Biomedical applications of electroactive polymer actuators

Biomedical Applications of Electroactive Polymer Actuators Biomedical Applications of Electroactive Polymer Actuators E...

Author: Federico Carpi | Elisabeth Smela

19 downloads 966 Views 26MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

Biomedical Applications of Electroactive Polymer Actuators

Biomedical Applications of Electroactive Polymer Actuators Edited by Federico Carpi and Elisabeth Smela © 2009 John Wiley & Sons Ltd. ISBN: 978-0-470-77305-5

Biomedical Applications of Electroactive Polymer Actuators FEDERICO CARPI University of Pisa, Pisa, Italy ELISABETH SMELA University of Maryland, College Park, USA

A John Wiley and Sons, Ltd., Publication

This edition first published 2009 Ó 2009 John Wiley & Sons Ltd. Registered office John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, United Kingdom. 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. The right of the author to be identified as the author of this work has been asserted in accordance with the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs and Patents Act 1988, without the prior permission of the publisher. Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic books. 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. 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 any implied warranties of fitness for a particular purpose. This work is sold with the understanding that the publisher is not engaged in rendering professional services. The advice and strategies contained herein may not be suitable for every situation. In view of ongoing research, equipment modifications, changes in governmental regulations, and the constant flow of information relating to the use of experimental reagents, equipment, and devices, the reader is urged to review and evaluate the information provided in the package insert or instructions for each chemical, piece of equipment, reagent, or device for, among other things, any changes in the instructions or indication of usage and for added warnings and precautions. 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. No warranty may be created or extended by any promotional statements for this work. Neither the publisher nor the author shall be liable for any damages arising herefrom. Library of Congress Cataloging-in-Publication Data Biomedical applications of electroactive polymer actuators / [edited by] Federico Carpi, Elisabeth Smela. p. ; cm. Includes bibliographical references and index. ISBN 978-0-470-77305-5 (H/B) 1. Polymers in medicine. 2. Conducting polymers 3. Actuators. I. Carpi, Federico, 1975– II. Smela, Elisabeth. [DNLM: 1. Polymers—diagnostic use. 2. Polymers—therapeutic use. 3. Biomedical Technology. 4. Equipment and Supplies. QT 37.5.P7 B6147 2009] R857.P6B5485 2009 610.280 4—dc22 2008056029 A catalogue record for this book is available from the British Library. ISBN: 978-0-470-77305-5 (H/B)

Set in 10/12pt Times by Integra Software Services Pvt. Ltd, Pondicherry, India. Printed and bound in Great Britain by CPI Antony Rowe Ltd, Chippenham, Wiltshire.

Contents

Preface List of Contributors Introduction SECTION I 1

page xv xvii 1

POLYMER GELS

5

Polymer Gel Actuators: Fundamentals Paul Calvert

7

1.1 1.2

1.3 1.4

1.5 1.6

1.7

Introduction and Historical Overview Properties of Gels 1.2.1 Biological Gels 1.2.2 Mechanical Properties of Simple, Single-Phase Gels 1.2.3 Elastic Moduli 1.2.4 Strength 1.2.5 Multi-Phase Gels 1.2.6 Double Network Gels 1.2.7 Transport Properties 1.2.8 Drying Chemical and Physical Formation of Gels Actuation Methods 1.4.1 Thermally Driven Gel Actuators 1.4.2 Chemically Driven Gel Actuators 1.4.3 Gels Driven by Oscillating Reactions 1.4.4 Light Actuated Gels 1.4.5 Electrically Driven Gel Actuators 1.4.6 Electro- and Magneto-Rheological Composites 1.4.7 LC Elastomers Performance of Gels as Actuators Applications of Electroactive Gels 1.6.1 Gel Valves and Pumps 1.6.2 Light Modulators 1.6.3 Gel Drug Delivery 1.6.4 Gel Sensors Conclusions References

7 8 8 9 10 10 12 13 14 15 16 19 19 20 22 23 23 25 26 26 30 30 30 31 32 32 33

vi

2

Contents

Bio-Responsive Hydrogels for Biomedical Applications Tom McDonald, Alison Patrick, Richard Williams, Brian G. Cousins and Rein V. Ulijn

43

2.1 2.2 2.3 2.4 2.5

Introduction Chemical Hydrogels Physical Hydrogels Defining Bio-Responsive Hydrogels Bio-Responsive Chemical Hydrogels 2.5.1 Actuation Based on Changing the Cross-Linking Density 2.5.2 Actuation Based on Changes in Electrostatic Interactions 2.5.3 Actuation Based on Conformational Changes Bio-Responsive Physical Hydrogels 2.6.1 Enzyme-Responsive Physical Hydrogels Electroactive Chemical Hydrogels Conclusion References

43 44 44 44 46

Stimuli-Responsive and ‘Active’ Polymers in Drug Delivery Aram Omer Saeed, Jo´hannes Pa´ll Magnu´sson, Beverley Twaites and Cameron Alexander

61

3.1 3.2

61 62 62 63 63

2.6 2.7 2.8

3

3.3

3.4

Introduction Drug Delivery: Examples, Challenges and Opportunities for Polymers 3.2.1 Oral Drug Delivery Systems 3.2.2 Parenteral Drug Delivery 3.2.3 Topical and Transdermal Drug Delivery 3.2.4 Delivery Challenges for Biomolecular Drugs and Cell Therapeutics 3.2.5 Peptides and Proteins 3.2.6 Nucleic Acids 3.2.7 Cell Delivery Emerging State-of-the-Art Mechanisms in Polymer Controlled Release Systems 3.3.1 Technologies for Controlled Drug Release 3.3.2 Polymer–Drug Conjugates 3.3.3 Polymer–Protein Conjugates 3.3.4 Polymer–Nucleic Acid Conjugates 3.3.5 Polymer–Nucleic Acid Complexes Responsive or ‘Smart’ Polymers in Drug Delivery 3.4.1 Soluble Smart Polymers 3.4.2 Responsive Polymer–Drug Conjugates 3.4.3 Responsive Polymer–Protein Conjugates 3.4.4 Responsive Polymers for DNA Delivery

46 49 51 53 53 56 57 57

64 64 65 65 67 67 67 67 68 68 73 73 76 76 77

Contents

3.5 3.6

4

Recent Highlights of Actuated Polymers for Drug Delivery Applications Conclusions and Future Outlook References

Thermally Driven Hydrogel Actuator for Controllable Flow Rate Pump in Long-Term Drug Delivery Piero Chiarelli and Pietro Ragni 4.1 4.2 4.3

4.4 4.5

Introduction Materials and Methods Hydrogel Actuator 4.3.1 Thermo-Mechanical Gel Dynamics 4.3.2 Experimental Results Pump Functioning Conclusion References

SECTION II 5

78 80 81

89 89 90 90 91 93 97 98 98 101

IPMC Actuators: Fundamentals Kinji Asaka and Keisuke Oguro

103

5.1 5.2

103 104 104 105 108 110 113 116 117 118

5.3 5.4 5.5 5.6 5.7

6

IONIC POLYMER–METAL COMPOSITES (IPMC)

vii

Introduction Fabrication 5.2.1 Ionic Polymer 5.2.2 Plating Methods Measurement Performance of the IPMC Actuator Model Recent Developments Conclusion References

Active Microcatheter and Biomedical Soft Devices Based on IPMC Actuators Kinji Asaka and Keisuke Oguro 6.1 6.2 6.3 6.4

6.5

Introduction Fabrication of the IPMC Device Applications to the Microcatheter Other Applications 6.4.1 Sheet-Type Braille Display 6.4.2 Underwater Microrobot 6.4.3 Linear Actuators for a Biped Walking Robot Conclusions References

121 121 122 124 127 127 130 134 135 135

viii

7

Contents

Implantable Heart-Assist and Compression Devices Employing an Active Network of Electrically-Controllable Ionic Polymer–Metal Nanocomposites Mohsen Shahinpoor 7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8 7.9 7.10

8

9

Introduction Heart Failure Background of IPMNCs Three-Dimensional Fabrication of IPMNCs Electrically-Induced Robotic Actuation Distributed Nanosensing and Transduction Modeling and Simulation Application of IPMNCs to Heart Compression and Assist in General Manufacturing Thick IPMNC Fingers Conclusions References

IPMC Based Tactile Displays for Pressure and Texture Presentation on a Human Finger Masashi Konyo and Satoshi Tadokoro

137 137 139 140 141 142 144 146 149 155 157 157

161

8.1 8.2 8.3 8.4 8.5 8.6

Introduction IPMC Actuators as a Tactile Stimulator Wearable Tactile Display Selective Stimulation Method for Tactile Synthesis Texture Synthesis Method Display Method for Pressure Sensation 8.6.1 Method 8.6.2 Evaluation 8.7 Display Method for Roughness Sensation 8.7.1 Method 8.7.2 Evaluation 8.8 Display Method for Friction Sensation 8.9 Synthesis of Total Textural Feeling 8.9.1 Method 8.9.2 Experiments 8.10 Conclusions References

161 162 164 165 167 168 168 168 169 169 170 171 172 172 172 173 173

IPMC Assisted Infusion Micropumps Il-Seok Park, Sonia Vohnout, Mark Banister, Sangki Lee, Sang-Mun Kim and Kwang J. Kim

175

9.1 9.2 9.3

175 176 177 178

Introduction Background of IPMCs Miniature Disposable Infusion IPMC Micropumps 9.3.1 Configuration of the IPMC Infusion Pump

Contents

9.4

9.5

9.3.2 The Control System 9.3.3 Performance Testing Modelling for IPMC Micropumps 9.4.1 Equivalent Bimorph Beam Model for IPMC Actuators 9.4.2 IPMC Diaphragm Conclusions References

SECTION III 10

180 181 181 181 182 189 189 193

Conjugated Polymer Actuators: Fundamentals Geoffrey M. Spinks, Gursel Alici, Scott McGovern, Binbin Xi and Gordon G. Wallace

195

10.1 10.2 10.3 10.4 10.5

195 197 200 201

10.6 10.7

10.8

11

CONJUGATED POLYMERS

ix

Introduction Molecular Mechanisms of Actuation in ICPs Comparison of Actuation Performance in Various ICPs Electrochemistry of ICPs Effect of Composition, Geometry and Electrolyte on Actuation of PPy 10.5.1 Effect of the Dopant Ion 10.5.2 Effect of Solvent 10.5.3 Charge Transfer Processes 10.5.4 Effect of Porosity/Morphology Mechanical System Response Device Design and Optimization 10.7.1 How to Tailor Actuator Performance to Meet Design Requirements 10.7.2 Design of a Swimming Device 10.7.3 Device Testing Future Prospects References

204 204 206 208 212 212 217 217 219 221 222 223

Steerable Catheters Tina Shoa, John D. Madden, Nigel R. Munce and Victor X.D. Yang

229

11.1 11.2 11.3

229 229 231 231 232 232 232 234 234 235

11.4

Introduction Catheters: History and Current Applications Catheter Design Challenges 11.3.1 Biocompatibility 11.3.2 Small Size 11.3.3 Low Cost 11.3.4 Structural Rigidity Active Steerable Catheters 11.4.1 Non-EAP Based Steerable Catheters 11.4.2 EAP Based Steerable Catheters

x

Contents

11.5

12

Microfabricated Conjugated Polymer Actuators for Microvalves, Cell Biology, and Microrobotics Elisabeth Smela 12.1 12.2 12.3 12.4

12.5

12.6

12.7 12.8 12.9

13

14

11.4.3 Conjugated Polymer Based Steerable Catheters Discussion and Conclusion References

Introduction Actuator Background Microfabrication Single Hinge Bilayer Devices: Flaps and Lids 12.4.1 Bilayer Actuators 12.4.2 Drug Delivery 12.4.3 Cell Manipulation 12.4.4 Cell-Based Sensors Multi-Bilayer Devices: Positioning Tools 12.5.1 Microtools 12.5.2 Microrobot Swelling Film Devices: Valves 12.6.1 Out-of-Plane Actuation Strain 12.6.2 Microvalve Lifetime Integrated Systems Conclusions References

237 246 246

249 249 250 251 253 254 254 255 256 257 257 257 258 259 259 260 260 261 261

Actuated Pins for Braille Displays Geoffrey M. Spinks and Gordon G. Wallace

265

13.1 13.2 13.3 13.4 13.5 13.6

265 266 268 271 274 275 276 276

Introduction Requirements for the Electronic Braille Screen Mechanical Analysis of Actuators Operating Against Springs Polypyrrole Actuators for Electronic Braille Pins Other Polymer Actuation Systems for Electronic Braille Pins Summary Acknowledgements References

Nanostructured Conducting Polymer Biomaterials and Their Applications in Controlled Drug Delivery Mohammad Reza Abidian and David C. Martin 14.1 14.2

Introduction Nanostructured Conducting Polymers 14.2.1 Fabrication 14.2.2 Biomedical Application

279 279 280 280 282

Contents

14.3

14.4

15

Conducting Polymer Nanotubes for Controlled Drug Delivery 14.3.1 Electrospinning 14.3.2 Electrospinning of Dexamethasone-Loaded Template PLGA Nanofibers 14.3.3 Electrochemical Deposition of PEDOT Nanotubes 14.3.4 Controlled Drug Delivery from PEDOT Nanotubes Conclusions Acknowledgements References

285 286 287 288 289 293 293 293

Integrated Oral Drug Delivery System with Valve Based on Polypyrrole Thorsten Go¨ttsche and Stefan Haeberle

301

15.1 15.2 15.3

301 303 305 305 306 307 307 308 310 310 311 314 315 316

15.4

15.5

15.6

Introduction System Concept Osmotic Pressure Pump 15.3.1 Valve Closed 15.3.2 Valve Open Polypyrrole in Actuator Applications 15.4.1 Why PPy in the IntelliDrug System 15.4.2 Actuation of PPy Valve Concepts Evaluated in the Course of the IntelliDrug Project 15.5.1 Wafer-Level Fabricated Membrane Valve 15.5.2 Micro-Assembled Membrane Valve Total Assembly and Clinical Testing of the IntelliDrug System Acknowledgement References

SECTION IV 16

xi

PIEZOELECTRIC AND ELECTROSTRICTIVE POLYMERS 317

Piezoelectric and Electrostrictive Polymer Actuators: Fundamentals Zhimin Li and Zhongyang Cheng

319

16.1 16.2

319 320 320 321 323 324 325 326 326 328 328 330

16.3

16.4

Introduction Fundamentals of Electromechanical Materials 16.2.1 Piezoelectric Effect 16.2.2 Electrostrictive Effect 16.2.3 Other Effects Material Properties Related to Electromechanical Applications 16.3.1 Electromechanical Coupling Factor (k) 16.3.2 Elastic Response 16.3.3 Frequency and Temperature Responses Typical Electromechanical Polymers and Their Properties 16.4.1 Piezoelectric Polymers 16.4.2 Electrostrictive Polymers

xii

Contents

16.5

17

Miniature High Frequency Focused Ultrasonic Transducers for Minimally Invasive Imaging Procedures Aaron Fleischman, Sushma Srivanas, Chaitanya Chandrana and Shuvo Roy 17.1 17.2

17.3 17.4 17.5 17.6 17.7

18

Introduction Coronary Imaging Needs 17.2.1 Vulnerable Plaques 17.2.2 Stent Thrombosis High Resolution Ultrasonic Transducers 17.3.1 Polymer Transducers Fabrication Techniques Testing Methods Results Conclusion References

Catheters for Thrombosis Sample Exfoliation in Blood Vessels Using Piezoelectric Polymer Fibers Yoshiro Tajitsu 18.1 18.2 18.3 18.4 18.5

18.6

19

16.4.3 Maxwell Stress Effect Based Polymers 16.4.4 Practical Considerations Conclusions References

Introduction Piezoelectricity of Polymer Film and Fiber Simple Measurement Method for the Bending Motion of Piezoelectric Polymer Fiber Piezoelectric Motion of Poly-L-Lactic Acid (PLLA) Fiber Elementary Demonstration of Prototype System for Catheters Using Piezoelectric Polymer Fiber 18.5.1 Preliminary Demonstration 18.5.2 More Realistic Model for Application of Piezoelectric Polymer Fiber to Catheter Summary References

332 332 332 332

335

335 337 337 339 340 341 342 345 346 351 351

357 357 358 361 362 363 364 364 367 367

Piezoelectric Poly(Vinylidene) Fluoride (PVDF) in Biomedical Ultrasound Exposimetry Gerald R. Harris

369

19.1 19.2 19.3 19.4

369 370 371 372

Introduction Needle Hydrophone Design Spot Poled Membrane Hydrophone Design Application to Diagnostic Ultrasound

Contents

19.5 19.6

Application to Therapeutic Ultrasound Conclusion References

SECTION V 20

21

22

DIELECTRIC ELASTOMERS

xiii

374 377 378 385

Dielectric Elastomer Actuators: Fundamentals Roy Kornbluh, Richard Heydt and Ron Pelrine

387

20.1 20.2 20.3 20.4 20.5

387 388 389 391 392 393

Introduction Basic Principle of Operation Dielectric Elastomer Materials Transducer Designs and Configurations Operational Considerations References

Biomedical Applications of Dielectric Elastomer Actuators John S. Bashkin, Roy Kornbluh, Harsha Prahlad and Annjoe Wong-Foy

395

21.1 21.2 21.3 21.4 21.5 21.6 21.7 21.8

395 396 400 403 405 406 408 409 410

Introduction UMA Based Actuators and Their Application to Pumps Mechanical Stimulation Using Thickness-Mode Actuation Implantable Artificial Diaphragm Muscle Implantable Artificial Facial Muscles Limb Prosthetics and Orthotics Mechanical Actuation for ‘Active’ Cell Culture Assays Conclusions References

MRI Compatible Device for Robotic Assisted Interventions to Prostate Cancer Jean-Se´bastien Plante, Lauren Devita, Kenjiro Tadakuma and Steven Dubowsky 22.1 22.2

22.3

22.4

Introduction Prostate Cancer Therapy 22.2.1 Prostate Cancer Detection 22.2.2 Prostate Cancer Treatment 22.2.3 Needle Placement in MRI Systems Elastically Averaged Parallel Manipulator Using Dielectric Elastomer Actuators 22.3.1 Design Requirements 22.3.2 Manipulator Concept 22.3.3 Manipulator Analytical Model Results 22.4.1 Analytical Results 22.4.2 Experimental Results

411

411 413 413 414 415 415 415 417 418 420 421 423

xiv

Contents

22.5

23

A Braille Display System for the Visually Disabled Using a Polymer Based Soft Actuator Hyouk Ryeol Choi, Ig Mo Koo, Kwangmok Jung, Se-gon Roh, Ja Choon Koo, Jae-do Nam and Young Kwan Lee 23.1 23.2 23.3 23.4

23.5

23.6

24

Conclusions Acknowledgements References

Introduction Fundamentals of Actuation Principle Design of Tactile Display Device Braille Display System 23.4.1 Fabrication 23.4.2 System Outline 23.4.3 Experiments Advanced Applications 23.5.1 Wearable Tactile Display System 23.5.2 Virtual Reality Tactile Display Conclusions References

424 424 424

427

427 428 430 431 431 432 434 437 437 440 441 441

Dynamic Splint-Like Hand Orthosis for Finger Rehabilitation Federico Carpi, Andrea Mannini and Danilo De Rossi

443

24.1 24.2 24.3 24.4

443 444 445

Introduction Passive Dynamic Hand Splints: State of the Art Active Dynamic Hand Splints: State of the Art Proposed Concept: Dynamic Splint Equipped with Dielectric Elastomer Actuators 24.5 Splint Mechanics 24.6 Dimensioning of the Actuators 24.7 Prototype Splint 24.8 Performance of the Prototype Splint 24.9 Future Developments 24.9.1 Magnetic Resonance Imaging-Compatible Hand Splint 24.9.2 Electromyography-Controlled Hand Splint 24.10 Conclusions References Index

446 449 449 450 451 454 454 457 460 460 463

Preface

The great majority of traditional actuation technologies are based on thermochemical (combustion) motors, electromagnetic drives and hydraulic/pneumatic machines. However, these are inadequate to satisfy the diversity of new challenges presented in fields such as mechatronics, robotics and biomedical engineering. The biomedical field is particularly sensitive to the need for new types of actuators, since it includes applications from the nanoscale through to the macroscale, with requirements that differ enormously in terms of both structure and function. As an example, actuation devices span those designed to interact with single cells, and potentially organelles and even molecular structures, up to those used to replace limbs or to perform tele-operated surgery. Clearly, these different areas of application require different forces, displacements and speeds, as well as different durability, robustness and types of biocompatibility. Devices for cell manipulation might be single-use and disposable, whereas artificial hearts must sustain billions of cycles. New actuation materials and technologies should ideally have high work output, actuation strain, mechanical compliance, damage tolerance and efficiency. Depending on the application, they may also be required to be lightweight, have compact and simple structures that can readily be fabricated, and be reasonably low cost. In most scenarios, these new technologies will serve a complementary role working together with conventional actuators. In the last few decades this need for new actuators has drawn considerable effort towards the development of materials that can directly transduce an input energy into mechanical work. Much of this attention has been focused, and is increasingly being focused today, on electroactive polymers (EAPs). This is a large family of materials that includes many different chemical structures, actuation mechanisms and electromechanical performances. Although most EAP materials have been known for decades, before now they found limited translation from proof of concept demonstrations in the laboratory to actual use, despite their potential. This has changed with recent developments in materials science, processing, configuration design and driving strategies which are permitting serious efforts towards concrete exploitation, as this book describes. In fact, EAPs are opening the way to numerous new applications precluded by conventional actuation technologies. This book intends to provide a comprehensive and updated insight into both the fundamentals of each class of EAP, and examples of the most significant applications of EAP actuators in the biomedical field, either already demonstrated or currently under development. For this purpose, the book comprises five sections devoted to the most technologically mature EAPs, namely polymer gels, ionic polymer–metal composites, conjugated polymers, piezoelectric/electrostrictive polymers and dielectric elastomers. Each section is

xvi

Preface

introduced by a chapter that is focused on the fundamentals and which aims to provide a description of the main features of the technology and the current state of the art. These introductory chapters are followed by chapters describing specific applications. The contributors to this book are inventors and international leaders in the field. The broad and far-reaching range of applications covered by this book is intended not only to make it the first text on biomedical uses of the emerging EAP based actuation technologies, but also to serve as a source of inspiration for possible new applications aimed at improving health and well-being. Federico Carpi, University of Pisa Elisabeth Smela, University of Maryland October 2008

List of Contributors

Cameron Alexander, School of Pharmacy, University of Nottingham, UK Gursel Alici, ARC Centre of Excellence for Electromaterials Science and Intelligent Polymer Research Institute, University of Wollongong, Australia Kinji Asaka, National Institute of Advanced Industrial Science and Technology (AIST), Japan Mark Banister, Medipacs LLC, Tucson, USA John Bashkin, Fremont, CA, USA Paul Calvert, University of Massachusetts, Dartmouth, USA Federico Carpi, Interdepartmental Research Centre ‘‘E. Piaggio’’, School of Engineering, University of Pisa, Italy Chaitanya Chandrana, Cleveland Clinic, Lerner Research Institute, Department of Biomedical Engineering, Cleveland, USA Zhongyang Cheng, Materials Research and Education Center, Alkermes Inc., Auburn, Alabama, USA Piero Chiarelli, Institute of Clinical Physiology, CNR, Italy Hyouk Ryeol Choi, School of Mechanical Engineering, Chemical Engineering, Polymer System Engineering, Sungkyunkwan University, Korea Brian G. Cousins, School of Materials, Materials Science Centre and Manchester Interdisciplinary Biocentre (MIB), University of Manchester, United Kingdom Danilo De Rossi, Interdepartmental Research Centre ‘‘E. Piaggio’’, School of Engineering, University of Pisa, Italy Lauren Devita, Massachusetts Institute of Technology, USA

xviii


Steven Dubowsky, Massachusetts Institute of Technology, USA Aaron Fleischman, Cleveland Clinic, Lerner Research Institute, Department of Biomedical Engineering, Cleveland, USA Thorsten Go¨ttsche, Institut fu¨r Mikro- und Informationstechnik of the Hahn-SchickardGesellschaft (HSG-IMIT), Germany Stefan Haeberle, Institut fu¨r Mikro- und Informationstechnik of the Hahn-SchickardGesellschaft (HSG-IMIT), Germany Gerald R. Harris, Food and Drug Administration, Center for Devices and Radiological Health, USA Richard Heydt, SRI International, USA Kwangmok Jung, Pohang Institute of Intelligent Robotics, Korea Kwang J. Kim, University of Nevada, USA Sang-Mun Kim, University of Nevada, USA Masashi Konyo, Graduate School of Information Sciences, Tohoku University, Japan Ig Mo Koo, School of Mechanical Engineering, Chemical Engineering, Polymer System Engineering, Sungkyunkwan University, Korea Ja Choon Koo, School of Mechanical Engineering, Chemical Engineering, Polymer System Engineering, Sungkyunkwan University, Korea Roy Kornbluh, SRI International, USA Sangki Lee, University of Nevada, USA and Volvo Korea, South Korea Young Kwan Lee, School of Mechanical Engineering, Chemical Engineering, Polymer System Engineering, Sungkyunkwan University, Korea Zhimin Li, Pharmaceutical Chemistry, Auburn University, Cambridge, Massachusetts, USA John D. Madden, Advanced Materials and Process Engineering Laboratory and Department of Electrical & Computer Engineering, University of British Columbia, Vancouver, Canada Jo´hannes Pa´ll Magnu´sson, School of Pharmacy, University of Nottingham, UK Andrea Mannini, Interdepartmental Research Centre ‘‘E. Piaggio’’, School of Engineering, University of Pisa, Italy


xix

David C. Martin, Biomedical Engineering, Materials Science and Engineering and Macromolecular Science and Engineering, The University of Michigan, Ann Arbor, MI, USA Tom McDonald, School of Materials, Materials Science Centre, University of Manchester and Manchester Interdisciplinary Biocentre (MIB), University of Manchester, United Kingdom Scott McGovern, ARC Centre of Excellence for Electromaterials Science and Intelligent Polymer Research Institute, University of Wollongong, Australia Nigel R. Munce, Imaging Research, Sunnybrook Health Science Centre, University of Toronto, Canada Jae-do Nam, School of Mechanical Engineering, Chemical Engineering, Polymer System Engineering, Sungkyunkwan University, Korea Keisuke Oguro, National Institute of Advanced Industrial Science and Technology (AIST), Japan Il-Seok Park, University of Nevada, USA Alison Patrick, School of Materials, Materials Science Centre and Manchester Interdisciplinary Biocentre (MIB), University of Manchester, United Kingdom Ron Pelrine, SRI International, USA Jean-Se´bastien Plante, Universite´ de Sherbrooke, Canada Harsha Prahlad, SRI International, USA Pietro Ragni, Institute of Nuclear Chemistry, CNR, Italy Mohammad Reza Abidian, Biomedical Engineering, The University of Michigan, Ann Arbor, MI, USA Se-gon Roh, School of Mechanical Engineering, Chemical Engineering, Polymer System Engineering, Sungkyunkwan University, Korea Shuvo Roy, Cleveland Clinic, Lerner Research Institute, Department of Biomedical Engineering, Cleveland, USA Aram Omer Saeed, School of Pharmacy, University of Nottingham, UK Mohsen Shahinpoor, Biomedical Engineering Laboratories, Department of Mechanical Engineering, University of Maine, Orono, USA

xx


Tina Shoa, Advanced Materials and Process Engineering Laboratory and Department of Electrical & Computer Engineering, University of British Columbia, Vancouver, Canada Elisabeth Smela, Department of Mechanical Engineering, University of Maryland, USA Geoffrey M. Spinks, ARC Centre of Excellence for Electromaterials Science and Intelligent Polymer Research Institute, University of Wollongong, Australia Sushma Srivanas, Cleveland Clinic, Lerner Research Institute, Department of Biomedical Engineering, Cleveland, USA Kenjiro Tadakuma, Massachusetts Institute of Technology, USA Satoshi Tadokoro, Graduate School of Information Sciences, Tohoku University, Japan Yoshiro Tajitsu, Smart Structures and Materials Laboratory, Department of Electrical Engineering, Graduate School of Engineering, Kansai University, Japan Beverley Twaites, School of Pharmacy and Biomedical Sciences, University of Portsmouth, UK Rein V. Ulijn, University of Strathclyde, United Kingdom Sonia Vohnout, Medipacs LLC, Tucson, USA Gordon G. Wallace, ARC Centre of Excellence for Electromaterials Science and Intelligent Polymer Research Institute, University of Wollongong, Australia Richard Williams, School of Materials, Materials Science Centre and Manchester Interdisciplinary Biocentre (MIB), University of Manchester, United Kingdom Annjoe Wong-Foy, SRI International, USA Binbin Xi, ARC Centre of Excellence for Electromaterials Science and Intelligent Polymer Research Institute, University of Wollongong, Australia Victor X.D. Yang, Imaging Research, Sunnybrook Health Science Centre and Department of Electrical and Computer Engineering, Ryerson University, Toronto, Canada

Plate 1

A heart with an IPMNC compression band. (See Figure 7.17)


2.58 10 5

25

10 63 Laser beam line width: 0.2 (a)

(mm)

(b)

Plate 2 (a) Laser cutting machine and (b) a CAD design for interdigitated IPMC (Reproduced with permission from Vohnout, S., Kim, S.-M., Park, I.-S. and Banister, M., IPMC-assisted miniature disposable infusion pumps with embedded computer control, Proceedings of the SPIE conference 2007. Copyright (2007) SPIE). (See Figure 9.5)

MSC.Patran 2001 r2a 16-Aug04 10:12:28 Fringe: SC1:DIAPHRAGM, A14:Static Subcase: Displacements, Translational-(NON-LAYERED) (ZZ) Deform: SC1:DIAPHRAGM, A14:Static Subcase: Displacements, Translational

9.66–004 9.02–004 8.38–004 7.73–004 7.09–004 6.44–004 5.80–004 5.15–004

9.66–004

0.

4.51–004 3.87–004

+

3.22–004 2.58–004 1.93–004 1.29–004 6.44–005

Z

–1.16–010 Y X

default_Fringe: Max 9.66–004 @Nd1 Min 0. @Nd 316 default_Deformation : Max 9.66–004 @Nd1

(a)


0 –4.57–005 –9.14–005 –1.37–004 –1.83–004 –2.29–004 –2.74–004 –3.20–004

0. +

–3.66–004 –4.11–004 –4.57–004

–0

–5.03–004 –5.48–004 –5.94–004 –6.40–004 Z

–6.86–004 Y X

(b)

default_Fringe: Max 0. @Nd316 Min –6.86–004@Nd1 default_Deformation : Max 6.86–004 @Nd1

Plate 3 Deformed shapes of IPMC diaphragms: (a) circle-shaped electrode (radius of electrode ¼ 8.5 mm); (b) ring-shaped electrode (radial length of electrode ¼ 5.5 mm) (Reproduced with permission from Lee, S., Kim, K.J. and Park, H.C. (2006) Modeling of an IPMC Actuator-driven Zero-Net-Mass-Flux Pump for Flow Control, J. Intelligent Mat. Systems and Structures, 17, 6, 533–41. Sage Publications). (See Figure 9.12)

MSC.Patran 2001 r2a 20-Aug04 14:27:41 Fringe: SC1:DIAPHRAGM, A2:Mode 1: Freq. = 429.69: Eigenvectors, Translational-(NON-LAYERED) (MAG) Deform: SC1:DIAPHRAGM, A2:Mode 1: Freq. = 429.69: Eigenvectors, Translational

3.15+002 2.94+002 2.73+002 2.52+002 2.31+002 2.10+002 1.89+002 1.68+002

3.15+002

1.47+002

0.

1.26+002

+

1.05+002 8.41+001 6.31+001 4.21+001 2.10+001

Z

4.20–005 Y X

default_Fringe: Max 3.15+002 @Nd1 Min 0. @Nd 316 default_Deformation : Max 3.15+002 @Nd1

(a)

MSC.Patran 2001 r2a 20-Aug-04 14:28:51 Fringe: SC1:DIAPHRAGM, A2:Mode 2: Freq. = 1659.1: Eigenvectors, Translational-(NON-LAYERED) (MAG) Deform: SC1:DIAPHRAGM, A2:Mode 2: Freq. = 1659.1: Eigenvectors, Translational

3.02+002 2.81+002 2.61+002 2.41+002 2.21+002 2.01+002 1.81+002 1.61+002

0 3.02+0

2

0.

1.41+002 1.21+002

+

1.01+002 8.04+001 6.03+001 4.02+001 2.01+001

Z

3.81–006 Y X

(b)

default_Fringe: Max 3.02+002 @Nd261 Min 0. @Nd316 default_Deformation : Max 3.02+002 @Nd261

Plate 4 Normal mode analysis results for an IPMC diaphragm (radius of electrode ¼ 8.5 mm): (a) first mode; (b) second mode (Reproduced with permission from Lee, S., Kim, K.J. and Park, H.C. (2006) Modeling of an IPMC Actuator-driven Zero-Net-Mass-Flux Pump for Flow Control, J. Intelligent Mat. Systems and Structures, 17, 6, 533–41. Sage Publications). (See Figure 9.14)

Plate 5 Schematic diagrams illustrating the surface modification of neural microelectrodes to create nanotubular PEDOT: (A) electrospinning of biodegradable polymer (PLGA) fibers with well-defined surface texture (1) on the probe tip; (B) electrochemical polymerization of conducting polymers (PEDOT) (2) around the electrospun fibers; and (C) dissolving the electrospun core fibers to create nanotubular conducting polymers (3) [7] (Reprinted wih permission from Advanced Materials, Conducting polymer nanotubes for controlled drug release by Abidian, M. R., et al., 18, 4, 405–9. Copyright (2006) Wiley-VCH Verlag GmbH Co. KGaA). (See Figure 14.5)

Plate 6 Optical micrographs of: (E) the gold electrode site; (F) the electrode site after electrospinning showing the coverage of the PLGA electrospun nanoscale fibers; (G) the electrode after electrochemical deposition of PEDOT on the gold site and around the electrospun fibers; and (H) the electrode after removal of the core nanoscale fiber templates (Reprinted with permission from Advanced Materials, Conducting polymer nanotubes for controlled drug release by Abidian, M. R., et al., 18, 4, 4059. Copyright (2006) Wiley-VCH Verlag GmbH Co. KGaA). (See Figure 14.7)

A

C

B

D

E

F CE

CE

v

v

WE

WE

Electrolyte

Dexamethasone Anion Cation

CE Counter Electrode WE Working Electrode

Plate 7 Schematic illustration of the controlled release of dexamethasone: (A) dexamethasoneloaded electrospun PLGA; (B) hydrolytic degradation of PLGA fibers leading to release of the drug; and (C) electrochemical deposition of PEDOT around the dexamethasone-loaded electrospun PLGA fiber slows down the release of dexamethasone (D); (E) PEDOT nanotubes in a neutral electrical condition; (F) external electrical stimulation controls the release of dexamethasone from the PEDOT nanotubes due to contraction or expansion of the PEDOT (Reprinted with permission from Advanced Materials, Conducting polymer nanotubes for controlled drug release by Abidian, M. R., et al., 18, 4, 405–9. Copyright (2006) Wiley-VCH Verlag GmbH Co. KGaA). (See Figure 14.9)

1.0 Absorbance (AU)

Cumulative Mass Released (mg)

2.0 1.5 1.0 0.5 0.0

0

200 400 600 800 1000 1200 1400 Time (h) (a)

0.8 0.6 0.4 0.2 0.0

220 240 260 280 300 Wavelength (nm)

320

(b)

Plate 8 (a) Cumulative mass release of dexamethasone from: PLGA nanoscale fibers (black squares), PEDOT-coated PLGA nanoscale fibers (red circles) without electrical stimulation, and PEDOT-coated PLGA nanoscale fibers with electrical stimulation of 1 V applied at the five specific times indicated by the circled data points (blue triangles). (b) UV absorption of dexamethasoneloaded PEDOT nanotubes after 16 h (black), 87 h (red), 160 h (blue) and 730 h (green). The UV spectra of dexamethasone have peaks at a wavelength of 237 nm. Data are shown with a standard deviation (n ¼ 15 for each case) (Reprinted with permission from Advanced Materials, Conducting polymer nanotubes for controlled drug release by Abidian M.R., et al., 18, 4, 405–9. Copyright (2006) Wiley-VCH Verlag GmbH Co. KGaA). (See Figure 14.10(a) and 14.10(b))

Introduction Electroactive Polymers as Smart Materials for Actuation Federico Carpi 1 and Elisabeth Smela 2 1

Interdepartmental Research Centre ‘E. Piaggio’, University of Pisa, Italy 2 Department of Mechanical Engineering, University of Maryland, USA

I.1 Actuation: the Need for New Materials and Technologies Actuators are materials, devices or systems that are able to act upon their external environment by transducing input energy into external mechanical work. Biological muscles have long drawn interest as natural actuation systems, and many regard their functional properties as ideal models for artificial actuators. The conventional actuators used today, which mainly comprise thermochemical motors (combustion engines), electromagnetic drives and hydraulic/pneumatic machines, differ considerably from their natural counterparts. For example, their underlying physical principles of actuation share little with biological muscles. In fact, the latter are electro-chemo-mechanical actuators based on contractile proteins that use the body’s chemical energy (adenosine triphosphate, ATP) to generate motion, triggered by neuro-electric commands. Despite the advanced performance of conventional motor technologies, which in some respects outreaches that of natural muscles, they cannot satisfy the ever-increasing demand for new actuators in a large number of quite different areas of technology, spanning the range from mechatronics to biomedical engineering. In particular, new actuation technologies are needed that are easily scalable, structurally simple and mechanically compliant, while also having high power-to-weight and power-to-volume ratios, and fine control capability. To illustrate the need for new approaches to actuation, consider that despite years of effort to develop prostheses (artificial hands and arms) driven by electric motors, currently available systems are still stiff, heavy and noisy. The effort to develop ‘artificial muscles’


2


would benefit from new actuators with intrinsic properties similar to those of biological muscular tissue, in terms of both passive characteristics (namely high mechanical compliance and low mass density) and active behaviour (intrinsic actuation capabilities in response to electrical stimuli). Another consideration is the small size of the objects that it is wished to move or manipulate, for which miniaturization of conventional actuators is not going to be possible. For instance, insect-size robots that can fly, hop and crawl are one example of systems that cannot be built with conventional actuators. Fabrication paradigms and driving principles prevent conventional drives from being scaled down further. The inability of traditional actuation technologies to cover the entire spectrum of requirements for new, and vastly different, application areas has given rise to new materials, new drive principles and new devices. These are intended to play at least a complementary role, so as to compensate the deficits of the more consolidated technologies in specific domains. To satisfy the needs described above, new research avenues focused on electroactive polymers have been opened in recent years.

I.2 Electroactive Polymers: Classification Polymers are promising candidates for new actuation technologies. Several classes of polymeric compounds can convert electrical energy (or other sources of energy, such as heat, light and chemical gradients) into mechanical work, so as to effect the movement of loads. Piezoelectric and shape memory polymers are among the most well known representatives of polymer actuators. Electroactive polymers (EAPs) [1, 2] comprise a broad family not only of active materials like piezoelectrics that intrinsically change volume, but also of polymercontaining devices in which the polymer is passive. They share the capability of changing dimensions and/or shape in response to a stimulus, which is most preferably electrical. They can have sizable actuation strains compared with inorganic materials (although actuation stresses are lower), high compliance, low mass density and scalability. They also promise ease of processing and low cost. As a result, EAPs are one of the most promising classes of materials for ‘muscle-like’ actuators [1, 2]. EAPs can be divided into two main classes: ionic, which are activated by electricallyinduced transport of ions and/or solvent, and electronic, which are activated by electric fields [1, 2]. The former generally require low voltage but high currents, while the latter need high voltage but low currents. They thus have different application areas, as described in detail in the later chapters of this book. Another difference is that the ionic EAPs operate in a liquid electrolyte medium, while the electronic EAPs are used in air. The ionic EAPs are thus particularly attractive for use within biological environments. The electronic EAPs can operate at higher frequencies and with higher efficiencies in converting electrical energy into mechanical work, making them of more interest in applications such as robotics. No EAP technology can today play a general-purpose role, and the selection of the most suitable actuator should be carefully evaluated according to the specific requirements of the application. Within each group there are a wide variety of specific actuation mechanisms and related types of materials. In particular, ionic EAPs include polymer gels [3], ionic polymer–metal

Electroactive Polymers as Smart Materials for Actuation

3

composites (IPMC) [4], conjugated (conducting) polymers [5] and carbon nanotubes [6]. Electronic EAPs include piezoelectric polymers [7], electrostrictive polymers [8], flexoelectric polymers [9] and dielectric elastomers [10]. The different types of EAPs and the most significant materials used in each type are summarized in Table I.1. Table I.1

Electroactive polymer actuators: classification and representative materials

EAP class

EAP sub-categories

Ionic EAP

Polymer gels

Electronic EAP

Examples of materials

Poly(acrylic acid) (PAAc) Poly(vinyl alcohol) (PVA) Modified poly(acrylonitrile) (PAN) Ionic polymer–metal Metalized ion exchange membranes e.g. composites (IPMC) Nafion/Pt Conjugated polymers Polypyrrole (PPy) Polyaniline (PANi) Carbon nanotubes Single-walled nanotubes (SWCNT) Multi-walled nanotubes (MWCNT) Piezoelectric polymers Poly(vinylidene fluoride) (PVDF) Electrostrictive polymers PVDF based copolymers e.g.: Poly(vinylidene fluoride–trifluoroethylene) (PVDF–TrFE) Poly(vinylidene fluoride– hexafluoropropylene) (PVDF–HFP) Flexoelectric polymers Liquid crystal elastomers Dielectric elastomers Silicone elastomers Acrylic elastomers Polyurethane elastomers

Ref. [3] [4] [5] [6] [7] [8]

[9] [10]

The remainder of this book is organized into sections covering each of the main classes of EAPs. The first chapter of each section focuses on the fundamentals of each technology, describing the actuation mechanism and performance. Two of the classes listed in Table I.1, carbon nanotubes and flexoelectric polymer actuators, have not been included in this book because of their lower technological maturity.

References 1. Bar-Cohen, Y. (2004) (ed.), Electroactive Polymer (EAP) Actuators as Artificial Muscles: Reality, Potential, and Challenges, 2nd edn, SPIE Press Monograph, Vol. PM136. 2. Madden, J. D. W., Vandesteeg, N. A., Anquetil, P. A., et al. (2004) Artificial muscle technology: physical principles and naval prospects, IEEE J. Oceanic Eng., 29, 3, 706–28. 3. Tanaka, T., Nishio, I., Sun, S-T. and Ueno-Nishio, S. (1982) Collapse of gels in an electric field, Sci., 218, 467–9. 4. Asaka, K., Oguro, K., Nishimura, Y., et al. (1995) Bending of Polyelectrolyte MembranePlatinum Composites by Electric Stimuli, I. Response Characteristics to Various Waveforms, Polym. J. 27, 4, 436–40. 5. Baughman, R. H. (1996) Conducting polymer artificial muscles, Synth. Met., 78, 339–53. 6. Baughman, R. H., Changxing, C., Zakhidov, A. A., et al. (1999) Carbon nanotube actuators, Sci., 284, 1340.

4


7. Nalwa, H. S. (1995) Ferroelectric Polymers, Marcel Dekker, New York. 8. Zhang, Q. M., Bharti, V., Zhao, X. (1998) Giant electrostriction and relaxor ferroelectric behaviour in electron-irradiated poly(vinylidene fluoride–trifluoroethylene) copolymer, Sci., 280, 2101–3. 9. Lehmann, W., Skupin, H., Tolksdorf, C., et al. (2001) Giant lateral electrostriction in ferroelectric liquid-crystalline elastomers, Nat., 410, 447–50. 10. Pelrine, R., Kornbluh, R., Pei, Q. and Joseph, J. (2000) High-speed electrically actuated elastomers with strain greater than 100%, Sci., 287, 836–839.

Section I Polymer Gels


1 Polymer Gel Actuators: Fundamentals Paul Calvert University of Massachusetts, Dartmouth, USA

1.1 Introduction and Historical Overview One view of materials development is to search for what materials correspond to empty spaces on a hypothetical multi-dimensional map of the properties of available materials. Following a biomimetic philosophy, for instance, it can be seen that tough ceramics and moldable short fiber composites with high moduli are possible but absent from the list of available synthetic materials. Likewise artificial muscle is missing, where the properties are defined as a developed stress of over 300 kPa, a linear contraction of 25 % and a response time of below one second. Currently dielectric elastomers come closest but have disadvantages [1, 2]. The actin–myosin muscle system provides the performance target for electroactive polymer actuators [3, 4]. The process is driven chemically by the energy change from hydrolysis of the polyphosphate bond as ATP (adenosine triphosphate) binds to myosin and is converted to ADP (adenosine diphosphate) and phosphate. A simple chemical analogy suggests that muscle-like gels should be feasible but it is now clear that the task is much harder than it seems. Early work on gel actuation by Katchalsky–Katzir demonstrated that engines could be built using the chemical energy in diluting a lithium bromide solution to drive contraction and expansion of a collagen belt [5]. In essence, the collagen expands to take up the salt solution or contract to exclude the water. This change is both a molecular conformation change and a volume change. Other chemically driven gels, for instance polyacrylic acid fibers [6] which respond to a pH change, also rely on a solubility change giving rise to a volume change.


8


Work by Osada and others explored electrically-driven actuators where pH changes at each electrode drive local volume changes [7]. While electrical power should be a more convenient way of driving artificial muscles, it has proved to be difficult to couple the electrical energy input to provide a mechanical output. As a result the achievable efficiencies for energy conversion are very low. These systems driven by volume changes are rate-limited by the need for water or solvent to diffuse in and out of the gel. The force developed is also limited by the free energy associated with the change in solvent content. In this regard, the apparent analogy between muscle action and gel contraction breaks down. Muscle contraction is essentially a shape change driven by a conformation change. As such there seems to be no significant volume change and so no diffusion limitation. The energy source is dephosphorylation of ATP, which is also large compared to many solvation interactions. The fact that early work of Katchalsky used collagen, which so resembles muscle, may have been misleading. Subsequent work that has focused on similar artificial muscles has developed systems that are intrinsically too slow and too feeble to provide the force and speed associated with natural muscle. In this chapter the properties and applications of current gel actuators are summarized and the developments needed to produce a true artificial muscle are discussed.

1.2 Properties of Gels To be useful as actuators, gels must be stiff enough to exert the desired force and must be strong enough to carry the desired load. If the change in shape is caused by swelling, the fluid must flow into the gel in a short time. The transport of fluid will also be important to whether the actuator can be used without drying out. In addition, because gels cannot be readily machined or molded, it must be possible to form the gel into a suitable shape for an actuator by polymerization in situ. If gels are to be used as actuators, they must be mechanically robust. Most current applications of gels are in food, pharmaceuticals and cosmetics where mechanical properties are not a major concern. Muscle as an actuator develops a force of about 300 kPa and a strain of 25 %. If these are regarded as target numbers, a gel actuator should be strong enough to easily withstand the force developed, and so the fracture strength should be at least 1 MPa. In addition, it should be stiff enough that, in lifting a weight for instance, an active contraction is not simply cancelled by passive extension under the load. If it is said that the passive extension should be less than one-tenth of the contraction, the elastic modulus would be greater than 12 MPa. While a low force gel actuator may be useful for some applications, such as a microfluidics valve, the mechanical requirements on a muscle-like actuator are quite demanding. 1.2.1

Biological Gels

In biology there are many examples of structural gels in the marine environment, including seaweeds and the bodies of many invertebrates, such as sea anemones. In the human body, cartilage, cornea, the dermis and arterial wall are all fiber-reinforced gels. Although soft and not very strong, these materials are very tough and so survive the impacts of life in motion better than many hard materials in machines.

Polymer Gel Actuators: Fundamentals

9

Many biological tissues show a ‘J-shaped’ tensile stress–strain curve [8] that combines a low initial modulus with high strength. The reinforcing fibers initially rotate as the soft gel is stretched and then the fibers take up the load as they become parallel to the stress axis. Other variations occur depending on whether the fibers are bonded into a network. Unbonded fibers can flow with the gel under slow loading but give rigidity under fast loading [9]. Articular cartilage is a proteoglycan gel reinforced with about 20 % of collagen fibers. It has a strength of about 1.5 MPa and an extension to break of about 100 %. The structure is layered and the properties vary greatly with depth below the surface, with strength up to up to 30 MPa in layers with higher fiber contents [8, 10–13]. Costal (rib) cartilage has a strength of 5–7 MPa [14]. The large extension to break and large work of fracture (1 kJ/m2) [15] allow cartilage to function effectively under impact even though the average strength is not high. An unfamiliar aspect of the mechanical properties of gels is that they will tend to lose water under compression and take up water under tension. As a result, the mechanical properties will be different depending on the test speed. Thus the fast modulus of cartilage is 2.5 MPa while the equilibrium modulus, measured as water is displaced from the structure, is about 0.7 MPa [8, 12]. The transition between these two values will depend on sample size, as the water has to flow out of the gel. Likewise, testing under water will result in properties that differ from those measured in air. Cornea is another tissue that is reinforced with collagen fibers. As with cartilage, there is an immediate need for a synthetic substitute to replace damaged corneal tissue. Cornea contains about 20 % of collagen fibers in a gel matrix. The fiber diameter is of the order of 20 nm, so that light is not scattered and the material is transparent [16]. The tensile strength is about 4 MPa and the elastic modulus is about 6 MPa at 20 % strain on the J-shaped stress–strain curve [17]. The properties of these soft tissues and marine gels suggests that a combination of a higher polymer content and fiber reinforcement should let us form materials which retain the responsive properties of gels whilst having sufficient mechanical strength to be used in engineering systems. As a target system, we could envision seeking a synthetic gel with a tensile strength of 5 MPa and an extension to break of 100 % but we do also need to develop a better understanding of the mechanical properties of gels. 1.2.2

Mechanical Properties of Simple, Single-Phase Gels

Because gels are weak and so do not currently have many synthetic applications, there is not a large coordinated literature on their mechanical properties. For many materials we can consider elastic modulus and tensile strength as sufficient to characterize the mechanical properties. The first of these reflects the rigidity or degree of bending under stress, the second the ability to withstand static stress without breaking or deforming irreversibly. For hard materials that have to withstand impacts, we are also concerned with the toughness, often measured as the energy absorbed in propagating a crack through the material. These concepts are substantially derived from the consideration of metals and ceramics where their stress–strain relationship is essentially linear up to about 1 % strain and then yield or fracture occurs. They serve well in most engineering situations where objects are designed to be rigid.

10


Many tissues operate in a different regime, where there are large reversible strains and substantial impacts can be tolerated without damage. In this case the energy needed to produce damage may be more important than the strength. The stress–strain curve may be very nonlinear and the shape of the curve out to large strains becomes important. The same is true of rubbery materials, but these are also not very familiar in structural engineering. Thus, engineering with gels will put us into a regime that is unfamiliar to many mechanical engineers. 1.2.3

Elastic Moduli

One view of a gel is as a modified rubber. The properties of amorphous polymers change dramatically above the glass transition point where the chains become mobile. Since polymer chains are mobile in solution and we think of a simple gel as a cross-linked solution, we can regard a single phase gel as a dilute rubber. Dense polyacrylamide, for instance, is a glassy polymer. We do not want to compare the gel properties with this state but with the same polymer as a cross-linked rubber above its glass transition. Gels are soft materials, so we would expect elastic moduli to be below 10 MPa and we would expect the modulus to decrease as the volume fraction of solvent increases. As an example, a gelatin gel swollen to five times its dry weight, has a modulus of about 0.8 MPa and a fracture stress of about 70 kPa with an extension to break of 10 %. At a swelling of 40 times, the modulus is only 40 kPa and the strength 6 kPa with the extension to break 11 %. This soft, weak, brittle behavior is characteristic of most simple gels. The initial elastic modulus of a nonionic gel can be derived from rubber elasticity theory as: G¼A

2=3 RT v02 ðv2 Þ1=3 Mc

ð1:1Þ

where A is close to one, r is density, Mc is the average molecular weight between cross-links, v2 is the volume fraction of polymer in the swollen gel and v20 is the volume fraction in the gel as synthesized. Thus swelling after synthesis decreases the modulus relative to the Go, the modulus as synthesized [18, 19]: 1=3 G v2 ¼ ¼ Vr 1=3 ð1:2Þ G0 v02 Many natural gels are highly charged polyelectrolytes. It might be expected that there would be a strong difference in modulus between otherwise comparable ionic and nonionic gels. For weakly charged groups (acrylic acid), charge seems to have little effect on elastic properties [20]. Other studies show that charged groups increase the modulus and decrease the dependence of modulus on swelling [21]. 1.2.4

Strength

Synthetic gels, based on cross-linked soluble polymers, are mostly too weak to be used in any structural application. Many natural gel structures, such as are found in marine organisms, seem to be quite strong. As discussed next, possibly this difference arises from the microstructure of natural gels, which most synthetic gels lack.


11

In determining the strength of gels, there are important factors which can often be ignored in dense materials. Gels often fracture at much higher strains than conventional engineering materials, properties can be very time dependent and liquid may be taken up or lost during the test. Likewise, the properties of immersed gel samples differ from samples tested in air, as water is normally taken up in tension and exuded in compression. The degree of confinement and timescale of testing is also going to be important for the same reasons. At high compressive strains sample geometry will also be crucial, since friction at the platens will result in shear stresses which effectively put the sample into hydrostatic compression where failure cannot occur. Thus properties at high strain in compression must not be regarded as directly comparable to a true strength. Most hard amorphous polymers under tension show brittle fracture. The strength, s, is determined under the Griffith equation (Equation (1.3)) by the fracture surface energy, g, which in turn mostly depends on the energy absorbed by the plastic deformation and void formation (crazing) that occurs immediately at the tip of the crack. E is the elastic modulus and c is the crack length: 2 E 1=2 ¼ ð1:3Þ pc Since elastomers are essentially liquid polymers, the elastic modulus is low and crazing is not believed to occur. Most of the fracture energy probably goes into pulling individual chains out across the fracture and so the energy increases with the chain length between cross-links [22]. Fracture of rubbers does not follow the Griffith theory because of the higher extensions at fracture, but the role of fracture energy in limiting crack extension still applies. Most unreinforced elastomers lack significant energy absorbing mechanisms and so readily tear at any cut or notch. Based on the comparison between gels and elastomers, we would expect the strength of unstructured gels to be lower than that of rubbers with a similar cross-link density by factors reflecting the dilution of the gel by water or solvent and reflecting the degree of preextension of the chain due to swelling by the solvent. We thus expect gels to be weak and to get weaker as they swell more. There are exceptions to the generally low strength of elastomers where some energyabsorbing deformation can occur. One example is natural rubber, where the crystallization occurs under tension, resulting in increased stiffness at high stress and a large energy to fracture as chains slip through the ordered crystals. Large fracture energies are also obtained when diene rubber chains slip over the surface of reinforcing carbon black particles or through the hard regions of two-phase polyurethane elastomers. It may be possible to build similar energy-absorbing mechanisms into gels. Theoretical discussion of gel fracture has focused on gelatin gels, which are important in food. Both fracture mechanics and fracture energy approaches have been considered but understanding is still imperfect [23–25]. Synthetic gels based on acrylates are unstructured and so would also be expected to be weak. Natural gels, such as agarose [26] and the calcium alginates, do seem to form ordered regions of double helix or multiple helices [27–29]. As a result, the mechanical properties are very dependent on the extent of structure developed during gelation [30]. The disruption of these structures during fracture could be expected to be a source of energy

12


absorption and so increase strength and toughness. 2 % Agarose gels have a strength of about 0.14 MPa and a strain to failure of 40 %. In contrast, the strength of similar gelatin gels is about 1 kPa [31]. It is possible that similar ordered structures, to those in agarose, exist in gels of hydrogen-bonding polymers, such as hydroxyethylmethacrylate and vinylpyrrolidone, which do seem to be stronger than less polar synthetic gels. One area where there has been a vigorous search for improved mechanical properties is in gels for contact lenses but there is no clear picture of what determines strength [32, 33]. Contact lenses have water contents of 30–50 % and strengths of 2–4 MPa [34]. Tests on vinylpyrrolidone gels with low water contents (30–40 %) gave strengths up to 2 MPa, which is in the range of cartilage and so could be considered adequate for construction of equipment [35]. On the other hand, contact lens gels made from mixed acrylic and vinylpyrrolidone monomers with 40–70 % water content have strengths from 100–600 kPa [36]. There is no simple relationship between polymer structure or water content and gel strength but gels based on vinylpyrrolidone do tend to be stronger. Work on cross-linked acrylic acid gels for microfluidics showed similar strengths, with a dramatic decrease as the gel was swollen at high pH [37]. A cross-linked copolymer gel of hydrophilic and hydrophobic segments was reported to have a strength of 200–500 kPa [38]. 1.2.5

Multi-Phase Gels

Many gels are two-phase composite systems. Polyacrylamide gels are often quite turbid, suggesting phase separation into polymer-rich and polymer-poor regions. The cross-linked structure prevents large-scale separation, so unambiguous evidence for two phases is hard to obtain. Crystallizable synthetic polymers form solvent-containing gels, which apparently contain crystallites connected by segments of solubilized polymer chain. Similar combinations of regions of nanoscale order linked by disordered solution probably characterize many biological gels, such as gelatin, agarose and calcium alginate. In principle the phase behavior of a lightly cross-linked gel would be expected to be the same as that for a high molecular weight sample of the same polymer in the same solvent. Heavier cross-linking would restrict the entropy of the chain and might induce phase separation. The search for an artificial cartilage material has long driven the search for strong gels. Various multi-phase systems have been found that are much better than simple gels but, until recent unexpected results on ‘double network’ gels, none have been strong enough to look really promising. It is has been known for some time that the properties of poly(vinyl alcohol) and mixed polyacrylic acid/poly(vinyl alcohol) gels can be enhanced by a series of freeze–thaw cycles that drive more extensive aggregation of the polymer [39]. Exactly what happens is unclear but growth of ice crystals will probably concentrate the polymer to the crystal boundaries and drive formation of insoluble hydrogen-bonded complexes of the polymers [40]. Addition of DMSO as a co-solvent enhances the gel strength, possibly by limiting ice crystal size. Early work on two-phase freeze–thaw modified neutral gels of poly(vinyl alcohol) mixed with cationic and anionic polymers found a strength of 1 MPa at 85 % water [41]. More recently, such poly(vinyl alcohol) gels with water contents of around 80 % were found to fail in compression at a few MPa [42]. This freeze–thaw process produces a two-phase composite structure which has recently been studied in more detail [43, 44]. Composite gels with poly(vinyl alcohol) and other water-soluble polymers have also been studied [45].


13

There have been many recent studies of composite gels made by irradiation of mixed solutions of polymers and increases in strength have been reported when compared to single polymer gels [46]. It would be expected that the properties of these disordered systems would primarily follow the water content. A number of studies have considered reinforcement of gels with inorganic fibers or plates both added before gelation and grown in situ in the gel and a significant increase in modulus is certainly seen [47–49]. With exfoliated clays as reinforcement, moduli increased from 4 to 20 kPa as the clay was added and the tensile strength increased from 0.1 MPa to 0.3 MPa [50–52]. One very attractive approach, based on the analogy to collagen-reinforced biological gels, is to reinforce gels with fibrils of rigid-rod polymers [53]. This particular system does show a significant increase in modulus but from very low values and no strength data was given. Thus the full potential of composites of this type has yet to be fully explored. A simple variant of this approach is to reinforce a gel with a textile, such as non-woven polypropylene [54, 55]. 1.2.6

Double Network Gels

Gong et al. [56] have formed gels with compressive strengths up to 17 MPa at 90 % water content by forming an interpenetrating network of ionic and nonionic gels in a two-step process (Figure 1.1). This compares with a strength of 0.2 MPa in compression for the equivalent single-component gel. These gels are produced by forming a moderately tightly cross-linked network, then swelling this gel in a solution of a second monomer with a low ratio of cross-linking agent and carrying out a second polymerization. As a result of the high degree of swelling in the monomer solution, the first gel network is highly extended in the final product while the second network is relaxed. The weight fraction of the second network in the final gel is 10–20 times that of the first network. 20

PAMPS-PAAm DN gel

Stress (MPa)

15

10 PAAm gel PAMPS gel

5

0 0

20

40

60

80

100

Strain (%)

Figure 1.1 Stress–strain curves from DN gels show much higher strength than conventional gels (Reprinted with permission from Gong, J. P., Katsuyama, Y., Kurokawa, T. and Osada, Y. Double network hydrogels with extremely high mechanical strength, Advanced Materials, 15, 1155–58, Copyright (2003) Wiley-VCH Verlag GmbH).

14


Other gels with a more lightly cross-linked first network show yield and necking in tension with extensions to break over 10 (1000 % þ strain) and a strength of 0.3 MPa [57]. Using a tear test, a fracture energy of 300 J/m2 was measured, compared to 0.1–1 J/m2 for conventional gels [58]. Yield is characteristic of metals and semicrystalline polymers where molecular slipping sets in at high stresses. It is not normally seen in rubbers or gels where the permanent cross-linked network prevents slippage. While strength was greatly increased by the addition of the second component, the initial modulus was affected little when compared to a conventional gel at the same concentration. If these double gels are made with a linear polymer in place of the second network, the fracture strength and fracture energy rise steeply at a molecular weight over 106, where the second polymer becomes highly entangled and these entanglements can act as physical cross-links [59]. Cyclic loading tests do show hysteresis, with a loss of modulus after successive cycles to high strain. This implies that breakage or other irreversible loss of some highly strained cross-links is occurring [60]. Other workers have found similar enhanced strengths in poly(ethylene oxide)–polyacrylic acid double network gels [61, 62]. These gels have tensile strengths which range up to 12 MPa depending on composition and swelling. Molecular dynamics simulations have been carried out for these PEO–PAA gels and show that the elastic modulus rises suddenly at strains of about 100 % where the first network becomes fully stretched [63]. This combination of high stress with a large strain to break increases the fracture toughness. Other interpenetrating networks, formed without the extension of the first network, have not shown similar improvements in strength. For instance HEMA–gelatin gels reach strengths of 65 kPa, just slightly above gelatin alone [64] and polyacrylamide/polyN-isopropylacrylamide IPN gels reach strengths of just 10 kPa [65]. Templating gels on a colloidal crystal and then removing the colloid has also been reported to give good toughness, although the modulus and strength remain low [66]. Reinforcing polyacrylamide gels with a rigid-chain polyelectrolyte does lead to a large increase in modulus [53]. Shull has recently emphasized that there is a need to develop better understanding of fracture toughness in these double network gels and in biological gels [67]. The preceding discussion also shows that microstructure control can lead to greatly enhanced strength and toughness in gels. Improvements obtained by double networks, by freeze–thaw and by fiber reinforcement suggest that there are many possible routes to better properties. A J-shaped stress–strain curve, that is an increase in elastic modulus at high strain, seems to be one signature of better toughness and strength. In this view, cartilage is a similar cross-linked network of collagen microfibers with a second network of coiled proteoglycan chains that can absorb fracture energy. Thus it seems that modulus and strength of networks can be separately controlled in order to design suitable mechanical performance into any functional matrix [68]. A useful objective for future work would be to develop some design rules. 1.2.7

Transport Properties

Many potential applications of active gels, as muscles, for drug release or as sensors, depend on their ability to respond to external influences by changing volume or shape or by taking up or releasing small molecules. If this responsiveness is not important, their mechanical properties can be duplicated by a range of dense elastomers and there is no


15

reason to employ a gel. Small molecule transport properties are thus crucial. Since drug release has been well discussed earlier [69], the focus here is on the faster diffusion times appropriate for sensing and actuation. Iit would be expected that diffusion coefficients of small solutes in gels should be intermediate between those in solution and in an elastomer. The diffusion coefficients of solutes in dilute gels have been measured and do not differ dramatically from those in solution [70]. Diffusion processes in gels can also be conveniently studied by conductivity [42]. For low levels of soluble small molecule additives in an elastomer, the diffusion coefficient can be estimated from the properties of the polymer and the size of the molecule [71]. At the other extreme, of a highly swollen hydrogel, the diffusion of water and of soluble compounds in the water have been shown to be reduced roughly in proportion to the water content of the gel. A significant difference will result when the diffusion process causes swelling or deswelling of the gel. The resulting nonuniform volume changes through the gel will result in highly non-Fickian behavior and may also cause significant internal stresses or fracture. Tanaka and co-workers [72] studied many of these interactions. As gel concentration changes in a single-phase gel, diffusion and solubility of a solute will change, such that the resulting changes in permeability can be quite complex. Two-phase gels will be even more complicated and so should be a source of many complex changes in release or uptake of solutes. In building a gel based machine, it will probably be equally desirable that the gel can swell and deswell over a range without dramatic changes in the properties of the surface region. This requirement imposes a coupling of effective sample size, diffusion rate and response time on any gel device. The self-diffusion coefficient of water is 2 10–5 cm2s–1 and diffusion times can be roughly estimated as x2/2D, where x is the effective half-thickness of the gel. This gives a response time for a gel one millimeter thick of about one minute, which would probably be the longest acceptable response time for many devices. It is clear that any responsive structure that is larger could have a finely divided spongelike structure that allows for fluids to flow in and out through channels rather than simply diffusing. If a hypothetical microstructural scale of 100 microns thickness is adopted, a diffusion coefficient of about 3 10–7 cm2s 1 can be accepted in a working device. 1.2.8

Drying

Engineering materials are generally dry and there is a reluctance to use materials that may lose liquid and dry out. However, we do work with systems that need to retain liquids, such as foods, cosmetics, paints and inks. Our experience with houseplants suggests that it would be quite possible to develop long-lived systems that depend on occasional replenishment of a water reservoir if there were desirable and unique properties. This survey of recent work on gels focuses on whether it is possible to reach a combination of mechanical properties, stability and activity that would allow more use of gel devices and structures. Drying time is also clearly an issue in gel devices. The evaporation rate of water is very dependent on temperature, humidity and air flow but measurements on drying of snails give a rate of about 100 microns/h as typical for still air with an active snail

16


being able to reduce this by about 20-fold by maintaining a surface coating [73]. If some similar mechanism was available to gel devices with a size of about one centimeter, they would experience 10 % water loss in a week and so would need only occasional rehydration.

1.3 Chemical and Physical Formation of Gels While linear polymers can be purchased in bulk and processed in the melt or solution to the desired form, cross-linked gels must be chemically formed in situ. Gel actuators will have to be built as systems containing multiple layers or material, structured pores and electrodes. Processing methods that lend themselves to building these systems must be available. Many gels, are chemically cross-linked like vulcanized rubbers and this limits our ability to process shape them. Just as thermoplastic elastomers have allowed an expansion of the range of applications of rubbery materials in complex shapes, so meltable gels, like agarose, can be more readily shaped than chemically cross-linked gels. Most of the work on synthetic gels uses gels formed by free radical polymerization of the families of hydrophilic acrylate, methacrylate and acrylamide monomers plus vinylpyrrolidone. While these methods give a very versatile family of hydrogels, it is worth noting that the polymers are atactic or otherwise irregular and this limits formation of any microstructure that might give rise to toughness. Also, free radical polymerization is oxygen sensitive and this can make it difficult to get good control of the polymerization in small or air-exposed devices. For these reasons, it may be valuable to explore other approaches to forming synthetic hydrogels. The kinetics of free radical linear polymerizations has been thoroughly studied [74] and the relationships between molecular weight distribution and polymerization conditions are well known. Gels are made by incorporating a small fraction of bi-functional or multifunctional monomers that becomes part of more than one kinetic chain so that a network forms. The statistics of network formation are also well known. Two factors act against being able to make reproducible samples of muscles. These reactions are very sensitive to small amounts of impurities as well as oxygen, and these can be difficult to control in small samples. In addition, the structure of a network is much more difficult to characterize in detail than is a linear polymer. As a result, it is hard to know what is the real structure of a gel sample and usually they are not characterized to any significant extent. Many gel samples are formed by UV irradiation. Quite reproducible samples can be formed by irradiation of de-oxygenated samples held between glass plates under conditions where the UV is only slightly attenuated as it passes through the sample (Figure 1.2). On the other hand, samples made in air under strong UV will tend to have structures that change properties through the thickness and many residual unpolymerized chain ends in the network. The reaction kinetics will depend on the water content of the monomer solution during polymerization. This, in turn, will affect the network structure and so the swellability of the final gel network. Many studies of gel actuators use gels with low cross-link densities that show very large equilibrium swelling. To get good mechanical properties, it is preferable to work with gels that are more like contact lenses and only swell to become about 50 % water at equilibrium.


17

Figure 1.2 Gel microcantilever in water before (a) and after (b) UV irradiation (Reprinted with permission from Watanabe, T., Akiyama, M., Totani, K. et al. Photoresponsive hydrogel microstructure fabricated by two-photon initiated polymerization, Advanced Functional Materials, 12, 9, 611–4, Copyright (2002) Wiley-VCH Verlag GmbH).

To avoid the oxygen-sensitivity problem, Yoshioka and Calvert studied epoxy hydrogels for small artificial muscles and sensors (Figure 1.3) [75, 76]. Water-soluble polymers and hydrogels have also been made by ring-opening metathesis polymerization [77–80]. It would also be expected that it is possible to form stable hydrogels based on polyamides and other polymers formed by condensation chemistry that might form tough microstructures. 1.8

Thickness Changes (h/ho)

1.7 Gel 4

1.6 1.5 1.4 1.3 Gel 3

1.2 1.1 1 0

120

240

360

480

600 720 840 Time (s)

960 1080 1200 1320

Figure 1.3 Swelling–collapse behavior of two epoxy gels in 0.01 M NaCl at pH 4.3; gel dots are printed on the platinum plate, which contacted the anode side. Average dot diameter (d0) was approximately 460 mm and the initial thickness (h0) was 130 mm. Applied voltage was 6.0 V and current was 3.6 mA (Reprinted from dissertation work of Y. Yoshioka, Courtesy of University of Arizona).

18


It has long been know that many pairs of polymers self-assemble to form gels through interchain bonding. Gelled capsules can be formed by dripping a solution of cationic polymer into a solution of anionic polymer [81, 82] or by dripping alginate into calcium [83]. This process results in a thick-walled capsule with a nonuniform structure. There is no way of simply mixing the two components so that a uniform block of material is formed. Ionic selfassembly by sequential dipping into anionic and cationic polymers [84] or by repeated contact printing [85] does produce uniform thin films of ionic gels in a more controlled fashion. While these systems have been shown to have many potential applications, there has been little work on the structure and properties of the gels themselves. One structural study is from Lewis et al. [86] where a fine capillary was used to make micron-scale 3D structures by extruding a stream of solution into water at a pH that induced gelation (Figure 1.4). Such writing systems could be used to make a wide range of gel microdevices.

Figure 1.4 Printed polyelectrolyte complex filaments, written with a micron-sized nozzle. Scale bar ¼ 10 mm (Reprinted with permission from Gratson, G. M., Xu, M and Lewis, J. A. Microperiodic structures: Direct writing of three-dimensional webs, Nature, 428, 386, Copyright (2004) Nature Publishing Group).

Natural proteins form structures by a combination of ionic interactions, binding to multivalent cations and hydrogen bonding. It is possible to design synthetic proteins to form similar structures [87–89] and demonstrate that they gel. Thus there would seem to be considerable opportunities for better characterization of gelation of synthetic ionic


19

polymers and for the study of more structured synthetic gel systems formed by other polymerization chemistries. This should lead to a better understanding of the whole structure–mechanical properties map for gels.

1.4 Actuation Methods The merit of a muscle-like actuator depends on a fast response time, a high contractile force and a large actuation strain. The state of swelling of a gel is dependent on temperature, solvation and pH. These factors may in turn be influenced by the application of chemical changes, light, heat, electrical and magnetic fields. The most desirable way to control a gel actuator is via an electric field but electrically driven actuators have so far produced little force. The simplest form of control, for understanding the response, is to change the temperature. 1.4.1

Thermally Driven Gel Actuators

Polymer chain coils in solution will undergo conformations changes as the temperature changes [90]. In many cases, the coil contracts as the temperature is raised until the solution separates into polymer-rich and dilute phases at the lower critical solution temperature (LCST). A homopolymer gel with a low density of cross-links is thermodynamically similar to a solution of very high molecular weight polymer. Thus, the gel will contract under conditions where the linear polymer coils contract. Tanaka studied this process in gels using the concepts of phase transitions and distinguishing between gels which show a continuous change from the expanded to the contracted state and those which go through a phase separation [91]. Poly-N-isopropylacrylamide (NIPAM) solution in water precipitates above 40 °C, so NIPAM gels have provided the most fruitful example of a thermally driven phase change. A swollen cross-linked NIPAM gel deswells at the same temperature, 40 °C. Hompolymer gels show a discontinuous phase transition while copolymer gels with ionizable groups such as acrylic acid can show a continuous transition or critical point behavior. Having resolved the thermodynamics of the system, the kinetics of the phase change become important. The kinetics of gel swelling can be treated as a two-step process. Solvent diffuses into the gel causing some regions to swell and then there is an instantaneous shape change to minimize the elastic energy between the swollen and unswollen regions. For swelling involving small changes in volume in the continuous region of the phase diagram, the kinetics and dependence on the gel shape correspond to those expected for diffusion of solvent. Thus the response time of a long acrylamide gel cylinder 1.3 mm in diameter was about one hour, in agreement with a theory based on the coupling of the diffusion coefficient of water in the gel and the shear modulus of the gel [92]. In the two-phase region, or where there is a large volume change, the changes in diffusion coefficient and gel properties with the local state of swelling will lead to complex kinetics which depend on the details of the initial and final gel states. A number of workers have addressed the diffusion time by making the gels porous, so that simple diffusion only occurs over a small distance. The effective diffusion rate can be

20


linked to the degree of porosity [93]. For instance, polymerizing the gels under reduced pressure produces a macroporous gel that responds to temperature changes in a few minutes, about 10 times faster than normal gels [94]. After freeze drying this porous gel had an apparent pore size of 20 microns although the actual structure prior to drying was not studied. Viewed as an actuator, these porous systems have the disadvantage that the forces developed will also be reduced as the porosity increases. Polymerizing gels in the presence of poly(ethylene oxide) also yields porous, fast-responding gels [95], as does polymerization of gels on micron-scale liquid templates [96]. A related, second approach to increasing response time is to prepare a two-phase gel so that a nonresponsive matrix can allow fluid flow into and out from the gel. This has been shown for solutions of linear NIPAM as a block or graft copolymer with poly(ethylene oxide) [97]. The collapsed, precipitated state of the graft copolymer is more open than pure NIPAM and so allows more rapid water penetration and redissolution when the temperature is decreased. This type of approach addresses the problem that a rapidly shrinking gel tends to form a dense skin that inhibits water loss from the interior, both slowing the volume change and possibly leading to fracture of the gel under the resulting shrinkage stresses. The effect of such skin formation on deswelling kinetics has been studied by Hirose and Shibayama, who showed that pure NIPAM gels form a dense layer and shrink much more slowly than weakly charged copolymers of NIPAM and acrylic acid that retain more mobility for water in the collapsed state [98]. NIPAM has been most widely discussed but other aqueous gels are also thermally responsive, including poly N-vinylcaprolactam (PVCL) and hydroxypropylcellulose (HPC) [99]. In the unswollen (high temperature) state, PVCL is much stiffer than in the swollen state and resembles a glassy polymer, whereas in HPC the modulus decreases as it shrinks. These thermally driven gels have been widely studied with a view to use as actuators or in drug delivery. While the volume change is large, the response is slow, primarily because water must diffuse into and out from the gel. In addition, rapid heating may be quite easy to achieve but it can be difficult to cool the gel rapidly in any practical fashion. The stress developed on shrinkage can be estimated from a knowledge of the relationship between elastic modulus and degree of swelling, as discussed below. 1.4.2

Chemically Driven Gel Actuators

While electrical actuation would be the most practical method to control and drive gel actuators, gels primarily respond to the chemistry of their environment and so chemical actuation has proved to be the most efficient method to drive gel actuators. Muscle is a chemically driven actuator where the energy of the ATP to ADP conversion is used to drive a cyclic shape change in myosin that causes it to ‘walk’ along actin filaments. From the viewpoint of building an artificial muscle, it is important to note the points that: i. the energy to drive this process in the short term is stored locally in the cells as glucose and oxygen; ii. muscle develops force only while it is burning energy and will passively extend once the cells cease to produce ATP. It does not move between two states as gels do. Thus the analogies between real muscle and chemically contracting gels are not as close as they might seem [3, 4, 100].


21

In most cases the source of chemical energy for gel actuators is either a change in pH or a change in solvent. Many gels are based on cross-linked polyacrylic acid. The linear polymer is very soluble in water at pH values above about five where it is in the ionized form as a sodium or other salt. At low pH, below four, the acid form predominates and is only slightly soluble in water. Repulsion effects between adjacent ions on the polymer chain cause the pH range over which the polymer goes from the acid to salt form to be much wider than for a simple acid such as acetic acid [90]. The chain is also much more extended due to Coulombic repulsions in the salt form. The chain extension is sensitive to the ionic strength of the aqueous solution, since other ions in solution will screen the repulsions and allow the chain to coil. In terms of the phase transition treatment of Tanaka discussed above [72, 101], both commonly used gel muscles, polyacrylic acid and polymethacrylic acid, are soluble in water at all pH values and so the acid to ionized change is a continuous phase transition in water. In mixed solvents the change can be discontinuous. The response of polyacrylic acid and polymethacrylic acid gels to changes in pH and solvent have been widely studied [102]. Other acid gels [103] and chitosan gels [104] have also been studied. However, the number of systems explored only represents a tiny sample of the potential range of chemically actuating gels. These two monomers can be easily copolymerized with a wide range of other monomers to form gels with different solubility characteristics. It can be expected that: addition of a more hydrophobic monomer will induce phase separation at low pH; addition of a strongly ionized monomer, such as a sulfonated monomer, will limit shrinkage at low pH; and addition of a neutral water soluble monomer, such as hydroxyethylmethacrylate, will limit the maximum swelling at high pH. These remarks apply to random copolymers. Graft or block copolymers could be made with a wide range of responses to pH, ionic strength or solvent. An amine-functional monomer such as aminoethylmethacrylate can be expected to respond as a mirror-image to the acidic gels, swelling at low pH and expanding at high pH. Yoshioka and Calvert have studied the response of epoxy–amino gels [75, 105]. Amine gels could also be formed with a wide range of comonomers. One interesting but littlestudied case is amphoteric gels containing both amino and acid groups which may be swollen in acid, in base or in salt solutions [106–108]. Other chemistries may also be used to reversibly change the properties of hydrogels. Thus, gels with reversible disulfide cross-links have been studied [109–111]. Reversible metal-ligand cross-links with a self-repairing function have also been described [112]. Suitable modification of these systems could also be used to form chemically driven actuators. In most studies, pH is adjusted by exchange with a mineral acid such as hydrochloric acid and sodium hydroxide. Ammonium and potassium counter-ions would be expected to act very similarly but lithium or divalent ions, such as calcium, will not result in such large solubility changes. Lightly cross-linked gels with more than about 20 monomer units between cross-links will be expected to act as a polymer of infinite molecular weight as it responds to the medium by expanding and contracting. Higher cross-link densities will decrease the tendency of the gel to swell, both by limiting the extent to which the chains can coil and uncoil but also by limiting the solubility of the chains. Gels with more acidic side groups, such as sulfonated gels, will be less responsive because very low pH would be needed to deionize the gel. Phosphated gels would be expected to show several ionization steps but have not been much studied. In addition to

22


acid gels, gels with amine side groups respond to pH. They show a mirror of the behavior of acids, contracting in strong base but ionizing and expanding in acid [75]. Gels will also expand and contract as a result of being moved between water and a watermiscible organic solvent such as acetone. The expansion and contraction of the gel reflects the solubilization and precipitation of the equivalent linear polymer in the solvents. The best performance of chemically driven gels is by polyacrylonitrile fibers which are partially converted by heat and alkali to cross-linked polyacrylic acid fibers. Previous work has shown that these can be made into muscle-like actuators that respond to pH changes with good force and strain characteristics [6, 113]. The good characteristics of this material arise because the spun fibers are oriented and so the swollen gels are stronger than unoriented gels. Orientation must also affect the actuation properties but this does not seem to have been studied. Also, the fibers have a small diameter and so respond rapidly to pH changes. Recent work has found that commercial poly(acrylonitrile) (PAN) yarns treated in this way could give actuator strains of about 80 % and stresses of 100 kPa to 1 MPa with a response time of about 10 seconds [114]. The same fibers can be driven electrically at 5 V, using an electrode embedded in or adjacent to the fiber, to produce similar stress but the response time is much longer, about 10 minutes [113, 114]. Electrospinning has been used to make fibers of less than one micron diameter that could then be twisted into yarns [115]. Actuator strains of 38 % were obtained in cycling from pH 1 to pH 12 with a response time of about five seconds. The authors claim an actuator stress of about 10 kPa but the data presented suggest this value is actually closer to 1 MPa. Other similar stiff synthetic gels have also developed higher forces in response to chemical activation [116]. Actuation that relies on pH changes has the advantage that hydrogen ions diffuse rapidly. In biology, changes in calcium ion concentration often drive changes in molecular conformation. Forisomes, plant proteins responsible for opening and closing of leaf pores, produce a force of 11 kPa in response to calcium and pH [117]. An actuator based on changes in copper ion concentration has been described [118]. As the copper was oxidized to Cu2+ and reduced back to copper metal, the gel contracted and swelled. The oxidation and reduction could be driven chemically or electrochemically. 1.4.3

Gels Driven by Oscillating Reactions

Several groups have recently studied self-oscillating gels that swell and contract cyclically in response to an oscillating chemical reaction in the medium. Yoshida and co-workers have demonstrated patterns of contraction in NIPAM gels that cycle through the lower critical solution temperature (LCST) as the gel cycles between two oxidation states [119]. Yashin and Balasz have modeled the moving patterns of expansion and contraction set up in these gels [120]. Jones and co-workers have used oscillating pH reactions in the liquid surrounding an acidic block copolymer to drive a gel motor [121, 122]. However, their calculations of power density for these gels do give very low values. This approach does suggest that it might be possible to use an electrical signal to set off such a chemically driven oscillation. This would be a good mimic of natural muscle, where an electrical impulse triggers chemically driven contraction of the muscle.


1.4.4

23

Light Actuated Gels

With the availability of powerful LEDs and solid state lasers, light would seem to be an excellent way of communicating with gels in order to drive actuators or read out from sensors. Light can be used to heat an absorbing gel and cause a shape or volume change. Light could also be used to change pH in order to produce a volume change in a gel. Gels can also be made with light-sensitive groups that undergo a chemical conversion that results in a shape change. So far none of these systems has shown a strong enough response to be attractive as an actuator. Light can also be used as a heat source to swell and deswell thermally-sensitive gels. Suzuki and co-workers showed that a NIPAM gel containing copper chlorophyllin, a soluble dye derived from chlorophyll, can be induced to swell and contact as the light is turned on and off [123]. A copolymer of NIPAM with sodium polyacrylate showed similar behavior but with a large hysteresis, such that the gel could be switched to the unswollen state at high intensity but would remain in either the swollen or unswollen state over an intermediate intensity range [124]. Similar bi-stable behavior was shown with pH or temperature. In principle, this type of bi-stable switching would be of interest for many applications. The bi-stability was analyzed in terms of a Landau model for phase transitions and might also be viewed as relating to skin effects. In lieu of a dissolved dye, a similar response to heating from a near-IR laser was obtained in NIPAM gels containing dispersed gold nanorods [125]. Beebe and co-workers have produced gels containing gold nanoparticles that respond by swelling on absorption of selected wavelengths of light and so act as valves to open or close channels in a microfluidics system [126]. Similar gold-hydrogel composites also show a change to becoming electrically conducting as the collapse [127]. The photoisomerization of azobenzene between the cis and trans forms is well known and gels containing attached azobenzene units have been shown to respond to UV illumination by stiffening [128]. Leucocyanides are photoresponsive dyes that convert between ionized and unionized forms on irradiation with light and cause osmotic swelling and contraction of a gel [129]. Marder and co-workers developed a hydrogel that responds to UV irradiation with a keto to enol tautomerization that results in mechanical deflection of a cantilever [130]. Nitrocinnamate chemistry has been used to create a hydrogel which can be reversibly photocross-linked and photocleaved [131]. A copper cross-linked polyacrylic acid gel, containing titanium dioxide particles, has been shown to swell under UV light and contract again in the dark [132]. A similar system with silver-coated titanium dioxide has also been described [133]. The volume change is large but the response time is hundreds of minutes. 1.4.5

Electrically Driven Gel Actuators

A recent review [134] of electroactive gels focused particularly on polyelectrolyte gels based on cross-linked polymers of acrylic acid. Such acidic hydrogels can show a large volume change in water as the pH changes from acidic to basic and the gel becomes ionized. A pair of electrodes in a suitable salt solution will produce hydrogen at the cathode and oxygen at the anode. This results in decreasing the pH at the cathode and increasing pH at the anode. A pH-sensitive acid gel will thus shrink if it is attached to the cathode in a

24


75

0.2

50

0.1

25

Strain

0.3

0

5 Time (min)

10

Weight gain (%)

solution at neutral pH. Depending on whether one or both electrodes are embedded in the gel or are in the surrounding solution, a gel may shrink, expand or bend [135, 136]. Several groups have demonstrated actuation by these gels, while immersed in solution, with embedded or external electrodes. Osada and co-workers made gel ‘loopers’ that would crawl along a bar between electrodes [137, 138]. Shiga and co-workers carried out an extensive series of studies on poly(vinyl alcohol)– polyacrylic acid gel actuators driven electrically (Figure 1.5) [136, 139–143]. These gels are prepared by a freeze–thaw method that produces a strong and highly porous structure. A ‘fish’ was made that would swim along a channel between external electrodes (Figure 1.6) [39]. Although the forces are small, such gels can be used to transport ‘cargo’ as they expand and contract in a tube [144].

0

Figure 1.5 Strain in bending and weight gain of poly(vinyl alcohol)–poly(sodium acrylate) composite hydrogel under an electric field of 10 Vcm–1 (Reprinted with permission from Shiga, T., Hirose, Y., Okada, A. and Kurauchi, T. Bending of poly(vinyl alcohol)-poly(sodium acrylate) composite hydrogel in electric fields Journal of Applied Polymer Science, 44, 249–53, Copyright (1992) Wiley-VCH Verlag GmbH).

Natural muscle undergoes a linear contraction. A single-component gel placed symmetrically between electrodes will bend as one side expands and one contracts. A combination of layers of two different gels that expands and contracts in a linear fashion when the cathode is embedded in one layer and the anode in the other has been demonstrated [145]. This can be achieved by having gels of opposite ionic charge or by having a high modulus contractile gel layered onto a soft neutral gel. In this second case, as the stiff gel contracts the water is pushed into the neutral, soft gel and this expand in thickness. Almost all gel actuators work in solution or with embedded electrodes but gels have been demonstrated which respond to large electrical fields in air [146]. The gels showed a crawling motion driven by electrostatic response of the gels to the applied field.


25

Figure 1.6 An ‘eel’ gel of PAMPS moving in an oscillating electric field (Reprinted with permission from Osida, Y., Okuzaki, H. and Hori, H. A polymer gel with electrically driven motility, Nature, 355, 6357, 242–4, Copyright (1992) Nature Publishing Group).

1.4.6

Electro- and Magneto-Rheological Composites

Recently there has been much interest in composites of magnetic particles in a soft matrix which respond to a magnetic field by a change of shape and properties [147]. These materials are the elastomeric equivalents of magneto-rheological fluids and electrorheological fluids. The prototypical system is a dispersion of 25 % of micron-sized iron particles in gel of silicone rubber and silicone oil but polyurethanes and other rubbers have been investigated. When cured in a magnetic field the particles form chains [148]. After curing, the stiffness of the material increases in a magnetic field. At a strain of 5 %, an increase in stress of about 50 % is seen in a field of 123 kA/m (1500 Oe). In a study of magnetostriction, a field of 800 kA/m (1 Tesla) produced a strain of 0.3 % in a sample preloaded in compression to 100 kPa, a strain comparable to magnetostrictive alloys (Terfenol-D) [149]. These materials also show an electric field response if they can be formed to be nonconductive [150]. A shear stress of a few kPa can be produced at a field of 1 MV/m. A study on barium ferrite particles in carrageenan hydrogels showed a modulus decrease of about 75 % (from 20 MPa) on magnetization at 800 KA/m [151]. Strains of about 0.2 % could be produced in these gels. This effect is clearly not the same as that seen in the ironfilled elastomers since it is larger, in the opposite direction and irreversible.

26

1.4.7


LC Elastomers

Liquid crystalline (LC) elastomers show a contraction on heating from the nematic to isotropic phase and so can be pictured as actuators [152, 153]. Similar behavior can also be seen in liquid crystalline gels [154], which opens the possibility for combining electrical switchability, the shape change associated with liquid crystalline elastomers with the volume change associated with gels [155, 156]. Similar changes have been demonstrated in hydrogels [157]. These systems can deliver rapid large shape changes, stresses of 130 kPa have been reported for thermal actuation [158], which is enough to be useful as actuators.

1.5 Performance of Gels as Actuators The performance of muscle has been measured well on scallop [159]. Muscle is characterized by a response time of less than one second. The maximum actuator stress is about 300 kPa at a mid-point of contraction, corresponding to about 12 % shortening. The maximum strain is 25 %. The energy density of muscle is about 50 kJ.m–3. The power output is 200 W/kg peak or 50 W/kg for sustained cyclical contraction. Reaching or exceeding these characteristics is now the goal for an artificial muscle system. Unlike many mechanical systems, muscle uses power as long as stress is needed to hold a static position. As a result, the normal mechanical concept of efficiency is hard to apply. For instance, the position of the arm may be held by balanced force between a muscle pair trying to extend and contract the elbow. No motion occurs, but considerable energy is expended in this isometric exercise. As a result, discussions of efficiency can be misleading, since the conditions of measurement need to be well defined. The actin–myosin system is essentially the only type of actuator in animals but there are a few other examples of muscle-like tissues. The jelly bodies of jellyfish, sea anenomes and other sea animals contain proteoglycan hydrogel, elastic fibers, collagen and muscle fibers in various arrangements [160–162]. The sub-micron reinforcing fibers are loosely connected to the matrix and so provide strength at high extensions and slow the viscoelastic response of the tissue [8]. In sea cucumbers, starfish and other echinoderms, the mechanical properties of the gel can be quickly changed from soft to hard. This was thought to occur by release of calcium ions to harden the polyanionic matrix but actually seems to be due to the release of proteins that temporarily bond to and cross-link the collagen fibrillar network [163]. This response is essentially a change in elastic modulus, as opposed to an active contraction such as is characteristic of muscle. In legumes, specialized cells open and close to control fluid flow in the vascular system of the plant. These ‘forisomes’, which are 30 microns in diameter, contract in response to calcium ions in about 50 milliseconds and develop a force of 11 kPa [117]. The available energy density has been estimated at 0.5 MJ.m–3, which is close to what would be needed for an artificial muscle [164], and preliminary tests have been made of forisomes in a microfluidic system [165]. It may be that this system will provide a better model for artificial muscle than the much more complex actin–myosin system. The target performance characteristics for a biomimetic muscle would be a response time of about one second, an actuator strain of about 25 % and a developed stress of about


27

300 kPa. In the absence of any current good answers, any approximation to these values would be of interest. At present, the only systems that approach this are the converted PAN fibers acting as chemical actuators. The actuation stress for a gel of known elastic modulus can readily be estimated. Consider, for example, a gel that initially contains about 50 % water. This is a much lower water content than most experimental hydrogels but is similar to a contact lens formulation and would have an elastic modulus in the region of 1 MPa [36]. Assume the gel will act as an actuator in compression, by pushing a weight upwards. A change in thickness of 25 % would correspond to a volume change of 2, assuming width, length and thickness change equally. To develop a substantial force at this new thickness, the volume change would have to be greater, for instance to 3.4, which corresponds to an unconstrained thickness change of 50 %. Using the equation for modulus and swelling above, the elastic modulus will have dropped to about 660 kPa at this swelling. When this swollen gel deforms elastically back to a thickness of 25 % greater than the original value, the stress generated will be 165 kPa. In contrast, more lightly cross-linked gels, with water contents of 90 % or more, have moduli of a few kPa and can exert little force in this format [166]. This then shows that an elastic gel, acting as an actuator in tension or compression, can only generate a substantial force if the elastic modulus is quite high and the water content relatively low. The swelling pressure exerted by these dilute gels can be much higher than the modulus [20, 166]. This can be up to hundreds of kPa at low degrees of swelling but drops rapidly as the polymer swells toward equilibrium. Ionic gels show higher swelling pressures under conditions where they are charged [21]. To use this to drive an actuator, the gel would need to be confined, for instance in a rigid cylinder with porous walls and a moveable piston. It should also be possible to exploit the swelling pressure in a suitable composite structure or oriented structure where swelling is free along one axis but is very constrained in the other two directions. Some of the marine animal structures, with reinforcing fibers wound spirally at an angle to one axis, may fulfill this requirement by only allowing expansion along one axis. These preceding arguments apply however the change in gel swelling is brought about, whether through thermal, chemical or electrical energy input. Thermal activation provides a convenient method for driving the actuator, which can clearly be engineered by a number of different routes. Cooling will be very slow for actuation occurring at 40 °C but a higher temperature could allow both rapid heating and cooling. Nonetheless, the associated systems for removing heat could add substantially to the weight and volume. It is also not clear that the phase changes necessary for thermal actuation can occur in gels at the high solids loadings needed to produce a significant stress. Hinkley et al. report an estimated maximum work achievable by the thermally driven PVCL gels of 1 MJ.m–3 of dry gel, considerably higher than that for poly(vinyl alcohol) gels and comparable with muscle [99]. However, the response is very slow. This maximum work reflects the higher mechanical strength of these gels. In many other gels, the mechanical work done by thermal actuation is small because the gels have a low elastic modulus and low strength. A measure of the efficiency of these gels as actuators or transducers is the mechanical energy density and the power density, the rate at which energy can be delivered. Gels driven by solvent-induced contraction have achieved 135 J/kg but only 2 W/kg, which compares with 70 J/kg and 100–200 W/kg for muscle that is in turn similar to the energy storage density of a lithium battery (Figure 1.7) [167].


wv [Nmm/cm3]

28

160

150/15

140

125/40

120

130/15

100 80 60 40 20 0 0

100

200

300

400

500

σ [kPa] Figure 1.7 Working energy of poly(vinyl alcohol)/poly(acrylic acid) hydrogels, cross-linked under different conditions of temperature/time, versus applied stress (Reprinted with permission from Arndt, K., Richter, A., Ludwig, S. et al. Poly(vinyl alcohol)/poly(acrylic acid) hydrogels: FT-IR spectroscopic characterization of crosslinking reaction and work at transition point, Acta Polymerica, 50, 383–90, Copyright (1999) Wiley-VCH Verlag GmbH).

Chemical actuation, as discussed above, can give a response with acceptable force, strain and speed parameters but has the drawback that delivering the fluid reagents requires inconvenient amounts of associated pumping and piping. In this context the human body can be viewed as a large amount of muscle with a roughly equal amount of piping to provide the driving energy and a supporting skeleton. Electrical actuation may be much more convenient in principle but brings a further set of engineering problems. The current flow necessary to ensure expansion or contraction of a gel can be estimated on the basis of the number of ions that must be inserted or removed to convert from the acid to the ionized state, of a carboxylate gel for instance. One gram of acrylic acid gel at 50 % by weight of water, contains 7 mM of acid groups, which corresponds to 670 Coulombs or 11 amperes for one minute. Any practical system will thus demand substantial currents, which in turn places requirements on the conductivity of the surrounding liquid. If the associated electrode reactions also produce gas, there will 75 cm3 of gas produced per gramme of gel switched. The previous estimate assumes that all the ions generated at the anode react with the surrounding gel. This is probably true if the electrodes are far apart, so that the acid generated at the anode and the base generated at the cathode do not diffuse but essentially remain in separate compartments. However, this wide separation will add to the volume and weight of the system and reduce the current flow at any given potential difference. A further design problem is that the electrodes themselves must be flexible enough to expand and contract without significantly restraining the volume changes of the gel. This can be achieved with flexible wire electrodes or some other type of soft electrode system. Thus, in the two-gel multi-layer system made by Liu and Calvert [145], the performance is still constrained by current flow and the ability of the gels to contract and expand over


29

the wire electrodes. Likewise, the performance of electrically driven PAN fibers is quite inferior to that of similar fibers chemically driven [113] and it is hard to retain stability of the electrodes as the fibers cycle. The electrical response of many soft actuators has been demonstrated as a bending beam. Typically, a strip of gel will bend in alternating directions as the potential difference across the gel is cycled. These thin sheets typically show a fast response and large deflection. However, the actual strain in the surfaces of the gel can be quite small. It is less obvious how to assess these bending actuators as a muscle-mimic but one measure is the blocking force, the force delivered at the end of the actuator, as a fraction of the weight of the actuator. Ionic Polymer–Metal Composites (IPMCs), for example, can deliver a blocking force of about 50 the weight of the sample [168]. Muscle in linear contraction would deliver about 300 its own weight. The elastic modulus of most experimental gel actuators is so low that the blocking force also will be very low. Earlier work showed that the response of gels in an electric field depends on the composition of the gel itself, whether the gel is touching the electrode and on the composition of the solution [135]. More recently a number of finite element models have been developed to calculate the response of a gel to an electric field (Figure 1.8) [169–171]. These have the advantage of being able to take into account all the competing effects and provide a prediction for the response of a gel actuator under any conditions.

Figure 1.8 Bending curvature of hydrogels in an applied field, effect of gel modulus [200] (Reprinted with permission from Li, H., Ng, T. Y., Yew, Y. K. and Lam, K. Y. Meshless Modeling of pH-Sensitive Hydrogels Subjected to Coupled pH and Electric Field Stimuli: Young Modulus Effects and Case Studies, Macromol. Chem. Phys., 208, 1137–46, Copyright (2007) Wiley-VCH Verlag GmbH).

30


It should be remembered that the chemical changes can both cause volume and stiffness changes, such that the results of combining pH changes and applied force can be unexpected [172]. This coupling between swelling thermodynamics and mechanical stress leads to a number of other peculiar phenomena, such as negative Poisson’s ratios [173–175], and may result in strange responses to complex loads, such as bending. Thus the challenge for gel actuators is to devise a practical solution to the whole system. Three possible approaches are:

Develop a chemical actuator with a compact and reliable supply of the needed chemical energy, not necessarily depending on acid and base.

Develop a thermal actuator based on expansion against a piston of a gel within a porous but rigid enclosure, similar to an automobile radiator thermostat.

Develop a two-compartment fine scale electrically driven gel muscle as an extension of the electrically-driven PAN fibers.

1.6 Applications of Electroactive Gels Recent papers have explored producing faster responses by inducing porosity into the gels. While the application as artificial muscles is not practical, low stress applications, such as sensors and valves, are possible [176]. Lenses and light modulators have been demonstrated recently [177, 178]. 1.6.1

Gel Valves and Pumps

Active gels have been developed as valves for microfluidic systems, where swelling of the gel in response to light, pH, thermal or electrical stimulation can be used to close a fluid channel or act as a pump [37, 126, 176, 179–181]. Electrically driven gels essentially respond to local changes in pH or ionic strength caused by electrolysis of the water [179, 182]. Bassetti et al. show that, at high electrical fields, there is a fast response due to ion migration within the gel and then a slower response due to pH changes in the solution as hydrogen ions are released from the anode. These effects were also seen by Shiga and co-workers [136]. In a flowing system, the hydrogen ions are swept away and so only the fast response to ion migration in the gel remains. Advantages of this type of application are that large forces are not needed and the response time could be several minutes. 1.6.2

Light Modulators

A series of papers have described light modulators based on the temperature driven shrinkage of colored N-isopropylacrylamide in a second gel matrix [178, 183, 184]. As the colored particles expand at lower temperature, they extract water from the surrounding matrix and occupy most of the volume, thus cutting off light transmission (Figure 1.9). Clearly, an electrically driven system based on temperature or pH change could be constructed. Thermally responsive systems have also been developed for use as optical modulators. Particles of pigmented NIPAM gels are embedded in a dilute host gel that is selected to


31

Figure 1.9 A gel-in-gel thermally driven light modulator at 20 °C (a, c) and 60 °C (b, d) (Reprinted by permission from Tsutsui, H., Mikami, M. and Akashi, R. All-polymer-gel light modulator consisting of a ‘gel-in-gel’ system, Adv. Mater., 16, 1925–9, Copyright (2004) Wiley-VCH Verlag GmbH).

allow ready exchange of water with the particles without opposing osmotic forces. The external gel must also be formed by cross-linking a polymer precursor in order to avoid forming an interpenetrating network with the gel particles. An external temperature increase results in a decrease in the absorbance of a thin sheet of the gel composite within one second as the particles shrink and the pigment becomes more localized [178, 183, 184]. 1.6.3

Gel Drug Delivery

Controlled drug release has been a subject of intense academic interest for many years. The bulk of commercial controlled release systems depend on pH sensitive release in the stomach or intestine. The matrix polymer may also dissolve or swell slowly to allow sustained release. Similar controlled-release systems may also be used to release flavors in food and rinks or to release fragrances during laundry or cleaning processes. Most such systems are essentially solid polymers because hydrogels would normally release the active small molecule too rapidly for most applications. This has changed in recent years as methods are developed to deliver much larger and sensitive protein and peptide drugs. These drugs also tend to be active at very low concentrations, so there is less need to produce high loadings of the drug in the matrix. Among the niche applications now being considered for hydrogels are the protection of protein drugs in the digestive system, adhesion of drug delivery patches to the mucus membranes and the protection of nanoparticulate drugs from the immune system [185].

32


The fact that the swelling and contraction of these systems can be driven electrically, as discussed above, means that there are potential applications for adhesive patches [186] or implanted gels with electrically-triggered release. The thermo-responsive gels have been suggested for drug eluting stents, and here too an electrical mechanism might be used to trigger the temperature change and release [187]. 1.6.4

Gel Sensors

Many biochemical sensors depend on enzymes immobilized at an electrode surface, such that an analyte reacts to produce a compound that is readily oxidized or reduced at the electrode. Hydrogels bound to the electrode surface can be used to immobilize the enzyme without deactivating it [188]. In this case the gel is acting as a passive matrix but there are also systems where the responsiveness of the gel drives the sensing process. Gel-coated silicon microcantilevers can be used to detect swelling in response to changes in pH or other species [69, 189, 190]. Changes in swelling can also be sensed by attaching the gel to a pressure sensor. Sensors can also be based on fluorescence changes in sensitive molecules in the gel [191], on volume changes of gel particles changing the diffraction angle from a colloidal crystal [192, 193], by color changes in a gel hologram [194, 195] and other responses [196]. Sensors have been made which depend directly on the electroactivity of the gel. Glucose sensors have been made based on the binding of glucose by phenylboronic acid attached to a gel, where the change in ionic conductivity of the gel is measured as the binding reaction changes the swelling of the gel [197, 198]. Although the biocompatibility of gels is not an electroactive property, it should be noted that they may be important for implantable sensors. Many sensors have been developed that work well on implantation but almost all surfaces become coated with proteins and lose sensitivity over days after implantation [199]. Many gels are known to resist protein binding and may be able to protect implanted sensors and electrodes. A colloidal crystal contains an ordered array of sub-micron particles that will diffract light at an angle that depends on the spacing of the particles, following Bragg’s law. If the array is in a gel matrix, any volume change by the matrix will change the particle spacing and the diffraction angle, and so can be used as the basis of an optical sensor. This effect has been studied by many groups since the original work by Holtz and Asher [193, 192, 189].

1.7 Conclusions In just over 40 years since Katchalsky’s demonstration of a collagen engine, useful gel actuators remain elusive. A number of recent developments offer hope that this problem will be resolved before too long. Good chemically driven actuators can be made although they are not very practical for machinery. There have been significant advances in the understanding of gel responses to electrical fields and in the design of gels with superior mechanical properties. The challenge now is to design an electrochemical system that provides a suitable reversible chemical change to drive a gel and a mechanical design that


33

subdivides the gel enough to give a rapid response. Problems such as this have been solved in the battery field many times, but with much greater inputs of researchers’ effort than have been applied to gels.

References 1. See Chapter 20, Dielectric Elastomer Actuators: Fundamentals. 2. Cohen, Y. B. (2004) Electroactive Polymer (EAP) Actuators as Artificial Muscles: Reality, Potential, and Challenges, 2nd edn, SPIE Press Monograph, Vol. PM136. 3. Pollack, G. H., Blyakhman, F. A., Reitz, F. B., et al. (2004) Natural Muscle as a Biological System, in Electroactive Polymer (EAP) Actuators as Artificial Muscles: Reality, Potential, and Challenges, 2nd edn. (ed. Bar-Cohen, Y.), SPIE Press Monograph, Vol. PM136. 4. Alexander, R. M. (2001) Natural Muscle in Encyclopedia of Materials: Science and Technology (ed. Buschow, K. H. J.), Elsevier Science Ltd, Oxford. 5. Steinberg, I. Z., Oplatka, A. and Katchalsky, A. (1966) Mechanochemical engines, Nature, 210, 568–71. 6. Umemoto, S., Okui, N. and Sakai, T. (1991) Contraction behavior of polyacrylonitrile gel fibers, in Polymer Gels (eds DeRossi, D., Kajiwara, K., Osada, Y. and Yamauchi, A.), Plenum, NY. 7. Osada, Y. and Hasebe, M. (1985) Electrically Activated Mechanochemical Devices Using Poly-Electrolyte Gels, Chemistry Letters, 9, 1285–8. 8. Wainwright, S. A., Biggs, W. D., Currey, J. D. and Gosline, J. M. (1986) Mechanical Design in Organisms, Princeton University Press, Princeton, NY. 9. Gosline, J. (1969) A Macromolecular Model Of Mesoglea Of A Sea Anemone, American Zoologist, 9, 1115. 10. Roth, V. and Mow, V. (1980) The intrinsic tensile behavior of the matrix of bovine articular cartilage and its variation with age, J. Bone Joint Surg. Am., 62, 1102–17. 11. Mow, V. C., Howell, D. S. and Buckwalter, J. A. (1990) Structure-Function Relationships of Articular Cartilage and the Effects of Joint Instability and Trauma on Cartilage Function, in Cartilage Changes in Osteoarthritis (ed. Brandt, K.D.), Indiana University Press, Bloomington, IN. 12. Ker, R. F. (1999) The Design Of Soft Collagenous Load-Bearing Tissues, The Journal of Experimental Biology, 202, 3315–24. 13. Bellucci, G. and Seedhom, B. B. (2001) Mechanical behavior of articular cartilage under tensile cyclic load, Rheumatology, 40, 1337–45. 14. Guo, B.-Y., Dong-hua Liao, Li, X.-Y., et al. (2007) Age and gender related changes in biomechanical properties of healthy human costal cartilage, Clinical Biomechanics, 22, 292–7. 15. Simha, N. K., Carlson, C. S. and Lewis, J. L. (2003) Evaluation of fracture toughness of cartilage by micropenetration, J. Mater. Sci. Mater. Med., 14, 631–9. 16. Meek, K. M. and Leonard, D. W. (1993) Ultrastructure of the corneal stroma: a comparative study, Biophysical Journal, 64, 273–80. 17. Zeng, Y., Yang, J., Huang, K., et al. (2001) A comparison of biomechanical properties between human and porcine cornea, Journal of Biomechanics, 34, 533–7. 18. Okay, O. and Durmaz, S. (2002) Charge density dependence of elastic modulus of strong polyelectrolyte hydrogels, Polymer, 43, 1215–21. 19. Huglin, M. B. and Rehab, M.M.A.-M. (1987) Mechanical and thermodynamic properties of butyl acrylate-N-vinylpyrrolidone hydrogels, Polymer, 28, 2200–6. 20. Dubrovskii, S. A. (1996) Compressional and Shear Behaviour of Weakly Ionic Polyacrylamide Gels, Polymer Gels and Networks, 4, 467–80.

34


21. Horkay, F., Han, M.-H., Han, I. S., et al. (2006) Separation of the effects of pH and polymer concentration on the swelling pressure and elastic modulus of a pH-responsive hydrogel, Polymer, 47, 7335–8. 22. Kinloch, A. and Young, R. (1983) Fracture Behavior of Polymers, Appied Science Publishers, London . 23. vanVliet, T. and Walstra, P. (1995) Large deformation and fracture behaviour of gels, Faraday Discussions, 101, 359–70. 24. Clark, A. and Ross-Murphy, S. (1987) Structural and mechanical-properties of bio-polymer gels, Advances in Polymer Science, 83, 57–192. 25. Bot, A., van Amerongen, I. A., Grot, R. D., et al. (1996) Large Deformation Rheology of Gelatin Gels, Polymer Gels and Networks, 4, 189–227. 26. Guenet, J.-M. and Rochas, C. (2006) Agarose Sols and Gels Revisited, Macromol. Symp., 242, 65–70. 27. Morris, E. R., Rees, D. A. and Young, G. (1982) Chiroptical characterisation of polysaccharide secondary structures in the presence of interfering chromophores: Chain conformation of interjunction sequences in Calcium Alginate gels, Carbohydrate Research, 108, 181–95. 28. Rees, D. (1981) Polysaccharide shapes and their interactions – some recent advances, Pure & Appl. Chem., 53, 1–14. 29. Webber, R. E. and Shull, K. R. (2004) Strain Dependence of the Viscoelastic Properties of Alginate Hydrogels, Macromolecules, 37, 6153–60. 30. Aymard, P., Martin, D. R., Plucknett, K., et al. (2001) Influence of Thermal History on the Structural and Mechanical Properties of Agarose Gels, Biopolymers, 59, 131–44. 31. McEvoy, H., Ross-Murphy, S. B. and Clark, A. H. (1985) Large deformation and ultimate properties of biopolymer gels: 1. Single biopolymer component systems, Polymer, 26, 1483–92. 32. Peppas, N. and Yang, W. (1981) Properties-based optimisation of the structure of polymers for contact lens applications., Contact Intraoc. Lens Med. J., 7, 300–314. 33. Tighe, B. (1976) The design of polymers for contact lens applications, Br. Polym. J., 8, 71–7. 34. Stammen, J., Williams, S., Ku, D. and Guldberg, R. (2001) Mechanical properties of a novel PVA hydrogel in shear and unconfined compression, Biomaterials, 22, 799–806. 35. Guven, O. and Sen, M. (1991) Preparation and characterization of poly(n-vinyl 2-pyrrolidone) hydrogels, Polymer, 32, 2491–5. 36. Tranoudis, I. and Efron, N. (2004) Tensile properties of soft contact lens materials, Contact Lens & Anterior Eye, 27, 177–91. 37. Johnson, B. D., Beebe, D. J. and Crone, W. C. (2004) Effects of swelling on the mechanical properties of a pH-sensitive hydrogel for use in microfluidic devices, Materials Science and Engineering C, 24, 575–81. 38. Isayeva, I., Gent, A. and Kennedy, J. (2002) Amphiphilic networks part XVIII – Synthesis and burst strength of water-swollen immunoisolatory tubules, J. Polym. Sci. Pol. Chem., 40, 2075–84. 39. Shiga, T., Hirose, Y., Okada, A. and Kurauchi, T. (1993) Bending of Ionic Polymer Gel caused by Swelling under Sinusoidally Varying Electric Fields, J. Appl. Polym. Sci., 47, 113–9. 40. Nugent, M. J. D. and Higginbotham, C. L. (2006) Investigation of the influence of freeze-thaw processing on the properties of polyvinyl alcohol/polyacrylic acid complexes, J. Mater. Sci., 41, 2393–404. 41. Suzuki, M. (1991) Amphoteric Poly(vinyl alcohol) Hydrogel and Electrodynamic Control Method for Artificial Muscles, in Polymer Gels (ed. deRossi, D.) Plenum, New York. 42. Sheppard, M., Lesho, R., Tucker, and SalehiHad, S. (1997) Electrical conductivity of pHresponsive hydrogels, J. Biomater,. Sci.-Polym,. Ed., 8, 349–62. 43. Matsumura, K., Hyon, S., Oka, M., et al. (1998) Scanning electron microscopy and atomic force microscopy observations of surface morphology for articular cartilages of dog’s knee and poly(vinyl alcohol) hydrogels, Kobunshi Ronbunshu, 55, 786–90.


35

44. Shapiro, Y. (1999) H-1 NMR self-diffusion study of morphology and structure of polyvinyl alcohol cryogels, J. Colloid Interface Sci., 212, 453–65. 45. Nho, Y. and Park, K. (2002) Preparation and properties of PVA/PVP hydrogels containing chitosan by radiation, J. Appl. Polym. Sci., 85, 1787–94. 46. Relleve, L., Yoshii, F., dela Rosa, A. and Kume, T. (1999) Radiation-modified hydrogel based on poly(N-vinyl-2-pyrrolidone) and carrageenan, Angew. Makromol. Chem., 273, 63–8. 47. Gao, D., Heimann, R., Williams, M., et al. (1999) Rheological properties of poly(acrylamide)bentonite composite hydrogels, J. Mat. Sci., 34, 1543–52. 48. Xia, X., Yih, J., D’Souza, N, and Hu, Z. (2003) Swelling and mechanical behavior of poly (N-isopropylacrylamide)/Na-montmorillonite layered silicates composite gels, Polymer, 44, 3389–3393. 49. Yamaguchi, I., Itoh, S., Suzuki, M., et al. (2003) The chitosan prepared from crab tendons: II. The chitosan/apatite composites and their application to nerve regeneration, Biomaterials, 24, 3285–92. 50. Zhu, M., Liu, Y., Sun, B., et al. (2006) A Novel Highly Resilient Nanocomposite Hydrogel with Low Hysteresis and Ultrahigh Elongation, Macromol. Rapid Commun., 27, 1023–8. 51. Zhang, W., Liu, Y., Zhu, M., et al. (2006) Surprising Conversion of Nanocomposite Hydrogels with High Mechanical Strength by Posttreatment: From a Low Swelling Ratio to an Ultrahigh Swelling Ratio, J. Polym. Sci.: Part A: Polym. Chem., 44, 6640–5. 52. Haraguchi, K. and Song, L. (2007) Microstructures Formed in Co-Cross-Linked Networks and Their Relationships to the Optical and Mechanical Properties of PNIPA/Clay Nanocomposite Gels, Macromolecules, 40, 5526–6. 53. Philippova, O., Rulkens, R., Kovtunenko, B., et al. (1998) Polyacrylamide hydrogels with trapped polyelectrolyte rods, Macromolecules, 31, 1168–79. 54. Wu, M., Bao, B., Chen, J., et al. (1999) Preparation of thermosensitive hydrogel (PP-g-NIPAAm) with one-off switching for controlled release of drugs, Radiat. Phys. Chem., 56, 341–6. 55. Lopergolo, L., Lugao, A. and Catalaini, L. (2002) Development of a poly(N-vinyl-2-pyrrolidone)/poly (ethylene glycol) hydrogel membrane reinforced with methyl methacrylate-grafted polypropylene fibers for possible use as wound dressing, J. Appl. Polym. Sci., 86, 662–6. 56. Gong, J. P., Katsuyama, Y., Kurokawa, T. and Osada, Y. (2003) Double network hydrogels with extremely high mechanical strength, Adv. Mater., 15, 1155–8. 57. Na, Y.-H., Tanaka, Y., Kawauchi, Y., et al. (2006) Necking Phenomenon of Double-Network Gels, Macromolecules, 39, 4641–5. 58. Tanaka, Y., Kuwabara, R., Na, Y.-H., et al. (2005) Determination of Fracture Energy of High Strength Double Network Hydrogels, J. Phys. Chem. B, 109, 11559–62. 59. Tsukeshiba, H., Huang, M., Na, Y.-H., et al. (2005) Effect of Polymer Entanglement on the Toughening of Double Network Hydrogels, J. Phys. Chem. B, 109, 16304–9. 60. Webber, R. E., Creton, C., Brown, H. R. and Gong, J. P. (2007) Large Strain Hysteresis and Mullins Effect of Tough Double-Network Hydrogels, Macromolecules, 40, 2919–27. 61. Myung, D., Koh, W., Ko, J., et al. (2005) Characterization of Poly(Ethylene Glycol)– Poly(Acrylic Acid) (PEG–PAA) Double Networks Designed for Corneal Implant Applications, Invest. Ophthalmol. Vis. Sci. (Suppl.), 46, 5003. 62. Myung, D., Koh, W., Ko, J., et al. (2007) Biomimetic strain hardening in interpenetrating polymer network hydrogels, Polymer, 48, 5376–87. 63. Jang, S. S., William, I., Goddard, A., et al. (2007) Mechanical and Transport Properties of the Poly(ethylene oxide)-Poly(acrylic acid) Double Network Hydrogel from Molecular Dynamic Simulations, J. Phys. Chem. B, 111, 1729–37. 64. Santin, M., Huang, S. J., Iannace, S., et al. (1996) Synthesis and characterization of a new interpenetrated poly(2-hydroxyethylmethacrylate)-gelatin composite polymer, Biomaterials, 17, 1459–67.

36


65. Djonlagic, J. and Petrovic, Z. S. (2004) Semi-Interpenetrating Polymer Networks Composed of Poly(N-isopropyl acrylamide) and Polyacrylamide Hydrogels, J. Polym. Sci. Part B: Polym. Phys., 42, 3987–99. 66. Stachowiak, A., Bershteyn, A., Tzatzalos, E. and Irvine, D. (2005) Bioactive gels with an ordered cellular structure combine interconnected macroporosity and robust mechanical properties, Adv. Mater., 17, 399–403. 67. Shull, K. R. (2006) Fracture and Adhesion of Elastomers and Gels: Large Strains at Small Length Scales, J. Polym. Sci.: Part B: Polym. Phys., 44, 3436–9. 68. Kong, H. J., Wong, E. and Mooney, D. J. (2003) Independent Control of Rigidity and Toughness of Polymeric Hydrogels, Macromolecules, 36, 4582–8. 69. Peppas, N. A., Hilt, J. Z., Khademhosseini, A. and Langer, R. (2006) Hydrogels in Biology and Medicine: From Molecular Principles to Bionanotechnology, Adv. Mat., 18, 1345–60. 70. Pavesi, L. and Rigamonti, A. (1995) Diffusion constants in polyacrylamide gels, Physical Review E, 51, 3318–23. 71. van Krevelen, D. W. (1976) Properties of Polymers, Elsevier, Amsterdam. 72. Shibayama, M. and Tanaka, T. (1993) Volume Phase-Transition And Related Phenomena Of Polymer Gels, Adv. Polym. Sci., 109, 1–62. 73. Machin, J. (1996) The evaporation of water from helix aspersa IV. Loss from the mantle of the inactive snail, J. Exp. Biol., 45, 269–78. 74. Odian, G. G. (1991) Principles of Polymerization, John Wiley & Sons, Inc., New York. 75. Yoshioka, Y. and Calvert, P. (2002) Epoxy-based electroactive polymer gels, Experimental Mechanics, 42, 404–8. 76. Heller, A., Maidan, R. and Wang, D. (1993) Amperometric biosensors based on three-dimensional hydrogel-forming epoxy networks, Sensors and Actuators B, 13-14, 180–3. 77. Rankin, D. and Lowe, A. (2006) Synthesis of polyelectrolytes via ring opening metathesis polymerization, Polyelectrolytes And Polyzwitterions: Synthesis, Properties, And Applications, ACS Symposium Series, 937, 117–28. 78. Bell, B., Hamilton, J., Law, E. and Rooney, J. (1994) A one-pot synthesis of comb polymers and hydrogels by the ring-opening metathesis polymerization reaction, Macromol. Rapid Comm., 15, 543–50. 79. Hamilton, J., Law, E. and Rooney, J. (1997) Metathesis polymerization as a route to water soluble polymers and hydrogels, J. Molecular Catalysis A – Chemical, 115, 1–9. 80. Feast, W., Cacialli, F., Daik, R., et al. (1999) Precision and control in polymer synthesis – Why it’s important and some recent examples of how to do it, Macromolecular Symposia, 143, 81–93. 81. Simsek-Ege, F., Bond, G. and Stringer, J. (2002) Matrix molecular weight cut-off for encapsulation of carbonic anhydrase in polyelectrolyte beads, J. Biomaterials Sci. – Polym. Edn, 13, 1175–87. 82. Simsek-Ege, F., Bond, G. and Stringer, J. (2003) Polyelectrolye complex formation between alginate and chitosan as a function of pH, J., Appl. Polym. Sci., 88, 346–51. 83. Sriamornsak, P. and Kennedy, R. (2006) A novel gel formation method, microstructure and mechanical properties of calcium polysaccharide gel films, Int. J. Pharm., 323, 72–80. 84. Bertrand, P., Jonas, A., Laschewsky, A. and Legras, R. (2000) Ultrathin polymer coatings by complexation of polyelectrolytes at interfaces: suitable materials, structure and properties, Macromol. Rapid Comm., 21, 319–48. 85. Xueping, J., Haipeng, Z., Shoshana, G. and Hammond, P. (2002) Polymer-on-Polymer Stamping: Universal Approaches to Chemically Patterned Surfaces, Langmuir, 18, 2607–15 . 86. Gratson, G. M., Xu, M. and Lewis, J. A. (2004) Direct writing of three dimensional webs, Nature, 428, 386. 87. Zhang, S. (2002) Emerging biological materials through molecular self-assembly, Biotechnol. Adv., 20, 321–39.


37

88. Zhao, X. and Zhang, S. (2007) Designer self-assembling peptide materials, Macromol. Biosci., 7, 13–22. 89. Zhao, X. and Zhang, S. (2006) Self-assembling nanopeptides become a new type of biomaterial, Advances In Polymer Science (Polymers For Regenerative Medicine), 145–70. 90. Tanford, C., (1961) Physical chemistry of macromolecules., John Wiley & Sons, Inc., New York. 91. Li, Y. and Tanaka, T. (1992) Phase transitions of gels, Ann. Rev. Mat. Sci., 22, 243–77. 92. Li, Y. and Tanaka, T. (1990) Kinetics of swelling and shrinking of gels, J. Chem. Phys., 92, 1365–71. 93. Ikkai, F. and Shibayama, M. (1996) Microstructure and Swelling Behavior of Ion-Exchange Resin, J. Polym. Sci.: Part B: Polym. Phys, 34, 1637–45. 94. Zhang, X., Wang, F. and Chu, C. (2003) Thermoresponsive Hydrogel With Rapid Response Dynamics, J. Mat. Sci. – Materials In Medicine, 14, 451–55. 95. Zhang, X. and Zhuo, R. (2000) Preparation of fast responsive, thermally sensitive poly (N-isopropylacrylamide) gel, Eur. Polym. J., 36, 2301–3. 96. Zhang, J.-T. and Jandt, K. D. (2008) A Novel Approach to Prepare Porous Poly(N-isopropylacrylamide) Hydrogel with Superfast Shrinking Kinetics, Macromol. Rapid Commun., 29, 593–7. 97. Van Durme, K., Van Assche, G., Aseyev, V., et al. (2007) Influence of macromolecular architecture on the thermal response rate of amphiphilic copolymers, based on poly(N-isopropylacrylamide) and poly(oxyethylene), in water, Macromolecules, 40, 3765–72. 98. Hirose, H. and Shibayama, M. (1998) Kinetics of Volume Phase Transition in Poly(N-isopropylacrylamide-co-acrylic acid) Gels, Macromolecules, 31, 5336–42. 99. Hinkley, J. A., Morgret, L. D. and Gehrke, S. H. (2004) Tensile properties of two responsive hydrogels, Polymer, 45, 8837–43. 100. King, A. M., Loiselle, D. S. and Kohl, P. (2004) Force Generation for Locomotion of Vertebrates: Skeletal Muscle Overview, IEEE J. Oceanic Eng., 29, 684–91. 101. Tanaka, T. (1979) Phase transitions in gels and in a single polymer, Polymer, 20, 1404–12. 102. Osada, Y. (1987) Conversion Of Chemical Into Mechanical Energy By Synthetic-Polymers (Chemomechanical Systems), Advances In Polymer Science, 82, 1–46. 103. Rodriguez, E. and Katime, I. (2003) Behavior of acrylic acid-itaconic acid hydrogels in swelling, shrinking, and uptakes of some metal ions from aqueous solution, J. App. Polym. Sci., 90, 530–6. 104. Goycoolea, F., Heras, A., Aranaz, I., et al. (2003) Effect of chemical cross-linking on the swelling and shrinking properties of thermal and pH-responsive chitosan hydrogels, Macromolecular Bioscience, 3, 612–9. 105. Yoshioka, Y. (2001) Dissertation, University of Arizona. 106. Tamagawa, H. and Taya, M. (2000) A theoretical prediction of the ions distribution in an amphoteric polymer gel, Mat. Sci. Eng. A, 285, 314–25. 107. Tamagawa, H., Nogata, F., Watanabe, T., et al. (2002) Reversible Hardness Variance as a Commonly Observable Phenomenon for Various Amphoteric Gels, JSME Int. J. Series A, 45, 579–84. 108. Zhang, Z., Chao, T. and Jiang, S. (2008) Physical, Chemical, and Chemical-Physical Double Network of Zwitterionic Hydrogels, J. Phys. Chem. B, 112, 5327–32. 109. Ravi, N., Wan, K., Swindle, K., et al. (2006) Development of techniques to compare mechanical properties of reversible hydrogels with spherical, square columnar and ocular lens geometry, Polymer, 47, 4203–9. 110. Nuttelman, C. R., Rice, M. A., Rydholm, A. E., et al. (2008) Macromolecular monomers for the synthesis of hydrogel niches and their application in cell encapsulation and tissue engineering, Prog. Polym. Sci., 33, 167–79. 111. Rydholm, A. E., Bowman, C. N. and Anseth, K. S. (2005) Degradable thiol-acrylate photopolymers: polymerization and degradation behavior of an in situ forming biomaterial, Biomaterials, 33, 4495–506.

38


112. Kersey, F. R., Loveless, D. M. and Craig, S. L. (2007) A hybrid polymer gel with controlled rates of cross-link rupture and self-repair, J. R. Soc. Interface, 4, 373–80. 113. Schreyer, H., Gebhart, N., Kim, K. and Shahinpoor, M. (2000) Electrical activation of artificial muscles containing polyacrylonitrile gel fibers, Biomacromolecules, 1, 642–7. 114. Choe, K. and Kim, K. J. (2006) Polyacrylonitrile linear actuators: Chemomechanical and electro-chemomechanical properties, Sensors and Actuators A, 126, 165–72. 115. Lee, D. Y., Kim, Y., Lee, S.-J., et al. (2007) Characteristics of chemo-mechanically driven polyacrylonitrile fiber gel actuators, Mat. Sci. Eng. C: Biomimetic Mat. Sens. Syst, 28, 13, 2200–6. 116. Brock, D., Lee, W., D. (1994) Segalman and W. Witkowski, A dynamic model of a linear actuator based on polymer hydrogel, J. Intelligent Mat. Systems and Structures, 5, 764–71. 117. Knoblauch, M., Noll, G. A., Mu¨ller, T. et al. (2003) ATP-Independent Contractile Proteins From Plants, Nature Mats, 2, 600–3. 118. Takada, K., Tanaka, N. and Tatsuma, T. (2005) A redox actuator based on reversible formation of bond between poly(acrylic acid) gel and Cu2+ ion, J. Electroanalytical Chem., 585, 120–7. 119. Yoshida, R., Takahashi, T., Yamaguchi, T. and Ichijo, H. (1996) Self-Oscillating Gel, J. Am. Chem. Soc., 118, 5134–5. 120. Yashin, V. V. and Balazs, A. C. (2007) Theoretical and computational modeling of selfoscillating polymer gels, J. Chem. Phys, 126, 124707 (online). 121. Topham, P. D., Howse, J. R., Crook, C. J., et al. (2007) Antagonistic Triblock Polymer Gels Powered by pH Oscillations, Macromolecules, 40, 4393–5. 122. Howse, J. R., Topham, P., Crook, C. J., et al. (2006) Reciprocating Power Generation in a Chemically Driven Synthetic Muscle, Nano Letters, 6, 73–7. 123. Suzuki, A. (1994) Light-Induced Phase-Transition of Poly(N-Isopropylacrylamide-CoChlorophyllin) Gels, J. Intelligent Mat. Systems and Structures, 5, 112–6. 124. Suzuki, A., Ishii, T. and Maruyama, Y. (1996) Optical switching in polymer gels, J. App. Phys., 80, 131–6. 125. Shiotani, A., Mori, T., Niidome, T., et al. (2007) Stable incorporation of gold nanorods into N-isopropylacrylamide hydrogels and their rapid shrinkage induced by near-infrared laser irradiation, Langmuir, 23, 4012–8. 126. Sershen, S., Mensing, G., Ng, M., et al. (2005) Independent optical control of microfluidic valves formed from optomechanically responsive nanocomposite hydrogels, Adv. Mat., 17, 1366–8. 127. Zhao, X., Ding, X., Deng, Z., et al. (2005) Thermo switchable electronic properties of a gold nanoparticle/hydrogel composite, Macromol. Rapid Commun., 26, 1784–7. 128. Hosono, N., Furukawa, H., Masubuchi, Y., et al. (2007) Photochemical control of network structure in gels and photo-induced changes in their viscoelastic properties, Colloids and Surfaces B: Biointerfaces, 56, 285–9. 129. Irie, M. (1986) Photoresponsive Polymers. Reversible Bending of Rod-Shaped Acrylamide Gels in an Electric Field, Macromolecules, 19, 2890–2. 130. Watanabe, T., Akiyama, M., Totani, K., et al. (2002) Photoresponsive hydrogel microstructure fabricated by two-photon initiated polymerization, Adv. Functional Mat., 12, 611–4. 131. Micic, M., Zheng, Y., Moy, V., et al. (2002) Comparative studies of surface topography and mechanical properties of a new, photo-switchable PEG-based hydrogel, Colloid Surf. B-Biointerfaces, 27, 147–58. 132. Takada, K., Miyazaki, T., Tanaka, N. and Tatsuma, T. (2006) Three-dimensional motion and transformation of a photoelectrochemical actuator, Chem. Commun., 2024–6. 133. Tatsuma, T., Takada, K. and Miyazaki, T. (2007) UV-Light-Induced Swelling and VisibleLight-Induced Shrinking of a TiO2-Containing Redox Gel, Adv. Mater., 19, 1249–51. 134. Calvert, P., Electroactive polymer gels, in Electroactive Polymer (EAP) Actuators as Artificial Muscles: Reality, Potential, and Challenges, 2nd Edn (ed. Cohen, Y.B.), SPIE Press Monograph Vol. PM136 2004).


39

135. Doi, M., Matsumoto, M. and Hirose, Y. (1992) Deformation Of Ionic Polymer Gels By ElectricFields, Macromolecules, 25, 5504–11. 136. Shiga, T. (1997) Deformation and viscoelastic behavior of polymer gels in electric fields, Adv. Polym. Sci., 134, 131–63. 137. Osada, Y., Okuzaki, H. and Hori, H. (1992) A polymer gel with electrically driven motility, Nature, 355, 242–4. 138. Kaneko, D., Gong, J. P. and Osada, Y. (2002) Polymer gels as soft and wet chemomechanical systems—an approach to artificial muscles, J. Mater. Chem., 12, 2169–77. 139. Shiga, T. and Kurauchi, T. (1990) Deformation of polyelectrolyte gels under the influence of an electric field, J. App. Polym. Sci., 39, 2305–20. 140. Shiga, T., Hirose, Y., Okada, A. and Kurauchi, T. (1992) Bending of poly(vinyl alcohol)poly(sodium acrylate) composite hydrogel in electric fields, J. Applied Polym. Sci., 44, 249–53. 141. Shiga, T., Hirose, Y., Okada, A. and Kurauchi, T. (1992) Electric field-associated deformation of polyelectrolyte gel near a phase transition point, J. Applied Polym. Sci., 46, 635–40. 142. Shiga, T., Hirose, Y., Okada, A. and Kurauchi, T. (1993) J. Intelligent Material Systems and Structures, 4, 553–7. 143. Shiga, T., Hirose, Y., Okada, A. and Kurauchi, T. (1994) Deformation Of Ionic Polymer Gel Films In Electric-Fields, J. Mater. Sci., 29, 5715–8. 144. Yeghiazarian, L., Mahajan, S., Montemagno, C., et al. (2005) Directed motion and cargo transport through propagation of polymer-gel volume phase transitions, Adv. Mater., 17, 1869-. 145. Liu, Z. and Calvert, P., (2005) Multilayer hydrogels as muscle-like actuators, Adv. Mater., 12, 288–91. 146. Liang, S., Xu, J., Weng, L., et al. (2007) Electrically Induced Linear Locomotion of Polymer Gel in Air, J. Polym. Sci.: Part B: Polym. Phys, 45, 1187–97. 147. Filipcsei, G., Csetneki, I., Szilagyi, A. and Zrinyi, M. (2007) Magnetic field-responsive smart polymer composites, Adv. In Polym. Sc., 206, 137–89. 148. Bellan, C. and Bossis, G. (2002) Field dependence of viscoelastic properties of MR elastomers, Int. J. Modern Phys B, 16, 2447–53. 149. Ginder, J. M., Clark, S. M., Schlotter, W. F. and Nichols, M. E. (2002) Magnetostrictive phenomena on magnetorheological elastomers, Int. J. Modern Phs B, 16, 2412–8. 150. An, Y., Liu, B. and Shaw, M. T. (2002) Soft gels with ordered iron particles: fabrication and electrorheological response, Int. J. Modern Phys B, 16, 2440–6 . 151. Mitsumata, T., Sakai, K. and Takimoto, J.-i. (2006) Giant Reduction in Dynamic Modulus of K-Carrageenan Magnetic Gels, J. Phys. Chem. B, 110, 20217–23. 152. Li, M.-H. and Keller, P. (2006) Artificial muscles based on liquid crystal elastomers, Phil. Trans. R. Soc. A, 364, 2763–77. 153. Buguin, A., Li, M.-H., Silberzan, P., et al. (2006) Micro-Actuators: When Artificial Muscles Made of Nematic Liquid Crystal Elastomers Meet Soft Lithography, J. Am. Chem. Soc., 128, 1088–9. 154. Hikmet, R. A. M. (1999) Anisotropic networks and gels formed by photopolymerisation in the ferroelectric state, J. Mater. Chem., 9, 1921–32. 155. Urayama, K., Honda, S. and Takigawa, T. (2006) Deformation Coupled to Director Rotation in Swollen Nematic Elastomers under Electric Fields, Macromolecules, 39, 1943–9. 156. Yu, Y., Nakano, M. and Ikeda, T. (2004) Photoinduced bending and unbending behavior of liquid-crystalline gels and elastomers, Pure and App. Chem., 76, 1467–77. 157. Kaneko, T., Yamaoka, K., Osada, Y. and Gong, J. (2004) Thermoresponsive shrinkage triggered by mesophase transition in liquid crystalline physical hydrogels, Macromolecules, 37, 5385–8. 158. Spillmann, C. M., Naciri, J., Martin, B. D., et al. (2007) Stacking nematic elastomers for artificial muscle applications, Sensors and Actuators A, 133, 500–5. 159. Marsh, R. L., Olson, J. M. and Guzik, S. K. (1992) Mechanical performance of scallop adductor muscle during swimming (and R.M. Alexander (Commentary)), Nature, 357, 411–3.

40


160. Megill, W. M., Gosline, J. M. and Blake, R. W. (2005) The modulus of elasticity of fibrillin-containing elastic fibres in the mesoglea of the hydromedusa Polyorchis penicillatus, J. Exp.l Bio., 208, 3819–34. 161. Gosline, J. (1971) Connective Tissue Mechanics Of Metridium-Senile 1. Structural And Compositional Aspects, J. Exp. Bio., 55, 763–74. 162. Gosline, J. (1971) Connective Tissue Mechanics Of Metridium-Senile 2: Visco-Elastic Properties And Macromolecular Model, J. Exp. Bio., 55, 775–85. 163. Koob, T. J., Koob-Emunds, M. M. and Trotter, J. A. (1999) Cell-Derived Stiffening And Plasticizing Factors In Sea Cucumber (Cucumaria Frondosa) Dermis, J. Exp.l Bio., 2291–301. 164. Pickard, W., Knoblauch, M., Peters, W. and Shen, A. (2006) Prospective energy densities in the forisome, a new smart material, Mat. Sci. Eng. C-Biomimetic And Supramolecular Systems, 26, 104–12. 165. Shen, A., Hamlington, B., Knoblauch, M., et al. (2006) Forisome based biomimetic smart materials, Smart Structures and Systems, 2, 225–35. 166. Geissler, E., Hecht, A.-M., Horkay, F. and Zrinyi, M. (1998) Compressional Modulus of Swollen Polyacrylamide Networks, Macromolecules, 21, 2594–9. 167. Arndt, K., Richter, A., Ludwig, S., et al. (1999) Poly(vinyl alcohol)/poly(acrylic acid) hydrogels: FT-IR spectroscopic characterization of crosslinking reaction and work at transition point, Acta Polymerica, 50, 383–90. 168. Kim, K. J. and Shahinpoor, M. (2002) Ionic polymer–metal composites: II. Manufacturing techniques, Smart Mat. Struct., 11, 1–15. 169. Achilleos, E., Christodoulou, K. and Kevrekidis, I. (2001) A transport model for swelling of polyelectrolyte gels in simple and complex geometries, Comput. Theor. Polym. Sci., 11, 63–80. 170. Achilleos, E. C., Prud’homme, R. K., Kevrekidis, I. G., et al. (2000) Quantifying Deformation in Gel Swelling: Experiments and Simulations, AIChE J., 46, 2128–39. 171. Paxton, R. A. and Al-Jumaily, A. M. (2000) Swelling of polyelectrolyte hydrogels using a finite element model, Polymer, 47, 5997–6003. 172. Kim, S., Spinks, G., Prosser, S., et al. (2006) Surprising shrinkage of expanding gels under an external load, Nature Mat., 5, 48–51. 173. Hirotsu, S. (1991) Softening of bulk modulus and negative Poisson’s ratio near the volume phase transition of polymer gels, J. Chem. Phys., 94, 3949–57. 174. Li, C., Hu, Z. and Li, Y. (1993) Poisson’s ratio in polymer gels near the phase-transition point, Phys. Rev. E, 48, 603–6. 175. Takigawa, T., Morino, Y., Urayama, K. and Masuda, T. (1996) Osmotic Poisson’s ratio and equilibrium stress of Poly(acrylamide) gels, Polym. J., 28, 1012–3. 176. Beebe, D. J., Moore, J. S., Bauer, J. M., et al. (2000) Functional hydrogel structures for autonomous flow control inside microfluidic channels, Nature, 404, 588–90. 177. Dong, L., Agarwal, A. K., Beebe, D. J. and Jiang, H. (2006) Adaptive liquid microlenses activated by stimuli-responsive hydrogels, Nature, 443, 551–4. 178. Tsutsui, H., Mikami, M. and Akashi, R. (2004) All-polymer-gel light modulator consisting of a ‘gel-in-gel’ system, Adv. Mat., 16, 1925–9. 179. Bassetti, M., Chatterjee, A., Aluru, N. and Beebe, D. (2005) Development and modeling of electrically triggered hydrogels for microfluidic applications, J. Microelectromechanical Systems, 14, 1198–207. 180. Kim, D. and Beebe, D. (2007) Hydrogel-based reconfigurable components for microfluidic devices, Lab on a chip, 7, 193–8. 181. Harmon, M. E., Tang, M. and Frank, C. W. (2003) A microfluidic actuator based on thermoresponsive hydrogels, Polymer, 44, 4547–56. 182. Richter, A., Kuckling, D., Howitz, S., et al. (2003) Electronically Controllable Microvalves Based on Smart Hydrogels: Magnitudes and Potential Applications, J. Microelectromechanical Systems, 12, 748–53.


41

183. Tsutsui, H. and Akashi, R. (2006) Controlling optical properties of a novel light-modulation device consisting of colored N-isopropylacrylamide gel particles dispersed in a poly(vinyl alcohol) solution, J. App. Polym. Sci., 102, 362–8. 184. Tsutsui, H. and Akashi, R. (2006) Effects of pigments on the volume-phase-transition properties of N-isopropylacrylamide gels as light-modulation materials, J.Polym. Sci. A – Polym. Chem., 44, 4644–55. 185. Lin, C.-C. and Metters, A. T. (2006) Hydrogels in controlled release formulations: Network design and mathematical modeling., Adv. Drug Delivery Rev., 58, 1379–408. 186. Peppas, N. A., Bures, P., Leobandung, W. and Ichikawa, H. (2000) Hydrogels in pharmaceutical formulations, Eur. J. Pharm. and Biopharm., 50, 27–46. 187. Kavanagh, C. A., Rochev, Y. A., Gallagher, W. M., et al. (2004) Local drug delivery in restenosis injury: thermoresponsive co-polymers as potential drug delivery systems, Pharmacology & Therapeutics, 102, 1–15. 188. Mueller, A. (2005) Enzyme Electrodes for Medical Sensors, Mini-Rev. in Med. Chem., 5, 231–9. 189. Lavine, B., Kaval, N., Westover, D. and Oxenford, L. (2006) New approaches to chemical sensing-sensors based on polymer swelling, Analytical Letters, 39, 1773–83. 190. Gerlach, G., Guenther, M., Sorber, J., et al. (2005) Chemical and pH sensors based on the swelling behavior of hydrogels, Sensors and Actuators B (2005), 111–112, 555–61. 191. Ambade, A. V., Sandanaraj, B. S., Klaikherd, A. and Thayumanavan, S. (2007) Fluorescent polyelectrolytes as protein sensors, Polym. Int., 56, 474–81. 192. Holtz, J. and Asher, S. (1997) Polymerized colloidal crystal hydrogel films as intelligent chemical sensing materials, Nature, 389, 829–32. 193. Holtz, J. H., Holtz, J. S. W., Munro, C. H. and Asher, S. A. (1998) Intelligent Polymerized Crystalline Colloidal Arrays: Novel Chemical Sensor Materials, Anal. Chem, 70, 780–91 . 194. Marshall, A. J., Young, D. S., Kabilan, S., et al. (2004) Holographic sensors for the determination of ionic strength, Anal. Chim. Acta, 527, 13–20. 195. Kabilan, S., Marshall, A. J., Sartain, F. K., et al. (2005) Holographic glucose sensors, Biosensors and Bioelectronics, 20, 1602–10. 196. Linden, H. J. v. d., Herber, S., Olthuis, W. and Bergveld, P. (2003) Stimulus-sensitive hydrogels and their applications in chemical (micro)analysis, Analyst, 128, 325–31. 197. Arnold, F. H., Zheng, W. and Michaels, A. S. (2000) A membrane-moderated, conductimetric sensor for the detection and measurement of specific organic solutes in aqueous solutions, J. Membrane Sci., 167, 227–39. 198. Kikuchi, A., Suzuki, K., Okabayashi, O., et al. (1996) Glucose-Sensing Electrode Coated with Polymer Complex Gel Containing Phenylboronic Acid, Anal. Chem., 68, 823–8. 199. Li, C., Dong, H., Cao, X., et al. (2007) Implantable electrochemical sensors for biomedical and clinical applications: Progress, problems, and future possibilities, Current Med. Chem., 14, 937–51. 200. Li, H., Ng, T. Y., Yew, Y. K. and Lam, K. Y. (2007) Meshless Modeling of pH-Sensitive Hydrogels Subjected to Coupled pH and Electric Field Stimuli: Young Modulus Effects and Case Studies, Macromol. Chem. Phys. 208, 1137–46.

2 Bio-Responsive Hydrogels for Biomedical Applications Tom McDonald1,2, Alison Patrick1,2, Richard Williams1,2, Brian G. Cousins1,2 and Rein V. Ulijn3 1

School of Materials, Materials Science Centre, University of Manchester, United Kingdom Manchester Interdisciplinary Biocentre (MIB), University of Manchester, United Kingdom 3 University of Strathclyde, United Kingdom

2

2.1 Introduction The term hydrogel is derived from the Latin prefix hydro- referring to water and gel, geluor gelatus- to describe a frozen or immobilised structure. Hydrogels are super absorbent materials, which consist of over 90–99 % water and are prepared by chemical polymerisation or physical assembly of man made or natural resources. Hydrogels form an important class of materials that impact upon modern medicine and are rapidly becoming significant due to the greater life expectancy that people will reach in the twenty-first century. Such materials have found applications such as intraocular and soft contact lenses, pharmaceutical carriers, drug delivery devices, biological sensors, wound dressings and scaffolds for regenerative medicine [1, 2]. Such materials are often nonfouling due to their high hydrophilicity, allow for facile diffusion of solutes and are easily functionalised to promote specific interactions with and respond to the biological environment [3]. In this chapter, the focus is on hydrogel materials that are designed to facilitate communication between the material and the biological environment. There are two main classes of hydrogels (Figure 2.1), those composed of threedimensional networks of cross-linked polymer chain structures that are insoluble in water (chemical hydrogels) and those produced by the self-assembly of (macro) molecules to form noncovalent structures (physical hydrogels).


44


(a) Chemical hydrogel

(b) Physical hydrogel

Figure 2.1 Schematic diagram of the two main classes of hydrogels produced by chemically cross-linked polymers (a) and (macro) molecules forming noncovalent structures (b).

2.2 Chemical Hydrogels Chemical hydrogels are composed of three-dimensional networks produced by the reaction of one or more (macro) monomeric structures through chemical cross-linking, resulting in insoluble structures that may be further stabilised by noncovalent interactions such as ionic, hydrophobic, pi-stacking, hydrogen bonding and/or van der Waals interaction [4]. The use of synthetic hydrogels dates back to the 1950s when Wichterle and Lim synthesised the first hydrogels for biomedical applications to develop soft contact lenses based on a copolymer of 2-hydroxyethyl methacrylate and ethylene dimethacrylate to form poly(hydroxyethyl methacrylate) (PHEMA) [5]. The commercial success of contact lens materials stimulated vast interest in hydrogels and, more recently, led to the development of stimuli-responsive or ‘smart’ hydrogels that change their physical properties in response to changes in the local environment such as pH, temperature, ionic strength, solvent composition, pressure and electrical potential. Responsive chemical hydrogels typically change their hydration states, that is they swell or collapse in response to applied stimuli.

2.3 Physical Hydrogels Physical hydrogels are composed of molecular building blocks varying from small amphiphilic molecules to macromolecules, which form by self-assembly, exclusively through noncovalent cross-links. Water can penetrate throughout the physical structure in between molecular building blocks, which gives rise to hydrogel network structures. Physical hydrogels typically show gelation or dissolution behaviour in response to applied stimuli.

2.4 Defining Bio-Responsive Hydrogels By using the similar design principles for both hydrogel types, they can be tailored to incorporate recognition motifs that respond to specific biomolecular events.

Bio-Responsive Hydrogels for Biomedical Applications

45

Here, the term ‘bio-responsive’ is defined as stimuli responsive materials that change properties in response to a biomolecular recognition event, whereby these molecular interactions are translated into bulk changes in material properties that result in an observed response, which may be optical, electronic or chemical in nature (Figure 2.2).

Input

Design

Step 1: Stimulus The biochemical stimulus may be a small molecule (peptide, glucose, inorganic ions) or a biomacromolecule (antibody, enzyme).

Step 2: Molecular Biorecognition Recognition of the stimulus by the incorporated receptor/sensing element.

Step 3: Molecular Actuation A molecular change triggers a macroscopic transition such as collapse, swelling, selfassembly.

Step 4: Response The macroscopic transition causes emission of an optical, electronic or chemical signal. Output

Figure 2.2 A simplified scheme of events that define and influence a bio-responsive hydrogel system. An input of a stimulus (step 1) leads to the binding of the receptor to the ligand within the gel network resulting in molecular biorecognition events (step 2). Molecular actuation is stimulated in step 3 to trigger further micro and macroscopic events, i.e. swelling. Molecular actuation results in a response (step 4) that can be mechanical, chemical, optical or electronic.

Indeed, bio-responsive hydrogels have gained significant interest over the last several years for applications related to drug delivery, diagnostics, wound healing and tissue regeneration. This exciting field is vast and the emergence of new articles appears on a daily basis and continues to increase rapidly. The authors have therefore chosen a relatively small number of studies from the literature to outline each example in detail. The reader is encouraged to locate further examples of bio-responsive systems cited elsewhere in the literature [1, 3, 6, 7].

46


2.5 Bio-Responsive Chemical Hydrogels This section looks at three examples of actuation for bio-responsive chemical hydrogels and describes how these methods have been used in the development of biosensor, drug delivery and tissue scaffold systems. These three actuation examples are classified in Table 2.1 as follows: changes in cross-linking density (1–8), electrostatic interactions (9–13) and molecular conformation (14, 15). 2.5.1

Actuation Based on Changing the Cross-Linking Density

Cross-linking density may be modified by either the cleavage of polymer chains (irreversible) or via competition at the binding sites of the cross-linking moieties (reversible). An antigen-responsive hydrogel has been described by Uragami and co-workers [8], (Table 2.1, entry 1). This semi-interpenetrating network (semi-IPN) hydrogel contains both grafted antigens and their corresponding specific antibody. Antigen–antibody binding causes the formation of cross-links within the polymer. Free antigens in the solution around the hydrogel create competition between the grafted antigens leading to a decrease in crosslinking density, thus an increase in swelling of the hydrogel. This increase in swelling is reversible. By removing the hydrogel from the antigen solution and washing, the polymer would return to approximately its original volume. This response also corresponds to an increase in permeability of the polymer, and by using the polymer as a membrane, selective permeation of a protein was shown in response to the specific antibody (Figure 2.3). More recently a similar design was employed to produce an antigen-responsive membrane that has gating properties for selective diffusion in response to the presence of a free antigen [9] (Table 2.1, entry 2). A glucose responsive hydrogel has been developed by using the formation of crosslinking complexes between poly(glucosyloxy-ethy1methacrylate) poly(GEMA) and concanavalin A (Con A) (Table 2.1, entry 3) [10]. When there is free glucose present the hydrogel responds through an increase in swelling. Systems based on the cleavage of polymer chains as the mechanism for changing crosslinking density include Table 2.1, entries 4–6: a chymotrypsin-responsive hydrogel developed by Moore and co-workers [11], a cell-responsive hydrogel demonstrated by Hubbell’s group in which enzymes secreted by the cells cleave cross-links within the polymer [12], and also a calcium-responsive hydrogel by Golbart and Kost. Here an inactive form of the enzyme is activated by calcium ions from solution leading to the degradation of the polymer matrix [13]. Messersmith and co-workers have developed a system in which the assembly of short peptides is controlled by the formation of chemical cross-links. The stimulus is the crosslinking enzyme, Transglutamase. Eleven peptide residues of alternating hydrophobic/ hydrophillic charge are used. The biorecognition sites are intermolecular glutamine and lysine residues. The molecular activation is the formation of covalent cross-links, which stabilise the peptides into fibrils that provide the response, a hydrogel [14] (Table 2.1, entry 7). In a second system from Lutolf’s group, the lysine donor is connected by a peptide sequence, GPQGIWGQ, a biorecognition motif for the cellularly excreted stimulus, the enzyme MMP-1 [15] (Table 2.1, entry 7). The molecular response in this case is the enzymatic hydrolysis of peptide cross-links. By incorporation of peptides that can be


47

Table 2.1 Examples of some bio-responsive hydrogels described in the literature

1 2 3 4 5 6 7 8

9

Stimulus

Actuation

Antigen

Cross-linking: Antigen– antibody binding

Hydrogel

Response

Poly-acrylamide Increased swelling (reversible) Antigen Cross-linking: Antigen– Dextran Gating of antibody binding polymeric membrane network (reversible) Glucose Cross-linking: Poly(GEMA) Increased Con A– glucose binding swelling (reversible) Enzyme Cross-linking: Cleavage Poly-acrylamide Gel dissolution (Chymotrypsin) (irreversible) Enzyme (MMPs) Cross-linking: Cleavage PEG Gel dissolution (irreversible) Cellulose-Starch Degradation of Calcium Cross-linking: Matrix matrix degradation (low cross-link (irreversible) density) Cross-linking: Cleavage of Peptide Hydrogel Enzyme hydrogel formation (Transglutaminase/ strands MMP-1) PNIPAM-coChange in Biotin Cross-linking: focus of image Displacement of molecules AAC projected through material Glucose

10 Glucose

11 Enzyme (Protease)

12 Enzyme (Protease)

13 Specific ligand 14 Specific ligand

Electrostatic: GOx oxidation of glucose causing pH responsive swelling Electrostatic: GOx oxidation of glucose causing pH responsive collapse of gating polymer Electrostatic: Protease cleavage of zwitterions resulting in a net charge and swelling Electrostatic: Same as entry 11 except net charge upon zwitterion cleavage is the same as the charge of the protein to be released Change in conformation: Incorporated protein Change in conformation: Incorporated protein

Ref. [8] [9] [10] [11] [12] [13] [14, 15] [16]

Poly (HEMA-co- Increased swelling DMAEMA (reversible)

[18]

Membrane of PVDF with grafted PAAC

Gating of membrane (reversible)

[19]

PEGA

Increased swelling (irreversible)

[20]

PEGA

Increased swelling (irreversible)

[22]

PEG and the protein CaM

Increased swelling (reversible) Increased swelling (reversible)

[24]

PEG and the protein CaM

[25]

48

Biomedical Applications of Electroactive Polymer Actuators (a)

H O AAm

N C CH

APS/ TEMED

Modified antibody H O N C CH

Antigen

Polymerized antibody (c)

CH2

Modified antigen (δ ), AAm MBAA, APS/TEMED

Equilibrium swelling ratio (m3/m3)

(b)

1.15 Antigen-antibody semi-IPN hydrogel

Antigen

1.10

1.05

1.00

0.95

PAAm semi-IPN hydrogel

0

2

4

6

8

10

Antigen concentration (mg ml–1)

Figure 2.3 Antigen-responsive hydrogel. (a): Synthesis of the antigen–antibody semi-IPN hydrogel. (b): Effects of the free antigen concentration on the hydrogel swelling ratio. (c): Diagram of a suggested mechanism for the swelling of an antigen–antibody semi-IPN hydrogel in response to a free antigen. (Reprinted with permission from [8]. Copyright (1999) Nature Publishing Group.)

cleaved by cell-secreted proteases, cells can be encouraged to migrate through a gel by digesting it locally, similar to cell migration in natural extracellular matrices. By including these motifs in artificial matrix mimics, it improves the biological compatibility. Cells can be cultured on artificial matrix, and then begin to replace the artificial matrix with a natural one. In another example of actuation by a change in cross-linking density, the optical properties of spherical hydrogel particles are exploited as flexible microscopic lenses that change their focus upon changes in swelling (Table 2.1, entry 8). Kim and co-workers [16] developed a system in which a reaction between the biomolecule and cross-links in the hydrogel structure leads to swelling at the surface of the material. Hydrogels of poly(N-isopropylacrylamide-co-acrylic acid), P(NIPAM-co-AAc), were prepared and functionalised with biotin and aminobensophenone. These hydrogels were adsorbed onto a surface. Anti-biotin molecules were photochemically cross-linked into the structure at the surface, linking the biotin and aminobenzophenone. This cross-linking controls the degree of swelling at the surface. When free biocytin (substitute for biotin) is present in solution it displaces the anti-biotin from the anchored hydrogel biotin, breaking the cross-links and changing the degree of swelling at the surface, as shown in Figure 2.4. The degree of swelling at the hydrogel surface controls the optical properties. Actuation results in a


49

change in the focus of an image projected through the material, observed by optical microscopy. This can be used as the output method to determine that a change in swelling has occurred. Some results are shown in Figures 2.4c and 2.4d before and after treatment with biocytin respectively. Before treatment the hydrogel focuses the image so a white edge and dark centre is seen. After treatment the hydrogel de-focuses and a white centre is seen.

P(NIPAM-co-AAc) H2 C

H C C

H2 C n O

NH

uv

H C C

m O

(b)

OH

CH H3C

(c)

(a)

CH3

Biotin

Aminobenzophenone

Anti-Biotin

(d)

Biocytin

Cross-linked area

Figure 2.4 (a) Structure of P(NIPAM-co-AAc). (b) Schematic diagram showing hydrogels functionalised with biotin and aminobenzophenone, adsorbed to a surface. Anti-biotin crosslinks are formed to control degree of swelling at the surface. Optical microscopy images: (c) before treatment with biocytin, (d) after treatment with biocytin (scale bar ¼ 2 mm).

Other bio-responsive hydrogels have been developed to give an optical output that use different actuation methods. These include sensors for a-cyclodextrin, in which a reaction with polydiacetylene liposomes embedded in poly(ethylene glycol) diacrylate (PEGDA) leads to a visible colour change of the hydrogel from blue to red [17]. 2.5.2

Actuation Based on Changes in Electrostatic Interactions

A second mechanism to actuate a change in swelling is through electrostatic repulsion/ attraction of polymer chains. This method of actuation is widely studied in the context of glucose responsive polymers for the treatment of diabetes. Generally, glucose oxidase (GOx) is immobilised in the hydrogel, along with catalase, an enzyme that is used to prevent the build up of hydrogen peroxide. When glucose is present it is converted to gluconic acid and hydrogen peroxide, which lowers the pH within the hydrogel. There are two different designs in which this decrease in pH is used to actuate a change in swelling: matrix type systems, where the enzyme and insulin are contained within a bulk polymer, or membrane type systems, where the drug is contained in a reservoir within a membrane. Kost and co-workers produced an example of the matrix system where the insulin and enzyme were contained uniformly throughout a solid polymer [18] (Table 2.1, entry 9). This polymer contained amine groups and had a low cross-linking density, so that at low pH the amines become ionised leading to an increase in swelling. In Figure 2.5, GOx catalyses the reaction of glucose to gluconic acid forming hydrogen peroxide (equation 1). A build up of hydrogen peroxide leads to inhibition of the enzyme, and because oxygen is needed to form gluconic acid a shortage of oxygen leads to slower

50


swelling rates. For this reason this system also contained catalase, which serves to convert hydrogen peroxide to water and oxygen (equation 2). The device was implanted in rats and these in vivo experiments indicated that some of the entrapped insulin retained its active form and was effective in reducing blood glucose levels. A membrane type system has been created by Liang and co-workers [19] (Table 2.1, entry 10). Here, poly(acrylic acid) (PAAC) was grafted to a porous membrane of poly(vinylidene fluoride) (PVDF) and GOx was covalently bound to the PAAC. The PAAC chains ‘gate’ the membrane pores and in the presence of glucose the chains collapse and the pores are opened.

GOx Glucose + O2 + H2O

2H2O2

Figure 2.5

→

Gluconic acid + H2O2

Catalase → O2 + 2H2O

(1)

(2)

Reactions driven by the enzymes used in glucose responsive hydrogels.

A different type of electrostatically actuated responsive hydrogel was developed by Ulijn and co-workers [20] (Table 2.1, entry 11). The stimulus in this system is an enzyme, specifically a protease (an enzyme that hydrolyses peptide bonds). Enzymes are well suited to use as a stimulus as they can be targeted to highly specific substrates thanks to unique chemo-, regio- and enantio-selective mechanisms [21]. They work under aqueous physiological conditions, that is high ionic strength, pH 5–8 and at 37 °C. These systems are easily matched to applications through a wide range of functions and perform key roles as selective catalysts in cell pathways and disease states. Here the response is governed by a peptide actuator which consists of two components: a neutral dipeptide flanked by oppositely charged actuating amino acids, creating an overall neutral, zwitterionic peptide chain. The central peptide is the enzyme cleavable linker (ECL), which can be matched to be cleaved exclusively by a target protease. When hydrolysis occurs, a mobile anionic peptide fragment is cleaved, leaving a cationic peptide fragment attached to the hydrogel. Electrostatic repulsion between these groups leads to an increase in swelling, causing the entrapped payload to diffuse out of the hydrogel into the surrounding solution. More recently, this system has been further developed to release proteins [22] (Table 2.1, entry 12). This was achieved by tailoring the design of the peptide actuator to give a net charge after ECL cleavage that is matched to the charge on the protein. Electrostatic repulsion then assists the release of the protein. A biosensor system that uses such ECL recognition followed by actuation has been developed for the detection of elastase. It uses fluorescence resonance energy transfer (FRET) as a method of displaying the actuation [23] (Table 2.1, entry 14). In this biosensor example a FRET pair is incorporated into the hydrogel, poly(ethylene glycol acrylamide) (PEGA), and is separated by an ECL. Enzyme action cleaves the ECL, removing the acceptor molecule of the FRET pair, switching on fluorescence. These hydrogels are also functionalised with negatively charged residues that cause electrostatic repulsion and lead to increased swelling of the hydrogel for easier diffusion of the enzyme into the particle. As elastase is positively charged at neutral pH, the negative charges added into the hydrogel are bifunctional and attract the elastase to the ECL once inside the hydrogel (Figure 2.6).


51

Figure 2.6 (a) Peptide designed for the release of negatively charged protein molecules. Two positively charged amino acids (arginine) are separated from two negative amino acids (aspartic acid) by the neutral ECL (two alanine residues). A single net negative charge remains on the particle following enzyme hydrolysis. (b) Exclusion of albumin from the negatively charged swollen particle occurs following hydrolysis of the bond between alanine residues by thermolysin. (c) Two photon microscopy images of Texas Red labelled albumin being released from the particles (scale bars ¼ 75 mm). (Reprinted with permission from [22]. Copyright (2008) Royal Society of Chemistry.)

2.5.3

Actuation Based on Conformational Changes

The third mechanism of actuation is based on changes in conformation of natural proteins (Table 2.1, entries 13 and 14). These systems incorporate a natural protein into the hydrogel that undergoes a conformation change and thus alters the characteristics of the material. The use of proteins as actuators is a new development in bio-responsive hydrogels. An example of this system was developed by Mrksich and co-workers [24]. In this study the functional nature of the hydrogel was given by the conformational properties of the protein calmodulin (CaM). Calmodulin has two distinct conformational states. In the presence of calcium ions, CaM has an extended, dumb-bell shaped conformation (extended CaM).

52


Figure 2.7 Ligand-responsive hydrogel that relies on conformational changes. (a) The two conformational states of CaM: an extended conformation in the presence of calcium ions (left) and a collapsed conformation upon binding to a ligand (right). (b) Shown is a hydrogel with CaM in a ligand-free state (left) and the same gel with CaM in a ligand-bound state (right) (scale bars ¼ 1 mm). (c) Hydrogels were exposed to TFP ligand and the volume was measured at various intervals for two hours. The gel was then washed repeatedly and incubated in a calcium-containing buffer to restore the extended CaM conformation. (Reprinted with permission from [27]. Copyright (2006) American Chemical Society.)


53

This calcium-bound extended CaM undergoes a transition from an extended dumb-bell to a collapsed conformation (collapsed CaM) upon the binding of ligands. The CaM protein is incorporated into the hydrogel by covalently reacting it with four armed PEG molecules. The hydrogel shows a macroscopic decrease in volume when exposed to the trifluoperazine ligand (TFP). TFP binds specifically to CaM, causing the CaM to undergo a conformation change from extended to collapsed (Figure 2.7). This decrease in volume could be reversed by chelating the calcium ions, thus removing the calcium-bound ligand. Numerous cycles between extended and collapsed material were possible, demonstrating the reversible nature of the hydrogel. More recently, this concept has been developed to incorporate a photochemical assembly, allowing spatial control of the location of dynamic proteins [25].

2.6 Bio-Responsive Physical Hydrogels The spontaneous assembly of (macro) molecules to form physical hydrogels is an area of increasing importance. It is desirable to introduce a level of control in these systems, enabling the triggering of hydrogel formation or dissolution upon application of selective stimuli that are uniquely found in certain biological situations. This can be achieved by employing biorecognition elements into (some of) the molecular building blocks in the direct vicinity of ionisable groups, or hydrophobic regions. Stimuli that can be exploited in this context include altering of physical conditions, that is temperature, pH, ionic strength and oxidative species, or using enzymes, such as proteases, kinases, phosphatases or ligases. Enzymatic actuation leads to controlled intramolecular events, such as shortening/straightening of a peptide chain, ionisation, side chain modification or the exposure of hydrophobic regions, to redefine existing molecular interactions that enable new intermolecular associations [21]. The next section gives examples of mechanisms that may be used in physical gels based on small molecules or macromolecules. The biorecognition elements are variable depending upon enzymatic function, for example a peptide bond between specific amino acids (Table 2.2, entries 3, 4, 7–9), a (phosphate) ester group (entries 1–3, 5) or a specific sequence of DNA (entry 6). 2.6.1

Enzyme-Responsive Physical Hydrogels

The formation of physical assemblies from small peptide amphiphiles can be achieved by manipulating the structure of the peptides with enzymes. These reactions are typically the formation/breaking of covalent bonds altering the structures and thereby controlling the magnitude of weak structural forces, typically p–p interactions, hydrogen bonding and electrostatic interactions. By appropriate choice of pairs of enzymes, hydrogelation can be made reversible [26]. Typically, self-assembly is a one-way process, as the enzyme only operates in one direction governed by thermodynamic equilibrium, but by using two enzymes with opposite actions, one under thermodynamic control and the other involving a coupled reaction, this limitation may be overcome. For example, peptide (de-) phosphorylation can be exploited as an effective means of hydrogelation control, where a kinase in the presence of ATP (a natural source of energy) adds the phosphate group, while a phosphatase cleaves it off (Figure 2.8a and 2.8b). The phosphate group is used to prevent self-assembly of beta sheet forming

54


Table 2.2 Entries 1–7 highlight mechanism of physical hydrogel formation. Entries 8 and 9 show self-assembly directed through the formation/disruption of covalent cross-links Stimulus

Biorecognition Molecular Actuation

Response

Ref.

RGYSLG

Soluble b-sheets* Hydrogel formation* Hydrogel formation* Hydrogel formation Hydrogel formation Hydrogel formation Drug release Drug release Cell migration

[29]

1 Phosphatase/ Kinase 2 Phosphatase/ Kinase 3 Thermolysin/ Subtilisin 4 Thermolysin

Fmoc-X + FF

5 Penicillin G Amidase 6 T4 DNA ligase

Phenylacetic acid Base pairs

Triple elimination

7 Chymotrypsin 8 Urokinase 9 MMP-2

FA SGRSANA GTAGLIGQ

Bond cleavage b-sheet disruption Hydrophobic disruption

Y/ phosphate Y + L/F-OMe

Phosphate group cleavage/addition Phosphate group cleavage/addition Peptide synthesis/ ester cleavage Peptide synthesis

DNA ligation

[27, 30–32] [33] [26] [34] [28] [35] [36] [37]

Letters indicate single letter code for amino acids as follows: A ¼ Alanine, F ¼ Phenylalanine, G ¼ Glycine, I ¼ Isoleucine, K ¼ Lysine, N ¼ Asparagine, P ¼ Proline, Q ¼ Glutamine, R ¼ Arginine, S ¼ Serine, T ¼ Threonine W ¼ Tryptophan, X ¼ any amino acid and Y ¼ Tyrosine. * Indicates reversible two-enzyme system.

peptides or small aromatic peptide derivatives, typically 1–5 amino acids long. This system is used for a range of applications: sensing the presence of enzymatic inhibitors, the signal being prevention of gelation [30]; the reversible formation of an extracellular matrix mimic for cell culture in vivo [27]; and an antimicrobial system, in which the self assembly is triggered inside microbes using their own enzymes, killing the cell [31]. Table 2.2, entry 4 is a system based on subtilisin and thermolysin [33]. The protease thermolysin acts as the stimulus for assembly, operating to form dipeptide derivatives from Fmoc-threonine and either a leucine or a phenylalanine methyl ester (Figure 2.8c). The dipeptide derivatives formed give rise to self assembled nanofibrous hydrogels. The hydrogel can be dissolved by employing a second enzyme, subtilisin, which cleaves the methyl ester to leave a soluble Fmoc-dipeptide. This system could be used to entrap cells in a network using one enzyme, and release them upon application of the other. A system for controlled release of a payload from a physical gel is discussed in Table 2.2, entry 8 [36]. By encasing a drug within a network which can be degraded by an enzyme, a responsive system has been created – b-sheet segments (KLD)12 either side of an enzyme cleavable biorecognition motif, SGRSANA. A drug is linked to a similar sequence, and when mixed in aqueous solvent, form a gel through stacking of the b-sheets. The stimulus is the addition of the protease urokinase plasminogen activator. This acts from the outside of the gel, through cleavage of the biorecognition motif. Once cleaved, the resulting molecular activation is a weakening of the interactions and breaking down the self-assembled structure. The response to the stimulus is the gradual release of the encapsulated drug. Enzyme-responsive hydrogels have also been studied as extracellular matrix mimics. For example, Hartgerink and co-workers investigated the use of peptide ampiphiles which self-assemble into fibrous cylindrical micelles driven by the interactions of hydrophobic


55

Figure 2.8 Reversible self-assembly by the use of enzymes. (a) De-phosphorylation of the target molecule results in the formation of a self-assembled network. By introducing an inhibitor to the system, hydrogel formation is inhibited (Reprinted with permission from Yang, Z and Xu, B. Bio-responsive hydrogels for biomedical applications, Chem. Commun., 21, 2424–25, Copyright (2004) Royal Society of Chemistry). (b) A reversible system of assembly by dephosphorylating a peptide with phosphatase to trigger assembly, and reversing the assembly with the use of a kinase (in the presence of ATP) (Reprinted in part with permission from Yang, Z., Liang, G., Wang, L. and Xu, B. Using a Kinase/Phosphate Switch to Regulate a Supramolecular Hydrogel and Forming the Supramolecular Hydrogel in vivo, J. Am. Chem. Soc., 128, 9, 3038–43, Copyright (2006) American Chemical Society). (c) A reversible system in which the formation of a peptide bond results in an assembling Fmoc-dipeptide methyl ester. Assembly is reversed by the use of subtilisin to promote ester hydrolysis (Reproduced with permission from Das, K. A., Collins, R. J. and Ulijn R. V., Exploiting Enzymatic (Reversed) Hydrolysis in Directed Self-Assembly of Peptide Nanostructures, Small, 4, 279–87. Copyright (2008) Wiley-VCH Verlag GmbH).

tails and hydrophilic heads with water (Table 2.2, entry 9) [37]. These peptides contain RGD, a cell binding epitope, which is displayed on the surface allowing their use as a cell culture medium. To allow the naturally secreted enzyme MMP-2 to act as a stimulus for matrix remodelling, the biorecognition motif GTAGLIGQ is inserted between the tail and the head. The molecular response to the stimulus is the disruption of the hydrophobic effects and breaking down the fibres [38]. The response to the stimulus was that the cells could be observed migrating along the fibres.

56


2.7 Electroactive Chemical Hydrogels Electroactive chemically cross-linked hydrogels integrate both polymer hydrogels and conjugated (conducting) electroactive polymers, that is forming an electroactive hydrogel. In these systems, a biochemical reaction results in a current which can easily be read. This type of system is especially useful for observing the action of redox enzymes. An example of this combination of polymers is the work done by Brahim and co-workers [39] on a glucose sensing system, in which platinum electrodes were coated with cross-linked poly(hydroxyethyl methacrylate) (pHEMA) and polypyrrole components and contained GOx entrapped in the matrix (Figure 2.9). When the glucose is present it reacts with the enzyme and produces hydrogen peroxide. The steady state current produced by this reaction can be measured and gives a reading of the enzyme activity which is dependent on the enzyme substrate concentration. The pHEMA component creates a biocompatible environment for the entrapped enzyme, giving the sensor a lifetime of up to one year. Brahim and coworkers have also shown in this work that by switching the polypyrrole component for dimethylaminoethyl methacrylate (DMA) the system can release a drug upon stimulation, becoming a drug delivery system rather than a sensing system [40].

Figure 2.9 Schematic diagram showing platinum electrodes coated with cross-linked pHEMA and polypyrrole components with GOx entrapped to form glucose sensor that displays result as current.

This is an example of a biosensor that uses current as an output. Other biosensor systems have been developed that use similar methods for the detection of cholesterol, glucose and hydrogen peroxidise [41, 42].


57

2.8 Conclusion In summary, it has been demonstrated that bio-responsive hydrogels allow effective and selective response to a variety of biochemical stimuli, usually by displaying a swelling/ collapse or gelation/dissolution response. While many systems have been studied in well defined conditions, future development should allow devices to be created that can actively and rapidly respond to biological stimuli at physiological levels and in complex biological fluids. Due to lengthy regulatory and approval processes, it is likely that applications will focus on the use of US Food and Drug Administration (FDA) approved polymers, such as pHEMA, PEG and PGLA based hydrogels. In physical gels, the use of a wide range of enzymes to control gelation has been demonstrated. Many of these systems are at the proof of concept stage, but increasing numbers of examples of biological relevance have been demonstrated, typically in the realm of cell culture. Increasingly, the ability to selectively control the formation and growth of hydrogels will be shown to be of tremendous value for a range of applications, such as in vivo tissue repair, biosensing and drug delivery.

References 1. Peppas, N. A., Hilt, J. Z., Khademhosseini, A. and Langer, R. (2006) Hydrogels in biology and medicine: From molecular principles to bionanotechnology, Adv. Mat., 18, 1345–60. 2. Alexander, C. and Shakesheff, K. M. (2006) Responsive polymers at the biology/materials science interface, Adv. Mat., 18, 3321–8. 3. Ulijn, R. V., Bibi, N. and Jayawarna, V., et al. (2007) Bioresponsive hydrogels, Mat. Today, 10, 40–8. 4. Peppas, N. A., Bures, P., Leobandung, W. and Ichikawa, H. (2000) Hydrogels in pharmaceutical formulations, Eur. J. Pharm. Biopharm., 50, 27–46. 5. Wichterle, O. and Lim, D. (1960) Hydrophilic gels for biological use, Nature, 185, 117–8. 6. Hoffman, A. S. (2002) Hydrogels for biomedical applications, Adv. Drug. Deliv. Rev., 54, 3–12. 7. Chaterji, S., Kwon, I. K. and Park, K. (2007) Smart polymeric gels: redefining the limits of biomedical devices, Prog. Polym. Sci., 32, 1083. 8. Miyata, T., Asami, N. and Uragami, T. (1999) A reversibly antigen-responsive hydrogel, Nature., 399, 766–9. 9. Zhang, R., Bowyer, A., Eisenthal, R. and Hubble, J. (2007) A smart membrane based on an antigen-responsive hydrogel, Biotech. and Bioeng., 97, 976–84. 10. Miyata, T., Jikihara, A., Nakamae, K. and Hoffman, A. S. (1996) Preparation of poly(2glucosyloxyethylmethacrylate) concanavalin A complex hydrogel and its glucose-sensitivity, Macromolecular Chem., and Phys, 197, 1135. 11. Plunkett, K. N., Berkowski, K. L. and Moore, J. S. (2005) Chymotrypsin responsive hydrogel: application of a disulfide exchange protocol for the preparation of methacrylamide Containing Peptides, Biomacromolecules., 6, 632–7. 12. Lutolf, M. P., Raeber, G. P., Zisch, A. H., et al. (2003) Cell-responsive synthetic hydrogels, Adv. Mat., 15, 888–92. 13. Goldbart, R. and Kost, J. (1999) Calcium responsive bioerodible drug delivery system, Pharm. Res., 16, 1483–6. 14. Collier, J. H. and Messersmith, P. B. (2003) Enzymatic modification of self-assembled peptide structures with tissue transglutaminase. Bioconjugate Chem., 14, 748–55.

58


15. Ehrbar, M., Rizzi, S. C., Schoenmakers, R. G., et al. (2007) Biomolecular hydrogels formed and degraded via site-specific enzymatic reactions, Biomacromolecules, 8, 3000–7. 16. Kim, J. S., Singh, N. and Lyon, L. A. (2007) Displacement-induced switching rates of bioresponsive hydrogel microlenses, Chem. Mat., 19, 2527–32. 17. Lee, N. Y., Jung, Y. K. and Park, H. G. (2006) On-chip colorimetric biosensor based on polydiacetylene (PDA) embedded in photopolymerized poly(ethylene glycol) diacrylate (PEG-DA) hydrogel, Biochem. Eng. J., 29, 103. 18. Traitel, T., Cohen, Y. and Kost, J. (2000) Characterization of glucose-sensitive insulin release systems in simulated in vivo conditions, Biomat., 21, 1679–87. 19. Chu, L. Y., Li, Y., Zhu, J. H., et al. (2004) Control of pore size and permeability of a glucoseresponsive gating membrane for insulin delivery, J. Controlled Release, 97, 43–53. 20. Thornton, P. D. M., Mart, R. J. and Ulijn, R. V. (2007) Enzyme-responsive polymer hydrogel particles for controlled release, Adv. Mat., 19, 1252–6. 21. Ulijn, R. V. (2006) Enzyme-responsive materials: a new class of smart biomaterials, J. Mat. Chem., 16, 2217–25. 22. Thornton, P. D. M., Mart, R. J., Webb, S. J. and Ulijn, R. V. (2008) Enzyme-responsive hydrogel particles for the controlled release of proteins: designing peptide actuators to match payload, Soft Matter, 4, 821–7. 23. Patrick, A. G. and Ulijn, R. V. (2007) Fluorescent hydrogel sensor particles for detection of elastase, Mat. Res. Soc. Proc., 1063–PP06–05. 24. Murphy, W. L. D., Dillmore, W. S., Modica, J. and Mrksich, M. (2007) Dynamic hydrogels: translating a protein conformational change into macroscopic motion, Angew. Chem. Int. Edn., 46, 3066–9. 25. Sui, Z., King, W. J. and Murphy, W. L. (2007) Dynamic materials based on a protein conformational change, Adv. Mat., 19, 3377–80. 26. Toledano, S., Williams, R. J., Jayawarna, V., and Ulijn, R. V. (2006) Enzyme-triggered selfassembly of peptide hydrogels via reversed hydrolysis, J. Am. Chem. Soc., 128, 1070–1. 27. Yang, Z. M., Liang, G. L., Wang, L. and Bing, X. (2006) Using a kinase/phosphatase switch to regulate a supramolecular hydrogel and forming the supramoleclar hydrogel in vivo, J. Am. Chem. Soc., 128, 3038–43. 28. Um, S. H., Lee, J. B., Park, N., et al. (2006) Enzyme-catalysed assembly of DNA hydrogel, Nature Mat., 5, 797–801. 29. Winkler, S., Wilson, D. and Kaplan, D. L. (2000) Controlling beta-sheet assembly in genetically engineered silk by enzymatic phosphorylation/dephosphorylation, Biochem., 39, 12739–46. 30. Yang, Z. M. and Xu, B. (2004) A simple visual assay based on small molecule hydrogels for detecting inhibitors of enzymes, Chem. Comm., 2424–5. 31. Yang, Z., Liang, G., Guo, Z. and Xu, B. (2007) Intracellular hydrogelation of small molecules inhibits bacterial growth, Angew. Chem. Int. Edn., 46, 8216–9. 32. Yang, Z. M., Liang, G. L., Ma, M. L., et al. (2007) In vitro and in vivo enzymatic formation of supramolecular hydrogels based on self-assembled nanofibers of a beta-amino acid derivative, Small, 3, 558–62. 33. Das, A. K., Collins, R., Ulijn, R. V. (2008) Exploiting enzymatic (reversed) hydrolysis in directed self-assembly of peptide nanostructures, Small, 4, 279–87. 34. Alder-Abramovich, L., Perry, R., Sagi, A., et al. (2007) Controlled assembly of peptide nanotubes triggered by enzymatic activation of self-immolative dendrimers, Chembiochem., 8, 859–62. 35. van Bommel, K. J. C., Stuart, M. C. A., Feringa, B. L. and van Esch, J. (2005) Two-stage enzyme mediated drug release from LMWG hydrogels, Organic and Biomol. Chem., 3, 2917–20. 36. Law, B., Weissleder, R. and Tung, C. H. (2006) Peptide-based biomaterials for proteaseenhanced drug delivery, Biomacromolecules., 7, 1261–5.


59

37. Jun, H. W., Yuwono, V., Paramonov, S. E. and Hartgerink, J. D. (2005) Enzyme-mediated degradation of peptide-amphiphile nanofiber networks, Adv. Mat., 17, 2612–7. 38. Yang, Z., Liang, G. and Xu, B. (2008) Enzymatic hydrogelation of small molecules, Acc. Chem. Res., 41, 315–26. 39. Brahim, S., Narinesingh, D. and Guiseppi-Elie, A. (2002) Bio-smart hydrogels: co-joined molecular recognition and signal transduction in biosensor fabrication and drug delivery, Biosensors & Bioelectronics, 17, 973–81. 40. Brahim, S., Narinesingh, D. and Guiseppi-Elie, A. (2002) Bio-smart materials: Kinetics of immobilised enzymes in p(HEMA)/p(pyrrole) hydrogels in amperometric biosensors, Macromolecular Symposia, 186, 63–73. 41. Ivekovic, D., Milardovic, S. and Grabaric, B. S. (2004) Palladium hexacyanoferrate hydrogel as a novel and simple enzyme immobilization matrix for amperometric biosensors, Biosensors and Bioelectronics., 20, 872–8. 42. Sun, Y. X., Zhang, H. T., Huang, S. W. and Wang, S. F. (2007) Hydrogen peroxide biosensor based on the bioelectrocatalysis of horseradish peroxidise incorporated in a new hydrogel film, Sensors and Actuators B-Chem., 124, 494–500.

3 Stimuli-Responsive and ‘Active’ Polymers in Drug Delivery Aram Omer Saeed1, Jo´hannes Pa´ll Magnu´sson1, Beverley Twaites2 and Cameron Alexander1 2

1 School of Pharmacy, University of Nottingham, UK School of Pharmacy and Biomedical Sciences, University of Portsmouth, UK

3.1 Introduction The efficacy of synthetic drug compounds in therapy is strongly dependent on their formulation into medicines and on their distribution, localisation and accumulation in different regions in the body. The ability to deliver a drug compound to the specific target site (organ, tissue, cell, intracellular compartment) remains a challenge. A century on from Paul Ehrlich’s visionary hypothesis, for which he received the Nobel Prize in 1908, there are still no ‘Magic Bullets’ [1] against most diseases. As a result, the delivery of drugs to the desired target sites in the body at the right time and in the right dose is still an unmet clinical need [2]. This leads to inefficient use of drugs, undesired side effects and greater medical intervention, with resulting burdens on patients, carer populations and healthcare budgets. Current targeting methods for many drugs either lack specificity or are not active for certain patient groups. As a consequence, drug delivery systems are gaining in importance – and many of these are based on polymers [2, 3]. In addition, new generations of therapeutics are emerging, and these too are often macromolecular or polymeric in nature. For example, in 2002 and 2003 the US Food and Drug Administration (FDA) approved more biotechnology products (proteins and antibodies) and drug delivery systems as marketed products than new low molecular weight drugs. However, these new classes of drugs and their conjugates, complexes and formulated vehicles – sometimes considered as Nanomedicines – and the related biotherapeutics, still urgently require technologies to


62


ensure their specific localisation in target sites [4, 5]. As a consequence, new modalities of ‘smart’ or active materials ranging from bioresponsive to electroactive polymer–drug delivery systems are being developed. In this chapter, a short introduction is given to key concepts in drug delivery, the specific issues for polymers and smart materials related to the transport and release of therapeutic compounds are considered, and some exciting recent examples where actuating polymers have been developed to help target and deliver drug molecules are highlighted. The focus is primarily on soluble polymers and nanoparticles rather than hydrogels, as there are already numerous excellent reviews on responsive hydrogels available, [6–16] and the design concepts in the soluble and nanoparticle polymer systems best exemplify the key responses that can be engineered into ‘smart’ materials.

3.2 Drug Delivery: Examples, Challenges and Opportunities for Polymers In general, drug delivery aims at optimising therapy by delivering bioactive agents at specific sites or specific rates to the patient. The field has evolved from simple topical waxes and delayed release formulations to the targeted delivery of therapeutic agents to specific cells and subcellular compartments. Traditional approaches to drug delivery have been based on simple formulation parameters. For low molar mass compounds (Mw 0 Þ ¼ aðT þ DT Þ ¼ 0

92


the following equality is obtained: aðT; t Þ ¼ ð1 yðt ÞÞa0 where y(t) is the unit step function. Hence, the application of a temperature step change is equivalent to freeing the gel sample, at time equal zero by the stress a0 , and letting it move toward the equilibrium volume, such as in a free swelling experiment. In the case of thermal stimulation, it is possible to perform free deswelling (a0 < 0) experiments depending on the sign of temperature change. By introducing Equation (4.2) in Equation (4.1), and by taking the divergence of both members and inserting the incompressibility of solid and liquid constituents, the equation of motion finally reads [7]: f @ eaa =@t ¼ ðk þ 4m=3Þ@ 2 eaa =@xi @ xi

ð4:4Þ

When the gel sample has the shape of a thin quasi-planar layer, assuming the z-axis is perpendicular to the gel layer plane, Equation (4.2) can be simplified as follows [6, 9]: @exx =@t ¼ D @ 2 exx =@z2

ð4:5aÞ

@eyy =@t ¼ D @ 2 eyy =@z2

ð4:5bÞ

@eaa =@t ¼ Db @ 2 eaa =@z2

ð4:5cÞ

where D ¼ m/f and Db ¼ (k þ 4m/3)/f. The spatio-temporal solutions for the strains exx and ezz are [9]:

exx ¼

1 4e0 X ð1Þn ð2n þ 1 Þ 2 t ð2n þ 1Þzp cos exp t a p n ¼ 1 2n þ 1

ezz ¼ eaa 2exx ¼

1 4e0 X ð 1Þ n ð2n þ 1Þzp cos a p n ¼ 1 2n þ 1

3 exp

ð4:6aÞ

ð4:6bÞ

ð2n þ 1 Þ 2 t ð2n þ 1 Þ 2 t 2 exp tb t

where t ¼ a2 / p2D and tb ¼ a2 / p2Db are the characteristic time constants for the ‘shear’ and ‘bulk’ diffusional gel readjustment, e0 is the initial uniform strain of the sample with respect the final one (at t ¼ 1) assumed as reference (e1 ¼ 0) and ‘a’ is the gel layer thickness at t ¼ 1. Moreover, since Db > D, it follows that tb < t From Equations (4.6a) and (4.6b), the length L(t) and the thickness a(t) of the gel are obtained as a function of time, respectively, to read [9]:

Thermally Driven Hydrogel Actuator for Controllable Flow Rate Pump

Lðt Þ ¼ L1

93

( ) 1 4e0 X ð 1Þ n ð2n þ 1Þ 2 t 1 þ exx ð Z ¼ 0 Þ ¼ L1 1 þ exp t p n ¼ 1 2n þ 1

(

aðt Þ ¼ a 1 þ 2

Z

a=2

ezz 0

ð4:7aÞ

)

(

) 1 8e0 X ð 1 Þ n ð2n þ 1 Þ 2 t ð2n þ 1Þ 2 t ¼a 1 þ 2 2 exp 3 exp tb t p n ¼ 1 2n þ 1

ð4:7bÞ

Given the above kinetics for the pore walls and assuming the global macroscopic gel dimension quasi-isomorphic to the pore diameter, and hence to the pore wall circumference, the gel actuator length will approximately follow the time law given by Equation (4.7a). For t > 9t in Equation (4.7a) the slower exponential relaxation prevails, so that the PVME actuator length reads: 2 h t i 4e0 4e0 p D Lðt Þ ffi L1 1 exp exp 2 t ð4:8Þ ¼ L1 1 t a p p By fitting the exponential length relaxation of the PVME samples it possible to obtain the characteristic time (t) and the gel diffusion coefficient (D). 4.3.2

Experimental Results

The dependence of the PVME response time constant, t ¼ a2/p2D, as a function of the temperature is reported in Figure 4.1. By introducing the temperature dependence of the

Response time constant (s)

20

• •

15

•

•

•

•

10

•

5

• •

0 20

25

30

35

40

45

Temperature (°C)

Figure 4.1 Readjustment time constant of the PVME gel as a function of temperature.

94


pore wall thickness, a ¼ a f(T), where f(T) ¼ L(T)/L(T ¼ 20°C) is reported in Figure 4.2 and where a ¼ a(T ¼ 20°C), the normalized diffusion coefficient, D/a2, shown in Figure 4.3 is obtained. The data refer to thermal free swelling experiments at different temperatures by applying a thermal sudden jump of 2 °C.

•

Fractional PVME length

1.0

•

• •

•

•

•

•

•

0.9

•

• • • • • •

0.8

• • •

0.7

• • •

0.6 15

20

25

30

• •

35

• •

40

45

Temperature (°C)

Normalized diffusion coefficient (s–1)

Figure 4.2 The equilibrium length L(T) of the PVME gel normalized by one at T ¼ 20 °C as a function of temperature.

0.014

•

0.012 0.010 •

0.008 0.006

• •

• •

0.004

• •

•

0.002 20

25

30

35

40

45

Temperature (°C)

Figure 4.3 The shear diffusion coefficient D/a2 of the PVME gel matrix as a function of temperature.


95

By using the relationship that links the Young’s elastic modulus (E) to the bulk (k) and to the shear (m) elastic moduli that reads: m ¼ Eð3=ð9 E=kÞÞ since, in gels, it is usual for k > E [1, 6, 8], it follows that: 3=8

E < m < 1=3 E

Therefore, by measuring the Young’s elastic modulus (E) of the PVME by means of independent force–elongation experiments, reported in Figure 4.4, it is possible to evaluate the shear elastic modulus of the material, reported in Figure 4.5, with a precision of about 5 %. Once the shear elastic modulus has been determined, as reported in Figure 4.6 the friction coefficient ‘f ’ is obtained by the relation D ¼ m /f. It is interesting to note that the shear elastic modulus (m) and the friction coefficient (f) show the typical dispersion sigmoid and bell shape, respectively.

Young’s elastic modulus (N/s2)

60000

• • •

50000

40000

•

30000 •

20000

•

• •

•

10000 20

25

30

35

40

45

Temperature (°C)

Figure 4.4 The Young’s elastic modulus (E) of the PVME gel as a function of temperature.

The force generated by a PVME gel strip of a known sectional area, is a consequence of the thermal stress (a(T)) of Equation (4.3). Actually, the effect of the thermal stress (a(T)) is to change the free gel rest length and, therefore, it is possible to obtain the isometric force generation of a PVME actuator, with free lateral surfaces, by means of the equilibrium PVME length measured as a function of the temperature shown in Figure 4.2 with the Young’s gel modulus reported in Figure 4.4. By posing that the PVME actuator exerts zero force at 36 °C, the stress generated by its temperature change is reported in Figure 4.7.

96


•

Shear elastic modulus (N/s2)

20000

• •

15000 •

10000

• • • •

•

5000

20

25

30

35

40

45

Temperature (°C)

Figure 4.5 The shear elastic modulus (m) of the PVME gel as a function of temperature.

Friction coefficient (Ns/m2)

3.5 × 106

•

3.0 × 106

•

• •

2.5 × 106

•

2.0 × 106 1.5 × 106

• •

1.0 × 106

•

20

•

25

30

35

40

45

Temperature (°C)

Figure 4.6 The fluid–matrix friction coefficient (f) of the PVME gel as a function of temperature.


97

PVME stress generation (N/m2)

10000 •

5000

• • •

•

0 •

–5000 • •

–10000 20

25

30 35 Temperature (°C)

40

•

45

Figure 4.7 The generated stress of the PVME gel actuator as a function of temperature when its zero force point is chosen at T ¼ 36°C.

4.4

Pump Functioning

The schematic drawing of a controllable pump for long-term drug delivery is shown in Figure 4.8. An internal spring is regulated for the maximal drug release that is achieved at the lowest device temperature (body temperature of 36 °C). Then, to lower the drug flow as requested, the temperature of the actuator must be appropriately raised by means of the thermal heating of the resistors. Since cooling of the PVME cell happens spontaneously and the lost heat power is fixed, while the heating power can be regulated by the electrical energy dissipated into the resistors, the heating–cooling cycle is not symmetric. Therefore, if the PVME contraction can be practically as fast as desired, its relaxation time is fixed and slow. This fact limits the application of the PVME gel motor (as all thermally driven mechanisms such as shape memory alloys) to phenomena that do not have very fast kinetics. The use of Peltier’s cells that actively cool the gel mover can lead to a faster relaxation response of the actuator and increase of drug delivery. An internal programming unit able to take into account of the thermal and mechanical inertia of the whole pumping system will definitely improve the timely regulated drug outflow.

98

Biomedical Applications of Electroactive Polymer Actuators long term reservoir

36°C

long term reservoir

PVME actuator

38°C

heaters

passive spring

long term reservoir

42°C

Figure 4.8

Schematic drawing of a controllable pump for long-term drug delivery.

4.5 Conclusion In this present chapter it has been shown how a thermally controlled hydrogel actuator can be used to let a drug infusion pump execute a time-defined tasks. By using a biphasic model, the characteristics of the hydrogel mover (as force density and time response) are explicitly defined together with their functional dependence by the geometrical and material parameters. The macroporous structure of the ‘hydrogel motor’ allows a quick contractile response to temperature changes regardless of its dimension. The use of thermoelectric cooling units can shorten the thermal cooling and the elongation time of the actuator.

References 1. Hirasa, O., Morishita, Y., Onomura, R., et al. (1989) Preparation and mechanical properties of thermo-responsive fibrous hydrogels made from poly(vinyl methyl ether)s, Kobunshi Ronbunshu, 46, 661–5.


99

2. Biot, M. A. (1956) Theory of propagation of elastic waves in a fluid-saturated porous solid, I low-frequency range, J. Acoust. Soc. Am., 28-2, 168–90. 3. Johnson, D. L. (1982) Elastodynamics of gels, J. Chem. Phys, 77, 1531–9. 4. Peters, A. and Candau, S. J. (1988) Kinetics of swelling of spherical and cylindrical gels, Macromolecules, 21, 2278–82. 5. Tanaka, T. and Fillmore, D. J. (1979) Kinetics of swelling of gels, J. Chem. Phys, 70, 1214–8. 6. Chiarelli, P. and De Rossi, D. (1988) Determination of mechanical parameters related to the kinetics of swelling in an electrically activated contractile gel, Prog. Coll. Polym. Sci., 78, 4–8. 7. Chiarelli, P. and De Rossi, D. (1992) Modeling and Mechanical Characterization of Thin Fibers of Contractile Polymer Hydrogel, J. Intelligent Materials System and Structures, 3, 398–417. 8. Suzuky, M. and Hirasa, O. (1993) An Approach to artificial Muscle Using Polymer gels Formed by micro-phase Separation, Adv.Polym. Sci., 110, 241–61. 9. Chiarelli, P., Domenici, C. and Genuini, G. (1993) Crazing Dynamics in the swelling of thermally cross-linked PVA–PAA films, J. Mat. Sci.: Mat. in Med., 4, 5–11.

Section II Ionic Polymer–Metal Composites (IPMC)


5 IPMC Actuators: Fundamentals Kinji Asaka and Keisuke Oguro National Institute of Advanced Industrial Science and Technology (AIST), Japan

5.1 Introduction Ionic polymer–metal composites (IPMCs), which are composed of ionic gel polymer plated with metal electrodes, are one of the most promising electroactive polymer (EAP) materials for the artificial muscle-like actuators. The image and the schematic drawing of the actuation of the IPMC driven by low voltages are shown in Figure 5.1. When applying a voltage, the counter cation moves the cathode side with dragging water, which results in the pressure gradient for the ionic gel polymer. A large ionic current gives the IPMC actuator a soft and relatively powerful motion, as shown in Figure 5.1. IPMC actuators have number of advantages that make them attractive to use for various biomedical and human affinity applications:

Low drive voltage (1–3 V) Relatively high response (up to several hundreds of Hertz) Large response Soft material The possibility and ease to miniaturize and to mould into any shape Can be activated in water or in wet condition. Possible to work in dry condition.

Historically, direct transformation from electrical energy to mechanical work using ionic polymer gel was firstly reported by Hamlen et al. in 1965 [1]. After that, many pioneer workers investigated the electric response of ionic gels [2]. Following these workers, Oguro et al. [3] firstly reported the bending response of the perfluorosulfonic acid membrane (Nafion 117) plated with platinum electrodes, being activated by low


104


Weight

Voltage IPMC actuator 1second

Electrode

Voltage

Ionic conductive Polymer : Polymer electrolyte (Anion)

: Water molecule

(a)

: Cation

(b)

Figure 5.1 (a) Photograph of the bending performance of the IPMC actuator. The Nafion/Au composite of 0.8 mm thickness lifts 10 g weight driven by a 3 V voltage. (b) Schematic representation of the structure of the IPMC actuator.

voltages, about 1 V, in 1992, which was named IPMC [4]. Shahinpoor et al. also reported a similar idea in 1992 [5]. The IPMC actuator was more durable and it had higher response than the electric response polymer gels that were known at that time. Hence, many researchers have been applying IPMC actuators to various applications since then. The IPMC is also known as the ionic conductive polymer gel film (ICPF) [6]. Described in this chapter, are the basic aspects of IPMC actuator fabrications, measurement methods for testing, actuator performances, physics-based models and recent development of the material of the IPMC-like ionic polymer based actuators, based on our previous works. Interested readers can refer more comprehensive review articles on the IPMCs [6–9].

5.2 Fabrication 5.2.1

Ionic Polymer

Ionic polymers usually used for the IPMC are perfluorosulfonic acid or perfulorocarboxylic acid polymers, of which the typical chemical structures are shown in Figure 5.2 [10]. Commercially available products of thin films made from perfluorosulfonic acid can be obtained from E.I. Dupont de Nemours Co. (Nafion). Several other companies supply similar compounds. Asahi Glass Co. produces perfluorocarboxylic acid type (Flemion). A thin film of the perfluorinated ionic polymers can be obtained by casting their dispersing solution and evaporating the solvent, or by hot moulding a thermoplastic form of their polymers and changing –SO2F to –SO3– by hydrolysis. The device can be fabricated in any shape by using the casting solution or the thermoplastic beads of commercially available products of perfluorosulfonic acid (Nafion) (Chapter 6).

IPMC Actuators: Fundamentals –(CF2

CF2)n –(CF

105

CF2)–

O CF2 CF–CF3 O (CF2)y X

Figure 5.2 Chemical structure of prefluorinated ion exchange polymer. X ¼ SO3–: perfluorosulfonic acid polymer; COO – perfluorocarboxylic acid polymer; or SO2F: thermoplastic polymer.

Hydration of the fluorinated ion exchange resins depends on the ionic form and ion exchange density [11–13]. Figure 5.3 shows the water content of the perfluorosulfonic acid (Nafion 117 (N-117), charge density 0.91 meq./g) and perfluorocarboxylic acid membranes (Flemion (F-1.44), charge density 1.44 meq./g, and Flemion (F-1.8), charge density 1.8 meq./g) of various ionic forms. In the case of the Nafion membrane, the water content decreases as the hydrophobicity of the counter cations increases. F-1.44 has lower water content than N-117, while F-1.8 has much higher water content in the case of every counter cation. The ionic conductivity of the three kinds of membranes, which were estimated by impedance measurement, are shown in Figure 5.4. The ionic conductivity also depends on the ionic size, the hydration, the charge density of the ionic polymer, and so on. The alkali and alkali earth cation-form polymers have larger conductivity than the alkyl ammonium cation-form polymers. As the size of the alkyl ammonium cation increases, the conductivity decreases. Flemion membranes have larger conductivity than Nafion membranes. These properties, in relation to the electric bending response of IPMC actuators, are discussed in detail later. 5.2.2

Plating Methods

The plating electrodes for optimum performance of the IPMC actuator should have the following criteria:

Good adhesion to the ionic polymer High electric conductivity Large electrochemical interfacial area Large electrochemical window (high over potential to redox reactions) Inertness Softness Nontoxicity.

An established method of electrode plating on the ion exchange membrane for fulfilling the above criteria is chemical plating with platinum or gold electrodes. Oguro et al. firstly found the bending response of Nafion 117 chemically plated with platinum electrodes. However, a platinum electrode has a narrower electrochemical window and is mechanically

Biomedical Applications of Electroactive Polymer Actuators g water/g dry membrane (%)

106

35 30 25 20 15 10 5 0

Li

a

N

K

s

C

C

a

Ba

g

M

A

TE

A

A rA TE TP

A TB

A

A rA TE TP

A TB

TM

A

r TP

A

A

TB

water content g water/g dry membrane (%)

(a) 30 25 20 15 10 5 0

Li

a

N

K

C

s

C

a

g

M

Ba

TM

g water/g dry membrane (%)

(b)

60 50 40 30 20 10 0

Li

a

N

K

C

s

C

a

g

M

Ba

TM Membrane Form (c)

Figure 5.3 Water content of (a) Nafion 117, (b) Flemion F-1.44 and (c) Flemion F-1.8 membranes of various ionic forms. Abbreviations used for alkyl ammonium ions are TMA, TEA, TPrA, TBA for tetramethyl, tetraethyl, tetrapropyl and tetrabutyl ammonium ions, respectively (Reprinted with permission from Asaka, K., Fujiwara, N., Oguro, K. et al. State of water and ionic conductivity of solid polymer electrolyte membranes in relation to polymer actuators, J. Electroanalytical Chem., 505 (1–2), 24–32. Copyright (2001) Elsevier).

harder than a gold electrode. Hence, a chemical plating method with a gold electrode has been developed for the IPMC actuator [14]. Two different methods, known as ‘reductant permeation’ (RP) and ‘impregnation reduction’ (IR), have been successfully developed for plating the electrodes for the ion exchange

IPMC Actuators: Fundamentals 0.1

K N-117

Na

Li

0.001 TEA

Mg Ca Ba

membrane conductivity (λ m /S cm–1)

membrane conductivity (λ m /S cm–1)

0.01

Cs

TMA

0.0001 –5

10

10–6

107

TPrA

TBA

F-1.44 NH4

0.0001

TPrA

TEA

Cs

Ca

TMA

0.001

K

Na

Li

0.01

Mg Ba

TBA

10–7 10 20 30 40 50 60 70 80 limiting equivalent conductivity (λ w /Scm2equiv–1)

10–5 10 20 30 40 50 60 70 80 limiting equivalent conductivity (λ w /Scm2equiv–1)

(a)

(b)

membrane conductivity (λ m/S cm–1)

0.1 F-1.8 K

Na

0.01

Li

Ca

TMA

TPrA

Cs

Mg

TEA

Ba

0.001 TBA

0.0001 10 20 30 40 50 60 70 80 limiting equivalent conductivity (λ w /Scm2equiv–1)

(c)

Figure 5.4 Dependence of the membrane conductivity of Nafion 117 (a), F-1.4(b), and F-1.8 (c) membranes of various ionic forms on the limiting ionic conductivity of each ion in water. Abbreviations used for alkyl ammonium ions in the figure is the same as Figure 5.3 (Reprinted with permission from Asaka, K., Fujiwara, N., Oguro, K. et al. State of water and ionic conductivity of solid polymer electrolyte membranes in relation to polymer actuators, J. Electroanalytical Chem., 505 (1–2), 24–32. Copyright (2001) Elsevier).

resins under wet conditions. In the RP method, a metal layer is formed on the surface of a membrane by permeation of reducing agents into the other side of the membrane when a metal complex solution and a reducing solution are placed on either side of the membrane, respectively. In the IR method, the cation exchange membrane with pre-exchanged cationic metal species is subsequently immersed in the reducing solution, which reduces and displaces the metal toward the outer surfaces of the membrane. The IR method is known to be better than the RP method for achieving the above criteria. Fujiwara et al. [14] firstly developed the gold plating method on the surfaces of the ion exchange membrane by using the IR method. The schematic representation of the chemical plating of gold on the surfaces of the ion exchange resin by the IR method is shown in Figure 5.5. After roughening the surface of the ion exchange resin membrane by dry blasting or emery paper, the ion exchange polymer is

108

Biomedical Applications of Electroactive Polymer Actuators Ion-exchange polymer

Au [AuL]+ SO3–

SO3– H+ +[AuL]+

H+ SO3–

H+

–H+

SO3–

SO3–

SO3–

[AuL]+

+X+

[AuL]+

–L

SO3–

SO3–

1.Ion-exchange

SO3–

X+ X+ SO3–

SO3–

X+

[AuL]+

H+

Au X+

2. Reduction

SO3– 3. Sequential

+ [AuL]+ : N

N Au Cl

Cl

Figure 5.5 Chemical plating of gold onto the surface of the ion exchange polymer using a cationic gold complex such as dichlorophenanthrolinegold (III) (Reprinted with permission from Fujiwara, N., Asaka, K., Nishimura, Y. et al. Preparation of Gold-Solid Polymer Electrolyte Composites As Electric Stimuli-Responsive Materials, Chem. of Mat., 12 (6), 1750–4. Copyright (1999) American Chemical Society).

immersed in the aqueous solution of gold complex salt. Then, the metal complex adsorbed in the ion exchange polymer is reduced by the reducing agent. The sequential plating technique has been developed for optimizing the electrode structure, which has large electrochemical area and soft mechanical property.

5.3 Measurement The measuring setup for the displacement of the IPMC actuator in response to electric signals is shown in Figure 5.6(a). Typically, the actuator strip – 15 mm length and 1 mm width – is clamped by two gold disks and the displacement at a point 10 mm away from the fixed point (free length) is measured by the laser displacement meter. The applied voltage and electric current are simultaneously measured. The curvature is evaluated from the measured displacement () by the following equation: 1=R ¼ 2=ðl2 þ 2 Þ;

ð5:1Þ

where R is the curvature radius and l is the free length. The blocking force is also measured by the setup, as shown in Figure 5.6(b). The relationship between the bending force (F) and the curvature (1/R) is: 1=R ¼ ðM

FlÞ=EI

ð5:2Þ

IPMC Actuators: Fundamentals

109

Load cell

Waveform generator/ potentiostat

Waveform generator/ potentiostat Electrode

Electrode Actuator element

Curvature radius: R Laser displacement meter

Laser

Actuator element

l

Force

Displacement: δ

(a)

(b)

Voltage (V)

1 0 –1

Current (mA)

20 0 –20

Displacement (mm)

Displacement (mm)

Current (mA)

Voltage (V)

Figure 5.6 Schematic representations of the measuring setup for the performance of the IPMC actuator. (a) Displacement measurement. (b) Force measurement.

0.2 0.0 –0.2 0

10

20

30

40

50

60

70

80

90

1 0 –1 20 0 –20

0.2 0.0 –0.2 0

10

20

30

40

50

60

Time (s)

Time (s)

+ a. Na form

b. (C4Hg)4N+ form

70

80

90

Figure 5.7 Performance of the Flemion/Au actuator: (a) Naþ form; (b) (C4H9)4Nþ form (Reprinted with permission from Onishi, K., Sewa, S., Asaka, K. et al. The effects of counter ions on characterization and performance of a solid polymer electrolyte actuator, Electrochimica Acta, 46 (8), 1233–41. Copyright (2001) Elsevier).

110


where M is the bending moment that is a function of input voltage, EI is the rigidity of the IPMC. In the case of the displacement measurement, Equation (5.2) is: 1=R ¼ M=EI

ð5:3Þ

since F ¼ 0. In the case of the blocking force measurement, Equation(5.2) is: F ¼ M=l

ð5:4Þ

since 1/R ¼ 0. The bending moment as a function of input voltage can be derived by the microelectromechanical transfer model, which is described in Section 5.5.

5.4 Performance of the IPMC Actuator During the past decade, much work has been done on the development of the IPMC actuator. In this section, typical examples of our previous experimental results with the Nafion/Au and Flemion/Au actuators are summarized. Shown in Figure 5.7 are the experimental curves of the input voltage, the electric current and the displacement of the Flemion/Au composite of the sodium ion (Naþ) form (a) and the tetrabutyl ammonium ion (C4H9)4Nþ from (b) [15]. The displacement response of the Nafion 117/Au composite actuators of the lithium (Liþ), Sodium (Naþ), Cesium (Csþ) and tetraethyl ammonium (TEA) ion (C2H5)4Nþ forms are shown in Figure 5.8 [16]. The displacement curves of alkali ion form of both Nafion and Flemion are a typical response of the IPMC actuator in application of step voltages. When applying the step voltages, the actuator strip bends to the anode side quickly (within 0.1 s), then slowly back to the cathode side (back motion), and finally stops in a stable position. When using an anion exchange membrane as the ionic polymer, the reverse behaviour takes place [17]. The back motion depends on the counter cation species of the Nafion or Flemion membranes as shown in Figures 5.7 and 5.8. When bulky alkyl ammonium cations, such as tetraethyl ammonium or tetrabutyl ammonium, are used, there can be no back motion as shown in Figures 5.7b and 5.8. The ionic forms of the Nafion and Flemion affect the response speed and the bending amplitude of the electromechanical response of the IPMC actuator as shown in

Displacement (mm)

8 6 4 2 0 0 –2 –4 –6 –8

Li Na Cs

TEA 0.1

0.2

0.3

0.4

0.5

Time (s)

Figure 5.8 Displacement response of the Nafion/Au actuator with Liþ (Li), Naþ (Na), Csþ (Cs) and (C2H5)4Nþ (TEA) as counter ions [16].


111

Figures 5.7 and 5.8. Alkyl ammonium ion IPMCs have larger displacement and slower response speed than the alkali metal ion IPMC. From these experimental results, a qualitative model of the IPMC actuator is proposed based on the ion cluster structure shown in Figure 5.9. It is well known that the fluorinated ion exchange polymer swollen with water has a hydrophilic channel-linked ion cluster structure surrounded with a hydrophobic perfluoro backbone polymer network. Counter ions and water transfer through hydrophilic narrow channel-linked ion cluster. If a hydrophilic small cation, such as Liþ, Naþ, Kþ and so on, transfers through the channel, large mobility and little electro-osmosis take place, which result in higher response and smaller displacement. In the case of hydrophobic large cations, such as TEA and TBA, the reverse occurs. It has been reported that the normalized displacement per charge of the Nafion 117/Pt was proportional to the water transference coefficient of the counter cation [18, 19]. These results will be explained quantitatively in the next section based on the micromechanical model of electric responsive ionic gel.

Figure 5.9 Schematic drawing of the ion cluster structure of the fluorinated ion exchange polymers with (a) alkali metal counter ion and (b) alkyl ammonium counter ion forms (Reprinted with permission from Onishi, K., Sewa, S., Asaka, K. et al. The effects of counter ions on characterization and performance of a solid polymer electrolyte actuator, Electrochimica Acta, 46 (8), 1233–41. Copyright (2001) Elsevier).

The structure of the plated electrode also affects the performance of the IPMC actuator [20]. Images of the cross-section of the Flemion/Au composite plated by different plating cycles are shown in Figure 5.10. By adsorption–reduction cycling, a fractal-like structure of gold with high interfacial area within the membrane was obtained. Figure 5.11 shows the plots of the electric double-layer capacitance of the Flemion/Au composites against the number of plating steps, which were estimated by the cyclic voltammogram. The double-layer

112


100 µm

100 µm

20 µm

20 µm

100 µm

20 µm

20 µm 8 plating cycles

6 plating cycles

4 plating cycles

2 plating cycles

100 µm

Figure 5.10 Scanning electron micrograph of the cross-section for the Flemion/Au composite of different plating cycles (Reprinted with permission from Onishi, K., Sewa, S., Asaka, K. et al. Morphology of electrodes and bending response of the polymer electrolyte actuator, Electrochimica Acta, 46 (5), 737–43. Copyright (2001) Elsevier).

Capacitance (µF cm–2)

2500 No.1

2000

No.2 1500 1000 500 0

0

1

2 3 4 5 6 7 Number of Plating Steps

8

Figure 5.11 Plots of the double-layer capacitance of the Flemion/Au composite of different plating cycles (Reprinted with permission from Onishi, K., Sewa, S., Asaka, K. et al. Morphology of electrodes and bending response of the polymer electrolyte actuator, Electrochimica Acta, 46 (5), 737–43. Copyright (2001) Elsevier).


113

capacitance increases as the number of plating steps, since the effective electrode area increases, as shown in the SEM images in Figure 5.10. The bending amplitude is proportional to the accumulated charge as shown in the model. Hence, the displacement is proportional to the double-layer capacitance. By the development described above, an IPMC actuator having an optimized performance for a specific application, such as a quick response (Figure 5.12a) and a very large response (Figure 5.12b), can be fabricated.

a. Na+ form

b. (C4Hg)4Na+ form

Figure 5.12 Images of electric response of the Flemion/Au composite actuator in water: (a) Naþ form (b); (C4H9)4Nþ form (b) in water (Reprinted with permission from Onishi, K., Sewa, S., Asaka, K. et al. The effects of counter ions on characterization and performance of a solid polymer electrolyte actuator, Electrochimica Acta, 46 (8), 1233–41. Copyright (2001) Elsevier).

5.5 Model In order to explain the actuation behaviour and develop its performance, many workers have modelled the IPMC actuator. Some have developed a black box model, in which the IPMC actuator is considered as a black box, and the response function for determining the relationship between input and output [21–24]. The black box model is useful for applications. However, we cannot understand the mechanism of the electromechanical response of the IPMC. In this section, a physics-based IPMC model based on the electro-responsive gel theory is introduced. We proposed an IPMC model, in which the bending response is attributed to the electro-osmotic flow in the ion gel film, in 2000 [19], taking into account only the water flow in the membrane. In the same year, de Gennes et al. [25] gave a more comprehensive theory using the phenomenological equation for the electric current ( je) and the water flux ( js) based on the irreversible thermodynamics: je ¼ e r js ¼ rp

lrp lr

ð5:5Þ ð5:6Þ

114


where e is the conductance, is Darcy’s permeability and l is Onsager’s coupling constant. The actuation and sensor properties were qualitatively discussed by de Gennes et al., but they have not done any quantitative analysis based on their equations. Yamaue et al. [26] developed the electrostress diffusion coupling (ESDC) model as more formal theory for the deformation dynamics of ion gels under electric field. The theory is a straightforward extension of the stress diffusion coupling model, which was proposed by Doi in 1990 [27]. In their formulation, by discussing the coupling of the network stress with the solvent permeation of the gel network, adding a kinetic equation for the ion flux and electric potential, a complete set of equations for the gel deformation was derived. In their model, the Onsagar coefficients in Equations (5.5) and (5.6) are represented by microscopic parameters: e ¼

l¼

¼

cp q2p X ci q2i þ ; p i i

qp ð1 p ð1

p Þ þ

ð5:7Þ

X ci qi wi ; i i

p Þ 2 X ci w2i þ ; cp p i i

ð5:8Þ

ð5:9Þ

where suffixes p, s and i denote polymer, solvent and ions, respectively, and c, q, z and w represent the concentration (number of molecules per unit volume), charges, friction constant related to solvent and the specific volume, respectively. The general model was applied to the ion cluster structure of fluorinated ion exchange membrane shown in Figure 5.9. The friction constant of free ions is given by the Stokes– Einstein law: i ¼ 6pai ;

ð5:10Þ

where is the viscosity of solvent and ai is the ion radius. To estimate zp, it was assumed that the gel consists of microchannels of characteristic length zb, which correspond to the diameter of the microhydrophilic channel shown in Figure 5.9. Then, the friction constant of polymer gel is given by: p ¼

6p : 2b cp

ð5:11Þ

Using Equations (5.10) and (5.11), the Onsager coefficients are written: cp q2p b 3 cp b þ e ¼ 6pb ai cp qp 2b l¼ 6p

(

ð1

p Þ

4p 3

ai b

2 )

ð5:12Þ

ð5:13Þ


¼

ð1

8 p Þ 2 2b < 1þ : 6p

4p 3ð1

p Þ

!2

cp 3b

ai b

9 5 = : ;

115

ð5:14Þ

Equations 5.12 and 5.13 show that the conductivity decreases and the electro-osmosis increases as the ion size increases in relation to the channel size. This is qualitatively in agreement with the displacement experiments shown in Figures 5.7 and 5.8. The initial curvature is given by the equation: 1 4 l ¼ Q RðtÞ h2 e

ð5:15Þ

R–1(ai/ξb,t = 0)/R–1(ai/ξb = 0.1,t = 0)

which means that initial curvature is driven by the pressure gradient in the ion gel due to the electro-osmotic flow. Hence, the initial curvature is a function of ionic charge (Q) and the water transference coefficient (l/e), which represents how many water molecules transfer per one counter cation transfer. Figure 5.13 shows the normalized value of initial curvature of the IPMC actuator (Nafion 117/Au) of various ionic forms, together with the theoretical curves. Experimental and theoretical values are in good agreement each other.

10 : experiment λ = λi : theory z=1

8

λ = λi z=2

6

λ = λp + λi z=1

5 Li+

λ = λp + λi

2 H+

0

0

0.1

K+

Ca++

Na+

0.3

0.2

z=2

0.4

0.5

ai/ξb

Figure 5.13 Calculated initial curvature of various counter ions to that of proton (Hþ): R(Hþ)/ R(0) (Reprinted with permission from Yamaue, T., Mukai, H., Asaka, K. and Doi, M., Electrostress Diffusion Coupling Model for Polyelectrolyte Gels, Macromolecules, 38 (4), 1349–56. Copyright (2005) American Chemical Society).

One more important conclusion derived from ESDC theory concerns the back motion of the IPMC actuator. As a result of the initial charging of the electric double-layer, a large pressure gradient is created near the electrodes, which results in the initial curvature as given by Equation (5.15). Accordingly, the water starts to flow from the high pressure region to the low pressure region. As a result, the IPMC

116


starts to bend back. According to Yamaue et al., the characteristic time of the back motion is given by: relax ¼

h2 ¼ D0

2

l e

h2

4 Kþ G 3

ð5:16Þ

D’ is the collective diffusion coefficient of the fluorinated ion exchange polymer, which is given by the effective solvent permeability coefficient and the mechanical constant of the polymer network. The effective solvent permeability coefficient depends on the ratio of the

2

log10 (τrelax/τrelax, H+)

: Experiment : Theory

TEA+

1.5 Z=1 Z=2

1 Ca++

0.5 Na+

0

H+

0

Li+

K+

0.1

0.3

0.2

0.4

0.5

ai/ξb

Figure 5.14 Calculated ratio of the relaxation time for various counter ions to that of proton (Hþ): relax / relax,Hþ (Reprinted with permission from Yamaue, T., Mukai, H., Asaka, K. and Doi, M., Electrostress Diffusion Coupling Model for Polyelectrolyte Gels, Macromolecules, 38 (4), 1349–56. Copyright (2005) American Chemical Society).

ion radius to the diameter of the microchannel. Hence, the characteristic time of the back motion depends on the counter cation, as shown in Figure 5.14. Experimental and theoretical results are good agreement with each other. If large cations such as TEA and TBA are used as counter cations, the characteristic time is infinitely long. Hence, the back motion of the IPMC of alkyl ammonium form cannot be observed. It is considered that the static response of the IPMC actuator can be attributed to the electrostatic and osmosis effect due to the electric double-layer charging. A theoretical model in consideration with these effects has also been developed [28].

5.6 Recent Developments Recently, applications of room temperature ionic liquids to electrochemical devices are very attractive in both scientific and technological fields. In the area of the EAP


117

actuators, it is very promising to use ionic liquids as a nonvolatile and high conductivity electrolytes for the ion conductive material of the conductive polymer based actuator [29]. Leo and co-workers [30] have developed a method of using ionic liquids as solvents for the IPMC actuators and sensors in order to use them in air. They successfully fabricated the IPMC actuator using the ionic liquid working in air for a long time. Fukushima et al. [31] have developed a dry actuator that can be fabricated simply through layer-by-layer casting with bucky gel, a gelatinous mixture composed of ionic liquid and a single-walled carbon nanotube. A configuration of the actuator, which has a bimorph structure with a polymer-supported internal ionic liquid layer sandwiched by bucky gel electrode layers, is shown in Figure 5.15.

Figure 5.15 Schematic drawing of a configuration of the bucky gel actuator (Reprinted with permission from Fukushima, T., Asaka, K., Kosaka, A. and Aida, T. Fully Plastic Actuator through Layer-by-Layer Casting with Ionic-Liquid-Based Bucky Gel, Ang. Chem. Int. Ed., 44 (16), 2410–3. Copyright (2005) Wiley-VCH Verlag GmbH).

The actuator film was fabricated by layer-by-layer casting of electrode-layer (singlewalled carbon nanotubes (SWNTs) and ionic liquids (ILs)) and electrolyte-layer (ILs) components in a gelatinous mixture of poly(vinylidene fluoride-co-hexafluoropropylene) (PVdF(HFP)) as a polymer support and a solvent. The actuator can be activated by low voltage (SO3– in KOH (0.15)/DMSO(0.35)/H2O (0.5) at 70 °C

Crystallization at 100–170 °C Immersion in fresh water for more than one night Conditioning in HCl/H2O2 solution at 70 °C Immersing in 12 % HNO3 (a)

(b)

Figure 6.3 Scheme for the preparation of the Nafion polymer by casting (a) and hot moulding (b) methods.

124


films of various thicknesses can be obtained easily. Preparing the Nafion film connecting with other materials can also be done by the same methods. The standard method of electrode plating for the IPMC is chemical plating of gold or platinum as described in Chapter 5. Recently, several physical technologies such as printing or pressing using metal or metal oxide nanoparticles were reported [15, 16]. In order to make an IPMC device that can move in multi degrees of freedom, the plating electrode needs to be patterned. In the laboratory, the plating electrode can be patterned by cutting the plating electrode by PC-controlled mechanical or optical (laser) cutting machine. Finally, a soft and smart device is fabricated by integrating the IPMC actuator with control and senor units. Recently, organic transistor control and sensor technologies have been developed. Hence, all organic, flexible electromechanical devices can be constructed using the IPMC and an organic transistor. In this chapter, such an example is described as a sheet-type Braille display. It was reported that the IPMC has a function of not only an actuator but also a sensor, like a piezo-material [17]. Yamakita et al. [18] studied the sensor property of Nafion/Au and developed the IPMC sensor/actuator system. The response of the IPMC actuator was controlled by the feedback of the same IPMC sensor signal.

6.3 Applications to the Microcatheter Catheter based diagnosis and therapy have become increasingly popular. Conventional catheters cannot move actively and the operations require certain human skill. If the direction of the tip of microcatheter and guide wires can be controlled outside the body, it would be very useful to use catheter based diagnosis and therapy. Such catheters are called ‘active catheters’. In order to develop active catheters, several micro-actuators were applied to the active microcatheter system. The IPMC actuator is soft and driven by low voltage. And it is easy to miniaturize. Hence, the IPMC actuator was applied to the active microcatheter for aneurysm surgery in the brain. Guo et al. [8, 9] developed the microcatheter system with active guide wire. They proposed a prototype model of microcatheter with active guide wire, which has two bending degrees of freedom. Prototype models were 3Fr, 4Fr, 6Fr (1Fr ¼ 1/3 mm) in diameter and consisted of catheter tube and active guide wire with IPMC actuator (Nafion 117 plated with platinum) on its front end as the servo actuator (Figure 6.4). The bending motions and bending angle of the active microcatheter system were measured by application of electricity in physiological saline solution using laser displacement sensors. They also carried out the modelling of this active microcatheter system, which is reasonable for practical applications. By using simulators, they also carried out simulation experiments in vitro using various blood vessel simulators. The system of blood vessel simulators consists of a blood vessel simulator, a pump for circulating physiological saline, an instrument for measurement and a heater (Figure 6.5). The specifications of blood vessel simulator are 4 mm, 5 mm and 8 mm in internal diameter, turning angle of 45–95° and sectional diameter of 2–8 mm in aneurysms. The catheters used for experiments are 6Fr, 4Fr, 3Fr in outer diameter. When the temperature of

Active Microcatheter and Biomedical Soft Devices Based on IPMC Actuators Lead wire

Micro tube

125

Electrode

φ dg

ICPF

Bond Length

Figure 6.4 Prototype model of a microcatheter system with active guide wire using IPMC (ICPF) actuator (Reprinted with permission from Guo, S., Fukuda, T., Kosuge, K. et al. Micro catheter system with active guide wire-structure, experimental results and characteristic evaluation of active guide wire catheter using ICPF actuator, Proceedings of the IEEE 5th International Symposium on Micro Machine and Human Science, 191–8. Copyright (2004) IEEE).

Instrument for measurement

Infusion pump

Heater

Turning point

Aneurysm Simulator Catheter

Figure 6.5 Schematic drawing of a system of blood vessel simulator for testing a microcatheter with the IPMC active guide wire (Reprinted with permission from Guo, S., Fukuda, T., Kosuge, K. et al. Micro catheter system with active guide wire-structure, experimental results and characteristic evaluation of active guide wire catheter using ICPF actuator, Proceedings of the IEEE 5th International Symposium on Micro Machine and Human Science, 191–8. Copyright (2004) IEEE).

physiological saline ranges from 20 to 36 °C, and the flow rate ranges between 50 and 650 ml/min, the inserting motion into each aneurysm and at divergence’s were confirmed as shown in Figure 6.6. According to Guo et al., these experimental results in vitro indicate that the proposed catheter with active guide wire works properly, and that it can improve the effectiveness of traditional procedure for intracavity operations. If the microcatheter itself is moveable, active microcatheter system is more compact and easier to control than the active guide wire. Oguro et al. [10] developed the tubular IPMC actuator with four electrodes and active microcatheter system without guide wire using the tubular IPMC actuator. Figure 6.7 shows a schematic drawing of the tubular IPMC actuator with four electrodes around the tube, which can drive the tubular resin to bend in multiple directions. The four electrodes were fabricated by laser ablation cutting after the chemical plating of a gold electrode. In order to bend the tubular actuator of 0.8 mm outer diameter for 90°, the perfluorocarboxylic acid polymer (Flemion) was used for the ionic gel instead

126


(b)

(c)

(d)

Figure 6.6 Bending motion of the active guide wire into aneurysm in blood vessel simulator (Reprinted with permission from Guo, S., Fukuda, T., Kosuge, K. et al. Micro catheter system with active guide wire-structure, experimental results and characteristic evaluation of active guide wire catheter using ICPF actuator, Proceedings of the IEEE 5th International Symposium on Micro Machine and Human Science, 191–8. Copyright (2004) IEEE).

Figure 6.7 Tubular IPMC actuator for controlling active microcatheter (Reprinted with permission from Oguro, K. et al. Proceedings of the SPIE 6th International Symposium on Smart Structures and Materials, Newport Beach, 3669, 64–71. Copyright (1999) SPIE).

of the perfluorosulfonic acid polymer (Nafion). By applying the sequential plating of the gold electrode and optimizing the counter cation as alkyl ammonium cation, we successfully developed a tubular actuator for the active catheter. The motion of the tube to all directions can be controlled with combined signals applied to four electrodes. For practical usage of the tip of active microcatheter for intravascular neurosurgery as shown in Figure 6.8, the tubular actuator of external diameter of 0.8 mm in the swollen state was attached to the end of catheter. The actuator is 2 cm long and the total length of the catheter is 1.5 m. Four conductive layers were made on the outside of the long

Active Microcatheter and Biomedical Soft Devices Based on IPMC Actuators

127

Figure 6.8 Schematic representation of active catheter system using the tubular IPMC actuator for intravascular neurosurgery (Reprinted with permission from Oguro, K. et al. Proceedings of the SPIE 6th International Symposium on Smart Structures and Materials, Newport Beach, 3669, 64–71. Copyright (1999) SPIE).

catheter with gold paste. The tip of the active catheter bent more than 90° within 10 seconds in physiological saline without gas evolution. Intravascular in vivo tests of the active microcatheter with an animal were also carried out.

6.4 Other Applications 6.4.1

Sheet-Type Braille Display

The IPMC actuator is soft, flexible and a low voltage drive. Hence, it is very safe for humans. If the IPMC film can be successfully patterned and integrated, and each microactuator strip can be controlled separately, a human-affinity tactile information communication system can be developed. Someya et al. [19] developed a large-area, flexible and lightweight sheet-type Braille display, integrating the IPMC actuator with their high quality organic transistors. Images of the Braille display are shown in Figure 6.9. An array of the rectangular IPMC actuator is mechanically processed using a numerically controlled (NC) cutting machine to form an array of 12 12 rectangular actuators whose size is 1 4 mm2. A small semi-sphere, which projects upward from the rubber-like surface of the display, is attached to the tip of each rectangular actuator. The effective display size is 4 4 cm2. Each Braille letter consists of 3 2 dots and 24 letters can be displayed. The total thickness and the weight of the entire device are 1 mm and 5.3 g, respectively.

128

Biomedical Applications of Electroactive Polymer Actuators (b)

Cover layer

(a)

(c) WL1

Transistor sheet

Actuator sheet

WL8

BL1

BL18

Figure 6.9 Images of a Braille display: (a) the flexible Braille sheet display using IPMC actuator array; (b) the device assembly; (c) circuit diagram of the Braille sheet display (Reprinted with permission from Kato, Y., Sekitani, T., Takamiya, M. et al. Sheet-Type Braille Displays by Integrating Organic Field-Effect Transistors and Polymeric Actuators, IEEE Trans on Electron. Devices, 54 (2), 202–9. Copyright (2007) IEEE).

Figure 6.10 A cross-sectional illustration of a single Braille dot. An organic transistor is connected to an IPMC actuator with silver paste patterned by a microdispenser (Reprinted with permission from Kato, Y., Sekitani, T., Takamiya, M. et al. Sheet-Type Braille Displays by Integrating Organic Field-Effect Transistors and Polymeric Actuators, IEEE Trans on Electron. Devices, 54 (2), 202–9. Copyright (2007) IEEE).


129

Each IPMC actuator is connected to one organic transistor. The circuit diagram of the Braille sheet display is shown in Figure 6.9c. The vertical and the horizontal lines represent bit and word lines, respectively. Figure 6.10 shows the cross-sectional structure of a single Braille cell composed of one transistor and one actuator. When a voltage is applied to the IPMC actuator, the IPMC actuator bends (Figure 6.10). The semi-sphere placed on the actuator rises with the voltage supply and pushes up a rubber-like surface. An organic transistor active matrix is used to address the pop-up dots. Figure 6.11a shows one of the Braille dots moving upwards and downwards. Four Braille letters displayed by the present device are also shown in Figure 6.11b. Someya et al. tested the readability of the sheet-like Braille display using the IPMC. Four visually impaired individuals participated in the reading tests. When the operator input ‘Na’ and ‘Wa’ in the Japanese Braille format, all four individuals were able to recognize the letters correctly.

Figure 6.11 Display operation. (a) Magnified pictures of one Braille dot moving upwards and downwards. The scale is 1 mm. (b) Pictures of the Braille sheet display showing the characters ‘l’, ‘w’, ‘b’ and ‘f’ in the American Braille style (Reprinted with permission from Kato, Y., Sekitani, T., Takamiya, M. et al. Sheet-Type Braille Displays by Integrating Organic Field-Effect Transistors and Polymeric Actuators, IEEE Trans on Electron. Devices, 54 (2), 202–9. Copyright (2007) IEEE).

130


This result shows that the generating force and the displacement of this device are sufficient to facilitate reading by the visually impaired. Further, it demonstrates the feasibility of this design, which integrates organic transistors and the IPMC actuators, for realizing Braille sheet displays. The issues that remain to be addressed are the readability and stability of the device. The IPMC actuators are usually operated in wet conditions, while organic transistors degrade easily in moisture and/or oxygen. The straightforward approach toward suppressing such degradation is to employ more sophisticated encapsulation techniques. Alternately, the use of the ionic-liquid-gel based polymer actuator technology, which can be operated in ambient air, is an option. 6.4.2

Underwater Microrobot

An underwater microrobot is very attractive for developed medical practice, both for diagnosis and for surgery. For instance, with medical technology a common application is to perform a delicate surgical operation supported using micromachines, thus avoiding unnecessary incisions. Microrobots can restrict their work to the affected part or the breakdown spot and do not unnecessarily influence their surroundings. Gou et al. [2–4] developed a fish-like underwater microrobot, using the IPMC actuator. Figure 6.12 shows a schematic representation of the structure of the microrobot, which consists of the body made of wood materials shaped as a fish, a pair of fins driven by the IPMC (ICPF) actuator, respectively, connecting the lead wires for supplying the electrical energy to the IPMC actuator. The IPMC actuator used was Nafion 117 plated with platinum or gold. The two fins were driven independently. The direction of swimming can be controlled by the frequency of both fins. Figure 6.13 shows photographs of the microrobot. Its overall size is 45 mm in length, 10 mm in width and 4 mm in thickness. The swimming characteristic of the underwater microrobot was measured by changing the frequency of input voltage from 0.1 to 5 Hz in water and the amplitude of the input voltage from 0.5 to 10 V. The experimental results indicate that Fin

ICPF actuator

Electrode Buoyancy adjuster (D) V+

Body (A)

V+

Left

Right

Com (C)

Electrode

ICPF actuator(B)

Fin

Figure 6.12 Schematic drawing of the structure of a microrobot using the IPMC (ICPF) actuator (Reprinted with permission from Guo, S., Fukuda, T. and Asaka K., A new type of fish-like underwater microrobot, IEEE/ASME Trans on Mechatronics, 8 (1), 136–41. Copyright (2004) IEEE).


131

Figure 6.13 Images of the microrobot using the IPMC actuator (Reprinted with permission from Guo, S., Fukuda, T. and Asaka K., A new type of fish-like underwater microrobot, IEEE/ASME Trans on Mechatronics, 8 (1), 136–41. Copyright (2004) IEEE).

changing the amplitude and the frequency of input voltage can control the swimming speed of the underwater microrobot using the IPMC actuator. Guo et al. [20] developed the micropump using the IPMC actuator as shown in Figure 6.14. A photograph of the prototype model of the micropump, which has a size of 10 mm in diameter and 20 mm in length, is shown in Figure 6.15. A flow of 4.5–37.8 ml/min was successfully obtained from the micropump by changing the frequency of the applied voltage. The micropump using the IPMC actuator, as shown in Figure 6.14, is able to make a microflow and is silent for driving, which is suitable for biomedical uses. Nakabo et al. [14] developed a snake-like swimming robot with a patterned-electrode IPMC. The aim is that a snake-like motion sweeps a smaller area than simple bending swimming. Thus, it is suitable for future swimming robots in thin tubes, such as blood vessels, as shown in Figure 6.16. A photograph of the patterned-electrode IPMC for the snake-like swimming robot is shown in Figure 6.17. The IPMC used in this study was a Nafion 117 membrane chemically plated five times with gold. The plated electrode was separated into seven segments by cutting the plated gold using a small hand chisel. For the swimming experiment using the

132

Biomedical Applications of Electroactive Polymer Actuators ICPF actuator

One-way valve

(a) Liquid from the Inlet

(b) Liquid from the Outlet

Figure 6.14 Mechanism of a micropump using the IPMC actuator (Reprinted with permission from Guo, S., Nakamura, T., Fukuda, T. and Oguro, K. Development of the micropump using ICPF actuator, Proceedings of the 1997 IEEE International Conference on Robotics and Automation, 1, 266–71. Copyright (1997) IEEE).

Figure 6.15 Image of a micropump using the IPMC actuator (Reprinted with permission from Guo, S., Nakamura, T., Fukuda, T. and Oguro, K. Development of the micropump using ICPF actuator, Proceedings of the 1997 IEEE International Conference on Robotics and Automation, 1, 266–71. Copyright (1997) IEEE).

snake-like IPMC robot, seven connectors and electric wires touched each segment with floats, which prevent the IPMC robot from sinking. The applied travelling wave voltage propels the snake-like IPMC robot in the water. Figure 6.18 shows a photograph of the propulsions of the IPMC snake-like robot. An optimal condition for propulsion was studied by changing the frequency and phase shift of the applied travelling wave voltage. The maximum speed was obtained at the frequency of 2 Hz. The propelling direction of the IPMC (forward or backward) was successfully

Active Microcatheter and Biomedical Soft Devices Based on IPMC Actuators Get Current position

133

Indicate objective position

Control of swimming

Thin tube Communicator, Patterned (blood vessel etc.) wave generator artificial muscle

Figure 6.16 Concept of a snake-like swimming robot in a thin tube (Reprinted with permission from Nakabo, Y. et al. Biomimetic soft robots using IPMC in Electroactive Polymers for Robotic Applications (eds Kim, K. J. and Tadokoro, S.), 165–98. Copyright (2007) Springer).

Figure 6.17 Patterned IPMC actuator film (Reprinted with permission from Nakabo, Y. et al. Biomimetic soft robots using IPMC in Electroactive Polymers for Robotic Applications (eds Kim, K. J. and Tadokoro, S.), 165–98. Copyright (2007) Springer).

Figure 6.18 Forward and backward propulsions of the snake-like robot using the patternedelectrode IPMC actuator (Reprinted with permission from Nakabo, Y. et al. Biomimetic soft robots using IPMC in Electroactive Polymers for Robotic Applications (eds Kim, K. J. and Tadokoro, S.), 165–98. Copyright (2007) Springer).

134


controlled by changing the direction and speed of propagation of waves along the body by advancing and delaying the phases of the sine waves. The maximum speed was obtained at the phase shift of 60°. The snake-like IPMC robot was also successfully turned right or left by applying a saw-tooth wave voltage. Nakabo et al. presented a propulsion model of the snake-like swimming motion to explain results of the propulsion and turning of the patterned-electrode IPMC robot. 6.4.3

Linear Actuators for a Biped Walking Robot

The IPMC actuator basically has a bending motion. In order to apply the robotics, an actuator that has a linear motion (linear actuator) is often very practical. Yamakita et al. [11–13] developed a linear actuator using the IPMC actuator. The proposed linear actuator is composed of many basic units connected in parallel and series so that enough force and displacement can be obtained (Figure 6.19). The elementary unit consists of four IPMC films (Figure 6.20). One side of the unit is formed from a pair of films that are connected by a flexible material or the same thin film. When an input voltage is applied to electrodes on the surface with the anode outside, each membrane bends outwards, then the actuator is constricted. The actuation force and displacement of each unit are small; however, the elementary units can be connected in parallel and series, so the actuator can realize the desired force and displacement. elementary unit

Figure 6.19 Concept of a linear actuator using the IPMC actuator (Reprinted with permission from Yamakita, M. et al., Robotic Application of IPMC Actuators with Redoping Capability in Electroactive Polymers for Robotic Applications (eds Kim, K. J. and Tadokoro, S.), 199–225. Copyright (2007) Springer).

Figure 6.20 Structure of the elementary unit of the proposed linear actuator (Reprinted with permission from Yamakita, M. et al., Robotic Application of IPMC Actuators with Redoping Capability in Electroactive Polymers for Robotic Applications (eds Kim, K. J. and Tadokoro, S.), 199–225. Copyright (2007) Springer).

Active Microcatheter and Biomedical Soft Devices Based on IPMC Actuators uh

135

uh mh r

b

f

ml a

ml

g

pulley

actuator

θ2

θ1

l =a+b (x, y)

Figure 6.21 Model of a biped walking robot using the IPMC linear actuator (Reprinted with permission from Yamakita, M. et al., Robotic Application of IPMC Actuators with Redoping Capability in Electroactive Polymers for Robotic Applications (eds Kim, K. J. and Tadokoro, S.), 199–225. Copyright (2007) Springer).

The basic characteristics of the elementary unit were studied. The IPMC actuator used in the experiment was Nafion plated with gold. A counter cation in the Nafion was sodium. Though the response of the actuator varies depending on its condition, it was confirmed that the unit whose total length is 40 mm is constricted by 10 mm with a step input voltage of 2.5 V on average. As shown in the chapter on the fundamentals of this technology, the bending characteristics of the IPMC actuator depend on the counter cation in the ionic polymer. Yamakita et al. developed the application of the IPMC linear actuator to a biped walking robot (Figure 6.21) and optimized the performance of the actuator by selecting the counter cation in the ionic polymer.

6.5 Conclusions In this chapter, methods of fabricating IPMC actuator devices and various biomedical applications of IPMC actuators have been described. As examples shown in this chapter, IPMC actuators have much potential for biomedical and human-affinity applications. Though there still remain some issues that must be solved, IPMC actuators are expected to be used in various practical biomedical applications in the near future.

References 1. Bar-Cohen, Y., Leary, S., Shahinpoor, M., et al. (1999) Electro-active polymer (EAP) actuators for planetary applications. Proceedings of the SPIE conference on Smart Structures and Materials, Electroactive Polymer Actuators and Devices, Newport Beach, 3669, 57–63. 2. Guo, S., Fukuda, T. and Asaka, K. (2002) Fish-like underwater microrobot with 3 DOF. Proceedings of the IEEE International conference on Robotics and Automation, 738–43.

136


3. Guo, S., Fukuda, T. and Asaka, K. (2003) A new type of Fish-Like Underwater Microrobot., IEEE/ASME Trans on Mechatronics, 8, 136–41. 4. Zhang, W., Guo, S. and Asaka, K. (2006) A new type of hybrid fish-like microrobot, Int. J. Automation and Computing, 3, 358–65. 5. Tadokoro, S., Fuji, S., Fushimi, S., et al. (1998) Development of a distributed actuation device consisting of soft gel actuator elements. Proceedings of the IEEE International Conference on Robotics and Automation, 2155–60. 6. Konyo, M., Tadokoro, S., Takamori, T. and Oguro, K. (2000) Artificial tactile display using soft gel actuators. Proceedings of the IEEE International Conference on Robotics and Automation, 3416–21. 7. Tadokoro, S., Yamagami, S., Ozawa, M., et al. (1999) Multi-DOF device for soft micromanipulation consisting of soft gel actuator elements. Proceedings of the IEEE International Conference on Robotics and Automation, 2177–82. 8. Guo, S., Fukuda, T., Kosuge, K., et al. (1994) Micro catheter system with active guide wirestructure, experimental results and characteristic evaluation of active guide wire catheter using ICPF actuator, 5th International Symposium on Micro machine and Human Science Proceedings, Nagoya, 191–7. 9. Guo, S., Fukuda, T., Kosuge, K., et al. (1995) Micro active guide wire catheter using ICPF actuator, IEEE International Conference on Intelligent Robotics and Systems (IROS 95), Pittsburgh, 2, 172–7. 10. Oguro, K., Fujiwara, N., Asaka, K., et al. (1999) Polymer electrolyte actuator with gold electrodes. Proceedings of the SPIE 6th Annual International Symposium on Smart Structures and Materials, Newport Beach, 64–71. 11. Yamakita, M., Kamamichi, N., Kaneda, Y., et al. (2004) Development of an Artificial Muscle Linear Actuator Using Ionic Polymer-Metal Composites. Adv. Robotics, 18, 383–99. 12. Kamamichi, N., Yamakita, M., Kozuki, T., et al. (2007) Doping effects on robotic systems with ionic polymer-metal composite actuators, Adv. Robotics, 21, 65–85. 13. Yamakita, M., Kamamichi, N., Luo, Z.-W. and Asaka, K. (2007) Robotic application of IPMC actuators with redoping capability, in Electroactive Polymers for Robotics Applications (eds Kim, K. J. and Tadokoro, S.), Springer, London. 14. Nakabo, Y., Mukai, T. and Asaka, K. (2007) Biomimetic soft robots using IPMC, in Electroactive Polymers for Robotics Applications (eds Kim, K. J. and Tadokoro, S.), Springer, London. 15. Akle, B. J., Bennett, M. D. and Leo, D. J. (2006) High-strain ionomeric-ionic liquid electroactive actuators, Sensors and Actuators A, 126, 173–81. 16. Levitsky, I. A., Kanelos, P. and Euler, W. B. (2004) Electromechanical actuation of composite material from carbon nanotubes and ionomeric polymer, J. Chem. Phys, 121, 1058–165. 17. Shahinpoor, M., Bar-Cohen, Y., Simpson, J. O. and Smith, J. (1999) Ionic polymer-metal composites (IPMC) as biomimetic sensors, actuators & artificial muscles – a review, Field Response Polym., American Chemical Society, 25–50. 18. Yamakita, M., Sera, A., Kamamichi, N., et al. (2006) Integrated design of IPMC actuator/ sensor, Proceedings of the 2006 IEEE International Conference on Robotics and Automation, 1834–9. 19. Kato, Y., Sekitani, T., Takamiya, M., et al. (2007) Sheet-type Braille displays by integrating organic field-effect transistors and polymeric actuators, IEEE Trans. on Electron Devices, 54, 202–9. 20. Guo, S., Fukuda, T., Nakamura, T. and Oguro, K. (1997) Development of the Micro pump using ICPF actuator, Proceedings of the 1997 IEEE International Conference on Robotics and Automation, Albuquerque, NM, 266–1.

7 Implantable Heart-Assist and Compression Devices Employing an Active Network of Electrically-Controllable Ionic Polymer– Metal Nanocomposites Mohsen Shahinpoor Biomedical Engineering Laboratories, Department of Mechanical Engineering, University of Maine, Orono, USA.

7.1 Introduction Congestive heart failure (CHF) is the number one killer of people all over the world. In the United States alone, cardiovascular and weak heart-related diseases cost the health care systems over US$ 300 billion each year. Currently there is no viable system to assist a weak heart with its ventricular compression deficiencies. There have been many attempts to build a safe and operational artificial heart in the past which we refrain from listing. However, in recent years there has been a number of advanced and totally implantable left-ventricular assist systems (LVAS), such as the Thoratec’s HeartMateÒ (Figure 7.1a), the Baxter’s NovacorÒ (Figure 7.1b) or Arrow International, Inc.’s LionHeartÒ (Figures 7.1c and 7.1d). All are rather invasive, end-stage, temporary and only help with the left ventricular compression and simply are far from being able to fight this deadly disease of humanity, namely CHF. Recently, with the pioneering work of Shahinpoor (1–20), the idea of helping a weak heart by compressing it from without in a soft, intact and intelligent manner by implantable polymeric artificial muscles has gained acceptance in the United States and international medical communities.


138


(b)

(c)

(d)

Figure 7.1 Totally implantable LVASs: (a) Thoratec’s HeartMateÒ II; (b) Baxter’s NovacorÒ; (c) and (d) Arrow International’s Lion HeartÒ.

The cost of cardiovascular diseases and stroke in the United States is estimated at well over US$300 billion/year (http://www.americanheart.org). In most severe cases the patient has only a few days to live and a donor heart will be required for any chance of survival, even though the patient’s body is likely to reject the donated transplanted hearts at a rate of 50%. It is in this spirit that it is proposed to develop a family of minimally invasive (without opening the chest of the patient) and thorascopically implantable, intelligent multi-fingered heart compression/assist systems equipped with soft and resilient electroactive polymeric artificial muscles. Shahinpoor has developed soft ionic polymeric artificial muscles that thrive in the wet and saline environment of the inside of the human body and, in particular, around the myocardium of the heart. Such heart compression/assist systems will be capable of selectively assisting the ventricles, and in particular the left ventricle, of a weak heart to produce more internal pressure and to pump more blood from one or more sides in synchrony with the natural systolic contraction of the ventricle, as well as providing arrhythmia control of the beating heart. Note that the heart weighs between 7 and 15 ounces (200–425 grams) and is a little larger than the size of a fist. By the end of a long life, a person’s heart may have beat (expanded and contracted) more than 3.5 billion times. In fact, each day, the average heart pumps about 2000 gallons (7571 liters) of blood and beats 100 000 times. The human heart is located between the lungs in the middle of the chest, behind and slightly to the left of the breastbone (sternum). A double-layered membrane called the pericardium

Implantable Heart-Assist and Compression Devices

139

surrounds the heart like a sac. The outer layer of the pericardium surrounds the roots of the heart’s major blood vessels and is attached by ligaments to the spinal column, diaphragm and other parts of the human body. The inner layer of the pericardium is attached to the heart muscle. A coating of fluid separates the two layers of membrane, letting the heart move as it beats, yet still be attached to the body. The heart has four chambers. The upper chambers are called the left and right atria, and the lower chambers are called the left and right ventricles. A wall of muscle called the septum separates the left and right atria and the left and right ventricles. The left ventricle is the largest and strongest chamber in the heart. The left ventricle’s chamber walls are only about a half-inch thick, but they have enough force to push blood through the aortic valve and into the body. Four types of valve regulate blood flow through the heart: The tricuspid valve regulates blood flow between the right atrium and right ventricle. The pulmonary valve controls blood flow from the right ventricle into the pulmonary arteries, which carry blood to the lungs to pick up oxygen. The mitral valve lets oxygen-rich blood from the lungs pass from the left atrium into the left ventricle. The aortic valve opens the way for oxygen-rich blood to pass from the left ventricle into the aorta, the body’s largest artery, where it is delivered to the rest of the body (Figure 7.2a). Electrical impulses from the heart muscle (the myocardium) cause the heart to contract. This electrical signal begins in the sinoatrial (SA) node, located at the top of the right atrium (Figure 7.2b). The SA node is sometimes called the heart’s ‘natural pacemaker’. An electrical impulse from this natural pacemaker travels through the muscle fibers of the atria and ventricles, causing them to contract. Although the SA node sends electrical impulses at a certain rate, heart rate may still change depending on physical demands, stress or hormonal factors.

(a)

(b)

Figure 7.2 Basic operation of the heart (a) and electrical impulses and activities (b).

7.2 Heart Failure Heart failure is a clinical syndrome in which the heart fails to maintain an adequate output, resulting in diminished blood flow and congestion in the circulation in the lung or other

140


parts of the body or both. Manifestations of engorged blood vessels in the lungs are referred to as left heart failure, and engorgement of veins and capillaries in other parts of the body are called right heart failure. Common causes of heart failure are high blood pressure, coronary artery disease and rheumatic heart disease. Clinical features may vary considerably. The main symptom of left heart failure is shortness of breath, which often occurs during mild exercise and which may result in periodic sudden attacks that frequently occur at night while the victim is lying flat. As heart failure progresses, shortness of breath becomes more difficult to relieve.

7.3 Background of IPMNCs Recent findings on ionic polymer conductor nanocomposites (IPCNCs) and ionic polymer–metal nanocomposites (IPMNCs) as biomimetic distributed nanosensors, nanoactuators and artificial muscles and electrically controllable polymeric network structures have been presented recently [1–20]. Basically, distributed nanosensing and nanoactuation means that these materials are active down to nano size level. In other words, if they are cut as small as nanoactuators and sensors in a few nanometer range, they will still sense and actuate under a voltage of a few microvolts. Furthermore, methods of fabricating several electrically and chemically active ionic polymeric gel muscles – such as polyacrylonitrile (PAN), poly(2-acrylamido2-methyl-1-propane sulfonic) acid (PAMPS) and polyacrylic-acid-bis-acrylamide (PAAM) – as well as a new class of electrically active composite muscle – such as Ionic Polymeric Conductor Composites (IPCCs) or Ionic Polymer–Metal Composites (IPMCs) made with perfluorinated sulfonic or carboxylic ionic membranes – have been introduced and investigated [7]. Several apparatuses for modeling and testing of the various IPMNC artificial muscles have been described to show the viability of the application of both chemoactive and electroactive muscles. Furthermore, fabrication methods of PAN fiber muscles in different configurations, such as spring-loaded fiber bundles, biceps, triceps, ribbon type muscles and segmented fiber bundles, to make a variety of biomimetic sensors and actuators have also been reported [1–20]. Theories and numerical simulations associated with ionic polymer gel electrodynamics and chemodynamics have also been discussed, analyzed and modeled for the manufactured material. In this chapter the focus is on perfluorinated sulfonic ionic multi-functional materials, as potentially powerful ionic polymers for biomimetic distributed nanosensing, nanoactuation, nanorobotics, nanotransducers for power conversion and harvesting, as well as artificial muscles for medical and industrial applications. It must be noted that widespread electrochemical processes and devices use poly (perfluorosulfonic acid) ionic polymers. These materials exhibit [1–20] good chemical stability, remarkable mechanical strength, good thermal stability and high electrical conductivity when sufficiently hydrated and made into a nanocomposite with a conductive phase such as metals, conductive polymers or graphite. As described elsewhere [7], a number of physical models have been developed to understand the mechanisms of water and ion transport in ionic polymers and membranes. Morphological features influence transport of ions in ionic polymers. These features


141

have been studied by a host of experimental techniques including: small and wide angle X-ray scattering, dielectric relaxation and a number of microscopic and spectroscopic studies. The IPMNCs are basically a two-phase system made up of a polar fluid (water) containing ion cluster network surrounded by a hydrophobic polytetrafluoroethylene (PTFE) medium. The integrity and structural stability of the membrane is provided by the PTFE backbones and the hydrophilic clusters facilitate the transport of ions and water in the ionic polymer. These nanoclusters have been conceptually described as containing an interfacial region of hydrated, sulfonate-terminated perfluoroether side chains surrounding a central region of polar fluids. Counter ions such as sodium (Naþ) or lithium (Liþ) are to be found in the vicinity of the sulfonates. It must be noted that the length of the side chains has a direct bearing on the separation between ionic domains, where the majority of the polar fluids resides, and the nonpolar domains. High-resolution NMR of some perfluoroionomer shows an unusual combination of a nonpolar, Teflon-like backbone, with polar and ionic side branches. Liu and SchmidtRohr [21] have obtained the first high-resolution NMR spectra of solid perfluorinated polymers by combining 28 kHz magic-angle spinning (MAS) with rotation-synchronized 19F pulses. Their NMR studies enable more detailed structural investigations of the nanometer-scale structure and dynamics of PTFE based ionomers. It has also been well established [1–21] that anions are tethered to the polymer backbone and cations (Hþ, Naþ, Liþ) are mobile and solvated by polar or ionic liquids within the nanoclusters of size 3–5 nanometers.

7.4 Three-Dimensional Fabrication of IPMNCs In a previous work, Kim and Shahinpoor [11] have reported a newly developed fabrication method that can scale up or down the IPMNC artificial muscles in a strip size of micro-to-centi-meter thickness, using the liquid form of perfluorinated ionic polymers. By meticulously evaporating the solvent (isopropyl alcohol) out of the solution, recast ionic polymer can be obtained. A number of these samples are shown in Figure 7.3. (a)

(b)

Figure 7.3 (a) An eight-finger synthetic muscle. It has a thickness of approximately 2 mm. (b) A coil-type synthetic muscle. This coil type muscle creates a linear actuation motion.

142


7.5 Electrically-Induced Robotic Actuation In perfluorinated sulfonic acid polymers there are relatively few fixed ionic groups. They are located at the end of side chains, so as to position themselves in their preferred orientation to some extent. Therefore, they can create hydrophilic nanochannels, so called cluster networks [Gierke and Hsu [22] and Gierke, Munn and Wilson [23]]. Such configurations are drastically different in other polymers, such as styrene/divinylbenzene families that limit, primarily by cross-linking, the ability of the ionic polymers to expand due to their hydrophilic nature. Basically, the cations attract water molecules and thus they separate from the polymer backbone charged pendant groups and gather around them a number of water molecules, thus expanding the network or swelling. Once an electric field is imposed on such a network, the conjugated and hydrated cations rearrange to accommodate the local electric field and thus the network deforms, where in the simplest of cases, such as in thin membrane sheets, spectacular bending is observed (Figures 7.4 and 7.5) under small electric fields such as tens of volts per millimeter.

Figure 7.4 A four-fingered IPMNC compression system in open configuration.

Figure 7.5 Typical deformation of strips (10 80 0.34 mm) of ionic polymers under a step voltage of 4 V.


143

2.0 ERI-S1 Muscle (5Pt-1Pt/PVP) Lo = 1 inch input = sine

1.8 1.6

0.1 Hz

1.4

δ/Lo

1.2

0.2 Hz

1.0 0.5 Hz

0.8 0.6 0.4 0.2 0.0 0

1000

2000

3000

4000 E (V/m)

5000

6000

7000

8000

2.0 ERI-S1 Muscle (5Pt-1Pt/PVP) Lo = 1.5 inch input = sine

1.8 1.6

δ/Lo

1.4 1.2

0.1 Hz

1.0

0.2 Hz

0.8 0.5 Hz

0.6 0.4 0.2 0.0 0

1000

2000

3000

4000 E (V/m)

5000

6000

7000

8000

Figure 7.6 Displacement characteristics of an IPMNC, ERI-S1 (d: arc length, Lo: effective cantilever beam length). Lo ¼ 1.0 inch (top) and Lo ¼ 1.5 inch (bottom).

Typical experimental deflection curves are depicted in Figures 7.6. Typical frequency dependent dynamic deformation characteristics of IPMNCs are depicted in Figure 7.7. Once an electric field is imposed on an IPMNC cantilever, in the cantilever polymeric network the hydrated cations migrate to accommodate the local electric field. This creates a pressure gradient across the thickness of the beam and thus the beam undergoes bending deformation (Figures 7.6) under small electric fields such as tens of volts per millimeter.

144


Tip Deflection in mm Tip Force in gramforce

40

30

Tip Deflection points Tip Force points

20

10

0

0

2

4

6

8

10

Step voltage (v)

Figure 7.7 Variation of tip blocking force and the associated deflection if allowed to move versus the applied step voltage for a 10 50 0.3 mm IPMNC Pt-Pd sample in a cantilever configuration.

Figure 7.7 depicts typical force and deflection characteristics of cantilever samples of IPMNC artificial muscles.

7.6 Distributed Nanosensing and Transduction Shahinpoor [19] has presented a review on sensing and transduction properties of ionic polymer conductor composites. Shahinpoor in 1995 [24] and 1996 [25], and recently [26], reported that IPMNCs by themselves and not in a hydrogen pressure electrochemical cells as reported by Sadeghipour, Salomom and Naogi [27] can generate electrical power like an electromechanical battery if flexed, bent or squeezed. Shahinpoor [24–25] reported the discovery of a new effect in ionic polymeric gels, namely the ionic flexogelectric effect, in which flexing, compression or loading of IPMNC strips in air created an output voltage like a dynamic sensor or a transducer converting mechanical energy to electrical energy. Keshavarzi, Shahinpoor, Kim and Lantz [28] applied the transduction capability of IPMNC to the measurement of blood pressure, pulse rate and rhythm measurement using thin sheets of IPMNCs. Motivated by the idea of measuring pressure in the human spine, Ferrara et al. [29] applied pressure across the thickness of an IPMNC strip while measuring the output voltage. Typically, flexing of such material in a cantilever form sets them into a damped vibration mode that can generate a similar damped signal in the form of electrical power (voltage or current) as shown in Figure 7.8. IPMNC sheets can also generate power under normal pressure. Thin sheets of IPMNC were stacked and subjected to normal pressure and normal impacts


IPMC Composite Sensor

145

Electrodes

Flip

0.008 0.006

E (volts)

0.004 0.002 0.000 –0.002 –0.004 –0.006 0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

Time (s)

Figure 7.8 A typical voltage response of an IPMNC strip (10 40 0.2 mm) under oscillatory mechanical excitations.

and were observed to generate large output voltage. Endo-ionic motion within IPMNC thin sheet batteries produced an induced voltage across the thickness of these sheets when a normal or shear load was applied. A material testing system (MTS) was used to apply consistent pure compressive loads of 200 N and 350 N across the surface of an IPMNC sheet (2 2 cm). The output pressure response for the 200 N load (73 psi) was 80 mV in amplitude and for the 350 N (127 psi) it was 108 mV. This type of power generation may be useful in the heels of boots and shoes or places where there are a lot of foot or car traffic. The output voltage of the thin sheet IPMNC batteries [26] under 200 N normal load is depicted in Figure 7.9. The output voltage is generally about 2 mV/cm length of the IPMNC sheet. IPMNCs also enjoy fairly consistent operation under oscillatory activation, as depicted in Figure 7.10. This characteristics is necessary for cyclic heart compression to the tune of billions of times.

146

Biomedical Applications of Electroactive Polymer Actuators IPMC response under compression (200N load) Sensor Output Load

200

250 200

150

150

100

100

50

50

0

0

–50

0

10

20

30

40

50

60

70

Load (N)

Sensor Output (MV)

250

–50 80

Time (s)

Figure 7.9 Output voltage due to normal 90 ° impact of a 200 N load on a 20 20 0.2 mm IPMNC sample.

8 1 V, 1/2 Hz, Step

7 6 PVP Treated IPMC-Nafion (Li+)

Fg (g)

5 4 3 2

Conventional IPMC-Nafioin (Li+)

1 0 0

1×109

2×109

3×109 4×109 5×109 No. of Cycles

6×109

7×109

8×109

Figure 7.10 Long cycles oscillation of IPMNCs versus blocking force (sample size 5 20 0.2 mm). The environmental chamber was maintained at T ¼ 25 °C and RH ¼ 50–55 %.

7.7 Modeling and Simulation As recently as 2000, Nobel Laureate Pierre de Gennes, Okumura, Shahinpoor and Kim [30] presented the first phenomenological theory for sensing and actuation in ionic polymer–metal composites. Asaka and Oguro [31] discussed the bending of


147

polyelectrolyte membrane–platinum composites by electric stimuli and presented a theory on actuation mechanisms in IPMNC by considering the electro-osmotic drag term in transport equations. Nemat-Nasser and Li [32] discussed a modeling on the electromechanical response of ionic polymer–metal composites based on electrostatic attraction/repulsion forces in IPMNCs. Later, Nemat-Nasser [33] proposed a revised version of their earlier paper and stressed the role of hydrated cation transport and mobility within the clusters and polymeric networks in IPMNCs. Nemat-Nasser and Wu [34] have proposed a discussion on the role of the backbone ionic polymer and, in particular, sulfonic versus carboxylic ionic polymers, as well as the effect of different cations such as potassium (Kþ), sodium (Naþ), lithium (Liþ), caesium (Csþ) and some organometalic cations on the actuation and sensing performance of IPMNCs. Tadokoro [35], Tadokoro, Yamagami, Takamori and Oguro [36] and Tadokoro, Takamori and Oguro [37] have presented an actuator model of IPMNC for robotic applications on the basis of physico-chemical phenomena. A recent comprehensive review by Shahinpoor and Kim [12] on modeling and simulation of ionic polymeric artificial muscles discusses the various modeling approaches for understanding the mechanisms of sensing and actuation of ionic polymers and the notion of ion mobility. The underlying principle of IPMNCs actuation and sensing capabilities, which can be described by the standard Onsager formulation using linear irreversible thermodynamics, can now be summarized. When static conditions are imposed, a simple description of mechanoelectric effect is possible based upon two forms of transport: ion transport (with a current density, J , normal to the material) and solvent transport (with a flux, Q ; it can be assumed that this term is water flux). The conjugate forces include the electric field, E , and p the pressure gradient, r . The resulting equation has the concise form of: J ðx; y; z; tÞ ¼ E ðx; y; z; tÞ L12 r pðx; y; z; tÞ

ð7:1Þ

Qðx; y; ztÞ ¼ L21 E ðx; y; z; tÞ K r pðx; y; z; tÞ

ð7:2Þ

where and K are the material electric conductance and the Darcy permeability, respectively. A cross coefficient is usually L ¼ L12 ¼ L21. The simplicity of the above equations provides a compact view of the underlying principles of both actuation, transduction and sensing of the IPMNCs. When measuring the direct effect (actuation mode) we work (ideally) with electrodes which are impermeable to ion species flux, and thus Q ¼ 0. This gives:

L Eðx; y; z; tÞ pðx; y; z; tÞ ¼ r K

ð7:3Þ

r , proportional to r pðx; y; z; tÞ will, in turn, induce a curvature, pðx; y; z; tÞ. The relationships between the curvature, , and pressure gradient, r pðx; y; z; tÞ, are fully derived and described in de Gennes, Okumura, Shahinpoor and Kim [30]. It should be mentioned that (1/rc) ¼ M(E)/YI, where M(E) is the local induced bending moment and is a function of the imposed electric field E, Y is the Young’s modulus (elastic stiffness) of the strip, which is a function of the hydration H of the IPMNC, and I is the

148


moment of inertia of the strip. Note that locally M(E) is related to the pressure gradient such that in a simplified scalar format: rpðx; y; z; tÞ ¼ ð2P=t Þ ¼ ðM=IÞ ¼ Y=rc ¼ Y :

ð7:4Þ

From Equation (7.4) it is clear that the vectorial form of curvature, E , is related to the imposed electric field E by: ¼ ðL=KYÞ E E

ð7:5Þ

Based on this simplified model the tip bending deflection, max, of an IPMNC strip of length lg should be almost linearly related to the imposed electric field due to the fact that: ffi ½2 max =ðl2g þ 2max Þ ffi 2 max =l2g ffi ðL=KYÞ E E

ð7:6Þ

The experimental deformation characteristics depicted in Figures 7.5 and 7.6 are clearly consistent with the above predictions obtained by the above linear irreversible thermodynamics formulation, which is also consistent with Equations (7.5) and (7.6) in the steady state conditions and has been used to estimate the value of the Onsager coefficient, L, to be of the order of 10–8 m2/V-s. Here, a low frequency electric field has been used in order to minimize the effect of loose water back diffusion under a step voltage or a DC electric field. Other parameters have been experimentally measured to be K 10–18 m2/CP, 1 A/mV or S/m. A more detailed set of data pertaining to Onsager coefficient, L, is depicted in Figure 7.11. It must be noted

4 Sample #1 Sample #2 Sample #3

L × 108 [(m/s)/(V/m)]

3

2

1

0 0

5000

10000

15000

20000

25000

E (V/m)

Figure 7.11

Experimental determination of Onsager coefficient, L, using three different samples.


149

Figure 7.12 Some movement capabilities of IPMNCs suitable for heart compression.

that IPMNCs are quite capable of undergoing complex movements (Figure 7.12) to accommodate the complex pumping action (systolic–diastolic) of the heart.

7.8 Application of IPMNCs to Heart Compression and Assist in General This chapter discusses the broad category of heart compression and assist and arrhythmia control devices, and more particularly the potential applications of ionic polymer–metal nanocomposite (IPMNCs) biomimetic sensors, actuators and artificial muscles integrated mechatronically as a heart compression device which can be implanted external to the patient’s heart, and partly sutured to the heart, without interfering or contacting with the internal blood circulation. Thus, the proposed IPMNC device can thereby avoid thrombosis and similar complications, which are common to current artificial heart, or heart-assist devices, which may arise when the blood flow makes repeated contacts with nonbiological or nonself surfaces. In compressing a heart ventricle the device must be soft and electronically robust in order not to damage the ventricle. This means that the device should contain control means, such as bradycardic (pacing) and tachyarrhythmic (cardioverting/ defibrillating), to facilitate device operation in synchronism with the left ventricular contraction and capable of transcutaneous recharging of the implanted batteries. The general idea is presented in Figure 7.13. Note that the device is implanted essentially in

150


42

5 3

30 46

12

10

44

Figure 7.13

General configuration for the proposed heart compression device.

the ribcage of the patient but is supported on a flexible stem that extends to the abdomen. It is also possible to place the supporting structure of the heart compression device on the diaphragm muscle. These details will be worked out during the clinical testing and operation of such devices. The slender flexible stem allows the systolic and diastolic cycles of the heart to continue and yet allows the body of the heat to make swinging motions to one side or the other without any obstacle. In Figure 7.13, 42 is the patient body, 44 is the abdomen area, 46 is the rib cage, 5 is the heart, 3 is the polymeric compression finger made with IPMNCs, 30 is the base of the compression device, 10 is a slender column carrying the electronic wires to the muscle and acting as a flexible support column as well and 12 is the power/battery housing placed in the abdomen. A more detailed drawing of the compression device itself is depicted in Figure 7.14. In Figure 7.14, again 3 denotes the compression fingers made with IPMNCs, 5 is the heart itself, 4 depicts an encapsulated enclosure filled with water to create a soft cushion for the compression fingers, 4d’s are IPMNC based sensors cilia to continuously monitor the compression forces applied to the heart and 3e and 3f are the associated wiring and electronics. Note that, assisting or soft compression of the left ventricle of a weak heart will produce more internal pressure to pump more blood in synchrony with the natural systolic contraction of the ventricle. Additionally, the proposed system will also provide


151

4a

5 4b

4

4d 3 3

4b 3c

3c 4a

4

4c 4c

3e 3d

3f

3f

Figure 7.14

3e

Heart compression device equipped with IPMNC fingers.

arrhythmia control of the beating heart and will be powered by IPMNC artificial muscles for cyclically actuating the resilient compression soft fingers, thereby cyclically pumping blood from the ventricles and allowing the ventricles to refill. These devices will be completely implantable in the body of a patient external of the heart, thereby avoiding thrombogenesis and other complications that may arise from contact between the blood flow and artificial, nonbiological surfaces. The soft fingers comprise suitably mounted platinum and gold electrodes for heart monitoring purposes by means of IPMNC capability to determine the ventricular stroke volume and/or pressure. Specifically, the proposed IPMNC based device will provide entirely electricallycontrollable and micro-processor-controlled multi-fingered resilient sphinctering heart compression devices that can be implanted inside the rib cage of a patient with weak heart and will gently squeeze the weak heart to enhance blood circulation and assist the weak heart. Other configurations are depicted in Figures 7.15 and 7.16.

(a)

(b)

Figure 7.15 Four-fingered heart compression device equipped with thick IPMNCs: (a) before compression; (b) after compression.

152


Figure 7.16 The upright configuration of the heart compression device.

The compression devices shown in Figure 7.15 and 7.16 were designed and fabricated from thick (2 mm) ionic polymer–metal composites and were subsequently 24-karat gold-plated. In compressing a heart ventricle the device must be soft and electronically robust in order not to damage the ventricle. This means that the device should contain control means such as bradycardic (pacing) and tachyarrhythmic (cardioverting/defibrillating) to facilitate device operation in synchronism with the left ventricular contraction, and should be capable of transcutaneous recharging of the implanted batteries. Note that the device is implanted essentially in the rib cage of the patient but is supported on a slender flexible stem that extends to the abdomen. The stem allows the systolic and diastolic cycles of the heart to continue and yet allows the body of the heart to make swinging motions to one side or the other without unnecessary restriction. It is also possible to place the supporting structure of the heart compression device on the diaphragm muscle. These details can be worked out during the clinical testing and operation of such devices. Again 3 denotes the compression fingers made with IPMNCs, 5 is the heart itself, 4 depicts an encapsulated enclosure filled with water to create a soft cushion for the compression fingers, 4d’s are IPMNC based sensor cilia to continuously monitor the compression forces applied to the heart and 3e and 3f are the associated wiring and electronics. As designed, this device produces assisting or soft compression of the left ventricle of a weak heart to produce more internal pressure and to pump more blood from one or more sides in synchrony with the natural systolic contraction of the ventricle. Additionally, the system can also provide arrhythmia control of the beating heart. The soft fingers incorporate suitably located electrodes for monitoring the ventricular stroke volume and pressure. A simpler design configuration uses a compression band to assist the heart in its systolic and diastolic cycles of compression– decompression as shown in Figure 7.17. Also, the compression band can be designed such that it can encircle the heart as shown in Figure 7.18.


Figure 7.17

153

A heart with an IPMNC compression band. (See Color Plate 1).

Figure 7.18 An IPMNC compression band in open and closed configurations.

Presented here are some preliminary data concerning a mini heart compression device equipped with IPMNCs. First measured is the force generated by each strip at 5 V, then measured is the pressure generated when squeezing a small balloon or an animal’s heart (Figure 7.19). Figures 7.20 and 7.21 depict the variation of pressure generated (in mm Hg) with the voltage applied. As discussed before, thick IPMNC polymeric muscles had to be manufactured to generate the required force for heart compression. These were measured experimentally on an animal heart such as a rat’s heart. These newly developed threedimensional IPMNCs have been fully discussed in [37–41] and are briefly described below.

154


Figure 7.19 Mini heart compression device equipped with IPMNC muscles.

10

9.6

9 Weight (g)

8

7.6

7.4

Surface (dm2)

7

Compression force developed (g)

6 5 4

4

3.6

3.51

3 2

2.7

2.34 1.53

1.17

1 0.28

0.18

1.02 0.42

1.2 0.36

0.55

0.24

0

Thick platinum

Thick gold

Figure 7.20

Very thick platinum

2× thick gold

3× thick gold

Pressure generation versus electrode thickness.

Thin gold


155

1.4

mmHg developed in the Langedorf

1.2

1 Pressure (mmHg) baseline 74 Pressure (mmHg) baseline 42 0.8

Pressure (mmHg) baseline 22 Pressure (mmHg) baseline 14 T 50 fatigue (min)

0.6

0.4

0.2

0 Thick platinum

Thick gold

Figure 7.21

very thick platinum

2× thick gold

3× thick gold

thin gold

Pressure generation versus electrode thickness.

7.9 Manufacturing Thick IPMNC Fingers The preparation of thick IPMNC fingers follows the procedure outlined in reference [38]. However, the reader is referred to references [14–18] for additional information on IPMNCs. Thus, thick IPMNC strips were manufactured based on the procedure reported in [37] and subsequently equipped with platinum electrodes and gold plating on both sides of the strip with a particle penetration depth of 20 mm. Photographs of the IPMNC samples with the platinum electrodes covered with the gold electrodes are shown in Figure 7.22. It must be mentioned that other types of organ compression, and in particular aortic peristaltic compression to enhance blood circulation and assist a weak heart, are also possible with IPMNC polymeric muscles, as depicted below in Figure 7.23. Furthermore, endoscopic surgical operations with IPMNCs bundled up and insertable through endoscopic incisions is also possible, as depicted in Figure 7.24.

156


(b)

(c)

Figure 7.22 Manufacturing sequence of the heart compression device: (a) four IPMNC fingers cut to scale; (b) the fingers assembled between two gold ring electrodes; (c) the fingers placed between the ring electrodes and closed.

24

3

Figure 7.23 IPMNCs configured to perform peristaltic compression on aortic blood flow and enhance blood circulation and assist a weak heart.


157

4

3c

4

Figure 7.24 Endoscopically surgically insertable configuration for IPMNC polymeric muscles for heart and organ compression.

7.10 Conclusions Design and development of a novel mammalian heart compression/assist device in the form of a multi-fingered robotic hand equipped with an assembly of thick ionic polymer—metal composite (IPMNC) fingers encased inside a water bag were described. The designed and laboratory-tested multi-fingered heart compression device (MFHCD) was shown to be entirely endoscopically implantable. It was further shown that these devices can be directly or transcutaneously energized by inductive magnetoresonant generator. Therefore, based on the background technologies and their successes and failures in treating weak hearts, it is highly desirable to develop a soft heart compression device for patients with CHF problems. In this connection, ionic polymer–metal composites as soft biomimetic sensors, actuators and artificial muscles present a tremendous opportunity.

References 1. Shahinpoor, M. and Osada, Y. (1995) Heart tissue Replacement with Ionic Polymeric Gels, Proceedings of the 1995 ASME Winter Annual Meeting, San Francisco, CA, 314–8. 2. Shahinpoor, M. (2002) Electrically-Controllable Multi-Fingered Resilient Heart Compression Devices, US Patent Office, US Patent 6,464,655, Issued 15 October 2002. 3. Tozzi, P., Shahinpoor, M., Hayoz, D. and L. von Segesser (2004) Electroactive Polymers to Assist Failing Heart: The Future Is Now, Proceedings of the Second World Congress On Biomimetics and Artificial Muscle (Biomimetics and Nano-Bio 2004), Albuquerque, NM, 5–8 December. 4. Shahinpoor, M., Electrically-Controllable Multi-Fingered Resilient Heart Compression Devices (CIP), (2007) US Patent Office, Patent No.7,198,594, CIP to US Patent 6,464,655, Issued 3 April 2007. 5. Shahinpoor, M. (2004) Artificial Muscles, in Encyclopedia of Biomaterials and Biomedical Engineering (eds Wnek, G. and Bowlin, G.), Marcel Dekker Publishers, NY.

158


6. Shahinpoor, M., Kim, K. J. and Mojarrad, M. (2004) Ionic Polymeric Artificial Muscles, ERI/ AMRI Press, Albuquerque, NM. 7. Shahinpoor, M., Kim, K. J. and Mojarrad, M. (2007) Artificial Muscles: Applications of Advanced Polymeric Nano Composites (1st edn), Taylor & Francis Publishers, London. 8. Kim, K. J. and Shahinpoor, M. (2007) Special Issue: Biomimetics, Artificial Muscles, and NanoBio 2004, J. Intelligent Material Systems and Structures, 7 (18), 101. 9. Shahinpoor, M. and Hans-Jo¨rg Schneider (eds), (2008) Intelligent Materials (1st edn), Royal Society of Chemistry (RSC) Publishers, Cambridge. 10. Shahinpoor, M., and Kim, K. J. (2001) Ionic Polymer-Metal Composites – I. Fundamentals, Smart Materials and Structures Int. J., 10, 819–33. 11. Kim, K. J. and Shahinpoor, M. (2003) Ionic Polymer-Metal Composites – II. Manufacturing Techniques, Smart Materials and Structures (SMS), Smart Materials and Structures Int. J., 12 (1), 65–79. 12. Shahinpoor, M. and Kim, K. J. (2004) Ionic Polymer-Metal Composites – III. Modeling and Simulation As Biomimetic Sensors, Actuators, Transducers and Artificial Muscles, Smart Materials and Structures Int. J., 13 (4), 1362–88. 13. Shahinpoor, M. and Kim, K. J. (2005) Ionic Polymer-Metal Composites – IV. Industrial and Medical Applications, Smart Materials and Structures Int. J., 14 (1), 197–214. 14. Choi, K., Kim, K. J., Kim, D., et al. (2006) Performance Characteristics of Electro-Chemically Driven Polyacrylonitrile Fiber Bundle Actuators, J. Intelligent Material Systems and Structures, 17 (7), 563–76, July. 15. Shahinpoor, M. (2005) Soft Plastic Robots and Artificial Muscles, Int. J. Adv. Robotic Systems, 2 (2), 161–74. 16. Shahinpoor, M. (2005) Smart Ionic Polymer Conductor Composite Materials as Multifunctional Distributed Nanosensors, Nanoactuators and Artificial Muscles, Am. Soc. Mech. Eng., Mat. Div. (Publication), MD, 485–9. 17. Shahinpoor, M. (2004) Artificial Muscles, in Encyclopedia of Biomaterials and Biomedical Engineering (eds Wnek, G. and Bowlin, G.), Marcel Dekker Publishers, 43–52, NY. 18. Shahinpoor, M. (2004) Electroactive Ion Containing Polymers, in Hand Book of Smart Systems, Institute of Physics (IOP) Publication, London. 19. Shahinpoor, M. and Guran, A. (2003) Ionic Polymer-Conductor Composites (IPCC) as Biomimetric Sensors, Actuators and Artificial Muscles, Selected Topics, in Structures and Mechatronics Systems (ed Belyaev, A. and Guran, A.), World Scientific Publishers, London, 417–436. 20. Shahinpoor, M. (2003) Ionic Polymer-Conductor Composites As Biomimetic Sensors, Robotic Actuators and Artificial Muscles-A Review, Electrochimica Acta, 48 (14–16), 2343–53. 21. Liu, S. F. and Schmidt-Rohr, K. (2001) High-Resolution Solid-State 13C NMR of Fluoropolymers, Macromolecules, 34, 8416–8. 22. Gierke, T. D., and Hsu, W. Y. (1982) The cluster-network model of ion clustering in perfluorosulfonated membranes, in Perfluorinated Ionomer Membranes (eds Eisenberg, A. and Yeager, H. L.), ACS, Washington, DC, 283–307. 23. Gierke, T. D., Munn, G. E. and Wilson, F. C. (1982) Morphology of perfluorosulfonated membrane products—wide-angle and small-angle X-ray studies, ACS Symp. Ser., 180, 195–216. 24. Shahinpoor, M. (1995) A New Effect in Ionic Polymeric Gels: The Ionic ‘Flexogelectric’ Effect, Proceedings of the SPIE 1995 North American Conference on Smart Structures and Materials, 28 February–2 March, San Diego, CA, 2441, 42–53. 25. Shahinpoor, M. (1996) The Ionic Flexogelectric Effect, Proceedings of the 1996 Third International Conference on Intelligent Materials, ICIM’96, and Third European Conference on Smart Structures and Materials, Lyon, France, June 1996, 1006–11.


159

26. Shahinpoor, M. (2004) Smart Thin Sheet Batteries Made With Ionic Polymer-Metal Composites (IPMCs), Proceedings of ASME-IMECE 2004 International Mechanical Engineering Congress and RD&D Exposition, 13–19 November, Anaheim, CA. 27. Sadeghipour, K., Salomon, R. and Neogi, S. (1992) Development of A Novel Electrochemically Active Membrane and Smart Material based Vibration Sensor/damper, Smart Materials and Structures J., 1, 172–9. 28. Keshavarzi, A., Shahinpoor, M., Kim, K. J. and Lantz, J. (1999) Blood Pressure, Pulse Rate, and Rhythm Measurement Using Ionic Polymer-Metal Composite Sensors, Elactroactive Polymers, SPIE (publication number 3669-36), 369–76. 29. Ferrara, L., Shahinpoor, M., Kim, K. J., et al. (1999) Use of Ionic Polymer-Metal Composites (IPMCs) As A Pressure Transducer In the Human Spine, in Electroactive Polymers, SPIE (publication, number 3669-45), 394–401. 30. de Gennes, P. G., Okumura, K., Shahinpoor, M., M. and Kim, K. J. (2003) Mechanoelectric Effects in Ionic Gels, Europhysics Letters, 50 (4), 513–8. 31. Asaka, K. and Oguro, K. (2000) Bending of Polyelectrolyte Membrane Platinum Composites by Electric Stimuli, Part II. Response Kinetics, J. Electroanal. Chem., 480, 186–98. 32. Nemat-Nasser, S. and Li, J. Y. (2000) Electromechanical Response of Ionic Polymer-Metal Composites, J. Applied Phys, 87 (7), 3321–31. 33. Nemat-Nasser, S. (2002) Micro-Mechanics of Actuation of Ionic Polymer-Metal Composites (IPMCs), J. Applied Phys, 92 (5), 2899–915. 34. Nemat-Nasser, S. and Wu, Y. (2003) Comparative Experimental Study of Ionic Polymer-Metal Composites with Different Backbone Ionomers And In Various Cation Forms, J. Applied Phys, 93, 5255–67. 35. Tadokoro, S. (2000) An Actuator Model of ICPF for Robotic Applications On the Basis of Physico-Chemical Hypotheses, Proceedings of the IEEE International Conference on Robotics and Automation, 1340–6. 36. Tadokoro, S., Yamagami, S., Takamori, T., and Oguro, K. (2000), Modeling of Nafion-Pt Composite Actuators (ICPF) by Ionic Motion, in Proceedings of the SPIE 7th Smart Structures and Materials Symposium, EAPAD Conference, San Diego, CA, March 2000, 3987, 92–102. 37. Tadokoro, S., Takamori, T. and Oguro, K. (2001) Application of the Nafion-Platinum Composite Actuator, in Proceedings of the SPIE 8th Smart Structures and Materials Symposium, EAPAD Conference, San Diego, CA, March 2001, 4329, 28–42. 38. Kim, K. W. and Shahinpoor, M. (2001) Development of 3-D Polymeric Artificial Muscles, in Proceedings of the SPIE 8th Smart Structures and Materials Symposium, EAPAD Conference, San Diego, CA, March 2001, paper number 4329–30. 39. Adolf, D., Shahinpoor, M., Segalman, D. and Witkowski, W. (1993) Electrically Controlled Polymeric Gel Actuators, US Patent Office, US Patent 5,250,167, Issued 5 October 1993. 40. Shahinpoor, M. and Mojarrad, M. (2000) Soft Actuators and Artificial Muscles, US Patent Office, US Patent 6,109,852, Issued 29 August 2000. 41. Shahinpoor, M. (1995) Spring-Loaded Ionic Polymeric Gel Linear Actuator, US Patent Office, US Patent 5,389,222, Issued 14 February 1995.

8 IPMC Based Tactile Displays for Pressure and Texture Presentation on a Human Finger Masashi Konyo and Satoshi Tadokoro Graduate School of Information Sciences, Tohoku University, Japan

8.1 Introduction Tactile sensation is an important cue for us to find a subtle difference in handling objects. For a robotic telemanipulation surgery, tactile displays, which produce virtual cutaneous sensation on human hands, provide haptic feedback to the operators for finding diseases carefully and manipulating tools dexterously. For virtual reality applications in medicine, such as rehabilitation and psychotherapy, a tactile display can also contribute to human emotional responses because tactile sensation is highly related to both our comfort and wonder. Many researchers have developed tactile displays and haptic interfaces, as reported by Benali-Khoudja [1] and Hayward [2]. However, it is difficult for conventional tactile displays to synthesize rich and complex tactile sensation arbitrarily. The most characteristic feature of tactile sensation is the variety of perceptual contents reflected, from physical factors such as rigidity, elasticity, viscosity, friction and the surface shape of the target material. It is interesting that tactile receptors in human skin cannot sense the physical factors directly. They only detect the inner skin deformation caused by contact with the object. This fact suggests that a tactile display does not have to reproduce the same physical factors of the material to represent virtual touch. In other words, virtual touch needs only to reproduce the internal deformations in the skin. Furthermore, reproduction of only the nervous activities of the tactile receptors can provide the virtual touch regardless of the inner


162


deformation. Based on this standpoint, several researchers have proposed tactile display methods to produce virtual touch by generating stimulation of each tactile receptor selectively. Shinoda [3] developed a tactile display using a magnetic oscillator and air pressure to make selective stimulation. Kajimoto [4] used electrocutaneous stimulation to control nervous excitements. Their methods could represent several complex tactile sensations but the problem remains of how to stimulate all kinds of tactile receptors. In addition, active perceptual processes based on touch by movement by hands are quite important for human tactile perception. Hand movement is used consciously or actively to clarify material properties. Gibson [5] reported that active touch is superior to passive touch in quality and quantity. For the conventional tactile displays, however, it was difficult to perform touch movement freely in a 3-D space due to the limits of the size and weight of the device. Electroactive polymer (EAP) materials have many attractive characteristics as a ‘soft’ and ‘light’ actuator for realizing a compact tactile display. The authors have developed a tactile display using ionic polymer–metal composite (IPMC) actuators [6–9]. We successfully developed a wearable tactile display presenting mechanical stimuli on a finger in response to hand movements [9]. In our research, the target of tactile information is quite different from conventional ones. Our display can synthesize several tactile sensations, such as pressure, roughness and friction [9]. It can also produce rich and complex touch feel, even including qualitative information such as a haptic impression or material feel when we stroke a surface of cloth [6, 9]. In this chapter, we introduce the wearable tactile display and the tactile synthesis method using IPMC actuators. Our display can represent texture feeling and pressure sensation by controlling three physical characteristics: roughness, softness and friction.

8.2 IPMC Actuators as a Tactile Stimulator Conventional tactile displays could hardly control delicate tactile sensation, because it was difficult to make fine distributed stimuli on a human skin under the limitations of their actuators, such as magnetic oscillators, piezoelectric actuators, shape memory alloy actuators, pneumatic devices, and so on. EAP materials have many attractive characteristics as a ‘soft’ and ‘light’ actuator for such a stimulation device. The ionic polymer–metal composite (IPMC, which is also known as ICPF in the robotics field) [10, 11] is one of the electroactive polymers that have shown potential for practical applications. The IPMC is an chemically plated electroactive polymer (EAP) material that bends when subjected to a voltage across its thickness (Figure 8.1). A Nafion–gold composite type IPMC [12], which contained the sodium ion, has a relatively good Electrode Pt or Au layer Bending movement PFS membrane

Figure 8.1 Ionic polymer–metal composite (IPMC) actuator.

IPMC Based Tactile Displays for Pressure and Texture Presentation

163

performance in both response speed and displacement. The IPMC generates a relatively small force where a cantilever-shaped actuator (2 10 0.18 mm) can generate about 0.6 mN, and therefore its applications need to be scoped accordingly. Some of the applications that were investigated for IPMCs include an active catheter system [13, 14], a distributed actuation device [15–17], and so on. The IPMC has many advantages for a tactile stimulator, including: 1. High spatial resolution: The required spatial resolution for stimulating sensory receptors, especially Meissner’s corpuscle in the fingertip, is less than 2 mm. IPMC films are easy to shape, and their simple operating mechanism allows a stimulator to be miniaturized to make a high-density distributed structure as shown in Figure 8.2a. Conventional actuators can hardly control such a minute force because of their heavy equivalent mass and high mechanical impedance. An IPMC has enough softness that special control methods are not required to use the passive material property.

Human skin Fingertip

Normal stress Shearing stress

Bending movement

IPMC actuator

IPMC actuators (a) Distributed stimulation

(b) Multi-directional stimulation

Figure 8.2 IPMC actuators as tactile stimulators.

2. Wide frequency range: A tactile display can stimulate several tactile receptors selectively by changing frequency ranges because each tactile receptor has different time response characteristics for vibratory stimulation [9]. The required frequency range is from 5 to 200 Hz to stimulate all kinds of tactile receptors. The response speed of an IPMC is fast enough to make a vibratory stimulation on a skin higher than 200 Hz. This means that an IPMC can stimulate all receptors selectively. 3. Stimuli in multiple directions: Each of the tactile receptors has selectivity for the direction of mechanical stimuli. Meissner’s corpuscle detects especially the shearing stress toward the skin surface. Figure 8.2b shows that bending motions of an IPMC, which contacts with a surface of skin in a tilted position, make a stress in both the normal direction and shearing direction. 4. Wearability: To generate the virtual reality of tactile feel, we should move our hand actively and freely, and receive appropriate stimuli in response to the hand movements. An IPMC based wearable display was successfully developed, which was made so small in size and weight that there was no interference with hand movements [8]. 5. Safety: The low driving voltage (less than 5 V) is safe enough to touch with a human finger directly.

164


8.3 Wearable Tactile Display A wearable tactile display was developed using IPMC actuators [8]. Figures 8.3a and 8.3b show a close-up view of the stimulator and of it mounted on a fingertip. The structure of the wearable stimulation device is shown in Figure 8.4. The ciliary part is provided with Au-Nafion composite actuators, where each cilium is 3 mm long and 2 mm wide, in 12 rows leaving 1 mm gaps horizontally and 1.5 mm gaps vertically. All cilia are tilted 45° to transmit mechanical stimuli both in the normal and the tangential directions to the surface of the skin efficiently, as shown in Figure 8.2. The power supply line of the IPMC is provided with a flexible wiring board in to minimize restrictions on the hand, so the fingertip can be bent. The use of silicon rubber of 25 25 8 mm applied to the base of the ciliary part has made it possible to lighten the device to approximately 8 g including the flexible wiring board.

(a)

(b)

Figure 8.3 Overviews of the wearable tactile display: (a) close-up of the IPMC stimulator; (b) mounted on a finger.

Figure 8.4 Structure of the tactile stimulator using IPMC actuators.


165

An IPMC needs to be kept moistened because its actuators are operated by ionic migration. Even in slightly moist air, however, the device can provide stimuli sufficiently for several minutes. The total display system is shown in Figure 8.5. The stimulation device is attached to the middle fingertip. The system is designed to read positional information of the hand using Polhemus FASTRAK, which can read information according to a magnetic field.

Figure 8.5 Wearable tactile display system in response to virtual contact motion.

8.4 Selective Stimulation Method for Tactile Synthesis In human skin, tactile receptors generate elementary sensations such as touch, pressure, vibratory sensation, pain, temperature sense, and so on. A tactile impression is an integrated sensation of these elementary sensations. To present tactile feel arbitrarily, stimuli applied to these receptors should be controlled selectively and quantitatively. As mentioned previously, tactile receptors cannot sense the physical factors of environments directly. They detect only the skin deformation caused by contacting objects. A tactile illusion can be provided by the reproduction of the activities of tactile receptors, regardless of the inner deformations. The concepts of the selective stimulation method are illustrated in Figure 8.6. There are four types of mechanoreceptors embedded in human fingers: FA I (Meissner’s corpuscle), SA I (Merkel corpuscle), FA II (Pacinian corpuscle) and SA II (Ruffini endings) [18]. It is known that each receptor has temporal response characteristics for mechanical stimulation and causes subjective sensation corresponding to its responsive deformation. For example, SA I detects static deformations of skin and produces static pressure sensation, and FA I detects the velocity of the deformation and produces the sense of fluttering vibration. Tactile impression is an integrated sensation of these elementary sensations. To present tactile feel arbitrarily, stimuli applied to these receptors should be controlled selectively.

166


Figure 8.6 Concept of the selective stimulation method for tactile synthesis.

Figure 8.7

Thresholds of tactile receptors for vibratory stimulus.

The first problem is how to stimulate each receptor selectively. We have focused on the frequency response characteristics of the tactile receptors. Figure 8.7 illustrates the human detection threshold against vibratory stimuli, which represents the sensitivity of each receptor to frequency variation [19]. A smaller amplitude threshold means higher sensibility. This figure shows that there are three frequency ranges in which the most sensitive receptor changes. In the lowest frequency range, SA I is most sensitive, relatively. The best becomes FA I in the middle range and FA II is best in the highest range. This suggests that the selective stimulation can be realized using these frequency characteristics, and arbitrary tactile feels can be produced by synthesizing several frequency components.


167

For the IPMC tactile display, selective stimulation is realized by changing drive frequencies, using the receptors’ response characteristics. It was confirmed by a subject’s introspection that the contents of sensation vary with the change of drive frequency as follows: 1. Less than 5 Hz: static pressure sensation (SA I). 2. 10–100 Hz: periodic pressing or fluttering sensation, as if the surface of a finger is wiped with some rough material (FA I). 3. More than 100 Hz: simple vibratory sensation (FA II).

8.5 Texture Synthesis Method We focused on the following three sensations to produce a total textural feeling related to the physical properties of materials: softness, roughness and friction. These sensations are fundamental to express the textural feel of cloth-like materials. The three sensations are produced by the following parameters based on the proposed method described later: 1. Softness sensation: the amount of pressure sensation when the finger contacts the surface (Section 8.6). 2. Roughness sensation: changes in the frequency and the amplitude caused by the relationship of the wavelength of the desired surface and hand velocity (Section 8.7). 3. Frictional sensation: changes in the amount of subjective sensation in response to hand acceleration when the finger slides across a surface (Section 8.8). The problem is how to connect the stimulation on each receptor with contact phenomena caused by hand movements and the physical properties of objects. We have proposed stimulation methods connected to the relationship between hand movements and the physical properties of objects [9]. The softness sensation can also represent concavo-convex surface using as a pressure distribution map as shown in Figure 8.8. When we stroke on the high-pressure area, which

Figure 8.8 Display of concavo-convex surface with texture feel using pressure distribution map.

168


is represented by gray-scaled colour, we can feel convex surface. If the tactile stimulation for texture feeling is superposed on the pressure map, we can feel the concavo-convex feel with the texture feel. For the roughness sensation, the frequencies of natural stimuli caused by contacting rough surfaces are changed in response to hand movements. Human beings have the possibility to use those changes of frequencies positively. It is known that the slope of the detection threshold of FA I is –1 in the range of less than 40 Hz. The activities of FA I reflects vibratory frequencies proportionally. This means that FA I can perform as a frequency analyser in a certain range. Based on this hypothesis, we proposed a frequency modulation method for displaying the roughness sensation in response to hand velocity, as described in Section 8.7.

8.6 Display Method for Pressure Sensation 8.6.1

Method

It is known that SA I detects static deformations of the skin and generates a static pressure sensation [18]. Therefore, selective stimulation on SA I can generate pressure sensations. As shown in Figure 8.7, the detection thresholds of SA I have flat frequency characteristics in the range of less than 100 Hz. In most of the range in Figure 8.7, FA I is more sensitive than SA I. However, in the range of less than 5 Hz, SA I becomes more sensitive than FA I. This means that very low frequency vibration can generate pressure sensations relatively larger than the sensation of FA I. The authors confirmed that this assumption was true when the amplitude of simulation was small enough not to sense the vibratory sensation.

8.6.2 Evaluation In this experiment [9], the wearable tactile display system shown in Figure 8.5 was used. The subjects put the device on the right middle finger. They could perform a stroking movement in the horizontal direction. The stimulation was simple sinusoidal vibrations at a frequency from 2 to 5 Hz. The stimulations were generated only when the hand velocity was higher than 25 mm/s despite the direction of movement. To measure the pressure sensation, the subjects pushed their left middle finger on a sponge that was set on an electric balance, controlling their finger to the same amount of the pressure sensation of the artificial pressure sensation for three seconds. Then, the amount of the pressure sensation was calculated as the mean of the force for three seconds. Figure 8.9 shows the relationship between the amplitude of vibration and the amount of pressure sensation at each frequency. The amounts of pressure sensation were calculated by a Z-score because the subjects had different sensitivities for the amount of the subjective sensation. The number in the parenthesis shows the mean value of actual forces at the frequency of 5 Hz as a reference. It was confirmed that as the amplitudes increase, the pressure sensations became larger for every frequency component.


169

Figure 8.9 Pressure sensation vs driving voltage of low-frequency stimulation for SA I.

Using this method, the softness of materials, which we feel instantaneously when the finger touches a surface, can be expressed by the parameter of amplitude for the frequency components of 5 Hz. If the pressure sensation is larger, the contacting object has more stiffness.

8.7 Display Method for Roughness Sensation 8.7.1

Method

As mentioned in Section 8.5, we suppose that human beings perceive the roughness sensation as the change in frequency detected by FA I in the relationship between their hand movements and the physical properties of the roughness of materials. The roughness of the surface is defined approximately as a sinusoidal surface, which has a given wavelength l as shown in Figure 8.10. When the finger slides on the sinusoidal surface

Figure 8.10 Definition of surface form using wavelength.

170


at a given velocity, v, the frequency of stimuli, f, which are generated in a fingertip, is expressed by a wave equation: f ¼

v l

ð8:1Þ

This equation shows that if the hand velocity becomes faster or if the wavelength, l, becomes smaller, the frequency, f, increases. We should consider the response characteristics of FA I, which is known as a tactile receptor related to the roughness sensation. It is known that FA I responds to the velocity of mechanical stimuli [18]. Here, when the finger slides across the surface, as shown in Figure 8.10, a displacement of stimulus, y, at a given time, t, is defined as a sinusoidal function: y ¼ a sinð2pftÞ

ð8:2Þ

where a is the amplitude of stimulation. Thus, the velocity of stimulation is expressed by substituting Equation (8.1) in the following equation: dy v v ¼ 2pa cosð2p tÞ dt l l

ð8:3Þ

This equation presents the information detected by FA I and shows that both the amplitude and the frequency change in response to the velocity, v. Based on this assumption, the roughness sensation can be presented by changing both the frequency and the amplitude of stimulation in accordance with hand velocity. In this manner, the roughness sensation can be defined by the wavelength, l. For practical use of this method, we applied phase adjustments to produce smooth outputs in response to changing frequencies with respect to each sampling time. The frequencies generated by the proposed method depend on the wavelength and hand movements. However, usual hand movement on the surface of several millimeter wavelengths generates the suitable frequency range for FA 1 consequently.

8.7.2

Evaluation

As evaluation indexes of the roughness sensation, nine kinds of close-set lead balls that had different diameters from 0.5 to 10 mm were used. The wearable tactile display system shown in Figure 8.5 was used. The amplitudes of the stimulations were fixed at 6.0 V ( ¼ the maximum input) and each offset was 0.5 V. The offset was needed to avoid an insensitive zone caused by shortage of amplitudes of the actuators. The subjects put the device on the right middle finger. They touched the index with their left hand at the same time. There was no restriction on time to explore. The subjects were six males in their twenties. Figure 8.11 shows the relationship between the defined wavelengths and the mean value of selected indexes with each error bar representing one standard deviation. The results showed that as the defined wavelength became longer, the roughness sensation seemed to increase when the two half groups were considered separately. Especially, as the


171

Figure 8.11 Roughness index selected vs wavelength of the stimuli.

wavelengths became shorter, the standard deviations became smaller and the roughness sensations were expressed clearly. From the results, it was confirmed that roughness sensation could be expressed by the parameter of the wavelength in the case of relatively short wavelengths. In addition to the wavelength, it is confirmed that the maximum amplitude of stimulus affects the amount of the subjective sensation of roughness.

8.8 Display Method for Friction Sensation To express a cloth-like textural feeling in response to contact motions, synthesis of both the roughness sensation and softness sensation is not enough. In this section, we introduce the friction sensation. In this study, the definition of friction sensation is not a usual description based on physical contact conditions. We assumed that the friction sensation can be produced as changes in the amount of subjective sensation in response to hand acceleration when the finger slides across the surface. Especially, the friction sensation is used to express the sticking tendency of materials at the beginning of sliding motion. The authors confirmed that stimulation of high-frequency components corresponding to the acceleration of hand movements could produce a natural sliding feeling [9]. It is known that FA II detects the acceleration of stimuli, and it seems that FA II is related to the detection of hand movements such as by a gyro sensor. Figure 8.12 illustrates the relationshipship between hand acceleration and amplitudes of the high-frequency component. The high-frequency component is fixed at 200 Hz, in which FA II become most sensitive. Therefore, the parameters of the friction sensation are the maximum and minimum values of the amplitude shown in Figure 8.12.

172


Figure 8.12 Relationship between the amplitude of high-frequency components for the friction sensation and the acceleration of hand movements.

8.9 Synthesis of Total Textural Feeling 8.9.1

Method

In this section, syntheses of total textural feeling related to the physical properties of materials based on the three methods described above were evaluated. The voltage inputs generated by the three methods were combined into a signal by a simple superposition. Four materials were selected as targets of the tactile syntheses. The artificial textural feelings were tuned subjectively by changing the parameters of the roughness, softness and friction sensations. The tunings of textural feelings were extremely easy compared with the author’s conventional study because each parameter was related to the physical properties of the materials. The following were the properties of the four materials and the tuned parameters: 1. Boa: shaggy, thick, uneven and very rough surface (l ¼ 10, a ¼ 5.0, P ¼ 0.0, Fmax ¼ 2.0) 2. Towel: rough surface, thick and soft (l ¼ 2.0, a ¼ 3.0, P ¼ 2.0, Fmax ¼ 1.0) 3. Fake leather: flat surface, thin, hard and high friction (l ¼ 8.0, a ¼ 1.0, P ¼ 4.0, Fmax ¼ 3.0) 4. Fleece: smooth surface, thin, soft and low friction (l ¼ 0.5, a ¼ 1.0, P ¼ 5.0, Fmax ¼ 1.0) 8.9.2

Experiments

Four artificial textures, which were tuned as mentioned above, were evaluated. The four real materials, that is boa, towel, fleece and fake leather, were used to compare the artificial tactile feels. The wearable tactile display system shown in Figure 8.5 was used. The subjects put the device on the right middle or index finger. They could perform stroking motions with their left hand in the horizontal direction. Before the experiments began, the


173

subjects had experience with the four textural feelings only once. The subjects compared each artificial texture with the corresponding real material. They were asked to evaluate the similarity of both feelings at five levels (1: Poor, 2: Fair, 3:Good, 4:Very Good and 5:Excellent). There was no restriction on time to explore the textures. The subjects were divided into two groups: three sight-restricted people (two females in their fifties and one female in her forties) and five ordinary persons (five males in their twenties). The sight-restricted people have more sensitive tactile sensation than ordinary persons. It was expected that the sight-restricted people could evaluate more correctly. Figure 8.13 shows the evaluation results for the sight-restricted people and the ordinary persons, respectively. Both the sight-restricted people and the ordinary persons judged a

Figure 8.13 Evaluation of artificial tactile feeling compared with the real materials.

score of more than three, that is ‘Good’, for the almost all artificial textures. These results demonstrated that the proposed methods could synthesize artificial textural feeling corresponding to the real materials. In addition, the sight-restricted people gave higher evaluations than the ordinary persons so that the synthesized textural feelings had the reasonable reality.

8.10 Conclusions In this chapter, we introduce the wearable tactile display and the tactile synthesis method using IPMC actuators. Our display can represent texture feeling and pressure sensation by controlling three physical characteristics: roughness, softness and friction. These parameters of textural feeling can be measured as physical properties. This means that artificial textural feelings could be synthesized automatically, if the tactile sensors could detect such physical parameters. The authors are also developing the tactile transmission system combining the tactile display and tactile sensors as a master–slave system [20].

References 1. Benali-Khoudja, M., Hafez, M., Alexandre, J. M. and Kheddar, A. (2004) Tactile interfaces: a state-of-the-art survey, Proceedings of the 35th International Symposium on Robotics (ISR 2004), Paris France, 23–26 March, 721–726.

174


2. Hayward, V. and MacLean, K. E. (2007) Do It Yourself Haptics, Part-I, IEEE Robotics and Automation Magazine, 14 (4), 88–104. 3. Asamura, N., Yokoyama, N. and Shinoda, H. (1998) Selectively Stimulating Skin Receptors for Tactile Display, IEEE Computer Graphics and Applications, 18 (6), 32–7. 4. Kajimoto, H., Inami, M., Kawakami, N. and Tachi, S. (2004) SmartTouch: Electric Skin to Touch the Untouchable, IEEE Computer Graphics & Applications, 23 (1), 36–43. 5. Gibson, J. J. (1962) Observation on Active Touch, Psychological Review, 69 (6), 477–91. 6. Konyo, M., Tadokoro, S., Takamori, T. and Oguro, K. (2000) Artificial tactile feel display using soft gel actuators, Proceedings of the IEEE Intel. Conference on Robotics and Automation, 3416–3421, San Francisco, April 24–28. 7. Konyo, M., Tadokoro, S., Hira, M. and Takamori, T. (2002) Quantitative Evaluation of Artificial Tactile Feel Display Integrated with Visual Information, Proceedings of the IEEE/RSJ International Conference on Intelligent Robotics and Systems, 3060–5, Lausanne, Switzerland, Sep 30–Oct 4. 8. Konyo, M., Akazawa, K., Tadokoro, S. and Takamori, T. (2003) Wearable Haptic Interface Using ICPF Actuators for Tactile Feel Display in Response to Hand Movements, J. Robotics and Mechatronics, 15 (2), 219–26. 9. Konyo, M., Yoshida, A., Tadokoro, S. and Saiwaki, N. (2005) A tactile synthesis method using multiple frequency vibration for representing virtual touch, Proceedings of the IEEE/RSJ International Conference on Intelligent Robotics and Systems, 1121–7, Edmonton, Canada, August 2–6. 10. Oguro, K., Kawami, Y. and Takenaka, H. (1992) Bending of an ion-conducting polymer filmelectrode composite by an electric stimulus at low voltage, J. Micromachine Society, 5, 27–30. 11. Shahinpoor, M. (1992) Conceptual Design, Kinematics and Dynamics of Swimming Robotic Structures using Ionic Polymeric Gel Muscles, Smart Materials and Structures, 1 (1), 91–4. 12. Fujiwara, N., Asaka, K., Nishimura, Y. et al. (2000) Preparation and gold-solid polymer electrolyte composites as electric stimuli-responsive materials, Chem. Mat., 12, 1750–4. 13. Guo, S., Fukuda, T., Kosuge, K., et al. (1995) Micro catheter system with active guide wire, Proceedings of the IEEE International Conference on Robotics and Automation, 79–84, Nagoya, Japan, May 24–26. 14. Onishi, Z., Sewa, S., Asaka, K., et al. (2000) Bending response of polymer electolyete acutator, Proc. SPIE Smart Structures and Materials 2000: Electroactive Polymer Actuators and Devices (EAPAD), Newport Beach, USA, March 6–8, 121–128. 15. Tadokoro, S., Murakami, T., Fuji, S., et al. (1997) An elliptic friction drive element using an ICPF (ionic conducting polymer gel film) actuator, IEEE Control Systems, 17 (3), 60–8. 16. Tadokoro, S., Fuji, S., Fushimi, M., et al. (1998) Development of a distributed actuation device consisting of soft gel actuator elements, Proceedings of the IEEE International Conference on Robotics and Automation, 2155–60, Leuven, Belgium, May 16–20. 17. Tadokoro, S., Fuji, S., Takamori, T. and Oguro, K. (1999) Distributed actuation devices using soft gel actuators, Distributed Manipulation, K. F. Bo¨hringer, H. Choset (Eds), Springer London, 217–235 (2000). (Kluwer Academic Publishers was absorbed by Springer). ˚ . B. and Johansson, R. S. (1984) Properties of cutaneous mechanoreceptors in the 18. Vallbo, A human hand related to touch sensation, Human Neurobiology, 3, 3–14. 19. Maeno, T. (2000) Structure and Function of Finger Pad and Tactile Receptors, J. Robot Society of Japan, 18 (6), 772–775. (In Japanese). 20. Okamoto, S., Konyo, M., Maeno, T. and Tadokoro, S. (2007) Roughness Feeling Telepresence System on the Basis of Real-time Estimation of Surface Wavelengths, Proceedings of the IEEE/ RSJ International Conference on Intelligent Robots and Systems, 2698–2703, San Diego, USA, Oct 27–Nov 2.

9 IPMC Assisted Infusion Micropumps Il-Seok Park1, Sonia Vohnout2, Mark Banister2, Sangki Lee1,3, Sang-Mun Kim1 and Kwang J. Kim1 1 University of Nevada, USA Medipacs LLC, Tucson, USA 3 Volvo Korea, South Korea

2

9.1 Introduction Ionic Polymer–Metal Composite (IPMC) is an attractive Electroactive Polymer (EAP), which is capable of soft actuation, self sensing and energy harvesting. Among the capabilities of the IPMCs, soft actuators and bio-robotic and/or biomimetic applications constitute especially interesting research fields due to the large bending ability, low driving voltage, easy processing and easy miniaturization of IPMCs. Thus, IPMCs are well considered and adapted to robotic actuators, artificial muscles and miniaturized propulsors. Recently, efforts have been conducted in significant biomedical applications, such as artificial muscles, surgical tools and micro medically-used pumps [1]. Other trials to develop the biomedical applications are currently being carried out in various biomedical disciplines: ophthalmology, proctology and urology [2]. In the future, it is expected that applications for IPMCs will broadly spread not only in small-sized biomedical devices but also in large-scale actuators for naval space’s propulsor, as well as in many industrial applications. In this chapter, a prototype infusion micropump is described, along with the relevant modelling of IPMC based micropumps [3, 4].


176

2


Background of IPMCs

IPMCs are smart materials that exhibit electromechanical (actuator) and mechanoelectrical (sensor) applications. Table 9.1 shows performance properties of state-of-the art IPMCs [5]. They bend quickly under a low voltage, as first reported by Oguro and his co-workers [6]. Later, Abe et al. introduced the important role of existent counter ions and their influence during the bending [7]. Asaka and Oguro introduced a theory of the actuation mechanisms [8]; Shahinpoor and Kim demonstrated that the ionic polymer actuator performance depends on the type of cation [9] and further developed a two-step fabrication method [10] in accordance with their findings. In addition, other groups tried to incorporate various metals as electrode materials to articulate physical properties or electrical responses [11–14]. Table 9.1 Performance properties of typical IPMCs (Reproduced with permission from I-S. Park, K. M. Jung, D. Kim, et al. Physical principles of ionic polymer–metal composites as electroactive actuators and sensors, MRS Bulletin, 33, 3, 190–5. Copyright (2008) MRS) Property

Typical

Estimated actuation speed (strain rate) Estimated bending strain Estimated work density Lifetime

3.3 (% sec 1) 0.5–3.3(%) 5.5 (kJ m 3) Up to 1 million cycles (estimated); 250 000 (experimentally reported)

IPMCs consist of both ion exchange polymers acting as base materials and nano-sized metal layers functioning as resistive and capacitive electrodes. Figure 9.1 shows the crosssectional view of typical platinum IPMC that represents the structure of the IPMC with an electrode layer and deposition/diffusion layer on membrane. The actuation and sensing performance of the IPMCs is highly dependent upon the components of the ion exchange polymer (ionic group and cation) and electrode material. Precious metals (or other conductive mediums), especially platinum and gold, have been adopted for electrode metals [15–21]. Also, there is an increasing interest in IPMC paints (Figure 9.2) [22].

Electrode Layer

Pt deposition/diffusion layer on Nafion

COMPO 15.0kV

X400

10µm

WD 15.2mm

Electrode Layer

Figure 9.1 A cross-sectional view of SEM micromorphology of a platinum IPMC that consists of an electrode and diffusion layer in both surface and membrane polymer as substrate.

IPMC Assisted Infusion Micropumps

177

Figure 9.2 A photograph of spraying Nafion using an airbrush (Reproduced with permission from Park, I.-S., Tiwari, R. and Kim, K. J. Sprayed Sensor Using IPMC Paint, Adv. Sci. Tech., 61, 59–64. Copyright (2008) Trans Tech Publications).

9.3 Miniature Disposable Infusion IPMC Micropumps The availability of safe, disposable and robust infusion pumps for intravenous fluid and drug delivery could provide a significant improvement in both private and public healthcare. An infusion pump should infuse medical fluids, bloods and nutrients into a patient without failure. Thus, there has been a demand for accurately controllable pump systems, from small capacity units such as insulin infusion to large volume feeding suppliers. Small volume infusion pumps, especially, are designed to be portable for use not only in a hospital but also for special purpose likes charity and battlefield use [23]. The former small unit infusion pumps are pressured by human blood pressure using a blood pressure cuff. However, this mechanism leads to serious problems, such as sudden change of the flow rate because the infusion flow rate is only dependent on the patient’s blood pressure. Thus, recently, small volume pumps have been usually operated by an automatically controlled motor or a small embedded system. In addition, for the portable infusion pump, other possible pump power sources have been developed, such as osmotic pressure and spring-powered systems. However, with osmotic power needs it is necessary to change a salt solution bag after finishing infusion and with spring power it is necessary to overcome the limitation of the flow rate for use in various conditions. Furthermore, in order to be a certified infusion operation, it is required that the pump can operate if the power cuts out or is even unplugged and can detect the change of flow rate or pressure even if the flow is blocked or kinked, taken backward and finished when an infusion bag or pump is empty. An IPMC has characteristics both as an actuator and sensor. In addition, it is easy to control the frequency (flow interval) and flow rate by changing of the electric signals with low voltage. We propose a new miniaturized disposal IPMC infusion pump with embedded computer controlling, which can control a micro flow with a low power source, batteries.

178

9.3.1


Configuration of the IPMC Infusion Pump

The prototype IPMC infusion pump was designed as shown in Figure 9.3. This prototype pump shows the assemblies for evaluating IPMC actuators. This system can be easily scalable. The cover, machined from transparent acrylic, has a series of ten indentations to secure the chambers between two fluid inlets and outlets. Two types of cover were designed, (a) oblong and (b) spherical shape as shown in Figure 9.4. Total volume for the two types was calculated using SolidWorks drawing software: • Spherical: 0.0243 in3 ¼ 0.398205 mL ¼ 398.21 L • Oblong: 0.0415 in3 ¼ 0.680063 mL ¼ 680.063 L

Figure 9.3 An illustration of the IPMC infusion micropump (Reproduced with permission from Vohnout, S., Kim, S.-M., Park, I.-S. and Banister, M., IPMC-assisted miniature disposable infusion pumps with embedded computer control, Proceedings of the SPIE conference 2007. Copyright (2007) SPIE).

(a)

(b)

Figure 9.4 Two types of pump chamber cover; (a) oblong; (b) spherical (Reproduced with permission from Vohnout, S., Kim, S.-M., Park, I. S. and Banister, M., IPMC-assisted miniature disposable infusion pumps with embedded computer control, Proceedings of the SPIE conference 2007. Copyright (2007) SPIE).


179

An ionic liquid treated platinum IPMC (IL-Pt IPMC) was cut using an Engravelab Laser Software and Venus V-12 carbon dioxide laser system (Figure 9.5) by matching the cover and chamber size. The top and general view of the laserfabricated IL-Pt IPMC with oblong type chamber cover are shown in Figure 9.6. To increase the pumping ability and effectiveness, another IPMC cutting design was made, the interdigitated 20-finger IPMC, which is shown in Figure 9.7 together with its 3-D micrographic view.

2.58 10 5

25

10 63 Laser beam line width: 0.2 (a)

(mm)

(b)

Figure 9.5 (a) Laser cutting machine and (b) a CAD design for interdigitated IPMC (Reproduced with permission from Vohnout, S., Kim, S.-M., Park, I.-S. and Banister, M., IPMC-assisted miniature disposable infusion pumps with embedded computer control, Proceedings of the SPIE conference 2007. Copyright (2007) SPIE). (See Color Plate 2).

(a)

(b)

Figure 9.6 (a) The top and (b) general view of laser-designed micropump IPMC with the oblong type chamber cover (Reproduced with permission from Vohnout, S., Kim, S.-M., Park, I.-S. and Banister, M., IPMC-assisted miniature disposable infusion pumps with embedded computer control, Proceedings of the SPIE conference 2007. Copyright (2007) SPIE).

180


(b)

IPMC electrode

Figure 9.7 (a) Pictures of the laser cutting designed interdigitated 20-finger IPMC and (b) its digital 3-D micrographic view (Reproduced with permission from Vohnout, S., Kim, S.-M., Park, I.-S. and Banister, M., IPMC-assisted miniature disposable infusion pumps with embedded computer control, Proceedings of the SPIE conference 2007. Copyright (2007) SPIE).

9.3.2

The Control System

A rudimentary controller for the IPMC infusion pump was designed and developed; it employed a Renesas M16C microcomputer (Figure 9.8). In the employed configuration, the controller was capable of actuating ten cells in various patterns. In order to operate the cell’s fingers in two opposing directions, each cell was actuated by a square wave with 50 % duty cycle (12 second cycle time). The finger cells actuated with approximately 5 V limited to a 10 mA signal under this controller. The polarity can be reversed or neutral (þ, 0, –) to allow control of the speed sequence and time of the actuation cycle.

Figure 9.8 The controller for the IPMC micropump (Reproduced with permission from Vohnout, S., Kim, S.-M., Park, I.-S. and Banister, M., IPMC-assisted miniature disposable infusion pumps with embedded computer control, Proceedings of the SPIE conference 2007. Copyright (2007) SPIE).


9.3.3

181

Performance Testing

For successful infusion performance, an infusion pump has to possess the pressure of about less than 55 kPa and 125 kPa in a vein and for epidural or subcutaneous purposes, respectively [24]. Thus, in order to meet the optimum pressure generation, the oblong pump chamber type was adopted. Two IPMC fingers are arrayed into each oblong section of the pump chamber and this configuration was expected to double the force generated. This 20-finger actuator successfully infused the liquid and sequenced with one and two at a time. For achievement of the entire sequence time of the pump within 1–3 s, we are still continuing to sequence the finger actuation program down to fractions of a second. Calculations estimate that the chamber volume can be displaced every three seconds and that we have to subtract the volume of at least one section for each sequence cycle. It would be calculated to 9 0.068 ml ¼ 0.612 ml per cycle time. Therefore, the pump should produce a flow rate of approximately 12.24 ml/min at 1 psi. 9.4

Modelling for IPMC Micropumps

Diaphragms are widely used to create stroke volumes [25, 26]. IPMCs are new, promising materials used for actuating diaphragms in micropumps. In this modelling, systematic design methods on an IPMC actuator-driven valve-less micropump are introduced [3, 4]. IPMC with a Nafion membrane was considered the best material for an actuating diaphragm. In order to estimate deformed shapes of a circle-shaped IPMC diaphragm, the equivalent bimorph beam model [27] for IPMC actuators was used in conjunction with the finite element method (FEM). Using this model, several parametric studies were performed to determine an optimal electrode shape of the IPMC diaphragm and to investigate the pressure effect on the stroke volume. 9.4.1

Equivalent Bimorph Beam Model for IPMC Actuators

To predict the behaviour of IPMC diaphragms, the equivalent bimorph beam model, which was recently introduced by Lee et al. [28], is adopted in this study. Here, the key ideas of the model are summarized. In the equivalent bimorph beam model (Figure 9.9), it is assumed that an IPMC has two virtual layers of the same thickness. Under an imposed electric field across the IPMC, the

Figure 9.9 Typical shape of a bimorph beam (Reproduced with permission from Lee, S., Kim, K. J. and Park, H. C. (2006) Modeling of an IPMC Actuator-driven Zero-Net-Mass-Flux Pump for Flow Control, J. Intelligent Mat. Systems and Structures, 17, 6, 533–41. Sage Publications).

182


upper layer and the lower layer of an IPMC expand or contract, opposing each other, to produce the IPMCs bending motion. Equation (9.1) shows a relationship between the input voltage (V) and the induced tip displacement (s) of an IPMC with an equivalent bimorph beam: s¼

3d31 VL2 2H 2

ð9:1Þ

where d31 (m/v) is the ‘effective’ electromechanical coupling coefficient for small deflections in which the subscripts 1 and 3 stand for the x-direction and the z-direction, respectively. From Equation (9.1), d31 is expressed as follows: d31 ¼

2sH 2 3VL2

ð9:2Þ

Substituting the experimentally measured tip displacement into Equation (9.2), d31 can be obtained for a given input voltage. In the equivalent bimorph beam model, the Young’s modulus (E) contributing to the bending stiffness of an IPMC is determined from the blocking force Equation (9.3) of a bimorph beam: Fb1 ¼

3WH 2 E d31 E3 8L

ð9:3Þ

where Fb1 is the measured blocking force, E3 is the electric field and d31 is previously calculated from Equation (9.2). To determine the equivalent of Young’s modulus (E), Equation (9.3) is rewritten as: E¼

8LFb1 3WH 2 d31 E3

ð9:4Þ

For all numerical analyses, a commercial finite element analysis program, MSC/ NASTRAN [29], was used in conjunction with the equivalent bimorph beam model. A thermal analogy technique proposed by Taleghani and Campbell [30] was used to implement the electromechanical coupling effect into the finite element model. In the thermal analogy technique, the electromechanical coupling coefficient (d31) is converted into the thermal expansion coefficient 1 as follows: 1 ¼

d31 t

ð9:5Þ

where t is the thickness of a layer across which an electric potential is applied, and then the electric potential (DV) is replaced by the temperature difference (DT). More details and verifications for the thermal analogy technique can be found in Lee et al. [27] and Lim et al. [31]. 9.4.2

IPMC Diaphragm

Deformations of circle-shaped IPMC diaphragms were analysed for the circle-shaped and ring-shaped electrode, respectively. Through parametric studies, an electrode shape was chosen for the optimal diaphragm, which generates maximum stroke volume. In order to show the effectiveness of the circle-shaped diaphragm, its stroke volume was compared to


183

that of a square-shaped diaphragm maintaining the same actuator area. Both the normal mode analysis and pressure effect on the selected IPMC diaphragm were introduced. 9.4.2.1

Circle-Shaped Electrode vs Ring-Shaped Electrode

Parametric studies on two kinds of electrodes for circle-shaped diaphragms with a radius of 10 mm were conducted with the material properties and thicknesses shown in Table 9.2 [3]. The material properties E and d31 of the IPMC in Liþ form were determined thorough the equivalent bimorph beam model [28]. The elastic modulus of Nafion in Liþ form and Poisson’s ratios were obtained from literature [32, 33]. Figure 9.10 shows the shape of the two electrodes in a one-quarter finite element model of diaphragms. Figure 9.10a is the circle-shaped electrode and 9.10b is the ring-shaped electrode. Table 9.2 Material properties and thicknesses for an IPMC diaphragm (Reproduced with permission from Lee, S., Kim, K. J. and Park, H. C. (2006) Modeling of an IPMC Actuatordriven Zero-Net-Mass-Flux Pump for Flow Control, J. Intelligent Mat. Systems and Structures, 17, 6, 533–41. Sage Publications) Elastic Modulus (GPa)

Poisson’s Ratio

*

þ

IPMC in Li form Nafion in Liþ form

1.158 0.05

d31 (m/V)

0.487 0.487

1.750 10 N/A

t (mm) 7

0.2 0.2

* This IPMC is heavily loaded with platinum (~6 vol-% Pt). The Pt loading technique was uniquely designed to enhance the humidity control of IPMC.

Sym.

Sym. Fixed B.C. along the edge

Fixed B.C. along the edge

Sym.

Sym.

Radius of electrode

Radius of electrode

Gray part: IPMC (or electrode) and black part: Nafion (a)

(b)

Figure 9.10 Shapes of electrodes for IPMC diaphragms (¼ finite element model): (a) circleshaped; (b) ring-shaped (Reproduced with permission from Lee, S., Kim, K. J. and Park, H. C. (2006) Modeling of an IPMC Actuator-driven Zero-Net-Mass-Flux Pump for Flow Control, J. Intelligent Mat. Systems and Structures, 17, 6, 533–41. Sage Publications).

184


The total number of elements (Quad4, MSC Software Corp.) [29] used for each model was 400. The symmetry boundary condition was applied to the vertical and horizontal lines, and fixed boundary condition to the outside edge. As shown in Figure 9.10, each IPMC diaphragm consists of the IPMC part and a Nafion part. Therefore, when a voltage is applied on an IPMC part, the vertical interface between IPMC and Nafion can rotate easily to produce large bending deformation, since Nafion has a low elastic modulus. Under an applied 2 V input and fixed boundary conditions along the outside edge, the centre displacements of the diaphragms were calculated with variations of the electrode length in the radial direction. The calculated results are provided in Figure 9.11. For the diaphragm with the circle-shaped electrode, in which the radius of electrode was 8.5 mm, the maximum centre displacement was 0.996 mm. The maximum centre was only 0.686 mm for the diaphragm with the ring-shaped electrode and is more efficient than the ring-shaped electrode in terms of deformation. The parametric studies suggest that there is an optimal radius and radial length to each electrode for the maximum deflections.

1 Center displacement (mm)

Center displacement (mm)

1 0.8 0.6 0.4 0.2 0

0

2

4

6

8

10

0.8 0.6 0.4 0.2 0

0

2

4

6

8

Radius of electrode (mm)

Radial length of electrode (mm)

(a)

(b)

10

Figure 9.11 Centredisplacement of IPMC diaphragms for each electrode case: (a) circle-shaped electrode; (b) ring-shaped electrode (Reproduced with permission from Lee, S., Kim, K. J. and Park, H. C. (2006) Modeling of an IPMC Actuator-driven Zero-Net-Mass-Flux Pump for Flow Control, J. Intelligent Mat. Systems and Structures, 17, 6, 533–41. Sage Publications).

For the two optimal electrode cases (radius, 8.5 mm for the circle-shaped electrode; radial length, 5.5 mm for the ring-shaped electrode), stroke volumes were calculated from the diaphragm with the deformed shapes shown in Figure 9.12. Note that the diaphragm with the circle-shaped electrode is bent upward while the diaphragm with the ring-shaped electrode is bent downward for the same electrical input of 2 V. Considering 2 V AC input, calculated stroke volumes (also the definition of stroke volume later in Figure 9.16) for the circle-shaped and the ring-shaped electrode cases were 216 and 104 mL, respectively.



185

9.66–004 9.02–004 8.38–004 7.73–004 7.09–004 6.44–004 5.80–004 5.15–004

9.66–004

0.

4.51–004 3.87–004

+

3.22–004 2.58–004 1.93–004 1.29–004 6.44–005

Z

–1.16–010 Y

default_Fringe: Max 9.66–004 @Nd1 Min 0. @Nd 316 default_Deformation: Max 9.66–004 @Nd1

X

(a)


0 –4.57–005 –9.14–005 –1.37–004 –1.83–004 –2.29–004 –2.74–004 –3.20–004

0. +

–3.66–004 –4.11–004 –4.57–004

–0

–5.03–004 –5.48–004 –5.94–004 –6.40–004 Z

–6.86–004 Y X

(b)

default_Fringe: Max 0. @Nd316 Min –6.86–004@Nd1 default_Deformation: Max 6.86–004 @Nd1

Figure 9.12 Deformed shapes of IPMC diaphragms: (a) circle-shaped electrode (radius of electrode ¼ 8.5 mm); (b) ring-shaped electrode (radial length of electrode ¼ 5.5 mm) (Reproduced with permission from Lee, S., Kim, K. J. and Park, H. C. (2006) Modeling of an IPMC Actuator-driven Zero-Net-Mass-Flux Pump for Flow Control, J. Intelligent Mat. Systems and Structures, 17, 6, 533–41. Sage Publications). (See Color Plate 3).

186


9.4.2.2

Circle-Shaped Diaphragm vs Square-Shaped Diaphragm

The shape effect of the diaphragm on the stroke volume was also investigated. A squareshaped diaphragm was modelled and analysed to calculate its centre displacement and stroke volume. The results of the square-shaped diaphragm were compared with those of the circle-shaped diaphragm. The areas of the IPMC and Nafion were maintained the same as for the circle-shaped diaphragm – the optimal case (i.e., the radius of diaphragm: 10 mm and the radius of electrode: 8.5 mm). Material properties and thicknesses are shown in Table 9.2. Figure 9.13 shows the shapes of the two diaphragms (¼ finite element model). For the finite element modelling, 400 and 324 elements (Quad4, MSC.Software Corp.) were used for the circle-shaped diaphragm and the square-shaped diaphragm, respectively. Sym.

Sym.

Sym.

Sym. Gray part: IPMC (or electrode) and black part: Nafion (a)

(b)

Figure 9.13 Shapes of IPMC diaphragms (¼ finite element model): (a) circle-shaped; (b) squareshaped (Reproduced with permission from Lee, S., Kim, K. J. and Park, H. C. (2006) Modeling of an IPMC Actuator-driven Zero-Net-Mass-Flux Pump for Flow Control, J. Intelligent Mat. Systems and Structures, 17, 6, 533–41. Sage Publications).

Under a 2 V input, for the square-shaped diaphragm’ the calculated centre displacement and stroke volume were 0.760 mm and 196 mL, respectively. Note that the calculated values for the circle-shaped diaphragm are 0.996 mm and 216 mL, respectively. From the results, it is evident that the use of the circle-shaped diaphragm is advantageous over the square-shaped one in order to generate larger stroke volumes. 9.4.2.3

Normal Mode Analysis

The normal mode analysis was performed for the optimal circle-shaped diaphragm with the circle-shaped electrode (radius of electrode: 8.5 mm). For the calculation, the density of Nafion in Liþ form was 2.078 103 kg/m3 [34] and that of IPMC in Liþ form was assumed to be 2.5 103 kg/m3. Figure 9.14 shows the first and second mode shapes of the diaphragm. The computed first (i.e., fundamental) and the second natural frequencies

IPMC Assisted Infusion Micropumps MSC.Patran 2001 r2a 20-Aug04 14:27:41 Fringe: SC1:DIAPHRAGM, A2:Mode 1: Freq. = 429.69: Eigenvectors, Translational-(NON-LAYERED) (MAG) Deform: SC1:DIAPHRAGM, A2:Mode 1: Freq. = 429.69: Eigenvectors, Translational

187

3.15+002 2.94+002 2.73+002 2.52+002 2.31+002 2.10+002 1.89+002 1.68+002

3.15+002

1.47+002

0.

1.26+002

+

1.05+002 8.41+001 6.31+001 4.21+001 2.10+001

Z

4.20–005 default_Fringe: Max 3.15+002 @Nd1 Min 0. @Nd 316 default_Deformation: Max 3.15+002 @Nd1

Y X

(a)

MSC.Patran 2001 r2a 20-Aug-04 14:28:51 Fringe: SC1:DIAPHRAGM, A2:Mode 2: Freq. = 1659.1: Eigenvectors, Translational-(NON-LAYERED) (MAG) Deform: SC1:DIAPHRAGM, A2:Mode 2: Freq. = 1659.1: Eigenvectors, Translational

3.02+002 2.81+002 2.61+002 2.41+002 2.21+002 2.01+002 1.81+002 1.61+002

02 3.02+0

0.

1.41+002 1.21+002

+

1.01+002 8.04+001 6.03+001 4.02+001 2.01+001

Z

3.81–006 Y X

(b)

default_Fringe: Max 3.02+002 @Nd261 Min 0. @Nd316 default_Deformation: Max 3.02+002 @Nd261

Figure 9.14 Normal mode analysis results for an IPMC diaphragm (radius of electrode ¼ 8.5 mm): (a) first mode; (b) second mode (Reproduced with permission from Lee, S., Kim, K. J. and Park, H. C. (2006) Modeling of an IPMC Actuator-driven Zero-Net-Mass-Flux Pump for Flow Control, J. Intelligent Mat. Systems and Structures, 17, 6, 533–41. Sage Publications). (See Color Plate 4).

188


are 430 and 1659 Hz, respectively. If we consider the driving frequency range of the IPMC diaphragm as less than 40 Hz, the calculated fundamental frequency is much higher than the driving frequency range. Therefore, the resonance will not affect the stroke volume in that driving frequency range. The results imply that we can linearly control the flow rates of an IPMC-driven micropump within the driving frequency (40 Hz) of interest. 9.4.2.4

Pressure Effect on Stroke Volume

The external pressure effect on the circle-shaped diaphragm having a circle-shaped electrode was investigated. The external pressure could be considered as the chamber pressure of a pump. In order to calculate numerically, the stroke volume under external pressure, uniform pressures were applied to the FE model for the optimal IPMC diaphragm as shown in Figure 13a. Figure 9.15 shows the estimated half stroke volumes of the optimal circle-shaped diaphragm under the upward pressures and 2 V input. In Figure 9.15, ‘Down Stroke’ indicates the half stroke volume when the diaphragm bends downward and ‘Up Stroke’ indicates the half stroke volume when the diaphragm bends upward, as shown in Figure 9.16 for the definition of half stroke volume. According to these results, in the case of the down stroke, the IPMC diaphragm could generate a stroke volume under the upward external pressure up to approximately 2300 Pa, which lies in a range of dynamic pressures of micro air vehicles (MAVs).

Figure 9.15 Half stroke volumes of the circle-shaped IPMC diaphragm (Reproduced with permission from Lee, S., Kim, K. J. and Park, H. C. (2006) Modeling of an IPMC Actuator-driven Zero-Net-Mass-Flux Pump for Flow Control, J. Intelligent Mat. Systems and Structures, 17, 6, 533–41. Sage Publications).


189

Aslot v

t2 t0 t1 ∆Vh_us

v = Flow (or jet) speed Aslot = Area of slot ∆Vh_us = Half up-stroke volume ∆Vh_ds = Half down-stroke volume ∆V = ∆Vh_us + ∆Vh_ds = Stroke volume

∆Vh_ds

Figure 9.16 Schematic of an IPMC-driven micropump (Reproduced with permission from Lee, S., Kim, K. J. and Park, H. C. (2006) Modeling of an IPMC Actuator-driven Zero-Net-MassFlux Pump for Flow Control, J. Intelligent Mat. Systems and Structures, 17, 6, 533–41. Sage Publications).

9.5 Conclusions In this chapter, we have described an IPMC-driven infusion micropump for recent biomedical applications. Even though the applications of IPMCs for biomedical fields require more trials and development methods, IPMCs are still attractive materials due to their electromechanical and mechanoelectric properties. A systematic design method of an IPMC-driven micropump was introduced. In order to properly estimate the deformed shapes of IPMC diaphragms, the equivalent bimorph beam model for IPMC actuators was conveniently used, in conjunction with the finite element method.

References 1. Shahinpoor, M. and Kim, K. J. (2005) Ionic polymer–metal composites: IV. Industrial and medical applications, Smart. Mat. Struct., 14, 197–214. 2. Kim, K. J. and Tadokoro, S. (2007) Electroactive Polymers for Robotic Application: Artificial Muscles and Sensors, Springer, London. 3. Lee, S., Kim, K. J. and Park, H. C. (2006) Modeling of an IPMC actuator-driven zero-net-massflux pump for flow control, J. Intelligent Mat. Systems and Structures, 17, 533–9. 4. Lee, S. and Kim, K. J. (2006) Design of IPMC actuator-driven valve-less micropump and its flow rate estimation at low Reynolds numbers, Smart Mat. Struct. 15, 1103–9. 5. Park, I-S., Jung, K. M., Kim, D., et al. (2008) Physical principles of ionic polymer–metal composites as electroactive actuators and sensors, MRS Bulletin, 33, 3, 190–5. 6. Oguro, K., Kawami, Y. and Takenaka, H. (1992) Bending of an ion-conducting polymer filmelectrode composite by an electric stimulus at low voltage, J. Micromachine So., 5, 27–30. 7. Abe, Y., Mochizuki, A., Kawashima, T., et al. (1998) Effect on bending behaviour of counter cation species in perfluorinated sulfonate membrane-platinum composite, Polym. Adv. Techn., 9, 520–6. 8. Asaka, K. and Oguro, K. (2000) Bending of polyelectrolyte membrane platinum composites by electric stimuli Part II. Response kinetics, J. Electroanal. Chem., 480 186–98.

190


9. Shahnpoor, M. and Kim, K. J. (2000) Effect of counter-ions on the performance of IPMCs, Proceedings of the SPIE 7th International Symposium on Smart Materials and Structures, Newport Beach, CA, 5–9 March 2000, 3987-18, 110–20. 10. Shahinpoor, M. and Kim, K. J. (2001) Ionic Polymer–Metal composites: I. Fundamentals, Smart Ma. Struct., 10 819–33. 11. Sadeghpour, K., Salomon, R. and Neogi, S. (1992) Development of a novel electrochemically active membrane and smart material based vibration sensor/damper, Smart Mat. Struct., 1, 172–9. 12. Guo, S., Nakamura, T. and Fukuda, T. (1996) Design and characteristic evaluation of micropump using ICPF actuator, Proceedings of the 7th International Symposium on Micro Machine and Human Science, Nagoya, Japan, 2–4 October 1996, 235–40. 13. Shahinpoor, M., Bar-Cohen, Y., Simpson, J. O. and Smith, J. (1998) Ionic polymer–metal composites (IPMCs) as biomimetic sensors, actuators and artificial muscles-a Review, Smart Mat. Struct., 7, R15–30. 14. Dogruer, D., Tiwari, R. and Kim, K. J. (2007) Ionic polymer metal composites as energy harvesters, Proceedings of the 2007 SPIE International Symposium on Smart Materials and Structures, San Diego, CA, 19–22 March 2007, 6524, 1C1–10. 15. Bennett, M. D. and Leo, D. J. (2003) Manufacture and characterization of ionic polymer transducers employing non-precious metal electrodes, Smart Mat. Struct., 12, 424–36. 16. Shahinpoor, M. and Kim, K. J. (2000) The effect of surface-electrode resistance on the performance of ionic polymer–metal composite (IPMC) artificial muscles, Smart Mat. Struct., 9, 543–51. 17. Nemat-Nasser, S. (2002) Micromechanics of actuation of ionic polymer–metal composites, J. Appl. Phys, 92, 2899–915. 18. Kim, S. M. and Kim, K. J. (2008) Palladium buffer-layered high performance ionic polymer– metal composites sensors and actuators, Smart Mat. Struct., 143, 343–351. 19. Lee, D. Y., Park, I.-S., Lee, M.-H., et al. (2007) Ionic polymer–metal composite bending actuator loaded with multi-walled carbon nanotubes, Sensors and Actuators A, 133, 117–27. 20. Bennett, M. D. and Leo, D. J. (2004) Ionic liquids as stable solvents for ionic polymer transducers, Sensors and Actuators A, 115, 79–90. 21. Park, I.-S. and Kim, K. J. (2007) Multi-fields responsive ionic polymer–metal composite, Sensors and Actuators A, 135, 220–8. 22. Park, I.-S., Tiwari, R. and Kim, K. J. (2008) Sprayed Sensor Using IPMC paint, Adv. Sci. Techn., 61, 59–64. 23. Vohnout, S., Kim, S.-M., Park, I.-S., et al. (2007) IPMC assisted miniature disposable infusion pumps with embedded computer control, Proceedings of the 2007 SPIE International Sympoium on Smart Materials and Structures, San Diego, CA, 19–22 March 2007, 6524, 1U1-7. 24. From http://en.wikipedia.org/wiki/Infuion_pump, Infusion pump. 25. Laser, D. J. and Santiago, J. G. (2004) A review of micropumps, J. Micromech. Microeng., 14, R35-64. 26. Woias, P. (2005) Micropumps – past, progress and future prospects, Sensors and Actuators B, 105, 28–38. 27. Lee, S., Park, H. C. and Kim, K. J. (2005) Equivalent modelling for ionic polymer–metal composite actuators based on beam theories, Smart Mat. Struct., 14, 1363–8. 28. Lee, S., Park, H. C., Kim, K. J. and Yoon, K. J. (2004) Equivalent beam and equivalent bimorph beam models for ionic polymer–metal composite actuators, J. Control, Automation, and Syst. Eng., 10, 1012–6. 29. MSC. Software Corp., MSC/NASTRAN user’s manual (2001). 30. Taleghani, B. K. and Campbell, J. F. (1999) Non-linear finite element modelling of THUNDER piezoelectric actuators, NASA/ TM, 209322. 31. Lim, S. M., Lee, S., Park, H. C., et al. (2005) Design and demonstration of a biomimetic wing section using a lightweight piezo-composite actuator (LIPCA), Smart Mat. Struct., 14, 496–503.


191

32. Park, H. C., Kim, K. J. Lee, S. and Chah, Y. J. (2004) Electromechanical flapping produced by ionic polymer–metal composites, Proceedings of the 2004 SPIE International Symposium on Smart Materials and Structures, San Diego, CA, March 2004, 5385–63. 33. Nam, J. D., Lee, J. H., Lee, J. H., et al. (2005) Water uptake and migration effects of electroactive IPMC (Ionic Polymer–Metal Composite) actuator, Sensors and Actuators A, 118, 98–106. 34. Nemat-Nasser, S. and Li, J. Y. (2000) Electromechanical Response of Ionic Polymer–Metal Composites, J. App. Phys, 87, 3321–31.

Section III Conjugated Polymers


10 Conjugated Polymer Actuators: Fundamentals Geoffrey M. Spinks, Gursel Alici, Scott McGovern, Binbin Xi and Gordon G. Wallace ARC Centre of Excellence for Electromaterials Science and Intelligent Polymer Research Institute, University of Wollongong, Australia

10.1 Introduction Inherently conjugated (or conducting) polymers (ICPs) are one of the main categories of electroactive polymers (EAPs) and fall within the ‘ionic’ category of EAPs. ICPs such as those shown in Scheme 10.1 have been extensively studied for a wide range of applications that use both their inherent conductivity (e.g. sensors [1] or electrostatic discharge protection [2]) and their facile electroactivity: their ability to be electrochemically switched between different states at low voltage with very large changes in properties. The conductivity can change by 10 orders of magnitude; the polymer’s colour changes; the polymer can switch from hydrophilic to hydrophobic; permeability to chemical species changes; the volume changes as does the mechanical properties (e.g. Young’s modulus). These property changes are then useful in a wide variety of devices including electrochromic displays [3], controlled release systems [4], membranes [5] and, of course, actuators. ICPs have been developed for a wide range of actuator applications, as reviewed recently [6, 7]. Compared with other actuator materials, ICPs produce relatively large strains and low/medium stresses and operate at low voltages. Two broad categories of ICP actuators have been developed: linear actuators and benders. The latter are produced when the ICP is


196

Biomedical Applications of Electroactive Polymer Actuators R +

+ A– N H

H N

A– S

n

Scheme 10.1

A–

n

m

m

m

Polypyrrole (PPy)

H N

Polythiophene (PTh)

Polyaniline (PANi)

Some simple ICPs (A represents a dopant anion species).

laminated onto a passive, flexible substrate [8]. The actuator system then works like a bimetallic strip so that volume changes in the ICP induce bending. Linear actuators are formed from free-standing films or fibres where actuation in one direction is principally used to perform work. One alternative variant of the linear actuator is those systems that use the thickness direction dimension changes of ICP coatings to affect the mechanical response [9]. One practical example of the latter is a pump system that uses concentric cylindrical layers of ICPs to generate a pumping pressure [10]. Companies dedicated to the development of artificial muscles based on conjugated polymers have also emerged in recent years. MicroMuscle based in Sweden and EAMEX from Japan are both actively pursuing actuators for biomedical and electronics applications. Santa Fe Science and Technology, USA, has produced continuous spun polyaniline fibres and demonstrated their use as linear actuators [11, 12]. Academic laboratories have also developed several demonstration products including a variable camber hydrofoil [13], a robotic fish propulsor fin [14], a gas valve [15], ‘microrobots’ [16] and a micropump [10], some of which are illustrated in Figure 10.1. An electronic Braille screen using ICP actuators is also described in this book [17]. Actuation in ICPs occurs through their reversible redox chemistry. The electrochemical reaction of polypyrrole, one of the most widely studied conjugated polymers, is illustrated Nitrogen

Oxygen cell

Amplifier

3 cm

+ Polymer Actuator

– Air flow

(a)

(b)

(c)

Figure 10.1 Some example devices using conjugated polymer actuators: (a) miniature pump system using concentric polypyrrole tubes; (b) prototype pectoral fin for biorobotic fish (Reproduced from Bioinspir. Biomim, 2, S6-S17, The application of conducting polymers to a birobotic fin propulsor by James Tangorra et al. with permission. Copyright (2007) Institute of Physics Publishing Limited); (c) a ball valve operated by a bender type actuator (Reproduced with permission from Sensors and Actuators A: Physical, 114, 1, Andrews, M.K., Jansen, M.L., Spinks, G. and Wallace, G.G. An integrated electrochemical sensor-actuator system. Copyright (2004) Elsevier; Photograph courtesy Dr Murray Jansen).

Conjugated Polymer Actuators: Fundamentals

197

in Scheme 10.2. The redox switching leads to the addition or removal of charge from the polymer backbone and, thus, the transport of ions into or out of the polymer to balance the charge. These ion movements are principally responsible for the volume changes that produce an actuation response. The extent and speed of actuation in ICPs, therefore, must be influenced by the electrochemical properties of the polymer, as well as the size of the ions that move into and out of the polymer and the speed at which the ions move. These influences will be explored in more detail in subsequent sections. Useful actuators must do work on their surroundings: either moving an attached load and/or generating a force that operates on an attached mechanism. Thus, the actuator material itself is subject to mechanical stress, so the mechanical behaviour of the ICP will also be important in determining the actuation behaviour. The ICP breaking strength limits the maximum stress that can be applied or generated, while the Young’s modulus of the material determines the extent to which the ICP deforms. The mechanical aspects of ICP actuator devices will also be further explored below. Finally, actuators must produce the desired movement accurately and repeatably. Thus, there is a need to develop control systems that ensure that the correct input stimulus is applied to achieve the desired output. Various approaches to modelling the control of ICP actuation are also summarised in this chapter. H N

H N N H

H N N H

N H

X

reduced state

+2e–/–2A–

–2e–/+2A–

oxidised state H N

H N N H

H N N H

–

A

Scheme 10.2

N H

A–

X

The electrochemical reaction of polypyrrole.

10.2 Molecular Mechanisms of Actuation in ICPs The electrochemistry in ICPs occurs by the application of small voltages (typically Pt > Au > ITO. SEM photographs of the corresponding films had similar microstructures on the solution side of the electrode. However, on the electrode side vast differences were evident. GC was seen to have an even distribution of fine pores approximately 1–2 mm in diameter, gold had a varied distribution ranging from 1–15 mm, platinum had an even distribution of pores approximately 8 mm in diameter and ITO had many fewer pores ranging from 5–15 mm in diameter. The anions within solution have ionic radii and solvation spheres in the nanometre range, and sub-micron sized structures that are more likely to be related to direct expansion were not visible under this magnification and may not be directly commented on. However, it was evident from the results that the strain rate of each of the polymers followed that: Pt > Au > ITO > GC. It is expected that the high prevalence of large pores in the polymer grown on platinum or gold enables easier passage of electrolyte into the bulk of the polymer, thus improving the rate of diffusion of ions into the sub-micron sized free volume of the polymer and improving the actuation strain rate. An earlier study by Pandey et al. [53] showed that the polymerization electrode also influenced the balance of anion/cation movement in the polymer. In this case PPy doped with naphthalene sulfonic acid (NSA) was prepared on three different electrodes. Anion movement was favoured in those films that had a more open, porous structure. The actuation was performed in aqueous NaCl electrolyte, so that the mobile cation (Naþ) was smaller than the mobile anion (NSA). As well as the electrode material, there are other factors that may help control the morphology of the deposited polymer. Electropolymerization by different means (cyclic voltammetry (CV), constant applied potential and constant applied current) are factors that are also known to affect the morphology of the deposited polymer. In general, constant applied potential generates films with a cauliflower structure that have a large free volume in contrast to the compact flat films that may be generated with growth by CV or constant current. By lowering the temperature or viscosity of the solvent, more compact films may also be generated that have improved conductivities and smaller pore sizes.

10.6 Mechanical System Response Ultimately, the actuation behaviour is determined by the mechanical output as manifest by a movement and/or a force. In practical actuator devices the required mechanical output


213

will be variable – the force applied by the external system will vary and the required forces/ movements will change depending upon the desired outcome. The ICP’s mechanical properties (especially the Young’s modulus) will, therefore, have a bearing on the actuation response. It is now well established that the Young’s modulus of ICPs depends on the oxidation state [19, 54], so that the modulus will change during the actuation cycle. A simple model is described that links the length change that occurs with no external forces acting on the ICP to the displacement produced with an applied load [19, 55]. The importance of the Young’s modulus of the actuator material is demonstrated by a simple example. As illustrated in Figure 10.10a., a common device geometry has the actuator attached to a restoring spring. This geometry is especially useful when the actuator

Spring

Final State

f=0

f>0

l = l0 E=Y

Actuator

l = lf E = Y’

Actuator Displacement (mm)

Initial State

1.2 1 0.8

B

0.6

C

0.4 0.2 0 0.01

0.1

100

70 Work per cycle (kJ/m^3)

2.5 Force Generated (N)

1 10 Stiffness Ratio

(b)

(a)

2 B

1.5 1

A

0.5 0 0.01

A

C

0.1


(c)

100

B

60 50 40 30

A

20 C

10 0 0.01

0.1


100

(d)

Figure 10.10 (a) A contractile linear actuator attached in series with a restoring spring; (b) final contraction (actuator stroke) achieved when operated against springs of increasing stiffness (stiffness ratio is the ratio of the spring stiffness to the actuator materials stiffness) and different Young’s moduli: Y is the Young’s modulus in the expanded (initial) state and Y’ is the Young’s modulus in the contracted (final) state; (c) force generated; and (d) work per cycle. In (b), (c) and (d) the labels for the curves represent: A: Y ¼ Y’; B: Y ¼ 1/2 Y’; C: Y ¼ 2Y’.

214


is long and slender, as in a film or fibre. In such geometries the actuator does not ‘push’ well due to the ease of buckling. To overcome this problem, the actuator is configured to operate in contraction and the restoring spring ensures expansion in the reverse cycle. The actuator/spring model is also approximately correct for the bender actuator, since the passive layer must be bent as the actuator contracts/expands and the bending of the passive layer produces a restoring force. In situations where the external spring force (fe) depends linearly on its displacement with a spring constant (ke), it has been shown [55] that the final change in length of the actuator will be: 1 DLf ¼ DL0 ð10:8Þ 1 þ r0 where DL0 is the actuation displacement when no external force is applied (the ‘free stroke’) and r’ is the ratio of the external spring stiffness (ke) to the stiffness of the actuator material (ki’). The internal stiffness is of the actuator material in its final state and takes into account any changes in Young’s modulus (Y’) that occurs during actuation: r0 ¼

ke k0i

and

k0i ¼

Y 0A L0

ð10:9Þ

where A and L0 are the initial cross-sectional area and length of the actuator, respectively. As illustrated in Figure 10.10b, the amount of actuation achieved depends significantly on the stiffness of the external spring. Low stiffness ratios correspond to a soft external spring, which then applies only a small force to the actuator. The actuator then produces close to its maximum displacement (DLf » DL0 ), since little elastic stretching by the spring occurs. When a stiff external spring is used (high r’), however, there can be little or no actuation displacement (DLf » 0), since the high force applied by small displacements of the spring cause elastic stretching of the actuator that counter balance the actuation contraction. The force generated by the actuator (Figure 10.10c) shows the opposite trend: low forces when a soft spring is used and maximum forces when a stiff spring is used. The work performed by the actuator in deforming the spring is given in Figure 10.10d and reaches a maximum when the external and internal stiffness ratios are equal (r’ ¼ 1). The data presented in Figure 10.10 are for a hypothetical actuator that generates a contraction of 1 mm, has original dimensions of L0 ¼ 20 mm and A ¼ 0.2 mm2 and initial Young’s moduli of Y ¼ 100 MPa. Also shown in Figure 10.10 is the effect of changing modulus of the actuator material on the actuation performance. The stiffness ratio on the x-axis is calculated assuming no change in modulus during actuation (Y’ ¼ Y ). The actuation displacement, force generated and work per cycle have all been calculated for three conditions: Y’ ¼ Y; Y’ ¼ 0.5Y and Y’ ¼ 2Y. It can be seen that for a given external spring stiffness, the actuator displacement decreases when the modulus decreases during the contractile actuation. The lower actuator modulus means that the material will stretch to a greater extent due to the external force applied by the spring. The net contraction is, therefore, smaller. Similarly, the force generated is smaller as a result of the smaller net contraction. In contrast, the actuator displacement, force generated and work per cycle all increase when the modulus increases during contractile actuation.


215

The calculated results given in Figure 10.8 all assume that the free stroke of the polymer does not change in each circumstance considered. Although experimental data does support the findings illustrated in Figure 8, it has generally been found that situations that change the modulus of the actuator also change the free stroke. For example, it is shown from Equation (10.8) that the actuation displacement would increase if the modulus in the contracted state was increased (and the free stroke unchanged). Addition of carbon nanotubes to polyaniline fibres certainly increased the modulus [56], however, the presence of the nanotubes also restricted the actuation of the PANi so that the free stroke also decreased. The result shown in Figure 10.11 was that the reinforced PANi gave higher actuations when higher external forces were applied, but the unreinforced PANi gave the larger actuation at smaller external forces [56]. These data were collected under isotonic conditions (rather than using a restoring spring), so that the actuation displacement is now given by Equation (10.10): f L0 1 1 DLf ¼ DL0 þ ð10:10Þ A Y0 Y where f is the applied isotonic force. The graph shown in Figure 10.11 approximates the linear behaviour predicted in Equation (10.10), with the intercept corresponding to the free stroke and the slope determined by the modulus shift during actuation. 0.2 0

Strain [%]

–0.2 –0.4 –0.6 –0.8 –1 –1.2 –1.4

0

25

50 75 Stress [MPa]

100

125

Figure 10.11 Actuation strain obtained under isotonic conditions for PANi fibre (square symbols) and carbon-nanotube reinforced PANi fibre (circles). A negative strain indicates contraction.

A number of studies on different ICPs have shown similar data to that given in Figure 10.11, with a decreasing actuation occurring when higher isotonic stresses are applied [19, 24, 57]. These observations reflect the fact that, in most cases, the modulus is smaller in the contracted state. The modulus of solvent swollen network polymers is known to be influenced by two factors. Firstly, the swelling by solvent reduces the concentration of load-bearing chains, so that the modulus tends to decrease. Secondly, as

216


chains become highly extended as a result of solvent swelling, they effectively become stiffer, which leads to an increase in modulus. This latter effect appears to dominate in most ICPs. The dependence of modulus on swelling is complex, however [58], and not linearly dependent upon the degree of swelling. It has been shown, for example, that there is almost no modulus change when a PPy material was subject to a small voltage range [19]. The same material, however, showed very large changes in modulus when a wide potential range was used to cause full oxidation and reduction of the polymer [19]. In this case the modulus changed by a factor of four, causing a very sharp drop in actuator strain as the applied stress increased. There has been at least one report of an increase in actuator strain with increasing stress. This situation arises when the modulus increases in the contracted state and occurred in a PTh film tested in an ionic liquid electrolyte [59]. Curiously, the same polymer showed the more typical ‘low modulus in the contracted state’ when tested in a conventional electrolyte, highlighting the complexity of the modulus shift phenomenon. Other mechanical effects also influence the actuation behaviour of ICPs. An unresolved problem at present is the creep or drift that is commonly observed. An example is shown in Figure 10.12, with cyclic potentials producing a larger expansion per cycle than the corresponding contraction [60]. The result is a net increasing in length of the sample over time. The rate of this drift is increased when higher loads are applied to the actuator, implying that the drift is related to a viscous or viscoelastic deformation occurring in the

20 Isotonic load (MPa) 34

Strain (%) curves offset for clarity

16

28 23

12

17 11

8

8.5 5.7

4

a c b

3.4 1.1

0

1.3% strain 0

100

200 300 Time (s)

400

500

Figure 10.12 Example data showing drift in actuation strain over time for polyaniline actuated at different applied isotonic stresses (Reproduced with permission from Synthetic Materials, Polyalinine actuators: Part 1. PANI(AMPS) in HCI by Smela, E., Lu, W. and Mattes, B.R., 151, 1, 25–42. Copyright (2005) Elsevier).


217

polymer. In the case of polyaniline, a significant decrease in the rate of drift has been achieved by adding carbon nanotube reinforcement [56]. However, as described above these reinforcing agents also reduce the amount of actuation occurring. More research is needed into the origin of the drift and the molecular processes occurring, before the problem can be rectified.

10.7 Device Design and Optimization The size and shape of the polymer actuators are as important as their electrochemomechanical properties when it comes to providing enough actuation power for practical applications. The actuation power depends on (i) force output, (ii) displacement output and (iii) speed of response. In this section, a practical approach is presented to link device design requirements to the performance parameters and to offer guidelines for the device design and optimization based on electroactive PPy bending-type actuators. The actuator considered is a one-end fixed and the other end-free bender. 10.7.1

How to Tailor Actuator Performance to Meet Design Requirements

The force and displacement outputs are created due to internal bending moment induced during the conversion of electrochemical energy into mechanical energy. The force (blocking force), F, can be estimated using a quantitative relationship based on the strain created as a function of the input voltage [61]: F¼

E1 b h1 ðh1 þ h2 Þ L

ð10:11Þ

where is the Young’s modulus of the PPy layers. For other variables in Equation (10.11), see Figure 10.13. The strain a in the PPy layers is a function of the strain to charge ratio and charge density in the PPy layers [62, 61]. As described in Equation (10.11), the force output h1

h2

Upper PPy layer

dx h1

y

Neutral axis Lower PPy layer dθ

R

F Figure 10.13 One end cantilevered actuator and the model parameters (Reproduced with permission from Bioinspir. Biomim., Establishment of a biomimetic device based on tri-layer polymer actuators-propulsion fins by Alici, G., et al., 2, S18–30. Copyright (2007) IOP).

218


Input Power (W)

Max Force (mN)

14 12 10 8 6 4 2 0 9

6

5

4

3

0.016 0.014 0.012 0.01 0.008 0.006 0.004 0.002 0

2

9

Actuator Width (mm)

6

5

4

3

2

Actuator Width (mm)

Figure 10.14 Variation of the maximum blocking force and the input power with the actuator width. The actuators used were 15 mm long with 30 m PPy thicknesses under 1 V. The electropolymerization of polypyrrole is achieved by submerging the sputter coated PVDF film in a solution of 0.1 M pyrrole, 0.1 M LiTFSI in Propylene Carbonate (PC) with 0.5 w/w% water (Reproduced from Electroactive Polymer Actuators and Devices (EAPAD) 2007, Proceedings of SPIE Vol. 6524, Tri-layer conducting polymer actuators with variable dimensions by Minato, R., Alici, G., McGovern, S. and Spinks, G., 6524, 6524J. Copyright (2007) SPIE).

is proportional to the width, and thickness, and inversely proportional to the actuator length. This follows that appropriately sized actuators can satisfy the force design requirement. For the force results presented in Figure 10.14, the force output increases with the actuator width. For the widths >4 mm, the actuator curls into a semi-cylindrical shape. This curling action decreases the bending displacement. However, the time-averaged electric power input increases with the actuator width [63]. The experimental bending moment (F L) of the PPy actuators as a function of PPy thickness is provided in Figure 10.15, where there is a non-negligible drop in the bending

Bending Moment (mN.cm)

1.4 1.2 Experimental

1

Theoretical 0.8 0.6 0.4 0.2 0 0

30

60 90 Thickness (µm)

120

Figure 10.15 Experimental and theoretical results showing the variation in the bending moment with the thickness for an actuator with the dimensions of 20 1 0.17 mm and for the salt TBA.PF6 0.25 M in the solvent propylene carbonate.


219

Bending Displacement (mm)

moment output when each of the PPy layers is thicker than 60 mm [64]. When the PPy layer is thicker than 60 mm, the bending moment begins to decrease. In this case, the PPy actuator is unable to overcome its increased flexural rigidity. For instance, an increase in the PPy thickness from 30 to 50 mm gives a 25 % increase to the total thickness of the strip, that is a 66 % increase in the volume of the PPy, but the flexural rigidity increases by 95 %. The corresponding bending displacement shows a similar trend. 8 7 6 5 4 3 2 1 0

Vertical Horizontal

0

0.2

0.4 0.6 Voltage (V)

0.8

1.0

Figure 10.16 The variation of the displacements with constant input voltages for actuator with the dimensions of 10 1 0.17 mm for the electrolyte TBA.PF6 0.05 M in the solvent propylene carbonate.

As depicted in Figure 10.16, the bending displacement or the tip displacement of the actuators is proportional to the input voltage [61], and the following differential equation describes the horizontal ‘x’ and vertical ‘v’ displacements of the actuators: " 2 # d2 v dv 3 V C ðh2 þ h1 Þ " 1þ ð10:12Þ 3 # ¼ 0 2 dx dx 2h1 þ h2 3 h2 þ ð E2 E 1 Þ 2 b L E1 2 2 where V is the input voltage, L is the length of the actuator, b is the width of the actuator, C is the capacitance, E1 and E2 are Young moduli of PPy and PVDF layers, respectively, h1 and h2 are described in Figure 10.13, and is the experimentally determined proportionality constant relating the internal stress s to the exchanged charge density. The experimental and theoretical results provided here suggest that, depending on the force, displacement and speed requirements of a practical device, the geometric parameters and shape of the actuator can be optimized suitably to satisfy the device requirements. As a case study, a swimming device propelled with the bending type actuators is presented next to demonstrate the influence of the actuator geometry on a functional system. 10.7.2

Design of a Swimming Device

As reported before [65, 66, 67, 68, 69], when designing a swimming device, there are two paramount factors to consider: (i) the shape, mechanical properties and the locations of propulsors/fins on the device; and (ii) their movement pattern. Fish is a good example to

220


follow in considering these factors appropriately. The propulsion efficiency or the swimming efficiency is governed by, the Froude number: ¼

TV P

ð10:13Þ

where T, V and P are the thrust generated, the velocity of the swimming body and power input. T and V are determined by the three design parameters outlined in the opening paragraph of this section. The thrust generated by a swimming device depends on four factors [67]: (i) the aspect ratio (the higher is the aspect ratio, the higher is the net thrust force to accelerate the device); (ii) the shape of the fins; (iii) the fin stiffness (the higher is the stiffness, higher is the thrust); and (iv) the oscillatory motion of the fins (the higher are the frequency and amplitude of the fin oscillations, the higher is the thrust). With reference to the performance results of the polymer actuators presented above, all these factors can be controlled except the fin stiffness, which highly depends on mechano-electro-chemical properties of the actuators. The topology of the proposed device is depicted in Figure 10.17 where eight fins are installed along both sides of a rigid body to move the device in a direction perpendicular to the longitudinal axis of the body. The rigid body is made of prepregnated carbon fibre strips of 0.3 mm thickness and hardened with resin. The device can be considered like a box fish having a carapace (rigid body) with side or paired fins running through the rigid body, like a fish having pectoral fins. The fins or polymer benders can be considered as individually controlled flexible membranes. Each fin is activated with sinusoidal inputs such that there is a phase lag between the movements of the successive fins, and the frequency and amplitude of the input can be changed to create enough thrust for propulsion. This generates an undulatory movement. This is in agreement with the finding in the literature [69] that the undulatory movement is superior over an oscillatory movement, from the propulsion efficiency point of view.

Paired fins

chord span

Direction of movement

Figure 10.17 Topology of the proposed biomimetic device (not to scale) (Reproduced with permission from Bioinspir. Biomim., Establishment of a biomimetic device based on tri-layer polymer actuators-propulsion fins by Alici, G., et al., 2, S18–30. Copyright (2007) IOP).


221

With reference to Figure 10.17, the aspect ratio, which is the ratio of the fin span to the fin chord, characterizes the movement of swimming bodies. A low aspect ratio fin generates a low velocity of movement, but results in great propulsion efficiency and manoeuvrability. As far as the force output of polymer actuators is concerned, low aspect ratio fins are not favourable. The biomimetic devices that are currently being worked on have high aspect ratios, which create less drag force. This follows that the devices with caudal fins are preferable. Further, future designs will also include fins at both ends of the body – carapace, like caudal fins. As reported in the biomimetic literature, such a device has better dynamic stability and endurance [67]. Depending on the direction of motion, either of both tail fins can be activated for cruising. 10.7.3

Device Testing

The assembled rigid frame and the fins are shown in Figure 10.18. Undulatory movements were employed to create enough thrust for propulsion [70]. This was achieved by activating each fin with a sinusoidal input such that there was a phase lag between the movement/operation of the successive fins. The best propulsion was observed at 1.5 Hz, 2 V peak to peak, with phase delays of 90o. A test was conducted when the first fin of one side of the prototype and the last fin of the other side were connected to the same power source, moving simultaneously. Similarly, the second fin of one side was connected to the second last fin from the other side, and so on. It was observed that each fin created sufficient undulatory movement to cause the prototype to rotate approximately 5°, though the platinum wires restricted the propulsion motion of undulating fins. This follows that the placement and method of attachment of the polymer actuator to the mechanical device determine the direction of movement and defines the actuator requirements.

Carbon Pre-preg Rigid Body Platinum Wire

Polymer Actuator Fins

Figure 10.18 The prototype swimming device with polymer actuators as the propulsion elements – fins (Reproduced with permission from Bioinspir. Biomim., Establishment of a biomimetic device based on tri-layer polymer actuators-propulsion fins by Alici, G., et al., 2, S18–30. Copyright (2007) IOP).

222


The next test was conducted by connecting the first fin of one side to the first fin of the other side of the prototype. All the other fins were connected in the same manner. The propulsions of the fins were large enough to produce a side-to-side motion along the longitudinal axis of the fins. Although this experiment was successful in a small container, it was not possible to observe the full distance the prototype could move under the propulsion force produced by the fins. More importantly, this test demonstrated that the fins made of bending type polymer actuators could generate enough thrust to move the device in an aquatic medium. In summary, the performance quantification of bending polymer actuators have been presented in terms of the force and displacement outputs, and a methodology has subsequently been proposed to design a swimming device propelled by the polymer actuators, which act as the propulsion elements of the device. All performance results presented and the successful testing of the device demonstrate the importance of understanding the application requirements and then tailoring the geometry and shape of the actuators, and the placement and attachment of the actuators for a successful attempt to widen the application areas of ICPs.

10.8 Future Prospects Since the first bending-type actuators from ICPs were demonstrated in the early 1990s, there have been considerable advances in actuator performance, applications and in the understanding of actuation mechanisms. In terms of performance improvements, most attention has been given to free-standing films. Around the turn of the century, the best ICP actuation performances (reported by several different groups) were around 5 % maximum strain, 1 %/s maximum strain rate and 5 MPa maximum sustained stress (isotonic). Since 2000, significant improvements have been reported in each of these areas:

maximum actuation strain: 12 % in 2003 [71]; 26 % in 2004 [72]; 40 % in 2005 [32] maximum actuation strain rate: 4 % in 2000 [44]; 15 % in 2003 [17] maximum sustained stress: 34 MPa in 2002 [73], 100 MPa in 2006 [56] While these improvements are impressive, the maximum performances in each area have not been achieved simultaneously. The highest stress actuators, for example, produce an actuation strain of only 2 % [56]. It is particularly useful to have actuators that give simultaneously high stroke, fast response and can operate against high stresses. A web based resource for tracking the published actuator performances of ICPs (and other actuator materials) has been developed by the Molecular Mechatronics Group at the University of British Columbia [74]. While the focus of much of the research to date has been in understanding mechanisms and improving the basic performance, several other areas require further work. Improvements can be achieved in the following areas:

Efficiency: the energy conversion efficiency (electrical to mechanical) of ICP actuators is poor at

Biomedical applications of electroactive polymer actuators

Biomedical Applications Polymer Blends

Frontiers in biomedical polymer applications, Volume 1

Frontiers in Biomedical Polymer Applications, Volume II

Biomedical Applications Polymer Blends (Advances in Polymer Science)