of both nonpermeable and permeable CPAs on ice matrix and their influence on the cryostability of LUVs. LUVs were used as model membranes because previous studies have shown that LUVs mimic cell membranes both qualitatively and quantitatively (McGrath 1984; Callow and McGrath 1985; Siow and others 2007b). The underlying cryoprotective mechanisms of nonpermeable and permeable CPAs were elucidated by using ternary partial phase diagrams.

Materials and Methods Preparation of Samples Sugar-Salt Solutions A series of sugar-salt solutions consisting of a sugar and several types of salts—namely, N-Tris(hydroxymethyl)methyl-2-aminoethanesulfonic acid (TES), ethylenediaminetetraacetic acid (EDTA), and sodium chloride (NaCl)—were prepared with a constant

From Water to Ice

553

Table 49.1. Composition of the sugar-salt solutions with constant weight ratios (Rs) of 9.09 and 36.34 Sugar to Sugar salt ratio, R (% wt/wt) 9.09

Salt (% wt/wt)

5.0000

10.0000 20.0000 30.0000 40.0000 50.0000 60.0000

0.5503

1.1006

2.2012

3.3018

4.4024

5.5030

6.6036

Water (% wt/wt) 94.4497 88.8994 77.7988 66.6982 55.5976 44.4970 33.3964 36.34

Salt (% wt/wt)

0.5503

Water (% wt/wt)

0.8255

1.1006

1.3758

1.6509

79.4497 69.1745 58.8994 48.6242 38.3491

Table 49.2. Composition of the dimethyl sulfoxide (DMSO)–ethylene glycol (EG) salt solutions with constant weight ratios (Rs) of 4.54 and 18.17 DMSO or EG DMSO/EG to salt ratio, R (% wt/wt) 4.54

Salt (% wt/wt)

2.5000

5.0000 10.0000 20.0000 30.0000 40.0000 50.0000

0.5503

1.1006

2.2012

4.4024

6.6036

8.8048

Water (% wt/wt) 96.9497 93.8994 87.7988 75.5976 63.3964 51.1952 18.17

Salt (% wt/wt) Water (% wt/wt)

0.5503

1.1006

1.6509

2.2012

2.7515

89.4497 78.8994 68.3491 57.7988 47.2485

weight ratio (R) of sugar to salt and varying water contents (Table 49.1). At low sugar concentrations, the sugar-salt solutions were made by dissolving sugar and salt in distilled water at room temperature. At sugar concentrations above 65% (wt/wt), the sugar, salt, and water were left in a temperature-controlled water bath at 80°C until the solutions became clear upon mixing. A quantity of water equivalent to that which evaporated during heating was added after cooling of the solutions. The compositions of sucrose-salt solution with R = 9.09 and R = 36.34 are listed in Table 49.1. DMSO-Salt and EG-Salt Solutions Two series of DMSO-salt solutions with constant weight ratios of DMSO to salt of 4.54 and 18.17 and varying water contents were prepared. The DMSO-salt solutions were made up of DMSO and salts: namely, TES, EDTA, and NaCl. Two series of EG-salt solutions with constant R values were also prepared as described for the DMSO-salt solutions. The compositions of both the DMSO-salt and EG-salt solutions are listed in Table 49.2. LUV Dispersions The dispersion of LUVs was prepared as described by Oliver and others (2001). In brief, 1,2-dipalmitoyl-rac-glycero-3-phosphocholine (DPPC) (10 mg) was dissolved in chloroform (0.5 mL), and the aliquot was dried under a stream of nitrogen to form a phospholipid film. The dried phospholipid film was left overnight in a vacuum oven to remove the residual chloroform. Subsequently, 0.5 mL of carboxyfluorescein (CF) solution (100 mM of CF in 10 mM of TES and 0.1 mM of EDTA, pH 7.4) was added

554

PART 2: Poster Presentations

to rehydrate the phospholipid film for 30 min to yield a phospholipid concentration of 20 mg/mL. The phospholipid dispersion was vortexed and then extruded through a handheld extruder (Avestin, Ottawa, Canada) (Macdonald and others 1991) through two layers of 100-nm polycarbonate membranes to produce LUVs of ∼100 nm in diameter. The external CF was removed by passing the aliquot of LUV dispersion through a Sephadex G75 column in TEN buffer (10 mM of TES, 0.1 mM of EDTA, and 50 mM of NaCl, pH 7.4). Thermal Analysis Sugar-Salt-Water Ice-melting endotherms of the sugar-salt solutions were measured by using a differential scanning calorimeter (Pyris 1; Perkin-Elmer, Norwalk, CT, USA) equipped with an Intracooler II cooling unit. A sample weight of 1 mg was used for sugar-salt solutions with sugar concentrations below 50% (wt/wt). At 60% (wt/wt) sugar concentrations, approximately 2 and 35 mg of sample was used for sugar solutions with the R values of 36.34 and 9.09, respectively. Ideally, the sample size for such an analysis should be kept small and constant for all the measurements. However, since a very low amount of freezable water was available at the higher sugar concentrations, especially for sugar solutions with R = 9.09 (Table 49.1), a larger sample size was used at 60% (wt/wt) sugar solutions to improve the sensitivity for the detection of the icemelting endotherm. Below 40% (wt/wt) of sugar, the sugar-salt solutions were cooled in sealed aluminum pans to −50°C at 5°C/min, held 1 min at −50°C, and heated to 10°C at 1°C/min. At 40% (wt/wt) or higher sugar concentrations, the sugar-salt solutions were cooled to −50°C at 5°C/min, held 1 min at −50°C, and heated to the respective annealing temperatures at 5°C/min. Trehalose, sucrose, and glucose solutions were annealed for 30 min at −42° (Pyne and others 2003), −34°, and −45°C, respectively (Liesebach 2003). After annealing, the samples were cooled to −50°C at 5°C/min, held 1 min at −50°C, and heated to 10°C at 1°C/min. For the 60% (wt/wt) sucrose of the sucrose-salt solutions with R = 9.09, the sucrosesalt solutions were cooled to −65°C at 5°C/min, heated to −34°C at 5°C/min, and cooled to −65°C at 10°C/min, followed by annealing at −65°C for 60 min; long enough for ice to form. The samples were then heated to −34°C at 10°C/min and annealed at −34°C for 40 min before cooling to −50°C at 5°C/min and heating to 10°C at 1°C/min. DMSO/EG-Salt-Water One milligram of DMSO-salt solutions at concentrations ranging between 2.5% and 20% (wt/wt), as well as 4 and 10 mg of 30% and 40% (wt/wt) DMSO-salt solutions, respectively, were weighed into aluminum pans and then hermetically sealed for thermal analysis. Similar sample weights of 1, 4, and 10 mg of EG solutions ranging between 2.5% and 40% (wt/wt) were used for thermal analysis. At concentrations below 30% (wt/wt) DMSO or EG, both DMSO-salt and EG-salt solutions in the sealed aluminum pans were cooled to −50°C at 5°C/min, held 1 min at −50°C, and heated to 5°C at 1°C/min. At 30% (wt/wt), the DMSO-salt and EG-salt

From Water to Ice

555

solutions were cooled to −60°C at 5°C/min and heated to −35°C at 5°C/min, followed by isothermal annealing at −35°C for 30 min. Samples were then cooled from −35° to −50°C at 5°C/min and heated to 5°C at 1°C/min. At 40% (wt/wt), the DMSO-salt and EG-salt solutions were cooled to −60°C at 5°C/min, held at −60°C for 120 min for the ice to form, and then heated to −35°C at 5°C/min, followed by isothermal annealing at −35°C for 30 min. Samples were then cooled from −35° to −50°C at 5°C/min and heated to 5°C at 1°C/min. Phase Volume of Ice and Unfrozen Matrix Freezing curves were obtained by plotting the mean of the peak temperatures of icemelting endotherms (Tm) against CPAs and salt concentrations. The phase volume of the ice and the unfrozen matrix was calculated based on the lever rule of the freezing curves that were extrapolated by using software for second-order polynomials (Fennema 1973) (Microsoft Excel, Mountain View, CA, USA). Stability of LUV Dispersions Approximately 20 mg of the aliquot of a mixture of CPA solutions (15 μL) and LUV dispersions (15 μL) was cooled from 20° to −40°C and held 1 min at −40°C, followed by heating to 20°C at 10°C/min by using the Pyris 1. LUV stability after freezing and thawing was measured (in triplicate) at room temperature with respect to the level of leakage of the encapsulated CF by using a fluorescence spectrophotometer (Oliver and others 2001) (Varian Cary Eclipse; Varian, Palo Alto, CA, USA), with excitation at 460 nm and emission at 550 nm. Samples that were kept at room temperature were used as controls. The percentage of CF leakage from the LUVs was calculated by using Equation 49.1. ⎡ ( initial fluorescence of treated sample − ⎤ ⎢ ⎥ initial fluorescence of control) % leakage = ⎢ ×100 ( final fluorescence of treated sample − ⎥ ⎢ ⎥ initial fluorescence of control) ⎣ ⎦

(49.1)

Results and Discussion Effect of Nonpermeable and Permeable CPAs on Temperature Depression Nonpermeable CPAs In the current study, freezing curves were constructed by using sugar-salt instead of sugar-water solutions (Young 1957; Green and Angell 1989; Blond and others 1997) to mimic the extraliposomal phase of the LUVs, to which sugar solution was added to the LUV dispersion. The colligative properties of sugars are shown in Figure 49.1, in which, for both R values, a stronger temperature depression by a monosaccharide such as glucose was observed compared to disaccharides such as trehalose and sucrose at equal mass concentrations. It should be noted that Tm of the series with R = 9.09 was more depressed than that of the series with R = 36.34 (Figure 49.1), which was a result of the higher salt-to-sugar ratio of the R = 9.09 series of sugar-salt solutions

556

PART 2: Poster Presentations

(a)

0 –10 R = 9.09

Tm (°C)

–20 –30 –40

Trehalose Sucrose Glucose

–50 –60 0

(b)

20

40

60 80 Sugar + salt (% [wt/wt])

0 R = 36.34 –10

Tm (°C)

–20 –30 –40 Sucrose Glucose

–50 –60 0

20

40

60 80 Sugar + salt (% [wt/wt])

Figure 49.1. (a) Partial phase diagram of trehalose-salt-water, sucrose-salt-water, and glucose-salt-water solutions with a constant weight ratio (R) value of 9.09 (n = 3). (b) Partial phase diagram of sucrose-salt-water and glucose-salt-water solutions with an R value of 36.34 (n = 3). Tm, melting temperature.

(Table 49.1). Similar temperature depression has been reported for the sucrose-NaClwater, DMSO-NaCl-water, and the EG-NaCl-water ternary systems in which the NaCl was the most effective in lowering the temperatures of the freezing curves compared to sucrose, DMSO, or EG (Mazur and others 1972; Cocks and Brower 1974; Gayle and others 1977). This is because the NaCl is a much smaller molecule and dissociates into two species in a solution (Woods and others 1999).

From Water to Ice

557

Table 49.3. Phase volume of ice and unfrozen matrix at −40°C and leakage of DPPC large unilamellar vesicles (LUVs) at −40°C CPA to salt ratio, R 9.09

Type of CPA

Initial CPA Phase volume Phase volume of Leakage of concentrations of ice (%) unfrozen matrix (%) LUV (%) (% wt/wt) 93.9

6.1

37 ± 3

Sucrose

92.9

7.1

37 ± 5

Glucose

92.1

7.9

29 ± 1

93.0

7.0

Trehalose

5

Average 36.34

*

*

Sucrose

Trehalose

20

77.8

22.2

1±0

Glucose

73.8

26.2

2±0

75.8

24.2

Average 4.54

DMSO

2.5

EG Average 18.17

DMSO

10

EG Average

*

93.1

6.9

21 ± 1

93.9

6.1

22 ± 3

93.5

6.5

77.4

22.6

14 ± 3

79.6

20.4

16 ± 3

78.5

21.5

DPPC, 1,2-dipalmitoyl-rac-glycero-3-phosphocholine. * Not determined.

Even though the Tm was more depressed for the series of sugar-salt solutions with R = 9.09 compared to that with R = 36.34 (Figure 49.1), the phase volume of ice at a defined temperature of −40°C was estimated to be higher for the series with R = 9.09 (Table 49.3) because of its higher initial water content (with 5% [wt/wt] of initial sugar) than that of the series with R = 36.34 (with 20% [wt/wt] of initial sugar) (Table 49.1). In contrast, the phase volume of the unfrozen matrix was lower for the series with R = 9.09 compared to that of the series with R = 36.34 (Table 49.3). Permeable CPAs The Tm was more depressed for the series with the R value of 4.54 compared to the R = 18.17 series (Figure 49.2) because of the higher salt-to-DMSO or salt-to-EG ratio (Table 49.2). A higher phase volume of ice, however, was found for the series with the R value of 4.54 compared to the series with the R value of 18.17 (Table 49.3) because the initial water content of the former series was higher than that of the latter (Table 49.2). Effect of Ice Reduction on the Stability of Frozen LUVs Nonpermeable CPAs Leakage from LUVs that had been cooled to −40°C was measured after the LUVs were thawed to room temperature. Below 5% (wt/wt) of trehalose or sucrose, higher leakage from LUVs was observed compared to the control LUVs, to which no sugars were added (Figure 49.3). The higher leakage from LUVs in 5% (wt/wt) of trehalose

558

PART 2: Poster Presentations

(a)

0 –10 R = 18.17

Tm (°C)

–20 –30

R = 4.54

–40 –50 –60 0

10

20

30

40

50

60

DMSO + salt (% [wt/wt]) (b)

0 –10

Tm (°C)

–20 R = 18.17

–30 R = 4.54

–40 –50 –60 0

10

20

30

40

50

60

EG + salt (% [wt/wt])

Figure 49.2. (a) Partial phase diagram of the dimethyl sulfoxide (DMSO)-saltwater solution with constant weight ratio (R) values of 4.54 (䉱) and 18.17(䉭). (b) Partial phase diagram of the ethylene glycol (EG)-salt-water solution with R values of 4.54 (䊏) and 18.17 (䊐).

or sucrose compared to LUVs with no added sugar could be related to the viscosity effect imposed by sugars. In the presence of a small amount of sugar, the depression of Tm may have allowed the mobility of the LUVs in the remaining unfrozen matrix, which probably led to LUV collision and leakage. This suggestion was made as aggregates of LUVs were observed in 5% (wt/wt) of trehalose at −40°C (Figure 49.4a). In comparison, control LUVs were completely embedded in the extraliposomal ice matrix; thus, a close approach and interaction between LUVs was prevented. At 10% (wt/wt) of trehalose or sucrose and 5% (wt/wt) of glucose, a decrease in LUV leakage was observed (Figure 49.3). The leakage was further reduced at 15% (wt/wt) of trehalose or sucrose and 10% (wt/wt) of glucose, since the phase volume

From Water to Ice

559

45 Trehalose Sucrose Glucose

40 35 Leakage (%)

30 25 20 15 10 5 0 0

5

10

15

20 25 30 CPA concentrations (% [wt/wt])

Figure 49.3. Leakage from large unilamellar vesicles (LUVs) at various concentrations of sugar solutions that were previously cooled to −40°C and thawed to room temperature (20°C). CPA, cryoprotective agent.

Figure 49.4. Scanning electron micrographs of large unilamellar vesicles (LUVs) that were cooled to −40°C at 10°C/min. (a) LUVs in 5% (wt/wt) trehalose. The arrow shows the aggregated or fused LUVs. (b) LUVs in 20% (wt/wt) trehalose.

of ice was substantially reduced in the presence of 20% (wt/wt) of sugar compared to 5% (wt/wt) of sugar (Table 49.3). Consequently, the phase volume of the unfrozen matrix increased, thereby leaving the LUVs neither freeze-concentrated nor aggregated from one and another. This was evident in the field emission electron micrographs in which the DPPC LUVs were spaced out at −40°C in the presence of 20% (wt/wt) of trehalose or glucose (Figure 49.4b).

560

PART 2: Poster Presentations

30

EG DMSO

Leakage (%)

25 20 15 10 5 0 0

5

10

15 20 CPA concentrations (% [wt/wt])

Figure 49.5. Leakage from large unilamellar vesicles (LUVs) with various concentrations of dimethyl sulfoxide (DMSO) or ethylene glycol (EG) solutions after the mixtures were cooled to −40°C, held 1 min at −40°C, and heated to 20°C at 10°C/min. CPA, cryoprotective agent.

In terms of sugar species, lower leakage from LUVs in the presence of glucose was observed compared to sucrose and trehalose in equal mass concentrations (Figure 49.3). At a defined subfreezing temperature of −40°C, the phase volume of ice was lower in the presence of 20% (wt/wt) glucose than in the presence of trehalose or sucrose (Table 49.3). In comparison, the phase volume of the unfrozen matrix of glucose was higher than that of the trehalose and sucrose. As a result, LUVs in the presence of glucose were not being freeze-concentrated in the unfrozen matrix, thus leading to a lower leakage from LUVs. Permeable CPAs LUV leakage was observed to decrease as the concentrations of DMSO or EG increased (Figure 49.5), probably as a result of the reduction in the phase volume of ice (Table 49.3). Overall, in 93%–94% of the phase volume of ice, the lower leakage from LUVs in DMSO or EG compared to sugars (Table 49.3) could be related to the increase in the predominating polar moiety of DMSO or EG at low temperature (Arakawa and others 1990; Westh 1994). The polar moiety of DMSO has been postulated to interact with phospholipid membrane, and this interaction is suggested to be important in cryopreserving the phospholipid bilayers (Anchordoguy and others 1991). As they are small molecules, DMSO and EG have the advantage of permeating LUVs and interacting with the phospholipid bilayers, which probably results in more stable LUVs compared to sugars as CPAs that are present only in the extraliposomal space. As the phase volume of ice decreased to around 76%–79%, there was almost no LUV leakage in sugar solutions but LUV leakage in 10% (wt/wt) of DMSO or EG was observed (Table 49.3). In the presence of DMSO or EG, LUV leakage could be due to structural

From Water to Ice

561

destabilization caused by the high concentration (10% [wt/wt]) of DMSO or EG. At high concentrations of DMSO, hydrophobic interactions of DMSO and protein predominate even at low temperature (Arakawa and others 1990). The hydrophobic interactions between the acyl groups of the compound and nonpolar regions of phospholipids destabilize the phospholipid bilayers. In comparison, the extremely low leakage from LUVs in the presence of sugars at −40°C suggested that the LUVs were separated from one another and that the LUVs were solid at −40°C.

Conclusions In this study, the partial ternary phase diagrams of CPAs-salt-water showed that CPAs depressed the freezing point of water, which led to a reduction of extraliposomal ice formation and thus an increase in the unfrozen matrix. As a consequence, LUVs were protected from being compressed or freeze-concentrated by the propagating ice front and the simultaneous reduction of the unfrozen matrix. The cryoprotective mechanisms of nonpermeable CPAs were effective only if the concentrations of the nonpermeable CPAs were above a defined sugar concentration. Above that concentration, nonpermeable CPAs were more effective than permeable CPAs in preventing LUV leakage. Below the defined sugar concentration, the remaining unfrozen matrix allowed a limited mobility of the LUVs, which may have led to LUV collision and leakage. This, however, was not anticipated when the LUVs were completely embedded in the unfrozen matrix.

Acknowledgments We thank the 10th ISOPOW organizing committee for awarding the travel bursary. Thanks to Liz Girvan for assistance with the field emission scanning electron microscopy. L.F.S. thanks the University of Otago for awarding the Dr. Sulaiman Daud Jubilee 125 Postgraduate Scholarship to carry out this work and the Postgraduate Publishing Award during the preparation of this chapter.

References Anchordoguy TJ, Cecchini CA, Crowe JH, Crowe LM. 1991. Insights into the cryoprotective mechanism of dimethyl sulfoxide for phospholipid bilayers. Cryobiology 28:467–73. Arakawa T, Carpenter JF, Kita YA, Crowe JH. 1990. The basis for toxicity of certain cryoprotectants: a hypothesis. Cryobiology 27:401–5. Bhatnagar BS, Cardon S, Pikal MJ, Bogner RH. 2005. Reliable determination of freeze-concentration using DSC. Thermochim Acta 425:149–63. Blond G, Simatos D, Catte M, Dussap CG, Gros JB. 1997. Modeling of the water-sucrose state diagram below 0 degrees C. Carbohydr Res 298:139–45. Callow RA, McGrath JJ. 1985. Thermodynamic modeling and cryomicroscopy of cell-size, unilamellar, and paucilamellar liposomes. Cryobiology 22:251–67. Cocks FH, Brower WE. 1974. Phase diagram relationships in cryobiology. Cryobiology 11:340–58. Fennema OR. 1973. Solid-liquid equilibria. In: Fennema OR, Powrie WD, Marth EH, editors. Lowtemperature preservation of foods and living matter. New York: Marcel Dekker. p 101–49. Gayle FW, Cocks FH, Shepard ML. 1977. H2O-NaCl-sucrose phase diagram and applications in cryobiology. J Appl Chem Biotechnol 27:599–607.

562

PART 2: Poster Presentations

Green JL, Angell CA. 1989. Phase relations and vitrification in saccharide-water solutions and the trehalose anomaly. J Phys Chem 93:2880–2. Liesebach J. 2003. Novel approach for determining the Cg. In: Determination of the unfrozen matrix concentration and its application to describe reaction kinetics during frozen storage [PhD diss]. Dunedin, New Zealand: Department of Food Science, University of Otago. p 35. Lovelock JE, Bishop MWH. 1959. Prevention of freezing damage to living cells by dimethyl sulphoxide. Nature 183:1394–5. Macdonald RC, Macdonald RI, Menco BP, Takeshita K, Subbarao NK, Hu LR. 1991. Small-volume extrusion apparatus for preparation of large, unilamellar vesicles. Biochim Biophys Acta 1061:297–303. Mazur P. 1984. Freezing of living cells: mechanisms and implications. Am J Phys 247(3 Pt 1):C125–42. Mazur P, Cole KW. 1985. Influence of cell concentration on the contribution of unfrozen fraction and salt concentration to the survival of slowly frozen human erythrocytes. Cryobiology 22:509–36. Mazur P, Leibo SP, Chu EHY. 1972. 2-Factor hypothesis of freezing injury: evidence from Chinese hamster tissue culture cells. Exp Cell Res 71:345–55. McGrath JJ. 1984. Cryomicroscopy of liposome systems as simple models to study cellular freezing response. Cryobiology 21:81–92. Miles CA, Mayer Z, Morley MJ, Houska M. 1997. Estimating the initial freezing point of foods from composition data. Int J Food Sci Technol 32:389–400. Morris GJ, McGrath JJ. 1981. The response of multilamellar liposomes to freezing and thawing. Cryobiology 18:390–8. Muldrew K, Acker JP, Elliott JAW, McGann LE. 2004. The water to ice transition: implications for living cells. In: Fuller BJ, Lane N, Benson EE, editors. Life in the frozen state. Boca Raton, FL: CRC. p 67–108. Nei T. 1981. Mechanism of freezing injury to erythrocytes: effect of initial cell concentration on the postthaw hemolysis. Cryobiology 18:229–37. Oliver AE, Hincha DK, Tsvetkova NM, Vigh L, Crowe JH. 2001. The effect of arbutin on membrane integrity during drying is mediated by stabilization of the lamellar phase in the presence of nonbilayerforming lipids. Chem Phys Lipids 111:37–57. Pegg DE. 2002. The history and principles of cryopreservation. Semin Reprod Med 20:5–13. Pegg DE, Diaper MP. 1989. The unfrozen fraction hypothesis of freezing injury to human erythrocytes: a critical examination of the evidence. Cryobiology 26:30–43. Pyne A, Surana R, Suryanarayanan R. 2003. Enthalpic relaxation in frozen aqueous trehalose solutions. Thermochim Acta 405:225–34. Rubinsky B, Lee CY, Bastacky J, Onik G. 1990. The process of freezing and the mechanism of damage during hepatic cryosurgery. Cryobiology 27:85–97. Sei T, Gonda T. 2006. Melting point of ice in aqueous saccharide solutions. J Cryst Growth 293:110–2. Siow LF, Rades T, Lim MH. 2007a. Effect of intra/extraliposomal distribution of sodium chloride on the stability of large unilamellar vesicles. Cryo Lett 28:429–44. Siow LF, Rades T, Lim MH. 2007b. Characterizing the freezing behavior of liposomes as a tool to understand the cryopreservation procedures. Cryobiology 55:210–21. Sterling C. 1968. Effect of low temperature on structure and firmness of apple tissue. J Food Sci 33:577–80. Westh P. 1994. Thermal expansivity, molar volume, and heat capacity of liquid dimethyl sulfoxide water mixtures at subzero temperatures. J Phys Chem 98:3222–5. Woods EJ, Zieger MAJ, Gao DY, Critser JK. 1999. Equations for obtaining melting points for the ternary system ethylene glycol/sodium chloride/water and their application to cryopreservation. Cryobiology 38:403–7. Young FE. 1957. D-glucose–water phase diagram. J Phys Chem 61:616–9.

50 Does Microencapsulation Improve Storage Stability of Cloudberry (Rubus chamaemorus) Ellagitannins? P. Laine, P. Kylli, M. Heinonen, and K. Jouppila

Abstract The cloudberry phenolics, of which ellagitannins (ET) are the major compounds, have been associated with many valuable functions such as antioxidant and antimicrobial activities. Our research studied the storage stability of microencapsulated cloudberry ET to determine whether storage stability depended on the physical state of the capsule matrix. Maltodextrins DE5–8 and DE18.5 were used as capsule materials, and microcapsules were produced by freeze drying. For analysis of physical state, water-sorption properties and glass transition temperature (Tg) of maltodextrins and microcapsules were determined. Results showed that microcapsules’ stability could not be explained solely by the physical state of the capsule matrix, since ET losses were observed even at temperatures below the Tg. The retention of ET was strongly dependent on storage RH and the capsule matrix used. At 33% RH, microcapsules made of maltodextrin DE5–8 retained all ET, whereas ET retention in microcapsules made of maltodextrin DE18.5 was poor. At 66% RH, neither maltodextrin protected ET. The antioxidant activity of all the samples remained almost the same as they were at the beginning of the study, irrespective of the storage RH or capsule material. The present study suggested that, at low RH, the storage stability of ET can be increased via encapsulation by maltodextrin DE5–8.

Introduction The interest in cloudberry (Rubus chamaemorus) phenolics has increased since previous studies demonstrated that phenolics can have not only antioxidative activity (Kähkönen and others 2001) and antimicrobial activity (Puupponen-Pimiä 2005) but also a preventive effect against cancer cell proliferation (Wu and others 2007). The most predominant phenolics in cloudberries are ellagitannins (ET), which may comprise 60%–80% of all phenolics (Kähkönen and others 2001). From a technologist’s perspective, it would be desirable to be able to extract ET from cloudberries and dry them into a powder. These powders would be convenient to handle and portion out. However, the tendency of ET toward chemical transformation (oxidation, degradation, polymerization) and their susceptibility to reactions with a variety of molecules (organic and inorganic) may cause problems during storage of such powders (Quideau and Feldman 1996; Haslam 1998; Lei 2002). Microencapsulation, which is often 563

564

PART 2: Poster Presentations

used to protect sensitive food components, may be a practical way of stabilizing ET during storage (Thies 2004). Microcapsules are defined as microsized (<800 μm) particles that consist of capsule material(s) and encapsulated core component(s) (Thies 2004). The main function of a microcapsule is to prevent diffusion of material into or from a microcapsule and consequently, to protect the encapsulated component (Thies 2004). Freeze drying (along with spray drying and extrusion cooking) is among the most commonly used encapsulation technologies in the food field. Core materials encapsulated by freeze drying are typically heat sensitive and include, for example, flavors (Che Man and others 1999) and colors (Serris and Biliaderis 2001). Maltodextrins are potential capsule materials because they can stabilize core materials by entrapping them within amorphous glassy matrices during freeze drying. However, the protective effect of the glassy matrix may be lost if capsules are stored above the glass transition temperature (Tg) of capsule material (Thies 2004). In that case, molecular mobility and diffusion increase, and consequently, the amorphous matrix transforms from a glassy to a rubbery state, resulting in collapse of the capsule structure (Roos 1995). Water plays an important role in that transition by plasticizing amorphous materials and thus reducing the Tg (Roos 1995). Considering the aforementioned issues, the ability of capsule material to protect encapsulated components during storage depends strongly on the storage conditions; for example, relative humidity (RH), temperature, and the presence of oxygen (Thies 2004). The aims of the present study were (a) to produce microencapsulated cloudberry ET by freeze drying using maltodextrins as capsule material, (b) to monitor the storage stability of encapsulated and nonencapsulated cloudberry ET at different RHs at 25°C, and (c) to evaluate whether the physical state of microcapsules could explain the storage stability of encapsulated ET.

Materials and Methods The following maltodextrins were used as capsule materials: C*Dry A 01318 with DE18.5 (MD18.5) and C*Dry MD 01955 with DE5–8 (MD5–8). The products were donated by Cerestar Finland/Cargill (Helsinki, Finland). The abbreviations MC5–8 and MC18.5 are used later in this text to describe microcapsules produced with MD5–8 and MD18.5, respectively. Cloudberries were purchased from a local store. Cloudberry phenolics were extracted with 70% acetone in water, and the extract was purified by using an Amberlite XAD column to remove sugars. The extract was freeze-dried to achieve powder that included ET (62%), hydroxycinnamic acids (10%), hydroxybenzoic acids (10%), and catechins (10%). To prepare microcapsules, maltodextrin extract (9% wt/wt) and cloudberry phenolic extract (1% wt/wt) were dissolved in distilled water and freeze-dried using a Lyovac GT2 freeze dryer (Amsco Finn-Aqua, Hürth, Germany). As a reference, an ET-water solution (without added maltodextrin) and a maltodextrin-water solution (without ET

Does Microencapsulation Improve Storage Stability of Cloudberry Ellagitannins?

565

added) were freeze-dried similarly. After drying, samples were placed in a desiccator over phosphorus pentoxide for a week (at least) to produce anhydrous samples. Water-sorption properties for microcapsules and freeze-dried maltodextrins were determined at 25°C by equilibrating dry samples (5–30 mg) in vacuum desiccators over saturated salt solutions: lithium chloride, potassium acetate, magnesium chloride, potassium carbonate, magnesium nitrate, sodium nitrite, and sodium chloride, providing RH conditions of 11%, 24%, 33%, 44%, 54%, 66%, and 76%, respectively. Water sorption was determined gravimetrically as a weight gain over time. The GuggenheimAnderson-de Boer (GAB) model was fitted to the data. Glass transition temperatures (Tg, onset values) of microcapsules and freeze-dried maltodextrins were determined after storage at 25°C at various RHs (0–76%). Tg was determined by using a differential scanning calorimeter (DSC823; Mettler Toledo, Greifensee, Switzerland). A heating rate of 5°C min−1 was used. The data were modeled by using the Gordon-Taylor equation. The encapsulated and nonencapsulated extracts were stored at 33% and 66% RH at 25°C for 64 days. Storage stability was monitored by analyzing ET and ellagic acid content with high-performance liquid chromatography (Kähkönen and others 2001) (Waters Alliance 2690; Waters, Milford, MA, USA) coupled with a diode-array detector (Waters 996 PDA) and by analyzing antioxidant activity by using a liposome oxidation model. The inhibition of liposome oxidation was calculated by measuring the formation of the primary (conjugate diene hydroperoxides) and secondary (hexanal) oxidation products.

Results and Discussion Water Sorption and Glass Transition Temperature The water-sorption isotherms of the maltodextrins and microcapsules had a sigmoidal shape typical of food material (Roos 1995) (Figure 50.1). Predictably, water sorption increased with increasing storage RH. MD5–8 sorbed slightly more water than did MD18.5. These results contrast with those reported previously (Roos 1993), showing that maltodextrins with lower average molecular weight (MW) sorbed more water than those with higher MW. The difference between the results may be explained by the somewhat different compositions of the maltodextrin preparations used in the two studies. The water-sorption behavior of microcapsules and maltodextrins differed. Microcapsules sorbed less water than did the corresponding maltodextrin matrix. This perhaps indicates that maltodextrins and extract interacted by forming complexes during microencapsulation and thus exhibited a more hydrophobic character compared to maltodextrins alone. It has been observed in previous studies that ET, as amphiphilic components, may form weak-bonded complexes with some carbohydrates (Haslam 1998). The Tg values of microcapsules and maltodextrins decreased with increasing water content, which is consistent with previous findings (Roos 1993) (Figure 50.1). Microcapsules had a slightly lower Tg compared to maltodextrins, indicating that

566

PART 2: Poster Presentations

Figure 50.1. Relationships between glass transition temperature (Tg), water activity (aw), and water content for maltodextrins (MDs) with DE18.5 and DE5–8 and microcapsules (MCs) produced with MDs DE18.5 and DE5–8. Experimental water-sorption data were fitted by using the Guggenheim-Anderson-de Boer (GAB) model. Glass transition data were modeled with the Gordon-Taylor equation.

cloudberry extract contained some low-MW component(s) that decreased the Tg of the microcapsules. ET, which were the major components of the cloudberry extract, might have a relatively high MW (up to 4000) (Haslam 1996; Clifford and Scalbert 2000), whereas low-MW phenolics and possibly traces of sugars, which were minor components in cloudberry extract, might reduce Tg. Critical water content and water activity (i.e., values that decrease the Tg of material to 25°C [Roos 1993]), were lower for MC18.5 than for MC5–8 (Figure 50.1). Storage Stability The storage stability of encapsulated and nonencapsulated ET at different RH at 25°C was monitored for 64 days, and the changes in ET/ellagic acid (EA) content and

Does Microencapsulation Improve Storage Stability of Cloudberry Ellagitannins?

567

Table 50.1. Storage stability of encapsulated and nonencapsulated cloudberry extract at different relative humidities (RHs) evaluated as retention of ellagitannins (ET), content of ellagic acid (EA), and antioxidative activity

Retention of ET (%)

EA content (%)

Inhibition of hexanal formation (%) Inhibition of conjugate diene formation (%)

Storage time (days)

MC5–8

Storage at 33% RH MC18.5

C

MC5–8

Storage at 66% RH MC18.5

C

0

100

100

100

100

100

100

32

101

76

89

26

27

80

64

119

31

63

10

25

34

0

100

100

100

100

100

100

32

105

97

108

91

90

148

64

172

182

116

120

248

159

0

77

81

80

77

81

80

32

71

72

72

76

73

77

64

82

75

80

84

81

80

0

69

68

68

69

68

68

32

62

62

64

68

64

69

64

75

67

72

76

72

73

Antioxidant activity was determined by inhibition of hexanal and conjugated diene formation. MC5–8, microcapsules prepared with maltodextrin DE5–8; MC18.5, microcapsules prepared with maltodextrin DE18.5; and C, nonencapsulated cloudberry extract.

antioxidative activity were analyzed. The retention of ET was strongly dependent on storage RH and the capsule material used (Table 50.1). The lower the RH, or the higher the MW of the maltodextrins, the greater was the amount of ET remaining in the microcapsules. Conversely, EA content increased in conjunction with higher storage RH and lower MW of maltodextrin. At 33% RH, MD5–8 retained all ET whereas ET retention in MC18.5 and in nonencapsulated extract decreased remarkably. The reason for the considerably better stability of encapsulated ET in MC5–8 might be due to the higher Tg value for microcapsules (Tg, 77°C) compared to the corresponding value for MC18.5 (Tg, 41°C). Regardless of glassy state, MD18.5 failed to protect the ET. A number of studies have found that the chemical stability of microencapsulated components can be related not only to physical state of microcapsules, but also to other properties such as their porous and heterogeneous structure (Serris and Biliaderis 2001; Yoshioka and Aso 2007). At 66% RH, neither maltodextrin promoted ET retention. The Tg of MC18.5 was below the storage temperature (Tg, 14°C), whereas the Tg of MC5–8 was close to the storage temperature (Tg, 33°C), which probably contributed to the poor protection of ET. Various speculations can be presented about the reason for loss of ET during storage. First, ET probably disappeared because of polymerization reactions during oxidation (Lei 2002). Second, the phenolics might interact with carbohydrates by forming complexes with them (Haslam 1998), perhaps influencing changes in the extractability of the phenolics. Third, the loss

568

PART 2: Poster Presentations

of ET might be due to its hydrolysis (at higher storage RH), which resulted in liberation of EA from the ET structure (Lei 2002). This explanation seems possible since ET losses and EA formation increased together with storage RH. Also, the fact that the color of microencapsulated and nonencapsulated extract changed from reddish to brown when stored at high RH may be related to hydrolysis (and other unspecified reactions) of ET (Lei 2002). The antioxidative activity of encapsulated and nonencapsulated cloudberry extract remained almost the same as at the beginning, or even improved slightly during, the storage period (Table 50.1). The values of the inhibition of conjugate dienes and hexanal formation varied from 62%–75% and 71%–84%, respectively, which means that both encapsulated and nonencapsulated cloudberry extract had moderate antioxidative activities. Neither capsule material nor storage conditions influenced the antioxidative activity of the samples. A variation or decrease in ET content during storage did not seem to affect the total antioxidative activities of either nonencapsulated or encapsulated extract. Perhaps the loss of originally occurring antioxidants was compensated for by the formation of new phenolic compounds with equal or improved antioxidative activities, as was reviewed recently (Nicoli and others 1999).

Conclusions The present study suggests that, at low RH, the storage stability of ET can be increased via encapsulation by maltodextrin DE5–8. On the contrary, maltodextrin DE18.5 failed to protect ET under any storage conditions tested. The results demonstrate that the protection of ET by microcapsules is associated not only with the physical state of microcapsule matrices, but also with some other factors, probably such as the porosity and heterogeneity of the matrix structure. Further research is needed to establish whether microencapsulated cloudberry ET can be produced that are stable at higher RH. More hydrophobic capsule material(s) might provide effective protection for cloudberry ET during storage.

References Che Man YB, Irwandi J, Abdullah WJW. 1999. Effect of different types of maltodextrin and drying methods on physico-chemical and sensory properties of encapsulated durian flavour. J Sci Food Agric 79:1075–80. Clifford MN, Scalbert A. 2000. Ellagitannins: nature, occurrence and dietary burden [Review]. J Sci Food Agric 80:1118–25. Haslam E. 1996. Natural polyphenols (vegetable tannins) as drugs: possible modes of action. J Nat Prod 59:205–15. Haslam E. 1998. Plant polyphenols: vegetable tannins revisited. Cambridge: Cambridge University Press. Kähkönen MP, Hopia AI, Heinonen M. 2001. Berry phenolics and their antioxidant activity. J Agric Food Chem 49:4076–82. Nicoli MC, Anese M, Parpinel M. 1999. Influence of processing on the antioxidant properties of fruit and vegetables. Trends Food Sci Technol 10:94–100. Puupponen-Pimiä R, Nohynek L, Hartmann-Schmidlin S, Kähkönen M, Heinonen M, Määttä-Riihinen K, Oksman-Caldentey K-M. 2005. Berry phenolics selectively inhibit the growth of intestinal pathogens. J Appl Microbiol 98:991–1000.

Does Microencapsulation Improve Storage Stability of Cloudberry Ellagitannins?

569

Quideau S, Feldman KS. 1996. Ellagitannin chemistry. Chem Rev 96:475–503. Roos YH. 1993. Water activity and physical state effects on amorphous food stability. J Food Process Preserv 16:433–47. Roos YH. 1995. Phase transitions in foods. San Diego: Academic. Serris GS, Biliaderis CG. 2001. Degradation kinetics of beetroot pigment encapsulated in polymeric matrices. J Sci Food Agric 81:691–700. Thies C. 2004. Microencapsulation: what it is and purpose and microcapsule characterization. In: Vilstrup P, editor. Microencapsulation of food ingredients. Surrey, UK: Leatherhead International. p 1–54. Wu QK, Koponen JM, Mykkänen HM, Törrönen AR. 2007. Berry phenolic extracts modulate the expression of p21WAF1 and Bax but not Bcl-2 in HT-29 colon cancer cells. J Agric Food Chem 55:1156–63. Yoshioka S, Aso Y. 2007. Correlations between molecular mobility and chemical stability during storage of amorphous pharmaceuticals [Review]. J Pharm Sci 96:960–81. Lei Z. 2002. Monomeric ellagitannins in oaks and sweetgum [PhD diss]. Blacksburg, VA: Faculty of Virginia Polytechnic Institute and State University, Wood Science and Forest Products.

51 Nonenzymatic Browning Reaction of Glassy Foods: Characterization of Local Reactions Independent of the Glassy Matrix K. Kawai, T. Suzuki, and K. Kajiwara Abstract The rate of the nonenzymatic browning reaction (NEB) of model freeze-dried foods was investigated at temperatures below their glass transition temperature (Tg). The model samples used were NEB reactants (lysine and glucose) embedded in various glassy matrices (e.g., saccharide and polymer), and NEB reactants (lysine and various reducing sugars) embedded in a glassy trehalose matrix. The Tg of the samples depended on the type of matrix, but the type of reducing sugar had little effect on the Tg. In comparison, the NEB rate of lysine and glucose embedded in glassy polymer matrices decreased with increased Tg of the polymer matrices. Although the Tg of saccharide matrices was much lower than that of polymer matrices, the NEB rates in saccharide matrices were lower than those in polymer matrices. In comparison, the NEB rate of lysine and various reducing sugars embedded in a glassy trehalose matrix depended strongly on the type of reducing sugar even though the samples had a nearly equivalent Tg. Previous studies reported that hydrogen bonds between saccharides and amino acids were formed efficiently in the dehydrated state, though hydrogen bonds between polymers and amino acids were restricted by steric hindrance. These observations suggest that the NEB rate in a glassy matrix is affected not only by Tg, but also by hydrogen bonds between the NEB reactants and the matrix.

Introduction For practical purposes, the prediction and control of the nonenzymatic browning reaction (NEB) in dry foods is important because the NEB causes various quality losses (brown color formation, flavor change, and loss of nutrition and functionality) in food products. It is known that most dry-food products are in an amorphous state, and that rates of chemical and physical degradation are affected by glass transition. Amorphous materials show a transition between the highly viscous “glassy” state and the fluid “rubbery” state at the glass transition temperature (Tg). At temperatures below the Tg, it is believed that the progress of the NEB is inhibited by its extremely low molecular mobility. In addition, it is thought generally that the higher the Tg is, the lower the molecular mobility is at a reference temperature, because Tg is the temperature at which the viscosity of amorphous materials becomes a constant (∼1012 Pa). Thus, the NEB rate is expected to decrease with an increase in the Tg. 571

572

PART 2: Poster Presentations

The relation between the NEB and the Tg of amorphous food products has been studied extensively. In the previous studies, it was evident that the NEB rate was much lower at temperatures below the Tg as compared with temperatures above the Tg (Karmas and others 1992; Roos and Himberg 1994; Buera and Karel 1995; Bell 1996; O’Brien 1996; Bell and others 1998; Lievonen and others 1998, 2002). However, it should be noted that the NEB has been observed to progress even at temperatures below the Tg (Bell 1996; Bell and others 1998; Schebor and others 1999). That is, NEB progress is not always prevented even in the glassy state. To obtain some insight into the subject, the NEB rate of various types of model glassy food was investigated.

Materials and Methods Preparation of the Glassy Food Model Two different types of samples were prepared. In one, NEB reactants (lysine and glucose) were embedded in various glassy matrices: trehalose; raffinose; dextran (MW = 10,000); polyvinylpyrrolidone (PVP) PVPk13–19 (MW = 10,000); PVPk30 (MW = 40,000); and PVPk90 (MW = 360,000). In the other, NEB reactants (lysine and various reducing sugars: arabinose, xylose, galactose, glucose, lactose, and maltose) were embedded in the glassy trehalose matrix. A 10% (wt/wt) aqueous solution of a reducing sugar/lysine/glass-matrix–former (1 : 1 : 98 weight, dry basis) was prepared, and the fraction (2 mL) was placed into a 10-mL glass vial, frozen for 12 h at −50°C, and transferred to a precooled freeze dryer (RLE-52; Kyowa Vacuum Engineering, Tokyo, Japan). The frozen formulation was freeze-dried at 3.0 × 10−2 Torr, and then gradually warmed from −40° to 20°C over a 2-day period. The residual moisture of the freeze-dried solid was removed by storing over diphosphorus pentoxide in a vacuum desiccator for 1 week at room temperature. The moisture content of the samples was confirmed to be below 0.5% (wt/wt). Evaluation of the Initial NEB Rate The samples were stored at temperature ranges between 40° and 80°C over a 2-month period, after which the samples were rehydrated with distilled water. The fraction (3 mL) of the solution was placed into a cell (10 × 10 mm), and then absorbance per gram of sample at 280 nm was measured as representing the extent of NEB. The measurement was performed in triplicate and averaged. A pseudo–zero-order reaction rate constant (k280) was evaluated by determining the initial rate of NEB from the initial time dependence of the averaged absorbance as in prior studies (Karmas and others 1992; O’Brien 1996; Lievonen and others 1998, 2002). DSC Measurement The Tg of the samples was investigated by using a differential scanning calorimeter (DSC) (DSC-50; Shimadzu, Kyoto, Japan). Indium and distilled water were used to calibrate the temperature and heat capacity of the DSC, and alumina powder was used as a reference material. Approximately 10 mg of the sample was weighed and then sealed in an aluminum DSC pan. Each sample was scanned at 5°C/min from 0° to 140°C.

Nonenzymatic Browning Reaction of Glassy Foods

573

5 PVPk15 (Tg = 81 ºC)

4

PVPk30 (Tg = 132 ºC)

3

PVPk90 (Tg = 165 ºC)

2

ln k280

1 0 –1 –2

trehalose (Tg = 90 ºC) rafinose (Tg = 93 ºC)

–3 –4 2.9

dextran (Tg = 158 ºC)

3.0

3.1

3.2

1000/T

Figure 51.1. Arrhenius plot of the initial nonenzymatic browning reaction (NEB) rate k280 for samples of varying glassy matrices. T, absolute temperature; and Tg, glass transition temperature.

The onset point of the endothermic shift produced by the glass transition was considered to be the Tg.

Results Glass Transition and the NEB Rate of the Samples of Various Glassy Matrices The temperature dependence of the initial rate of NEB (k280) is shown as an Arrhenius plot in Figure 51.1, as is the Tg of the samples. The Tg of the samples depended strongly on the type of glassy matrix and tended to increase with an increase in molecular weight of the matrix. As already mentioned, it is expected that the higher the Tg, the lower the NEB rate is at storage temperatures. Comparing the k280 among the three types of PVP matrix samples of varying molecular weights confirmed that the k280 decreased with an increase in the Tg. However, when the k280 of the PVPk90 sample was compared with that of the other glassy samples (trehalose, raffinose, and dextran), although PVPk90’s Tg (165°C) was the highest, its k280 was higher than that of the trehalose, raffinose, and dextran samples. In addition, the k280 of the dextran (Tg = 158°C) was close to that of the trehalose (Tg = 90°C) and raffinose (Tg = 93°C) samples. These results show that the progress of NEB is at least partially independent of the molecular mobility of glassy matrix.

574

PART 2: Poster Presentations

5 arabinose (Tg = 91 ºC) xylose (Tg = 93 ºC) galactose (Tg = 89 ºC)

4 3

ln k280

2 1 0 –1 –2

glucose (Tg = 90 ºC) lactose (Tg = 97 ºC) maltose (Tg = 96 ºC)

–3 –4 2.8

2.9

3.0

3.1

3.2

1000/T

Figure 51.2. Arrhenius plot of the initial nonenzymatic browning reaction (NEB) rate for samples of varying reducing sugars embedded in a glassy trehalose matrix. T, absolute temperature; Tg, glass transition temperature.

Glass Transition and the NEB Rate of Samples of Various Reducing Sugars The temperature dependence of the k280 of trehalose matrix samples of varying reducing sugars is shown as an Arrhenius plot in Figure 51.2, as is the Tg of the samples. The type of reducing sugar had a minor effect on the Tg of the samples, with the Tg varying between 89° and 96°C. The NEB rate of the samples, on the other hand, depended strongly on the type of reducing sugar. When the k280 of each sample was compared, it was higher in the samples in the following order: pentose (xylose and arabinose), hexose (glucose and galactose), and disaccharide (lactose and maltose). These results, again, show that NEB progress is at least partially independent of the molecular mobility of the glassy matrix and suggests that a local reaction promotes NEB progress in a glassy matrix. The origin of this local reaction is discussed in the Discussion section, below.

Discussion It is well known that a freeze-dried protein-sugar mixture forms intermolecular hydrogen bonds between protein and sugar (Carpenter and Crowe 1989), so it can be reasonably assumed that amino acids, reducing sugars, and the glassy matrix also form intermolecular hydrogen bonds when in the dehydrated state. Taking this into account, the following interpretation is suggested as a mechanism of the NEB progress in the glassy matrix. When intermolecular hydrogen bonds are formed between the glassy matrix and NEB reactants (amino acids and reducing sugars), NEB progress will be effectively prevented, because the NEB reactants break the hydrogen bonds during the molecular rearrangement. Because of the steric hindrance, however, some NEB reactants are not saturated with intermolecular hydrogen bonds. In addition, intermo-

Nonenzymatic Browning Reaction of Glassy Foods

reducing sugar

575

(a)

amino acid trehalose matrix polymer matrix hydrogen bond (b)

(c)

Figure 51.3. Model of a local reaction portraying the nonenzymatic browning reaction (NEB) progress in the glassy matrix: (a) reducing sugars and amino acids embedded in a glassy trehalose matrix, (b) reducing sugars and amino acids embedded in a glassy polymer matrix, and (c) the effect of types of reducing sugars embedded in a glassy trehalose matrix.

lecular hydrogen bonds between amino acids and reducing sugars are also formed. These NEB reactants will react readily because of their loose entrapment by the matrix. As a result, a local reaction promoting the progress of the NEB occurs in the glassy matrix. The NEB rate was strongly affected by the types of glassy matrix (Figure 51.1). When the glassy matrix can form many intermolecular hydrogen bonds, the formation of hydrogen bonds between amino acids and reducing sugars will relatively be limited, and the progress of the NEB will be effectively prevented. Since trehalose and raffinose matrices can form many intermolecular hydrogen bonds (Carpenter and Crowe 1989; Davidson and Sun 2001), they entrap sufficient NEB reactants (Figure 51.3a). Compared to trehalose and raffinose matrices, polymer matrices (dextran and PVP) form a few intermolecular hydrogen bonds because of the significant steric hindrance (Crowe and others 1994). Furthermore, the ability of PVP, which has one pyrrolidone group per monomer, to form intermolecular hydrogen bonds is much inferior to that of dextran, which has some hydroxyl groups per monomer. Thus, NEB reactants embedded in the PVP matrix can react readily even if the matrix has a high Tg (Figure 51.3b). The NEB rate, on the other hand, depends on the types of reducing sugars embedded in the glassy trehalose matrix (Figure 51.2). As already mentioned, the

576

PART 2: Poster Presentations

glassy trehalose matrix can entrap NEB reactants effectively because of its ability to form many intermolecular hydrogen bonds. The NEB progress in the glassy matrix, however, will be also affected by the ability of the NEB reactants to form intermolecular hydrogen bonds: the NEB reactants that can form more intermolecular hydrogen bonds require more energy to break the bonds during the molecular rearrangement (Figure 51.3c). When the reducing sugars were compared, the number of hydroxyl groups (reflecting the ability to form intermolecular hydrogen bonds) of the reducing sugars was higher in the following order: disaccharide (lactose and maltose), hexose (glucose and galactose), and pentose (xylose and arabinose). These concepts reasonably explain the NEB progress in the glassy matrix in the study.

Conclusion It was found that the progress of the NEB is at least partially independent of the molecular mobility of the glassy matrix, and the source of a local reaction promoting the NEB progress was suggested to be the formation of intermolecular hydrogen bonds among amino acids, reducing sugars, and the glassy matrix. A possible extension of this research would involve the collection of further fundamental data on the progress of the NEB in the glassy matrix (e.g., the effect of various types of amino acids on the NEB rate in the glassy matrix) and clarification of the local reaction in the glassy matrix caused by the intermolecular hydrogen bonds.

References Bell LN. 1996. Kinetics of nonenzymatic browning in amorphous solid systems: distinguishing the effects of water activity and the glass transition. Food Res Int 28:591–7. Bell LN, Touma DE, White KL, Chen YH. 1998. Glycine loss and Maillard browning as related to the glass transition in a model food system. J Food Sci 63:625–8. Buera MP, Karel M. 1995. Effect of physical changes on the rates of nonenzymic browning and related reactions. Food Chem 52:167–73. Carpenter JF, Crowe JH. 1989. An infrared spectroscopic study of the interactions of carbohydrates with dried proteins. Biochemistry 28:3916–22. Crowe JH, Leslie SB, Crowe LM. 1994. Is vitrification sufficient to preserve liposomes during freezedrying? Cryobiology 31:355–66. Davidson P, Sun WQ. 2001. Effect of sucrose/raffinose mass ratios on the stability of co-lyophilized protein during storage above the Tg. Pharm Res 18:474–9. Karmas R, Buera MP, Karel M. 1992. Effect of glass transition on rates of nonenzymatic browning in food systems. J Agric Food Chem 40:873–9. Lievonen SM, Laaksonen TJ, Roos YH. 1998. Glass transition and reaction rates: nonenzymatic browning in glassy and liquid systems. J Agric Food Chem 46:2778–84. Lievonen SM, Laaksonen TJ, Roos YH. 2002. Nonenzymatic browning in food models in the vicinity of the glass transition: effect of fructose, glucose, and xylose as reducing sugar. J Agric Food Chem 50:7034–41. O’Brien J. 1996. Stability of trehalose, sucrose and glucose to nonenzymatic browning in model systems. J Food Sci 61:679–82. Roos YH, Himberg MJ. 1994. Nonenzymatic browning behavior, as related to glass transition, of a food model at chilling temperatures. J Agric Food Chem 42:893–8. Schebor C, Buera MP, Karel M, Chirife J. 1999. Color formation due to non-enzymatic browning in amorphous, glassy, anhydrous, model systems. Food Chem 65:427–32.

52 Physical Properties of ProteinCarbohydrate Sheets Produced by a Twin-Screw Extruder R. A. Talja, K. Pehkonen, K. Jouppila, and Y. H. Roos

Abstract The present study investigated the mechanical and oxygen permeability (OP) properties of protein-carbohydrate sheets close to their glass transition temperature (Tg). Protein-carbohydrate sheets were prepared using a twin-screw extruder. The Nacaseinate-trehalose, Na-caseinate-trehalose-lactose, and gelatin–gum arabic–sucrose sheets were translucent, whereas the Na-caseinate-lactose sheet was opaque. The fresh sheets were flexible and elastic because of their high water content (13–22 g/100 g of solids). The sheets became rigid after equilibration at 33% relative vapor pressure (RVP), which was observed from low elongation at break values (around 1%–2%), indicating that the sheets were in a glassy state. The Tg values and relaxations of the sheets determined by differential scanning calorimeter, dielectric analyzer, and dynamic mechanical analyzer changed from 13° to 22°C, 2° to 24°C, and 25° to 31°C, respectively. The highest values for tensile strength and Young’s modulus were observed for the gelatin–gum arabic–sucrose sheet. The OP of the gelatin–gum arabic– sucrose sheet was the lowest among all of the sheets. The gelatin–gum arabic–sucrose sheet was supposed to be denser than the other sheets, resulting in higher tensile properties and lower OP properties. Overall, the gelatin–gum arabic–sucrose sheet had the best protection from oxygen although the Na-caseinate-lactose and Na-caseinatetrehalose-lactose sheets provided almost equal protection.

Introduction Low-moisture food ingredients may be protected from water and oxygen by encapsulating them in glassy protein-carbohydrate films or/and matrices. Hydrophilic proteincarbohydrate matrices can be plasticized by water, which results in increased molecular mobility and may decrease the quality of food ingredients. When protein-carbohydrate systems are used to protect food ingredients against gases and water, low gas and water diffusion in the matrix is desired. Oxygen permeability (OP) of films changes at the glass transition temperature (Tg), which can be predicted from the permeability data obtained at various temperatures (Arvanitoyannis and others 1994, 1996; Biliaderis and others 1999). Moreover, the OP of the films increases with increasing temperature (Arvanitoyannis and others 1994, 1996; Biliaderis and others 1999) and relative vapor pressure (RVP) (Krochta 2002). For example, gum arabic has been reported to be an efficient encapsulation matrix for limonene in freeze drying (Kaushik and Roos 2007), 577

578

PART 2: Poster Presentations

and it has been suggested that the gelatin–gum arabic–sucrose (1 : 1 : 1) mixture could be used for limonene encapsulation in freeze drying (Kaushik and Roos 2007). Extrusion is a cost-effective process widely used in the food industry to produce a wide variety of products. In the extrusion process, materials (e.g., protein, carbohydrate, and water) are mixed within a screw barrel and forced through the die in which the final shape of the extrudate (e.g., a sheet [thickness, >0.25 mm] or a film [thickness, <0.25 mm]) is controlled. The present study investigated mechanical and OP properties of protein-carbohydrate sheets close to their Tg.

Materials and Methods The protein-carbohydrate sheets were prepared using proteins, sodium (Na)-caseinate (Dairygold, Mitchelstown, Ireland), and gelatin (Italgelatine, Santa Vittoria d’Alba, Italy); and trehalose (dihydrate; Hayashibara, Okayama, Japan), α-lactose (lactose monohydrate, Pharmatose 200M; DMV International, Veghel, Netherlands), gum arabic (Valmar International, Gemenos, France), and sucrose (Danisco, Espoo, Finland). The protein-carbohydrate sheets and their composition are listed in Table 52.1. A twin-screw extruder (Thermo Haake, Hannover, Germany) at a screw speed of 100 rpm was used to produce the protein-carbohydrate sheets. The protein-carbohydrate mixture and water were fed in the extruder at different gates. The temperature profile of barrel and slit-die (dimensions, 150 × 0.5 mm) and total mass flow rate are listed in Table 52.1 and Table 52.2. The water content of the fresh and equilibrated sheets was determined by an oven method (105°C, 4 h). Prior to analyses, the sheet samples were equilibrated at 33% RVP (21°C) for at least 7 days. The sheet samples for OP analyses were, however, equilibrated at 50% RVP (23°C) for at least 7 days. A universal testing machine (model 4465; Instron, Norwood, MA, USA) was used to determine the mechanical properties of the sheets, which had an initial gauge length of 100 mm and cross-head speed of 100 mm min−1. The sheet samples used in the mechanical test were cut parallel to the direction of flow in the extruder.

Table 52.1. The protein-carbohydrate sheet abbreviations, composition (solid base), feed rate of the solid and water, and total feed rate Sheet

Feed rate (g min−1)

Composition Solids

Water

NT

1:1

12.3

3.0

Total 15.3

NL

1:1

12.2

3.7

15.9

NTL

1 : 0.5 : 0.5

12.3

2.9

15.2

GAS

1:1:1

12.4

2.4

14.8

NT, sodium (Na)-caseinate-trehalose; NL, Na-caseinate-lactose; NTL, Na-caseinate-trehalose-lactose; and GAS, gelatin–gum arabic–sucrose.

Physical Properties of Protein-Carbohydrate Sheets

579

Table 52.2. Temperature profile of the controlled zones along extruder barrel Sheet

Temperature profile (°C) 1

2

3

4

5

6

Die 1

Die 2

NT

35

75

90

100

100

100

100

100

NL

35

75

90

95

95

95

95

80

NTL

35

75

90

100

100

100

100

100

GAS

35

75

90

110

110

110

110

90

NT, sodium (Na)-caseinate-trehalose; NL, Na-caseinate-lactose; NTL, Na-caseinate-trehalose-lactose; and GAS, gelatin–gum arabic–sucrose.

The Tg of the sheets was analyzed using a differential scanning calorimeter (Mettler Toledo, Greifensee, Switzerland) at 5°C min−1 from −50° to 100°C. The relaxation behavior of the sheets was analyzed using a dielectric analyzer (TA Instruments, Malmö, Sweden) at a heating rate of 2°C/min from −150° to 100°C at frequencies varying from 0.1 to 1000 Hz. Permittivity (ε′) and loss factor (ε″) data were collected. A peak temperature of tan δ (ε″/ε′) at 1 Hz was correlated with Tg. The relaxation behavior of the sheets was also analyzed using a dynamic mechanical analyzer (Netzsch-Gerätebau, Selb, Germany) at a heating rate of 2°C min−1 from −70° to 100°C at frequencies varying from 1 to 20 Hz. Storage modulus (E′) and loss modulus (E″) data were collected. A peak temperature of E″ at 1 Hz was correlated with Tg. The OP was measured using an OP tester (Ox-Tran Twin; Mocon, Minneapolis, MN, USA) (American Society for Testing and Materials 2001). The sheet sample was exposed to 100% oxygen gas on one side and a carrier gas (98% nitrogen gas and 2% hydrogen gas) on other side. The flow rate of gases was 10 mL min−1. The OP (cm3 μm m−2 day−1 kPa−1) was calculated using Equation 52.1, where OTR is the oxygen transmission rate (cm3 m−2 day−1), p is the oxygen pressure (kPa), and l is thickness of the sheet (μm). OP = ( OTR p ) ∗ l

(52.1)

Results and Discussion Protein-carbohydrate sheets were prepared successfully by using the twin-screw extruder. A lower temperature had to be used in the production of the Na-caseinatelactose sheet because the slit-die became blocked at the higher temperatures used for the other materials. Fresh protein-carbohydrate sheets were elastic and flexible because of their high water content. The sheets turned brittle after equilibration at 33% RVP because of water loss (Figure 52.1). The fresh, equilibrated Na-caseinate-trehalose, Na-caseinate-trehalose-lactose, and gelatin–gum arabic–sucrose sheets were translucent, whereas the Na-caseinate-lactose sheet was opaque. The physical state of sheets equilibrated at 33% RVP was verified by various thermal analyses. The Tg values of the sheets taken from differential scanning calorimetry (DSC), dielectric analysis (DEA), and dynamic mechanical analysis (DMA) data were close to room temperature (Figure 52.1). This was in accordance with the

580

PART 2: Poster Presentations

(a)

(b) Fresh

Equilibrated

20 Tg (°C)

Water content (g/100g of solids)

25

15 10 5 0 NT

NL

NTL

GAS

40 35 30 25 20 15 10 5 0

DSC

NT

NL

DEA

DMA

NTL

GAS

Figure 52.1. Water content (g/100 g of solids) of fresh sheets and equilibrated sheets (33% relative vapor pressure at 21°C) (a). The glass transition temperature (Tg) of the equilibrated sheets determined by differential scanning calorimetry (DSC), dielectric analysis (DEA, 1 Hz), and dynamic mechanical analysis (DMA, 1 Hz) (b). NT, sodium (Na)-caseinate-trehalose; NL, Na-caseinate-lactose; NTL, Na-caseinate-trehaloselactose; and GAS, gelatin–gum arabic–sucrose.

Figure 52.2. Arrhenius plot of α relaxation and β relaxation (solid and open symbols, respectively) of the dielectric spectra (tan δ) for the protein-carbohydrate sheets. NT, sodium (Na)-caseinate-trehalose; NL, Na-caseinate-lactose; NTL, Na-caseinatetrehalose-lactose; GAS, gelatin–gum arabic–sucrose; and f, frequency of dielectric measurement.

results reported for freeze-dried and spray-dried protein-carbohydrate mixtures equilibrated at 33% RVP (Haque and Roos 2006; Singh and Roos 2006). In the present study, DMA gave the highest Tg values of all techniques used. The Tg values determined by various techniques differed because the measurement frequencies used in the DEA and DMA affect peak temperatures of dielectric and mechanical spectra. The α relaxation and β relaxation determined by DEA were found to be linearly dependent on frequency when modeled using the Arrhenius equation (Figure 52.2). Similarly, the α relaxation determined by DMA was found to be linear. The highest values of Young’s modulus and tensile strength were observed for the gelatin–gum arabic–sucrose sheet as compared to the other sheets (Figure 52.3).

Physical Properties of Protein-Carbohydrate Sheets

581

Figure 52.3. Mechanical and oxygen permeability (OP) properties determined for the protein-carbohydrate sheets: Young’s modulus (a), tensile strength (b), elongation at break (c), and oxygen permeability (d). NT, sodium (Na)-caseinate-trehalose; NL, Na-caseinate-lactose; NTL, Na-caseinate-trehalose-lactose; and GAS, gelatin–gum arabic–sucrose.

Moreover, in the mechanical tests, very low elongation at break values of the sheet samples was observed (Figure 52.3). This indicated that the sheets were in a glassy state, which was also confirmed by various thermal analyses (Figure 52.2). The brittle plastic samples usually lengthen 1%–2%, and the stress used to elongate the samples increases linearly with strain until the samples break (Sperling 1992). In the mechanical testing of brittle or glassy biopolymer films, elongation values at break have been reported to vary from 3% to 9% simultaneously with a high Young’s modulus and high tensile strength (Biliaderis and others 1999). In any event, slightly increased elongation at break values has been reported for unplasticized biopolymer films in the glassy state (Chang and others 2000; Lazaridou and Biliaderis 2002; Lazaridou and others 2003). In those films, water content varied from 5% to 15%, and they still remained in the glassy state in which brittle-to-ductile transition had occurred (Chang and others 2000; Lazaridou and Biliaderis 2002; Lazaridou and others 2003). Similar brittle-toductile transition induced by water has been reported for gelatinized starch in the glassy state (Nicholls and others 1995). The OP of the gelatin–gum arabic–sucrose sheet was the lowest among all of the sheets (Figure 52.3). The gelatin–gum arabic–sucrose sheet had to be denser than the other sheets, resulting in higher tensile properties and lower OP properties. The OP of the sheets in the present study was found to correlate with the OP values reported for glycerol plasticized protein films based on whey protein (McHugh and Krochta 1994) and gluten (Hochstetter and others 2006).

582

PART 2: Poster Presentations

Conclusions Brittle protein carbohydrate sheets produced by a screw extruder were obtained after storage at 33% RVP. The mechanical properties of the sheets were affected by their physical state. The Tg of the protein-carbohydrate sheets was verified to be near ambient temperature. Overall, the gelatin–gum arabic–sucrose sheet provided the best protection against oxygen although the Na-caseinate-lactose and Na-caseinatetrehalose-lactose sheets provided almost equal protection. Knowledge of the physical state of encapsulating matrices is needed when they are used to protect food ingredients against, for example, water and oxygen.

Acknowledgment Financial support from Higher Education Authority (HEA) of Ireland in the Programme for Research in Third-Level Institutions (PRTLI) Cycle 3 is greatly appreciated.

References American Society for Testing and Materials (ASTM). 2001. ASTM D 3986: standard test method for oxygen gas transmission rate through plastic film and sheeting using a coulometric sensor. West Conshohocken, PA: ASTM International. Arvanitoyannis I, Kalichevsky M, Blanshard JMV, Psomiadou E. 1994. Study of diffusion and permeation of gases in undrawn and uniaxially drawn films made from potato and rice starch conditioned at different relative humidities. Carbohydr Polym 24:1–15. Arvanitoyannis I, Psomiadou E, Nakayama A. 1996. Edible films made from sodium caseinate, starches, sugars or glycerol. Part 1. Carbohydr Polym 31:179–92. Biliaderis CG, Lazaridou A, Arvanitoyannis I. 1999. Glass transition and physical properties of polyol– plasticized pullulan–starch blends at low moisture. Carbohydr Polym 40:29–47. Chang YP, Cheah PB, Seow CC. 2000. Plasticizing-antiplasticizing effects of water on physical properties of tapioca starch films in the glassy state. J Food Sci 65:445–51. Haque K, Roos YH. 2006. Differences in the physical state and thermal behavior of spray-dried and freezedried lactose and lactose/protein mixtures. Innov Food Sci Emerg Technol 7:62–73. Hochstetter A, Talja RA, Helén HJ, Hyvönen L, Jouppila K. 2006. Properties of gluten-based sheet produced by twin-screw extruder. LWT Food Sci Technol 39:893–901. Kaushik V, Roos YH. 2007. Limonene encapsulation in freeze-drying of gum arabic–sucrose–gelatin systems. LWT Food Sci Technol 40:1381–91. Krochta JM. 2002. Proteins as raw materials for films and coatings: definitions, current status, and opportunities. In: Gennadios A, editor. Protein-based films and coatings. Boca Raton, FL: CRC. p 1–41. Lazaridou A, Biliaderis CG. 2002. Thermophysical properties of chitosan, chitosan-starch and chitosanpullulan films near the glass transition. Carbohydr Polym 48:179–90. Lazaridou A, Biliaderis CG, Kontogiorgos V. 2003. Molecular weight effect on solution rheology of pullulan and mechanical properties of its films. Carbohydr Polym 52:151–66. McHugh TH, Krochta JM. 1994. Sorbitol- vs glycerol-plasticized whey protein edible films: integrated oxygen permeability and tensile property evaluation. J Agric Food Chem 42:841–5. Nicholls RJ, Appelqvist IAM, Davies AP, Ingman SJ, Lillford PJ. 1995. Glass transitions and the fracture behaviour of gluten and starches within the glassy state. J Cereal Sci 21:25–36. Singh KJ, Roos YH. 2006. State transitions and freeze concentration in trehalose-protein-cornstarch mixtures. LWT Food Sci Technol 39:930–8. Sperling LH. 1992. Introduction to physical polymer science. 2nd ed. New York: John Wiley & Sons. p 506–9.

53 Thermal Transitions, Mechanical Properties, and Molecular Mobility in Cornflakes as Affected by Water Content A. Farroni, S. B. Matiacevich, S. Guerrero, S. Alzamora, and M. P. Buera Abstract Common cornflakes (CCF) and sugar-frosted cornflakes (SCF) were ground and stored at relative humidities (RHs) of 11%–80% at 25°C to achieve a water content (WC) between 5% and 20% dry basis (db). Time-resolved proton nuclear magnetic resonance (TR-1H NMR) measuring the spin-spin transverse relaxation time (T2) was used to determine the mobility of water and solids. T2 analysis following a single 90° pulse (free-induction decay [FID]) was used to evaluate molecular mobility coupled to the solid matrix in the temperature range of 4°–90°C. Water mobility was evaluated using a spin-echo (90°-τ-180°) sequence. The FID analysis in both types of cornflakes (CF) showed a single T2. At a WC close to the Guggenheim-Anderson-de Boer (GAB) monolayer value (m0 = 5% db), the spin-echo analysis showed a short-relaxing T2 component (T2S) at 25–30 μs that was attributed to the mobility of protons from solids and the most closely associated water molecules. T2S increased at temperatures above the glass transition temperature (Tg), indicating higher proton mobility due to plasticization of solids by water. A long-relaxing T2 component (T2L) was observed at RHs above m0 and increased as WC increased in the range of 0.4–2.5 ms, manifesting the higher mobility of water protons. At very low WC, water acts as an antiplasticizer, but beyond a critical level, the plasticizing effect became evident in mechanical properties. The temperature dependence of water and solid molecular mobility can be analyzed by TR-1H NMR.

Introduction The shelf-life stability and textural properties of food products are closely related to their mechanical and thermal properties. The physical state, water mobility, and interactions of water and solids affect the storage stability, textural properties, and functional properties of food. Adsorption of water by biopolymers reduces the possibility of brittle fracture accompanied by dissipation of elastic energy and relaxation of stress (Poliszko 1995). Water molecules in biological materials can be classified into several fractions based on their molecular mobility or the extent of their association with or binding to macromolecules within the material (Fullenton and Cameron 1988). Spinspin transverse relaxation time (T2) can be measured by proton nuclear magnetic resonance (TR-1H NMR). This technique provides useful information on the effect of water-solid interactions on molecular mobility and water populations. Low-moisture 583

584

PART 2: Poster Presentations

products remain brittle in the glassy state, which can be studied by differential scanning calorimetry (DSC) (Slade and Levine 1995). As part of a comprehensive study on the impact of corn types on CF quality, our work analyzed the thermal and mechanical properties and molecular mobility of cornflakes.

Materials and Methods Common CF (CCF) and sugar-frosted CF (SCF) were ground and equilibrated over saturated salt solutions at relative humidities (RHs) in the range of 11%–80% for 20 days at 25°C to achieve different water contents (WCs). After equilibration, the water activity (aw) of systems was determined by dew point by using an Aqualab (Decagon Devices, Pullman, WA, USA), and WC was determined gravimetrically based on the difference in mass before and after samples had been dried in an oven at 100°C for 48 h. The water-sorption isotherms were obtained and fitted to the Guggenheim-Anderson-de Boer (GAB) equation, which can be considered a good model over a wide range of aw (Lomauro and others 1985) because it enables the calculation of WC at the monolayer (m0), which represents the saturation with 1 water molecule for each of the more active polar sites of the adsorbate. Approximately 15 mg of equilibrated systems were weighed on a Mettler Toledo balance, sealed in aluminum pans (40-μL capacity), and loaded into a differential scanning calorimeter (model 822; Mettler Toledo, Columbus, OH, USA). All experiments were performed in duplicate following the same protocol. The instrument was calibrated using indium, zinc, and lead. An empty pan was used as a reference. The DSC thermograms were obtained from −50° to 150°C at a heating rate of 10°C/ min. The glass transition temperature (Tg) was determined as the onset temperature of discontinuities in the curves of heat flow vs temperature (indicating a change in specific heat). All thermograms were analyzed using STARe software version 6.1 (Mettler Toledo thermal analysis system). Mechanical properties were measured by a compression test using a universal testing machine (Instron, Canton, MA, USA). Samples were compressed with a 30-mm-diameter piston at 50 mm/min up to a 30% deformation. The maximum force of compression was recorded, and replicates were averaged. Compression tests are those most commonly used because of their similarities with the mastication process (Roudaut and others 2002). Time-resolved proton nuclear magnetic resonance (TR-1H NMR) measuring the spin-spin transverse relaxation times (T2) was used to determine the mobility of water and solids. A Bruker Minispec mq20 spectrometer (TR-1H NMR) (Bruker AXS, Madison, WI, USA), operating at 20 MHz and thermostated at 40°C, was used. Freeinduction decay (FID) analysis following a single 90° pulse was used to evaluate molecular mobility coupled to the solid matrix in the range of 4°–90°C (Lin and others 2006). Relaxation times obtained under these conditions (without a 180° refocus pulse) are apparent relaxation time constants ( T2*). When fast relaxation times in samples of solids were analyzed, the T2* was very close to the T2 (Fullenton and Cameron 1988). The T2 associated with slow-relaxing protons was studied by using the Hahn spin-echo

Thermal Transitions, Mechanical Properties, and Molecular Mobility in Cornflakes

585

(90°-τ-180°) sequence to evaluate water populations that interacted less strongly with solids.

Results and Discussion The water-sorption isotherm of CCF showed higher WC than did SCF at the same aw (data not shown). This result was attributed to the presence of crystalline sucrose with very little adsorbed water in the RH range studied. However, the WC at the GAB m0 was close to 5% dry basis (db) for both the CCF and the SCF systems. Figure 53.1 shows the DSC thermograms obtained for CCF and SCF at WCs of ∼7% and ∼5% db, respectively, as an example. The different thermal transitions are indicated by arrows. The DSC analysis showed two endothermic events that were independent on the RH. A reversible transition centered at −25°C and the other at close to 50°C. The thermal transition at −20°C was not significantly different between samples with regard to the CF types or the RH. This event was related to lipidic components (Matiacevich and others 2006). The endothermic event observed at close to 50°C disappeared in a rescan but reappeared after the samples were aged for 2 days at 25°C. This event was previously described as a sub-Tg endotherm in starch-containing systems (Thiewes and Steeneken 1997). As expected, the Tg of CCF decreased as WC increased, and the Tg was slightly lower than that for gelatinized cornstarch (Zhon and

Tg

lipids h heat flux (← endo)

Heat flux (← endo)

sub-Tg endotherm

0.01 mW

sugar caramelization

–40 –35 –30 –25 –20 –15 –10

0.1mW

CCF wc: 7.5% SCF wc: 5.2%

–70

–45

–20

5

30

55

80

105 130 155

Temperature (ºC) Figure 53.1. Differential scanning calorimetry (DSC) thermograms of common cornflakes (CCF) and sugar-frosted cornflakes (SCF) at 7.5% and 5.2% (dry basis [db]) water content (wc), respectively. The lipid-related transition and sub–glass transition temperature (Tg) endotherm (endo) are shown by the arrows at −25°C and at close to 50°C, respectively. The inset shows a detail of the transition at −25°C. The onset of the Tg is indicated by the arrow in the CCF thermogram. The endothermic transition (peak at 140°C) corresponding to the melting and caramelization of sugar crystals is shown in the SCF samples.

586

PART 2: Poster Presentations

(a)

(b)

0.013

wc: 4.1% wc: 6.4% wc: 11.0% wc: 16.4%

0.012

0.011

T2* (ms)

T2* (ms)

wc: 2.75% wc: 4.3% wc: 10.9% wc: 14.6%

0.012

0.011 0.010 0.009

0.010 0.009 0.008

0.008 0.007

0.013

0.007 0

20

40

60

T (°C)

80

100

0

10 20 30 40 50 60 70 80 90 100

T (°C)

Figure 53.2. Relationship between the relaxation time constant (T2* ) determined by free-induction decay (FID) analysis and temperature for (a) common cornflakes and (b) sugar-frosted cornflakes. A change in slope detected by time-resolved proton nuclear magnetic resonance (TR-1H NMR) was close to the glass transition temperature (Tg) determined by differential scanning calorimetry (DSC) with the same water content (wc).

Sun 2005). In SCF samples, a thermal transition that began at 70°C was attributed to sugar caramelization, and, in these samples, the Tg was not detectable because it was overlapped by other thermal events. Freezable water was not detected in any samples in the RH range studied. The study of TR-1H NMR T2 as a function of WC and temperature provides information about the mobility of water protons. A single T2 was obtained by FID analysis. Figure 53.2 shows the values at different WCs along the temperature scale in the CCF (a) and the SCF (b). The T2 increased as WC and temperature increased, reflecting the higher mobility of polymers (Ruan and others 1998) and their associated water. Beyond 11% db, a change in slope was observed in the T2 vs temperature plots for CCF samples at a value close to the Tg determined by DSC at the same WC. At lower WC, this change in slope was not observed because the Tg of the systems was out of the temperature range that could be analyzed in the experimental conditions because of water evaporation. Nevertheless, in more complex food systems such as the SCF samples (Figure 53.2b), a change in slope was less marked than in CCF but still could be detected, although the Tg values could not be compared with DSC measurements because the Tg of the thermograms could not be determined. The T2 evaluated using a spin-echo sequence was used to analyze water mobility as a function of WC. The two sets of values for relaxation constants obtained were related to water populations displaying different degrees of interaction with solids. One of them, with stronger interactions and short relaxation times, was designated T2S. The other, with longer relaxation times more characteristic of multilayer or bulk water, was designated T2L. In both systems (Figure 53.3), at moisture contents of 5% db, close to the GAB m0, the spin-echo analysis showed the rapidly relaxing T2 com-

Thermal Transitions, Mechanical Properties, and Molecular Mobility in Cornflakes

(b) 1.50 1.25 1.00 0.75 0.50 0.25

T2 (ms)

T2 (ms)

(a)

587

0.03

T2L T2S 2

4

6

8

10 12 14 16 18 20 22

Water content (% db)

1.50 1.25 1.00 0.75 0.50 0.25 0.03

T2L T2S 2

4

6

8

10 12 14 16 18 20

Water content (% db)

Figure 53.3. Hahn spin-echo transverse relaxation time (T2) values vs water content for (a) common cornflakes and (b) sugar-frosted cornflakes (error bars represent standard deviation). The cluster for the short relaxation time was defined as T2S (triangles). The longer relaxation times (characteristic of more mobile protons, such as those of bulk water), were assigned T2L (squares). db, dry basis.

ponent (T2S) at 25–30 μs. T2S increased at WC that corresponded to samples of Tg close to the measurement temperature (40°C). The increased mobility indicated plasticization by water. A slow relaxation component (T2L) was observed at WCs above the m0, which increased to 0.4–2.5 ms as WC increased. The T2L was attributed to water protons with higher mobility that are probably located in the multilayer region. Both relaxation times (T2S and T2L) increased slightly up to a WC of ∼12% db, reflecting the overall increase in mobility of the system. In a parallel study, mechanical properties as a response to compression were studied for different WCs. At low WC, the deformation force increased as WC increased, indicating an antiplasticizing effect of water (Gondek and Lewicki 2006). When the WC reached a critical level, the force of compression decreased. The mobility of small molecules, like water, would be affected mainly by local viscosity (at the molecular level) rather than by the global viscosity that would be related to the glass transition (Chinachoti 1997). As WC increased, the mobility of water molecules increased and plasticizing effects became evident in the mechanical properties. When samples were equilibrated at temperatures close to the glass transition, thermal and mechanical properties showed coupled behavior (Figure 53.4). Although CF systems are complex, the thermal transitions and molecular mobility of their main components and their changes produced by increasing WC could be analyzed by DSC and NMR. However, for SCF, DSC studies did not reflect clear changes such as those observed in CCF, and probably TR-1H NMR represented a better alternative for determining the Tg in these samples. Thus, TR-1H NMR complemented DSC data in determining the temperature dependence of mobility for water and solids in assessing the quality of laminated corn products. The results of the present work indicate that while compression force showed a maximum as a function of WC, Tg decreased

588

PART 2: Poster Presentations

25

160 F

140

20 120 100

15

80 10

60

Temp (°C)

Force (kgF )

Tg

40

5

20 0 0

10 20 Water content (% db)

0 30

Figure 53.4. The maximum force of compression (kgF) and onset of the glass transition temperature (Tg) at different water contents (wc) for common cornflakes (error bars represent ± standard deviation). db, dry basis; F, for force plot; and T, temperature plot.

progressively as WC and water mobility increased (characterized by proton transverse relaxation times).

Acknowledgments The authors acknowledge the financial support of the Agencia Nacional de Promoción Científica y Tecnológica (PICT 20545) (Argentine National Agency of Scientific and Technological Promotion), the Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET, PIP 5799) (Argentine National Scientific and Technical Research Council), the University of Buenos Aires (X226), and the Instituto Nacional de Tecnología Agropecuaria (INTA) (National Institute of Agricultural Technology). We also thank Dr. Myriam Casanello for helpful contributions to this analysis.

References Chinachoti P. 1997. Water migration and food storage stability. In: Taub IA, Singh RP, editors. Food storage stability. Boca Raton, FL: CRC. p 245–67. Fullenton GD, Cameron IL. 1988. Relaxation of biological tissues. In: Wehrli S, Kneeland, editors. Biomedical magnetic resonance imaging: principles, methodology, and application. New York: VCH. p 115–55. Gondek E, Lewicki P. 2006. Antiplasticization of cereal-based products by water. Part II: Breakfast cereals. J Food Eng 77:644–52. Lin X, Ruan R, Chen P, Chung M, Ye X, Yang T, Doona C, Wagner T. 2006. NMR state diagram concept. J Food Sci 71:R136–45. Lomauro CJ, Bakshi AS, Labuza TP. 1985. Evaluation of food sorption isotherms equations. I. Fruit, vegetable and meat products. LWT Food Sci Technol 18:111–7.

Thermal Transitions, Mechanical Properties, and Molecular Mobility in Cornflakes

589

Matiacevich SB, Castellión ML, Maldonado SB, Buera MP. 2006. Thermal transitions of quinoa embryos and seeds as affected by water content. In: Buera M del P, Welti-Chanes J, Lillford PJ, Corti HR, editors. Water properties of food, pharmaceutical, and biological materials. Buenos Aires: CRC. p 565–70. Poliszko S, Klimek D, Molinski W. 1995. Acoustic emission activity of rehydrated corn extrudates. In: Lewicki W, editor. Properties of water in foods. Warsaw: Warsaw Agricultural University Press. p 25–30. Roudaut G, Dacremont C, Vallès Pàmies B, Colas B, Le Meste M. 2002. Crispness: a critical review on sensory and material science approaches. Trends Food Sci Technol 13:217–27. Ruan R, Long Z, Song A, Chen P. 1998. Determination of glass transition temperature of food polymers using low field NMR. LWT Food Sci Technol 31:516–21. Slade L, Levine H. 1995. Water and the glass transition: dependence of the glass transition and chemical structure—special implication for flour functionality on cookie baking. J Food Eng 22:143–88. Thiewes HJ, Steeneken PAM. 1997. The glass transition and the sub-Tg endotherm of amorphous and native potato starch at low moisture content. Carbohydr Polym 32:123–30. Zhon Z, Sun S. 2005. Thermal characterization and phase behavior of cornstarch studied by differential scanning calorimetry. J Food Eng 69:453–9.

54 Texture of Glassy Tapioca-Flour–Based Baked Products as a Function of Moisture Content R. Kulchan, P. Suppakul, and W. Boonsupthip

Abstract Texture change of tapioca-flour–based baked products as impacted by moisture content (MC) was investigated along with the products’ glass transition. The baked products were initially in a glassy state with water activity (aw), MC, and midpoint glass transition temperature (Tg,m) of 0.38, 3.9% and 148°C, respectively. While aw remained lower than 0.54 (6% MC), the products’ crispness was satisfactorily preserved (score ≥ 5). A mechanical test showed a concave dependence of the products’ hardness and work (area under the hardness peak) with increasing MC. When the products’ hardness or work reached a maximum and began to reduce, at aw ≈ 0.54 (MC ≈ 6% and Tg,m = 132.8°C), the products’ texture became slightly soft (score = 5, unacceptable texture). This suggests that maximum hardness and work could be used to identify critical aw, which results in an unacceptable loss of crispness. The sorption isotherm indicates the amount of water in the monolayer was ∼2.5% MC (aw ≈ 0.1). This amount of water was much lower than that at the critical aw for acceptable crispness. A plasticizing effect of water on the products’ glass transition in the aw range of 0–0.86 resulted in a decrease in Tg,m from 173.7 to −10.1°C. Such texture deterioration occurred even though the products were still in a glassy state.

Introduction Tapioca (Manihot esculenta Crantz) flour is used as a key ingredient of several dry, crisp products, such as potato chips and puffed curls. In addition, Asians and Latin Americans are interested in its use as a partial substitute for wheat flour (Hudson and Ogunsua 1976; Holt and others 1992; Pachico 1997; Afolabi and others 2001; Lopez and others 2004; Mohamed and others 2006). To consumers, the high crispness of such products indicates not only good quality but also freshness (Szczesniak and Kleyn 1963, 1971; Rohm 1990). Unfortunately, little information has been reported on the creation and preservation of crispness in tapioca-flour–based products (Van den Berg and others 1975; Chang and others 2000) and it is especially rare for multicomponent systems. In the texture study, crispness was perceived as a combination of the sound generated and the fracture of the product as it was bitten completely through with the back molars (Duizer and others 1998). Different instrumental and sensory approaches have been applied to study this quality attribute and a large amount of 591

592

PART 2: Poster Presentations

experimental data has been generated (Roudaut and others 2002). Unfortunately, no sound conclusion can be drawn about the relationship between the instrumental and sensory results because many definitions of crisp have been applied (Roudaut and others 2002) and the results of only a few studies of sensory data have been reported (Mohamed and others 1993; Van Hecke and others 1995; Ferriola and Stone 1998; Thakur and Saxena 2000; Roudaut and others 2002). The crispness of dry, crisp products is controlled by product composition and structure (Roudaut and others 2002). The conditions under which the process is performed also affects the final moisture content (MC), which governs the crispness of the final product (Roudaut and others 2002). During storage, water adsorption from the atmosphere or through mass diffusion from neighboring components can cause a loss of crispness (Nicholls and others 1995). Component properties in terms of hydration and movement capability are important. Water is one of the components most involved, acting as an excellent plasticizer (Van den Berg and others 1975; Fontanet and others 1997) reducing crispness and glass transition temperature. Large and entangled molecules (e.g., starch, fiber, and protein) have high glass transition temperatures, whereas those of small molecules (e.g., sugar, salt, and mineral) are low (Slade and Levine 1995). The effects of each component on crispness and glass transition are complex, especially for multicomponent or multiphase products because of the interactions among the components. Our work investigated the mechanical and sensory crispness of a tapioca-flour– based food system as a function of glass transition and MC. This system contained mainly tapioca flour. Minor components included coconut milk (fat source), sucrose, and water.

Materials and Methods Sample Preparation Glassy tapioca-flour–based baked samples were prepared by using tapioca flour from Choheng (Nakon Prathom, Thailand), and coconut milk, egg, and sucrose from retailers. First, a mixture of 18.4% coconut milk and 23% sucrose was heated at 90°C until 40% of the sample weight was lost. This mixture was added to 1.15% egg yolk and 55.2% flour and kneaded into dough with a domestic mixer (model KM 410; Kenwood Electronics, Walford, UK) at a minimum speed. The dough was stored in a tightly sealed container at room temperature overnight. Then, after the addition of a further 2.3% water, the dough was kneaded to obtain a homogeneous distribution before being divided into small balls (∼1 cm in diameter) that were placed on a greased pan and baked at 300°F (149°C) for 20 min. After baking, they were porous and had expanded to ∼1.5-cm diameter. The baked products were left to cool and kept in a tightly sealed container for further use. Proximate Analysis The samples were analyzed for moisture, protein, carbohydrate, starch, fat, ash, and fiber by AOAC methods (Lane 1998). All determinations were carried out in triplicate.

Texture of Glassy Tapioca-Flour–Based Baked Products

593

Sorption Isotherm A standard gravimetric methodology was used in determining the adsorption isotherm. The baked products were crushed and completely dried in a vacuum oven at 70°C and 76 mmHg for 48 h and then in a desiccator over phosphorus pentoxide for 2 weeks. The dried samples (in triplicate) were equilibrated over saturated salt solutions inside desiccators at 30°C for 4 weeks. The salts were lithium chloride, potassium acetate, magnesium chloride, potassium carbonate, magnesium nitrate, potassium iodide, sodium chloride, potassium chloride, and potassium sulfate, producing saturated solutions of known relative humidity, 11.3%, 21.6%, 32.4%, 43.2%, 51.4%, 67.9%, 75.1%, 83.6%, and 92.3%, respectively (Greenspan 1977). MC was then measured in quadruplicate by drying weighed samples of the product in an oven at 105°C for 3 h (Lane 1998), and then reweighing was expressed as g H2O/100 g dry sample for the dry-weight basis or g H2O/100 g wet sample for the wet-weight basis. Water activity (aw) was determined by using a Testo 650 to measure modular humidity (Testo, Lenzkirch, Germany). Dynamic Differential Scanning Calorimetry (DDSC) The glass transition temperature of the crushed samples was determined by using a Netzsch DSC 204 F1 (LMS Instruments, Göttingen, Germany). Indium was used for calibrating the temperature, heat flow, and furnace. Empty aluminum pans were used as reference baselines. About 11–15 mg of each sample was weighed and hermetically sealed in an aluminum pan used for differential scanning calorimetry (DSC). The differential scanning calorimeter was operated in a dynamic mode with a scan rate of 5°C/min over temperature ranging from −50° to 230°C, at an amplitude of 1°C, and for a 30-s equilibration period. A minimum of three samples were scanned to ensure reliability of the measurement. Texture Assessment: Crispness, Hardness, and Work Sensory Evaluation Twelve panelists were trained to ensure the same perception of crispness (as defined by Duizer and others [1998]). Crispness was rated on a nine-point category scale (from 1 for not crisp or soggy to 9 for very crisp). The panelists evaluated the samples randomly three times over three sessions. A three-way analysis of variance (products, panelists, and replications [with block factor]) ensured that the products and the panelists did not interact. This confirmed the validity of the crispness evaluation by the trained panel. Mechanical Measurement Mechanical measurement was performed with a texture analyzer (TM LRX S/N 10313; Lloyd Instruments, Fareham, UK). A sample was placed atop the lower hollow cylinder. A flat cylindrical plunger (4.77-mm diameter) was set to a cross-head speed of 10 cm/min and a load of 50 kgF. Force and deformation data were recorded. Each sample was measure in 15–20 replicates. Hardness (kgF) was defined as the maximum force required to break the product, and work (kgF · mm) as the integral area under the force-and-deformation curve (Li and others 1998).

594

PART 2: Poster Presentations

Results and Discussion Food Properties The chemical composition of the tapioca-flour–based baked samples was 84.78% ± 0.09% carbohydrate (∼60.08% starch), 0.63% ± 0.07% protein, 10.24% ± 0.09% fat, 3.74% ± 0.05% water, 0.32% ± 0.01% ash, and 0.29% ± 0.03% fiber (wet basis). These dry, crisp samples were high in starch and fat. After being baked, cooled down, and analyzed for aw, MC, and midpoint glass transition temperature (Tg,m), the samples’ properties were 0.38, 3.9%, and 148°C, respectively. Based on the Tg,m value, these samples were initially in a glassy state. Adsorption Isotherm When the water-sorption capacity of glassy tapioca-flour–based baked samples was examined, it was found that the equilibrium MC (dry basis) increases exponentially with increasing aw, as shown in Figure 54.1. As aw increased from 0.00 to 0.75, equilibrium MC increased only gradually. At higher aw (aw > 0.75), the increase became sharp. According to the Brunauer-Emmett-Teller (BET) classification, the curve takes the shape of a type II isotherm (Brunauer and others 1940). This isotherm form has often been reported for products with high starch content; for example, oatmeal biscuits (Cadden 1988; Durakova and Menkov 2005; Erbas and others 2005). In our study, the samples were found to contain a high starch content (∼60%). 60 Exp. % MC (dry basis)

50 GAB 40 30 20 10 0 0

0.2

0.4

0.6

0.8

1

aw

Figure 54.1. Sorption isotherm: experimental data of glassy tapioca-flour–based baked product (open triangles). aw, water activity; and MC, moisture content. The solid line was the predicted isotherm based on the Guggenheim-Anderson-de Boer (GAB) equation, expressed as X = X0CKAW [( 1− KAW ) (1− KAW + CKAW )]

(54.1)

where X is the equilibrium moisture content (kg moisture/kg dry solid), X0 is the monolayer moisture content (kg moisture/kg dry solid), C is a constant (GAB isotherm model), K is a constant (GAB isotherm model), and aw, is water activity.

Texture of Glassy Tapioca-Flour–Based Baked Products

595

The glassy tapioca-flour–based baked samples have low adsorption capacity (Figure 54.1). This was also reported for many agricultural crops with high fat content (Iglesias and Chirife 1982). It is probably due to the hydrophobic effect of fat molecules that hinder an attachment of water molecules from the atmosphere to the product surface. However, in some cases (e.g., oat flakes with 30.6% fat [McMinn and others 2007]), fat has shown an inverse effect by supporting sorption capacity. The Guggenheim-Anderson-de Boer (GAB) equation (Equation 54.1), one of the most reliable ones for sorption isotherms (Van den Berg and Bruin 1981), was applied to fit the isotherm of the baked samples. Best fit was observed with R2 = 0.9969 (Figure 54.1). The values of the GAB constants X0, C, and K were 2.4598, 49.0599, and 1.0904, respectively. Relationship Among Texture, Water Activity, and Glass Transition Crispness of the baked samples was determined using a sensory approach. The fresh samples were highly crispy, with a score of 7.8, and very moisture sensitive. As the samples adsorbed more water, the crispness acceptability sharply declined linearly with an increase in aw (Figure 54.2). The products’ crispness was acceptable (score ≥ 5) while containing a small amount of water (aw < 0.54 or MC [dry basis] < 6%). These specific values of aw and MC (0.54 and 6%) could be considered

Hardness

Work

MC

Tg

9

250

8

200

7

150

6

100

5

50

4

0

3

–50

2

–100

1

–150

0

Tg (°C), %MC (dry basis)

Crispness (scale), Hardness (kgF), Work (kgF-mm)

Crispness

–200 0

0.2

0.4

0.6

0.8

1

aw

Figure 54.2. Relationships between water activity (aw) and texture properties, glass transition, and moisture content (MC) of the glassy tapioca-flour–based baked product. kgF, maximum force of compression; and Tg, glass transition temperature.

596

PART 2: Poster Presentations

a critical point of crispness loss. This critical aw value corresponded to that of ∼0.5 reported for other dry, crisp starch-protein-matrix products (Katz and Labuza 1981; Roos and others 1998; Hough and others 2001). The tapioca baked samples also contained protein and starch, which would be anticipated to form a crispy matrix during baking. This matrix was significantly softened by the plasticization effect of water adsorbed above the critical level (Katz and Labuza 1981; Martinez-Navarrete and others 2004). The aw at the critical level (0.54) is much higher than that with monolayer water content (2.5% dry basis or aw ≈ 0.12). In addition, at the critical aw, the products were still in a glassy state (Tg,m ≈ 132.8°C). The products’ texture was also examined by a mechanical approach. The results showed a concave dependence of the products’ hardness and work with increasing aw (Figure 54.2). Each increased to a maximum at aw ≈ 0.54. As already pointed out, this value of aw is at the critical point of crispness loss (score = 5). This suggests that the maximum of mechanical hardness and work could be used to identify the critical aw of crispness loss as sensory data. The same conclusion has been reported for extruded flat bread (Roudaut and others 1998). The baked sample’s transition to a glassy state was investigated in relation to product aw (Figure 54.2). The Tg,m decreased gradually from 173.7° to 102.7°C with an increase in aw from 0 to 0.74. Once aw was higher than 0.74, water strongly plasticized the sample, leading the Tg,m to plunge sharply with increasing aw. For example, an increase in aw from 0.74 to 0.85 produced a dramatic decrease in Tg,m from 102.7 to −10.1°C. The Tg,m values of these products were governed largely by two main components: starch (∼60%) and sucrose (∼23%), besides water (as a strong plasticizer). Starch is a large molecule, which strongly contributes to the high glass transition temperature of the product. An addition of sucrose, whose molecular weight is much smaller than starch, would retard the glass transition (Roos and Karel 1991; Bhandari and others 1997). However, sucrose has an adverse effect, which is reduction of the product aw. This reduction would facilitate the glass transition. Therefore, the Tg,m of these products was still high. An increase in water content in the products resulted in a decrease in the glass transition temperature and crispness over the range of aw, but for hardness and work, the decrease occurred only at aw higher than 0.54. Such reductions are caused by molecular mobility, which is facilitated by water molecules (Roudaut and others 2002). However, at lower aw (0.23–0.54), the hardness and work increased sharply with an increase in aw. This is because at such a low aw (such as 0.23, close to the 0.12 of monolayer water), the products had only a small number of water molecules and an addition of water might only fill the free volume (at a microscopic level), leading to an increase in products’ density (Vrentas and others 1988; Benczedi 1999; Seow and others 1999) and interactions among water and other component molecules (Roudaut and others 2002). These phenomena cause steep increases in hardness and work (a higher puncture force is required). However, such an increase in water had only a slight impact on glass transition temperature, corresponding to a slight increase in molecular mobility.

Texture of Glassy Tapioca-Flour–Based Baked Products

597

References Afolabi WAO, Oguntona CRB, Fakunmoju BB. 2001. Acceptability and chemical composition of bread from beniseed composition flour. Nutr Food Sci 31:310–4. Benczedi D. 1999. Estimation of the free volume of starch water barriers. Trends Food Sci Technol 10:21–4. Bhandari BR, Datta H, Crooks R, Rigby S, Howes T. 1997. A semi-empirical approach to optimize the quantity of drying aids required to spray dry sugar rich foods. Drying Technol 15:2509–25. Brunauer S, Deming LS, Deming WE, Roller E. 1940. On the theory of Van der Waals adsorption of gases. J Am Chem Soc 62:1723–32. Cadden AM. 1988. Moisture sorption of several food fibres. J Food Sci 53:1150–5. Chang YP, Cilean PB, Seow CC. 2000. Variations in flexural and compressive fracture behavior of a brittle cellular food (dried bread) in response to moisture sorption. J Texture Stud 31:525–40. Duizer LM, Campanella OH, Barnes GRG. 1998. Sensory, instrumental and acoustic characteristics of extruded snack food products. J Texture Stud 29:397–411. Durakova AG, Menkov ND. 2005. Moisture sorption characteristics of chickpea flour. J Food Eng 68:535–9. Erbas M, Ertugay MF, Certel M. 2005. Moisture adsorption behaviour of semolina and farina. J Food Eng 69:191–8. Ferriola D, Stone M. 1998. Sweetener effects on flaked millet breakfast cereals. J Food Sci 63:726–9. Fontanet I, Davidou S, Dacremont C, Lemeste M. 1997. Effect of water on the mechanical behaviour of extruded flat bread. J Cereal Sci 25:303–11. Greenspan L. 1977. Humidity fixed points of binary saturated aqueous solutions. J Res Natl Bur Stand [A] 81:89–96. Holt SD, Mcwatters KH, Anna VA. 1992. Validation of predicted baking performance of muffins containing mixtures of wheat, cowpea, peanut, sorghum, and cassava flours. J Food Sci 57:470–4. Hough B, Buera MP, Chirife J, Moro O. 2001. Sensory texture of commercial biscuits as a function of water activity. J Texture Stud 32:57–74. Hudson BJ, Ogunsua AO. 1976. The effects of fibre, starch damage and surfactants on the baking quality of wheat/cassava composite flours. Int J Food Sci Technol 11:129–36. Iglesias HA, Chirife J. 1982. Handbook of food isotherms: water sorption parameters for food and food components. New York: Academic. Katz EE, Labuza TP. 1981. Effect of water activity on the sensory crispness and mechanical deformation of snack products. J Food Sci 46:403–9. Lane RH. 1998. Cereal foods. In: Official methods analysis of the Association of Official Analytical Chemists. 16th ed. Gaithersburg, MD: AOAC International. p 1–37. Li Y, Kloeppel MK, Hsieh F. 1998. Texture of glassy corn cakes as a function of moisture content. J Food Sci 63:1–4. Lopez ACB, Pereira AJG, Junqueira RG. 2004. Flour mixture of rice flour, corn and cassava starch in the production of gluten-free white bread. Braz Arch Biol Technol 47:63–70. Martinez-Navarrete N, Moragu G, Talens P, Chiralt A. 2004. Water sorption and the plasticization effect in wafers. Int J Food Sci Technol 39:555–62. McMinn WAM, McKee DJ, Magee TRA. 2007. Moisture adsorption behavior of oatmeal biscuit and oat flakes. J Food Eng 79:481–93. Mohamed S, Abdullah N, Muthu MK. 2006. Physical properties of keropok (fried crisps) in relation to the amylopectin content of the starch flours. J Sci Food Agric 49:369–77. Mohamed S, Add Hamid N, Abdul Hamid M. 1993. Food components affecting the oil absorption and crispness of fried batter. J Sci Food Agric 78:39–45. Nicholls RJ, Appelqvist IAM, Davies AP, Ingman SJ, Lillford PJ. 1995. Glass transition and the fracture behavior of gluten and starches within the glassy state. J Cereal Sci 21:25–36. Pachico D. 1997. Annual report 1997: linking small farms with growth markets to build sustainable livelihoods in rural areas. Cali, Colombia: International Center for Tropical Agriculture. Rohm H. 1990. Consumer awareness of food texture in Austria. J Texture Stud 21:363–73.

598

PART 2: Poster Presentations

Roos Y, Karel M. 1991. Plasticizing effect of water on thermal behaviour and crystallization of amorphous food models. J Food Sci 56:38–43. Roos Y, Roininen K, Jouppila K, Tuorila H. 1998. Glass transition and water plasticization effects on crispness of a snack food extrudate. Int J Food Properties 1:163–80. Roudaut G, Dacremont C, Le Meste M. 1998. Influence of water on the crispness of cereal based foods acoustic, mechanical, and sensory studies. J Texture Stud 29:199–213. Roudaut G, Dacremont C, Pamies BV, Colas B, Meste ML. 2002. Crispness: a critical review on sensory and material science approaches. Trends Food Sci Technol 13:217–27. Seow C, Cheah PB, Chang YP. 1999. Antiplasticization by water in reduced-moisture food systems. J Food Sci 64:576–81. Slade L, Levine H. 1995. Water and the glass transition: dependence of the glass transition on composition and chemical structure—special implications for flour functionality in cookie baking. J Food Eng 24:431–509. Szczesniak AS, Kleyn DH. 1963. Consumer awareness of texture and other food attributes, II. Food Technol 63:74–7. Szczesniak AS, Kleyn DH. 1971. Consumer awareness of texture and of other food attributes. J Texture Stud 2:196–206. Thakur S, Saxena DC. 2000. Formulation of extruded snack food (gum based cereal-pulse blend): optimization of ingredients levels using response surface analysis. Food Sci Technol 33:354–61. Van den Berg C, Bruin S. 1981. Water activity and its estimation in food systems. In: Rockland LB, Stewart F, editors. Water activity: influence on food quality. New York: Academic. p 147–77. Van den Berg C, Kaper FS, Weldring JAG, Wolters I. 1975. Water binding by potato starch. J Food Technol 10:589–602. Van Hecke E, Allaf K, Bouvier JM. 1995. Texture and structure of crispy-puffed food products. I. Mechanical properties in bending. J Texture Stud 26:11–25. Vrentas JS, Duda JL, Ling HC. 1988. Antiplasticization and volumetric behavior in glassy polymers. Macromolecules 21:1470–5.

55 Effects of Excipients on the Storage Stability of Freeze-Dried Xanthine Oxidase P. Srirangsan, K. Kawai, N. Hamada-Sato, M. Watanabe, and T. Suzuki

Abstract This study aimed to elucidate the effects stabilizers have on the storage stability of freeze-dried xanthine oxidase (XOD). Disaccharides and/or polymers are known to be effective stabilizers, and their stabilizing mechanisms are known as preferential exclusion, water substitution, and glass transition stabilization. XOD and selected stabilizer(s) (sucrose, sucrose with dextran, and sucrose with bovine serum albumin) were freeze-dried in potassium phosphate buffer. The samples were stored at 25°C up to 90 days. After storage, the samples were rehydrated with distilled water, and then XOD activity was investigated spectrophotometrically. All formulations containing stabilizers showed glass transition at a temperature above 25°C. The glass transition temperatures of formulations containing polymers (dextran or bovine serum albumin) were much higher than that of sucrose alone. All added stabilizers greatly protected XOD from activity loss compared to the dried XOD alone. However, the stabilizing effects of sucrose was diminished by adding dextran. Taking water substitution into account, it is suggested that hydrogen bonding between XOD and dextran is restricted by molecular steric hindrance.

Introduction The freshness of fish can be tested simply with freshness testing paper (FTP) based on the K value concept. FTP contains some freeze-dried enzyme (i.e., xanthine oxidase [XOD]), and enables the relative content of nucleotides in fish muscle to be evaluated. The enzyme, however, is thermally unstable, so FTP cannot be preserved at ambient temperature. Thus, for practical reasons, it is important to stabilize the freeze-dried enzyme during shipping and long-term storage. Freeze drying generates both freezing and drying stresses that can destabilize proteins to various degrees. In order to minimize the destabilization of freeze-dried proteins, various types of excipients have been used as stabilizers. Preferential exclusion and freeze-concentrated glass formation are the stabilizing mechanisms of the excipients on freeze-dried proteins during freezing. The former involves excipients that are 599

600

PART 2: Poster Presentations

preferentially excluded from the protein’s hydration shell, thus the protein’s unfolding is prevented. The latter involves excipients that form a highly viscous glassy matrix by freeze concentration and decrease the degradation rate (Anchordoquy and others 2001; Wang 2003; Ohkuma and others 2008). The mechanisms of stabilization of freeze-dried protein by excipients during dehydration are water substitution and glass transition stabilization. Water substitution postulates that the nativelike structure of protein is maintained by the formation of hydrogen bonds between the excipients and proteins in place of the removed water molecules (Schebor and others 1997; Kreilgaard and others 1998; Suzuki and others 1998; Kawai and others 2006). The concept of glass transition stabilization arises from the postulation that the glassy matrix initially formed by freeze concentration maintains its state even at an elevated temperature because of the decrease in the number of water molecules, which play the role of a plasticizer, in the freeze drying process. In the glass transition stabilization concept, the glass transition temperature (Tg) of stabilizers is one of the most significant parameters because the formulations change from glassy to liquidlike rubbery state at temperatures above their Tg. Among the excipients, disaccharides and/or polymers are known to be effective stabilizers (Prestrelski and others 1993; Chang and others 1996). For example, recombinant human factor XIII was completely stabilized by the addition of 100 mM sucrose prior to freeze drying and subsequent 3-month storage at 40°C (Kreilgaard and others 1998). Polymers (such as dextran and bovine serum albumin [BSA]) stabilize freeze-dried protein adequately. However, a combination of disaccharides and polymers improves the stabilization of dried protein synergistically (Allison and others 1998). Since few studies have investigated the effects of stabilizers on the storage stability of the enzymes used for FTP, this study, as the first step, aimed to elucidate the effects of stabilizers on the long-term storage stability of freeze-dried XOD.

Materials and Methods Preparation of Freeze-Dried Xanthine Oxidase Samples XOD (from butter milk [EC 1.1.3.22]; Wako Pure Chemical Industries, Osaka, Japan) was dialyzed against 20 mM potassium phosphate buffer (pH 7.5) for 48 h at 4°C to remove stabilizing agents. The XOD activity of the dialyzed solution was determined to be 0.6 U mL−1. Three types of stabilizer were used: 200 mM sucrose, 200 mM sucrose with 1% dextran (molecular weight, 10,400), and 200 mM sucrose with 1% BSA. As controls, an additive-free formulation was also prepared. Aliquots of 1 mL of each solution were placed into a 2-mL polyethylene tube and frozen instantaneously with liquid nitrogen for at least 1 min. The frozen solids were transferred to a precooled freeze dryer. The freeze drying was performed by warming from −40° to 25°C at 3.0 × 10−2 mmHg over a 2-day period. After freeze drying, the residual water in all samples was removed over phosphorus pentoxide in a vacuum desiccator for 1 week at room temperature.

Effects of Excipients on the Storage Stability of Freeze-Dried Xanthine Oxidase

601

Determination of Water Content A Karl Fischer coulometer (model 737 KF; Metrohm, Herisau, Switzerland) was used to measure the residual moisture content (percentage of grams of water per grams of total dried solid) of the samples. Differential Scanning Calorimetry The Tg of each sample was investigated by using a differential scanning calorimeter (DSC) (DSC-50; Shimadzu, Kyoto, Japan). Indium and distilled water were used to calibrate the temperature and heat capacity for the DSC. Alumina powder was used as the reference material. The sample (∼15 mg) was weighed and sealed in an aluminum DSC pan. All measurements were performed at 5°C/min from 0° to 150°C. Assay of Xanthine Oxidase Activity The freeze-dried samples were stored at 25°C over a 90-day period. After storage, the samples were rehydrated with distilled water to their original concentration. The XOD activity was determined by the enzymatic conversion of the substrate xanthine to uric acid. The assays were performed by adding 10 μL of XOD sample solution to 300 μL of 0.12 mM xanthine (sodium salt) in 20 mM sodium phosphate buffer (pH 7.5), and the absorbance at 292 nm at 25°C was measured immediately by using an ultravioletvisible (UV-Vis) spectrophotometer (model V-630BIO; Jasco, Tokyo, Japan). From the results, the initial reaction rate was determined and estimated to be the XOD activity. The remaining XOD activity was expressed as a percentage of the activity prior to freeze drying.

Results All formulations containing stabilizers showed glass transition at a temperature above 25°C. The Tg values and moisture contents of the formulations are listed in Table 55.1. Because of the plasticizing effect of the buffer and residual moisture, the Tg of the sucrose formulation was lower than that of pure sucrose (Tg = 68°C [Kawai and others 2005]). The Tg values of sucrose + polymer (BSA or dextran) formulations were higher than that of the sucrose formulation because compared to sucrose, the polymers have a higher Tg and greater resistance to a decrease in Tg induced by the plasticizing effect of water.

Table 55.1. Values for glass transition temperature (Tg) and the moisture contents of xanthine oxidase with various excipients Tg (°C)

% Moisture contenta

200 mM of sucrose

60.6

1.36 ± 0.14

200 mM of sucrose + 1% bovine serum albumin

66.2

1.18 ± 0.43

200 mM of sucrose + 1% dextran

71.9

0.95 ± 0.30

Formulation

a

The values are mean ± standard deviation (n = 3).

602

PART 2: Poster Presentations

After preparation 120

Remaining activity (%)

100 Additive free

80 60

Suc

40

Suc+1%BSA

20 Suc+1%Dex 0 0

20

40

60

80

100

Storage time (days)

Figure 55.1. Remaining activity of xanthine oxidase (XOD) with various excipients after preparation and subsequent storage at 25°C. The values are mean ± standard deviation (n = 3). BSA, bovine serum albumin; Dex, dextran; and Suc, sucrose.

The remaining XOD activity of the formulations after preparation and storage is shown in Figure 55.1. The XOD activity of the additive-free formulation decreased to 22% after its preparation and to 4% during subsequent storage. On the other hand, the formulations containing stabilizers maintained much higher XOD activity than did the additive-free formulation. The XOD activity of the sucrose formulation decreased to ∼63% after its preparation and to 45% during storage. The addition of 1% BSA to the sucrose formulation improved XOD stability: XOD activity decreased to 85% after preparation of this variant and to 58% during storage. In contrast, the addition of 1% dextran to the sucrose formulation diminished the storage stability: the XOD decreased to 47% during preparation and to 28% during storage.

Discussion It is known that polymers such as dextran and BSA have greater preferential exclusion ability and higher freeze-concentrated and/or dehydrated glass transition temperatures than do disaccharides such as sucrose. On the other hand, the effect of water substitution by the polymer is inferior to that of the disaccharide because of the polymer ’s molecular steric hindrance. So it is thought that polymers effectively stabilize protein in the frozen state, whereas disaccharides effectively stabilize protein in the dehydrated state. Based on these assumptions, the results presented in Figure 55.1 were interpreted as follows. In the frozen state, the stabilizing effect of sucrose + polymers on XOD is greater than that of sucrose alone. Furthermore, BSA is known to be a good cryoprotectant

Effects of Excipients on the Storage Stability of Freeze-Dried Xanthine Oxidase

Frozen state

603

Dehydrated state

XOD

(a)

H2O

Ice

(b)

XOD

(c)

XOD

H-bond

sucrose

BSA

dextran

Figure 55.2. Schematic illustration of freeze-dried xanthine oxidase (XOD) in different glassy matrices: (a) sucrose, (b) sucrose + bovine serum albumin (BSA), and (c) sucrose + dextran.

relative to dextran (Nema and Avis 1992), probably because BSA has great preferential exclusion ability. On the other hand, in the dehydrated state, the stabilizing effect of sucrose is greater than that of sucrose + polymers because the molecular steric hindrance of polymers interferes with the intermolecular hydrogen bonding between XOD and sucrose (Figure 55.2). As a result, the XOD activity after preparation was higher in the following order: sucrose + BSA > sucrose > sucrose + dextran. However, the remaining XOD activity still decreased gradually during storage. To determine how to maintain XOD activity completely, further studies are required.

References Allison SD, Randolph TW, Manning MC, Middleton K, Davis A, Carpenter JF. 1998. Effects of drying methods and additives on structure and function of actin: mechanisms of dehydration-induced damage and its inhibition. Arch Biochem Biophys 358:171–81. Anchordoquy TJ, Izutsu K, Randolph TW, Carpenter JF. 2001. Maintenance of quaternary structure in the frozen state stabilizes lactate dehydrogenase during freeze-drying. Arch Biochem Biophys 390:35–41.

604

PART 2: Poster Presentations

Chang BS, Beauvais RM, Dong A, Carpenter JF. 1996. Physical factors affecting the storage stability of freeze-dried interleukin-1 receptor antagonist: glass transition and protein conformation. Arch Biochem Biophys 331:249–58. Kawai K, Hagiwara TR, Suzuki T. 2005. Comparative investigation by two analytical approaches of enthalpy relaxation for glassy glucose, sucrose, maltose, and trehalose. Pharm Res 22:490–5. Kawai K, Hagiwara T, Takai R, Suzuki T. 2006. The role of residual water for the stability of protein freeze-dried with trehalose. In: Buera MDP, Welti-Chanes J, Lillford PJ, Corti HR, editors. Water properties of food, pharmaceutical, and biological materials. Boca Raton, FL: CRC. p 543–50. Kreilgaard L, Frokjaer S, Flink JM, Randolph TW, Carpenter JF. 1998. Effects of additives on the stability of recombinant human factor XIII during freeze-drying and storage in the dried solid. Arch Biochem Biophys 360:121–34. Nema S, Avis KE. 1992. Freeze-thaw studies of a model protein, lactate dehydrogenase, in the presence of cryoprotectants. J Parenter Sci Technol 47:76–83. Ohkuma C, Kawai K, Viriyarattanasak C, Mahawanich T, Tantratian S, Takai R, Suzuki T. 2008. Glass transition properties of frozen and freeze-dried surimi products: effects of sugar and moisture on the glass transition temperature. Food Hydrocolloids 22:255–62. Prestrelski SJ, Arakawa T, Carpenter JF. 1993. Separation of freezing- and drying-induced denaturation of lyophilized proteins using stress-specific stabilization. II. Structural studies using infrared spectroscopy. Arch Biochem Biophys 303:465–73. Schebor C, Burin L, Buera MP, Aguilera JM, Chirife J. 1997. Glassy state and thermal inactivation of invertase and lactase in dried amorphous matrices. Biotechnol Prog 13:857–63. Suzuki T, Imamura K, Fujimoto H, Okazaki M. 1998. Relation between thermal stabilizing effect of sucrose on LDH and sucrose-LDH hydrogen bond. J Chem Eng Jpn 31:565–70. Wang W. 2003. Lyophilization and development of solid protein pharmaceuticals. Int J Pharm 203:1–60.

56 Water Properties in Bread Produced with an Innovative Mixer E. Curti, E. Vittadini, A. Di Pasquale, L. Riviera, F. Antoniazzi, and A. Storci

Abstract This work studied the effect of an innovative dough-making process on the physicochemical properties, water properties, and staling of white bread. White bread was produced by using either a standard dough maker (standard bread, SB) or an innovative dough maker (innovative bread, IB). The volume and porosity of IB and SB loaves turned out to be comparable, although IB crumbs were significantly softer than SB crumbs after 7 days of storage. Water properties were affected by the innovative mixing, which suggested a stronger interaction between water and solids in IB. Significantly less moisture was extractable from the crumbs of IB than from SB at up to 5 days of storage, and more ice crystals melted at lower temperature in IB (differential scanning calorimetry [DSC] peak shape) than in SB. Proton nuclear magnetic resonance (1H NMR) mobility indicated the presence of two detectable 1H populations (fast [A] and slow [B]): the 1H population A was predominant in IB (ca. 80%–90%; transverse relaxation time [T2A] = 7–9 ms) and was a minority in SB (10%–30%; T2A = 2–6 ms). The 1H self-diffusion coefficient was lower in the IB samples. Amylopectin recrystallization was comparable in IB and SB after 7 days of storage. The innovative mixing process affected water properties significantly, inducing a stronger water-solid interaction that produced a softer texture that was preserved for longer periods.

Introduction Bread quality and stability are affected by many factors, including formulation, processing, and storage conditions. Ingredients in bread production are generally mixed discontinuously: the ingredients (solid and liquid) are introduced in a chamber where a rotating shaft provides mechanical energy allowing for hydration of solid components, formation of the gluten network, and incorporation of air bubbles, leading to the development of a viscoelastic dough. A recently designed innovative mixer is shown in Figure 56.1. This mixer (Bakmix; Storci, Collecchio, Parma, Italy) introduces dry (stored in 2 and volumetrically dosed in 3 [Figure 56.1]) and liquid (dosed with a pump and delivered through 4) ingredients simultaneously into a chamber (5) called Premix (Storci 2005) containing a fast-stirring mechanism. Dry and liquid ingredients are subjected to a centrifugal force that causes their dispersion as dust and aerosol, respectively, in air. 605

606

PART 2: Poster Presentations

Figure 56.1. Schematic representation of Bakmix: 1, motor; 2, semolina feeding; 3, volumetric semolina doser; 4, inlet of liquid ingredients; 5, mixing chamber; 6, outlet; 7, twin-screw; and 8, outlet die.

The dispersed materials come in contact with each other in the Premix chamber, inducing a uniform hydration of the surface of each individual dry particle and forming an incoherent mass (2–5 s). This mixture is immediately extracted from the chamber (through 6) and led into a twin-screw low-pressure extruder (7) (5 atm and 100 rpm) that favors the transformation of the incoherent wet mass into a dough that can undergo the normal bread-processing procedure. The Bakmix mixing process is significantly different from the traditional mixing process not only for its lack of extensive kneading action but also for allowing a more even exposure of semolina particles to water and for its significantly shorter processing time (ca. 5–10 s vs ∼10 min). It is hypothesized that this innovative mixing process may affect water-solid interactions and the state of water in the dough, and consequently, may influence bread properties and stability. The objective of this work was to study the effect of the innovative mixer on physicochemical properties, water properties, and staling of white bread.

Materials and Methods Bread Formulation and Production Bread was produced according to method 10-10A of the American Association of Cereal Chemistry (AACC 2000b) using the following formulation on a semolina basis: wheat semolina (100%), water (65%), sugar (6%), yeast (3%), sunflower oil (3%), and salt (2%).

Water Properties in Bread Produced with an Innovative Mixer

607

The ingredients were mixed using either a standard mixer for standard bread (SB) (KitchenAid 5KSM5; KitchenAid Europe, Brussels, Belgium) or, for innovative bread (IB) (Bakmix). Bread loaves were allowed to cool to room temperature for 2 h prior to being placed in sealed plastic bags and stored at 25°C for 7 days. The loaves were analyzed at 0, 1, 3, 5, and 7 days after production. Bread Characterization Water Activity Water activity was measured with an Aqualab 2000 (Decagon Devices, Pullman, WA, USA) at 25°C. At least triplicate samples of crumbs, crumbs just below the crust, and crust were analyzed for each bread loaf. Moisture Content Moisture content of crumbs (from the loaf center) and crust were determined in triplicate by weight loss in samples dried at 105°C to constant weight. DSC-Freezable Water and Retrograded Amylopectin Thermograms were obtained with a differential scanning calorimeter (DSC Q100; TA Instruments, New Castle, DE, USA) under a nitrogen atmosphere (50 mL/min). Crumb samples (10–15 mg) were compressed and then placed in hermetically sealed aluminum pans (Perkin Elmer, Shelton, CT, USA), quench cooled to −50°C, and then heated to 130°C at 5°C/min. At least duplicated samples were analyzed for each bread. Differential scanning calorimetry (DSC)-freezable water and retrograded amylopectin were obtained as described by Vittadini and Vodovotz (2003). Loaf Volume Loaf volume was determined by rapeseed displacement by AACC method 10-05 (AACC 2000a). Texture Analysis Bread-crumb hardness was measured on eight (2 × 2 × 2 cm) samples extracted from the center of bread loaves by compression (40%) with a cylindrical probe (P/35 DIA cylinder aluminum) using a TAX T2i texture analyzer (Stable Micro Systems, Godalming, UK). Peak force at 40% compression was taken as crumb hardness (in grams). Image Analysis Crumb grain was evaluated by means of a digital image analysis system as described by Chiavaro and others (2008). NMR Measurements A low-field spectrometer (Minispec; Bruker Biospin, Milan, Italy) operating at 20 MHz and 25°C was used to acquire the proton (1H) free-induction decay (FID) (using the

608

PART 2: Poster Presentations

standard 90° pulse) and Carr-Purcell-Meiboom-Gill (CPMG) spin-echo decay (1H transverse relaxation [T2]) (Meiboom and Gill 1958). T2 populations, their relaxation times, and relative 1H amounts were obtained by fitting CPMG decays with a biexponential model using SigmaPlot version 6.00 (Systat Software, Chicago, IL, USA). Proton nuclear magnetic resonance (1H NMR) mobility experiments were performed on four crumb samples obtained from the center of each bread loaf and placed in a 10-mm-diameter NMR tube sealed with parafilm to prevent moisture loss. Statistical Analysis An independent Student’s t-test analysis was used to identify differences between breads produced with different mixers at the same storage time (Statistica software version 5.5; Statsoft, Tulsa, OK, USA).

Results and Discussion Bread made with the innovative mixer had an overall appearance comparable to that of the product that underwent the traditional mixing. Loaf volume for SB and IB (939 ± 38 mL and 906 ± 52 mL, respectively) was not significantly different (P < 0.05), nor was the development of the crumb structure (comparable pore size distribution, data not shown). Hardness of the fresh crumb was about 120 g in the two products (Table 56.1) and was observed to increase during storage, as expected. IB and SB had a similar hardening rate during the first 5 days of storage, but IB was found to be significantly softer (482 ± 163 g) than SB (626 ± 135 g) at 7 days of storage. To better understand the difference in hardness in the stored product, other parameters that were reported to have a role in bread staling were investigated. The degree of recrystallized amylopectin was not significantly different in the two products at 7 days of storage (i.e., 2.2 ± 0.4 J/g in SB and 1.5 ± 0.5 J/g in IB), and therefore a role played by the discontinuous starch phase in the harder texture of SB over IB is unlikely (Hallberg and Chinachoti 2002). As the innovative mixing process was hypothesized to possibly affect water-solid interactions and the state of water in the product, the parameters characterizing water

Table 56.1. Hardness (in grams) of innovative bread and standard bread crumb during storage Time (days)

Innovative bread

Standard bread

0

133.5 ± 40.9

117.7 ± 25.3

1

260.9 ± 72.6

295.2 ± 55.2

3

357.0 ± 43.6

338.2 ± 34.5

5

497.1 ± 143.7

588.9 ± 77.2

7

482.7 ± 163.8*

626.4 ± 135.2*

An asterisk following numbers indicates significant differences between innovative bread and standard bread at the same storage period (P < 0.05).

Water Properties in Bread Produced with an Innovative Mixer

609

properties were analyzed thoroughly. The crust of both fresh breads had a moisture content of ∼20% (g water/100 g sample), and for the crumbs it was found to be ∼40%. During storage, moisture content progressively increased in the crust and decreased in the crumb in both samples, as expected (Baik and Chinachoti 2001; Gray and Bemiller 2003). The moisture content of IB crumbs remained slightly lower than that of SB for up to 5 days of storage, possibly suggesting a stronger water-solid interaction in this product. Similarly, the water activity of each of the IB bread fractions (crumbs, crumbs just below the crust, and crust) was mostly found to be slightly (but not significantly) lower than those of the corresponding SB fractions (data not shown). Water was also characterized in terms of thermal properties with a focus on the ice-melting transition. Although no significant differences were found between IB and SB regarding the enthalpy of the transition (% FW [g frozen water/100 g sample] was found to be 40.1% ± 2.6% in IB and 39.2% ± 3.2% in SB fresh samples), the transition was shifted toward lower temperatures in IB as compared to SB, indicating a lower melting-ice phase in IB, possibly induced by a stronger water-solute interaction and/or altered ice crystal distribution in this sample. NMR spectroscopy was used to study molecular mobility, and multiple 1H-NMR experiments were performed to characterize the very rigid 1H fraction (FID) and the more mobile 1H fraction (T2). The 1H FIDs of IB and SB (both fresh and stored for 7 days) are shown in Figure 56.2a: a fast decay of the first portion of the FID curve (<0.3 ms) is indicative of the presence of a very rigid 1H population. The 1H-FID decay of fresh IB was only slightly faster than in fresh SB. The 1H-FID rigid component became progressively more relevant during storage (as previously reported [Sereno and others 2007]) in both breads, more significantly in SB as compared to IB, as shown in Figure 56.2a (7 days of storage). This lower number of rigid protons may play a role in the observed softer texture of the IB product at 7 days of storage. 1 H T2 CPMG relaxation curves were well fitted with a biexponential model (R2 > 0.99) in all samples, indicating the presence of two 1H populations with distinct mobility: the less mobile 1H population A and the more mobile 1H population B. 1H population A was found to be significantly predominant (80%–90%) in IB but played a minor role (10%–30%) in SB, as shown in Figure 56.2b. The percentage of 1H population A slightly increased in IB after 3 days of storage but slightly decreased in SB, possibly because of a different water redistribution among bread components in the two breads. 1H relaxation rates (T2A and T2B) were found to be, respectively, ca. 6–7 ms and ca. 30–40 ms in IB, and underwent only minor variations during storage. Conversely, the T2A of SB was very fast (∼1 ms) at days 0 and 1 of storage and then increased to ∼6 ms; the T2B of SB increased progressively from ∼10 to ∼30 ms during the 7 days of storage (Ruan and Chen 2001). It may be speculated that the 1H population A was associated with an amorphous phase that may have been more evenly hydrated in the IB than in the SB product because of a more even exposure of semolina particles to water. The hydrated amorphous phase may have retained plasticity better at longer storage times, resulting in a softer product.

Figure 56.2. (a) Proton (1H) free-induction decay of innovative bread (IB) and standard bread (SB) fresh and after 7 days of storage. (b) 1H relaxation rate (T2A) populations for IB (population A, black; and population B, gray) and SB (population A, black stripes on white; and population B, gray stripes on white) during 0, 1, 3, 5, and 7 days of storage.

610

Water Properties in Bread Produced with an Innovative Mixer

611

Conclusions The innovative mixing apparatus appeared to have affected physicochemical properties of bread, resulting in a softer product (IB) at 7 days of storage. The innovative mixing process may have induced a more even hydration of the amorphous phase of the semolina particles, possibly favoring stronger water-solid interaction in IB and possibly inducing a more effective plasticization at longer storage times, which resulted in a softer product.

References American Association of Cereal Chemists (AACC). 2000a. Method 10-05: guidelines for measurement of volume by rapeseed displacement. St Paul, MN: AACC. American Association of Cereal Chemists (AACC). 2000b. Method 10-10B: optimized straight-dough bread-making method. St Paul, MN: AACC. Baik MY, Chinachoti P. 2001. Effects of glycerol and moisture gradient on thermomechanical properties of white bread. J Agric Food Chem 49:4031–8. Chiavaro E, Vittadini E, Musci M, Bianchi F, Curti E. 2008. Shelf-life stability of artisanally and industrially produced durum wheat sourdough bread (“Altamura bread”). LWT Food Sci Technol 41:58–70. Gray JA, Bemiller JN. 2003. Bread staling: molecular basis and control. Compr Rev Food Sci Food Saf 2:1–21. Hallberg LM, Chinachoti P. 2002. A fresh perspective on staling: the significance of starch recrystallization on the firming of bread. J Food Sci 67:1092–6. Meiboom S, Gill D. 1958. Modified spin-echo method for measuring nuclear relaxation times. Rev Sci Instrum 29:688–91. Ruan RR, Chen PL. 2001. Nuclear magnetic resonance techniques. In: Chinachoti P, Vodovotz Y, editors. Bread staling. Boca Raton, FL: CRC. p 113–27. Sereno NM, Hill SE, Mitchell JR, Scharf U, Farhat IA. 2007. Probing water migration and mobility during the aging of bread. In: Farhat GAB, Webb GA, editors. Magnetic resonance in food science: from molecules to man. Gateshead, UK: Royal Society of Chemistry. p 89–95. Storci A, inventors; Storci SRL, assignee. 2005 Dec 29. High speed mixing and homogenization of solid and liquid in e.g. food-, pharmaceutical- and paint manufacture, atomizes liquid and mixes rapidly with powder dispersion in air. German Patent DE102005025016. Vittadini E, Vodovotz Y. 2003. Changes in the physicochemical properties of wheat- and soy-containing breads during storage as studied by thermal analyses. J Food Sci 68:2022–7.

57 Evaluation of Deformation and Shrinking of Potato Slabs During Convective Drying R. Campos-Mendiola, C. Gumeta-Chávez, J. J. Chanona-Pérez, L. Alamilla-Beltrán, A. Jiménez-Aparicio, and G. F. Gutiérrez-López

Abstract This work studied macroscopic shrinking and deformation in potatoes by means of image and fractal analysis during convective drying. Slab-shaped sections of the material were subjected to image capture (lateral and from above) during drying. Shrinking was evaluated as the coefficient of area/initial area. Two stages of shrinkage were observed. The first may be associated with bending of rigid structures and shrinkage of nonrigid structures. This stage may be related to the removal of water from pores and capillaries. The second stage may be due to cell wall collapse. A descriptive model incorporating microscopic and macroscopic changes during deformation and shrinkage is presented, and the role of rigid and nonrigid structures in those phenomena is partially explained.

Introduction Vegetable foods are complex structural systems formed by different tissues responsible for the systems’ composition and structural characteristics. When vegetables are dried, structural changes occur that affect mass and heat-transport properties (Aguilera and Stanley 1999). Shrinkage has been studied by evaluating parameters such as porosity and density, which have been correlated with water content. Volume decrease or shrinkage has been studied in several vegetables, such as carrot, peach, potato, sweet potato, and garlic (Lozano and others 1980; Zogzas and others 1994; Karathanos and others 1996), and predictive models for the deformation of finite potato cylinders have also been reported. The models indicate that the center shrinks more than the surface (Yang and others 2001). Tools such as image analysis have been used to study structural changes in several foods (Aguilera 2003; Chanona and others 2003). Image analysis is an interesting tool for quantifying parameters related to cellular dimensions such as projected area, Feret diameter, and fractal dimension. Also, fractal geometry may be useful for studying interfaces and their roles as transfer-controlling barriers (Campos and others 2007). Given its high degree of accuracy, image analysis has advantages over conventional methods for obtaining dimensional measurements (Martynenko 2006). 613

614

PART 2: Poster Presentations

The aim of the present work was to obtain information on the nonisotropic shrinkage of potato slabs during convective dehydration, and to evaluate variant conditions by image analysis.

Materials and Methods Materials Potatoes (Solanum tuberosum var. Alpha) purchased in a local market were cut into circular slabs of 40-mm diameter (D) and 3-mm thickness (T). Thickness was selected based on the report by Rovedo and others (1995), which recommended a T/D ratio of <1:10. Slabs were submerged in a 0.05% solution of sodium metabisulfite (Sigma 7681-57-4; Sigma-Aldrich, St Louis, MO, USA) for 2 min to control browning. Drying Equipment and Image Acquisition System Two slices were dehydrated simultaneously in an experimental air dryer as described previously in another work (Campos and others 2007). Air temperature and velocity surrounding the samples were monitored by means of a thermoanemometer (model 8330-M; TSI, Shoreview, MN, USA). Four air-drying temperatures (40°, 50°, 60°, and 70°C), four air velocities (1, 1.5, 2.0, and 2.5 m/s), and four thicknesses of sample (1, 2, 3, and 4 mm) were tested. One of the slices was used to evaluate shrinkage-deformation by using two digital cameras (Kodak DC290 zoom) to capture digital images of the lateral and top views of samples during drying. Both cameras were connected to a serial port of a generic personal computer fitted with a video-recording card (All in Wonder Rage 128pro; ATI Technologies, Ottawa, KS, USA) and image recording software (ATI version 6.2 Multimedia Center). The other slab was used to evaluate weight loss and moisture of the sample during drying experiments. The weight of samples was evaluated in an analytic balance (Analytical Plus; Ohaus, Pine Brook, NJ, USA). The initial moisture of slabs was determined according to Association of Official Analytical Chemists method 32.1.03 (AOAC 1995). Images obtained from the side (lateral) and top (superior) views of the slabs were stored in a 640- × 480pixel bitmap format. Image Processing ImageJ software 1.34 (National Institutes of Health, Bethesda, MD, USA) was used to image the processing of slabs during drying. The steps included transformation into a gray scale of captured images, followed by segmentation of the lateral boundary to obtain the projected shape of the slabs. Extraction features included the dimensionless values of lateral and superior views of the projected area (Alat and Asup), the projected perimeter (Plat and Psup), and the Feret diameter (Flat and Fsup), which were taken as indicators of nonisotropic shrinkage-deformation during drying to determine a correlation with moisture removal from the sample.

Evaluation of Deformation and Shrinking of Potato Slabs

615

Results and Discussion Image Analysis in Evaluation of Shrinkage and Deformation The Plat, Psup, Flat, and Fsup decreased with drying time until constant values were reached. The behavior of these parameters was the same (Figure 57.1). Measurement of Plat, Psup, Flat, and Fsup enabled the shrinkage to be evaluated. This shrinkage pattern was similar to that found by several authors (e.g., see Lozano and others 1983). Figure 57.2 shows the variations in Alat of the images as a function of drying time. Alat increased during drying and could be used to evaluate the magnitude of deformation of potato slabs during drying, whereas Asup could be used to measure only overall shrinkage. Influence of Drying Conditions on Shrinkage and Deformation Figure 57.2 shows the influence of drying temperature on the kinetics of Alat, which has peaks at different drying times. A first peak appears before 20 min and indicates the first deformation stage. A second peak at ∼60 min indicates the deformation stage. Approaching the final drying process, the Alat values remained constant for all drying temperatures. Additionally, both peaks depended on drying temperature. At lower drying temperatures (40° and 50°C), the first peaks presented at the

Figure 57.1. Kinetics of Plat, Flat, Psup, and Fsup (defined in the text) of potato slabs during drying at 55°C, air velocity = 1.7 m/s, and thickness = 2.5 mm. l/l0 is the lateral boundary as compared to the initial sample.

616

PART 2: Poster Presentations

Figure 57.2. Effect of temperature on kinetics of the lateral projected area (Alat) of potato slabs during drying at air velocity = 1.7 m/s and thickness = 2.5 mm.

former drying time, whereas, at higher drying temperatures (60° and 70°C), these peaks appeared at the latter drying time. The first peak deformation may be associated with shrinkage caused by evaporation of nonbound water that may be related to water diffusion into the pores and capillaries of the material, as reported by Krokida and Maourulis (2001). This drying stage is typically called isotropic shrinkage. The second peak deformation may be caused by other factors, such as cellular wall collapse (Prothon and others 2003). Some reports (Prothon and others 2003) indicated that lower drying rates may cause the cell walls to collapse because of the separation of the medium lamella mainly during the latter drying stages. This stage (typically nonisotropic shrinkage) has been studied by others (Ng and Waldron 1997; Van Dijk and others 2002; Prothon and others 2003). In general, lower drying temperatures promote major deformation, and this behavior could be mainly associated with microstructural changes in the cellular wall. Figure 57.3 compares the effect of the temperature and airflow conditions. The shrinkage-deformation behavior during drying of the potato slabs did not show a regular pattern with respect to airflow. At lower temperatures and airflows, the deformation increased smoothly with drying time until reaching a maximum, whereas, when the material was dried at high temperatures and airflows, its deformation increased rapidly until reaching a maximum and then decreased, reaching initial Alat values.

Evaluation of Deformation and Shrinking of Potato Slabs

617

Figure 57.3. Variation in the lateral projected area (Alat) of potato slabs during drying at the conditions indicated. Alat/Alat,0 is the projected lateral area (Alat) as compared to the initial sample (Alat,0).

Conclusions Image processing was helpful in evaluating the shrinkage-deformation changes in potato slices during the drying process. Alat kinetics showed that there are two shrinkage-deformation stages during the convective drying of potato slabs. The magnitude of shrinkage-deformation depends on selection of the drying conditions. These results can be helpful for understanding the structural changes (shrinkage, deformation, and collapse) that occur during the drying of vegetable tissues.

Acknowledgments This work was sponsored by Secretaria de Investigación y Posgrado del Instituto Politécnico Nacional (SIP-IPN) (Secretary of Research and Graduate Institute of National Polytechnic) projects 20070631 and 20071011; Consejo Nacional de Ciencia y Tecnología (CONACYT) (National Council for Science and Technology) projects 48061-Z and 59730; and Comisión de Operación y Fomento de Actividades Académicas (COFAA) (Committee on Operation and Development of Academic Activites).

References Aguilera JM. 2003. Drying and dried products under the microscope. Food Sci Technol Int 9:137–43. Aguilera JM, Stanley DW. 1999. Simultaneous heat and mass transfer: dehydration. In: Microstructural principles of food processing and engineering. 2nd ed. London: Elsevier Applied Science. p 373–412.

618

PART 2: Poster Presentations

Association of Official Analytical Chemists (AOAC). 1995. Official methods of analysis. 16th ed. Arlington, VA: AOAC International. Campos MR, Hernàndez SH, Chanona PJJ, Alamilla BL, Jiménez AA, Fito P, Gutiérrez LGF. 2007. Nonisotropic shrinkage and interfaces during convective drying of potato slabs within the frame of the systematic approach to food engineering systems (SAFES) methodology. J Food Eng 83:285–92. Chanona PJJ, Alamilla BL, Farrera RRR, Quevedo R, Aguilera JM, Gutiérrez LGF. 2003. Description of the convective air-drying of a food model by means of the fractal theory. Food Sci Technol Int 9:207–13. Karathanos V, Kanellopoulos NK, Belessiotis VG. 1996. Development of porous structure during air drying of agricultural plant products. J Food Eng 29:167–83. Krokida MK, Maourulis ZB. 2001. Structural properties of dehydrated products during rehydration. Int J Food Sci Technol 36:529–38. Lozano JE, Rotstein E, Urbicain MJ. 1980. Total porosity and open-pore porosity in the drying of fruit. J Food Sci 45:1403–7. Lozano JE, Rotstein E, Urbicain MJ. 1983. Shrinkage, porosity and bulk density of foodstuffs at changing moisture contents. J Food Sci 48:1497–502. Martynenko AI. 2006. Computer-vision system for control of drying processes. Drying Technol 24:879–88. Ng A, Waldron KW. 1997. Effect of steaming on cell wall chemistry of potatoes (Solanum tuberosum cv. Bintje) in relation to firmness. J Agric Food Chem 45:3411–8. Prothon F, Ahrné L, Sjöholm I. 2003. Mechanisms and prevention of plant tissue collapse during dehydration: a critical review. Crit Rev Food Sci Nutr 43:447–79. Rovedo CO, Suarez C, Viollaz PE. 1995. Drying of foods: evaluation of drying model. J Food Eng 26:1–12. Van Dijk C, Fischer M, Beekhuizen JG, Boeiu C, Stolle-Smits T. 2002. Texture of cooked potatoes (Solanum tuberosum). 3. Preheating and the consequences for the texture and cell wall chemistry. J Agric Food Chem 50:5098–106. Yang H, Sakai N, Watanabe M. 2001. Drying model with non-isotropic shrinkage deformation undergoing simultaneous heat and mass transfer. Drying Technol 19:1441–60. Zogzas NP, Maroulis ZB, Marinos-Kouris D. 1994. Densities, shrinkage and porosity of some vegetables during air drying. Drying Technol 12:1653–66.

58 Effects of Different Cut-Induced Microstructural and Macrostructural Arrays on Convective Drying of Agave atrovirens Karw C. Gumeta-Chávez, J. J. Chanona-Pérez, L. Alamilla-Beltrán, G. Calderón-Domínguez, A. Vega, P. Ligero, J. A. Mendoza-Pérez, and G. F. Gutiérrez-López

Abstract Mexican agave (Agave atrovirens Karw) is important for the Mexican tequila industry. After the juice has been extracted, however, the plants have not been used for byproducts, and their potential for lignin and cellulose extraction needs to be studied further. The drying step preceding lignin and cellulose extraction was evaluated for agave slices cut transversely and longitudinally to the fiber edge. Structural changes during drying were analyzed by image and fractal analysis. The drying kinetics were evaluated for both materials. Morphological evolution was monitored by means of charged-coupled device (CCD) cameras to capture images laterally and superiorly. Shrinkage and deformation during drying were evaluated, as well as effective diffusion coefficients. The cut type influenced the type of structures exposed to drying air and affected the drying rates. Transversely cut slabs were more prone to deformation and shrinking, whereas longitudinally cut samples were more resistant to changes in form and size. Different structural arrays play important roles in moisture transfer, which in turn affects lignin and cellulose extraction. The lower levels of shrinkage and deformation achieved presented advantages over extraction.

Introduction Traditionally, several products can be obtained from Agave atrovirens Karw. Examples include pulque and aguamiel (beverages similar to diluted sugar syrup). We could potentially use this plant for the cellulose, lignin, and sugars remaining in its leaves after the extraction of the aguamiel juice (Idarraga and others 1999). An important step in the process of obtaining cellulose and lignin is convective drying, because the water and solvent diffusion will be directly affected by macrostructural and microstructural organization. During convective drying, the cells and pores collapse from water loss. The collapse of capillary structures and pores could increase the water loss and enhance the shrinkage and collapse of the material’s structures (Genskow 1990). The shrinkage in vegetable tissues is often considered anisotropic, but in traditional drying studies, the shrinkage has been studied as a function of the water content by 619

620

PART 2: Poster Presentations

means of mathematical models based on the initial and final moisture contents of the materials, and thus has been considered isotropic (Susuki and others 1976). Evaluation of the characteristic structure and composition of some vegetables can be helpful in understanding structural changes such as shrinkage, deformation, and collapse that occur during the drying process. One of these vegetables is agave, because its structural arrangement involves long fibers of cellulose (long, rigid structures) organized longitudinally as channels that enable the efficient conduction of nutrients and water along the plant tissues. These rigid structures or fibers also serve a support function for other cells of the vegetable tissue. This particular structural arrangement of the agave tissue supplies an attractive material model wherein the effect of cut type on water-diffusion transport can be studied, and the role of rigid structures in the shrinkage and deformation phenomena in this biomaterial can be evaluated. This work evaluated the effect of the cut type used for agave slices on the macrostructural and microstructural changes during convective drying and the water transport within the slice.

Materials and Methods Convective Drying Experiments Leaves of the Agave atrovirens Karw plant were used. The cuticle of the leaves was removed by a stainless-steel knife. Agave tissue was cut transversely and longitudinally to the fiber direction, and circular 3-mm-thick, 3.8-diameter slices were removed with an electric slicer. Drying kinetics were evaluated by using two 32 experimental designs to study the influence of the type of cut on the material shrinkage and deformation. The face centered–central composite design was used with two independent variables at three levels (air-drying temperature: 50°, 60°, and 70°C; and airflow: 1, 2, and 3 m/s) and triplicate central point. Three slices were dehydrated simultaneously in an experimental drying tunnel; this equipment has been described previously in another study (Campos and others 2007). The moisture content of the agave slices was measured using a thermobalance (model MB200; Ohaus, Pine Brook, NJ, USA). During drying, one slice was used to evaluate weight loss, by means of an analytic balance (Analytical Plus, Ohaus). The material was dried until it reached a constant weight. The effective diffusion coefficients of water removal were evaluated by means of Fick’s second law of diffusion for an infinite slab, based on a report by Saravacos (1995). Image Processing and Fractal Analysis Used to Study Shrinkage and Deformation Two charged-coupled device (CCD) digital cameras (Kodak DC290 zoom) were used to evaluate a second slice for macroscopic shrinkage and deforming evolution during convective drying; lateral and superior views were captured during the drying. Both cameras were connected to a serial port of a personal computer (Hewlett-Packard HP Compaq dc 5700 Microtower; Intel Pentium 4 CPU 3 GHz, 1 GB of RAM, and 80-Gb

Effects of Different Cut-Induced Microstructural and Macrostructural Arrays

621

hard drive) equipped with a video-recording card (All in Wonder Rage 128pro; ATI Technologies, Ottawa, KS, USA) and image-recording software (ATI version 6.2 Multimedia Center). Image processing and fractal analyses were performed with the ImageJ 1.34 software (National Institutes of Health, Bethesda, MD, USA) to evaluate the macrostructural changes in agave slices during drying. Images of both views of the slices were captured in RGB (red, green, blue) color and stored in bitmap format at 640 × 480 pixels. The image processing included cropping, turn-to-gray level, and thresholding to obtain binary images. The parameters extracted were the projected area (PA), major length (ML), and fractal dimension of the slice contour (FDC). These parameters were selected as indicators of nonisotropic shrinkage-deformation of agave slices during drying. The third slice was dried until it reached a constant weight, and its microstructural changes were observed by a scanning electron microscope (model JBM-5900LV; JEOL, Tokyo, Japan) (Bolin and Huxsoll 1987).

Results and Discussion Drying Kinetics In each experiment, the initial moisture content of agave fresh samples was 94.5% ± 0.5%. Figure 58.1a displays drying kinetics at 60°C and 2 m/s for two types of cut. Drying kinetics were obtained from calculated dimensionless moisture data vs drying time. Figure 58.1b shows the effect of the cut type on the effective diffusion coefficient (Deff) of water under several drying conditions. Figure 58.1 is presented to highlight the effect of the cut type on the moisture loss. Moisture loss rate and Deff were higher in the longitudinally cut, than the transversely cut slices (Figure 58.1). The drying time required to reach equilibrium moisture was less for the longitudinal cut (60 min) than for the transverse cut (100 min). This effect could be due to the differences in structural organization between the types of cut used. A transverse cut produces slices with short rigid structures, whereas a longitudinal cut produces long rigid structures. Because of these structural arrangements of the material, shrinkage and deformation patterns are different for each type of cut during the drying, and this could influence the moisture diffusion into material (Crossley and Aguilera 2001). Similar behavior was observed in the rest of the drying conditions. Shrinkage-Deformation Kinetics Figure 58.2 depicts a gallery of binary images of agave slices cut longitudinally and transversely at different convective drying times at 60°C and 2 m/s. PA, ML, and FDC changed during drying. The transversely cut slices displayed more shrinkage (77%) than did the longitudinally cut slices (48.4%) in the superior views. PA decreased 48.2% in the transverse cut and lateral view, but increased 83.4% in the longitudinal cut and lateral view because of drastic deformation of the material. ML in the transverse cut hardly decreased, but in the longitudinal cut it decreased smoothly in both views. The final irregularity of the contour measured as an FDC value was higher for the transverse cut than for the longitudinal cut in both views (Figure 58.2). The effects of

Figure 58.1. (a) Drying kinetics of agave slices at 60°C and 2 m/s for two types of cut. (b) Graphic tendencies of the experimental values of Deff as a function of the drying conditions and type of cut. X/X0, moisture content/initial moisture content.

622

(a)

LONGITUDINAL CUT

Superior view

Lateral view

Long rigid structures

PA = 10.84 ± 0.169

PA =1.27 ± 0.062

0 min

FDC =1.083 ± 0.002

FDC =1.022 ± 0.00

ML =3.80 ± 0.00

ML =3.80 ± 0.00 PA =2.0 ± 0.493

PA =6.93 ± 0.465 FDC =1.080 ± 0.002

FDC =1.058 ± 0.015

40 min

ML =3.36 ± 0.015

ML =3.50 ± 0.075

PA =2.33 ± 0.574 PA =5.60 ± 0.390

160 min

FDC =1.055 ± 0.022

FDC =1.089 ± 0.005

ML =3.25 ± 0.239

ML =3.50 ± 0.059

(b)

TRANSVERSE CUT

Superior view

Lateral view

Short rigid structures

PA = 11.06 ± 0.157

0 min

PA =1.45 ± 0.089

FDC =1.081 ± 0.001

FDC =1.048 ± 0.005

ML =3.80 ± 0.00

ML =3.80 ± 0.00

PA =4.16 ± 0.111

40 min

FDC = 1.116 ± 0.011

ML =2.45 ± 0.117

ML= 2.51 ± 0.081

PA =1.16 ± 0.192

PA =2.55 ± 0.218 FDC =1.137 ± 0.006 ML =2.09 ± 0.052

PA =1.24 ± 0.262 FDC =1.077 ± 0.011

160 min

FDC =1.097 ± 0.007 ML =1.97 ± 0.234

Figure 58.2. Gallery of binary images from superior and lateral views of agave slices during convective drying at 60°C and 2 m/s: (a) longitudinal cut and (b) transverse cut. PA, projected area; FDC, fractal dimension of the slice contour; and ML, major length.

623

624

PART 2: Poster Presentations

cut type on the shrinkage and deformation could be attributed to the different arrangement and length of the rigid structures in each type of slice (Fernàndez and others 2005). In the longitudinal slice, the long rigid structures avoid drastic shrinkage and deformation (Figure 58.2). In the transverse cut, the rigid structures are shorter and less resistant to structural changes or deformations because of the drying process, so this arrangement causes more shrinkage and deformation in all the directions of the cut. Microstructural Images Figure 58.3 shows scanning electron microscopic images of dried slices for both types of cut. Visual inspection of the agave cell structure showed that the transverse cut resulted in more extensive shrinkage, ruggedness, and deformation than did the longitudinal cut. Also, the transverse cut caused more damage to cell structure than did the longitudinal cut. The arrangement of rigid structures in the agave slices might influence the shrinkage and deformation processes, as well as the water transport within the material.

Figure 58.3. Scanning electron microscopic images of dried agave slices (drying conditions, 60°C and 2 m/s): (a and b) longitudinal cut, and (c and d) transverse cut.

Effects of Different Cut-Induced Microstructural and Macrostructural Arrays

625

Conclusions The highest levels of shrinkage and deformation, and lowest values of the effective diffusion coefficients of water obtained during the drying of agave slices were associated with short rigid structures in transversely cut material. Microstructural changes and cell collapse that occur during drying may increase the internal irregularity of the material, which could cause resistance to water transport. The results establish a guideline for selection of the best cut type, which may improve the drying process. Image processing and fractal analysis were helpful in evaluating the structural changes in agave slices during drying. The results are important in explaining the role of long rigid cellular structures in the shrinkage and deformation phenomena during convective drying.

Acknowledgments This work was sponsored by Secretaria de Investigación y Posgrado del Instituto Politécnico Nacional (SIP-IPN) (Secretary of Research and Graduate Institute of National Polytechnic) projects 20070631 and 20071011; Consejo Nacional de Ciencia y Tecnología (CONACYT) (National Council for Science and Technology) projects 59730 and 48061-Z; and Comisión de Operación y Fomento de Actividades Académicas (COFAA) (Committee on the Operation and Development of Academic Activities). C.G.-C. thanks CONACYT for her study grant.

References Bolin HR, Huxsoll CC. 1987. Scanning electron microscope/image analyzer determination of dimensional postharvest changes in fruit cells. J Food Sci 52:1649–98. Campos MR, Hernàndez SH, Chanona PJJ, Alamilla BL, Jiménez AA, Fito P, Gutiérrez LGF. 2007. Nonisotropic shrinkage and interfaces during convective drying of potato slabs within the frame of the systematic approach to food engineering systems (SAFES) methodology. J Food Eng 83:285–92. Crossley J, Aguilera JM. 2001. Modelling the effect of microstructure on food extraction. J Food Proc Eng 24:161–77. Fernàndez L, Castillero C, Aguilera JM. 2005. An application of image analysis to dehydration of apple discs. J Food Eng 67:185–93. Genskow LR. 1990. Considerations in drying consumer products. In: Mujumdar AS, Roques MA, editors. Drying ’89. New York: Hemisphere. p 38–45. Idarraga G, Ramos J, Zuñiga V, Sahin T, Young RA. 1999. Pulp and paper from blue agave waste from tequila production. J Agric Food Chem 47:4450–5. Saravacos GD. 1995. Mass transfer properties of foods. In: Rao MA, Rizvi SSH, editors. Engineering properties of food. 2nd ed. New York: Marcel Dekker. p 169–221. Susuki K, Kubota K, Hasegawa T, Hosaka H. 1976. Shrinkage in dehydration of root vegetables. J Food Sci 41:1189–93.

59 Study of White-Bread Structural Evolution by Means of Image Analysis and Associated Thermal History and Water-Loss Kinetics A. Pérez-Nieto, J. J. Chanona-Pérez, G. Calderón-Domínguez, R. Farrera-Rebollo, L. Alamilla-Beltrán, and G. F. Gutiérrez-López Abstract The objective of this work was to use image analysis to study structural changes in bread crumb during baking and to derive correlations with moisture loss and the thermal history of samples. Samples of white-bread dough were baked at 220°C, removed after a preset baking time, and stored frozen for analysis. Digital images were taken from samples representing every 50 s of the baking process for up to 1200 s. Bread-crumb grain features obtained from the images were related to temperature kinetics and moisture loss. Two stages were observed during baking that were related to the size evolution of the cell. The first stage, corresponding to a marked enlargement of loaves, was characterized by bubbles that were considerably larger than those visualized in the second stage, in which large bubbles coalesce and a homogeneous bread-crumb structure prevails. Water and ethanol escape during the second stage, thus reducing internal pressure. During the first stage, nonlinear moisture loss kinetics were observed that may be partially due to the formation of bread crumb from dough. During the second stage, moisture loss had a more linear tendency, perhaps because of a more homogeneous dehydration of the samples.

Introduction When a piece of dough is baked in a conventional oven, heat is transported from the surface to the center of the loaf, inducing the transformation of dough to bread. This transformation has been proposed to occur in three steps: (a) the expansion of the dough, (b) the transformation of the dough foam structure into an elastic crumb sponge structure, and (c) the setting of the sponge structure (Bloksma 1990; Scanlon and Zghal 2001; Chevallier and others 2002; Singh and Bhattacharya 2005; Sommier and others 2005). The dough expansion has been attributed to three processes that occur as temperature increases: release of carbon dioxide (CO2) (He and Hoseney 1991), growth of bubbles in the dough (Singh and Bhattacharya 2005), and reduction of dough density (Chevallier and others 2002). The transformation of dough foam structure into an elastic crumb sponge starts at 65°C. During this stage, starch gelatinization promotes an increase in the dough 627

628

PART 2: Poster Presentations

viscosity (He and Hoseney 1991). Also, the elasticity of the dough increases and its extensibility decreases, increasing gas pressure and tensile stress in the cells. Chevallier and others (2002) reported that during the second baking stage considerable heat is required for water evaporation. Nowadays, the study of dough crumb that develops during baking can be conducted through invasive (He and Hoseney 1991) and noninvasive procedures (Lagrain and others 2006). In both, image processing has been used to evaluate the crumb grain features, such as cell size, cell size distribution, cell number per unit area, cell wall thickness distribution, void fraction, and shape factor, among others (Sapirstein and others 1994; Crowley and others 2000). The aim of this work was to study the structural changes that occur in bread crumb during baking by means of digital image analysis correlating the resulting crumb grain features with the thermal and mass-loss kinetics of the samples.

Materials and Methods Ingredients Low-protein commercial wheat flour (100 g, 14% wet basis) without additives (9.4% protein N × 5.7, 0.56% ash, and 55.8% Farinographic absorption) was used. Table salt (1.50 g), sugar (6.09 g), powdered milk (4.05 g), instant yeast (0.84 g), and shortening (3.03 g) were commercial grade, as well. Dough Preparation Mixing was performed in a 300-g-capacity Farinograph (Brabender, Duisburg, Germany). Fermentation (30°C and 85% relative humidity) was divided into three steps (80, 45, 25 min) plus one proving step (55 min). A 40-g dough sample loaf size was selected after being compared with a 171.4-g bread-loaf sample (100 g of flour), wherein similar results for full baked samples were obtained. Baking Experiments Dough was baked at 220° ± 2°C in an electrical resistance oven (Henry Simon, Cheshire, UK). Sample loaves were removed from the oven every 50 s of baking time up to 400 s. After this period, samples were removed from the oven at 600, 900, and 1200 s. Immediately after removal, the sample loaves were tempered for 30 min at room temperature (20°–22°C) and then frozen to −18° ± 2°C (model CFM091K4W0; Kelvinator, Honea Path, SC, USA) for 24 h (Sapirstein and others 1994). All procedures were performed in triplicate. Temperatures Measurement Bread-dough temperatures were measured using J-type thermocouples, with 30gauge T/C wire attached to a scanning thermometer (Digi-Sense 92000-00; Eutech Instruments, Singapore). The thermocouples were fixed inside the dough sample at three different locations, as shown in Figure 59.1. Temperature was recorded every 10 s.

Study of White-Bread Structural Evolution by Means of Image Analysis

629

Figure 59.1. Temperature kinetics inside dough during baking at three different locations (1.6 [•], 3.3 [䊊], and 4.9 [䉬] cm from the bottom) and dough mass loss (䊏). db, dry basis; U, upper; C, center; L, lower; and Toven, oven temperature.

Mass-Loss Kinetics The mass loss of dough was measured in real time (Sommier and others 2005). The weight of the sample loaf was registered every 50 s during the first 400 s and then was measured at 400, 600, 900, and 1200 s. Initial moisture content of the dough was evaluated using the Bidwell method, AACC methods 40–50 [American Association of Cereal Chemists methods (AACC) 2000]. Mass loss was reported as moisture content (grams of water per grams of dry solid) at each time interval. Real-Time Bread-Height Measurements Loaf height was monitored by means of a video camera (Sony charged-coupled device [CCD] DCR-TRV120). Samples in the oven were illuminated by cold light via an optic fiber (endoscope; Karl Storz, Tuttlingen, Germany). The glass pan used for this baking test had the same dimensions as the aluminum pan mold. Image Processing A central slice of each frozen sample (7.5 mm thick) was used for image analysis. Images were captured by flatbed scanner (model 5000; BenQ, Irving, CA, USA) at a resolution of 300 dpi, real-color mode, and stored in a bitmap format. Images from the central slice of each loaf were cropped to a single rectangular field of view of 629 × 469 pixels. The preprocessing, segmentation, and crumb grain measurement

630

PART 2: Poster Presentations

were carried out using ImageJ 1.34 software (National Institutes of Health, Bethesda, MD, USA). Two crumb grain features are reported: mean cell area (square millimeter) and cell density (number of cells per square centimeter).

Results and Discussion Temperature Kinetics Figure 59.1 shows the change in dough temperature during baking (TB), measured at three different points inside the bread dough sample: U (upper), C (center), and L (lower). Dough temperature measured at location U increased faster compared to that at points L and C. The temperature profile at the three locations displayed a sigmoid behavior typical of heat transfer in the transient state and similar to internal temperature kinetics reported by Bloksma (1990). Temperature profiles in Figure 59.1 show the presence of two inflection points that, for the purpose of this work, have been named (a) cold inflection point and (b) hot inflection point. The cold inflection point, measured at location C, occurred around 200 s, when the dough was at a mean temperature of 37.1°C, with a heating rate of 2.6°C/min. During the same period, the temperature at U reached a value of 47.1° ± 7.25°C, with a heating rate of 12.4°C/min. A temperature difference of 10°C was found between points C and U. At L, the heating rate was of 4.6°C/min, while the temperature difference between the center and the lower dough measurement points reached 5.7°C. These results showed that the heating rate at U was higher than that at C and L. Similar results have been reported by Sommier and others (2005). The hot inflection point, measured at location C, was reached at 400 s at a dough temperature of 79.1° ± 2.77°C, with a heating rate of 9.6°C/min; at location U, the dough temperature was 86.2° ± 2.26°C, with a heating rate of 3.4°C/min; whereas at the lower dough measurement point (L) the results were similar to those obtained at C. Sommier and others (2005) reported a hot inflection point at 480 s with a maximum heating rate of 10°C/min and a center dough temperature of 89°C. The larger difference in temperature (ΔT = 24.6°C) between locations C and U was near 300 s of baking time; similar results have been reported by Sommier and others (2005) under similar experimental conditions. Mass-Loss Kinetics The mass-loss kinetics (Figure 59.1) enabled three baking stages to be identified. The first stage lasts from the onset of baking to 200 s, the second stage occurs between 200 and 400 s of baking, and the third stage ranges from 400 to 1200 s. The first stage showed a larger mass-loss rate than did the second stage, which may be associated with the evaporation of the water contained in the most external breaddough surface, as well as with CO2 release. The smallest change in mass loss, observed during the second period, may be related to the gelatinization and protein denaturation that occur during baking. These processes require water to proceed, making this water unavailable and consequently decreasing the mass-loss rate (He and Hoseney 1991).

Study of White-Bread Structural Evolution by Means of Image Analysis

631

The third stage lasted from 400 s to the end of baking (1200 s). During this period, the dough mass loss was the largest. This may be associated with the fast evaporation of the water contained in the dough center and in the pores of the dough/bread; similar results have been reported by Sommier and others (2005). Changes in Crumb During Baking The changes in dough/bread-crumb structure and center temperature, as function of baking time, are shown in Figure 59.2. This figure shows that bubbles coalesce during the first 100 s of baking (Figure 59.2b and c) while the temperature is still low (30°C). Below this temperature, the CO2, because of its greater volatility as compared with ethanol or water vapor, may be easily released from the dough, promoting bubble expansion and coalescence. As baking continues (150–200 s), the loaf height increases noticeably, 7.5 mm more than the nonbaked sample, and remains invariable during this heating period. At the same time, the largest bubble coalescence occurs (Figure 59.2d and e), resulting in a big hole inside the dough without a significant increase in dough center temperature. Ethanol, water vapor, and CO2 thermal expansion, as well as the CO2 desorption from the aqueous medium and the continuous generation of this gas by yeast, may increase

GL image

Baking time 0s

Temperature measurements

Baking time 300s

a

GL image g

Tc = 29.70 ± 0.2 50s

Tc = 48.80 ± 2.2 350s

b

h

Tc = 29.79 ± 0.2 100s

Tc = 64.85 ± 1.6 400s

c

i

Tc = 30.06 ± 0.2 150s

200s

250s

d

e

f

Tc = 79.12 ± 1.1 600s

Tc = 30.94 ± 0.4 900s Tc = 31.36 ± 0.6 1200s Tc = 37.45 ± 1.5

Temperature measurements

j

k

l

Tc = 91.18 ± 0.1

Tc = 92.32 ± 0.1

Tc = 92.79 ± 0.1

Figure 59.2. Dough/bread-crumb evolution and center temperature during baking. Tc, center temperature; and GL image, computer graphic library format for images.

632

PART 2: Poster Presentations

dough internal pressure, promoting the coalescence that enhances the big bubble formation and the increment in loaf height. All of these complex changes occur during the first stage of baking. During the second stage of baking (200–400 s), the big hole is filled by smaller bubbles with thinner walls (Figure 59.2f–i). This expansion phenomenon may be promoted by the dough’s higher temperature (>79°C), increasing the evaporation of the ethanol-water solution (Chevallier and others 2002). Throughout this period, the bread’s characteristic porous structure is produced. During the third stage of baking (400–1200 s), the dough surface grows golden brown, and the crumb loses moisture, as has been described in detail previously (Chevallier and others 2002; Singh and Bhattacharya 2005; Sommier and others 2005). During this stage, water evaporated faster than in the other two stages, and bubbles in the crumb did not show relevant changes in size or shape (Figure 59.2i–l). Image Processing Figure 59.3 displays the results obtained by image processing. Mean cell area (MCA) and cell density (CD) were plotted as functions of dough center temperature (Figure 59.3a), and MCA was also plotted as a function of mass loss (Figure 59.3b). The MCA maximum value and the CD minimum value were recorded at 33.2°C and 0.59 g water/g dry solid (200 s), as 62.0 mm2 and 0.13 cells/cm2, respectively (Figure 59.3a and b). These results correspond to the first stage of baking in which the big bubble is produced. During the second stage of baking (200–400 s), the larger pores are filled by smaller bubbles, resulting in greater CD (Figure 59.3a) while MCA decreases (Figure 59.3a and b). These results may be explained by the ethanol and water diffusion through the dough/bread structure; in fact, the water loss rate increased in the second and third stages as compared with the first stage. During the third stage of baking, MCA and CD remained invariable (Figure 59.3a and b), perhaps as a consequence of the water diffusion and the dough setting.

Conclusions Three stages of baking were confirmed by using relationships between physical properties measurements and dough/bread-crumb digital images. During the first stage of baking, different phenomena, such as cell expansion and coalescence, were observed, as well as cellular growth and the presence of a big bubble in the center of the dough/ bread. The second stage was characterized by a big bubble that was filled by smaller cells and an increase in the CD, possibly caused by ethanol and water diffusion. Finally, in third stage, the remainder of the water in the dough evaporated, and the final structure and color of the bread were fixed. The use of image processing and the measurement of physical properties were efficient for characterizing the evolution of the dough/bread-crumb structure and providing a more complete interpretation of the baking process.

Study of White-Bread Structural Evolution by Means of Image Analysis

(a)

80

633

80

60

60

40

40

20

20

0

Cellular density (No. cells/cm2)

Mean cell area (mm2)

Toven = 220 °C

0

20

30

40

50

60

70

80

90

100

Core temperature of dough/bread (°C) (b)

80

B 200s

Mean cell area (mm2)

60

40 Toven = 220 °C 20 1200s 0

0s

0.65

0.60

0.55

0.50

0.45

0.40

Dough moisture content (g water/g dry solid)

Figure 59.3. Image-processing results: (a) mean cell area (MCA) and cell density (CD) vs dough center temperature, and (b) MCA vs mass loss. Toven, oven temperature; circles, mean cell area; and diamonds, cellular density.

Acknowledgments This work was sponsored by Secretaria de Investigación y Posgrado del Instituto Politécnico Nacional (SIP-IPN) (Secretary of Research and Graduate Institute of National Polytechnic) projects 20070631 and 20070335, Consejo Nacional de Ciencia y Tecnología (CONACYT) (National Council for Science and Technology) project

634

PART 2: Poster Presentations

59730, and Comisión de Operación y Fomento de Actividades Académicas (COFAA) (Committee on the Operation and Development of Academic Activities).

References American Association of Cereal Chemists (AACC). 2000. Cereal laboratory methods. 10th ed. St Paul, MN: AACC. Bloksma AH. 1990. Dough structure, dough rheology, and baking quality. Cereal Foods World 35:237–44. Chevallier S, Dellavalle G, Colonna P, Broyart B, Trystram G. 2002. Structural and chemical modifications of short dough during baking. J Cereal Sci 35:1–10. Crowley P, Grau H, Arendt EK. 2000. Influence of additives and mixing time on crumb grain characteristics of wheat bread. Cereal Chem 77:370–5. He H, Hoseney RC. 1991. Gas retention in bread dough during baking. Cereal Chem 68:521–5. Lagrain B, Boeckx L, Wilderjans E, Delcour JA, Lauriks W. 2006. Non-contact ultrasound characterization of bread crumb: application of the Biot-Allard model. Food Res Int 39:1067–75. Sapirstein HD, Roller R, Bushuk W. 1994. Instrumental measurement of bread crumb grain by digital image analysis. Cereal Chem 71:383–91. Scanlon MG, Zghal MC. 2001. Bread properties and crumb structure. Food Res Int 34:841–64. Singh AP, Bhattacharya M. 2005. Development of dynamic modulus and cell opening of dough during baking. J Texture Stud 36:44–67. Sommier A, Chiron H, Colona P, Della Valle G, Rouille J. 2005. An instrumented pilot scale oven for the study of French bread baking. J Food Eng 69:97–106.

60 Effect of Hydrothermal Treatment on the Rheological Properties of HighAmylose Rice Starch P. Khunae, T. Tran, and P. Sirivongpisal

Abstract High-amylose rice starch from Chiang rice was subjected to hydrothermal treatment under different heat-moisture conditions: from 18% to 27% and at a temperature of 100°C for 16 h. All treated rice starches exhibited a combined hysteresis loop indicating a time-dependent shear-thinning and thixotropic behavior. The thixotropic behavior decreased at higher moisture levels of hydrothermal treatment. The pasting temperature of native rice starch increased following heat-moisture treatment (HMT). The pasting temperature increased linearly (R2 = 0.99) with increasing moisture levels of the hydrothermal treatment. A polynomial decrease (R2 = 0.99) was observed in setback with increasing levels of moisture.

Introduction Hydrothermal treatment of starch has been commonly used because it is considered natural and safe when compared with chemical modifications. Heat-moisture treatment (HMT), a type of hydrothermal treatment, involves treatment of starch granules at low moisture levels (<35% moisture, wt/wt) over a specific period (e.g., 15 min to 16 h) and at a specific temperature (84°–120°C) above the glass transition temperature (Tg) but below the gelatinization temperature. Under the foregoing conditions, changes in X-ray pattern, crystallinity, granule swelling, viscosity, gelatinization parameters, retrogradation, and acid and enzyme hydrolysis have been shown to occur in cereal tuber and legume starches (Hoover and Vassanthan 1994; Takaya and others 2000). The magnitude of these changes was found to depend on the moisture content during heat treatment and on the starch source. Relatively little is known about the impact of HMT on the rheological properties of high-amylose rice starch. Our work investigated the effects of HMT on the rheological properties of pasting and of flow behavior of high-amylose rice starch. It is hoped that the data generated from this study will provide efficient utilization strategies for the rice starch.

Materials and Methods Chiang rice was obtained from the Rice Research Center, Pattalung, Thailand. Starch from Chiang rice (20.16% amylose content) was prepared by the method reported by Sawai and Morita (1968). 635

636

PART 2: Poster Presentations

Heat-Moisture Treatment HMT was carried out using the method of Hoover and Manuel (1996). The moisture levels of rice-starch samples were increased to 18%, 21%, 24%, and 27%. The heatmoisture–treated rice starches were then sealed in cans and placed in a hot-air oven at 100°C for 16 h. Afterward, the treated samples were dried at 45°C to uniform moisture content (∼12%). Flow Behavior The heat-moisture–treated rice-starch slurry (4%, dry basis) was heated at 95°C in boiling water for 15 min with continuous stirring. The flow behavior of the gelatinized rice starch was determined by using a rotational rheometer (Haake RheoStress RS75; Thermo Scientific, Karlsruhe, Germany) equipped with coaxial cylinder geometry (Z41). The flow behavior of rice-starch paste was determined at 60° ± 0.1°C by increasing the shear rate to the range of 0–300 s−1. Pasting Properties The pasting properties of heat-moisture–treated rice starches were determined with a viscograph (Brabender, Duisburg, Germany). Starch paste (6%, wt/wt) was heated from 50° to 95°C, held at that temperature for 15 min, cooled to 50°C, and held at that temperature for 30 min.

Results and Discussion Flow Behavior After HMT, rice-starch pastes exhibited shear-thinning behavior (Figure 60.1) that followed the power-law equation (R2 = 0.99, P < 0.05). The consistency coefficient (K) of treated rice starches decreased, whereas the flow behavior index (n) increased when compared with native rice starch. The treatments increased the moisture level, causing a decrease in K but an increase in n following a polynomial (R2 = 0.98, P < 0.05) for rice starches, denoting a slight shear thinning (Figure 60.2). Rather than simple shear-thinning behavior, treated rice starches exhibited a combined hysteresis loop indicating a time-dependent shear-thinning and thixotropic behavior (Figure 60.3). The thixotropic behavior decreased at higher moisture levels of treatment. Pasting Properties After HMT, the pasting temperature of native high-amylose rice starch increased. The pasting temperature of treated rice starch increased linearly (R2 = 0.99, P < 0.05) with the increasing moisture level of the treatments (Figure 60.4). Polynomial reduction in setback (R2 = 0.99, P < 0.05) of rice starches with increasing moisture level of the treatments was observed (Figure 60.4). Hot-paste viscosity (Hv) and cold-paste viscosity (Cv) of rice starch correlated through a second-order polynomial (R2 = 0.99, P < 0.05) with treatment moisture level.

0.80 Native HMT18 HMT21

Viscosity (Pa · s)

0.60

HMT24 HMT27

0.40

0.20

0.00 1

10

100

1000

Shear rate (1/s) Figure 60.1. Relationship of apparent viscosity and shear rate of native rice starch and heat-moisture–treated rice starch at 18% (HMT18), 21% (HMT21), 24% (HMT24), and 27% (HMT27) moisture content over the shear rate range of 0–300 s−1 and at 60°C.

1.00 0.95

K

0.75

0.60

n

K (Pa · sn)

0.80

n

0.40 0.55

0.20

0.00

0.35

17

19

21

23

25

27

29

Moisture content (%)

Figure 60.2. Change in consistency index (K) and flow behavior index (n) of heatmoisture–treated rice starch at various moisture levels of the treatment.

637

25 Native

Shear stress (Pa)

20

15 HMT18 HMT21

10

HMT24 5

HMT27

0 0

100

200

300

400

Shear rate (1/s) Figure 60.3. Relationship of shear stress and shear rate of native rice starch and heatmoisture–treated rice starch at 18% (HMT18), 21% (HMT21), 24% (HMT24), and 27% (HMT27) moisture content over the shear rate range of 0–300 s−1 and at 60°C.

180

Setback

o

Pasting temperature

85

120

80

60

75 17

Setback viscosity (BU)

Pasting temperature ( C)

90

0 20

23

26

29

Moisture content (%) Figure 60.4. Relationship of pasting temperature and setback to moisture content of treated rice starch. BU, Brabender unit.

638

Effect of HTT on the Rheological Properties of High-Amylose Rice Starch

639

Conclusion After HMT, moisture level of the treatments increased and the pasting temperature of the high-amylose rice starches increased linearly. However, with increasing moisture level of the treatments, a polynomial reduction in setback of rice starch was observed. Heat-moisture–treated rice-starch pastes showed shear-thinning behavior following the power-law model. The consistency coefficient of rice starches decreased with the increasing moisture level of the treatments. Rather than solely shear-thinning behavior, all treated rice starches exhibited a combined hysteresis loop indicating a timedependent shear-thinning and thixotropic behavior. The thixotropic behavior decreased at the higher moisture levels of treatment.

Acknowledgments The authors express their sincere thanks to the Thailand Research Fund (TRF) for master research grants, and the Faculty of Agro-Industry, Prince of Songkla University, Songkla, Thailand, for research fund support.

References Hoover R, Manuel H. 1996. The effect of heat-moisture treatment on the structure and physicochemical properties of legume starches. Food Res Int 29:731–50. Hoover R, Vassanthan T. 1994. Effect of heat-moisture treatment on the structure and physicochemical properties of cereal, tuber, and legume starches. Carbohydr Res 252:33–53. Sawai H, Morita Y. 1968. Studies on rice glutelin. Part I: Isolation and purification of glutelin from rice endosperm. Agric Boil Chem 32:76–80. Takaya T, Sano C, Nishiami K. 2000. Thermal studies on the gelatinization and retrogradation of heatmoisture treated starches. Carbohydr Polym 41:97–100.

61 Influence of Glass Transition on Oxygen Permeability of Starch-Based Edible Films D. Thirathumthavorn, S. Charoenrein, and J. M. Krochta

Abstract Starches from different sources have distinctly different molecular weights. In addition, the high temperatures and shear stresses involved in starch-film solution casting, compression molding, and extrusion reduce the starch molecule chain length. This research investigated the effect of starch chain length on oxygen permeability (OP) and glass transition temperature (Tg) of starch films. Solution-cast starch films made from tapioca-based and rice-based starches having different chain lengths were used in this study. Starch, noncrystallizing sorbitol (the partial hydrolysis product of starch with subsequent hydrogenation under pressure), and deionized water were mixed, heated to 95°C and held for 30 min to gelatinize the starch, cast on high-density polyethylene plates, and dried at ambient conditions to form films. Results showed that both tapioca-starch and rice-starch chain length affected OP. Films made from longchain starch had lower OP than did those made from short-chain starch. The Tg of starch tended to decrease with decreasing degree of polymerization. Shorter-chain starch had lower Tg, which correlated with an increase in the OP of starch films. From these experiments, it can be concluded that the Tg of starch could be used to explain the change in OP of starch films related to starch chain length.

Introduction Starch films and coatings are good oxygen barriers (Mark and others 1966; Roth and Mehltretter 1967). Edible oxygen-barrier films can be used to protect foods that are susceptible to oxidation (rancidity, loss of oxidizable vitamins, etc.). Film glass transition temperature (Tg) values were also found to be inconsistent with oxygen permeability (OP) properties (Sothornvit and others 2002). The permeation of gas and vapor molecules through a film is higher above the Tg where polymer chains are more mobile and thereby enhance the diffusion and permeability of gases (McHugh and Krochta 1994; Cherian and others 1995; Biliaderis and others 1999). Above the Tg, polymeric materials exist in a soft, rubbery state, which impairs barrier properties; whereas, below the Tg, polymers assume a glassy, low-permeable state (Anker and others 1999). This research investigated the effect of starch chain length on OP and Tg of starch films. 641

642

PART 2: Poster Presentations

Materials and Methods Materials Tapioca starch was from Tapioca Development (Bangkok, Thailand). Rice starch was from Choheng (Nakon, Prathom, Thailand). Noncrystallizing sorbitol was provided by SPI Pharma (New Castle, DE, USA). Methods Preparation of Starch with Various Chain Lengths Starch was treated with 1 N HCl, following the method of Thirathumthavorn and Charoenrein (2005). Preparation of Starch Films Starch suspension was prepared at a concentration of 2%–5% wt/wt in deionized water. The amount of plasticizer was 40% wt/wt of total solid. The starch suspension and plasticizer were mixed and heated up to 95° ± 5°C for 30 min to gelatinize the starch, degassed by vacuum, and cast on high-density polyethylene plates. All casting plates were on a level surface at ambient conditions (23° ± 3°C and 35% ± 5% relative humidity [RH]) until the films could be peeled from the plates. All films were stored in a controlled cabinet (RH, 50% ± 5%; and temperature, 23° ± 2°C) for 3 days prior to testing. Three replications were done in this experiment. Oxygen Permeability OP of starch films was measured at 23°C and 50% ± 1% RH by using an Ox-tran 2/20 ML modular system (Modern Controls, Minneapolis, MN, USA) according to the American Society of Testing and Materials Standard Methods D3985 (ASTM 1995). Starch films were conditioned and stored in the chamber at 23° ± 2°C and 50% ± 5% RH by using magnesium nitrate salt. Glass Transition Temperature A Pyris 1 differential scanning calorimeter (Perkin-Elmer, Norwalk, CT, USA) was used for the measurement of Tg. Starch suspension was heated to 95°C for 30 min and then cast on plastic plates and dried at ambient conditions. The samples were conditioned at 50% RH in the chamber containing saturated magnesium nitrate. Samples of 30–35 mg were weighed in stainless-steel pans and hermetically sealed; an empty pan was used as a reference. Samples were heated from −60° to 160°C at a 10°C/min heating rate. The inflection point of the step in the heat-flow curve was taken as the Tg.

Results and Discussions Oxygen Permeability The transfer of oxygen to food from the environment has an important effect on food quality and shelf life. Edible films and coatings can prevent deterioration of many food products because the films often possess excellent oxygen-barrier properties

Influence of Glass Transition on Oxygen Permeability of Starch-Based Edible Films

643

Table 61.1. Oxygen permeability of 40% noncrystallizing sorbitol–plasticized starch films at various degrees of polymerization tested at 65% relative humidity after storage at 50% ± 5% relative humidity and 23° ± 2°C for 3 days Samples Native rice-starch–based film (DP 1630)

Oxygen permeability (mL μm/m2 · day kPa) 11.89 ± 0.30NS

Acid-treated rice-starch–based films having DP 1200

12.67 ± 0.32

DP 770

12.35 ± 0.50

Native tapioca-starch–based film (DP 6210) Acid-treated tapioca-starch–based films having

9.73 ± 0.33b

DP 3080

9.79 ± 0.64b

DP 1340

11.54 ± 0.63a

DP 620

12.65 ± 0.99a

a, b

The same letter in this column of each starch type indicates that the oxygen permeability showed no significant difference (P > 0.05). DP, degree of polymerization; and NS, no significant difference.

(Sothornvit and Krochta 2000). The OP of starch films was tested at 65% RH (Table 61.1) because, at 50% RH used for conditioning the films and testing tensile properties, the OP was below the limit of detection (0.05 cm3 μm/m2 day kPa) of the Ox-Tran 2/20 ML. There was no significant difference in the OP of rice-starch film with degree of polymerization (DP) in the range of 770–1630 (P > 0.05). In contrast, the oxygen permeability values of tapioca-starch–based films significantly increased with decreasing DP (P < 0.05). The oxygen permeability values of rice-starch–based and tapiocastarch–based films at a similar DP were comparable. The barrier properties of polymer films are generally related to the physical and chemical nature of the polymers, perhaps because the shorter-chain molecules of starch increase the flexibility in amorphous parts of starch films. The shorter-chain molecules of starch decreased the Tg (Figure 61.1), which is similar to the data reported by Shamekh and others (2002). The permeation of gas and vapor molecules through a film is higher above the Tg, where polymer chains are more mobile, and thereby enhance the diffusion and permeability of gases (McHugh and Krochta 1994; Cherian and others 1995; Biliaderis and others 1999). Glass Transition Temperature The Tgs of plasticized starch films were normally observed in both the upper transition and the lower transition. The upper transition is due to a starch-rich phase, whereas the lower transition is a starch-poor phase or plasticizer-rich area (Forssell and others 1997). The upper transition of 40% noncrystallizing sorbitol–plasticized starch films cannot be determined because the signal is too low for the differential scanning calorimeter to detect the transition. Forssell and others (1997) reported that the Tg at the upper transition can be observed at water contents of less than 6% for

644

PART 2: Poster Presentations

100

Tg (°C)

98 96 Tapioca Rice

94 92 90 0

2000

4000

6000

8000

Degree of polymerization

Figure 61.1. Glass transition temperatures (Tg) of starch films having different degrees of polymerization.

the 14% glycerol-content mixtures and 3% for the 20% glycerol-content mixtures. In the absence of plasticizer, both amylose and amylopectin showed only the upper glass transition (Myllärinen and others 2002). To observe the effect of starch chain length on the Tgs of starch films, starch films containing no plasticizer were tested. Without any plasticizers, the Tg of starch film tended to decrease (Figure 61.1) with decreasing DP. The results were in accordance with those reported by Shamekh and others (2002), who studied the Tgs of hydrolyzed potato starch. At a similar DP, the Tgs of native rice starch and acid-treated tapioca starch (DP 1340) were comparable. These results suggest that the molecular weight of starch influences the Tgs of nonplasticized starch films.

Conclusion The OP of starch films was affected by starch chain length. Films made from longchain starch had lower OP but higher Tg than did those made from short-chain starch. OP thus decreases as Tg increases.

References American Society for Testing and Materials (ASTM). 1995. Annual book of American standard testing methods. Philadelphia: ASTM. Anker M, Stading M, Hermansson AM. 1999. Effects of pH and the gel state on the mechanical properties, moisture contents, and glass transition temperatures of whey protein films. J Agric Food Chem 47:1878–86. Biliaderis CG, Lazaridou A, Arvanitoyannis I. 1999. Glass transition and physical properties of polyolplasticized pullulan-starch blends at low moisture. Carbohydr Polym 40:29–47. Cherian G, Gennadios A, Weller C, Chinachoti P. 1995. Thermomechanical behavior of wheat gluten films: effect of sucrose, glycerin and sorbitol. Cereal Chem 72:1–6.

Influence of Glass Transition on Oxygen Permeability of Starch-Based Edible Films

645

Forssell PM, Mikkilä JM, Moates GK, Parker R. 1997. Phase and glass transition behavior of concentrated barley starch–glycerol–water mixtures, a model for thermoplastic starch. Carbohydr Polym 34:275–82. Mark AM, Roth WB, Mehltretter CL, Rist CE. 1966. Oxygen permeability of amylomaize starch films. Food Technol 20:75–7. McHugh TH, Krochta JM. 1994. Permeability properties of edible films. In: Krochta JM, Baldwin EA, Nisperos-Carriedo MO, editors. Edible coatings and films to improve food quality. Lancaster, PA: Technomic. p 139–87. Myllärinen P, Buleon A, Lahtinen R, Forssell P. 2002. The crystallinity of amylose and amylopectin films. Carbohydr Polym 48:41–8. Roth WB, Mehltretter CL. 1967. Some properties of hydroxypropylated amylomaize starch films. Food Technol 21:72–4. Shamekh S, Myllärinen P, Poutanen K, Forssell P. 2002. Film formation properties of potato starch. Starch/ Stärke 54:20–4. Sothornvit R, Krochta JM. 2000. Plasticizer effect on oxygen permeability of β-lactoglobulin films. J Agric Food Chem 48:6298–302. Sothornvit R, Reid D, Krochta JM. 2002. Plasticizer effect on the glass transition temperature of betalactoglobulin films. Trans Am Soc Agric Eng 45:1479–84. Thirathumthavorn D, Charoenrein S. 2005. Thermal and pasting properties of acid-treated rice starches. Starch/Stärke 57:217–22.

62 Molecular Mobility and Seed Longevity in Chenopodium quinoa M. Castellión, S. Maldonado, and M. P. Buera

Abstract This work analyzed the relationship between water mobility and quinoa seed longevity. Transverse relaxation times (T2s) by time-resolved nuclear magnetic resonance (NMR) of four genotypes of quinoa seeds (cultivars Chadmo, NL-6, Sajama, and 2-Want) equilibrated at relative humidities (RHs) in the range of 22%–83% were measured. Quinoa seed germination and viability were tested. The curves obtained from seeds equilibrated at low RHs were fitted best to monoexponential decay, with T2 values in the range of 100–130 ms. The amplitude of the signal was related to the relative amount of lipid and was unaffected by RH. At ∼50% RH, curves began to fit biexponential decay best. The second term of the biexponential curves showed T2 values in the range of 1.13–2.54 ms, increasing as RH increased. The analysis of isotherm shapes and glass transition temperatures of quinoa seeds revealed that the second term could be attributed to protons both of water in multilayers and of plasticized solids above the glass transition temperature. Germination tests showed different storage behaviors between cultivars. Water mobility alone did not explain quinoa seed longevity after storage under these experimental conditions.

Introduction Temperature and moisture play a fundamental role in the storage longevity of seeds. The pattern of seed aging is described in terms of its water content (WC) during storage (Walters 1998). The different properties of water within the seeds, in terms of bound water, water activity (aw), water potential, and/or the presence of glassy aqueous phases, have been widely recognized as factors that determine the mechanisms and kinetics of seed longevity (Vertucci 1992; Williams 1994). The motion and thermal properties of the intracellular viscosity that is typical of glassy phases severely slows molecular diffusion; thereby decreasing the probability of chemical diffusion. The glassy states of water in seeds have been used to describe the temperature and WC at which seeds can be preserved (Sun and Leopold 1993). Time-resolved nuclear magnetic resonance (NMR) spectroscopy has been used to study water status and metabolic changes in many biological systems. The transverse relaxation behavior of water protons has been used to describe compartmentation of 647

648

PART 2: Poster Presentations

water in tissues (Ratcliffe 1994). Mobile and less-mobile water molecules can be distinguished by their different relaxation rates, and their relative amounts can be calculated (Krishnan and others 2003, 2004). Compared with gravimetric methods, NMR techniques can provide more detailed information on the amount and state of water in seeds at different WCs and relative humidities (RHs) during storage (Krishnan and others 2004). As Chenopodium quinoa is a species that has adapted well to extreme environmental conditions with regard to altitude, amount of annual precipitation, soil salinity, and minimum temperatures (Tapia 1999), its different genotypes seem to constitute appropriate material to study in determining the role of water status and metabolic changes in seed longevity. This work analyzed the relationship between water mobility and seed longevity in quinoa seeds that are originally from contrasting environments. Differential storage behavior had been preliminarily determined in two other quinoa genotypes of contrasting origin. Transverse relaxation times (T2s) of four genotypes of quinoa seeds (cultivars [cv.] Chadmo, NL-6, Sajama, and 2-Want) and equilibrated at different RHs were measured by time-resolved NMR using the Carr-Purcell-Meiboom-Gill (CPMG) method. Quinoa seeds were also subjected to storage at different RHs, and their germination and viability were tested.

Materials and Methods Seed Material Quinoa seeds of four genotypes were obtained at experimental greenhouses of the University of Buenos Aires. Chadmo belongs to the sea level adaptation group (Bertero and others 2004) and has adapted to humid conditions (800–1500 mm of annual rain), fertile soils, and temperatures above 5°C. The NL-6 genotype was obtained from the University of Wageningen (Wageningen, The Netherlands) from genotypes originally from the sea-level area. Sajama and 2-Want belong to the Altiplano adaptation group (Bertero and others 2004) and have adapted to very arid conditions characteristic of the salty soils of the Chilean Altiplano with less than 250 mm of annual rain and a minimum temperature of −1°C. Seed Storage Behavior Seed Storage Seeds were equilibrated at 43% or 75% RH and stored at 32° ± 2°C in vials. Samples were withdrawn regularly for seed testing. Seeds in all experiments were presorted by hand; excessively small or damaged seeds were discarded. Germination Assessment Briefly, four replicates of 50 seeds each were preincubated in 100% RH at 5°C for 3 h to avoid imbibitional damage, surface disinfected in a sodium hypochlorite solution (5.5 g Cl · L−1) for 10 min, washed thoroughly with distilled water, and set to

Molecular Mobility and Seed Longevity in Chenopodium quinoa

649

germinate over wet paper at 24° ± 1°C with a 12-h photoperiod for 3 days. Germination was evaluated according to International Seed Testing Association (2005) rules.

Viability Test The tetrazolium test for viability was conducted on stored quinoa seeds. Seeds were embedded in a 1% 2,3,5-triphenyl-2H-tetrazolium chloride (Merck, Rahway, NY, USA) solution and incubated at 25°C for 24 h. Results were evaluated according to International Seed Testing Association (2005) rules.

NMR Relaxation Measurements Seed Conditioning Seed samples were incubated in duplicate at RHs in the range of 22%–83% at 27° ± 1°C. To achieve different WCs, the saturated salt solutions (analytic grade; Mallinckrodt, Hazelwood, MO, USA) used were lithium chloride (11% RH); magnesium chloride (33% RH); potassium carbonate (43% RH); sodium bromide (63% RH); sodium chloride (75% RH); potassium chloride (87% RH); and potassium nitrate (96% RH) (Greenspan 1977). The aw of the samples was measured after 4 weeks by using an Aqualab aw meter (Decagon Devices, Pullman, WA, USA). Spin-Spin Relaxation Time (T2) Measurement Seed samples were placed in 10-mm-diameter, 4-cm-high tubes, sealed, and positioned in the probe of a Bruker NMS 120 pulsed NMR spectrometer (20 MHz) (Bruker AXS, Madison, WI, USA). T2 was measured by the CPMG method. The settings for data points (256), pulse separation (0.3 ms), dummy echoes (1), scans (16), and gain were maintained for all measurements. The gain was adjusted to maximize the signalto-noise ratio. The relaxation curves were analyzed to determine the parameters of the fitting equations (Equations 62.1 and 62.2). The distribution of these parameters was determined as a function of RH. −t

M(t ) = A T2 −t

(62.1) −t

M(t ) = AaT2 a + AbT2b

(62.2)

Sorption Isotherms After NMR measurements, the WC of seed samples was determined gravimetrically by using a Mettler Toledo Analytical Balance (Mettler Toledo, Greifensee, Switzerland). Seeds were placed in vials and dried in an oven at 97°C for 3 days. The isotherms measured were fitted to the Guggenheim-Anderson-de Boer (GAB) equa-

650

PART 2: Poster Presentations

tion (Equation 62.3) by using the least-squares method for minimizing the absolute differences between measured and calculated moisture content. m=

m0 Ckaw

(62.3)

(1 − kaw ) (1 + (C − 1) kaw )

Results Quinoa seeds stored at 43% RH (8.6%–9.6% WC) maintained high germination and viability values (Figures 62.1 and 62.2). Following 14 weeks of storage at 75% RH (12.6%–14.3% WC), seeds from cv. Chadmo showed high viability and germination values (67% and 78%, respectively). On the other hand, those values for cultivars Sajama, NL-6, and 2-Want dropped to between 0 and 30%. Quinoa seed moisture-sorption isotherms were similar among the four genotypes. The GAB parameters were found to be comparable among seeds of different genetic origin (Table 62.1). Monolayer moisture contents (m0s) were in the range of 6.234%– 6.927% (dry basis). The obtained k and C values were similar to those reported by Matiacevich and others (2006). Monoexponential and biexponential decay curves were proved to fit experimental data. The curves obtained from seeds equilibrated at low RHs were fitted best to monoexponential decay, with T2 values in the range 100–130 ms (Figure 62.3). The amplitude of the signal was 11.5–14.6 (cv. Chadmo), 16.3–21.6 (cv. NL-6), 22.1–30.5 (cv. Sajama), and 24.5–30.1 (cv. 2-Want) (Figure 62.4). The amplitude was related

Chadm 43%RH

NL-6 43%RH

Sajama 43%RH

2-Want 43%RH

Chadm 75%RH

NL-6 75%RH

Sajama 75%RH

2-Want 75%RH

100% 90%

Germination

80% 70% 60% 50% 40% 30% 20% 10% 0% 0

2

4

6

8 10 Storage time (weeks)

12

14

16

18

Figure 62.1. Percent germination of four quinoa genotypes during storage at 32°C and 43% or 75% relative humidity (RH). Cultivar Chadmo maintained high germination values after storage at 75% RH.

Molecular Mobility and Seed Longevity in Chenopodium quinoa

651

Chadm 43%RH

NL-6 43%RH

Sajama 43%RH

2-Want 43%RH

Chadm 75%RH

NL-6 75%RH

Sajama 75%RH

2-Want 75%RH

100% 90% 80%

Viability

70% 60% 50% 40% 30% 20% 10% 0% 0

2

4

6

8 10 12 Storage time (weeks)

14

16

18

Figure 62.2. Percent viability of four quinoa genotypes assessed by tetrazolium test. Cultivar Chadmo showed high viability values after storage at 75% relative humidity (RH).

Table 62.1. Sorption isotherms of four genotypes of quinoa seedsa m0 (% dry basis)b

Cc

kd

%Ee

Chadmo

6.927 ± 0.928

10.92 ± 6.57

0.8101 ± 0.0462

2.6069

NL-6

6.269 ± 0.706

13.24 ± 8.18

0.8178 ± 0.0378

2.7987

Sajama

6.329 ± 0.684

10.33 ± 5.36

0.8485 ± 0.0336

1.8618

2-Want

6.234 ± 0.511

11.32 ± 4.6

0.8332 ± 0.0266

2.1571

Genotypes

a

Guggenheim-Anderson-de Boer (GAB) parameters were similar among different genotypes. Moisture content needed to cover the entire surface with a unimolecular layer. c Factor-correcting properties of the multilayer molecules with respect to the bulk liquid. d Guggenheim constant. e Relative percent deviation modulus (E). b

to the relative amount of lipid and was unaffected by RH. Protons of the less mobile water molecules could not be detected by the spin-echo method used. At ∼50% RH, curves began to fit biexponential decay best. The parameters obtained from monoexponential curves detected remained almost constant in the biexponential curves at RH values in the range of 50%–83% (Figures 62.3 and 62.4). The second term of the biexponential curves showed T2 values in the range of 1.13–2.54 ms, increasing as RH increased (Figure 62.5). The amplitudes ranged from 15.6–76.0, 12.2–76.1, 11.4–84.1, and 11.3–80.1 for the genotypes Chadmo, NL-6, Sajama, and 2-Want, respectively (Figure 62.4).

Figure 62.3. Relaxation times (T2s) of the lipid component of quinoa seeds of four genotypes with different water activities. T2 values were not affected by water activity.

Figure 62.4. Amplitude of the signal of the lipid component of quinoa seeds of four genotypes with different water activities. Values remained unaffected by increasing water activity suggesting this component can be attributed to lipids.

652

Figure 62.5. Relaxation times (T2s) of the water component of quinoa seeds of four genotypes with different water activities. T2 values increased as water activity increased, indicating higher water-proton mobility.

Figure 62.6. Amplitude of the signal of the water component of quinoa seeds of four genotypes with different water activities. Values increased as water activity increased due to the increasing relative amount of water.

653

654

PART 2: Poster Presentations

Discussion Different quinoa genotypes showed differential storage behavior. Germination and viability tests proved the foremost tolerance to storage was at 75% RH in seeds of cv. Chadmo. Water in multilayers was detected in quinoa seeds from ∼10% WC. At aw of ∼0.50, curves began to fit biexponential decay best. T2 values, as well as their corresponding amplitudes, increased as WC increased, indicating higher water-proton mobility and WC, respectively. At lower aw, T2 for this water population was not present or could not be detected by this method. Freezable water was not detected in the range analyzed, because the data could not be fitted to triexponential decay curves. These results are consistent with the analysis of isotherm shapes and glass transition temperatures of quinoa seeds (Matiacevich and others 2006). Protons of the less mobile water molecules could not be detected by the spin-echo method used. Regarding genotypic differences, cv. Chadmo showed lower relaxation times at high aw, indicating lower water-proton mobility. A lipid component was identified since its parameters held almost constant as WC increased. The curves obtained from seeds equilibrated at low RHs were fitted best to monoexponential decay. The parameters obtained by analysis of monoexponential curves remained almost constant for the biexponential curves at RHs in the range 50%–83%. The amplitude of the signal was related to the relative amount of lipid present in the seeds of each quinoa genotype and was unaffected by RH. The findings of our study on different quinoa genotypes support the conclusion that Krishnan and others (2003) developed with respect to wheat and soybean seeds. Genotypes showing higher T2 values were more sensitive to high moisture content during storage. Thus, the information gained from proton spin relaxation times can be useful in predicting quinoa seed longevity. Contrary to expectations, results for the four genotypes did not correlate with their differing adaptation group.

References Bertero D, de la Vega AJ, Correa G, Jacobsen S, Mujica A. 2004. Genotype and genotype-by-environment interaction effects for grain yield and grain size of quinoa (Chenopodium quinoa Willd.) as revealed by pattern analysis of international multi-environment trials. Field Crops Res 89:299–318. Greenspan L. 1977. Humidify fixed points of binary saturated aqueous solutions. J Res Natl Bur Stand [A] 81:89–96. International Seed Testing Association (ISTA). 2005. International rules for seed testing. Bassersdorf, Switzerland: ISTA. Krishnan P, Joshi DK, Maheswari M, Nagarajan S, Moharir AV. 2004. Characterisation of soybean and wheat seeds by nuclear magnetic resonance spectroscopy. Biol Plant 48:117–20. Krishnan P, Nagarajan S, Dadlani M, Moharir V. 2003. Characterisation of wheat (Triticum aestivum) and soybean (Glycine max) seeds under accelerated ageing conditions by proton nuclear magnetic spectroscopy. Seed Sci Technol 31:541–50. Matiacevich S, Castellión M, Maldonado S, Buera P. 2006. Water-dependent thermal transitions in quinoa embryos. Thermochim Acta 448:117–22. Ratcliffe RG. 1994. In vivo NMR studies of higher plants and algae. Adv Bot Res 20:44–123.

Molecular Mobility and Seed Longevity in Chenopodium quinoa

655

Sun WQ, Leopold AC. 1993. The glassy state and accelerated ageing of soybeans. Physiol Plant 89:767–74. Tapia ME 1999. Agro-ecological zoning of the cultivation of quinoa (Chenopodium quinoa Willd.). In: Jacobsen SE, Portillo Z, editors. First international workshop on quinoa genetic resources and production systems. Lima, Peru: Regional Office for Latin America and the Caribbean, Food and Agriculture Organization of the United Nations, Santiago, Chile. Available from: http://www.rlc.fao.org/es/ agricultura/produ/cdrom/contenido/libro14/cap1.2.htm. Vertucci C. 1992. A calorimetric study of the changes in lipids during seed storage under dry conditions. Plant Physiol 99:310–6. Walters C. 1998. Understanding the mechanisms and kinetics of seed ageing. Seed Sci Res 8:223–44. Williams RJ. 1994. Methods for determination of glass transitions in seeds. Ann Bot 74:525–30.

63 Analyzing the Effect of Freeze-Thaw Cycle on the Off-Aroma of Pineapple by Using an Electronic Nose Technique S. Charoenrein and T. Kaewtathip

Abstract Aroma is one of the major determinants of fruit quality but can be easily modified or even greatly reduced during processing. This study used an electronic nose (e-nose) technique to investigate the aroma changes in pineapple during freezing and thawing. Two pineapple cultivars (Smooth Cayenne and Queen) were used to verify the technique. The results demonstrated that e-nose could clearly distinguish the fresh and frozen-thawed (freeze-thaw) pineapples. The more freeze-thaw cycles applied, the greater the difference from fresh sample observed. Also, while the aroma of the two pineapple cultivars could be distinguished by means of e-nose data, a similar aroma classification was found for both cultivars. The off-aroma scores for freeze-thaw pineapple from sensory evaluation corresponded with the e-nose data. This is the first study on the use of an e-nose to detect aroma changes in frozen fruit.

Introduction Aroma is one of the major determinants of fruit quality but can be easily modified or even greatly reduced during processing (Torreggiani and Maestrelli 2006). Our previous studies indicated that lower sensory scores in terms of flavor were obtained for frozen pineapple than for fresh pineapple (Wannawilas and others 2004). The main components of the overall sensation of flavor are taste and aroma (Salunkhe and others 1991). Soluble and volatile constituents in food reach receptors on the tongue and in the upper nasal cavities and trigger responses perceived as flavor. Fruit flavors are often assessed by sensory analysis in which the actual human senses are used (Skrede 1995). Problems associated with this method of sensory evaluation, especially in odor evaluation, include the standardization of measurements, the correctness of training, and the stability and the reproducibility of the evaluation (Berna and others 2004). Analytic techniques like gas chromatography and mass spectrometry, based on chemical identification and quantification of the compounds present in fruits, are also used. If the chemical identities for aroma or odor have been determined, gas chromatography is very accurate and useful for both quantitative and qualitative analyses. Otherwise, interpretation of chromatograms of accurately represented odor-active components is a very challenging problem (Grate and Abraham 1991). 657

658

PART 2: Poster Presentations

In recent years, sensor array–based aroma analysis technology has been developed that complements human sensory analysis. This technology, the so-called electronic nose (e-nose), uses an instrument that comprises an array of electronic chemical sensors with partial specificity for measurement of volatiles and an appropriate pattern-recognition system capable of recognizing simple or complex odors (Gardner and Bartlett 1994). Most commercially used sensors include a metal-oxide semiconductor (MOS), conducting polymer, bulk acoustic wave, surface acoustic wave, and quartz microbalance (Innawong and others 2004). The e-nose set of sensors afforded a large amount of information, and the processing of the data generated by the system was an essential part of the concept of electronic olfactometry (Martin and others 2001). Principal component analysis (PCA) is the pattern-recognition method most widely used for the elaboration of the sensor responses. It can be used to characterize the variations of an aroma as a function of time from initiation of process and change of the volatile compounds of a food (Capone and others 2001). Although considerable work using the e-nose has already been carried out; for example, in meat, coffee, mushrooms, cheese, fish, and orange juice (Schaller and others 1998), no application of the e-nose has been reported in aroma changes of frozen fruits. During freezing and frozen storage, water is the molecule most involved in deterioration reactions. Freezing contributes to cell rupture on a fruit’s surface, so interactions between previously separated substances then become possible. It provides a medium for diffusion of all other molecules involved in deterioration reactions. Water can also participate directly in deterioration reactions, including off-flavor production (Torreggiani and others 2000). In this chapter, we report on the aroma changes in pineapple during freezing and thawing that were investigated by using an e-nose technique. Two pineapple cultivars were used to verify the technique.

Materials and Methods Sample Preparation Pineapples (cultivars Smooth Cayenne and Queen) were purchased from a local market. Fruits were selected to have total soluble solids between 13 to 15 °Brix. The pineapples were washed, and an 80-mm-diameter borer was used in peeling them. A 30-mm-diameter borer was used to remove the core. Slices (15 mm thick) were cut perpendicularly to the fruit axis. Each slice was equally cut into eight pieces and packed individually in a 70-μm-thick nylon (polyamides/dry lamination/linear lowdensity polyethylene) bag. The air inside the bag was manually excluded as much as possible before sealing. Freezing and Thawing The pineapple samples were frozen in a chest freezer at −20°C (refrigerator model MDF-U536d; Sanyo, Tokyo, Japan) for 7 days and then thawed in a 4°C lowtemperature incubator for 12 h. This freeze-thaw cycle was repeated up to three times.

Analyzing the Effect of Freeze-Thaw Cycle on the Off-Aroma of Pineapple

659

E-nose Measurements A MOS-based gas-analyzer-array e-nose detector combined with a headspace automatic sampler (Fox 3000; Alpha MOS, Toulouse, France) was used in this study. The instrument comprised an array of 12 sensors placed in two chambers connected in series. The carrier gas of dried synthetic air was used. The data collected by the e-nose were analyzed by using Alpha software. Frozen-thawed pineapple (2 g) was placed in a 10-mL headspace vial. Eight samples were analyzed for each treatment. A crimping tool was used to seal each sample vial with a fitted cap and septum (polytetrafluoroethylene [PTFE]-coated silicone rubber). The sample vial was placed in the automatic sampler for the headspace septum. Each vial was incubated at 40°C for 10 min under agitation (500 rpm). A gas-tight 2.5-mL syringe kept at 50°C was used to withdraw 2500-μL volumes from the headspace. Sensory Evaluation A seven-member panel trained for study of freeze-thaw pineapple conducted sensory analysis for off-aroma. The panel members were trained to recognize the fresh pineapple aroma (no off-aroma) and changes in the aroma in fresh and freeze-thaw samples. Scores were assigned on a scale of 0–6, in which 0 was undetectable offaroma and 6 was the maximum detectable off-aroma. Statistical Analysis We used a completely randomized design. The difference between means was determined by using the Duncan’s new multiple range test. All statistical analyses were performed using SPSS 12.0 for Windows.

Results and Discussion The results demonstrated that the e-nose could clearly distinguish between the fresh and freeze-thaw pineapples by using PCA as a data-treatment technique. The projection of the two first principal axes is shown in Figure 63.1. The differentiation between the fresh and freeze-thaw pineapples can be seen from the first and second axes, explaining 58.49% and 27.42% of the variance, respectively. The e-nose seems to be sensitive to changes in pineapple aroma during freezing and thawing. The data were clearly classified into four groups: fresh, one freeze-thaw cycle, two freeze-thaw cycles, and three freeze-thaw cycles. Two and three freeze-thaw cycles induced a much greater increase in aroma changes than did one freeze-thaw cycle, as indicated by the greater variance from fresh. After two freeze-thaw cycles, the pineapple seemed to have reached its maximum freezing damage; therefore, the aroma changes after two and three cycles did not differ much. Furthermore, by means of the e-nose data the aromas of the two pineapple cultivars could be differentiated. Changes in the aroma of frozen fruits could signal either a loss of characteristic volatiles of fresh fruits during freezing (Ueda and Iwata 1982), or an increase or formation of some undesirable volatiles during freezing and thawing (Hirvi 1983). Unfortunately,

660

PART 2: Poster Presentations

10

Q

8

SC Q

6 4

DF2-27.418%

Freeze-Thaw 2 cycles

Fresh

2

SC

0

Q

–2

SC

Q

–4 –6 –8

Freeze-Thaw 1 cycle

SC

–10

–12 –11 –10 –9 –8 –7 –6 –5 –4 –3 –2 –1

Freeze-Thaw 3 cycles

0

1

2

3

4

5

6

7

8

9

10

DF1-58.491%

Figure 63.1. Principal component analysis plot for fresh and freeze-thaw Smooth Cayenne (SC) and Queen (Q) pineapple cultivars determined by e-nose. DF1 and DF2, discriminant functions 1 and 2. Table 63.1. Off-aroma scores of fresh and freeze-thaw (FT) pineapples determined by trained panels Average off-aroma scoresa

Cultivar Fresh Smooth Cayenne Queen

1 FT cycle

2 FT cycles

3 FT cycles

0bB

1.2bB

3.5aA

3.2aA

±0.0

±0.8

±1.8

±1.3

0bB

0.8bB

3.2aA

2.7aA

±0.0

±0.8

±1.6

±0.8

The values reported as mean ± standard deviation. a and b: Mean values in each row with different superscripts are significantly different (P ≤ 0.05). A and B: Mean values in each column with different superscripts are significantly different (P ≤ 0.05).

a

the e-nose data could not be used to determine whether the changes in frozen pineapple aroma were caused by the loss or gain in volatile constituents. The sensory evaluation data clearly indicated that our trained panels could distinguish between fresh pineapple aroma and freeze-thaw pineapple aroma (Table 63.1). Although two and three freeze-thaw cycles produced higher off-aroma scores than did one freeze-thaw cycle, the results of each freeze-thaw cycle had a relatively large variation (high standard deviation), indicating that the panels assigned different scores.

Analyzing the Effect of Freeze-Thaw Cycle on the Off-Aroma of Pineapple

661

The off-aroma scores from sensory evaluation of freeze-thaw pineapple corresponded relatively with the e-nose data.

Conclusion Our results suggest that changes in pineapple aroma during freezing and thawing, as well as pineapple cultivar effects, can be differentiated by means of an e-nose. The more freeze-thaw cycles applied, the greater was the difference observed from fresh sample. This is the first report of the use of an e-nose to detect aroma changes in frozen fruit.

Acknowledgments We greatly acknowledge the Thailand Research Fund–Master Research Grants (TRFMAG) under project MRG495S023. We also thank Sithiporn Associates for providing the e-nose, and Miss Sirinan Klaewthanong for advice on e-nose technique.

References Berna AZ, Lamnertyn J, Saevels S, Natale CD, Nicolai BM. 2004. Electronic nose systems to study shelf life and cultivar effect on tomato aroma profile. Sens Actuators [B] 97:324–33. Capone S, Epifani M, Quaranta F, Siciliano P, Taurino A, Vasanelli L. 2001. Monitoring of rancidity of milk by means of an electronic nose and a dynamic PCA analysis. Sens Actuators [B] 78:174–9. Gardner J, Bartlett P. 1994. A brief history of electronic nose. Sens Actuators [B] 18–19:211–20. Grate JW, Abraham MH. 1991. Solubility interactions and the design of chemically selective sorbent coatings for chemical sensors and arrays. Sens Actuators [B] 3:85–111. Hirvi T. 1983. Mass fragmentographic and sensory analyses in the evaluation of the aroma of some strawberry varieties. LWT Food Sci Technol 16:157–61. Innawong B, Mallikarjunan P, Marcy JE. 2004. The determination of frying oil quality using a chemosensory system. LWT Food Sci Technol 37:35–41. Martin YG, Oliveros MCC, Pavon JLP, Pinto CG, Cordero BM. 2001. Electronic nose based on metal oxide semiconductor sensors and pattern recognition techniques: characterisation of vegetable oils. Anal Chem Acta 449:69–80. Salunkhe DK, Bolin HR, Reddy NR. 1991. Sensory and objective quality evaluation. In: Salunkhe DK, Bolin HR, Reddy NR, editors. Storage, processing, and nutritional quality of fruits and vegetables, vol 1. 2nd ed. Boca Raton, FL: CRC. p 181–204. Schaller E, Bosset JO, Escher F. 1998. Electronic noses and their application to food. LWT Food Sci Technol 31:305–16. Skrede G. 1995. Fruits. In: Jeremiah LE, editor. Freezing effects on food quality. New York: Marcel Dekker. p 183–244. Torreggiani D, Lucas T, Raoult-Wack A-L. 2000. The pre-treatment of fruits and vegetables. In: Kennedy CJ, editor. Managing frozen foods. Boca Raton, FL: CRC. p 57–80. Torreggiani D, Maestrelli A. 2006. Quality and safety of frozen fruits. In: Sun D-W, editor. Handbook of frozen food processing and packaging. Boca Raton, FL: CRC. p 417–40. Ueda Y, Iwata T. 1982. Undesirable odour of frozen strawberries. J Jpn Soc Hortic Sci 51:219–23. Wannawilas K, Noiwong N, Lachachotiroj M, Charoenrein S. 2004. Freezing of two pineapple cultivars: Queen and Smooth Cayenne. Bangkok: Industrial Project for Undergraduate Students, Thailand Research Fund. p 146–7.

64 Water Uptake and Solid Loss During Soaking of Milled Rice Grains P. Chatakanonda and K. Sriroth

Abstract Water uptake and solid loss of milled rice grains from three Thai rice varieties, most commercially available—waxy, jasmine, and Sao Hai rice, with varied amylose contents of 7.17%, 16.63%, and 29.06%, respectively—were analyzed in the temperature range of 35°–85°C. An increasing rate of water uptake was observed at the beginning of the soaking process, and a reducing rate was detected thereafter, in particular at temperatures above 60°C for waxy and jasmine rice and above 75°C for Sao Hai rice. The parameters of the kinetics involving the simultaneous unsteady-state water diffusion and first-order irreversible water-starch interaction were evaluated by using five empirical models and the Fickian diffusion model. The Page model was found to suitably describe the water-absorption characteristics of milled rice grains. Water absorption rate, as well as solid loss, expressed as a power function of soaking duration were found to decrease with increasing amylose content of rice grains.

Introduction Milled rice usually requires 20–30 min to cook. This relatively long time is caused by the slow rate at which water diffuses into the kernel. The rice grain kernel is tightly packed; it has no air space or other channel for water to penetrate it. For the rice to cook, water must penetrate to the center of the kernel with sufficient heat capacity to gelatinize the starch, which constitutes about 90% of milled rice’s dried weight. Moisture absorption, therefore, plays an important role during the cooking of milled rice grains. Water absorption of cereals and legumes has been described mainly by Fick’s laws of diffusion (Sayar and others 2001; Bello and others 2004, 2007; Resio and others 2005; Muramatsu and others 2006). These laws of diffusion involve numerous functions and parameters, thus making them inconvenient for practical uses under most conditions. Exponential empirical equations, therefore, have been preferentially employed to model water-absorption characteristics of food materials in many cases (Kashaninejad and others 2007; Yadav and Jindal 2007) because of their simplicity and the fit of their applications. Many researchers have studied water absorption and starch gelatinization of rice during parboiling before milling (Bakshi and Singh 1980) and water absorption during cooking or soaking in hot water (Metcalf and Lund 1985; Yadav and Jindal 2007). Moisture absorption and loss of solids from rice grains have been reported to vary 663

664

PART 2: Poster Presentations

among varieties containing different amylose contents (Juliano 1972; Metcalf and Lund 1985; Yadav and Jindal 2007). These studies also indicated influences of other physicochemical properties including gelatinization, gel consistency, alkali spread, and protein content, as well as the presence of other ingredients. The variability observed among studies revealed diverse soaking characteristics of different cultivars of rice grains. In this work, water uptake and solid loss of selected Thai rice varieties with different amylose contents were characterized in the temperature range of 35°–85°C. Experimental data on the water-absorption characteristics of rice grains as affected by temperature were described by using different mathematical models.

Materials and Methods Rice Samples Grains from three different rice varieties—waxy, jasmine, and Sao Hai—were obtained commercially. These rice grains contained 12%–14% moisture content, and amylose contents of 7.17%, 16.63%, and 29.06%, respectively, as determined by the iodine colorimetric method (Juliano 1971). Broken and unripe grains (visually observed) were removed by hand before analysis. Determination of Moisture Content During Soaking Milled rice grains were soaked in excess water (rice-water ratio, 1:10) at 35°–85°C for various times. Soaked rice grains were removed at specific intervals (5–180 min), quickly blotted with tissue paper to remove surface water, weighed, and the moisture content was determined. Water uptake of rice grains was expressed in terms of the increment in moisture content on a dry-weight basis. The moisture ratio (MR) of milled rice grains as a function of soaking time at different temperatures was fitted to five mathematical models (Table 64.1) by nonlinear regression analysis (Equation 64.1): MR =

M t − Me Mo − Me

(64.1)

where Mt is the moisture content at soaking time t (g/g solid), Me is the equilibrium moisture content (g/g solid), Mo is the initial moisture content (g/g solid), t is the Table 64.1. Mathematical models for predicting the moisture ratio (MR) of rice grains Model

Equation*

Exponential model

MR = exp(−kt)

Henderson and Pabis model

MR = A exp(−kt)

Page model

MR = exp(−ktn)

Modified Page model

MR = exp(−kt)n

Two-term exponential model

MR = A exp(−k1t) + B exp(−k2t)

* See text for explanation of terms.

Water Uptake and Solid Loss During Soaking of Milled Rice Grains

665

soaking time (min), and k and n are regression constants. Goodness of fit was expressed by regression coefficient (R2) and percentage of relative mean square error (%RMSE) (Equation 64.2): ⎧1 %RMSE = 100 ⎨ ⎩N

12

⎛ M measured − M predicted ⎞ ⎫ ⎟⎠ ⎬ ∑ ⎜⎝ M measured i =1 ⎭⎪ 2

N

(64.2)

Determination of Diffusion Coefficient of Water The water-diffusion process in rice grains during soaking was determined by using Fick’s second law solution for a sphere, based on the assumptions that the effective diffusion coefficient is independent of moisture content, the grain volume does not change during water absorption, and the grain surface reaches the equilibrium moisture content instantly when immersed in water. The series equation for diffusion out of a sphere is as follows (Crank 1975) (Equation 64.3): MR =

M t − Me 6 = 2 Mo − Me π

∞

1

∑n n =1

2

2 2 ⎛ n Dπ ⎞ t⎟ exp ⎜ − ⎝ ⎠ r2

(64.3)

where D is the effective diffusion coefficient (m2/s), and r is equivalent to the spherical radius of rice grain (m). Experimental data were fitted to this equation by using nonlinear least-squares analysis to determine D values. Determination of Solid Loss During Soaking The soluble solids of cooked rice were analyzed; the fractions in soaked water were completely dried to determine the weight of solids leached during the soaking of rice grains. The loss of solids from milled rice grains as a function of soaking time at each temperature was described by a power-function model (Equation 64.4): SL = t a

(64.4)

where SL is the percentage of solid loss (dry basis), t is the soaking time (min), and a is the regression constant.

Results and Discussion Water Uptake of Rice Grains During Soaking The moisture content of rice grains at various temperatures as a function of soaking time is shown in Figure 64.1. The water-uptake rate of milled rice grains of all varieties was initially rapid and then relatively slow thereafter while approaching equilibrium. The sharp increase in moisture content of rice grains has been attributed to cracks present on the grain surface and internal fissure caused by milling (Ahromrit

5

Waxy rice

Moisture content (%, db)

o

35 C 45 oC 55 oC 60 oC o 65 C 75 oC 85 oC

4

3

2

1

0 0

50

100

150

200

Time (min)

Moisture content (g/g db)

5

35 oC 45 oC 55 oC 60 oC 65 oC 75 oC 85 oC

4

3

Jasmine rice

2

1

0 0

20

40

60

80

100

120

140

160

180

200

Time (min)

Moisture content (g/g db)

5

Sao Hai rice

45 oC 55 oC 60 oC 65 oC 75 oC 85 oC

4

3

2

1

0

0

20

40

60

80

100

120

140

160

180

200

Time (min)

Figure 64.1. Comparison of experimental and predicted moisture contents of milled rice grains from waxy, jasmine, and Sao Hai varieties as a function of soaking time at various temperatures. db, dry basis.

666

Water Uptake and Solid Loss During Soaking of Milled Rice Grains

667

and others 2006). The water uptake of rice grains was affected by temperature: wateruptake rate increased as a function of temperature, in particular at temperatures above 60°C, for waxy and jasmine rice and above 75°C for Sao Hai rice. These increases in hydration rate in the vicinity of gelatinization temperatures of rice starches from each variety (60°–80°C, 65°–80°C, and 70°–85°C for waxy, jasmine, and Sao Hai varieties, respectively) were caused mainly by irreversible changes in starch granules upon gelatinization. The moisture absorbed by rice grains clearly varied among the varieties selected. The saturated moisture content of milled rice grains at any soaking time and temperature was found to decrease with increasing amylose content. Changes in the moisture content of milled rice grains in terms of moisture ratio were fitted by five exponential models selected, and the results of the statistical analysis of fitting are presented in Table 64.2. The Page model provided higher regression coefficients (R2) and lower percentages of root mean square error (%RMSE) than did other models, indicating better prediction of water-uptake characteristics of milled rice for all varieties in this experiment. The Page model parameters further applied to predict moisture content of rice grains were in good agreement with the moisture content measured (Figure 64.1). The k, corresponding to the rate of moisture uptake, increased with soaking temperature up to the gelatinization temperature of rice starches (Table 64.3). The temperature dependence of k was determined in the temperature range of 35°–60°C. The reduction in k observed above the gelatinization temperature of all samples may be associated with the increased loss of solids induced by gelatinization. Values of k were inversely related to the amylose content of rice and the same trend was observed for equilibrium moisture content (Me), demonstrating a greater hydration capacity of rice containing lower amylose content. These results agree with those from previous studies observing a negative correlation between water uptake and amylose content (Metcalf and Lund 1985; Yadav and Jindal 2007). Juliano

Table 64.2. Regression coefficient (R2) and percentage of root mean square error (%RMSE) obtained from model fitting using exponential, Henderson and Pabis, Page, modified Page, and two-term exponential models Model Exponential model

Fitting results 2

R

%RMSE Henderson-Pabis model

R2 %RMSE

Page model

R2

Two-term exponential model

Jasmine

Sao Hai

0.945 ± 0.045

0.951 ± 0.060

0.959 ± 0.021

6.01 ± 4.42

6.09 ± 3.74

6.52 ± 5.52

0.942 ± 0.051

0.952 ± 0.060

0.964 ± 0.016

9.14 ± 8.32

13.04 ± 15.30

8.44 ± 8.62

0.988 ± 0.004

0.979 ± 0.030

0.984 ± 0.012

4.70 ± 3.73

5.86 ± 4.33

4.38 ± 3.62

R2

0.961 ± 0.035

0.977 ± 0.030

0.979 ± 0.013

%RMSE

17.64 ± 26.71

18.26 ± 19.86

17.14 ± 19.86

R2

0.962 ± 0.035

0.975 ± 0.028

0.984 ± 0.010

6.99 ± 8.23

10.95 ± 15.29

7.50 ± 9.75

%RMSE Modified Page model

Rice variety Waxy

%RMSE

668

PART 2: Poster Presentations

Table 64.3. Page model parameters (Me, k, and n) for water uptake by milled rice grains at different temperatures Rice variety Waxy

Jasmine

Sao Hai

Temp. (°C)

Me (g/g solid)

k (min−1)

n

35

0.530

1.043

0.492

45

0.523

1.093

0.649

55

0.535

1.053

0.435

60

0.558

1.180

0.362

65

1.779

0.160

0.208

75

4.459

0.013

0.910

85

4.297

0.020

0.928

35

0.445

0.805

0.426

45

0.452

0.906

0.383

55

0.466

1.124

0.289

60

0.533

0.917

0.257

65

1.065

0.293

0.204

75

3.214

0.010

1.226

85

3.355

0.006

1.449

35

0.349

0.707

0.442

45

0.346

0.738

0.463

55

0.361

0.866

0.380

60

0.384

0.955

0.270

65

0.448

0.682

0.244

75

0.839

0.157

0.499

85

2.887

0.019

0.751

Me, equilibrium moisture content; and k and n, regression constants.

(1972), however, reported the increased capacity of starch granules to absorb water for a rice variety with higher amylose content. Water Diffusion in Rice Grains During Soaking The effective diffusion coefficients (D) determined from the relationship between moisture ratio (MR) and soaking time at the temperatures studied are presented in Table 64.4. These values appeared to increase with soaking temperature up to 60°C, consistent with the findings from earlier studies (Muramatsu and others 2006; Bello and others 2007; Kashaninejad and others 2007). The drop in D was, however, observed at higher temperature for all rice varieties. This observation corresponds to the results of a previous study by Ahromrit and others (2006), who reported a decrease in the D of waxy rice at temperatures above 60°C and attributed this finding to starch gelatinization, which might restrict the kinetics of water transport. Solid Loss of Rice Grains During Soaking Changes in the loss of solids upon soaking of rice grains are shown in Figure 64.2. The extent of solid loss was significantly affected by amylose content: varieties with

Water Uptake and Solid Loss During Soaking of Milled Rice Grains

669

Table 64.4. Effective diffusion coefficient (D) for milled rice grains at different temperatures Rice variety Waxy

Jasmine

Sao Hai

Temp. (°C)

D (m2/s)

35

7.518 × 10−9

45

8.855 × 10−9

55

8.034 × 10−9

60

1.132 × 10−9

65

1.911 × 10−10

75

6.738 × 10−11

85

9.279 × 10−10

35

9.146 × 10−10

45

9.036 × 10−10

55

1.157 × 10−9

60

8.679 × 10−10

65

2.673 × 10−10

75

8.641 × 10−11

85

1.189 × 10−10

35

7.668 × 10−10

45

8.851 × 10−10

55

1.018 × 10−9

60

8.451 × 10−10

65

5.948 × 10−10

75

1.068 × 10−10

85

2.115 × 10−11

a lower amylose content had greater loss. A relatively lower extent (<15%) and a lower rate of solid loss were determined at ≤65°C for jasmine and waxy varieties and at ≤75°C for Soa Hai variety. The extent of solid loss by rice grains generally exhibited a power relationship with soaking duration at the soaking temperatures studied. The increments were determined in exponent and R2 values mainly at higher temperatures, except for waxy rice (Table 64.5).

Conclusion The water-absorption rate, as well as loss of solids, expressed as an exponential and power function respectively, of soaking duration at varied temperatures in excess water, were found to decrease with increasing amylose content of rice grains. The Page model was found to be a suitable empirical model for expressing moisture-uptake characteristics during soaking for all rice varieties. These results indicate that the kinetics of water absorption by rice grains differs depending on the rice variety, and that this can be used to predict the moisture content and ease of cooking of rice grains at different stages of soaking temperatures.

60 50 Solid loss (%, db)

Waxy rice

35 ºC 45 ºC 55 ºC 60 ºC 65 ºC 75 ºC 85 ºC

40 30 20 10 0

0

20

40

60

80

100 120 140 160 180 200

Time (min) 60 50 Solid loss (%, db)

Jasmine rice

35 ºC 45 ºC 55 ºC 60 ºC 65 ºC 75 ºC 85 ºC

40 30 20 10 0

0

20

40

60

80

100 120 140 160 180 200

Time (min) 60 50 Solid loss (%, db)

Sao Hai rice

35 ºC 45 ºC 55 ºC 60 ºC 65 ºC 75 ºC 85 ºC

40 30 20 10 0

0

20

40

60

80

100 120 140 160 180 200

Time (min)

Figure 64.2. Comparison of experimental and predicted solid loss of milled rice grains from waxy, jasmine, and Sao Hai varieties as a function of soaking time at various temperatures. db, dry basis.

670

Water Uptake and Solid Loss During Soaking of Milled Rice Grains

671

Table 64.5. Power-function model parameters (a and R2) for loss of solids from milled rice grains at different temperatures Rice variety Waxy

Jasmine

Sao Hai

Temp. (°C)

a

R2

35

0.199

0.969

45

0.227

0.937

4.88

55

0.221

0.674

13.91

60

0.335

0.731

15.10

65

0.530

0.964

10.48

75

0.757

0.529

33.76

85

0.791

0.656

32.20

35

0.039

0.283

20.78

45

0.045

0.331

21.28

55

0.121

0.639

17.16

60

0.171

0.671

30.60

65

0.307

0.784

44.59

75

0.687

0.965

32.63

85

0.714

0.961

36.45

35

0.079

0.669

12.86

45

0.071

0.522

16.68

55

0.099

0.622

21.13

60

0.106

0.598

21.55

65

0.184

0.759

16.52

75

0.453

0.952

20.86

85

0.649

0.888

44.37

%RMSE 5.47

a, regression constant; R2, regression coefficient; and %RMSE, percentage of root mean square error.

References Ahromrit A, Ledward DA, Niranjan K. 2006. High pressure induced water uptake characteristics of Thai glutinous rice. J Food Eng 72:225–33. Bakshi AS, Singh RP. 1980. Kinetics of water diffusion and starch gelatinization during rice parboiling. J Food Sci 45:1387–92. Bello MO, Tolaba MP, Suarez C. 2004. Factors affecting water uptake of rice grain during soaking. LWT Food Sci Technol 37:811–6. Bello MO, Tolaba MP, Suarez C. 2007. Water absorption and starch gelatinization in whole rice grain during soaking. LWT Food Sci Technol 40:313–8. Juliano BO. 1971. A simplified assay for milled rice amylose. Cereal Sci Today 16:334–8. Juliano BO. 1972. The rice caryopsis and its composition. In: Houston DF, editor. Rice chemistry and technology. 1st ed. St Paul, MN: American Association of Cereal Chemists. p 16–74. Kashaninejad M, Maghsoudlou Y, Rafiee S, Khomeiri M. 2007. Study of hydration kinetics and density changes of rice (Tarom Mahali) during hydrothermal processing. J Food Eng 79:1383–90. Metcalf SL, Lund DB. 1985. Factors affecting water uptake in milled rice. J Food Sci 50:1676–9, 1684. Muramatsu Y, Tagawa A, Sakaguchi E, Kasai T. 2006. Water absorption characteristics and volume changes of milled and brown rice during soaking. Cereal Chem 83:624–31.

672

PART 2: Poster Presentations

Resio ANC, Aguerre RJ, Suarez C. 2005. Analysis of simultaneous water absorption and water-starch reaction during soaking of amaranth grain. J Food Eng 68:265–70. Sayar S, Turhan M, Gunasekaran S. 2001. Analysis of chickpea soaking by simultaneous water transfer and water-starch reaction. J Food Eng 50:91–8. Yadav BK, Jindal VK. 2007. Water uptake and solid loss during cooking of milled rice (Oryza sativa L.) in relation to its physicochemical properties. J Food Eng 80:46–54.

65 Microstructural, Physical, and Rehydration Properties of Maltodextrin Powders Obtained by Spray Drying A. L. Muñoz-Herrera, V. Tejeda-Hernández, A. Jiménez-Aparicio, J. Welti-Chanes, J. J. Chanona-Pérez, L. Alamilla-Beltrán, and G. F. Gutiérrez-López Abstract Operating conditions in spray drying have a strong influence in rehydration of food powders. Moisture content, bulk and tapped density, angle of repose, dispersability, wettability, and microstructure of powders are important in mixing, transport, storage, and rehydration. This work studied the influence of three different concentrations of maltodextrin solutions and two spray-drying air temperatures on rehydration and microstructure characteristics of powders. Maltodextrin solutions containing 20%, 30%, and 40% total solids (TS) were subjected to cocurrent two-nozzle atomization spray drying at 140° and 249°C inlet drying-air temperature, outlet drying-air temperature of 70° and 120°C, and feed rate of 1.2 L/h. Physical and rehydration properties of maltodextrin powders can be improved when they are obtained by spray-drying solutions at higher temperatures and the highest concentration of feed. High air-drying temperature (249°/120°C) and high feed concentrations (40% TS) have a strong influence on the microstructure of particles produced during spray drying and improved the “instant” properties of powders.

Introduction Food processing generates products that should have specific quality and safety characteristics. Quality and safety for consumer health and wellness is the main aim in any food process (Lund 2002; Trystram and Bimbenet 2002). Product development and product improvement are largely based on creating structures in which components have different functions. Spray drying of foods is an example of how microstructures or nanostructures are created as powders or agglomerates (Aguilera and Stanley 1999). Food drying is a complex process in which the interaction of mass- and heattransfer mechanisms have a strong influence on the removal of water. Water plays an important role in the chemical, physical, and structural changes that occur during drying (Sanjuán and others 1999; Campos-Mendiola and others 2007). Spray drying has been used traditionally to produce food powder products. The microstructure of products developed either naturally or by processing has a direct effect on the sensory, physical, chemical, and transport properties of food 673

674

PART 2: Poster Presentations

powders (Jha and others 2002; Nijdam and Langrish 2005; Shittu and Lawal 2007). The quality of food powder properties obtained by spray drying depends on the operating conditions of the dryer, and rehydration properties (wettability, sinkability, dispersability, and solubility) are affected considerably by the process (Jha and others 2002; Goula and Adamopoulos 2005; Nijdam and Langrish 2005; Shittu and Lawal 2007). In the same way, flow properties (moisture content, bulk density, angle of repose, etc.) depend on the specific drying conditions used. Food powders are often rehydrated before being consumed, and the addition of water should produce a fast and uniform solution. For ideal wettability, a low contact angle must be developed between liquid and solid, which implies a high polarity on the surface (Lazghab and others 2005). Powder sinkability arises if the density of the product is higher than the density of the solvent that flows through the channels or pores, replacing the occluded air in the powder. If the correct properties are achieved, dispersion and a homogeneous solution are produced (Aguilera and Stanley 1999, Hogekamp and Schubert 2003; Lazghab and others 2005). Better rehydration and physical properties can be achieved by controlling the microstructure during the drying process. The role of structure in dehydrated products appears evident to the understanding of transport mechanisms and in the design of functional properties (Aguilera 2005). Image analysis can be applied to observe structures and obtain physical data in nanoscopic and microscopic fields. The information provided through image analysis can be used in the study of the structure-function relationships of dried foods and the rehydration of food powders (Aguilera and Stanley 1999; Aguilera 2005). This work assessed the effect of spray-drying inlet and outlet air temperature and the concentration of feed on the microstructural, physical, and rehydration properties of a maltodextrin powder.

Materials and Methods Testing Materials The testing materials were solutions of maltodextrin 20DE (Arancia Corn Products, Mexico City, Mexico) at wt/vol concentrations of 20%, 30%, and 40% total solids (TS). Maltodextrin solutions were dried by a two-fluid nozzle laboratory-scale spray dryer (SPAGA9601; Instituto Politécnico Nacional [IPN], Mexico City, Mexico) (Alamilla-Beltrán and others 2005). In all experiments, the atomizer air pressure and the feed rate were kept at 1.2 kg/cm2 and 1.2 L/h, respectively. The inlet/outlet drying air temperatures were 140°/70°C and 249°/120°C. Microstructure and Morphology of Particles The food powder product collected into the glass jar at the bottom of the cyclone of the spray dryer was observed (1000 ×) by means of a scanning electron microscope (SEM) (JEOL model JBM-5900LV) (Alamilla-Beltrán and others 2005).

Microstructural, Physical, and Rehydration Properties of Maltodextrin Powders

675

Physical Properties The average moisture content of these powders was evaluated by means of a thermobalance (Brainweigh model MB300). Bulk and tapped density were evaluated by applying the method reported by Okaka and Potter (1979). Samples of 20 g of powder were poured into a 100-mL measuring cylinder and the tapped 100 times on a flat platform. The mass of the empty, and filled measuring cylinder and the final volume occupied by each sample were obtained, and bulk and tapped density values were calculated. Angle of repose was evaluated applying the method reported by Shittu and Lawal (2007). A 250-mL round-mouth measuring cylinder was filled with 200 mL of the powder. The container was covered with a flat plate and removed abruptly, thus allowing the powder to fall in a uniform cascade to form a cone. The distance between the container with falling powder and the plate should be ∼20 cm. The angle of repose was calculated as the arc tangent of the ratio of height to the base radius of the heap formed. Rehydration Properties Dispersability was determined following the International Dairy Federation (IDF) standard. Powder (25 g) was mixed with 250 mL of water at 25°C in a mixer at 100 rpm for 90 s. The mix was filtered through a sieve of 150-μm apertures. The percentage of material not retained was calculated. Wettability was determined by following the modified method of Wollny and Schubert (1999). The wetted height of bulk material over time and its critical height were evaluated. The cone apparatus, which is a circular powder-sample container with a sieve cloth covering the bottom, was used. The powder is deposited into this apparatus in the form of a vortex. The container with the sample is placed into a container with water. The liquid is imbibed into the sample, and a growing spot of wetted material in the center of the powder is then observed. The wetted height can then be calculated.

Results and Discussion Air-drying temperatures, feed concentration, moisture content, bulk and tapped density, and angle of repose for the various tests are listed in Table 65.1. The moisture content of maltodextrin powders was in the range of 0.04–0.07 kg water/kg dry solid. The moisture content achieved during spray drying was lower at higher feed concentrations and air-drying temperatures. The lowest moisture content in powders was found in materials dried at the highest air-drying temperature and at the highest concentration of feed. At high spray-drying temperatures, moisture evaporated quickly and the resulting skin formed was hard and thin. Hollow nonshrinking particles tend to be formed under these thermal conditions (Figure 65.1a). Nonshrinkage was induced by the impossibility of vapor to diffuse out of the particle and condensation of the same into it (Nijdam and Langrish 2005). At low spray-drying temperatures, dried particles were small with a rough surface, and shrinkage was observed. The particles became less spherical and more surface

Table 65.1. Moisture content, bulk and tapped density, and angle of repose of powders evaluated at different feed concentrations and drying-air temperatures Drying-air temperature(°C) Inlet

Outlet

Feed concentration (%TS)

Moisture content kg water/kg dry solid

Density (g/mL)

Angle of repose (°)

Bulk

Tapped

140

70

20

0.06

0.36

0.60

43

140

70

30

0.06

0.42

0.64

51

140

70

40

0.07

0.43

0.63

43

249

120

20

0.04

0.27

0.59

47

249

120

30

0.04

0.33

0.52

47

249

120

40

0.04

0.21

0.38

45

TS, total solids.

Figure 65.1. Microstructure of powders obtained at 40% total solids (TS) feed concentration and inlet/outlet drying air: (a) 249°C/120°C and (b) 140°C/70°C.

676

Microstructural, Physical, and Rehydration Properties of Maltodextrin Powders

677

shrivel was evident as the drying temperature was reduced. Under these conditions, powders showed the highest moisture content (Figure 65.1b). Bulk and tapped density results ranged from 0.21–0.43 g/mL and 0.38–0.64 g/mL, respectively. The effect of drying-air temperatures and concentration of feed on powder bulk and tapped density were dependent on moisture content. At higher moisture contents, a stronger tendency of particles to stick together was observed, leaving more space between them and producing a larger bulk volume. Goula and Adamopoulos (2005) observed that increasing the drying-air temperature reduced bulk and tapped density. When drying occurred at the lowest temperature, powders had the highest bulk and tapped density, and the morphology of particles showed organized arrangements with reduced space between the particles. The angle of repose is affected by moisture content, particle size, and particle morphology. Low angles of repose were related to low moisture contents and irregular forms of particles, typical of the low temperature and low concentration of feed used. The angle of repose varied from 43° to 47°, which indicates that the powder was free flowing or cohesive (Shittu and Lawal 2007). Powders with high resistance to flow can be obtained by increasing the concentration of feed and drying at high temperatures. High spray-drying air temperatures resulted in poor and limited powder wettability, but dispersability was the highest reaching 95%–98% (Table 65.2). Wettability, evaluated by using the height of wetted powder (cone method), was 0.6–0.8 cm. The wettability rate was related to the feed concentration, and the highest values of this parameter were observed at 20% TS feed concentration. The presence of channels or pores formed in the structure during drying allowed for high wetting rates, so better wettability properties for a powder could be obtained at low air-drying temperatures, which diminished dispersability. Wettability increased when drying solutions at high temperatures and high concentrations of feed. Water vapor wet the surface of the particles, reducing the cohesion force among them. This allowed a fast flow of water through channels or pores, eliminating air. Liquid molecules close to the three-phase contact line (solid-liquid-gas) pushed away Table 65.2. Dispersability and wettability evaluated at different feed concentrations and drying-air temperatures in the production of powders Drying-air temperature (°C)

Feed concentration (%TS)

Dispersability (%)

Wettability (cm)

Inlet

Outlet

140

70

20

77.63

0.59

140

70

30

76.77

0.67

140

70

40

72.87

0.81

249

120

20

95.23

0.32

249

120

30

96.77

0.37

249

120

40

98.57

0.52

TS, total solids.

678

PART 2: Poster Presentations

gas or vapor adsorbed at the solid’s surface, and attached to the solid by forming bonds with molecules at the solid’s surface (Lazghab and others 2005). In this case, strong solid-liquid adhesive forces predominate (Lazghab and others 2005; Nijdam and Langrish 2005). Spherically shaped particles induce a low degree of interstitial air as the small particles in the size distribution fill void spaces. Powders obtained at 120°C inlet air-drying temperature with a feed concentration of 20% TS showed the highest wetting rate. This may indicate that liquid molecules close to the contact line moved away the air adsorbed and adhered to the solid by forming bonds. In this particular case, a strong solid-liquid adhesive force predominated. Liquid penetrated spaces among particles as channeling was enabled by particle morphology.

Conclusion Physical and rehydration properties of wall materials used in microencapsulation, like maltodextrin powders, can be improved when obtained by spray-drying solutions at higher temperatures and high feed concentration. High air-drying temperatures and high feed concentration have a strong influence on the microstructure of spray-dried particles and have marked influence on “instant” properties of test powders.

Acknowledgments This work was supported by the National Polytechnic Institute (IPN) of Mexico research projects SIP-IPN (20070544 and 20082618) and by COFAA-IPN and CONACYT (National Council for Science and Technology, Mexico) projects 2004-C-1-4861-Z and 52577-Z.

References Aguilera JM. 2005. Why food microstructure? J Food Eng 67:3–11. Aguilera JM, Stanley DW. 1999. Microstructural principles of food processing and engineering. 2nd ed. Gaitherburg, MD: Aspen. Alamilla-Beltrán L, Chanona-Pérez JJ, Jiménez-Aparicio AR, Gutiérrez-López GF. 2005. Description of morphological changes of particles along spray drying. J Food Eng 67:179–84. Campos-Mendiola R, Hernández-Sánchez H, Chanona-Pérez JJ, Alamilla-Beltrán L, Jiménez-Aparicio A, Fito P, Gutiérrez-López GF. 2007. Non-isotropic shrinkage and interfaces during convective drying of potato slabs within the frame of the systematic approach to food engineering systems (SAFES) methodology. J Food Eng 83:285–92. Goula AM, Adamopoulos KG. 2005. Spray drying of tomato pulp in dehumidified air. II. The effect on power properties. J Food Eng 66:35–42. Hogekamp S, Schubert H. 2003. Rehydration of food powders. Food Sci Technol Int 9:223–35. Jha A, Ambalal PA, Bijoy RR. 2002. Physico-chemical properties of instant Kheer mix. Lait 82:501–13. Lazghab M, Saleh K, Pezron I, Guigon P, Komunjer L. 2005. Wettability assesment of finely divided solid. Powder Technol 157:79–91. Lund DB. 2002. Food engineering for the 21st century. In: Welti-Chanes J, Barbosa-Cánovas GV, Aguilera JM, editors. Engineering and food for the 21st century. Boca Raton, FL: CRC. p 3–14. Nijdam JJ, Langrish TAG. 2005. An investigation of milk powders produced by a laboratory-scale spray dryer. Drying Technol 23:1043-56.

Microstructural, Physical, and Rehydration Properties of Maltodextrin Powders

679

Okaka J, Potter N. 1979. Physicochemical and functional properties of soybean powders processed to reduce flavor. J Food Sci 44:1235–40. Sanjuán N, Simal S, Bon J, Mulet A. 1999. Modelling of broccoli stems rehydration process. J Food Eng 42:27–31. Shittu TA, Lawal MO. 2007. Factors affecting instant properties of powdered cocoa beverages. Food Chem 100:91–8. Trystram G, Bimbenet JJ. 2002. Trends in food engineering. In: Welti-Chanes J, Barbosa-Cánovas GV, Aguilera JM, editors. Engineering and food for the 21st century. Boca Raton, FL: CRC. p 15–34. Wollny M, Schubert H. 1999. Eine neue Methode zur Bestimmung des zeitlichen Benetzungsverhaltens von Partikelschûttungen [A new method for determining the timing of bulk particle wetting]. DECHEMA Jahrestagungen 99 [DECHEMA Annual Conference 99]. Weisbaden, Germany: Kurzfassungen, 2:369–71.

66 Nanostructures and Minimum Integral Entropy as Related to Food Stability L. A. Pascual-Pineda, E. Flores-Andrade, C. I. B. Guevara, L. Alamilla-Beltrán, J. J. Chanona-Pérez, E. Azuara-Nieto, and G. F. Gutiérrez-López Abstract This work evaluated the minimum integral entropy of the water adsorbed in the matrix of different nanostructured (NSM) and non-nanostructured (NNS) food-model systems and its relation to stability. Sorption isotherms were determined at 25° and 35°C for sucrose-calcium powder obtained by spray drying paprika-containing alginic acid capsules (AAs), and for sucrose-calcium NSM powder obtained by a cryogenic process and for paprika-containing alginic acid capsules, which included zeolite Valfor 100. The minimum integral entropy was assessed by means of the variation in the available adsorption surface in relation to water activity. The micropore volume was determined according to the Dubinin-Radushkevich relationship. The carotenoid red fraction loss was determined by a spectrophotometric assay. The physical and chemical stability of NSM was maintained at high relative humidities during storage. Food product stability can be improved by inducing nanostructures and thus facilitating entropic control of water adsorption.

Introduction Nanotechnology has been the framework within which new food materials have been developed, and the use of nanostructures in the range of 2–1000 nm have been reported (Sanguansri and Augustin 2006). This technology enables the development of food materials with improved physical, chemical, and biological properties, as well as better stability. Until now, it has not been investigated how nanostructures may help microstructure control and help to modify the stability of dehydrated products. Foods are biological complex systems whose biochemical and physicochemical properties are modified by the type of processing that they undergo. Nevertheless, these systems are governed by the same principles and mechanisms that physicists, biologists, and biochemists study. Therefore, the findings derived through nanotechnology might have applications for the food industry. Parameters such as water activity (aw) and glass transition temperature (Tg) have been used simultaneously to predict the optimal conditions for preserving foods (Roos 1993). However, predicted storage conditions often do not correspond with real stability parameters, so more accurate methodologies to predict optimal storage conditions 681

682

PART 2: Poster Presentations

been developed, such as the methodology developed by Beristain and Azuara (1990) that enables suitable storing for various materials of known moisture content to be selected based on conditions at which water adsorption is controlled by entropyrelated parameters. The minimum integral entropy (ΔSint) appears when the energy interactions between the adsorbent (food matrix) and the adsorbate (water vapor) are strong; that is, when water molecules are adsorbed in the most active sites. This thermodynamic criterion has been demonstrated experimentally by showing that the best storage conditions regarding food–water vapor interactions are those at which the system has its minimum integral entropy. Beristain and others (2002), who evaluated the oxidation of essential orange oil microencapsulated with mesquite gum, observed that the oxidation intensity was less at a moisture content corresponding to the point (temperature and relative humidity of air) of minimum integral entropy, even though the samples were in the rubbery state. Recent work (Domínguez and others 2007) has shown that the optimal storage conditions that minimize lipid oxidation and maximize color and texture retention of macadamia nuts depend on the minimum integral entropy zone. The results reported by Azuara and Beristain (2006) also suggest that the moisture content in the minimum integral entropy coincides with the moisture corresponding to the micropore volume determined with the Dubinin-Radushkevich model. In addition, their study indicates that the interactions between water molecules and the food solid matrix can be explained by enthalpic and entropic mechanisms and by investigating water-nanopore interactions. It has been demonstrated that two control stages influence the interactions of water molecules with food. Enthalpic control is related to energy aspects of the sorption phenomena, whereas entropic control is related to changes in microstructure (Beristain and others 1996). Our work evaluated the minimum integral entropy of the water adsorbed in the matrix of different food systems that have nanostructural and non-nanostructural matrices and we relate our findings to food stability.

Materials and Methods Sucrose-Calcium Powders Cryogenic Process Sucrose-calcium powders were obtained by the Zeller and Saleeb (1996) method, in which cryogenic ethanol is used. A sucrose-calcium solution (33% of solids with a molar ratio of 1 : 1.25) was sprayed on ethanol at −40° (E-40) and −80°C (E-80) by using an ethanol/solution ratio of 30 : 1. The resulting powder was subjected to vacuum pressure of 0.133 mbar over 1 day to eliminate the ethanol. Spray Drying The sucrose-calcium solutions (33% of solids) were spray-dried in a Büchi Mini Spray Dryer (model 190; Büchi, Flawil, Switzerland) using inlet/outlet temperatures of 170° and 110°C, respectively.

Nanostructures and Minimum Integral Entropy as Related to Food Stability

683

Paprika Capsules Blank capsules were made with alginic acid (AA), whereas nanostructured AA capsules (zoelite AA capsules [ZAA]) were made with 75% Na-A zeolite (Valfor 100 zeolite; PQ, Malvern, PA, USA) and 25% AA as wall material. Paprika was added to the solution at a ratio of 1 : 0.5 (solids/paprika). Materials were dripped slowly onto a 2% calcium chloride solution to form coacervates. Materials were vacuum oven-dried (575 mbar) at 45°C over 1 day. Yogurt and Macadamia Nut The water-sorption data of lyophilized yogurt and macadamia nut were taken from reports by Azuara and Beristain (2006) and Domínguez and others (2007), respectively. Sorption Isotherms The sorption isotherms were determined at 25° and 35°C by means of a gravimetric method reported by Lang and others (1981). Moisture-Sorption Model The Guggenheim-Anderson-De Boer (GAB) model was used to fit sorption data. Thermodynamic Parameters The variation of the molar integral entropy (ΔHint) and the changes in the integral entropy (ΔSint) were determined according to the Beristain and Azuara method (1990). Isochromic Red Fraction of Carotenoids Absorption at 472 and 508 nm of acetone extracts containing carotenoid red fraction (CR) was carried out in a Beckman spectrophotometer (model UV7500i; Beckman Coulter, Fullerton, CA, USA) according to the method reported by Hornero and Mínguez (2001).

Results and Discussion The integral entropies of macadamia (M), yogurt (Y), zeolite clinoptilolite (ZCL), and zeolite Valfor (ZS) are shown in Figure 66.1. As can be seen, all materials presented a minimum integral entropy zone that does not change considerably within a defined range of moisture content. The minimum integral entropy of M, ZCL, Y, and ZS is 1.46, 6.33, 8.81, and 23.5 g water/g dry solids, respectively. Y and ZS have structures with a higher micropore volume, especially nanopores with diameters of <2 mm. This indicates that the control of the porous nanostructure of food materials induced by the increment in the micropore volume displaces the minimum integral entropy water aw to higher moisture contents, thus promoting stability. As water molecules are adsorbed on the adsorbent surface, the area available for adsorption decreases. Therefore, the available sorption surface (Ssp) is the area of dry

684

PART 2: Poster Presentations

0

M Y –20

ZCL ZS

(ΔSint)T(J/mol K)

–40

–60

–80

–100

–120 0

5

10

15

20

25

30

Moisture content (g water/100g dry solids)

Figure 66.1. Integral entropy ΔSint of sorption as a function of moisture content of the different products tested at 35°C. M, macadamia; Y, yogurt; ZCL, zeolite clinoptilolite; and ZS, zeolite Valfor.

adsorbent minus the area that has been covered by water. The Ssp is calculated with the equation of Kiselev (Berezin and Kiselev 1972) and dSsp/daw is obtained as the derivative of the area (Ssp) with respect to aw. Therefore, another way of determining the point of minimum integral entropy or most suitable arrangement of water molecules on the material surface is by graphically analyzing the variation in the available sorption surface in relation to aw as dSsp/daw vs aw (Figure 66.2), where dSsp/daw is a measure of the interactions of the surface area of the adsorbent and water molecules in the system. Before the minimum value of integral entropy, water sorption is dominated by the surface energy of the adsorbent. The minimum dSsp/daw should indicate the surface area that maximizes water-surface interactions. It is likely that the maximum water-surface interaction, therefore minimum dSsp/daw, corresponds to organized arrangement of the adsorbed molecules with minimum integral entropy. The minimum value of dSsp/daw agrees with the minimum integral entropy of the analyzed systems. This new analysis method described here for compensation entropyenthalpy has the advantage of needing only the determination of the sorption isotherm at one temperature. In our work, the minimum entropy and its relation to the degree of structure modification on the nanometric scale (induction of nanostructures) was analyzed by means of the variation in the surface available for adsorption.

Nanostructures and Minimum Integral Entropy as Related to Food Stability

1.5

685

–30 (ΔSint)T(J/mol K) M

1.4

–dSsp/daw M

1.3 –34

1.2

1.1

–36

(ΔSint)T(J/mol K)

–dSsp/daw (m2 x 10–2)

–32

1.0 –38 0.9

–40

0.8 0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

aw

Figure 66.2. Adsorption surface and integral entropy changes as a function of moisture content. aw, water activity.

This can be seen in Figure 66.3 where the value of aw occurs for the best arrangement of water molecules for sucrose-calcium powders of equal chemical composition but with different degrees of nanostructuration. The structure was modified on the nanometric scale by means of a cryogenic process that facilitates the formation of water crystals of nanometric size in the matrix of the sucrose-calcium particles. The creation of a major micropore volume in E-40 and E-80 compared with the sucrose-calcium spray-drying sample (SA) enables the point of minimum entropy to be displaced to a higher aw. The minimum integral entropy or best arrangement of water molecules in foods is useful in predicting the range of relative humidities at which a product should be stored; thus enhancing its physical stability. Chemical stability and its relation to minimum entropy were studied by the degradation of the CR in paprika capsules with and without nanostructures. The best arrangement of water molecules from a consideration of dSsp/daw as a function of aw for AA and ZAA was observed at aw 0.307 and 0.587, respectively. Regarding the stability of the CR in the paprika capsules (AA and ZAA), the carotenoids in capsules with nanostructures were more stable. A 100% carotenoid retention was observed at aw as high as 0.9, whereas the carotenoids in the AA capsules began to degrade at aw 0.495. This analysis was performed after 1 week of storage.

686

PART 2: Poster Presentations

4.0 SA E-40

3.5

–dSsp/daw (m2 × 10–2)

E-80 3.0 2.5 2.0 1.5 1.0 0.5 0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

aw

Figure 66.3. Adsorption surface changes (−dSsp/daw) as a function of water activity (aw) of sucrose-calcium powders at 25°C. SA, by spray drying; E-40, by cryogenic ethanol at −40°C; and E-80, by cryogenic ethanol at −80°C.

It is important to note that the nanostructured capsules retained 2.6 times more CR (2654.57 mg/kg) than did the non-nanostructured ones, perhaps because the nanopores interact more strongly with the carotenoids than do the larger pores.

Conclusions The physical and chemical stability of the different nanoporous systems analyzed in this work was related to the minimum integral entropy or best arrangement of adsorbed water molecules on food surface. The variation in the surface available for adsorption in relation to aw (dSsp/daw vs aw) proved to be an appropriate and simple tool for determining the point of minimum integral entropy by using a single sorption isotherm. This analysis revealed that the modification of the structure of the materials on the nanometric scale induced better stability at higher relative humidities.

Acknowledgments This work has been supported by the Instituto Politécnico Nacional (IPN) (National Polytechnic Institute) of Mexico; Secretaria de Investigación y Posgrado del Instituto Politécnico Nacional (SIP-IPN) research projects; and Comisión de Operación y Fomento de Actividades Académicas (COFAA)-IPN and Consejo Nacional de Ciencia

Nanostructures and Minimum Integral Entropy as Related to Food Stability

687

y Tecnología (CONACYT) (National Council for Science and Technology, Mexico) project 2004-C-1-4861-Z.

References Azuara E, Beristain CI. 2006. Enthalpic and entropic mechanisms related to water sorption of yogurt. Drying Technol 24:1501–7. Berezin GI, Kiselev AV. 1972. Adsorbate-adsorbate association on a homogenous surface of a nonspecific adsorbent. J Colloid Interface Sci 38:227–33. Beristain CI, Azuara E. 1990. Estabilidad máxima en productos deshidratados. Ciencia 41:229–36. Beristain CI, Azuara E, Vernon-Carter EJ. 2002. Effect of water activity on the stability to oxidation of spray-dried encapsulated orange peel oil using mesquite gum (Prosopis juliflora) as wall material. J Food Sci 67:206–11. Beristain CI, García HS, Azuara E. 1996. Enthalpy-entropy compensation in food vapor adsorption. J Food Eng 30:405–15. Domínguez IL, Azuara E, Vernon-Carter EJ, Beristain CI. 2007. Thermodynamic analysis of the effect of water activity on the stability of macadamia nut. J Food Eng 81:566–71. Hornero D, Mínguez MI. 2001. Rapid spectrophotometric determination of red and yellow isochromic carotenoid fractions in paprika and red pepper oleoresins. J Agric Food Chem 49:3584–8. Lang KW, McCune TD, Steinberg MP. 1981. Proximity equilibration cell for rapid determination of sorption isotherms. J Food Sci 46:936–8. Roos YH. 1993. Water activity and physical state effects on amorphous food stability. J Food Process Preserv 16:433–7. Sanguansri P, Augustin MA. 2006. Nanoscale materials development: a food industry perspective. Trends Food Sci Technol 17:547–56. Zeller BL, Saleeb FZ. 1996. Production of microporous sugars for adsorption of volatile flavors. J Food Sci 61:749–52.

Index

A AA. See Alginic acid Accessible volume, antiplasticization phenomenon and, 309 Acetone, dielectric constants of, 159t Acevedo, N. C., 21 Acid gels, structure of, 242 Acid-hydrolyzed cassava starch, differential scanning calorimetry data for, 265t Acrylic monomers, superabsorbent polymers based on, 386 Active ingredients, dual role of glassy carbohydrates in stabilization and delivery of, 354t Active pharmaceutical ingredient hydrates correlation of water mobility as determined by NMR, with that determined by DSC and watersorption measurements, 31, 34–36 ease of evaporation for hydration water as determined by DSC and watersorption isotherm measurements, 31 molecular mobility of hydration water, as determined by NMR, 26–30 molecular mobility of water in, 26–36 usefulness and limitation of water mobility by NMR and, 36 Active pharmaceutical ingredients chemical stability or instability of, in aqueous solutions or suspension formulations, 316 moisture uptake in amorphous pharmaceutical matrices and lipid vehicles and profound effects on physical and chemical stability of, 332

solid, apparent correlations between stability of, and molecular mobility of water, 25–26 water’s importance in, as solvent for, 315 Actomyosin, gel formation and, 241 Acyl-transfer reactions, water uptake and mobility, pharmaceutical stability in amorphous solids and, 325 Adam-Gibbs approach, enthalpy relaxation time of trehalose-glucose-lysine system and, 448 Additives, enthalpy relaxation influenced by, 420 Adhesive forces, surface tension and, 225 Adobe Photoshop 7.0, 516 Adsorption behavior, hysteresis between desorption behavior and, 238 Adsorption cycles, typical, for biopolymers and foods, 238 Adsorption isotherm, of glassy tapiocaflour-based baked sample, 594–596 Adsorption rates desorption rates vs., for starch-rich model bread crust, 169, 169–170 for starch-rich model bread crust during oscillatory sorption experiments, 172 Adya, A. K., 72, 77 Agarose, ice crystallization during rewarming observed with, 373, 381–382 Agave atrovirens Karw effects of different cut-induced microstructural and macrostructural arrays on convective drying of, 619–625 conclusions, 625 introduction, 619–620

689

690

Index

Agave atrovirens Karw (continued) materials and methods, 620–621 convective drying experiments, 620 image processing and fractal analysis used to study shrinkage and deformation, 620–621 results and discussion, 621, 624 drying kinetics, 621 microstructural images, 624 shrinkage-deformation kinetics, 621, 624 Agave slices during convective drying, gallery of binary images from superior and lateral views of: longitudinal cut and transverse cut, 623 dried, scanning electron microscopic images of, longitudinal and transverse cuts, 624, 624 drying kinetics of, at 60°C and 2 m/s for longitudinal and transverse cuts, 622 Agave tissue, structural arrangement of, 620 Aggregation heat gelation and, 239 relative kinetics of denaturation and, 240 Aguamiel juice, extraction of, 619 Ahromrit, A., 668 Al-Bezweni, M., 484 Albumins, heat gelation and, 239 Alginate film, describing sorption dynamics of, 166 Alginate gel, subjected to EDTA, 392 Alginic acid, paprika capsules prepared with, 681, 683 Alkali halide-saccharide-water ternary system, physicochemical characteristics of sodium ions and potassium ions on, 474 Alkyl carbon atoms, in monocaprylin or tricaprylin, radial distribution function for water oxygen with, 331 Almond powder, plasticizing effect of water and addition of, to soy bread, 183

Almonds health benefits with, 176 schematic of changes occurring in soy bread, with and without, during storage, 182 water state and distribution during storage of soy bread, with and without, 175–176 Almond-soy bread “freezable” water in, 181, 181 moisture content and specific loaf volume of soy bread and, 179t Alpha-lactose, protein-carbohydrated sheets prepared with, 578 Alpha relaxation of molecules, 161 Alpha-relaxation process, as main glass transition, 116–117 Altoids car interior temperature rise and effect on, 101 humidifying, review of observations, 102–104 AMBER 8, 329 Ambient food, initial freezing point of, 212 Ambient systems, water measurement and, 204 Ambient temperature systems, relative vapor pressure as control factor and, 212 American Association of Cereal Chemistry, 606 American Society of Testing and Materials Standard Method D882-91, 454 American Society of Testing and Materials Standard Method D1434-82, 454 American Society of Testing and Materials Standard Method D3985, 642 American Society of Testing and Materials Standard Method E96-85, 454 Amino acids model of local reaction promoting nonenzymatic browning reaction progress in glassy matrix and, 574, 575 protein film-forming properties and role of, 453

Index

Amorphous dry foods, glass-rubber transition and affect on chemical and physical stabilities of, 445 Amorphous food matrix, distribution of microstructural domains within, 59 Amorphous food powders factors related to different behavior of crystalline food powders and, 45 stickiness and caking of, 344–345 Amorphous foods relaxation phenomena of, 339–341 enthalpy relaxations, 340 structural relaxations, 341 Amorphous glass self-association of water at higher moisture levels observed in, 332 water uptake and its implications in, 318–326 water mobility in simulated PVP glasses, 322–325 water uptake and its distribution in PVP, 318–322 water uptake and mobility on pharmaceutical stability in amorphous solids, 325–326 Amorphous glassy state, description of, 419 Amorphous growth ring, starch granule, 254 Amorphous lactose, water-sorption and timedependent crystallization and recrystallization of, 347 Amorphous materials, dry-food products and, 571 Amorphous saccharides, different dissolution behavior of crystalline samples and, 42 Amorphous solids, implications of water uptake and mobility on pharmaceutical stability in, 325–326 Amorphous sucrose, pure, first experiments on state changes of, 95–96 Amorphous sugars, crystallization of, 346–348 Amorphous systems, antiplasticizing effect of water when polymers used for stabilization of, 401

691

Amoxicillin correlation between onset temperature and T2 for, 34 correlation between T2 and temperature dependence of T2 for, 35 differential scanning calorimetric thermograms for, 32 Gaussin-type decay and Lorentzian decay exhibited by, 29 Amphiphilic biopolymers, 50 Ampicillin correlation between onset temperature and T2 for, 34 correlation between T2 and temperature dependence of T2 for, 35 differential scanning calorimetric thermograms for, 32 Gaussin-type decay and Lorentzian decay exhibited by, 29 Ampoule method, sticky point by, 100 Amylopectin improved understanding of, 252–253 retrograded, DSX-freeazble water, bread characterization and, 607 Amylopectin clusters, starch granule, 254 Amylopectin lamellae, formation of, 253 Amylose in cassava starch, seasonality and, 258 improved understanding of, 252–253 Amylose content of milled rice varieties, 664 solid loss of rice grains during soaking and, 668–669 Anandaraman, S., 367 Anderson, B. D., 318 Anderson, Philip W., 111 Angel floss, 95 Angle of repose, for maltodextrin powders, evaluating, 675 Anhydrobiosis, trehalose and, 132 Annealing, 419 Annealing step, adding to freeze-drying cycle, 144, 146 Annealing treatments, time dependence of ice formation and, 348 Anomalous antiplasticization effects, of water in relation to some mechanical properties, 122

692

Index

Anomalous diffusion, 166 Antibiotic hydrates correlation between onset temperature and T2 for, 34 correlation between T2 and temperature dependence of T2 for, 35 ease of evaporation for, 31 time courses of spin-spin relaxation observed for, 29–30 Antimicrobials, controlled kinetics of Maillard reaction and, 11 Antioxidants, controlled kinetics of Maillard reaction and, 11 Antiplasticization, 115 depiction of feasible explanation for association between plasticization and, based on bulk free volume, 311, 311–312 external mechanical, by water, 116 hole-filling mechanism and, 362 MCC compaction and, 308 new perspectives on, 408 plasticization vs.: polymer-diluent interactions, 116–119 polymer-plasticizer systems and, 402 polymers and, 302 various theories and models of, in synthetic polymer systems, 117–118 Antiplasticization effect of water in raw and roasted coffee beans, 491 textural properties of raw and roasted coffee and, 494 Antiplasticization-plasticization phenomena, factors affecting physical manifestation of, 120, 122 Antiplasticization range, compressionforce-water-activity relationships for extruded flat breads and, 120, 122 Antiplasticized modulus vs. diluent content curve, of solid polymer-diluent blends, 118 API protons, spin-spin relaxation time of water protons and, 27 APIs. See Active pharmaceutical ingredients Apple drying methods, reaction kinetics and, 21

Apple-juice powder, onset of glass transition temperature of food powders measured by TMCT, DSC, and TMA, 434t Aqualab 2000, 607 Aqualab dew-point hygrometer, 493 Aqueous pharmaceutical formulations, solubility and stability of API issues, 315 Aqueous phase, 91 of dough, 50, 52, 53 Aqueous systems, quantifying fraction of water contributing to component of spectroscopic signal in, 216 Arabinose, intermolecular hydrogen bonds, NEB progress in the glassy matrix and, 576 Aroma, as major determinant of fruit quality, 657 Arrhenius equation, 111 Arrhenius plot of alpha relaxation and beta relaxation of dielectric spectra for proteincarbohydrate sheets, 580, 580 of initial nonenzymatic browning reaction rate for samples of varying glassy matrices, 573, 573 of initial nonenzymatic browning reaction rate for samples of varying reducing sugars embedded in glassy trehalose matrix, 574 for temperature dependence of water diffusion coefficient for different carbohydrate concentration and different molecular weights, 364, 365, 366 Arrhenius relationship, 66, 98 Aspergillus flavus combined effect of cinnamon essential oil and water activity on, 545–550 introduction, 545 materials and methods, 546–547 results and discussion, 547, 549–550 effect of cinnamon and relative humidity on growth inhibition of Rhizopus stolonifer on bread during storage at 30°C, 549

Index

effect of cinnamon oil and water activity on inhibition of, on bread model agar, 546, 547, 548, 549 minimum inhibitory concentrations of essential oils and potassium sorbate against, 547t Aspirin matrix average fractal dimension of edge of, within first 30s of their disintegration and diffusion, 516, 516 average fractal dimension of side face of, within first 30s of their disintegration and diffusion, 517, 517 evaluation of disintegration and diffusion of, by image processing and nonlinear dynamics, 515–520 images of, in degasified distilled water and without agitation, showing rings, halo, and vortices around, 517, 518 proposed mathematical simulation for one of the vortices during diffusion of, image at 10s; amplification of modeled vortex; digitalized vortex; iterative model; 20° clockwiseturned vortex, 519, 519–520 at 16s after being placed in degasified distilled water and without agitation, vortices arrangement in opposite-sense movement around halo formed, 518, 519 showing disintegration process comparable to opening of petals of flower, with fractures during such opening, 517, 520 Attenburrow, G. E., 123 Average mobility of water, 256 Avrami equation analysis of release characteristics of encapsulated flavor by, in response to various influence factors, 502 moisture content modeling for crystallizing amorphous lactitol with, 273 rate of crystallization of polymers modeled with, 346–347 Azuara, E., 682, 683

693

B Baik, M. Y., 181 Bakmix (innovative bread mixer), 607 recent design of, 605 schematic representation of, 606 Bakmix mixing process, traditional mixing process vs., 606 Barford, J. P., 53, 55t Barrier properties, of duck egg-white film made from shell eggs, 457, 458t BBSRC. See Biotechnology and Biological Sciences Research Council Beef, cooked, change in distribution of water-proton relaxation times during compression of, 244 Benczédi, D., 126 Benzamide (single solute molecule), in molecular dynamics simulation of 60% tricaprylin-40% 1-monocaprylin lipid mixture saturated with water at 37°C, 329, 330 Benzene, dielectric constants of, 159t Berberine chloride differential scanning calorimetric thermograms for, 32 endothermic peaks for, 31 Beristain, C. I., 682, 683 Bertram, A. K., 470 Beta-carotene, encapsulated, glass transition temperatures as function of mass fraction of water for polymeric and trehalose matrices and, 19 Beta-cyclodextrin, water-sorption isotherms of beta-cyclodextrincinnamaldehyde and betacyclodextrin-thymol complexes and, 152 Beta-cyclodextrin-cinnamaldehyde system, differential scanning calorimetry thermograms of, 153–154, 154 Beta-cyclodextrins materials and methods, 150–151 differential scanning calorimetry, 151 preparation of solid complexes, 150–151 sorption isotherms, 151 storage study, 151

694

Index

Beta-cyclodextrins (continued) release of thymol and cinnamaldehyde during storage, 154–155 results and discussion, 152–154 differential scanning calorimetry, 153–154 sorption moisture studies, 152–153 water content and structure of, 149 water-sorption properties and stability of inclusion complexes of thymol and cinnamaldehyde with, 149–155 introduction, 149–150 Beta-cyclodextrin-thymol system, differential scanning calorimetry thermograms of, 153, 153 Beta-lactoglobulin gels, 10%, lightmicroscopic images of, at pH 5.3, 241 Beta relaxation, 419 of molecules, 161 Beta-suppression effect, 118 BET value. See Brunauer-Emmett-Teller value BFV. See Bulk free volume Bhandari, B. R., 160, 195, 430 Bidwell method, mass-loss kinetics for bread dough calculated with, 629 Biexponential decay curves, for quinoa seeds equilibrated at low RHs, 650 Binary aqueous system temperature-composition phase diagram for, 213 temperature-composition state diagram for, 214 Biodegradable food-packaging, chitosan and, 459 Biological materials, water-activity theory on stability of, 158–159 Biological systems water in, 157 water’s crucial role in pressure-induced denaturation of, 80 Biologics, hypothetical states of matter of, 90 Biomaterials antiplasticizing effects of water on, 123–124 design of new, biomimetics and, 385

Biomimetics, influence and role of, 385, 395 Biomolecular dehydroprotectant agents, optimizing efficiency of, 18 Biomolecular functionality or structure, controlled kinetics of Maillard reaction and, 11 Biomolecules, encapsulation of, in amorphous matrix formed during dehydration process, 18–19 Biopolymer-based films, anomalous antiplasticization effects of water and mechanical properties in, 122 Biopolymer films prepared from aqueous solutions fracture behavior of, 291–296 introduction, 291 materials and methods, 291–292 mechanical tests, 292 preparation of specimens and conditioning, 291–292 visualization techniques, 292 results and discussion, 292–296 effect of water content on physicochemical and structural properties of biopolymer films, 292–293 post-deformation plastic zone of biopolymer films, 295 prediction of dependence of fracture characteristics on water content for: brittle-ductile transition, 295–296 in situ visualization of crack propagation, 293–294 Biopolymer gels, ice crystallization during rewarming observed with, 373 Biopolymeric matrix, fluid transport and complex flow path presented by, 232 Biopolymers edible films developed from, for over 40 years, 453 heterogeneity of, microscopically, 192 natural interaction between water and, 291 typical adsorption-desorption cycles for, 238

Index

Biotechnology and Biological Sciences Research Council, 296 Biscuit, dry, Fickian diffusion used to describe sorption dynamics of, 166 Blanched slices, of potato chips, 526 Bligh and Dyer methodology, potato chip oil content determined by, 526 Blocklets, starch granule structure, 253, 254 Blomberg, A., 76 Blond, G., 105 Blood cholesterol levels, soy-fortified diet and lowering of, 176 Blood plasma gels, moisture loss as function of temperature and salt concentration for, 243 Boiling-point curves, on state diagram, 91 Boonyai, P., 430 Borges, A., 120 Bouchon, P., 525, 527 Bound water, 115, 127 discrete-model assumption and describing free water vs., 256 Bound water fraction, 51 Bourne, M. C., 105, 121t Bovine serum albumin effects of, on storage stability of xanthine oxidase, 599, 600, 601, 601t, 602, 602t, 603 mixtures of, in probiotic cultures, 285 Bragg 2-theta angle peaks, powder X-ray diffraction of initial dry cotton candy and, 96, 97 Branch chains, distribution of, in amylopectin molecules, 253 Bread, extruded flat, compression-forcewater-activity relationships for, showing antiplastication range, 120, 122 Bread crust dry, crispy, water sorption and transport in, 165–173 conclusion, 173 introduction, 165–166 materials and methods, 166–167 results and discussion, 167–173 oscillatory sorption experiment on, 170, 170–171

695

starch-rich adsorption and description rates from best fits between relative change of weight, during isotherm experiment and single exponential model as function of RH, 169 adsorption and desorption rates for, during oscillatory sorption experiments, 172 measured relative change in weight of, and adjusted external relative humidity, 167–168, 168 Bread model agar combined effect of selected essential oils and water activities on inhibition of molds on, 545 effect of cinnamon oil and water activity on inhibition of Rhizopus stolonifer and Aspergillus flavus on, 546 Bread produced in innovative mixer bread characterization, 607–608 DSC-freezable water and retrograded amylopectin, 607 image analysis, 607 loaf volume, 607 moisture content, 607 NMR measurements, 607–608 statistical analysis, 608 texture analysis, 607 water activity, 607 materials and methods, 606–608 bread formulation and production, 606–607 water properties in, 605–611 conclusions, 611 introduction, 605–606 results and discussion, 608 Bread production, discontinuous ingredient mixing in, 605 Bread quality and stability, factors related to, 605 Bread shelf life application of cinnamon and controlled relative humidity in extension of, 546–547, 549–550 bread preparation, 546

696

Index

Bread shelf life (continued) inoculation, incubation, and growth measurement, 546–547 statistical analysis of data, 547 effect of cinnamon essential oil and water activity on growth inhibition of Rhizopus stolonifer and Aspergillus flavus and possible extension of, 545–550 Bread spoilage, mold growth and, 545 Bread Volume Measurer BVM-L370, 177 Breakfast cereals expanded solid foams and, 247 moisture content and loss of crispness in, 106 Breeding programs, cryopreserved gametes used in, 551 Brittle-ductile transition, 106 determining, 107 prediction of dependence of fracture characteristics on water content for biopolymer films associated with, 295–296 Brittle-ductile transition temperature, 106 Brittle-ductile transition temperature plot, as function of moisture at constant temperature for sugar snap cookie, 108 Brittle fracture, distinguishing, 106 Brookfield model RVDV III viscometer, viscosity of egg-white solutions determined with, 454 Brooks and Corey model, RC described in soil science by, 226 Broth culture, influence of glucose on fructose uptake in, 53 Browning development, in freeze-dried potato discs, 439 Browning index defined, 540 of 50% wt/vol systems during storage at 70°C, 541, 542 Browning rate as a function of temperature and mass fraction of water, 15 Maillard reaction and, 11, 14 Browning rate constants, for lactose, milk, and lactose-starch systems, 14

Brunauer-Emmett-Teller (BET) monolayer, 89 Brunauer-Emmett-Teller value, coffee beans, antiplasticization effect of water and, 492 Brunauer-Emmett-Teller water-sorption isotherms, 478 BSA. See Bovine serum albumin Bulk acoustic wave sensor, aroma analysis and, 658 Bulk and tapped density, of maltodextrin powders, evaluating, 675 Bulk free volume accessible volume playing role of, 309 depiction of feasible explanation for association between antiplasticization and plasticization, based on, 311, 311–312 Bulk water fraction, 50–51 Butter fat effect of water and fat contents on enthalpy of dissolution of, 42 pure, measured enthalpy of dissolution of, 43 C Caking of amorphous food powders, 344–345 food storage and, 93 moisture pickup and, 95 during storage of powders containing amorphous sugars, 100 Caking point temperature, advance, 100 Calcium, water content of, at 10% RH and 60% RH, 37t Calcium lactate effect of water content and, on phasetransition characteristics of mungbean starch in first endotherm, 509–510, 509t mung-bean starch and effect of, on transition temperatures in second endotherm at different levels of water content, 510 mung-bean starch granules and findings related to, 512

Index

Calibration curves, water content determination and, 207 Calvet calorimeter, enthalpy of dissolution for model food powders quantified isothermally in, 43 Cambridge Crystallographic Data Centre, 478 Campbell, G. S., 228 Candy, glass-rubber transitions measured for, 435 Candy formulation, temperature effect on relaxation behavior of, 425t Capillarity, defined, 225 Capillarity and tension head, flow in unsaturated porous media and, 224–225 Capillary flow mass transport in hierarchical structure and, 394 in porous media, 222–224 schematic illustration of mass transport by, 395 Capillary-flow approach to flow of liquids into dried foods, 233 for modeling temperature and anisotropic effects during rehydration of tea leaves, 231–232 Capillary imbibition theory, for modeling rehydration of foods, 231 Carbohydrate classes, effects of water on structure of, 356–357, 359 Carbohydrate hole sizes, suggested mechanisms leading to, increasing with water content in glassy state, 358 Carbohydrate matrix, varying molecular weight of, influencing membrane phase-transition temperature by, 368 Carbohydrate molecular weight, effect of, on matrix density and on glass transition temperature, 367 Carbohydrate polymers, disparate effect of water and low molecular weight sugars as plasticizers for, 362 Carbohydrates in amorphous states, 353–370 conclusions, 369–370 dynamic properties close to glass transition, 364, 366

697

effects of water on structure of carbohydrate glasses, 356–357, 359 glassy carbohydrates in food and pharmaceutical stability, 353–356 molecular packing in glassy carbohydrates, 359, 361–362 technological implications, 366–369 Carbohydrate-water systems dynamic properties close to glass transition, 364, 366 outline of phase transitions and physical states of, in nonfrozen state at constant temperature, 354 snapshot from molecular dynamics simulation of, 359 Carboxymethyl cellulose, water content of, at 10% RH and 60% RH, 37t Carboxymethyl starch, water content of, at 10% RH and 60% RH, 37t Car interior temperature, hard candy exposure to, 101–102, 102, 104 Carotene degradation, 1820 magnitude of kinetic constants in, 19 Carotenoids, isochromic red fraction of, 683 Carrot fiber, Tg from DSC and corresponding DEA frequencies for, in freezedried amorphous fructose, 195t Carrot fiber as carrier in spray drying of fructose, 191–196 conclusion, 196 introduction, 191–193 materials and methods, 193–194 results and discussion, 194–196 carrot fiber powder size and sugar content, 194 DSC and DEA, 194–195 spray drying of fructose mixtures, 195–196 Carrot fiber-fructose, spray-dried, Tg and yield from DSC of maltodextrinfructose and, at different ratios, 196t Carrot fiber powder size, sugar content and, 194 Carrots, volume decrease or shrinkage studied in, 613

698

Index

Carr-Purcell-Meiboom-Gill (CPMG) pulse sequence coffee samples analyzed with, 494 evaluating NEB inhibitions or accelerations by magnesium chloride, water availability and saccharide-specific interactions with, 539, 541 gelatinization of cassava starch from drought and rainy seasons and water mobility measured by, 263 NMR experiment, microdomain distribution in food matrices, 60 starch study and T2 measurement by, 258–259 water mobility measured by, acidmodified cassava starches diluted with water during heating and, 266 water mobility measured by, for cassava starches saturated with water during freezing, 264 Carr-Purcell-Meiboom-Gill echo train, hydration kinetics of cracker samples and, 414 Casein cold gelation and, 242 unique structure of, 237 Casein films and composites anomalous antiplasticization effects of water and mechanical properties in, 122 sorption dynamics of, 166 Case II diffusion defined, 166 water sorption and transport in dry, crispy bread crust and, 165, 173 Cassava, as unique drought-resistant crop, 252 Cassava starch acid-hydrolyzed, differential scanning calorimetry data for, 265t acid-modified, diluted with water during heating, water mobility measured by Carr-Purcell-Meiboom-Gill (CPMG) pulse sequence, 266

from drought and rainy seasons, water mobility measured by Carr-PurcellMeiboom-Gill (CPMG) pulse sequence, 263 effect of water content on physicochemical and structural properties of, 292 five freeze-thaw cycles, T2 distribution of protons in, 267, 268 gelatinized, moisture-content measurements for, 186 granular swelling upon gelatinization of, grown in drought and rainy seasons and harvested at 6 and 12 months, 260, 261 saturated with water during freezing, water mobility measured by CarrPurcell-Meiboom-Gill (CPMG) pulse sequence, 264 study of mechanical response to tensile loading of, 291 undergoing gelatinization, SEM and TEM micrographs of, 255 versatility and differing functionality of, 258 Cassava starch film fracture characteristics of, brittle-ductile transition and, 296 infrared maps for, in the hydrated state, 295, 295 multicrack fracture mechanism of, 293 in situ visualization of stable crack propagation for, at 69% relative humidity by using the high-speed camera, 294, 294 Castor oil, water content of, 328 CCF. See Commercial cornflakes CCF designs. See Composite face-centered designs CD. See Cell density CDs. See Cyclodextrins Cefazidime, differential scanning calorimetric thermograms for, 32 Cefazolin sodium correlation between onset temperature and T2 for, 34 correlation between T2 and temperature dependence of T2 for, 35

Index

differential scanning calorimetric thermograms for, 32 Gaussin-type decay and Lorentzian decay exhibited by, 29 water-sorption isotherms for, 33 Cefazolin sodium hydrates, water-sorption isotherms observed for, 31 Cefoxitin sodium, times for 10% degradation of, compared with measurements of structural relaxation and enthalpy relaxation times, 326t Ceftazidime correlation between onset temperature and T2 for, 34 correlation between T2 and temperature dependence of T2 for, 35 Gaussin-type decay and Lorentzian decay exhibited by, 29 Ceftazidime hydrates, temperature dependence of T2 for, 34 Cell density image processing of bread dough and, 632, 633 Cell mortality, changes in fluidity of plasma membrane and, 76 Cells effect of combined physical stresses on: role of water, 71–83 example 1: effects of combined hyperosmotic and temperature perturbation, 72–77 example 2: effects of combined high hydrostatic pressure, low temperature, and hyperosmotic perturbations, 77–82 Cellular wall collapse, shrinkagedeformation behavior during drying of potato slabs and, 613, 616, 617 Cellulose, in fibers, 192 Cellulose extraction, Agave atrovirens Karw, convection drying and, 619 Center of Excellence on Supramolecular Biomaterials: Structure Dynamics and Properties, 385 Center temperature, changes in dough/ bread-crumb structure and, 631, 631

699

Cephalothin, hydrolysis rate of, in freezedried formulations, molecular mobility of water and, 26 Cereal cracker sample, water-sorption isotherms for, 416 Cereals, Fick’s laws of diffusion and water absorption of, 663 Cereal snacks, dry, perception of crispness of, 105 Chadmo quinoa seeds amplitude of signal of lipid component of, with different water activities, 652 genotype description of, 648 high viability and germination values for, 650 percent germination of, during storage at 32°C and 43% or 75% relative humidity, 650, 650 percent viability of, assessed by tetrazolium test, 651 relaxation times of lipid component of, with different water activities, 652 sorption isotherms of, 651t storage behavior of, 654 Chang, Y. P., 120, 121t, 122, 131 Charge-coupled device (CCD) digital cameras, shrinkage and deformation of agave slices evaluated during convective drying with, 619, 620 Charmathy, S. P., 126 Charoenrein, S., 642 Chatakanonda, P., 258, 440 Chemical titration, quantification of water by, 207 Chen, P. L., 259 Cheng, L. H., 125, 132 Chenopodium quinoa adaptability of, to extreme environmental conditions, 648 materials and methods, 648–649 seed material, 648 molecular mobility and seed longevity in, 647–654 discussion, 654 introduction, 647–648 results, 650–651

700

Index

Chenopodium quinoa (continued) NMR relaxation measurements, 649 seed conditioning, 649 spin-spin relaxation time measurement, 649 seed storage behavior, 648–649 germination assessment, 648–649 seed storage, 648 viability test, 649 sorption isotherms, 649–650 Chevallier, S., 628 Chiang rice, hydrothermal analysis of highamylose rice starch from, 635 Chinachoti, P., 181 Chirife, J., 14, 439 Chitosan, uses and applications with, 459 Chitosan film describing sorption dynamics of, 166 disadvantages with, 459 Fickian diffusion for, 166 Fourier transform spectroscopy spectra of, 461, 461 improving water-resistant properties of, 459 scanning electron micrographs of surfaces and fracture surfaces of, 462, 462 Chitosan/MPEG-Beta-PCL blend homogeneous flims materials and methods, 460–461 blend films preparation, 460 chitosan and blend films characterization, 460–461 materials, 460 results and discussion, 461–463 FTIR analysis, 461–462 morphology, 462 water-vapor permeability, 462–463 water-vapor permeability of, 459–463 conclusions, 463 introduction, 459–460 Chitosan/MPEG-Beta-PCL blend ratio, rate of water-vapor permeability, 462–463, 463t Chocolate wafers, moisture content and transition temperatures of, 110 Chromatographic techniques, water content estimating with, 207

Chuy, L., 100 Cicerone, M. T., 132 Cinnamaldehyde in cinnamon essential oil, 150 release of, during storage, 154–155, 155 sorption characteristics of betacyclodextrins formed with, and their release, 149 storage study on, 151 water sorption and stability of betacyclodextrin complexes with, 155 Cinnamon, ground, application of, and controlled relative humidity in extending bread’s shelf life, 546, 549–550 Cinnamon bark oil, minimum inhibitory concentrations of and potassium sorbate against Rhizopus stolonifer and Aspergillus flavus, 547t Cinnamon essential oil, combined effect of water activity and, on growth inhibition of Rhizopus stolonifer and Aspergillus flavus and possible application in extending bread’s shelf life, 545–550 Cinnamon oil antifungal activity of, 547 effect of water activity and, on inhibition of Rhizopus stolonifer and Aspergillus flavus, on bread model agar, 547, 548, 549 preliminary examination of, for antifungal activities by an agar dilution method, 545 Citrus juice, studies of spray drying of, with maltodextrin as a carrier, 196 Citrus oils antifungal activity of, 547 carbohydrate-encapsulated, oxygen uptake by, 367–368, 368 Clathrate, comparison of orientation about oxygen-oxygen hydrogen bond for, 206 Clathrate cage lattice, schematic of, 205 Clathrate hydrates, 204 Clathrate lattice, ice lattice vs., 204 Clostridium botulinum, 89

Index

Cloudberry ellagitannins microencapsulation and improving storage stability of, 563–568 conclusions, 568 introduction, 563–564 materials and methods, 564–565 results and discussion, 565–568 storage stability, 566–568 water sorption and glass transition temperature, 565–566 Cloudberry extract, encapsulated and nonencapsulated, storage stability of, at different relative humidities evaluated as retention of ellagitannins, content of ellagic acid, and antioxidative acitivity, 567, 567t Cloudberry phenolics antioxidative and antimicrobial activity of, 563 extraction of, 564 CLSM. See Confocal laser scanning microscopy Coarcervates, formation of, 242 Coffea arabica, 492 Coffea canephora, 492 Coffee raw normalized differential scanning calorimetry thermograms of, rehydrated at water activity 0.11 and 0.94, 495, 495–496 uniform-penalty inversion T2 relaxograms of, rehydrated at different water contents, 496, 496–497 raw and roasted, GuggenheimAnderson-de Boer parameters computed by fitting of sorption isotherms of, 494t rehydrated green and roasted, normalized fracture force of, as a function of water activity, 495, 495 Coffee beans materials and methods, 492–494 DSC measurements, 493 materials, 492 NMR relaxation measurements, 494

701

textural properties, 493 water-sorption isotherm, 492–493 results and discussion, 494–497 DSC results, 495–496 NMR results, 496–497 textural properties, 494–495 water-sorption discussions, 494 water state and mobility affecting mechanical properties of, 491–497 conclusions, 497 introduction, 491–494 Cohesive forces, surface tension and, 225 Cojazzi, G., 381 Cold gelation, 242 Cold inflection point, for bread dough, 630 Collagen, protein films developed from, 453 Collapse defined, 341–342 in freeze drying, 343 Collapse phenomena in food systems, 341–345 collapse temperatures, 345 effect of thermal and water plasticization on mechanical, flow and, on relaxation times and timedependent changes in food solids, 340 equilibrium and nonequilibrium states and, 337 importance of, 335 reasons for occurrence of, 342 stickiness and caking, 344–345 time dependence and, 336 Collapse properties, of amorphous food solids in dehydration at various water contents and, 343 Collapse temperatures, of freeze-dried materials, 345 Collisions, between reactant molecules, 158 Colloidal phase separation, in dough, 50 Colorants, cyclodextrins and solubility, stability, and bioavailability of, 150 Color changes nonenzymatic browning and, 445 in potato slices during frying, 525, 529, 530

702

Index

Color evolutions, of blanched potato slices fried at 180°C, 529, 530 Color function calculations, in freeze-dried potato disc study, 439 Color values of duck egg-white film made from shell eggs, 456t of duck egg-white films, 455 Commercial cornflakes differential scanning calorimetry thermograms of, at 7.5% and 5.2% water content, 585, 585 Hahn spin-echo transverse relaxation time values vs. water content for, 586– 587, 587 maximum force of compression and onset of glass transition temperature at different water contents for, 587, 588 relationship between relaxation time constant determined by freeinduction decay analysis and temperature for, 586, 586 water content for, 583, 584 Compaction, balance of factors related to, 308 Compaction pressure plots of porosity vs., compressibility of microcrystalline cellulose under different conditions of relative humidity and, 306, 306 relationship between tensile strength and, in microcrystalline cellulose, 309 Compartmentalized water concept, for water in polymer gels vs. from water in pores, 373 Compartmentalized water in polymer gels factors influencing ice crystallization of, 377 freezing scheme for, 379–380 scheme for ice crystallization of, 380t Composite face-centered designs, 143–144 Compression-force-water activity relationships, for extruded flat breads showing wide antiplasticization range, 122

Compression pressure, effect of, on mechanical strength of MCC compacts, 308–309 Concentrated solution, liquid state of water and, 205, 206 Conducting polymer sensor, aroma analysis and, 658 Confectionary product changes availability of water, lactitol crystallization and, 271–281 conclusions, 281 glass transition temperature and, 278, 280–281 introduction, 271–272 materials and methods, 272–275 determination of moisture content, 273 determining melting enthalpy, melting temperature, and glass transition temperature, 274 modeling glass transition temperature, 274–275 modeling of moisture content of crystallizing amorphous lactitol, 273–274 preparation of amorphous lactitol, 272 preparation of saturated salt solutions, 272–273 sample preparation for differential scanning calorimetry, 274 results and discussion, 275–278 melting temperatures of lactitilol crystallites, 275–276 rate and extent of crystallization, 276–278 Confectionary products sugar as major component of, 271 water activity level and maintaining quality of, 271–272 Confined water relaxation time of, 408–409 restricted mobility of, 409 Confocal laser scanning micrographs, of mung-bean starch gel containing 60% water content (wt/wt) prepared by autoclaving at 121°C, 511, 511

Index

Confocal laser scanning microscope jellyfish images taken with, 389, 390 microstructure of jellyfish characterized by, 385 Confocal laser scanning microscopic images of cells and fibers in mesoglea, 391 showing area of mesoglea consisting of open network structure composed of polysaccharides and proteins, 392, 392 Conoidial suspension, A. flavus and R. stolonifer, fungal strains and preparation of, 546 CONTIN computer program, continuummodel approach pursued with, 258 Continuous models application of, to NMR relaxometry, 259 of proton relaxation, in interpretation of exponential decay curve, 257 Continuum models, for water mobility (proton spin-spin relaxation time, T2), 62–64 Convective drying evaluating deformation and shrinking of potato slabs during, 613–617 experiments on Agave atrovirens Karw, 620 image processing and fractal analysis used to study shrinkage and deformation of agave slices during, 620–621 Convenience foods, continued demand for, 219 Cookies dynamic mechanical thermal analysis of, 106 moisture content and loss of crispness in, 106 sugar snap brittle-ductile transition study on, 107 brittle-ductile transition temperature plot as function of moisture at constant temperature for, 108 wafer, stress/strain plot for, at 0.05% relative humidity and 23°C, 109 Cool, dry storage, for hard candies, 105 Cooling rate, free volume of molecules within amorphous matrix and, 338

703

Cooperative alpha relaxation, 419 Copolymeric food packaging, anomalous antiplasticization effects of water in relation to some mechanical properties demonstrated in, 122 Copovidone, water content of, at 10% RH and 60% RH, 37t Corn, leveling of melting enthalpies for, 277 Corn curls, extruded, loss of crispness in, 105 Corn flakes thermal transitions, mechanical properties and molecular mobility in, as affected by water content, 583–588 introduction, 583–584 materials and methods, 584–585 results and discussion, 585–588 Corn powder hydration experiments on, 414 water-sorption isotherms and dependence of relative mobile tdNMR signal on hydration for, 417 Cornstarch, native, glass-rubber transitions measured for, 435 Corn-syrup solids representative calorimetric thermograms of, 422 representative differential scanning calorimetric thermograms of, indicating increased enthalpy relaxation overshoot at higher water activity, 424, 424 Corn syrup-sucrose mixture enthalpy relaxation as function of corn syrup-sucrose ratio, 426 as function of temperature, 427 as function of water activity, 424–425 glass transition temperature value of, 421t analysis of enthalpy relaxation data, 421–423 kinetics of enthalpy relaxation in, 419–427 conclusion, 427 introduction, 419 Kohlrausch-Williams-Watts parameters for, 423t

704

Index

Corn syrup-sucrose mixture (continued) materials and methods, 420–421 determination of enthalpy relaxation kinetics, 420–421 sample preparation in glassy state at various water activities, 420 observed and predicted (KohlrauschWilliams-Watts model) enthalpy relaxation values for, at aging temperature of 22°C, 426 representative differential scanning calorimetric thermograms of, at various proportions, 425 results and discussion, 421–427 temperature effect on relaxation behavior of, 425t Corn syrup-sucrose mixture, representative calorimetric thermograms of, 422 Corn syrup-sucrose ratio, enthalphy relaxation as function of, 426 Corn zein, protein films developed from, 453 Cosmetic industry, cyclodextrins used in, 150 Cotton candy, 88, 111 fresh, X-ray diffraction (XRD) pattern for, 96, 97 history behind, 95 initial dry, powder X-ray diffraction of, 97 sugar recrystallization in storage of foods: sucrose and, 95–100 X-ray diffraction powder pattern of, stored for 2 hours at 45% relative humidity and 23°C, 97 Covalent bonds, 158 CPAs. See Cryoprotective agents CPMG pulse sequence. See Carr-PurcellMeiboom-Gill pulse sequence Crackers dynamic mechanical thermal analysis of, 106 moisture content and loss of crispness in, 106 Cracker sample, porous, water-sorption isotherms for, 415

Crack propagation in situ visualization of, 293–294 stable, in situ visualization of, for hydroxypropyl cellulose film and cassava starch film at 69% relative humidity by using high-speed camera, 294, 294 Crank, J., 167 Crispness consumer appreciation for, 165 of dry, crisp products, control of, 592 of glassy tapioca-flour-based baked samples, sensory approach to, 595–596 loss of food storage and, 105–108, 111 moisture pickup and, 95 potato chips processing and, 526 in tapioca-flour-based baked products, 591–592 rating for, 593 Crisp snacks, 88 Crispy Cracks: Creating and Retaining the Crispness of Food Symposium (University of Wageningen, Netherlands), 105 Cross-linked dextrans containing small amount of water, glass transition of, 381 ice crystallization exotherm during rewarming observed with, 374– 375, 377 Crowe, J. H., 16, 368 Cryogenic ethanol, at −40°C and −80°C, adsorption surface changes as function of water activity of sucrose-calcium powders at 25°C by, 686 Cryomicroscopy, freeze-thaw behavior of aqueous glucose solutions within concentration range of 10%-60% wt/wt studied by, 466 Cryopreservation description of process, 551–552 ice and freeze injury of cells during, 551 improving cell survival during, 552

Index

Cryoprotective agents nonpermeable, types of, 551 permeable, types of, 551 Crystal hydrates, 204 Crystal lattice, 204 Crystalline food powders, factors related to different behavior of amorphous food powders and, 45 Crystalline samples, different dissolution behavior of amorphous saccharides and, 42 Crystallinity amylopectin lamellae, 253 polymer-diluent interactions and, 119 Crystallinity index, 126 Crystallization of polyols, more research needed for, 272 in pure-water ice, generally observed in the hexagonal dendrite, 466 short-range mobility and, 94 of sugars, food properties and, 271 time-dependent, of amorphous lactose, 347 Crystallization phenomena, 345–348 crystallization of amorphous sugars and, 346–348 as glass transition-related timedependent change in low-water and frozen foods, 336 ice formation, recrystallization and, 348 Cutins, in fibers, 192 Cyclodextrins composition of, 149 uses for, 149–150 Cysteine effect of water content and, on phasetransition characteristics of mungbean starch in first endotherm, 509–510, 509t influence of, on starch network architecture at intermediate water content, 511–512 mung-bean starch and effect of, on transition temperatures in second endotherm at different levels of water content, 510

705

D Dairy foods, dehydrated, lactose glass transition in, 336 Dairy powders, sticking or caking, during storage of, 100 Darcy equation, fluid flow in porous media and, 233 Darcy flow of gases, due to pressure, 233 Darcy flow of liquid, due to gas and capillary pressures, 233 Darcy law, flow equation in saturated media and, 227 Davies, R. J., 51 DEA. See Dielectric analysis Decagon, 90 Decoupling of mobility of water, from mobility of carbohydrate molecules in approach to glass transition, technological implications of, 366–369 Deep-fat-fried potato chips, loss of crispness in, 105 Defects, in glassy structure, 419 Deff-moisture-content curve, form of, resembling that of modulusmoisture-content curve of antiplasticized system, 129 Dehydrated foods Arrhenius relationship and reactions limiting storage stability of, 98 stability of, main requirements for, 342–343 Dehydration manipulating food composition to control collapse phenomena in, 335 membrane structure and fluidity and, 71 sequence of hyperosmotic perturbation of yeast cells and, 72 understanding plasma membrane changes occurring during, 77 Dehydration characteristics, importance of collapse phenomena in food systems and, 335 Dehydration process encapsulation of biomolecules, in amorphous matrix formed during, 18–19 freezing process vs., 338

706

Index

Dehydration step, interdependence of changes occurring in rehydration step and, 74 De la Fuente, G., 55t Del Nobile, M. A., 166 Denaturation heat gelation and, 239 relative kinetics of aggregation and, 240 Dense glass, endotherm around glass transition and, 340 DENT specimens. See Double-edged notched tension specimens de Pablo, J. J., 543 Desiccated oils, lower solubility for estradiol and testosterone in, vs. in hydrated oils, 330, 332 Desiccators, saturated salt solutions used to maintain fixed relative humidity levels in, 303t Desorption behavior, hysteresis between adsorption behavior and, 238 Desorption cycles, typical, for biopolymers and foods, 238 Desorption rates adsorption rates vs., for starch-rich model bread crust, 169, 169–170 for starch-rich model bread crust during oscillatory sorption experiments, 172 Deteriorative reactions, analyzing kinetics of, 21 Deterministic chaos, solid matrix diffusion studied in frame of, 520 Deuterium T1 of, correlation between, and crystallization rate of amorphous nifedipine in solid dispersion formulations, 26, 27 T1 of, correlation between and hydrolysis rate of cephalothin in freeze-dried formulations, 26, 26 Devitrification, 348 Dextran gels, cross-linked, ice crystallization exotherm during rewarming observed with, 374–375, 377

Dextrans cross-linked glass transition of, containing small amount of water, 381 ice crystallization during rewarming of, 373 effects of, on storage stability of xanthine oxidase, 599, 600, 601, 601t, 602, 602t, 603 Dickenson, E., 245 Dielectric analysis, 192 finding glass transition temperature from freeze-dried mixtures of carrot fiber plus fructose with use of, 191, 194–195 physical state of protein-carbohydrate sheets verified by, 579 Dielectric constants, of various compounds, 159t Dielectric disturbance, time-dependent response of amorphous materials to, 341 Dielectric relaxation, in amorphous materials below and around glass transition, 342 Dielectric relaxation spectroscopy, 25 probing properties of carbohydrates in glassy state with, 355 Differential scanning calorimetric thermograms for active pharmaceutical ingredient hydrates, 32 of corn-syrup solids, indicating increased enthalpy relaxation overshoot at higher water activity, 424, 424 of corn syrup-sucrose mixture at various proportions, 425 Differential scanning calorimetry, 10, 14, 25, 192 amorphous lactitol sample preparation for, 274 denaturation measured by, 239 determining heat needed to melt thymol or cinnamaldehyde in complexes with use of, 151 ease of evaporation for hydration as determined by, and water-sorption isotherm measurements, 31

Index

fibrous and dense protein composites inspected by, 247 finding glass transition temperature from freeze-dried mixtures of carrot fiber plus fructose with use of, 191, 194–195 findings related to characterization of water fraction of soy bread with, 175 freezable water in soy bread and almondsoy bread obtained from, 181, 181 freeze-thaw behavior of aqueous glucose solutions within concentration range of 10%–60% wt/wt studied by, 466 glass transitions for potato discs determined with, 438 measuring distribution of water in bread with, 176 onset of glass transition temperature of food powders measured by, 434t physical state of protein-carbohydrate sheets verified by, 579 solvent water in dough and, 51 state diagram generation and use of, 214 thermal properties of food-grade starches determined with, 508 Differential scanning calorimetry data, for acid-hydrolyzed cassava starch, 265t Differential scanning calorimetry rewarming traces obtained with gelatinized potato-starch gels, 382 obtained with Sephadex gels dependent on prior freezing, 376 obtained with Sephadex G25 gels containing small amount of water, 381 Differential scanning calorimetry studies of amorphous sugars, 346 on freezing behavior of polymer gels, 373 Differential scanning calorimetry thermograms for cinnamaldehyde and beta-cyclodextrincinnamaldehyde complex, 153– 154, 154

707

of commercial cornflakes and sugarfrosted cornflakes, at 7.5% and 5.2% water content, 585, 585 for cornflakes study, 584 enthalpy relaxation kinetics in corn syrupsucrose mixtures and, 420, 421 for thymol and beta-cyclodextrin-thymol complex, 153, 153 Differential thermal analysis, of solvent water in dough, 51–52 Diffusion, schematic illustration of mass transport by, 395 Diffusion model, rehydration studies and, 221–222 Diffusion rates, size of diffusing molecule and affect on, 94 Dilatometry, probing properties of carbohydrates in glassy state with, 355 Dilute solution, liquid state of water and, 205–206 Dilution of the glass, 42 Dimethyl sulfoxide addition of, to freezing solutions, 551, 552 dielectric constants of, 159t Dimethyl sulfoxide salt solution, composition of ethylene glycol salt solution and, with constant weight ratios of 4.54 and 18.17, 553t Dimethyl sulfoxide-salt-water, partial phase diagram of, with constant weight ratio values of 4.54 and 18.17, 558 Dimethyl sulfoxide solution, leakage from large unilamellar vesicles with various concentrations of, after cooling and heating of mixture, 560 Dipole-dipole force, solvent-water concept and, 50 Disaccharides as effective stabilizers, 600 intermolecular hydrogen bonds, NEB progress in the glassy matrix and, 576 mixtures of, in probiotic cultures, 285 retention of LGG viability during freezing, freeze drying, and storage with, 289 survival during drying and, 141

708

Index

Discrete models of proton relaxation, in interpretation of exponential decay curve, 257 Dispersability of maltodextrin powders, determining, 675 maltodextrin powders evaluation, at different feed concentrations and drying temperatures in production of powders, 677t quality of food powder properties obtained by spray drying and, 674 Dispersive model, dough sample and use of, 61 Disposable baby diapers, superabsorbent polymers used in, 386 Dissolution, use of term, relative to food powders, 42 Dissolution of soluble food powders, coupling between heat and mass transfer during process of, 46 Dissolution process, complexity of, 47 Distilled water, isoinactivation of Escherichia coli, vs. pressure and temperature in, 81 D-limonene encapsulated, 499 effect of water activity on and n values for release of from spray-dried powder at 50°C storage, 503 effect of water activity on release-rate of, at 50°C, 503 release time courses in spray-dried powders stored at various relative humidities, 501 encapsulated flavor powder preparation, 500 DMA. See Dynamic mechanical analysis DMSO. See Dimethyl sulfoxide DMTA. See Dynamic mechanical thermal analysis Double-edged notched tension specimens, 291 Dough aqueous phase of, 52 pasta, heated at 5°C, 246 solvent water in, 51–52 transforming to bread, steps in, 627

Dough/bred-crumb evolution, center temperature and, during baking, 631, 631 Dough foam structure, transformation of, into elastic crumb sponge, 627–628 Dough manufacture, 52 Dough structure, 50 Downton, G. E., 344 DPPC. See 1,2-dipalmitoyl-rac-glycero-3phosphocholine DPPC large unilamellar vesicles, phase volume of ice and unfrozen matrix at −40°C and leakage of, at −40°C, 557, 557t Dried foods new approaches with, and other advances in rehydration of, 232–234 rehydration of, as fundamental unit operation in food industry, 219 Dried solid matrix, empirical or semiempirical models and liquid transport into, 233 Drink mixes, sticking or caking, during storage of, 100 Drought-resistant crops, 252 DRS. See Dielectric relaxation spectroscopy Drugs, hypothetical states of matter of, 90 Drug solubility in lipid vehicles, water uptake, distribution, and effects on those composed of triglycerides and monoglycerides, 326–330, 332 Dry foods importance of prediction and control of NEB in, 571 stored, nonenzymatic browning and stability of, 445 Drying, effects of various perturbations on, 72 Drying conditions, influence of, on shrinkage and deformation of potato slabs, 615–616 Drying kinetics agave samples and, 621 of agave slices, at 60°C and 2 m/s for longitidinal and transverse cuts, 622

Index

Dry pasta systems, Maillard reaction in, 162 DSC. See Differential scanning calorimetry DTA. See Differential thermal analysis Dubinin-Radushkevitch relationship, 681, 682 Duck egg white description of, 453 film-forming ability of, and its watervapor barrier property, 453–458 conclusion, 457–458 introduction, 453 Haugh unit, viscosity, and pH of, stored at room temperature for 18 days, 455, 455t materials and methods, 454 barrier properties, 454 egg freshness, 454 film preparation, 454 physical properties, 454 statistical analysis, 454 viscosity, 454 results and discussion, 455–457 barrier properties, 457 color properties, 455 egg freshness, 455 film formation, 455 mechanical properties, 456–457 separated from shell eggs, viscosity and film-forming ability of, 456t Duck egg-white film made from shell eggs, barrier properties of, 457, 458t made from shell eggs, mechanical properties of, 457t Ductile materials, brittle materials vs., 106 Duncan’s new multiple range test, pineapple sample statistical analysis and use of, 659 DVS. See Dynamic vapor sorption Dye, W. B., 95, 96, 97, 346 Dynamic differential scanning calorimetry, tapioca-flour-based baked products study and determining glass transition temperature with, 593 Dynamic mechanical analysis distribution of water in bread measured with, 176 physical state of protein-carbohydrate sheets verified by, 579

709

Dynamic mechanical thermal analysis, 106, 124, 427, 429 Dynamic vapor sorption, sorption behavior followed by, 484 E Edible films, starch-based, 642 influence of glass transition on oxygen permeability of, 641–644 conclusion, 644 introduction, 641 materials, 642 methods glass transition temperature, 642 oxygen permeability, 642 preparation of starch films, 642 preparation of starch with various chain lengths, 642 results and discussion, 642–644 glass transition temperature, 643–644 oxygen permeability, 642–643 Edible films and coatings chitosan and, 459 for future food packaging, 453 Ediger, M. D., 18 EDTA, fresh and dialyzed jellyfish as well as alginate gel subjected to, 392 Effective diffusion coefficient, 221 for milled rice grains at different temperatures, 669t EG. See Ethylene glycol Egg-albumen films, water-vapor permeability values for, 453 Egg-white solutions, viscosities of, after addition of sorbitol, 455 Elbing, K., 55t Electrical conductivity, enzyme stability and measurements of, 17–18 Electronic nose (e-nose) data on sensitivity of, to pineapple aroma during freezing and thawing, 659–661 measurements of frozen-thawed pineapple samples, 659 off-aroma scores from sensory evaluation of freeze-thaw pineapple corresponding to data from, 661

710

Index

Electronic nose (e-nose) (continued) principal component analysis plot for fresh and freeze-thaw Smooth Cayenne and Queen pineapple cultivars determined by, 659, 660 Electronic nose (e-nose) technique, analyzing effect of freeze-thaw cycle on pineapple off-aroma with, 657–661 Electron paramagnetic resonance in combination with PALS, for insight into structure of carbohydratewater system, 359 investigating dynamic properties close to glass transition and, 364 Electron spin resonance probing properties of carbohydrates in glassy state with, 355 TEMPO dissolved in dough matrix and monitored with, 64–66 Electron spin resonance spectrometer, microdomain distribution in food matrices and measured values in experiment with, 60–61 Elizalde, B. E., 20 Ellagitannins, 563 encapsulated and nonencapsulated, storage stability of, 566–567 at low RH, increased storage stability of, via encapsulation by maltodextrin DE5-8 vs. maltodextrin DE18.5, 563, 567, 568 Elongation, softening of a polymer and, 402 Empirical models liquid transport into dried solid matrix and, 233 rehydration studies and, 220 Emulsifiers controlled kinetics of Maillard reaction and, 11 water interactions and, 157 Emulsion stabilization, proteins and, 245 Endocytic vesicles, deep plasma membrane invaginations and formation of, 74 Endothermic enthalpy relaxations, changes in volume associated with around glass transition, 341

Endothermic trend, origin of, observed prior to exotherm during rewarming, 377–379 Enthalpy relaxation, 419 as function of corn syrup-sucrose ratio, 426 as function of water activity, 424–425 physical changes caused by, 445 time and temperature dependence of, 420 Enthalpy relaxation kinetics, determination of, in corn syrup-sucrose mixtures, 420–421 Enthalpy relaxations, of amorphous materials and, 340 Enthalpy relaxation time, macroscopic molecular mobility of glass and, 446 Environmental moisture, tablet making and, 302 Environmental scanning electron microscopy, appearance of freezedried Sephadex beads detected through, 380 Enzyme stability, 16–18 EPNOE. See European Polysaccharide Network of Excellence EPR. See Electron paramagnetic resonance Equilibrium freezing-point curves, on state diagram, 91 Equilibrium relative humidity (ERH), 88 modeling approach to predict sorption behavior of nonelectrolytic mixtures and, 483 thermodynamic approach and, 208 Erythorbic acid NMR and EXR experiments with wheat flour dough and, 60–61 reaction kinetics and monitoring of, in dough matrices at different moisture contents and temperatures, 64 Escherichia coli baroresistance, effect of low temperature and hyperosmotic perturbation on, 79–80 Escherichia coli inactivation pressure and temperature in a binary medium at a water activity of 0.85 vs., 80

Index

pressure and temperature in binary medium at a water activity of 0.99 vs., 79 pressure and temperature in distilled water vs., 81 Escherichia coli K12TG1 cells, unusual pattern of survival of, after combined high-pressure and subzero-temperature treatments, 78 Escherichia coli survival, combined high hydrostatic pressure and low temperature on, 77–79 Essential oils antifungal activity of, 546 cyclodextrins and solubility, stability, and bioavailability of, 150 Estradiol, water content and solubility of, 330 ET. See Ellagitannins Ethanol, dielectric constants of, 159t Ethylene glycol, addition of, to freezing solutions, 551, 552 Ethylene glycol salt solution, composition of dimethyl sulfoxide salt solution and, with constant weight ratios of 4.54 and 18.17, 553t Ethylene glycol solution, leakage from large unilamellar vesicles with various concentrations of, after cooling and heating of mixture, 560 Ethylene vinyl alcohol, anomalous antiplasticization effects of water and relation to mechanical properties in, 122 Eudragit E100 components in, 402 effect of low concentrations of water on, 404 effect of low concentrations of water on glass transition temperature of, 405 permeability of water in, as function of relative humidity, 406 Eudragit E100-indomethacin mixture, effect of low concentrations of water on glass transition temperature of, 408 Eukaryotic cells, physical or physicochemical environment changes and stress on, 71

711

European Polysaccharide Network of Excellence, 296 Excipients effects of, on storage stability of freezedried xanthine oxidase, 599–603 mechanisms of stabilization of freezedried protein by, during dehydration, 600 in pharmaceutical formulations, 302 Exothermic enthalpy relaxations, changes in volume associated with, around glass transition, 341 Expanded solid foams, 247–249 Exponential decay curve, discrete and continuous models of proton relaxation in interpretation of, 257 Exponential model, 165 adsorption and desorption curves from oscillatory experiments on model bread crust described by, 165, 168–171 adsorption and desorption curves from oscillatory experiments with dry, crispy bread crust described by, 165, 167 bread crust experiments and summary of fitted diffusional coefficient and goodness of fits of diffusion model and, 171t curve fitting of rehydration data and, 221t oscillatory adsorption and desorption curves for bread crust best described by, 173 predicting moisture rate of rice grains with, 664t rehydration studies and, 220 rice varieties and, regression coefficient and percentage of root mean square error obtained from model fitting with, 667t External mechanical antiplasticization, by water, 116 Extruded flat breads, compression-forcewater-activity relationships for, 120, 122 Extruded snacks, proteins and, 237

712

Index

Extruded wheat flour, light micrograph of, showing starch as continuous structure with protein particles included, 249 Extrusion cost-effectiveness of, in food industry, 578 structuring proteins by, 246–249 expanded solid foams, 247–249 fibrous and dense composites, 246–247 intermediate to high water content, 246 F Fairy floss, 95 Fat content ehthalpy of dissolution measurements for, in model food powders, 43 enthalpy of dissolution for food powder samples tested as function of water activity and, 44 enthalpy of dissolution of all food powder samples tested as function of water contents and, 46 enthalpy of food powder dissolution and effects of, 41, 42, 43, 44, 45, 46 Fat crystallization, glass transitions overlapping with, at some levels of water plasticization, 337 FDA. See Food and Drug Administration FDC. See Fractal dimension of the slice contour Feret diameter, image analysis and qualtifying parameters related to, 613 Fermi model, relaxation times above glass transition modeled with, 339 Fibers, composition of, and use in spray drying, 192–193 Fibrous structures, formation of, from proteins, 246 Fickian diffusion model, 165, 167 defined, 166 isotherm and oscillatory simulation of in sphere, summary of fitted diffusional coefficient and goodness of fit of diffusion model and exponential model for, 171t

Fick’s laws of diffusion, 166, 232 first law, diffusion model and, 221–222 liquid transport into dried solid matrix and application of, 233 second law, 220, 533 diffusion model and, 221–222 permeability of water through polymer obtained with, 404–405 water absorption of cereals and legumes described by, 663 transport phenomena in foods and drawbacks with, 222 Fick’s second law solution for a sphere, water-diffusion process in rice grains during soaking determined with, 665 FID. See Free-induction decay Finney, J. L., 238 First-order kinetics model curve fitting of rehydration data and, 221t rehydration studies and, 220 Fish, testing freshness of, 599 Flake extrudate, aligned, 246, 247 Flavor encapsulation definition of and advantages with, 499 effect of water activity on release characteristics, 499–505 conclusion, 504–505 introduction, 499–500 materials and methods, 500–501 encapsulated flavor powder preparation, 500 materials, 500 release of encapsulated flavor from spray-dried powder, 501 results and discussion, 501–504 analysis of release rate and release mechanism by Avrami equation, 502–504 effect of relative humidity on release of D-limonene from the powder, 501 effect of relative humidity on release of L-menthol from the powder, 502

Index

Flavor in food cyclodextrins and solubility, stability, and bioavailability of, 150 main components of overall sensation of, 657 nonenzymatic browning and changes in, 445 Flink, J. M., 345 Flomoxel in gelatin gel, with kanamycin, hydrolysis rate of, 26 Flour, water interacting with polymers in, 51 Flour suspensions, that have undergone five freeze-thaw cycles, T2 distribution of protons in, 268 Flowability, water of hydration, API hydrate and, 26 Flow behavior, of high-amylose rice-starch samples, 636 Flow in porous media theory, 233 Flow properties of particles in dehydration, collapse phenomena in food systems and, 335 Fluidity, hydraulic conductivity of saturated porous medium and, 227 Fluorescence emission spectra, of 50% wt/vol sugar systems at 70°C for 168h/240h, glucose and trehalose systems respectively, 541, 542 Fontanet, S., 121t Food and Drug Administration, 176 Food biopolymer gels, ice crystallization during rewarming observed with, 381–382 Food colloids, composition of, 251 Food dehydration, main requirements for, 342–343 Food drying process, complexity of, 673 Food gels, 119 Food industry, decoupling of mobility of water from mobility of carbohydrate molecules in approach to glass transition and implications for, 366–369 Food matrix, stability of, factors related to, 59 Food packaging, future, edible films and coatings for, 453

713

Food polymer science, basic premise behind, 116 Food polymer-water blend properties of, 119–120, 122–126 gas transport, 126 mechanical, 119–120, 122–124 thermal, 124–126 Food powders dissolution kinetics of, critical importance of, 41 effect of water and fat contents on enthalpy of dissolution of, 41–47 conclusions, 46–47 experimental methods, 42–43 introduction, 41–42 results and discussion, 43, 45–46 enthalpy of dissolution for all samples, tested as function of their water and fat contents, 46 enthalpy of dissolution of all samples tested as a function of water activity, 44 exothermic responses from dissolution of, at all conditions tested, 43 less exothermic responses related to slower dissolution rate in, 46 model, characterization of, 42t thermodynamics and predicting dissolution of, 46 Food processing, quality and safety characteristics in, 673 Foods as complex examples of soft condensed matter, 87 hypothetical states of matter of, 90 Food safety, quality and stability, distribution of microstructural domains and assessment of, 66 Food sciences carbohydrate polymers and role in, 301 water as most common plasticizer in, 401 Foods in storage, physical state changes in, 93–94 Food solids, effect of thermal and water plasticization on mechanical, flow, and collapse phenomena on relaxation times and timedependent changes in, 340

714

Index

Food stability general rules with respect to measured aw and, 89–90 nanostructures and minimum integral entropy as related to, 681–687 relative vapor pressure and, 212 Food systems heterogeneity of, 61 proteins in, 237 Forssell, P. M., 643 Fourier transform infrared spectroscopy, 126 apparent crystallinity index of microcrystalline cellulose, obtained from, 310, 310 chitosan and blend films characterization with, 460–461 in combination with PALS, for insight into structure of carbohydratewater system, 359 post-deformation plastic zones used with, in monitoring fracture behavior of biopolymer films prepared from aqueous solutions, 292 water-replacement hypothesis partially corroborated by use of, 355 Fourier transform spectroscopy spectra, of chitosan film, 90 : 10 blend flim, 80 : 20 blend film, 70 : 30 blend film, and MPEG-Beta-PCL powder, 461 Fractal analysis macroscopic shrinking and deformation in potatoes studied with, 613 shrinkage and deformation of agave slices evaluated during convective drying with, 620, 625 Fractal dimension, image analysis and qualifying parameters related to, 613 Fractal dimension of the slice contour, agave slices during convective drying and, 621 Fractal geometry evaluating complexity of biological systems with, 515 studying interfaces and their roles as transfer-controlling barriers with, 613

Fracture response of food material, determining, 107 Fracture stress and fracture strain, changes in, for gluten and starch, as function of moisture content, 124 Free-induction decay, 15, 60, 251 for ceftazidime and cefazolin sodium hydrates, 29 hydration kinetics of cracker samples and, 414 for Na2HPO4 · 12H2O and Na2HPO4 · 2H2O, 28 for quinidine sulfate and scopolamine hydrobromide hydrates, 30 water mobility in model dough system and, 62 Free-induction decay analysis molecular mobility in cornflakes evaluated with, 584 relationship between relaxation time constant determined by, and temperature for commercial cornflakes and sugar-frosted cornflakes, 586, 586 Free-induction decay curve, nonmonoexponential, evaluating, 257 Free volume defined, 94 gas transport properties and, 126 plasticization-antiplasticization threshold of water in microcrystalline cellulose based on, 301, 309–312 water and increase in low molecular weight materials and polymers, 407 Free-volume reduction, 118 Free water, 115, 128 discrete-model assumption and describing bound water vs., 256 “Freezable” water, in soy bread and almondsoy bread, 181 Freeze-concentrated glass formation, 599 Freeze-concentrated systems, ice formation in, 348 Freeze concentration, starch fraction and effect of initial water content and freezing temperature on rate of, 188

Index

Freeze-dried amorphous food powder samples large exothermic responses observed for, 43, 45 typical dissolution calorimetry curves of, containing different amounts of fat, 44 Freeze-dried enzymes, analyzing stability of, over range of temperature and water-content conditions, 16 Freeze-dried food powder samples, hydration of hydrophilic groups in, and highly exothermic values for, 45 Freeze-dried formulations, mannitol in, 20 Freeze-dried products, appearance of, factors related to, 142 Freeze drying bacterial preservation and, for food and pharmaceutical applications, 285 collapse in, 343 collapse phenomena in food systems and, 335 entrapment of probiotic bacteria in frozen cryoprotectants and viability in, 285–289 stresses generated by, 599 successful, requirements for, 343–344 uses for, 141 water activity of sucrose-cell formulations after, 144, 145 Freeze drying of Lactobacillus coryniformis Si3 focus on water, 141–147 introduction, 141–142 materials and methods, 142–144 cell survival, 143 fermentation and sample preparation, 142 freeze-drying protocol, 143 statistical analysis, 143–144 thermal analysis, 142–143 water activity, 143 results and discussion, 144, 146 important factors for freeze-drying survival, 144 water and solute properties, 144, 146

715

Freeze-drying survival effects plot with error bars showing 95 confidence level relative to, 144, 146 important factors for, 144 Freezing effect of initial water content and freezing temperature on starch-fraction moisture content after, 187–188 effects of various perturbations on, 72 understanding effects of, in starch-based food, 185–186 Freezing behavior of polymer gels, investigation of, 373–383 of water-saturated starch granules, 263–264 Freezing processes, dehydration processes vs., 338 Freezing rate freeze-drying survival and, 144 water activity of sucrose-cell formulations after freeze-drying and, 144, 145 Freezing scheme for compartmentalized water, in polymer gels, 379–380 Freezing temperatures, various, evolution of starch-fraction moisture content in function of freezing time at, 189 Freshness testing paper, 599 Frozen desserts, ice recrystallization and, 348 Frozen dough affinity constants for glucose uptake by Saccharomyces cerevisiae, 55t concentration of glucose and fructose in, made with 2 yeast and 2 added sucrose fermenting at 30°C, 54 factors related to shelf life of, 49 materials and methods, 52 manufacture of, 52 sugar extraction and analysis, 52 proofing power and, 49 results and discussion about, 52–53, 55 “solvent-water” concept and, 50–51 solvent water in, 51–52 structure of, 50 summary of mathematical model parameters, 55t

716

Index

Frozen-dough segment of bakery industry, yeast-leavened part of, 49 Frozen foods crystallization of sugars in, 347 differing stabilities of, 212 phase separation of ice crystals in starchbased systems during freezing and practical implications for, 185–190 Frozen systems, water measurement and, 204 Fructose as antiplasticizer, 130 approximate solubility of, in glycerol at 60°C after equilibration period of 10 days, 160t carrot fiber as carrier in spray drying of fructose, 191–196 conclusions, 196 introduction, 191–193 materials and methods, 193–194 results and discussion, 194–196 concentration of, in dough made with 2 yeast and 2 sucrose fermenting at 30°C, 54 first-order phase transition (melting) of, by use of common sugar crystals, 433t freeze-dried, amorphous, derivative of capacitance vs. temperature for, 194 freeze-dried amorphous, Tg from DSC and corresponding DEA frequencies for carrot fiber in, 195t spray drying of, difficulties with, 191 Fructose-carrot fiber, spray-dried powders of, compared to those of fructosemaltodextrin, 191, 193, 195, 196 Fructose content, glucose content and, in doughs, 53 Fructose-maltodextrin, spray-dried powders of fructose-carrot fiber compared to those of, 191, 193, 195, 196 Fructose mixtures, spray drying of, 195–196 Fruit flavors, sensory evaluation and assessment of, 657 Fruit juice, spray drying of, difficulties with, 191

Fruit quality, aroma as major determinant of, 657 Fruit storage, chitosan and, 459 FTIR. See Fourier transform infrared spectroscopy Fugiwara, T., 153 Fukuoka, M., 537 G GAB equation. See GuggenheimAnderson-de Boer equation GAB value. See Guggenheim-Anderson-de Boer value Galactose, intermolecular hydrogen bonds, NEB progress in the glassy matrix and, 576 Galatin samples, NMR study of water motion and relaxation in Gallant, D. J., 253 starch granule structure redrawn from, 254 Garlic, volume decrease or shrinkage studied in, 613 Gas chromatography, fruit flavor assessment and, 657 Gas transport behavior, biopolymer-based packaging films and coatings and, 126 Gas transport properties, of food polymerwater blend, 126 Gaudin, S., 132 Gaussian decay, 29, 30 antibiotic hydrates and, 35, 36 Gaussian peaks, three best-fit, soy bread and, obtained from deconvolution, 180 Gay-Lussac-type pycnometer, density of sodium- or potassium-glucosewater ternary system measured with, 473, 474 Gelatin effect of water content on physicochemical and structural properties of, 292 ice crystallization during rewarming observed with, 381–382 protein-carbohydrated sheets prepared with, 578 study of mechanical response to tensile loading of, 291

Index

Gelatin films infrared maps for, in the hydrated state, 295, 295 unstable crack propagation in, brittleductile transition and, 295–296 Gelatin gel, transmission electron microscopic image of freezeetched replica of, 393 Gelatin-gum arabic-sucrose sheet best protection against oxygen with, 582 highest values of Young’s modulus and tensile strength observed for, 580 Gelatinization acid hydrolysis, destruction of amorphous chains and, 265 SEM and TEM micrographs of cassava and potato starches undergoing, 255 starch-chain mobility during, 260–263 starch-chain mobility T2 distribution as measured by proton 90° pulse freeinduction decay of deuterated cassava starch saturated with heavy water during, 262 terminal extent of, 535 of water-saturated starch granules, 260 Gelatinized potato-starch gels, differential scanning calorimetry rewarming traces obtained with, 382 Gelatinized starch five freeze-thaw cycles, T2 distribution of protons in, 268 in glassy state, brittle-to-ductile transition induced by water reported for, 581 Gelatinized starch paste, polydispersion of, in nature, 507 Gelatin-whey protein mixture heating, 244 light-microscopic images of, at pH 5.4, induced by temperature and highpressure processing, 245 Gelation, 239–249 emulsion stabilization, 245 mixed systems, 243–245 single-protein systems, 239–242 cold gelation, 242 heat gelation, 239–242

717

structuring by extrusion, 246–249 expanded solid foams, 247–249 fibrous and dense composites, 246–247 intermediate to high water content, 246 water holding, 242–243 Gelling state, in many biological systems, 373 Gels beta-lactoglobulin, 10, light-microscopic images of, at pH 5.3, 241 particulate at pH 5.5 and transparent finestranded at pH 7.5, 240 Gennadios, A., 453, 457 Gervais, P., 76 Gibbs-Duhem equation, ice crystals estimated with, 377 Gidley, M. J., 124 Glalactose, embedding of, in glassy trehalose matrix, 572 Glass enthalpy relaxation time and macroscopic molecular mobility of, 446 solid state of water and, 204 Glass formation, food activity and role of, 130, 132 Glass-rubber transition diagram, 93, 93 Glass-rubber transitions, measurements for pasta, candy, native cornstarch, and individual rice kernels, 435 Glass transition of amorphous food solids in dehydration at various water contents, 343 brittle-ductile transition temperature comparable to, 106–107 carbohydrate-water systems and dynamic properties close to, 364, 366 changes in volume associated with endothermic and exothermic enthalpy relaxations around, 341 of cross-linked dextrans containing small amount of water, 381 enthalpy relaxations around, factors related to, 340 flow in glassy and liquid states and, 336–339 of food solids, collapse phenomena and, 335 increase in water content, loss of crispness and correlation with, 105–106

718

Index

Glass transition (continued) influence of, on oxygen permeability of starch-based edible films, 641–644 mechanical and dielectric relaxations in amorphous materials below and around, 341, 342 NEB rate and, for samples of various glassy matrices, 573 for potato discs, 438 relationship among moisture content, texture properties, water activity and, for glassy tapioca-flour-based baked product, 595, 595–596 stickiness in amorphous food materials around, 345 Glass transition curve defining dividing line for optimal storage conditions for hard candy with, 104–105 on state diagram, 91 Glass transition line need for, in state diagram, 111 for sucrose, 101 Glass transition point, defined, 93 Glass transition stabilization, 599, 600 Glass transition temperature, 98, 100, 192, 256 of anhydrous lactitol and water, 274 of corn syrup-sucrose mixtures, 421t effect of carbohydrate molecular weight, on matrix density and on, 367 effect of starch chain length on, 641 of Eudragit E100, effect of low concentrations of water on, 405 as function of mass fraction of water for apple, cabbage, and potato in which Maillard reaction was developed, 13 as function of mass fraction of water for polymeric, lactose, and trehalose matrices in which Maillard reaction develops, 12 as function of mass fraction of water for polymeric and trehalose matrices in which beta-carotene was encapsulated, 19 as function of % relative humidity for hard-ball candy, 102

of indomethacin, effect of low concentrations of water on, 404 indomethacin-Eudragit E100 mixture and effect of low concentrations of water on, 408 of lactitol, using to predict crystallization occurring at different storage temperatures and relative humidities, 278, 280–281 Maillard reaction in absence of water but below, 162 microdomain distribution in food matrices: model dough system and, 61 microdomain distribution in food matrices and, 59, 66 in model dough systems, 59 molecular mobility, sub-T9 relaxation and, 161 plasticization effect of water and decrease in, 491 of plasticized starch films, upper transition and lower transition, 643 relaxation times above, and effect of water plasticization on glass transition and relaxation times, 339, 339 of starch films having different degrees of polymerization, 644 state changes and temperature of storage in relationship to, 100 for sucrose, 142 values for moisture content and, of xanthine oxidase with various excipients, 601t Glass transition temperature curve for sugar, application of, on top of phase diagram of sugar-water mixture, 87 Glass transition temperature of glassy polymer water and, 401–409 conclusions, 408–409 materials and methods, 402–403 results and discussion, 404–408 Glass transition theory, 93 Glassy carbohydrates dual role of, in stabilization and delivery of active ingredients, 354t in food and pharmaceutical stability, 353–356

Index

local structure of, at molecular level, 355 molecular packing in, 359, 361–362 Glassy foods materials and methods, 446 glass transition and enthalpy relaxation, 446 nonenzymatic browning reaction, 446 sample preparation, 446 nonenzymatic browning reaction and enthalpy relaxation of, 445–450 discussion, 449–450 introduction, 445–446 nonenzymatic browning reaction of, 571–576 results, 447–448 enthalpy relaxation time of glassy trehalose-glucose-lysine system, 447–448 glass transition and enthalpy relaxation, 447 progress of NEB in glassy trehalose matrix, 448 Glassy matrices, glass transition and NEB rate of samples of, 573 Glassy polymer below a critical concentration, water decreases molecular mobility of, 409 LMM compound or diluent added to, 117 Glassy stable amorphous solid state, food storage and transition from, to rubbery amorphous solid state, 93 Glassy state flow in, glass transition, 336–339 free volume of carbohydrate system matrix independent of degree of polymerization in, 361 types of relaxation occurring within, 116–117 viscosity in, 338 water diffusivity, rubbery state vs., 94 Glassy system, estimation of free volume available within, 94 Glassy tapioca-flour-based baked product, sorption isotherm: experimental data of, 594, 595 Glassy trehalose matrix, progress of nonenzymatic browning in, 448

719

Globulins cold gelation and, 242 heat gelation and, 239 Glove bag, 273 Glucose, 130 addition of, to freezing solutions, 551 approximate solubility of, in glycerol at 60°C after equilibration period of 10 days, 160t concentration of, in dough made with 2 yeast and 2 sucrose fermenting at 30°C, 54 embedding of, in glassy trehalose matrix, 572 first-order phase transition (melting) of, by use of common sugar crystals, 433t fructose uptake and, in doughs, 53 glassy food model preparation and, 572 Glucose-lysine-trehalose solution glass transition and enthalpy relaxation properties investigated, 446 nonenzymatic browning reaction for, 446 Glucose-salt-water solutions partial phase diagram of, with constant weight ratio value of 9.09, 556 partial phase diagram of, with R value of 36.34, 556 Glucose solutions concentrated aqueous, ice formation in, 465–470 intensity ratio of isothermal transformation measurement at varying temperature, 469, 469 introduction, 465–466 materials and methods, 466–467 results and discussion, 467, 469–470 simultaneous X-ray diffractory and differential scanning calorimetry of 53% glucose solutions, 467, 468 thermogram and image from cryomicroscope at temperatures indicated, 467, 468 leakage from large unilamellar vesicles at various concentrations of, that were previously cooled to −40°C and thawed to room temperature, 559

720

Index

Glucose uptake in dough, hyperbolic equation, 52–53 Glucose-water mixtures, temperature dependence of ratio of rotational time of glucose and rotational time for water for various glucose concentrations in, 365 Gluten changes in fracture stress and fracture strain for, as function of moisture content, 124 inspection of DSC behavior of, 247 Gluten proteins in dough, 50 scanning calorimetry of, as function of moisture content, 248 Glycerol, 130 approximate solubility of some sugars in, at 60°C after equilibration period of 10 days, 160t dielectric constants of, 159t Maillard reaction promoted by, 162 melting-enthalpy-moisture-contentrelationships of konjac glucomannan films plasticized with sorbitol and, 125 mixtures of, in probiotic cultures, 285 processing of, for extruded rich starch study, 484–485 sorption behavior of extruded rice starch in presence of, 483–489 strongly polar water and solubility of, 160 tensile modulus of tapioca-starch films at different water activity as function of, 131, 131 water and interactive plasticizingantiplasticizing effects of, 130–132 Glycerol osmotic stresses, strain-dependent response to, 76 Glycerol-water mixture, ESR use and increase in rotational mobility of spin probes within, 94 Gondek, E., 121t Gontard, N., 122 Gordon-Taylor equation glass transition of cross-linked dextrans analyzed by means of, 381 modeling glass transition temperature of anhydrous lactitol, water and, 274

Gordon-Taylor fit, of effect of moisture content on glass transition temperatures of lactitol, 278, 280 Gordon-Taylor relationship, 337 Granulation processes, surface plasticization and dehydration as basis of, 343 Granule-associated proteins, molecular alterations and, applications of, 513 Gravimetric determination, of actual water content, 203 Gravimetric nuclear magnetic resonance, testing usefulness of, with range of food materials equilibrated at different water activities, 411 Gravimetric stepwise oscillatory experiments, on model bread crusts, 166–167 Green coffee, rehydrated, normalized fracture force of, as a function of water activity, 494, 495 Greenspan, L., 151 Green superabsorbents, developmental research on, 386 Guerrero, S. J., 126 Guggenheim-Anderson-de Boer equation application of, to fit isotherm of glassy tapioca-flour-based baked samples, 595 bread crust isotherms described by, 168 coffee study and water-sorption isotherms fitted by, 493 corn flakes study, water-sorption isotherms obtained and fitted to, 584 quinoa seeds and sorption isotherms fitted to, 649–650 sorption behavior of food materials represented by, 483–484 water-sorption isotherms for potato powder at 23°C, fitted by, 439, 440 Guggenheim-Anderson-de Boer (GAB) form, protein hydration studies and, 238 Guggenheim-Anderson-de Boer model, 683 Guggenheim-Anderson-de Boer parameters fitting of sorption isotherms of raw and roasted coffee with, 494t for rice-glycerol extrudates, 487, 487t

Index

Guggenheim-Anderson-de Boer value, 14 Gum arabic as efficient encapsulation matrix for limonene in freeze drying, 577 protein-carbohydrated sheets prepared with, 578 Gupta, A. S., 196 Guyot, S., 76 H Hahn spin-echo data, analysis of, in potato powder samples, as function of water content, 441, 441 Hahn spin-echo sequence, evaluating water populations in cornflake study with, 584–585 Hahn spin-echo transverse relaxation time values, water content vs., for commercial cornflakes and sugarfrosted cornflakes, 586–587, 587 Halek, G. W., 121t Handa, A., 453 Hard-ball candy, 88, 111 car interior temperature and effect on, 101–102, 102, 104 glass transition temperature as function of % relative humidity for, 102 making, 101 onset temperature for crystallization of, as function of relative humidity, 103 states of, stored at different abuse conditions as compared to storage condition at or above Tg, 103 stickiness of, 100–105 visualization of states of, after humidification at 33% relative humidity and then temperature abused, 104 Hardening, food storage and, 93 Hardness of bread-crumbs, measuring, 607 of innovative bread and standard bread crumb during storage, 608t loss of, storage of food and, 105–108, 111 rating, for tapioca-flour-based baked products, 593 Harris, M., 121t Hatley, R. H. M., 51

721

Haugh unit of duck egg-white stored at room temperature for 18 days, 455, 455t egg freshness and measurement of, 454 Heat, dissolution of soluble food powders and coupling of mass transfer and, 46 Heat gelation, 239–242 Heat-moisture-treated rice starch change in consistency index and flow behavior index of, at various moisture levels of treatment, 637 relationship of shear stress and shear rate of native rice starch and, at 18%, 21%, 4%, and 27% moisture content over shear rate range of 0–300 s−1 and at 60°C, 638 shear-thinning behavior in, following power-law model, 639 Heat-moisture treatment, of high-amylose rice-starch samples, 635, 636 Heat transfer, local, role in dissolution process, 47 Heldman, D. R., 121t Hemicellulose, in fibers, 192 Henderson-Pabis model moisture rate of rice grains predicted with, 664t rice varieties and, regression coefficient and percentage of root mean square error obtained from model fitted with, 667t Hen egg white, protein films developed from, 453 Hermansson, A-M., 239 Heterogeneity of food systems, 61 Heterogeneous system, complications in relaxation of, 257–258 Hexagonal ice, defined, 465 Hexose, intermolecular hydrogen bonds, NEB progress in the glassy matrix and, 576 High-density materials, antiplasticization and, 120 Highly polar solvents, dielectric constant of, 159 High-mobility water, ratio of, to lowmobility water in water-excipient mixtures, 37t

722

Index

High resolution proton nuclear magnetic resonance, 26 High temperature-short time extrusion processes, flowing plasticized mass of food solids produced by thermal and water plasticization in, 344 Hills, B. P., 238, 239, 258 Hiltner, A., 123 Hirst, A. G., 387 HMT. See Heat-moisture treatment Hofmeister series, 18 “Hole filling,” by diluent, 118 Hole size carbohydrate, suggested mechanisms leading to, increasing with water content in glassy state, 358 of thermally annealed maltodextrin DE12 matrices prepared by solvent casting, as function of temperature for various water contents, 356, 357 Hole volume of annealed maltodextrin matrices with given equilibration as function of temperature for various molecular weight distributions, 360 as function of weight fraction of water for a range of maltopolymer-maltose matrices equilibrated at water activities between 0 and 0.75, 363 Honey powder, onset of glass transition temperature of food powders measured by TMCT, DSC, and TMA, 434t Hoover, R., 636 Hot inflection point, for bread dough, 630 Hot-pressed-pullulan-starch blends study, antiplasticizing effects of water in, 123–124 Howes, T., 195 Hsieh, F., 105 HTST processes. See High temperature-short time extrusion processes Humidity hard-ball candy exposed to, 101–104 high, stored food transitioning into rubbery zone as function of, 95 Humidity caking, reasons for, 345

Hydrated beta-cyclodextrins, structure of, 150 Hydrated food material, description of total 1 H NMR signal amplitude obtained from, 412 Hydrated oils, lower solubility for estradiol and testosterone in desiccated oils vs. in, 330, 332 Hydrate formation, effects of water content on drug solubility in lipid vehicles and, 330 Hydrates, clathrate, 204 Hydration cracker results showing limited amount of material in carbohydrate matrix transferred to mobile phase during, 416 main components of vegetable powders used for hydration experiments vs. fraction of matrix component mobilized during, 417t in starch gelatinization, 255 Hydration shells, formation of, 82 Hydration water ease of evaporation for, as determined by DSC and water-sorption isotherm measurements, 31 molecular mobility of, as determined by NMR, 26–30 Hydraulic conductivity, of saturated porous medium, value of, 227 Hydric stresses, protein stability and, 16 Hydrodynamic flow mass transport in hierarchical structure and, 394 schematic illustration of mass transport by, 395 Hydrogen bond model of local reaction promoting nonenzymatic browning reaction progress in glassy matrix and, 574, 575 solvent-water concept and, 50 Hydrogen-bonding interactions, trehalose, enzyme stability and, 18 Hydrophilic biopolymers, 50 Hydrophilic protein-carbohydrate matrices, plasticizing of, by water, 577

Index

Hydrophobic hydration, solvent-water concept and, 50 Hydrostatic pressure, high, Escherichia coli survival relative to combination low temperature and, 77–79 Hydrothermal treatment of starch, common use of, 635 Hydroxyethyl starch (HES)-glucose mixtures, effects of lowering average molecular weight of, 368– 369, 369 Hydroxypropyl cellulose (HPC) film effect of water content on physicochemical and structural properties of, 292 infrared maps for, in the hydrated state, 295, 295 notched, study of mechanical response to tensile loading of, 291 in situ visualization of stable crack propagation for, at 69% relative humidity by using high-speed camera, 294, 294 water content of, at 10% RH and 60% RH, 37t Hydroxypropylmethyl cellulose, water content of, at 10% RH and 60% RH, 37t Hyperosmotic perturbation effect of low temperature and, on Escherichia coli baroresistance, 79–80 sequence of, on yeast cells, 72, 74 Hyperosmotic shock, cell shrinkage in response to, 76–77 Hysteresis, in adsorption-desorption, 238, 239 I IB. See Innovative bread Ice comparison of orientation about oxygen-oxygen hydrogen bond for, 206 defined, 204 Ice cream, crystallization of amorphous lactose in, 347–348 Ice crystal lattice, schematic of, 205

723

Ice crystallization of compartmentalized water in polymer gels, scheme for, 380t in gels and foods manipulated by polymer network, 373–383 during rewarming observed with various food biopolymer gels, 381–382 Ice crystal reduction and stability of frozen large unilamellar vesicles, 551–561 composition of dimethyl sulfoxideethylene glycol salt solutions with constant weight ratios of 4.54 and 18.17, 553t composition of sugar-salt solutions with constant weight ratios of 9.09 and 36.34, 553t conclusions, 561 effect of ice reduction on stability of frozen LUVs, 557–560 nonpermeable CPAs, 557–560 permeable CPAs, 560–561 effect of nonpermeable and permeable CPAs on temperature depression, 555–557 nonpermeable CPAs, 555–557 permeable CPAs, 557 introduction, 551–552 materials and methods, 552–555 phase volume of ice and unfrozen matrix, 555 preparation samples, 552–554 DMSO-salt and EG-salt solutions, 553 LUV dispersions, 553–554 sugar-salt solutions, 552–553 results and discussion, 555–561 stability of LUV dispersions, 555 thermal analysis, 554–555 DMSO/EG-salt-water, 554–555 sugar-salt-water, 554 Ice crystals materials and methods measurement of freezable and bound water by differential scanning calorimetry, 186–187 preparation of pregelatinized starch, 186 phase separation of, in starch-based systems during freezing, 185–190 conclusion, 189–190

724

Index

Ice crystals (continued) introduction, 185–186 materials and methods, 186–187 results and discussion, 187–188 effect of initial water content and freezing temperature on rate of freeze concentration, 188 effect of initial water content and freezing temperature on starchfraction moisture content after freezing, 187–188 Ice formation, recrystallization and, 348 Ice formation in concentrated aqueous glucose solutions, 465–470 introduction, 465–466 materials and methods, 466–467 results and discussion, 467, 469–470 Ice lattice, clathrate lattice vs., 204 Ice melting, endothermic trend due to, verifying assumption, 378 Ice recrystallization, frozen dessert quality and, 348 Idealized modulus vs. diluent content curve, of solid polymer-diluent blends, 118 IDF. See International Dairy Federation IGC. See Inverse gas chromatography IIFs. See Independent interaction factors Image analysis bread dough study, 629–630, 632 in evaluation of shrinkage and deformation of potato slabs during drying, 615 macroscopic shrinking and deformation in potatoes studied with, 613 observing structures and obtaining data in nanoscopic and microscopic fields, 674 ImageJ 1.34 software, 621 ImageJ 1.37 software, 516 Image processing evaluation of disintegration and diffusion of pharmaceutical solid matrices by, 515–520 shrinkage and deformation of agave slices evaluated during convective drying with, 620, 625 IMFs. See Intermediate moisture foods

Independent interaction factors, for rice starch-glycerol mixtures, 483, 488, 489 Indomethacin effect of low concentrations of water on glass transition temperature of, 404 effect of water on glass transition temperature of, 401, 402 Indomethacin-Eudragit E100 mixture effect of low concentrations of water on glass transition temperature of, 408 water and antiplasticization relative to, 407 Infant formula, sticking or caking in, during storage of, 100 Infrared spectroscopy, vibrational motions measured with, 216 Inhomogeneous reaction kinetics in dough matrix, investigating, 64–66 Innovative bread, 605, 607 hardness of, during storage, 608t proton free-induction decay of, fresh and after 7 days of storage, 609, 610 Innovative mixing analysis of parameters characterizing water properties in fresh and stored breads, 608–609 softer bread product at 7 days of storage as result of, 611 In situ visualization comparing water-dependent fracture behavior of different biopolymer films and visualizing crack propagation with, 291, 292 of crack propagation, 293–294 of stable crack propagation for hydroxypropyl cellulose film and cassava starch film at 69% relative humidity by using high-speed camera, 294, 294 Instant dairy powders, manufacture of, surface plasticization and dehydration as basis of, 343 Instron universal testing machine, sample taken from mesoglea, for puncture test where stress and strain were recorded in, 388

Index

Insulin degradation, differential coupling of overall rate constant for, and fraction dimerization to Tg in amorphous lactose formulations at 25°C and varying percent water, 327 Intermediate moisture foods, 90 International Dairy Federation, 675 International Symposium on the Properties of Water central theme, since first meeting of, 203 continuing advancement in understanding water’s role through, 217 first meeting of, 89 protein hydration lectures at, 237 RVP measures and correlations as theme at, 212 Intrinsic permeability of a medium, as function of pore structure and geometry, 227 Inulin, in fibers, 192 Inverse gas chromatography determination of glass transition temperature by, 403 glass transition temperature of polymer as function of environmental RH as determined by, 402 Ion-dipole force, solvent-water concept and, 50 “Islands of mobility” densification of glass and, 118 in glassy structure, 419 Isoflavone profile, of soy bread, altering, 175 ISOPOW. See International Symposium on the Properties of Water Isoteric heat of sorption, moisture-content plots of food colloids exhibiting a maximum Qs value vs., 128 Isothermal solution calorimetry, 41 Isotherm sorption experiments, on model bread crusts, 166 Isotropic shrinking, 616 Isoviability diagram, of Saccaromyces cerevisiae vs. temperature and water activity, 75 Israkarn, K., 508

725

J Jasmine rice, 663, 664 comparison of experimental and predicted solid loss of milled rice grains from waxy milled rice, Sao Hai and, as function of soaking time at various temperatures, 670 effective diffusion coefficient for, at different temperatures, 669t experimental and predicted moisture contents of, as function of soaking time at various temperatures, 666 Page model parameters for water uptake by, at different temperatures, 668t power-function model parameters for loss of solids from, at different temperatures, 671t regression coefficient and percentage of root mean square error obtained from model fitting using various models, 667t temperature and water uptake of, 667 Jellyfish (Aurelia aurita) composition of polysaccharides in, 392 dialyzed and freeze-dried, high-resolution magic angle spinning nuclear magnetic resonance images of, 386 grainy surface of, and transport canals, 390 life cycle and accessibility of, 387, 387 microstructure of, 385, 389–393 structural analysis of, 386–387 water-holding and texture properties of, 385, 386, 387–389 water retention by superabsorbent polymers and, as function of salt concentration, 388 Jellyfish network, transmission electron microscopic image of freezeetched replica of, 393 Johari, G. P., 118 Jouppila, K., 100, 479 JSM-6460LV scanning electron microscope, 460

726

Index

K Kaffir lime peel oil, minimum inhibitory concentrations of, and potassium sorbate against Rhizopus stolonifer and Aspergillus flavus, 547t Kaffir lime peels, preliminary examination of, for antifungal activities by an agar dilution method, 545 Kaminski, W., 484 Kamman, J. F., 162 Kapsalis, J. G., 121t Karel, M., 87, 88, 91, 98, 286, 345, 346, 347, 348 Karl Fischer coulometer, residual moisture content of xanthine oxidase samples measured with, 601 Karl Fischer titration, 207 Karl Fisher method, 29 Katz, E. E., 105, 121t Kelvin effect, 377 Kelvin equation, 230 water activity converted to tension head by, 229 KGM films. See Konjac glucomannan films Kinetic properties, temperature and, 208 Kinetics of potato slabs, during drying at 55°C, air velocity = 1/7 m/s, and thickness = 2.5 mm, 615, 615 Kinsho, T., 81 Kohlrausch-Williams-Watts approach, enthalpy relaxation time of trehalose-glucose-lysine system and, 448 Kohlrausch-Williams-Watts equation, enthalpy relaxation data and, 422, 423, 425 Kohlrausch-Williams-Watts parameters, for corn syrup-sucrose mixtures, 423t Koichi, Y., 275 Kollidon K12, molecular weight for, 318 Konjac glucomannan anomalous antiplasticization effects of water in relation to some mechanical properties demonstrated in, 122 changes in tensile strength and strain at break of, as function of water activity, 123

Konjac glucomannan-carboxy methylcellulose films, changes in tensile strength and strain at break of, as function of water activity, 123 Konjac glucomannan films, plasticized with glycerol and sorbitol, meltingenthalpy-moisture-content relationships of, 125 Kostaropoulos, A. E., 129 Krishnan, P., 654 Krokida, M. K., 616 Krusch, L., 98 Arrhenius plot for crystallization of lactose, at three different moisture contents from data of, 99 KWW equation. See Kohlrausch-WilliamsWatts equation L Labuza, Katherine, 101 Labuza, T. J., 107 Labuza, T. P., 100, 105, 106, 121t, 162 Lactic acid bacteria, freeze drying and stabilization of, 141 Lactitol amorphous, preparation of, 272 BET water-sorption isotherms of, 478 crystallizing amorphous, modeling moisture content of, 273–274 freeze-dried comparison of water sorption and crystallization behaviors of, 477–482 experimental glass transition temperature and water contents after storage of 98h for, 479, 479–480 Gordon-Taylor fit of effect of moisture content on glass transition temperatures of, 278, 280, 281 increased use of, in foods, 271 materials and methods in comparative study, 478 melting temperatures of, stored at 20°C at 49%, 58%, and 81% RH and 32°C at 41%, 64%, and 81% RH, 276

Index

moisture content and Tg as useful tools in predicting stability of, 281 moisture sorption of, stored at various temperatures and relative humidities, 279 sample preparation for differential scanning calorimetry, 274 water content of, during storage at 25°C at various relative vapor pressures, 480, 480–481 X-ray powder diffraction patterns of, at end of storage at 25°C at various relative vapor pressure of 66%, 481, 481 Lactitol crystallites melting temperatures of, 275–276 leveling off, 277 Lactitol crystallization melting enthalpy measured as function of time for, 277 plateau melting enthalpies for, at 20° and 32°C storage temperatures and relative humidities, measured after initial period of, 278t rate and extent of, 276–278 thermograms of, at 20° and 49% RH, 275 Lactitol samples, sample water peak melting enthalpy and moisture content of, grouped according to their storage conditions and relative humidity, 280 Lactobacilli strains, dehydration tolerance differences among, 141 Lactobacillus coryniformis Si3 freeze drying of: focus on water, 141–147 cell survival, 143 fermentation and sample preparation, 142 freeze-drying protocol, 143 important factors for freeze-drying survival, 144 statistical analysis, 143–144 thermal analysis, 142–143 water activity, 143 water and solute properties, 144, 146 materials and methods, 142–144

727

Lactobacillus plantarum, synergism between high pressure and subzero temperature on, 78 Lactobacillus rhamnosus GG effects of freezing, freeze drying and storage relative water-vapor pressure on viability of, 285–289 glass transition temperature at various relative humidities of freeze-dried systems in presence and absence of, 287t scanning electron microscopy of, encapsulated in trehalose and lactose-trehalose glass, with data for colony forming units, 289 Lactose amorphous state diagram of, 349 water-sorption and time-dependent crystallization and recrystallization of, 347 approximate solubility of, in glycerol at 60°C after equilibration period of 10 days, 160t Arrhenius plot for crystallization of, at three different moisture contents, 99 BET water-sorption isotherms of, 478 embedding of, in glassy trehalose matrix, 572 endothermic response and recrystallization of, in food samples, 47 first-order phase transition (melting) of, by use of common sugar crystals, 433t freeze-dried comparison of water sorption and crystallization behaviors of, 477–482 experimental glass transition temperature and water contents after storage of 94h for, 479, 479–480 glass transition of maximally freezeconcentrated solutes and onset temperature of ice melting of, 287t

728

Index

Lactose (continued) intermolecular hydrogen bonds, NEB progress in the glassy matrix and, 576 materials and methods in comparative study, 478 water content of, during storage at 25°C at various relative vapor pressures, 480, 480–481 X-ray powder diffraction patterns of, at end of storage at 25°C at various relative vapor pressure of 54%, 481, 481 Lactose crystallization modified Williams-Landel-Ferry plot for, 99 proteins and, 17 retarding of, proteins or biopolymers and, 11, 14 Roos and Karel’s work on, 98 Lactose glass transition, in dehydrated dairy foods, 336 Lactose glass transition temperature, as function of mass fraction of water, 14 Lactose-trehalose glass, scanning electron microscopy images of Lactobacillus rhamnosus GG encapsulated in trehalose glass and, with data for colony forming units, 289 Lactose-trehalose solutions, glass transition of maximally freeze-concentrated solutes and onset temperature of ice melting of, 287t Laminar flow, Poiseuille equation for, 223 Land, L. M., 328, 330 Lang, K. W., 683 Laplace equation capillary flow in porous media and, 223 height of capillary rise expressed as, 225 Laplace-Young equation, ice crystals estimated with, 377 Large unilamellar vesicles calculation of percentage of CF from, 555 effect of ice crystal reduction on stability of, 551–561 conclusions, 561

introduction, 551–552 materials and methods, 552–555 results and discussion, 555–561 frozen, effect of ice reduction on stability of, 557–561 nonpermeable CPAs, 557–560 permeable CPAs, 560–561 leakage from added with various concentrations of dimethyl sulfoxide or ethylene glycol solutions, after cooling of mixture, 560 at various concentrations of sugar solutions previously cooled to −40°C and then thawed to room temperature, 559 scanning electron micrographs of those cooled to −40°C at 10°C/min, 559 Laroche, C., 76 Lawal, M. O., 675 Lazar, M. E., 344 Lebedeva, T. L. F., 321 Legumes, Fick’s laws of diffusion and water absorption of, 663 Legume starches, limited swelling capability of, 507 Leslie, R. B., 129 Levine, H., 90, 98, 158, 212, 470 Lewicki, P. P., 121t LGG. See Lactobacillus rhamnosus GG Li, Y., 121t Lifschitz-van der Waals interactions, 305 Lignin Agave atrovirens Karw, convection drying and, 619 in fibers, 192 Lillford, Peter, 25 Limbach, H. J., 355 Lindner, K., 153 Linear low-density polyethylene, first-order phase transition (melting) of, by use of common sugar crystals, 433t Lipid-based drug-delivery vehicles formulation properties and consequences of water’s presence in, 328

Index

self-association of water at higher moisture levels observed in, 332 water present in natural oils and other components of, 327 Lipid phase separation, endovesicle formation and scission of membrane invaginations promoted by, 74 Lipids, in fibers, 192 Lipophilic drugs, cyclodextrins and solubility, stability, and bioavailability of, 150 Liquefying, 42 Liquid model, adapting to dough, 49 Liquid state flow in, glass transition, 336–339 of water, 205–206 Liquid uptake, modeling, 220 Lithium chloride salts, in tapioca-flour-based baked products study, 593 Liu, J., 74 Lloyd, R. J., 100 L-menthol, encapsulated, 499 effect of water activity on and n values for release of from spray-dried powder at 50°C storage, 503 effect of water activity on release-rate of, at 50°C, 503 flavor powder preparation, 500 release time of, in spray-dried powders stored at various relative humidities, 502 LMM diluents. See Low molecular mass diluents Local reaction, NEB progress in glassy matrix promoted by, 574, 575 Lorentzian decay, 29, 30 antibiotic hydrates and, 35, 36 Lourden, D., 130, 132 Loveday, S. M., 52 Low-density hydroxyproply cellulose, water content of, at 10% RH and 60% RH, 37t Low-density materials, antiplasticization and, 120 Low-density polyethylene, first-order phase transition (melting) of, by use of common sugar crystals, 433t

729

Low-mobility water, ratio of, to highmobility water in water-excipient mixtures, 37t Low-moisture foods brittleness of, in glassy state, 583–584 knowledge of water plasticization and crystallization behaviors of sugars and sugar alcohols important for use in, 482 protection for, by encapsulation in glassy protein-carbohydrate films and/or matrices, 577 Low-moisture systems, molecular mobility changes and, 216 Low molecular mass diluents antiplasticization of food polymer systems by, 115–133 as antiplasticizers at T less than Tg, 130 conclusion, 133 introduction, 115–116 moisture content reduction in food polymer system and, 119 nonwater diluents as antiplasticizers at T less than Tg, 130–132 polymer-diluent interactions: plasticization vs. antiplasticization, 116–119 properties of diluent (water), 126–129 thermodynamics of water sorption, 128–129 water diffusivity, 129 properties of food polymer-water blend, 119–120, 122–126 gas transport properties, 126 mechanical properties, 119–120, 122–124 thermal properties, 124–126 uses for, 115 Low molecular weight solutes, in dough, 50 Low-water foods, control of physicochemical and flow properties of, 335 Low-water food systems collapse phenomena, 341–345 collapse requirements, 345 stickiness and caking, 344–345 crystallization, collapse, and glass transition in, 335–349 conclusions, 349

730

Index

Low-water food systems (continued) introduction, 335–336 state diagrams, 348–349 crystallization phenomena, 345–348 crystallization of amorphous sugars, 346–348 ice formation and recrystallization, 348 flow in glassy and liquid states, 336–341 glass transition, 336–339 glass transition as characteristic property of, 337 relaxation phenomena of amorphous foods, 339–341 enthalpy relaxations, 340 structural relaxations, 341 Lucas, C. H., 387 Lucas-Washburn equation rehydration modeling and, 223 variation of, for modeling rehydration of dried foods, 233 Lucifer yellow stained cells, Saccharomyces cerevisiae, hyperosmotic treatments and increase in, 73, 74 LUVs. See Large unilamellar vehicles Lycopene solubility, in complex microemulsion, addition of water and changes in, 329 Lyophilized protein formulations, correlations observed for stability of, 26 Lyovac GT2 freeze dryer, 564 Lysine embedding of, in glassy trehalose matrix, 572 glassy food model preparation and, 572 M Macadamia nuts integral entropies of, 683, 684 lyophilized, water-sorption date for, 683 Magnesium chloride effect of, on sucrose curves, 17 evaluating effect of, on kinetics of nonenzymatic browning, 539–544 in tapioca-flour-based baked products study, 593 Magnesium nitrate, in tapioca-flour-based baked products study, 593

Magnetic resonance imaging findings related to characterization of water fraction of soy bread with, 175 monitoring water distribution inside food body with, 533 Maier, A., 55t Maillard reaction, 10–15, 157, 354, 544 controlling kinetics of, 11 factors governing rate of, 539 fried potato color and, 526 glass transition temperatures as function of mass fraction of water for polymeric, lactose and trehalose matrices and development of, 12 glass transition temperatures as function of water for apple, cabbage, and potato and development of, 13 influences on, 11 molecular mobility and, 161–162 in absence of water but below glass transition temperature, 162 potato disc study and, 438 structural effects related to, 20–21 water uptake and mobility, pharmaceutical stability in amorphous solids, 325 Maize starch starch granule-associated proteins and, 508 waxy, Fickian diffusion used to describe sorption dynamics of, 166 Major length, fractal analysis of agave slices during convective drying and, 621 Makower, B., 95, 96, 97, 346 Maltitol BET water sorption isotherms, 478–479 freeze-dried comparison of water sorption and crystallization behaviors of, 477–482 experimental glass transition temperature and water contents after storage of 98h for, 479, 480 materials and methods in comparative study, 478

Index

water content of, during storage at 25°C at various relative vapor pressures, 480, 480–481 X-ray powder diffraction patterns of, at end of storage at 25°C at various relative vapor pressure of 54%, 481, 481 Maltodextrin DE 12, thermally annealed, hole size of, in glassy and rubbery states, 356, 357 Maltodextrin DE 21, effect of water and fat contents on enthalpy of dissolution of, 42 Maltodextrin-fructose mixture spray-dried, Tg and yield from DSC of carrot fiber-fructose and, at different ratios, 196t spray-dried powders of fructose-carrot fiber compared to those of, 191, 193, 195, 196 Maltodextrin matrix annealed, hole volume and specific volume of, with given equilibration as function of temperature for various molecular weight distributions, 360 average fractal dimension of edge of, within first 30s of their disintegration and diffusion, 516, 516 average fractal dimension of side face of, within first 30s of their disintegration and diffusion, tangential perspective showing three layers, 517, 517 evaluation of disintegration and diffusion of, by image processing and nonlinear dynamics, 515–520 showing disintegration process comparable to opening of petals of flower, with fractures during such opening, 517, 520 Maltodextrin powders dispersability and wettability evaluated at different feed concentrations and drying air temperatures in production of, 677t

731

influence of high air-drying temperatures and high feed concentration on, 678 materials and methods, 674–675 microstructure and morphology of particles, 674 physical properties, 675 rehydration properties, 675 testing materials, 674 microstructural, physical, and rehydration properties of, obtained by spray drying, 673–678 conclusion, 678 introduction, 673–674 results and discussion, 675, 677–678 microstructure of, obtained at 40% total solids feed concentration and inlet/ outlet drying air: 249°C/120°C and 140°C, 676 moisture content, bulk and tapped density, and angle of repose of, evaluated at different feed concentrations and drying air temperatures, 676t Maltodextrins as capsule materials, 564 relationships between glass transition temperature, water activity, and water content of, 566 of varying molecular weight distribution, oxidation of citrus oil encapsulated in, at 45°C, 367–368, 368 Maltodextrin-water mixture, ESR use and increase in rotational mobility of spin probes within, 94 Maltooligomers, Arrhenius plot of diffusion coefficient of water in, for three carbohydrate concentrations, 364, 365, 366 Maltooligomer-water mixtures, temperature dependence of free-volume fraction of, at carbohydrate concentration of 90wt%, 362 Maltopolymer-maltose matrices, hole volume as function of weight fraction of water for range of, equilibrated at water activities between 0 and 0.75, 363

732

Index

Maltose approximate solubility of, in glycerol at 60°C after equilibration period of 10 days, 160t BET water-sorption isotherms of, 478 embedding of, in glassy trehalose matrix, 572 first-order phase transition (melting) of, by use of common sugar crystals, 433t freeze-dried comparison of water sorption and crystallization behaviors of, 477–482 experimental glass transition temperature and water contents after storage of 94h for, 479, 479–480 intermolecular hydrogen bonds, NEB progress in the glassy matrix and, 576 materials and methods in comparative study, 478 water content of, during storage at 25°C at various relative vapor pressures, 480, 480–481 X-ray powder diffraction patterns of, at end of storage at 25°C at various relative vapor pressure of 76%, 481, 481 Mandala, I. G., 121t Mannitol, in freeze-dried formulations, 20 Manuel, H., 636 Maourulis, Z. B., 616 Marine-inspired water-structured biomaterials, 385–395 gel structures, hierarchy, and mass transport, 393–395 introduction, 385–386 jellyfish (Aurelia aurita), 386–387 microstructure of, 389–393 water-holding and texture properties of, 387–389 life cycle and accessibility, 387 Marles, C., 91 Marzec, A., 121t Mass-loss kinetics, for bread dough, 629, 630–631

Mass spectrometry, fruit flavor assessment and, 657 Mass transfer dissolution of soluble food powders and coupling of heat and, 46 limited crystal growth and, 272 Mass transport, illustration of, by hydrodynamic flow, capillary flow, and diffusion, 394, 395 Mateus, M. L., 497 Mathematical modeling for rehydration of dried foods, 220–222 diffusion model, 221–222 empirical and semiempirical models, 220 Matric potential, 225 Maximum force of compression, onset of glass transition temperature and, at different water contents for commercial cornflakes, 587, 588 Maximum forces moisture content vs., for control potato slices fried at given temperatures, 529, 529 Mazzobre, M. F., 16, 18, 104 MC. See Moisture content MCA. See Mean cell area MCC. See Microcrystalline cellulose McLaren, C., 101 McMinn, W., 487 Mean cell area, image processing of bread dough and, 632, 633 Meat protein, protein films developed from, 453 Mechanical disturbance, time-dependent response of amorphous materials to, 341 Mechanical properties, 133 complex food products wherein water acts as antiplasticizer on, 121t for cornflakes, measurement of, 584 of duck egg-white film made from shell eggs, 456–457, 457t of food polymer-water blend, 119–120, 122–124 shelf-life stability and textural properties of food products and, 583

Index

Mechanical relaxation, in amorphous materials below and around glass transition, 342 Mechanistic enzyme competitive inhibition equations, 53 Melting, glass transitions overlapping with, at some levels of water plasticization, 337 Melting enthalpy, thermal properties of KGM films and, 125, 125 Melting-enthalpy-moisture-content relationships, of konjac glucomannan films plasticized with glycerol and sorbitol, 125 Melting-point curves, on state diagram, 91 Membrane state, survival of osmotic stresses linked to, 77 Mesoglea, 385, 389, 390 capillary flow and diffusion in, 394 confocal laser scanning microscopic images of cells and fibers in, 391 different cation contents of seawater, dialyzed jellyfish and, 392 filamentous structures in, with interdispersed cells, close to surface, 390 gel structures, hierarchy, and mass transport in, 393–395 illustration of possible structural roles of proteoglucans in, 394 sample taken from, for puncture test where stress and strain were recorded in Instron universal testing machine, 388, 388 TEM and CLSM images of, 392, 392 Metal-oxide semiconductor sensor, aroma analysis and, 658 Methanol dielectric constants of, 159t mixtures of, in probiotic cultures, 285 Methyl cellulose, water content of, at 10% RH and 60% RH, 37t Metrics for water, 206–217 amount of water, 207–208 molecular mobility approach to, 212–217 phase and state diagrams, 212–216 spectroscopic approach to, 216–217

733

sorption isotherms, 211–212 thermodynamic approach to, 208–211 Mettler Toledo balance, 584 Mexican agave. See Agave atrovirens Karw Mexican National Council of Science and Technology, 520 Mexican tequila industry, Mexican agave’s importance for, 619 MF. See Maximum forces MIC. See Minimum inhibitory concentration Microbial inactivation parallel changes with pressure and temperature of protein behavior, water structure and, 80–82 synergistic effects of high pressure and low or subzero temperatures on, 78 Microbial stabilization, lactic acid bacteria commercialization and, 141 Microcapsules, defined, 564 Microcrystalline cellulose, 126 apparent crystallinity index of, obtained from x-ray diffraction and Fourier transform infrared spectroscopy, 310 compaction properties of, 302 materials and methods mechanical properties, 303 microcrystalline cellulose, 302 preparation of compacts, 302–303 solid fraction and porosity, 304 true density, 304 moisture-induced antiplasticization of: permeability of water and octane, tensile strength, Young’s modulus, and crystallinity, 127 moisture-sorption isotherm of, measured at 25°C, 305 plasticization-antiplasticization threshold of water in, 301–313 conclusions, 313 introduction, 301–302 materials and methods, 302–304 results and discussions, 304–312 plots of porosity vs. compaction pressure showing compressibility of, under different conditions of relative humidity, 306, 306

734

Index

Microcrystalline cellulose (continued) relationship between tensile strength and compaction pressure of, 309 typical rehydration data for, at 25°C, 231 typical water-sorption and water-retention data for, 230 typical water-sorption isotherm for, 229 unique place of, in pharmaceutical applications, 301 water content of, at 10% RH and 60% RH, 37t Microcrystalline cellulose compacts, tensile strength of, effect of relative humidity on at constant porosity of 0.3 and 0.1, 307 Microcrystalline cellulose powder, packing ability of, 305 Microdomain distribution in food matrices conclusions, 66 glass transition temperature, water mobility, and reaction kinetics evidence in model dough system, 59–66 ESR experiments, 60–61 introduction, 59–60 mathematical model, 61 NMR experiments, 60 sample preparation, 60 results and discussion glass transition temperature, 61 reaction kinetics, 64–66 water mobility, 62–64 Microemulsions, complex microstructures of, 329 Microorganism viability, role of water in, 71 Microstructural domains, defined, 59 Microstructure of porous foods, X-ray microtomography and observation of, 234 Mille, Y., 74 Milled rice, cooking time for, 663 Milled rice grains materials and methods, 664–665 determination of diffusion coefficient of water, 665 determination of moisture content during soaking, 664–665

determination of solid loss during soaking, 665 rice samples, 664 results and discussion, 665, 667–669 solid loss of rice grains during soaking, 668–669 water diffusion in rice grains during soaking, 668 water uptake of rice grains during soaking, 665, 667–668 water uptake and solid loss during soaking, 663–671 conclusion, 669 introduction, 663–664 Miller, D. P., 543 Milli-Q grade water, lactitol mixed with, 272 Minimum inhibitory concentration antifungal activity of essential oils and, 546 of essential oils and potassium sorbate against Rhizopus stolonifer and Aspergillus flavus, 547t Minimum integral entropy occurrence of, 682 as related to food stability, assessment of, 681, 682, 684, 685, 686 Minitab 13 software, 516 Miscible components in foods, compositiondependent glass transition behavior in, 336–337 Mitutoyo Digimatic Thickness Gauge, eggwhite film thickness determined with, 454 ML. See Major length MLRs. See Multiple linear regressions Modified Page model moisture rate of rice grains predicted with, 664t rice varieties and, regression coefficient and percentage of root mean square error obtained from model fitting with, 667t Moisture, storage longevity of seeds and, 647 Moisture content adsorption surface and integral entropy changes as function of, 684, 685 of agave slices, 620

Index

of bread crumbs, 607 of bread dough, 629 changes in, for starch fraction during freezing at various temperatures for samples with 30%, 40%, and 50% initial moisture content, 187 changes in fracture stress and fracture strain for gluten and starch as function of, 124 in dough samples, ESR spectral patterns for TEMPO and, 64 effect of, on thermodynamic response of dissolving powders, 42 enthalpy of food powder dissolution and effects of, 41, 42, 43, 45, 46, 47 expanded solid foams and, 247 for gelatinized Cassava starch, 186 Gordon-Taylor fit of effect of, on glass transition temperatures of lactitol, 278, 280, 281 integral entropy of sorption as function of, for macadamia, yogurt, zeolite clinoptilolite, and zeolite Valfor, 683, 684 of lactitol samples, grouped according to their storage conditions and relative humidity, 280 loss of crispness relative to, 105–106 of maltodextrin powders, evaluating, 675 maximum forces vs., for control potato slices fried at given temperatures, 529, 529 mechanical properties of food polymerwater blend and, 120 of milled rice varieties, 664 modeling of, for crystallizing amorphous lactitol, 273–274 potato chips during frying and, 527 reduction of, in food polymer system, 119 relationship among texture, water activity, glass transition and, for glassy tapioca-flour-based baked product, 595, 595–596 rice starch-glycerol mixtures, 488, 488 scanning calorimetry of gluten protein as function of, 248 soaking of milled rice grains and, 664–665

735

soy bread samples, migration during storage and, 179 in spray drying of fructose mixtures, 195–196 stability of food matrix and, 59 for standard and innovative bread crumb samples, 609 in starch fraction, after freezing at −3, −40, and −50°C for samples with 30%, 40%, and 50% overall moisture content, 188t of starch fraction, effect of freezing temperature on time needed to freeze 50% and 100% of available freezable water and corresponding amount of, 189t texture of glassy tapioca-flour-based baked product as function of, 591–596 transition temperatures of chocolate wafers as function of, 110 values for glass transition temperature and, for xanthine oxidase with various excipients, 600, 601 Moisture dependency, mechanical properties and, 120 Moisture increase, influence of, on state change, 93 Moisture in foods, typical Deff values for, 222 Moisture pickup, transition from glassy to rubbery state in stored food and, 95 Moisture ratio, of milled rice grains, 664, 664t Moisture sorption, variations in mechanical behavior in response to, 120 Moisture-sorption isotherms, 211 designing stable multicomponent food materials and, 411 of microcrystalline cellulose measured at 25°C, 305 schematic representation of, 211 Moisture toughening, 119 Mojonnier method, fat content of model powder samples determined by, 43 Mold growth, bread spoilage and, 545

736

Index

Molds, defeating, Lactobacillus coryniformis Si3 and, 142 Molecular diffusion of gases, transport flow in porous media and, 233 Molecular dynamics simulations, 353 for exploring molecular organization of lipid components, water, and solute in lipid-based drug-delivery vehicles, 329 investigating dynamic properties close to glass transition and, 364 observations from PALS experiments verified with, 361 PALS studies on physics of glassy carbohydrate matrices related with, 355–356 snapshot from, of 60% tricaprylin-40% 1-monocaprylin lipid mixture saturated with water at 37°C, 329, 330 snapshot of carbohydrate-water system, 359 spatial distributions of water molecules obtained in, from two instantaneous structures of PVP glasses containing 0.5% and 10% water by weight, 320, 320 water’s role in nonaqueous pharmaceutical systems and, 317–318 Molecular entities, types of motions for, 158 Molecular mobility glass transition temperature, sub-T9 relaxation and, 161 Maillard reaction and, 161–162 quinoa seed longevity and, 647–654 understanding, 158 Molecular mobility approach to metrics for water, 212–217 phase and state diagrams, 212–216 spectroscopic approach to, 216–217 Molecular mobility of water in API hydrates, 26–36 correlation of, as determined by NMR with that as determined by DSC and water-sorption isotherm measurements, 31, 34–36

as determined by NMR, 26–30 ease of evaporation for hydration water as determined by DSC and watersorption isotherm measurements, 31 apparent correlations between stability of solid APIs and, 25–26 coexistence of, with pharmaceutical excipients, 36 Molecular packing, 356 in glassy carbohydrates, 359, 361–362 Molecular relaxation, 161 Monocaprylin-OH oxygen groups, radial distribution function for water oxygen with, 331 Monocaprylin oxygen atoms, radial distribution function for water oxygen with tricaprylin C=O oxygen atoms and, 331 Monoexponential decay curves, for quinoa seeds equilibrated at low RHs, 650 Monoglycerides, water uptake, distribution, and effects on drug solubility in lipid vehicles composed of, 326– 330, 332 Monoglyceride-water mixtures, freezefracture electron microscopy study of organization of lipid molecules in, 329 Monomolecular layer water, 159 Monosaccharides mixtures of, in probiotic cultures, 285 survival during drying and, 141 Moraru, C. I., 121t Moreira, R. G., 527 Morel-Seytoux, H. J., 228, 229 Morita, Y., 635 Morris, V. J., 245 Morrison, William, 95 MPEG-Beta-PCL Fourier transform spectroscopy spectra of, 461, 461 purifying, 460 synthesis of, for blend homogeneous film with chitosan, 459 MR. See Moisture ratio MSIs. See Moisture-sorption isotherms

Index

Multicomponent systems, predicting sorption profile of, 484 Multidomain foods, storage problems with, 94 Multiple linear regressions, 144 Mung-bean (MB) starch granules effect of modifier on transition temperatures in second endotherm at different levels of water content, 510 in first endotherm, effect of water content and protein modifiers on phasetransition characteristics of, 509– 510, 509t light micrograph of, under visible light and cross-polarized light, 510 materials and methods, 508–509 gel microstructure, 508–509 statistical analysis, 509 thermal properties, 508 role of protein modifiers on, as less significant than water’s role on, 513 water and protein modifier effects on phase transitions and microstructures of, 507–513 conclusions, 513 introduction, 507–508 results and discussion, 509 Mung-bean starch gel confocal laser scanning micrographs of, containing 60% water content (wt/wt) prepared by autoclaving at 121°C for 15 min., 511, 511 micrographs of, containing 50% water content (wt/wt) prepared by autoclaving at 121°C for 15 min. under same observation field, 512 Mung bean (Vigna radiata), as major starch crop in Thailand, 507 Murray, B. J., 466, 470 Mustapha, W. A. W., 162 Mycotoxins, molds, bread spoilage and, 545 Myllärinen, P., 486 Myosin, gel formation and, 241

737

N Na-caseinate-lactose sheets, protection against oxygen with, 582 Na-caseinate-trehalose-lactose sheets, protection against oxygen with, 582 Nanostructured (NSM) food-model systems, minimum integral entropy as related to food stability and, 681 Nanotechnology, new food materials development and, 681 National Center for Genetic Engineering and Biotechnology (Thailand), 513 National Council for Science and Technology (Mexico), 687 National Polytechnic Institute of Mexico, 678, 686 Natural flavors, controlled kinetics of Maillard reaction and, 11 Navier-Stokes analog, fluid flow in porous media and, 233 N-deacetylation of chitin, chitosan produced by, 459 NDS. See Nondissolvable solids NEB. See Nonenzymatic browning Neutron scattering in combination with PALS, for insight into structure of carbohydratewater system, 359 investigating dynamic properties close to glass transition and, 364 probing properties of carbohydrates in glassy state with, 355 Newton’s equation of motion, molecular dynamics simulations and, 317 Ngamdee, P., 460 Nicholls, R. J., 107, 123 Nifedipine, amorphous, crystallization rate in solid dispersion formulations, 26, 27 Nikoladis, A., 106 Nimmo, J. R., 228, 229 NL-6 quinoa seeds amplitude of signal of lipid component of, with different water activities, 652 genotype description of, 648

738

Index

NL-6 quinoa seeds (continued) percent germination of, during storage at 32°C and 43% or 75% relative humidity, 650, 650 percent viability of, assessed by tetrazolium test, 651 relaxation times of lipid component of, with different water activities, 652 sorption isotherms of, 651t viability and germination values for, 650 NMR. See Nuclear magnetic resonance Nonaqueous pharmaceutical systems, water’s role in, 315–332 Noncovalent bonds, 158 Noncrystalline solids, glass transition as characteristic property of, 337 Nondissolvable solids, rehydration process and use of, 219–220 Nonenzymatic browning, 11, 90 food quality in processing and storage and importance of, 437–438 in glassy matrix, model of local mobility promoting progress of, before storage at 25°C, storage at 70°C, and storage at 50°C and 60°C, 450, 450 importance of prediction and control of, in dry foods, 571 inhibition or acceleration by magnesium chloride, 539–544 conclusion, 543–544 introduction, 539–540 materials and methods, 540–541 results and discussion, 541–543 model of local reaction promoting progress of in glassy matrix, 574, 575 progress of, in glassy trehalose matrix, 448 proton NMR studies of molecular mobility in potato systems in relation to, 437–442 quality changes caused by, 445 rate of, related to degree of interactions between water and solid molecules and with physical state of food matrix, 442

Nonenzymatic browning reaction of glassy foods conclusions, 576 discussion, 574–576 introduction, 571–576 materials and methods, 572–573 DSC measurement, 572–573 evaluation of initial NEB rate, 572 preparation of glassy food model, 572 results, 573–574 glass transition and NEB rate of samples of various glassy matrices, 573 glass transition and NEB rate of samples of various reducing sugars, 574 Nonenzymatic browning reaction rate Arrhenius plot of for samples of varying glassy matrices, 573, 573 for samples of varying reducing sugars embedded in glassy trehalose matrix, 574 glass transition and, for various reducing sugars, 574 Nonenzymatic Maillard browning, color development due to, in waterless environment at 100°, 115°, and 130°C, 163 Nonequilibrium amorphous structures, higher enthalpy and volume in, vs. in equilibrium crystalline structures under same conditions, 337, 338 Nonequilibrium states, collapse phenomena in foods resulting from timedependent flow in, 337 Nonisotropic shrinkage, 616 Nonlamellar phases, low water concentrations and, 77 Nonlinear dynamics applications of, 515 evaluation of disintegration and diffusion of pharmaceutical solid matrices by, 515–520 Non-nanostructured (NNS) food-model systems, minimum integral entropy as related to food stability and, 681

Index

Nonpermeable cryoprotective agents effect of, on temperature depression, 555–557 effect of ice reduction on stability of frozen LUVs and, 557–560 permeable cryoprotective agents vs., 552 Nonpolar bonds, 158 Nonpolar solvents, dielectric constant of, 159 Nonsolvent fractions, 50 Nonsolvent water, 115, 127 frozen dough and, 49 in semisolid foods, 53, 55 Nonstoichiometric water uptake of bulk drug active pharmaceutical ingredient as function of percent relative humidity at 25°C, 319 moisture uptake in pharmaceutical solids and, 318 Nonwater diluents, as antiplasticizers at T less than Tg, 130 Noodles, cooked, texture of and watercontent distribution in after boiling, 533 Normalized Weibull distribution function curve fitting of rehydration data and, 221t rehydration studies and, 220 Nova-Sina, 90 Noyes, A., 41 Nuclear magnetic resonance, 25 in combination with PALS, for insight into structure of carbohydratewater system, 359 correlation of water mobility as determined by, with that as determined by DSC and watersorption isotherm measurements, 31, 34–36 for determination of Tg, glass transition temperature and, 61 higher mobility of protons as measured by, in rubbery vs. glassy state, 94 hydration effects at level of molecular mobility studied with, 411 investigating dynamic properties close to glass transition and, 364 measuring distribution of water in bread with, 176

739

molecular mobility of hydration as determined by, 26–30 probing properties of carbohydrates in glassy state with, 355 proton-relaxation data from dough fitted with continuous spectra featuring three distinct peaks shown by, 51 sensitivity of technique, to rotational and translational motions of water, 216 usefulness and limits with determining water mobility by, 36 Nuclear magnetic resonance relaxation measurements quinoa seed conditioning and, 649 quinoa seed samples and spin-spin relaxation time measurement, 649 Nuclear magnetic resonance spectra, for dialyzed and freeze-dried jellyfish, 386–387 Nuclear magnetic resonance spectrometers, water movement and, 251–252 Nuclear magnetic resonance spin-spin transverse relaxation, analysis as function of presence of magnesium chloride and/or brown pigment, 543, 543t Nuclear magnetic resonance T2 distribution, discriminating, 257–258 Nucleation, 94 Nutritional value, nonenzymatic browning and changes in, 445 Nylon film, sorption dynamics of, 166 O Ofelt, C. W., 176 Off-aroma scores, of fresh and freeze-thaw Smooth Cayenne and Queen pineapples, 660t Oil fractions, absorption mechanisms in fried potato cylinders and, 525–526 Oil-uptake fractions, kinetics of, in control potato slices during frying at 180°C, 527, 528 Okada, C. Y., 74 Okaka, J., 675 Oligosaccharides in dough, 50 in fibers, 192

740

Index

Oliveira, A. C., 80 Olive oil, dielectric constants of, 159t Oliver, A. E., 553 1,2-dipalmitoyl-rac-glycero-3phosphocholine, LUV dispersions and investigation of effect of ice formation on stability of phospholipid bilayers with, 551, 553 Onset temperatures, ease of evaporation for hydration and factors related to, 31 Oostergetel, G. T., 253 Oscillatory sorption experiment adsorption and desorption rates of starch-rich model bread crust during, 172 on bread crust, 170, 170–171 Osmotic stress cell shrinkage in response to, 77 membrane state linked to survival of, 77 yeast survival and effect of combined thermal stress and, 75–77 Oven drying, water measurement and, 207 Oxidation foods susceptible to, edible oxygen-barrier films and, 641 unsaturation in structure of carotenes and, 18–19 Ox-tran 2/20 ML modular system, 642 Oxygen-oxygen hydrogen bond, comparing orientation about, for ice and clathrate, 206 Oxygen permeability effect of starch chain length on, 641 of 40% noncrystallizing sorbitolplasticized starch films at various degrees of polymerization, 643t for protein-carbohydrate sheets, 581 close to their glass transition temperature, 577, 578 of starch films discussion on, 642–643 measuring, 642 Oxygen permeability tester, 579 Oxygen uptake, by carbohydrateencapsulated citrus oils, 367–368, 368

P PA. See Projected area Page model predicting moisture rate of rice grains with, 664t rice varieties and, regression coefficient and percentage of root mean square error obtained from model fitting with, 667, 667t Page model parameters, for water uptake by milled rice grains at different temperatures, 668t PALS. See Positron annihilation lifetime spectroscopy Paprika capsules preparation of, with alginic acid, 681, 683 with and without nanostructures, chemical stability and its relation to minimum entropy studied for, 685 Parametric models, for relating water content to tension head, 226, 227 Partial pressure of water, sample, general configuration of a cell for determination of, 209 Particulate protein gels at pH 5.5, 240 pores in, 242 Pasta cooked, texture of and water-content distribution in after boiling, 533 glass-rubber transitions measured for, 435 rehydration of, radial nuclear magnetic resonance study of, 224 Pasta dough, heated at 5°C, 246 Pasta systems, dry, Maillard reaction in, 162 Pasteurization, effects of various perturbations on, 72 Pasting properties, of heat-moisture-treated rice starches, 636 Pasting temperature relationship of setback to moisture of treated rice starch and, 638 of rice starch, 635 Payne, C., 107 PCA. See Principal component analysis Peaches, volume decrease or shrinkage studied in, 613 Pectin gel, syneresis study of, 267

Index

Pedreschi, F., 527 Peleg, M., 108, 120, 345 Peleg’s model curve fitting of rehydration data and, 221t liquid transport into dried solid matrix and, 233 Penetrated surface oil, absorption mechanisms in fried potato cylinders and, 526 Pentose, intermolecular hydrogen bonds, NEB progress in the glassy matrix and, 576 Peptone, mixtures of, in probiotic cultures, 285 PerkinElmer Spectrum GX Series Fourier transform infrared spectrophotometer, 460 Permeability softening of a polymer and, 402 water’s antiplasticizing effect and, 404–405 Permeability values of water in Eudragit E100, as function of relative humidity, 406 in Eudragit E100, as function of water content, 405 Permeable cryoprotective agents effect of, on temperature depression, 557 effect of ice reduction on stability of frozen LUVs and, 560–561 nonpermeable cryoprotective agents vs., 552 Pharmaceutical excipients, molecular mobility of water coexisting with, 36 Pharmaceutical industry cyclodextrins used in, 150 decoupling of mobility of water from mobility of carbohydrate molecules in approach to glass transition and implications for, 366–369 Pharmaceutical products, knowledge of water plasticization and crystallization behaviors of sugars and sugar alcohols for, 482 Pharmaceutical sciences carbohydrate polymers and role in, 301 water as most common plasticizer in, 401

741

Pharmaceutical solid matrices evaluation of disintegration and diffusion of, by image processing and nonlinear dynamics, 515–520 conclusions, 520 introduction, 515 methods, 515–516 results and discussion, 516–517, 519–520 Pharmaceutical stability, glassy carbohydrates in, 353–356 Pharmacology, cyclodextrins used as drugcarrier systems in, 150 Phase diagrams molecular mobility approach in water measurement and, 212–214 temperature-composition, for binary aqueous system, 213 Phase-separated water, in soy bread with and without almond during storage and, 182 Phase separation, gelation of mixed systems and, 245 Phase-transition phenomena, interaction of high pressure and subzero temperature in microbial inactivation and, 78 Phosphate salts, mannitol crystallization during freeze drying and, 20 Physical aging, 419 Physical product stability, criteria for, 366t Physical state changes, of foods in storage, 93–94 PID principles. See Proportional integral derivative principles Pigments, controlled kinetics of Maillard reaction and, 11 Pineapple off-aroma effect of freeze-thaw cycle on, by using electronic nose technique, 657–661 conclusion, 661 introduction, 657–658 results and discussion, 659–661 materials and methods, 658–659 electronic nose measurements, 659 freezing and thawing, 658

742

Index

Pineapple off-aroma (continued) sample preparation, 658 sensory evaluation, 659 statistical analysis, 659 Pipemidic acid differential scanning calorimetric thermograms for, 32 endothermic peak for, 31 water-sorption isotherms for, 33 Pipemidic acid hydrates, temperature dependence of T2 for, 34 Pittia, P., 121t, 493 Pizzoli, M., 381 Plasma membrane, cell mortality and changes in fluidity of, 76 Plasticization, 122 antiplasticization vs.: polymer-diluent interactions, 116–119 depiction of feasible explanation for association between antiplasticization and, based on bulk free volume, 311, 311–312 water, glass transition temperature of glassy polymer and, 401–402 Plasticization threshold, 117, 120 Plasticizers increased molecular mobility and, 158 primary effects of, 116 Plasticizer threshold gas transport properties and, 126 glycerol and, 132 Plasticizing effect of water loss of sensory brittleness in foods relative to, 108, 111 in raw and roasted coffee beans, 491 textural properties of raw and roasted coffee and, 494 PMMA. See Polymethyl methacrylate Poiseuille equation, for laminar flow, 223 Polarity, relative, of various compounds, 159t Polarity of solvent, determining, 159–160 Polyacrylamide gels cross-linked, ice crystallization during rewarming of, 373 ice crystallization exotherm during rewarming of, 377

Polyethylene, first-order phase transition (melting) of, by use of common sugar crystals, 433t Poly (lactide-co-glycolide), anomalous antiplasticization effects of water in relation to some mechanical properties demonstrated in, 122 Poly-L-hydroxyproline, water acting as antiplasticizer by increasing rigidity to maximum value, 123 Polymer-diluent interactions, plasticization vs. antiplasticization, 116–119 Polymer gels freezing behavior and, 373–383 conclusions, 382 freezing scheme for compartmentalized water in polymer gels, 379–380 glass transition of cross-linked dextrans containing small amount of water, 381 ice crystallization during rewarming observed with various food biolpolymer gels, 381–382 ice crystallization exotherm during rewarming observed with crosslinked dextran gels, 374–375, 377 introduction, 373–374 origin of endothermic trend observed prior to exotherm during rewarming, 377–379 vitrified water in a frozen G25 gel, 379 physical state of unfrozen water in, 374 scheme for ice crystallization of compartmentalized water in, 380t Polymeric glassy matrices, enzyme stability and, 16 Polymerization, glass transition temperatures of starch films with different degrees of, 644 Polymer matrix intermolecular hydrogen bonds and, 575 model of local reaction promoting nonenzymatic browning reaction progress in glassy matrix and, 574, 575

Index

Polymers based on phenylpropane units, in fibers, 192 as effective stabilizers, 600 elongation, permeability, and softening of, 402 importance of, in various disciplines, 301 inhibiting crystallization and, 17 plasticizers and softening effect in, 401 synthetic enthalpy relaxations in, 340 similarities between collapse temperature and glass transition behavior of, 345 Polymethyl methacrylate, crack propagation by void formation in CS films compared with crazing mechanism observed in, 294 Polyols, 163 acting as antiplasticizers in variety of biopolymer systems, 131 crystallization of, 272 use of, in food industry, 483 waterlike solvation property of, 159–161 Polypropylene, first-order phase transition (melting) of, by use of common sugar crystals, 433t Polysaccharides in fibers, 192 in jellyfish, 389, 390, 392 Polyunsaturated fatty acids, oxygen uptake and, 367 Polyvinylacetate, water uptake profiles for, 318 Polyvinyl chloride, crack propagation by void formation in CS films compared with crazing mechanism observed in, 294 Polyvinylpyrrolidone, 315 intermolecular hydrogen bonds and, 575 radial distribution functions between oxygen atoms in water and carbonyl oxygen atoms in or hydrogen atoms in PVP methylene groups at 0.5% water content, 320–321, 321

743

representative displacement profiles vs. simulation time for two water molecules in, at 0.5% wt/wt water and 298K, 322–323, 323 self-association of water at higher moisture levels observed in, 332 water uptake and its distribution in, 318–322 Polyvinylpyrrolidone carbonyl oxygen atom, probability distributions for number of water molecules surrounding, or given water molecule in simulated PVP glass with 10% wt/wt water, 322 Pommet, M., 124 Poole, P. L., 238 Popcorn, hot-oil puffed, loss of crispness in, 105 Pore size index, 226 Pore sizes, in particulate protein gels, 242, 243 Porosity changes in blanched potato slices fried at 180°C, 529, 530 plots of compaction pressure vs., compressibility of microcrystalline cellulose under different conditions of relative humidity and, 306, 306 potato chip frying and, 526 of tablet, defined, 304 Porous media. See also Unsaturated porous media capillary flow in, 222–224 developing theory of flow in, 228–231 prediction and validation of retention curve from water-activity data, 230–231 retention curve, 230 retention-curve validation, 231 water-sorption isotherm, 229 kinetics of liquid penetration and characterization of, 232 rehydration of foods and use of, 231–232 water-retention curve and, 225–227

744

Index

Positron annihilation lifetime spectroscopy, 353 exploring molecular structure of amorphous carbohydrates in glassy state with, 355, 356, 357 insight into structure of carbohydratewater system by combining other techniques with, 359 Post-deformation plastic zone, of biopolymer films, 295 Post-deformation plastic zones, Fourier transform infrared spectroscopy used with, in monitoring fracture behavior of biopolymer films prepared from aqueous solutions, 292 Potassium acetate salts, in tapioca-flourbased baked products study, 593 Potassium carbonate, in tapioca-flour-based baked products study, 593 Potassium chloride, in tapioca-flour-based baked products study, 593 Potassium chloride-glucose-water mixture, at 25°C, differences in relative viscosities of ternary solutions with, 474, 475 Potassium iodide, in tapioca-flour-based baked products study, 593 Potassium ions effects of, on viscosities in sodium/ potassium-glucose-water ternary system, 473–475 sodium ions vs., roles of, in living bodies, 473–474 Potassium sulfate, in tapioca-flour-based baked products study, 593 Potato, volume decrease or shrinkage studied in, 613 Potato chips crispness of, 526 effect of violent drying during frying on texture, color, oil content, and porosity of, 525, 526, 527, 529–530 frying of control slices, observed and predicted water loss during, 527, 528

loss of crispness in, 105 low-fat and fat-free, consumers’ preferences for, 525 materials and methods, 526–527 analysis, 526–527 frying conditions, 526 materials, 526 modeling water loss, 527 moisture content and loss of crispness in, 106 tapioca flour in, 591 water content and physical properties of, 525–530 conclusions, 529–530 introduction, 525 results and discussion, 527, 529 Potato discs, color development at 70°C in, at different water activity, changes in luminosity value, 439 Potato growth rings, 253 Potato powder spin-spin relaxation time obtained by single 90° pulse by proton NMR for, as a function of temperature, 440, 441 spin-spin relaxation time obtained by spinecho Hahn sequence for, as function of water content, 441, 441 water-sorption isotherms for, at 23°C, fitted by Guggenheim-Andeson-de Boer equation, 439, 440 Potato slabs effect of temperature on kinetics of lateral projected area of, during drying at air velocity = 1.7 m/s and thickness = 2.5 mm, 615, 616 kinetics of, during drying at 55°C, air velocity = 1.7 m/s, and thickness = 2.5 mm, 615, 615 two shrinkage-deformation stages during convective drying of, 613, 615, 616, 617 variation in lateral projected area of, during drying at conditions indicated, 616, 617

Index

Potato slabs during convective drying evaluation of deformation and shrinking of, 613–617 conclusions, 617 introduction, 613–614 materials and methods, 614 drying equipment and image acquisition system, 614 image processing, 614 materials, 614 results and discussion, 615–616 image analysis in evaluation of shrinkage and deformation, 615 influence of drying conditions on shrinkage and deformation, 615–616 Potato slices blanched and fried at 180°C changes in porosity of, 529, 530 color evolutions of, 529, 530 control kinetics of oil-uptake fractions and total oil in, during frying at 180°C, 527, 528 maximum forces vs. moisture content for, fried at given temperatures, 529, 529 effective moisture diffusivity during frying of, at given temperatures, 527, 528 Potato starch leveling of melting enthalpies for, 277 undergoing gelatinization, SEM and TEM micrographs of, 255 Potato-starch extrudates, amorphous, water as antiplasticizer and, 126 Potato-starch gels, gelatinized, differential scanning calorimetry rewarming traces obtained with, 382 Potato starch synthesis, impact of drought on, 252 Potato systems proton NMR studies of molecular mobility in, relative to nonenzymatic browning, 437–442 conclusions, 442 introduction, 437–438

745

materials and methods, 438–439 results and discussion, 439–442 Potter, N., 675 Povidone, water content of, at 10% RH and 60% RH, 37t Powder agglomeration, surface plasticization and dehydration as basis of, 343 Powdered anhydrous milk systems, Maillard reaction below glass transition temperature in, 162 Power-function model parameters, for loss of solids from milled rice grains at different temperatures, 671t Power-law model, 165 shear-thinning behavior in heat-moisturetreated rice-starch pastes according to, 639 sorption dynamics for dry, crispy bread crust described by, 165, 167 Prado, S. M., 20 Preferential exclusion mechanism, 599 Premix chamber, 605, 606 Pressure potential, 225 Principal component analysis elaboration of sensor responses and, 658 for fresh and freeze-thaw Smooth Cayenne and Queen pineapple cultivars determined by e-nose, 659, 660 Probiotic bacteria entrapment of, in frozen cryoprotectants and viability in freeze drying, 285–289 introduction, 285–286 materials and methods, 286–287 results and discussion, 287–289 Probiotic bacteria cells, entrapment of, in frozen and freeze-dried carbohydrate (cryoprotectant) matrices, 288 Progesterone, water content and solubility of, 330 Progressive chain association, gels, syneresis and, 242 Projected area fractal analysis of agave slices during convective drying and, 621 image analysis and qualifying parameters related to, 613

746

Index

Prokaryotic cells, physical or physicochemical environment changes and stress on, 71 Proofing (or proving) stage, frozen dough, 49 Propeller method, advance caking point by, 100, 101 Property curve, in state diagram for binary aqueous system, 214 Propidium iodide/Lucifer yellow doublestained cells, viability of Saccharomyces cerevisiae vs. water activity and use of, 73, 74 Proportional integral derivative principles, 430 Propylene glycol, dielectric constants of, 159t Protective media for biological systems/ ingredients, controlled kinetics of Maillard reaction and, 11 Protein behavior, parallel changes with pressure and temperature of, plus microbial inactivation and water structure, 80–82 Protein-carbohydrate sheets abbreviations, composition, feed rate of the solid and water and total feed rate, 578t Arrhenius plot of alpha relaxation and beta relaxation of dielectric spectra for, 580, 580 mechanical and oxygen permeability properties determined for, 580– 581, 581 Protein-carbohydrate sheets produced by twin-screw extruder physical properties of, 577–582 conclusions, 582 introduction, 577–578 materials and methods, 578–579 results and discussion, 579–581 Protein films, development of, 453 Protein hydration in structure creation, 237–249 gelation, 239–249 emulsion stabilization, 245 mixed systems, 243–245 single-protein systems, 239–242

structuring by extrusion, 246–249 water holding, 242–243 hydration studies, 238–239 introduction, 237 Protein modifiers, influences of, on starch network architecture at intermediate water content, 511 Proteins beneficial effects with prior denaturation of, during raw material preparation, 247 crispiness in tapioca baked samples and, 596 delay in lactose crystallization and, 17 freeze-dried, mechanisms of stabilization of, by excipients during dehydration, 600 functional performance of, 237 gelation of, determination of, 239 mixed systems and, 243–245 protein-carbohydrated sheets prepared with, 578 structure of water and thermal denaturation of, 81 water holding capacity for, 242–243 Protein stability, in solid state vs. in liquid state, 16 Proteoglycans in mesoglea, illustration of possible structural roles of, 394 water-holding capacity in jellyfish and, 392–393 Proton nuclear magnetic resonance, 26 findings related to characterization of water fraction of soy bread with, 175 spin-spin relaxation time obtained by 90° pulse by, for potato powder as a function of temperature, 440, 441 spin-spin transverse relaxation time measured by, 583, 584 water mobility determination with, in magnesium chloride and kinetics of NEB analysis, 539, 541 Proton relaxometry, continuous model of, in application to food systems, 257 Proton spin-spin relaxation time, T2, water mobility in model dough system and, 62–64

Index

P6 gel, ice crystallization exotherm during rewarming of, 377 PSO. See Penetrated surface oil PUFAs. See Polyunsaturated fatty acids Puffed cereal products, sensorily perceived moisture hardening or toughening in, 122 Puffed curls, tapioca flour in, 591 Puffed-rice cakes, loss of crispness in, 105 Pullulan, anomalous antiplasticization effects of water in relation to some mechanical properties demonstrated in, 122 Pulque, extraction of, from Agave atrovirens Karw, 619 Pulsed nuclear magnetic resonance, mobility of water in foods investigated by, 540 Pummelo peel oil, minimum inhibitory concentrations of and potassium sorbate against Rhizopus stolonifer and Aspergillus flavus, 547t Pummelo peels, preliminary examination of, for antifungal activities by an agar dilution method, 545 PVAc. See Polyvinylacetate PVC. See Polyvinyl chloride PVP. See Polyvinylpyrrolidone PVP glass relationship between diffusion coefficient for water in, containing 0.5% or 10% wt/wt water and the number of first-shell water molecules, 324, 324–325 simulated, water mobility in, 322–325 PVP systems, Maillard reaction rate and, 11 Pyris Diamond differential scanning calorimeter, 485 Pyris 1 differential scanning calorimeter, 642 Q Quality of food, distribution of microstructural domains and assessment of, 66 Quartz microbalance sensor, aroma analysis and, 658

747

Queen pineapple, 657 freezing and thawing of sample, 658 fresh and freeze-thaw off-aroma scores of, 660t principal component analysis plot for, as determined by e-nose, 659, 660 sample preparation, 658 Quinidine sulfate differential scanning calorimetric thermograms for, 32 free-induction decay exhibited by, 30 Gaussian-type decay exhibited by, 29 Quinidine sulfate hydrate, endothermic peak for, 31 Quinine hydrochloride differential scanning calorimetric thermograms for, 32 endothermic peaks for, 31 R Raffinose adding to cotton candy, onset time of crystallization and, 98 gelatin and inhibiting crystallization of, 17 Raffinose matrices, intermolecular hydrogen bonds and, 575 Rahman, M. S., 91 Raman spectroscopic studies, on Sephadex gels, 379 Raman spectroscopy, vibrational motions measured with, 216 Reactant molecules, collisions between, 158 Reaction kinetics correlations between details of moisturesorption isotherm and, 211 important modifiers of, 21 Reaction kinetics evidence microdomain distribution in food matrices: model dough system, 64–66 in model dough systems, 59 Reaction rate constant (k) distribution of, as function of moisture content in dough at 288 k, 65 distribution of, as function of temperature for 35% moisture dough, 65 microdomain distribution in food matrices and, 59 Rechsteiner, M., 74

748

Index

Recombinant human factor XIII, stabilization of, 600 Recrystallization ice formation and, 348 time-dependent, of amorphous lactose, 347 Redispersion, 42 Redissolution, 42 Reducing sugars intermolecular hydrogen bonds and, 576 model of local reaction promoting nonenzymatic browning reaction progress in glassy matrix and, 574, 575 Rehydration, 42 membrane structure and fluidity and, 71 understanding plasma membrane changes occurring during, 77 Rehydration data, empirical models frequently used in curve fitting of, 221t Rehydration modeling, interdisciplinary approaches to, 223 Rehydration modeling of food particulates principles of water transport in porous media and, 219–234 capillary flow in porous media, 222–224 conclusions, 234 diffusion model and, 221–222 empirical and semiempirical models, 220 flow in unsaturated porous media, 224–232 introduction, 219–220 mathematical modeling, 220–222 new approaches and other advances, 232–234 research needs, 234 Rehydration of food powders, operating conditions in spray drying and, 673 Rehydration of pasta, radial nuclear magnetic resonance study of, 224 Rehydration process, new interdisciplinary insights into mechanisms involved in, 233 Rehydration ratio, 220 Rehydration step, interdependence of changes occurring in dehydration step and, 74

Reid, D. S., 207 Reidy, G. A., 121t Reifenberger, E., 55t Reineccius, G. A., 367, 500 Relative humidity adsorption and desorption rates, and relative change of weight of starch-rich bread crust during isotherm experiment and single exponential model as function of, 169 crystallization of lactose dependence on water content and, during storage, 272 effect of, on release of D-limonene from spray-dried powder, 501 effect of, on release of L-menthol from spray-dried powder, 502 glass transition temperature of lactitol, and predicting crystallization at different levels of, 278, 280–281 levels in desiccators, saturated salt solutions used in maintenance of, 303t low and high, summary of fitted diffusional coefficient and goodness of fits of diffusion and exponential models for, 171t onset temperature for crystallization of, as function of, 103 permeability of water in Eudragit E100 as function of, 406 sample water peak melting enthalpy and moisture content of lactitol samples grouped according to storage conditions and, 280 saturated salt solutions at varying levels of, at 20° and 32°C, 273t water sorption and transport in dry, crispy bread crust relative to, 167 Relative vapor pressure, 208 food stability and, 212 schematic representation of moisturesorption isotherm and, 211 water content of lactose, maltose, lactitol, and maltitol during storage at 25°C and, 480, 480–481

Index

Relative water content model. See also Water content application of, to wheat-flour dough during boiling, 534 water migration in multiphase food systems described by, 533 Relaxation, types of, occurring within glassy state, 116–117 Rennet milk gels, structure of, 242 Residual saturation, defined, 228 Residual water content, defined, 226 Response surface methodology, 143 Retention curve developing theory of flow in porous media and, 230 prediction and validation of, from wateractivity data, 230–231 Retention-curve validation, flow in porous media and, 231 Rewarming ice crystallization during, observed with various food polymer gels, 381–382 ice crystallization exotherm during, as observed with cross-linked dextran gels, 374–375, 377 origin of endothermic trend observed prior to exotherm during, 377–379 RH. See Relative humidity Rheological properties, of high-amylose rice starch, 635–639 Rhizopus stolonifer combined effect of cinnamon essential oil and water activity on, 545–550 introduction, 545 materials and methods, 546–547 results and discussion, 547, 549–550 effect of cinnamon and relative humidity on growth inhibition of Aspergillus flavus on bread during storage at 30°C, 549 effect of cinnamon oil and water activity on inhibition of, on bread model agar, 546, 547, 548, 549 minimum inhibitory concentrations of essential oils and potassium sorbate against, 547t

749

Rhodotorula rubra, baroprotective effect of solutes and, 79 Rice, cooked, texture of and water-content distribution in after boiling, 533 Rice-based starches, oxygen permeability and, 641 Rice flour, five freeze-thaw cycles, T2 distribution of protons in, 267, 268 Rice-glycerol extrudates, GuggenheimAnderson-de Boer parameters obtained for, 487, 487t Rice grains, diverse soaking characteristics of different cultivars of, 664 Rice kernels, glass-rubber transitions measured for, 435 Rice Research Center (Thailand), 635 Rice starch extruded, sorption behavior of in presence of glycerol, 483–489 conclusions, 488–489 introduction, 483–484 extruded, sorption isotherms for, with 0%, 5%, 10%, 20% glycerol, 485–486, 486 extruded, X-ray diffraction patterns for, with 0%, 5%, 10%, and 20% glycerol, 485, 486 five freeze-thaw cycles, T2 distribution of protons in, 267, 268 heat-moisture-treated, relationship of apparent viscosity and shear rate of native rice starch and, at 18%, 21%, 4%, and 27% moisture content over shear rate range of 0–300 s−1 and at 60°C, 637 materials and methods, 484–485, 635–636 DSC, 485 DVS, 485 flow behavior, 636 heat-moisture treatment, 636 pasting properties, 636 WAXD, 485 results and discussion, 485–488, 636 DVS, 485–486 flow behavior, 636 modeling sorption isotherms, 487 pasting properties, 636

750

Index

Rice starch (continued) prediction of sorption isotherms, 487–488 WAXD, 485 Rice starch, high-amylose effect of hydrothermal treatment on rheological properties of, 635–639 conclusion, 639 introduction, 635 relationship of apparent viscosity and shear rate of heat-moisture-treated rice starch at 18%, 21%, 4%, and 27% moisture content over shear rate range of 0–300−1 and at 60°C, 637 treated, relationship of pasting temperature and setback to moisture content of, 638 Rice-starch-based film, oxygen permeability values of, 643, 643t Rice starch-glycerol mixture interaction factor and error function for, 487, 487t interaction factors vs. moisture content of, 488, 488 Richards equation boundary, initial conditions and, 227 solving, retention-curve validation and, 231 Roe, K., 101 Roos, J., 100 Roos, Y., 87, 88, 91, 98, 106, 160, 286, 345, 346, 347, 348, 479 Roozen, M. J. G. W., 94, 100 Rosenberg, M., 500, 504 Rossi, C., 228 Rotational diffusion, 419 Rotational motion, 158, 161 liquid state of water and, 205 of water, nuclear magnetic resonance techniques and sensitivity to, 216 Rotronics, 90 Roudaut, G., 105, 106, 121t RR. See Rehydration ratio RSM. See Response surface methodology Ruan, R. R., 61, 259

Rubbery amorphous solid state, food storage and transition from glassy stable amorphous solid state to, 93 Rubbery state free volume of carbohydrate system matrix independent of degree of polymerization in, 361 water diffusivity, glassy state vs., 94 Rupley, J. A., 238 Rusk roll, sorption experiments performed on crusts of, 166–167 Rutnakornpituk, M., 460 RVP. See Relative vapor pressure RWC model. See Relative water content model Ryshkewitch analysis, 312 tensile strength as function of relative humidity at porosity of zero obtained with, 308 Ryshkewitch equation, tensile strength of MCC compacts at zero porosity obtained with, 307 S Saccharin sodium, differential scanning calorimetric thermograms for, 32 Saccharin sodium hydrates, endothermic peaks for, 31 Saccharomyces cerevisiae affinity constants for glucose uptake by, as reported in various studies, 55t cell resistance to thermal and osmotic stress and surface topology of, 77 isoviability diagram of, vs. temperature and water activity, 75 synergism between high pressure and subzero temperature on, 78 variations in average cell volume and cell viability of, after osmotic shock, 73 viability of, vs. water activity by use of two probes: Lucifer yellow and propidium iodide, 73 Saccharomyces cerevisiae cells, glucose and fructose in dough and, 53 Saccharomyces cerevisiae membrane, isoanisotropy of, vs. temperature and water activity, 76

Index

Saenger, W., 153 Safety of food, distribution of microstructural domains and assessment of, 66 Sajama quinoa seeds amplitude of signal of lipid component of, with different water activities, 652 genotype description of, 648 percent germination of, during storage at 32°C and 43% or 75% relative humidity, 650, 650 percent viability of, assessed by tetrazolium test, 651 relaxation times of lipid component of, with different water activities, 652 sorption isotherms of, 651t viability and germination values for, 650 Salinity changes, jellyfish water-holding and texture properties relative to, 388, 388–389 Salt in bread formulation and production, 606 carotene encapsulation, mannitol and, 20 effect of, phase diagram of sugars, 17 sugar crystallization kinetics and enzyme inactivation and, 18 Salt concentration, water retention by superabsorbent polymers and jellyfish as function of, 388 Saltine crackers, loss of crispness in, 105 Salt solutions, saturated, preparation of, 272–273 Sao Hai rice, 663, 664 comparison of experimental and predicted solid loss of milled rice grains from waxy milled rice, jasmine rice and, as function of soaking time at various temperatures, 670 effective diffusion coefficient for, at different temperatures, 669t experimental and predicted moisture contents of, as function of soaking time at various temperatures, 666 Page model parameters for water uptake by, at different temperatures, 668t power-function model parameters for loss of solids from, at different temperatures, 671t

751

regression coefficient and percentage of root mean square error obtained from model fitting using various models, 667t temperature and water uptake of, 667 SAPs. See Superabsorbent polymers Saravacos, G. D., 129 Sarcoplasmic proteins, gel formation and, 241 Saturated hydraulic conductivity, 226, 231 Saturated salt solutions preparation of, 272–273 of varying relative humidity measured at 20° and 32°C, 273t Sauvageot, F., 105 Sawai, H., 635 SB. See Standard bread Scanning electron micrograph images, of surfaces and fracture surfaces of chitosan film, 90 : 10 blend film, 80 : 20 blend film, and 70 : 30 blend film, 462, 462 Scanning electron microscopic images, of agave dried slices, longitudinal and transverse cuts, 624, 624 Scanning electron microscopy crystalline amylopectin lamellae alignment as evidenced by, 253 images of Lactobacillus rhamnosus GG encapsulated in trehalose and lactose-trehalose glass, with data for colony forming units, 289 Scannng electron micrographs, of cassava and potato starches undergoing gelatinization, 255 SCF. See Sugar-frosted cornflakes Schebor, C., 162 Schmidt, A., 91 Schmidt, S. J., 216 Schubert, H., 675 SCM. See Soft condensed matter Scopolamine hydrobromide differential scanning calorimetric thermograms for, 32 endothermic peaks for, 31 water-sorption isotherms for, 33

752

Index

Scopolamine hydrobromide hydrates free-induction decay exhibited by, 30 Gaussian-type decay exhibited by, 29 Scott, W. J., 89 Seafood products, chitosan and storage of, 459 Secondary relaxation processes, glassy state and, 117 SEM. See Scanning electron microscopy Semicrystalline lamellae, starch granule, 254 Semiempirical models liquid transport into dried solid matrix and, 233 rehydration studies and, 220 Semimoist foods, Arrhenius relationship and reactions limiting storage stability of, 98 Semipolar solvents, dielectric constant of, 159 Semisolid foods, solvent and nonsolvent water in, 55 Semisolid phase, in dough, 50 Semisolid viscoelastic phase, in dough, 50 Sensor array-based aroma analysis technology, 658 Sensory crispness, ability of water to plasticize polymer structure related to, 108 Sensory crispness intensity, as function of temperature for wafer cookie at 5% moisture content and intersection representing critical point for onset of crispness loss, 110 SENT specimens. See Single-edge notched tension specimens Seow, C. C., 116, 129 Sephadex G25 gel differential scanning calorimetry rewarming traces obtained with, 375, 375 differential scanning calorimetry rewarming traces obtained with, containing small amount of water, 381 frozen and containing 50% water, twodimensional x-ray diffraction images obtained with, 378

ice crystallization during rewarming observed with, 382 ice crystallization exotherm during rewarming observed with, 374– 375, 377 vitrified water in, 379 Serrano, R., 55t Shalaev, E. Y., 74 Shamblin, S. L., 426 Shamekh, S., 643, 644 Shear-thinning behavior, of rice-starch pastes after heat-moisture treatment, 636 Shelf life of food mechanical and thermal properties and, 583 predicting, temperature dependence of physical and chemical reactions and, 98 texture loss and reduction in, 105 water distribution in foods and, 533 Shiozawa, S., 228 Shittu, T. A., 675 Shogren, R. L., 130 Short-range mobility, crystallization and, 94 Shrinkage isotropic, 616 nonisotropic, 616 studying in vegetables, 613 Shrinkage and deformation of agave slices during convective drying, image processing and fractal analysis used in study of, 620–621 of potato slabs during drying, image analysis in evaluation of, 615 Shrinkage-deformation kinetics, agave samples and, 621, 624 Sigma Scan 5.0, 516 Sigma Stat 2.0 software, 516 Silage, Lactobacillus coryniformis Si3 as additive to, 142 Single-edge notched tension specimens, 291 Sinkability, quality of food powder properties obtained by spray drying and, 674

Index

Skimmed milk, mixtures of, in probiotic cultures, 285 Skim-milk powder effect of water and fat contents on enthalpy of dissolution of, 42 onset of glass transition temperature of food powders measured by TMCT, DSC, and TMA, 434t Slade, L., 90, 98, 158, 212, 470 Slaninova, I., 74 Smooth Cayenne pineapple, 657 freezing and thawing of sample, 658 fresh and freeze-thaw off-aroma scores of, 660t principal component analysis plot for, as determined by e-nose, 659, 660 sample preparation, 658 SMP. See Skim-milk powder SO. See Surface oil Sodium, water content of, at 10% RH and 60% RH, 37t Sodium-caseinate, protein-carbohydrated sheets prepared with, 578 Sodium chloride maximum load of 2.4 mm probe penetrating jellyfish at speed of 0.5 mm/s at increasing additions of, 389, 389–390 in tapioca-flour-based baked products study, 593 Sodium chloride-glucose-water mixture, at 25°C, differences in relative viscosities of ternary solutions with, 474, 475 Sodium ions effects of, on viscosities in sodium/ potassium-glucose-water ternary system, 473–475 potassium ions vs., roles of in living bodies, 473–474 Sodium/potassium-glucose-water ternary system effects of sodium and potassium ions on viscosities in, 473–475 introduction, 473–474 materials and methods, 474 results and discussion, 474–475 Soesanto, T., 344

753

Soft condensed matter background, 88–93 conclusions, 111 examples of use of state diagrams and glass transition curve, 94–108, 111 loss of crispness or hardness, 105–108, 111 stickiness of hard candy, 100–105 sugar recrystallization in storage of foods: sucrose and cotton candy, 95–100 introduction, 87–88 perspective on physics of food states and stability, 87–111 physical state changes in foods in storage, 93–94 Softening, food storage and, 93 Softening of crispy food in storage, moisture pickup and, 95 Soil science, retention curve in, 226 Soles, C. L., 132 Solid active pharmaceutical ingredients, apparent correlations between stability of, and molecular mobility of water, 25–26 Solid foams, expanded, 247–249 Solid food materials: thermal mechanical compression test, 429–436 applications, 433–434, 436 design of sample cell, 431 determining onset of glucose melting temperature based on mechanical behavior measured by TMCT, 433 diagram of sample cell, 430 introduction, 429–430 methodology, 430–433 design and setup of TMCT, 430–431 operation protocol of TMCT, 431–433 onset of melting temperature of sugar crystals and some plastic polymer beads measured by TMCT and DSC, 433 performance of heater controller, 431 standard operation protocols for TMCT, 432 Solid foods, rheological and textural properties of, 119 Solid fraction, of a tablet, defined, 304

754

Index

Solid polymer-diluent blends, idealized and antiplasticized modulus vs. diluent content curve of, 118 Solid state of water, 204 Solid system states, solid food matrix state, 90 Solubility approximate, of some sugars in glycerol at 60°C after an equilibration period of 10 days, 160t quality of food powder properties obtained by spray drying and, 674 water of hydration, API hydrate and, 26 Solubility curve of solute, thermodynamic phase diagram for binary aqueous system and, 214, 215 Soluble solids of cooked rice, analysis of, 665 Soluble starch, antiplasticizing effect of water and, 311 Solution viscosities, influence of sodium ions and potassium ions on, 474 Solvent fractions, 50 Solvent water in dough, 51–52 food freezing and removal of, 338 in semisolid foods, 55 Solvent-water concept in dough, 49, 50–51 frozen dough and, 49 Sommier A., 630 Soottitantawat, A., 500, 502, 504 Sorbitol anhydrous, sucrose crystal dissolving in, at 100°C, 160 dielectric constants of, 159t melting-enthalpy-moisture-contentrelationships of konjac glucomannan films plasticized with glycerol and, 125 as plasticizer, in egg-white film preparation, 454 tensile strength for egg-white films plasticized with 50% and 60%, 456–457 tensile strength of starch-based films at low moisture contents and, 132

viscosities of egg-white solutions after addition of, 455 water-vapor permeability values of duck egg-white film with addition of, 457 Sorption isotherms determining for beta-cyclodextrins, 151 for extruded rice starch with 0%, 5%, 10%, and 20% glycerol, 485–486, 486 modeling for rice-glycerol extrudates, 487 prediction of, for rice starch-glycerol mixtures, 487 for quinoa seed samples, 649–650 Sorption mechanisms, distinguishing between, 166 Sorption moisture studies, for betacyclodextrins, 152–153 Soya mixed systems in, 244 non-denatured, inspection of DSC behavior of, 247 Soy bread “freezable” water in, 181, 181 magnetic resonance proton-intensity signal images of, on days 0 and 6 of storage at 4°C, 179 moisture content and specific loaf volume of almond-soy bread and, 179t regular, formulations of, 177t schematic of changes occurring in, with and without almond during storage, 182 Soy bread with and without almond materials and methods differential scanning calorimetry, 178 preparation of samples, 176–177 specific loaf volume measurement, 177 thermogravimetric analysis, 177–178 results and discussion, 178–183 differential scanning calorimetry, 181–183 loaf volume, 178 moisture loss and migration during storage, 179 water state, thermogravimetric analysis, 180–181

Index

water state and distribution during storage of, 175–183 introduction, 176 Soy protein history behind addition of, to bread, 176 inclusion of, in U.S. diet, 175 protein films developed from, 453 Spaghetti, describing sorption dynamics of, 166 Specific volume, of annealed maltodextrin matrices with given equilibration as function of temperature for various molecular weight distributions, 360 Spectroscopic approach, to metrics for water, 216–217 Spectroscopy, measurement of frequencies and amplitudes of signals with, 216 Spiess, W. E. L., 493 Spin-lattice relaxation, for Na2HPO4 · 12H2O and Na2HPO4 · 2H2O, 27, 28 Spin-spin relaxation measuring rotational mobility of water molecules and, 62 molecular mobility of hydration water in API hydrates determined by, 27 usefulness and limitation of determination of water mobility by NMR and, 36 Spin-spin transverse relaxation time, measuring of, by proton nuclear magnetic resonance, 583, 584 Sponge cake, Fickian diffusion used to describe sorption dynamics of, 166 Spray-dried food powders, onset of glass transition temperature of food powders measured by TMCT, DSC, and TMA, 434, 434t Spray drying adsorption surface changes as function of water activity of sucrose-calcium powders at 25°C by, 686 benefits with, 191 of citrus juice, with maltodextrin as a carrier, 196

755

drying aids and, 192 of fructose mixtures, 195–196 main requirements for, 342–343 microstructural, physical, and rehydration properties of maltodextrin powders obtained by, 673–678 operating conditions in, and rehydration of food powders, 673 quality of food powder properties obtained by, 674 Stability of food, distribution of microstructural domains and assessment of, 66 Standard bread, 605, 607 hardness of, during storage, 608t proton free-induction decay of, fresh and after 7 days of storage, 609, 610 Starch changes in fracture stress and fracture strain for, as function of moisture content, 124 crispiness in tapioca baked samples and, 596 from different sources, different molecular weights for, 641 enzyme stability and, 16 hydrothermal treatment of, common use, 635 ice crystallization during rewarming observed with, 373, 381–382 influence of water behavior in, 251 water mobility in, 256 Starch acid hydrolysis treatment, 264–265, 267 Starch-based food, understanding effects of freezing in, 185–186 Starch-chain mobility, during gelatinization, 260–263 Starch chain mobility T2 distribution, as measured by proton 90° pulse free-induction decay of deuterated cassava starch saturated with water during gelatinization, 262 Starch films and coatings, as good oxygen barriers, 641

756

Index

Starch fraction effect of freezing temperature on time needed to freeze 50% and 100% of available freezable water, and corresponding moisture content of, 189t moisture content in, after freezing at −3, −40, and −50°C, for samples with 30%, 40%, and 50% overall moisture content, 188t Starch-fraction moisture content effect of initial water content and freezing temperature on, after freezing, 187–188 evolution of, in function of freezing time at various freezing temperatures, 189 Starch gelatinization, hydration in, 255 Starch genetics and biosynthesis, drought and, 252 Starch glass transition, phase separation of ice crystals in starch-based systems during freezing and effects on moisture content and, 185–190 Starch granule-associated proteins, defined, 508 Starch granules in dough, 50 understanding nature of, and seasonal changes affecting properties of, 268 water-saturated freezing behavior of, 263–264 gelatinization behavior of, 260 Starch granule structure, redrawn from Gallant and others, 254 Starch growth ring, components of, 253 Starch retrogradation, 119 Starch samples, syneresis and freeze-thaw stability in, 267 Starch state diagram, to show freeze concentration of starch fraction during freezing for system initially at 50% moisture content, 186 Starch-water films, sorbitol added to, 132 Starchy food during cooking method, 534–535 modeling, 534–535 water-holding capacity profile, 535

predicting water migration in, 533–544 conclusion, 538 introduction, 533–534 nomenclature, 538 results and discussion, 535–537 starch gelatinization and conversion of, into multiphase body, 533 STARe software, 584 State change influence of temperature and moisture increase on, 93 key to understanding of, 100 State diagrams of amorphous lactose, 349 generating, 214 low-water food systems and use for, 348–349 molecular mobility approach in water measurement and, 212–216 need for glass transition line in, 111 optimizing efficiency of biomolecular dehydroprotectant agents and development of, 18 showing equilibrium and amorphous regions: sugar mix, cereal dough, soft cookie or caramel, hard-ball candy, cotton candy, or crisp cookie, 92 showing saturated, ambient temperature; eutectic temperature; glass transition temperature; solutespecific glass transition temperature; and melting temperature, 92 state changes and use of, 94–108, 111 loss of crispness or hardness, 105–108, 111 stickiness of hard candy, 100–105 sugar recrystallization in food storage: sucrose and cotton candy, 95–100 temperature-composition, for binary aqueous system, 214 theoretical development of, 91 Steponkus, P. L., 74 Sterilization, effects of various perturbations on, 72 Steroids, water content and solubility of, 330

Index

Stickiness of amorphous food powders, 344–345 of amorphous food solids in dehydration at various water contents and, 343 collapse phenomena in food systems and, 335 food storage and, 93 of hard candy, 100–105 Sticky point temperature, 100 StO. See Structural oil Stoichiometric water uptake of bulk drug active pharmaceutical ingredient as function of percent relative humidity at 25°C, 319 moisture uptake in pharmaceutical solids and, 318 Strain, changes in, at break of tapioca starch, konjac glucomannan, and konjac glucomannan-carboxy methylcellulose films as function of water activity, 123 Structural oil, absorption mechanisms in fried potato cylinders and, 525 Structural relaxation, 161, 419 amorphous materials and, 341 Student’s t-test analysis, for identifying differences between breads produced with different mixers at same storage time, 608 Sub-Tg (or secondary relaxation), brittleductile transition of synthetic amorphous or partially crystalline polymers and, 117 Sub-T9 relaxation, molecular mobility, glass transition temperature and, 161 Sucrose addition of, to freezing solutions, 551 amorphous, crystallization of, 346 approximate solubility of, in glycerol at 60°C after equilibration period of 10 days, 160t cotton candy and: sugar recrystallization in storage of foods, 95–100 effects of, on storage stability of xanthine oxidase, 599, 601, 601t, 602, 603 first-order phase transition (melting) of, by use of common sugar crystals, 433t

757

freeze-dried products and, 142 glass transition line for, 101 protein-carbohydrated sheets prepared with, 578 time to crystallization of, as function of water activity, 96 Sucrose-calcium powders, 681 adsorption surface changes as function of water activity of, at 25°C, 686 cryogenic process, 682 spray drying, 682 Sucrose-cell formulations, water activity of, after freeze-drying, 144, 145 Sucrose crystals density of, 101 dissolving of, in anhydrous sorbitol at 100°C, 160 Sucrose glass transition temperature, as function of mass fraction of water: sucrose solubility curve; and water-melting curve, 17 Sucrose hydrolysis, in mixed doughs, 53 Sucrose recrystallization, raffinose and, 98 Sucrose-salt-water solution partial phase diagram of, with constant weight ratio value of 9.09, 556 partial phase diagram of, with R value of 36.34, 556 Sucrose solution, leakage from large unilamellar vesicles at various concentrations of, that were previously cooled to −40°C and thawed to room temperature, 559 Sucrose-water mixture, ESR use and increase in rotational mobility of spin probes within, 94 Sugar amorphous, crystallization of, 346–348 in bread formulation and production, 606 molecular structures and properties of sugar alcohols and, 477 Sugar alcohols, molecular structures and properties of sugar and, 477 Sugar concentrations in dough, expressions of, 49 Sugar content, carrot fiber powder size and, 194

758

Index

Sugar crystallization, enzyme stability and, 16 Sugar crystals, first-order phase transition (melting) of crystalline food solids by use of, 433t Sugar extraction, dough manufacture and analysis of, 52 Sugar-free confectionary products, with lactitol, moisture content and Tg as useful tools in predicting stability of, 281 Sugar-free products, market demand for, 272 Sugar-frosted cornflakes differential scanning calorimetry thermograms of, at 7.5% and 5.2% water content, 585, 585 Hahn spin-echo transverse relaxation time values vs. water content for, 586– 587, 587 relationship between relaxation time constant determined by freeinduction decay analysis and temperature for, 586, 586 water content for, 583, 584 Sugar glasses, glycerol as antiplasticizer and enhancing bioprotective properties of, 132 Sugar recrystallization in storage of foods, sucrose and cotton candy, 95–100 Sugars acting as antiplasticizers in variety of biopolymer systems, 131 crystallization of, food properties and, 271 Maillard reaction and, 11 protecting enzymes during drying and storage with, 16 Sugar-salt-water systems, concentrated, measurements of electrical conductivity in, 17 Sugar snap cookies brittle-ductile transition study on, 107 brittle-ductile transition temperature plot as function of moisture at constant temperature for, 108 Sugar-water mixtures, thermal, mechanical, and dynamic properties of, 465

Sulfated polysaccharides, in jellyfish, 392 Sulpyrine differential scanning calorimetric thermograms for, 32 endothermic peak for, 31 Sun, W. Q., 101 Sunflower oil, in bread formulation and production, 606 Superabsorbent gel, transmission electron microscopic image of freezeetched replica of, 393 Superabsorbent polymers acrylate-based, drawbacks with, 386 transmission electron microscopic image of freeze-etched replica of, 393 water retention by jellyfish and, as function of salt concentration, 388 Superabsorbents developmental research on, 385–386 water uptake reduction in, by addition of salt, 387 Supercooled liquids glass transition for, 336 spatially heterogeneous dynamics in, 18 Surface acoustic wave sensor, aroma analysis and, 658 Surface integration, limited crystal growth and, 272 Surface oil, absorption mechanisms in fried potato cylinders and, 526 Surface tension, 225 Surfactants, water interactions and, 157 Suwonischon, T., 121t Sweet potato, volume decrease or shrinkage studied in, 613 T Tablet making environmental moisture and, 302 mechanical strength of, 301 Tablets, as special type of dispersion, 304 Takahashi, K., 78 TAM. See Thermal activity monitor Tang, H. R., 258 Tapioca flour, uses for, in various food products, 591

Index

Tapioca-flour-based baked products glassy, relationships between water activity, texture properties, glass transition , and moisture content of, 595, 595–596 glassy, texture of, as function of moisture content, 591–596 introduction, 591–592 materials and methods, 592–593 dynamic differential scanning calorimetry, 593 mechanical measurement, 593 proximate analysis, 592 sample preparation, 592 sorption isotherm, 593 texture assessment: crispness, hardness, and work sensory evaluation, 593 results and discussion, 594–596 adsorption isotherm, 594–595 food properties, 594 relationship among texture, water activity, and glass transition, 595–596 Tapioca starch anomalous antiplasticization effects of water in relation to some mechanical properties demonstrated in, 122 changes in tensile strength and strain at break of, as function of water activity, 123 oxygen permeability of, 641 Tapioca-starch-based film oxygen permeability values of, 643, 643t tensile modulus of, at different water activity as function of glycerol content, 131, 131 Taylor, L. S., 318, 321 Tea leaves, capillary-flow approach for modeling temperature and anisotropic effects during rehydration of, 231–232 TEG. See Terminal extent of gelatinization TEM. See Transmission electron microscopy Temperature bread-dough study, 628 brittle fracture and, 106

759

critical state for powders during storage and, 100–101 defining/predicting kinetic coefficients of desirable and undesirable changes in foods, 9–10 extent of freeze concentration and, 338 influence of, on state change, 93 isoanisotropy of Saccharomyces cerevisiae membrane, vs. water activity and, 76 isoviability diagram of Saccharomyces cerevisiae vs. water activity and, 75 low combination of high hydrostatic pressure and, on Escherichia coli survival, 77–79 effect of hyperosmotic perturbation and, on Escherichia coli baroresistance, 79–80 Maillard reaction rate, PVP systems and, 11 solvent water in dough and, 51 stability of food matrix and, 59 storage, glass transition temperature of lactitol, and predicting crystallization at different levels of, 278, 280–281 storage longevity of seeds and, 647 thermodynamic and kinetic properties influenced by, 208 transition, of chocolate wafers as function of moisture content, 110 vapor pressure and, 210 water uptake of rice grains and, 667 Temperature-concentration diagrams, 213 Temperature dependence of chemical reactions, Arrhenius relationship and, 98 of partial pressure of water in system and partial pressure above pure water at same temperature, 210 Temperature depression nonpermeable cryoprotection agents and effect on, 555–557 permeable cryoprotection agents and effect on, 557

760

Index

Temperature kinetics for bread dough, 630 inside dough during baking at three different locations, 628, 629 TEMPO dissolving of, in dough matrix, reaction kinetics and, 64–66 NMR and ESR experiments with wheat flour dough and, 60–61 Tensile strength changes in, at break of tapioca starch, konjac glucomannan, and konjac glucomannan-carboxy methylcellulose films as function of water activity, 123 for egg-white films plasticized with 50% and 60% sorbitol, 456–457 highest value of, observed for gelatin-gum arabic-sucrose sheet as compared with other sheets, 580 relationship between compaction pressure and, in microcrystalline cellulose, 309 Tensile strength of microcrystalline cellulose compacts, effect of relative humidity on, at constant porosity of 0.3 and 0.1, 307 Tension head, 225 parametric models for relating water content to, 226, 227 water activity converted to, by Kelvin equation, 229 Tension head and capillarity, flow in unsaturated porous media and, 224–225 Terminal extent of gelatinization, 535 Testosterone equilibrium solubility of, in triglyceride oils that were either desiccated or water saturated, 331 water content and solubility of, 330, 332 Tetrazolium test for viability, for quinoa seeds, 649 Texture, product quality and loss of, 105 Texture Expert Exceed software, 303 Texture of food, water distribution in foods and, 533

Texture properties mechanical and thermal properties and, 583 of raw and roasted coffee, water plasticization and antiplasticization effects and, 494 relationship among moisture content, water activity, glass transition and, for glassy tapioca-flour-based baked product, 595, 595–596 small changes in moisture and, 483 TGA. See Thermogravimetric analysis Thailand Research Fund, 639 Thermal activity monitor, water uptake and mobility on pharmaceutical stability in amorphous solids and measurements by, 325 Thermal expansion coefficient, free volume available within glassy system and, 94 Thermal mechanical analysis, 429, 430 onset of glass transition temperature of food powders measured by, 434t Thermal mechanical compression test accuracy and robustness of, in analysis of solid food samples, 436 development of, 430 as ideal tool for measuring phase transition of solid food materials, 433 onset of glass transition temperature of food powders measured by, 434t Thermal mechanical compression test: solid food materials, 429–436 applications, 433–434, 436 introduction, 429–430 methodology, 430–433 design and setup of TMCT, 430–431 operation protocol of TMCT, 431–433 Thermal plasticization, effect of, on mechanical, flow, and collapse phenomena on relaxation times and time-dependent changes in food solids, 340 Thermal properties of food polymer-water blend, 124–126 shelf-life stability and textural properties of food products and, 583

Index

Thermal stress protein stability and, 16 yeast survival and effect of combined osmotic stress and, 75–77 Thermodynamic approach, to metrics for water, 208–211 Thermodynamic behavior of soluble food powders, demonstration of, 41–46 Thermodynamic phase diagram, for binary aqueous system, 214, 215 Thermodynamic properties of water, critical importance of, in survival of microorganisms, 82–83 Thermodynamics, of water sorption, 128–129 Thermogravimetric analysis findings related to characterization of water fraction of soy bread with, 175 measuring distribution of water in bread with, 176 soy bread, typical weight-loss firstderivative curve from, 180, 180 Thirathumthavorn, D., 642 Thixotropic behavior, of high-amylose rice starch, 635, 636, 639 Three-point beam-bending method, mechanical properties of microcrystalline cellulose tablets measured with, 303 Three-point-bending test, 120, 126 Thymol release of, during storage, 154–155, 155 sorption characteristics of betacyclodextrins formed with, and their release, 149 storage study on, 151 in thyme essential oil, 150 water sorption and stability of betacyclodextrin complexes with, 155 Time domain (td) proton nuclear magnetic resonance combining classic moisture-sorption isotherms with, 411 microdomain distribution in food matrices and measured values ascertained with, 60

761

Time-resolved NMR spectroscopy, studying water status and metabolic changes in biological systems with, 647–648 Timmermann, E. O., 14, 439 TMA. See Thermal mechanical analysis TMCT. See Thermal mechanical compression test To, E. T., 345 Tolstoguzov, V., 245 Tomato powder hydration experiments on, 414 water-sorption isotherms and dependence of relative mobile tdNMR signal on hydration for, 417 water-sorption isotherms for, 415–416, 417 Total oil (TO), kinetics of, in control potato slices during frying at 180°C, 527, 528 Transition temperatures, of chocolate wafers as function of moisture content, 110 Translational diffusion, 419 Translational motion, 94, 158, 161 glass transition temperature, collapse phenomena and, 336 liquid state of water and, 205 of water, nuclear magnetic resonance techniques and sensitivity to, 216 of water molecules in pharmaceutical glasses, 322 Transmission electron micrographs, of cassava and potato starches undergoing gelatinization, 255 Transmission electron microscopic images of freeze-etched replica of superabsorbent polymer, superabsorbent gel, gelatin gel, and jellyfish network, 393 showing area of mesoglea consisting of open network structure composed of polysaccharides and proteins, 392, 392 Transmission electron microscopy crystalline amylopectin lamellae alignment as evidenced by, 253 microstructure of jellyfish characterized by, 385, 391

762

Index

Transparent fine-stranded gel, at pH 7.5, 240 Transport phenomena in foods, Fickian approach in modeling of, 222 Trehalose, 142 addition of, to freezing solutions, 551 amorphous, mean radius of holes in, as function of water content at 25°C, 356, 358 anhydrobiosis and, 132 enzyme stability and, 18 glass transition of maximally freezeconcentrated solutes and onset temperature of ice melting of, 287t protecting enzymes during drying and storage with, 16 protein-carbohydrated sheets prepared with, 578 raising Tg of, inhibiting rate of crystallization of lactose and, 104 scanning electron microscopy images of Lactobacillus rhamnosus GG encapsulated in lactose-trehalose glass and, with data for colony forming units, 289 Trehalose-glucose-lysine system glass transition and enthalpy relaxation results for, 447 glassy, enthalpy relaxation time of, 447–448 time courses of ABS280 for, at given temperatures, 448, 449 Trehalose hydrolysis, interactions of, with magnesium chloride and/or with water as role in browning inhibition, 541 Trehalose matrix intermolecular hydrogen bonds and, 575, 576 model of local reaction promoting nonenzymatic browning reaction progress in glassy matrix and, 574, 575 Trehalose-salt-water solution, partial phase diagram of, with constant weight ratio value of 9.09, 556

Trehalose solution, leakage from large unilamellar vesicles at various concentrations of, that were previously cooled to −40°C and thawed to room temperature, 559 TRF. See Thailand Research Fund Tricaprylin C=O oxygen atoms, radial distribution function for water oxygen with monocaprylin oxygen atoms and, 331 Tricaprylin-monocaprin, water uptake in, at 37°C and 100% relative humidity as function of molar concentration of monoglyceride component, 328, 328 Tricaprylin-monocaprylin, water uptake in, at 37°C and 100% relative humidity as function of molar concentration of monoglyceride component, 328, 328 Triglyceride-monoglyceride-water mixtures freeze-fracture electron microscopy study of organization of lipid molecules in, 329 self-association of water at higher moisture levels observed in, 332 Triglycerides, water uptake, distribution, and effects on drug solubility in lipid vehicles composed of, 326–330, 332 True density of microcrystalline cellulose powder, measuring, 304 TS. See Tensile strength T2 distribution of values of probability density function for, in microsecond range, for 30% moisture dough at different temperatures, 63 distribution of values of probability density function for, in millisecond range for 30% moisture dough at different temperatures, 63 T2 line, state diagram and equilibrium curves combined with, 91 T2 values, water mobility in dough samples and comparison of, 62–63 Twin-screw extruder, protein-carbohydrate sheets prepared with, 577, 579

Index

Two-term exponential model predicting moisture rate of rice grains with, 664t rice varieties and, regression coefficient and percentage of root mean square error obtained from model fitting with, 667t 2-Want quinoa seeds amplitude of signal of lipid component of, with different water activities, 652 genotype description of, 648 percent germination of, during storage at 32°C and 43% or 75% relative humidity, 650, 650 percent viability of, assessed by tetrazolium test, 651 relaxation times of lipid component of, with different water activities, 652 sorption isotherms of, 651t viability and germination values for, 650 U Ubbelohde viscometer, 474 Unfreezable water, 115, 119, 127 Unfreezable water fraction, 51 Unfrozen water, physical state of, in polymer gels, 374 Uniform-penalty inversion program, coffee samples analysis and, 494 Uniform-penalty inversion T2 relaxograms, of raw coffee samples rehydrated at different water contents, 496, 496–497 Unsaturated porous media flow, 224–232 capillarity and tension head, 224–225 developing theory of flow in porous media, 228–231 prediction and validation of retention curve from water-activity data, 230–231 retention curve, 230 retention-curve validation, 231 water-sorption isotherm, 229 rehydration of foods using porous media, 231–232 Richards equation, boundary, and initial conditions, 227–228 water-retention curve, 225–227

763

UPEN. See Uniform-penalty inversion program Urbani, R., 424 Urea, 130 UTM dynamometer model 4301, 493 V Vacuum oven drying, measuring amount of water and, 207 van Bruggen, E. F. J., 253 van der Waals forces, 158 van der Waals interactions, solvent-water concept and, 50 van Genuchten model, for relating water content to tension head, 226, 227 Van Hecke, E., 121t van Vliet, T., 242 Vapor pressure measuring, 209 solvent-water concept and, 50 temperature and, 210 Vapor state of water, 206 Vegetable foods, complexity of, 613 Vegetable powders hydration experiments on, 414 main (dry-mass based) components of, used for hydration experiments vs. fraction of the matrix component mobilized during hydration, 417t water-sorption isotherms and dependence of relative mobile tdNMR signal on hydration for, 417 water-sorption isotherms for, 415 Vegetable tissues, understanding structural changes occurring during drying of, 617 Vertical liquid penetration, studies of, 232 Vibrational motion, 158 liquid state of water and, 205 Vickers, Z. M., 105 Viscosity changes in stickiness and caking in amorphous food materials and, 344 structural relaxations around and above glass transition and, 341

764

Index

Viscosity (continued) of film-forming ability of duck egg-white separated from shell eggs, 456t of ternary solutions, with sodium chlorideglucose-water and potassium chloride-glucose-water at 25°C, 474, 475 Vitamins, cyclodextrins and solubility, stability, and bioavailability of, 150 Vitrification, of protective sugars, 16 Vitrified water, in frozen G25 gel, 379 Vogel-Tamman-Fulcher model, relaxation times above glass transition modeled with, 339 Volume decrease, studying in vegetables, 613 Volume relaxation, 419 Vuataz, G., 346 W Wafer cookie sensory crispness intensity as function of temperature for, at 5% moisture content and intersection point representing critical point for onset of crispness loss, 110 stress/strain plot for, at 0.05% relative humidity and 23°C, 109 Walstra, P., 242 Washburn equation, capillary flow in porous media and use of, 222–223 Washburn-Rideal equation, rehydration modeling and, 223 Water as antiplasticizer on mechanical properties in complex food products, 121t biopolymer interaction with, 291 bound, 115, 127 in bread formulation and production, 606 dielectric constants of, 159t external mechanical antiplasticization by, 116 free, 115, 128 glass transition temperature of glassy polymer and, 401–409 glycerol and interactive plasticizingantiplasticizing effects of, 130–132

heat gelation and, 239 heterogeneous nature of, in various starch domains, 268 loss of sensory brittleness of foods and plasticizing nature of, 108, 111 measurement of influence of, 203 monomolecular layer, 159 in most biological systems, 157 as most common plasticizer in pharmaceutical and food sciences, 401 nonsolvent, 115, 127 nonwater diluents as antiplasticizers at T less than Tg, 130–132 phase-separated, in soy bread with and without almond during storage and, 182 as plasticizer of choice, in preparation of compacts, 302 properties of, 126–129 properties of, as diluent thermodynamics of water sorption, 128–129 water diffusivity, 129 properties of food polymer-water blend, 119–120, 122–129 gas transport properties, 126 mechanical properties, 119–120, 122–124 thermal properties, 124–126 role of, in food drying process, 673 salts and tetrahedral hydrogen-bond network of, 18 as solvent of life, 203 soy-almond bread and plasticizing effect of, 183 structure of carbohydrate classes and effects of, 356–357, 359 as ubiquitous diluent, as antiplasticizer at T less than Tg, 119–132 ubiquity of, in food systems, 203 unfreezable, 115, 119, 127 Water activity, 203 effect of, on n values for release of encapsulated D-limonene and L-menthol from spray-dried powder at 50°C, 503

Index

effect of, on release-rate constant of encapsulated D-limonene and L-menthol at 50°C, 503 enthalpy dissolution of food powder samples tested as function of, 44 enthalpy relaxation as function of, 424–425 isoanisotropy of Saccharomyces cerevisiae membrane, vs. temperature and, 76 isoviability diagram of Saccharomyces cerevisiae vs. temperature and, 75 measured, food stability and general rules with respect to, 89–90 relationship among moisture content, texture properties, glass transition and, for glassy tapioca-flour-based baked product, 595, 595–596 structure of beta-cyclodextrins and, 150 of sucrose-cell formulations after freezedrying, 144, 145 tensile strength and strain changes at break of tapioca starch, konjac glucomannan, and konjac glucomannan-carboxy methylcellulose films as function of, 123 time to crystallization of sucrose as function of, 96 typical dissolution calorimetry curves of food powder samples containing 14.3% fat equilibrated at various levels of, 45 Water-activity data, prediction and validation of retention curve from, 230–231 Water activity stability diagram, of relative relation rate vs. aw, 89 Water-activity theory, 157 on stability of biological materials, 158–159 Water content for biopolymer films, prediction of dependence of fracture characteristics on: brittle-ductile transition, 295–296 changes in crystallinity of microcrystalline cellulose as function of, 313

765

effect of, on physicochemical and structural properties of biopolymer films, 292–293 enthalpy of dissolution of all food powder samples tested as function of fat contents and, 46 Hahn spin-echo transverse relaxation time vs., for commercial cornflakes and sugar-frosted cornflakes, 586–587, 587 hydrate formation and effects of, on drug solubility in lipid vehicles, 330 illustration of typical behavior of the Sw/Sm as function of, 413, 413 of lactose, maltose, lactitol, and maltitol during storage at 25°C at various relative vapor pressures, 480, 480–481 low, complex lipid phase behavior and, 77 moisture-sorption isotherm and plotting of, against relative vapor pressure, 211 pattern of seed aging and, 647 permeability values of water through Eudragit E100 as function of, 405 physical properties of potato chips and, 525–530 of protein-carbohydrate sheets, 578 fresh sheet and equilibrated sheet, 580 residual, 226 thermal transitions, mechanical properties, and molecular mobility in corn flakes as affected by, 583–588 various levels of, glass transition, stickiness, and collapse properties of amorphous food solids in dehydration at, 343 Water content of systems, defining/ predicting kinetic coefficients of desirable and undesirable changes in foods and, 9–10 Water crystallization, degree of, freeze-dried products and, 142 Water-diffusion process, in rice grains during soaking, 665, 668

766

Index

Water diffusivity, 129 effective, comparison of, for granular hydrates and gelatinized amioca and hylon 7, obtained from dryingcurve data at 60°C, dry basis, 130 rubbery vs. glassy state, 94 Water distribution in food biopolymers, 251 NMR relaxation and, 256 Water-excipient mixtures, ratio of highmobility water to low-mobility water in, 37t Water-holding capacity defined, 534 for proteins, 242–243 of starchy food, extent of starch gelatinization and, 535 Water-holding capacity profiles for calculation of transient water-content profile in slab of wheat-flour dough during boiling, linear, convex, and concave schematics, 535, 535 in starchy food during cooking, 533 with three bending points, in slab of wheat-flour dough, for boiling time and dry basis, 537, 537 Water-imbibition theory, applicability of, in modeling rehydration of dried porous food, 219, 224 Waterless environment, color development due to nonenzymatic Maillard browning in, at 100°, 115°, and 130°C, 163 Waterless system waterlike functions of other biological compounds in, 157–163 conclusions, 162–163 glass transition, molecular mobility, and sub-T9 relaxation, 161 introduction, 157–158 molecular mobility and the Maillard reaction, 161–162 understanding molecular mobility, 158 water-activity theory on stability of biological materials, 158–159 waterlike solvation property of polyols, 159–161

Water loss, potato chips and modeling of, 527 Water measurement, 203–217 conclusion, 217 introduction, 203–204 metrics for water, 206–217 amount of water, 207–208 molecular mobility approach, 212–217 sorption isotherms, 211–212 thermodynamic approach, 208–211 states of water, 204–206 liquid state, 205–206 solid state, 204 vapor state, 206 Water mobility chemical reactivity in low-moisture and intermediate-moisture systems and, 10, 21 correlation of, as determined by NMR with that as determined by DSC and water-sorption isotherm measurements, 31, 34–36 decoupling from mobility of carbohydrate molecules in approach to glass transition, technological implications of, 366–369 Maillard reaction rate and analysis of, 15 in model dough systems, 59 quinoa seed longevity and, 647–654 in simulated PVP glasses, 322–325 in solid pharmaceuticals, 25–38 molecular, coexisting with various pharmaceutical excipients, 36 molecular, in API hydrates, 26–36 usefulness and limitation of determination of, by NMR, 36 Water mobility (proton spin-spin relaxation time T2) microdomain distribution in food matrices: model dough system and, 62–64 microdomain distribution in food matrices and, 59 Water molecules, as universal solvent supporting all living systems, 251 Water partitioning in colloidal systems as determined by nuclear magnetic resonance, 251–268 acid hydrolysis, 264–265, 267

Index

amylose, amylopectin, and more, 252–253 cassava starch, 258 conclusions, 268 discriminating NMR T2 distribution, 257–258 freezing behavior of water-saturated starch granules, 263–264 gelatinization behavior of watersaturated starch granules, 260 hydration, 255 introduction, 251–252 NMR, 258–260, 259 nondiscriminating properties, 256 starch, 252 starch-chain mobility during gelatinization, 260–263 syneresis and freeze-thaw stability, 267 water distribution, 256–257 Water peak, 207 Water plasticization, effect of, on mechanical, flow, and collapse phenomena on relaxation times and time-dependent changes in food solids, 340 Water-proton relaxation times, change in, during compression of cooked beef, 244 Water protons, spin-spin relaxation time of API protons and, 27 Water-related effect of combined physical stresses on cells, 71–83 combination of high hydrostatic pressure and low temperature on Escherichia coli survival, 77–79 conclusions, 82–83 based on combination of high pressure, temperature, and osmosis, 82 effects of combined high hydrostatic pressure, low temperature, and hyperosmotic perturbations, 77–82 effects of combined hyperosmotic and temperature perturbation: example 1, 72–77 combined osmotic and thermal stress, 75–77

767

conclusions on combined osmosis and temperature, 77 hyperosmotic perturbation of yeast cells, 72, 74 effects of low temperature and hyperosmotic perturbation on Escherichia coli baroresistance, 79–80 introduction, 71–72 parallel changes with pressure and temperature of protein behavior, microbial inactivation, and water structure, 80–82 Water relationships, food polymer science approach to study of, 116 Water-replacement hypothesis, 355 Water retention determining, traditional method of, 226 by superabsorbent polymers and jellyfish, as function of salt concentration, 388 Water-retention curve, porous media and, 225–227 Water-retention data, typical, for microcrystalline cellulose, 230 Water-saturated oil phase, 208 Water-saturated starch granules freezing behavior of, 263–264 gelatinization behavior of, 260 Water sorption stability of beta-cyclodextrin complexes with thymol and cinnamaldehyde and, 155 thermodynamics of, 128–129 Water sorption and transport in dry, crispy bread crust, 165–173 conclusion, 173 introduction, 165–166 materials and methods, 166–167 results and discussion, 167–173 Water-sorption behaviors, in pharmaceutical excipients, 36 Water-sorption data, typical, for microcrystalline cellulose, 230 Water-sorption isotherm data, 25

768

Index

Water-sorption isotherm measurements correlation of water mobility as determined by NMR with that as determined by DSC and, 31, 34–36 ease of evaporation for hydration as determined by differential scanning calorimetry and, 31 Water-sorption isotherms, 256 for active pharmaceutical ingredient hydrates, 33 of beta-cyclodextrin and betacyclodextrin-cinnamaldehyde and beta-cyclodextrin-thymol complexes, 152 for cereal cracker sample, 416 developing theory of flow in porous media and, 229 enabling better understanding of, food hydration behavior and, 417–418 for microcrystalline cellulose, 229 for porous cracker sample, 415 for raw and roasted coffee, 494 for two vegetable powders displaying distinct differences, 415–416, 417 Water-sorption isotherms of food materials molecular mobility interpretation of, by means of gravimetric NMR, 411–418 conclusions, 417–418 discussion, 416–417 hydration experiments, 414 introduction, 411–412 results, 415–416 theory, 412–414 T2 analysis, 414–415 Water’s role in nonaqueous pharmaceutical systems, 315–332 conclusions, 332 implications of water uptake and mobility on pharmaceutical stability in amorphous solids, 325–326 introduction, 315–317 molecular dynamics simulations, 317–318 uptake, distribution, and effects on drug solubility in lipid vehicles composed of triglycerides and monoglycerides, 326–330, 332

water uptake and its implications in amorphous glass, 318–326 water mobility in simulated PVP glasses, 322–325 water uptake and its distribution in PVP, 318–322 Water state and mobility mechanical properties of coffee beans at different hydration degrees related to, 491, 497 in soy-containing breads, for improved dough handling and loaf homogeneity, 180 Water structure, parallel changes with pressure and temperature of protein behavior, microbial inactivation, and, 80–82 Water substitution, 599, 600 Water uptake, of rice grains during soaking, 665, 667–668 Water-vapor permeability, of chitosan and blend films, 462–463, 463t Water-vapor permeability values of duck egg-white film, effect of increased storage time and sorbitol content on, 457 for egg-albumen films, 453 WAXD. See Wide-angle X-ray diffraction Waxes, in fibers, 192 Waxy milled rice, 663, 664 comparison of experimental and predicted solid loss of milled rice grains from jasmine, Sao Hai and, as function of soaking time at various temperatures, 670 effective diffusion coefficient for, at different temperatures, 669t experimental and predicted moisture contents of, as function of soaking time at various temperatures, 666 Page model parameters for water uptake by, at different temperatures, 668t power-function model parameters for loss of solids from, at different temperatures, 671t

Index

regression coefficient and percentage of root mean square error obtained from model fitting using various models, 667t temperature and water uptake of, 667 Waxy rice starch, five freeze-thaw cycles, T2 distribution of protons in, 267, 268 Webb, S. W., 229 Webb, T., 51 Weibull distribution function curve fitting of rehydration data and, 221t liquid transport into dried solid matrix and, 233 rehydration studies and, 220 Weight determination, for amount of water, 207 Weller, C. L., 453 Wettability of maltodextrin powders, determining, 675 maltodextrin powders evaluation, at different feed concentrations and drying temperatures in production of powders, 677t quality of food powder properties obtained by spray drying and, 674 Wetting process, food powder samples and exothermic nature of, 45 Wharton, John C., 95 Wheat flour extruded, light micrograph of, showing starch as continuous structure with protein particles included, 249 sorption kinetics of, 166 tapioca flour as partial substitute for, 591 Wheat flour crackers, hydration experiments on, 414 Wheat-flour dough during boiling relative water content model applied to, 534 water-holding capacity profiles used for calculation of transient watercontent profile in, 535, 535 water-content profile in slab of, calculated using water-holding capacity profiles in, 536, 536–537 water-holding capacity profile with three bending points in, for boiling time and dry basis, 537, 537

769

Wheat gluten anomalous antiplasticization effects of water in relation to some mechanical properties demonstrated in, 122 protein films developed from, 453 Wheat growth rings, 253 Wheat semolina, in bread formulation and production, 606 Wheat starch, five freeze-thaw cycles, T2 distribution of protons in, 267, 268 Whey, mixed systems in, 244 Whey powder comparison of glass transition temperature measured by TMCT, DMTA/ TMA, and DSC for, 434 onset of glass transition temperature of food powders measured by TMCT, DSC, and TMA, 434t Whey-protein gel mixture light-microscopic images of, at pH 5.4, induced by temperature and highpressure processing, 245 moisture loss as function of temperature and salt concentration for, 243 White-bread structural evolution confirmation of three stages of baking in, 627, 632 image-processing results: mean cell area, cell density vs. dough center temperature, and MCA vs. mass loss, 632, 633 materials and methods, 628–630 baking experiments, 628 dough preparation, 628 image processing, 629–630 ingredients, 628 mass-loss kinetics, 629 real-time bread-height measurements, 629 temperatures measurement, 628 results and discussion, 630–632 changes in crumb during baking, 631–632 image processing, 632 mass-loss kinetics, 630–631 temperature kinetics, 630

770

Index

White-bread structural evolution (continued) study of, by means of image analysis and associated thermal history and water-loss kinetics, 627–634 conclusions, 632 introduction, 627–628 Whitney, W. R., 41 Whorton, C., 499 Wide-angle X-ray diffraction, starch conversion assessed with, 484 Williams, M. C., 344 Williams, M. L., 98 Williams-Landel-Ferry equation, 98, 111 Williams-Landel-Ferry plot, modified, for lactose crystallization, 99 Williams-Landel-Ferry power law, relaxation times above glass transition modeled with, 339 Winger, R. J., 52 WLF power law. See Williams-Landel-Ferry power law Wolf, W. R., 493 Wollny, M., 675 WVP values. See Water-vapor permeability values X Xanthine oxidase freeze-dried, effects of excipients on storage stability of, 599–603 discussion, 602–603 introduction, 599–600 results, 601–602 freeze-dried, illustration of, in different glassy matrices: sucrose, sucrose + bovine serum albumin, and sucrose + dextran, 603, 603 materials and methods, 600–601 assay of xanthine oxidase activity, 601 determination of water content, 601 differential scanning calorimetry, 601 preparation of freeze-dried xanthine oxidase samples, 600

remaining activity of, with various excipients after preparation and subsequent storage at 25°C, 602, 602 values for glass transition temperature and moisture contents of, with various excipients, 601t Xiang, T-X, 318 XOD. See Xanthine oxidase X-ray diffraction apparent crystallinity index of microcrystalline cellulose, obtained from, 310, 310 freeze-thaw behavior of aqueous glucose solutions within concentration range of 10%–60% wt/wt studied by, 466 X-ray diffraction patterns for extruded rice starch, with 0%, 5%, 10%, and 20% glycerol, 485, 486 of model food powders, 43 X-ray diffraction powder patterns for cotton candy stored for 2h at 45% relative humidity and 23°C, 97 for fresh cotton candy, 96, 97 for lactose, maltose, lactitol, and maltitol at end of storage at 25°C at relative vapor pressures of 54%, 76%, 66%, and 54% respectively, 481, 481 X-ray diffraction studies, on amorphous sucrose, 95 X-ray microtomography (MCT), microstructure of porous foods observed with, 234 X-ray scattering in combination with PALS, for insight into structure of carbohydratewater system, 359 investigating dynamic properties close to glass transition and, 364 XRD-DSC measurements, endothermic trend due to ice melting, verifying assumption with, 378

Index

Xylose embedding of, in glassy trehalose matrix, 572 intermolecular hydrogen bonds, NEB progress in the glassy matrix and, 576 Y Yang, P., 238 Yeast, in bread formulation and production, 606 Yeast activity, shelf-life of frozen dough and retention of, 49 Yeast cells in dough, 50 sequence of hyperosmotic perturbation of, 72, 74 Yeast survival combined osmotic and thermal stresses on, 75–77 osmosis and temperature combined for improvement in, 77 Yielding, brittle fracture vs., 106

771

Yogurt integral entropies of, 683, 684 lyophilized, water-sorption date for, 683 Yoshii, H., 500, 502 Yoshioka, S., 325 Young’s modulus, 124, 126 fracture response of food material and, 107 highest value of, observed for gelatin-gum arabic-sucrose sheet as compared with other sheets, 580 microcrystalline cellulose compact and, 303 Z Zein film, sorption dynamics of, 166 Zeolite clinoptilolite, integral entropies of, 683, 684 Zeolite Valfor, integral entropies of, 683, 684 Zeolite Valfor 100, in paprika-containing alginic acid capsules, 681 Zografi, G., 101, 426 Zygosaccharomyces bailii, baroprotective effect of solutes and, 79