Polymer Electrolytes: Fundamentals and Applications

Polymer electrolytes © Woodhead Publishing Limited, 2010 Related titles: Materials for energy conversion devices (IS...

Author: Cesar Sequeira | Diogo Santos

382 downloads 3422 Views 9MB 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

Polymer electrolytes

© Woodhead Publishing Limited, 2010

Related titles: Materials for energy conversion devices (ISBN 978-1-85573-932-1) The term electroceramic is used to describe ceramic materials that have been specially formulated with specific electrical, magnetic or optical properties. Electroceramics are of increasing importance in many key technologies including: microelectronics, communications and energy conversion. This book is the first comprehensive survey on major new developments in electroceramics for energy conversion devices. It presents current research from leading innovators in the field. Materials for fuel cells (ISBN 978-1-84569-330-5) This authoritative reference work provides a comprehensive review of the materials used in hydrogen fuel cells, which are predicted to emerge as an important alternative energy option in transportation and domestic use over the next few years. The design and selection of the materials is critical to the correct and long-term functioning of fuel cells and must be tailored to the type of fuel cell. The book looks in detail at each type of fuel cell and the specific material requirements and challenges. Chapters cover material basics, modelling, performance and recyclability. Solid oxide fuel cell technology: Principles, performance and operations (ISBN 978-1-84569-628-3) High-temperature solid oxide fuel cell (SOFC) technology is a promising power generation option, which features high electrical efficiency and low emissions of environmentally polluting gases such as CO2, NOx and SOx. It is ideal for distributed stationary power generation applications where both high efficiency electricity and high-quality heat are in strong demand. This book presents a systematic and indepth narrative of the technology from the perspective of fundamentals, providing comprehensive theoretical analysis and innovative characterisation techniques for SOFC technology. The book covers the development of SOFC technology, from cell materials and fabrication, to performance analysis and stability and durability issues. Details of these and other Woodhead Publishing materials books can be obtained by: • •

visiting our web site at www.woodheadpublishing.com contacting Customer Services (e-mail: [email protected]; fax: +44 (0) 1223 893694; tel.: +44 (0) 1223 891358 ext. 130; address: Woodhead Publishing Limited, Abington Hall, Granta Park, Great Abington, Cambridge CB21 6AH, UK)

If you would like to receive information on forthcoming titles, please send your address details to: Francis Dodds (address, tel. and fax as above; e-mail: francis. [email protected]). Please confirm which subject areas you are interested in.


Polymer electrolytes Fundamentals and applications Edited by César Sequeira and Diogo Santos

Oxford

Cambridge

Philadelphia

New Delhi


Published by Woodhead Publishing Limited, Abington Hall, Granta Park, Great Abington, Cambridge CB21 6AH, UK www.woodheadpublishing.com Woodhead Publishing, 525 South 4th Street #241, Philadelphia, PA 19147, USA Woodhead Publishing India Private Limited, G-2, Vardaan House, 7/28 Ansari Road, Daryaganj, New Delhi – 110002, India www.woodheadpublishingindia.com First published 2010, Woodhead Publishing Limited © Woodhead Publishing Limited, 2010 The authors have asserted their moral rights. This book contains information obtained from authentic and highly regarded sources. Reprinted material is quoted with permission, and sources are indicated. Reasonable efforts have been made to publish reliable data and information, but the authors and the publisher cannot assume responsibility for the validity of all materials. Neither the authors nor the publisher, nor anyone else associated with this publication, shall be liable for any loss, damage or liability directly or indirectly caused or alleged to be caused by this book. Neither this book nor any part may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, microfilming and recording, or by any information storage or retrieval system, without permission in writing from Woodhead Publishing Limited. The consent of Woodhead Publishing Limited does not extend to copying for general distribution, for promotion, for creating new works, or for resale. Specific permission must be obtained in writing from Woodhead Publishing Limited for such copying. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation, without intent to infringe. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library. ISBN 978-1-84569-772-3 (print) ISBN 978-1-84569-977-2 (online) The publisher’s policy is to use permanent paper from mills that operate a sustainable forestry policy, and which has been manufactured from pulp which is processed using acid-free and elemental chlorine-free practices. Furthermore, the publisher ensures that the text paper and cover board used have met acceptable environmental accreditation standards. Typeset by Toppan Best-set Premedia Limited, Hong Kong Printed by TJI Digital, Padstow, Cornwall, UK


Contents

Contributor contact details Preface

xi xv

Part I Types and development of polymer electrolytes

1

1

3

1.1 1.2 1.3 1.4 1.5 1.6 1.7 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 2.10

Introduction to polymer electrolyte materials C. A. C. Sequeira and D. M. F. Santos, Technical University of Lisbon, Portugal Introduction Categories of polymer electrolytes Structure and its implications Conductivity measurements Applications in practical devices Conclusions References Ceramic polymer electrolytes J. S. Syzdek, Warsaw University of Technology, Poland Introduction Experimental approaches First composites – conductive fillers Development of insulating fillers Impact of the filler surface on the transport properties Interfacial concerns Other types of ceramic–polymer systems Conclusions Acknowledgements References

3 5 17 20 39 51 52 62 62 67 70 72 75 77 80 82 83 84 v


vi

Contents

3

Polymer electrolytes based on natural polymers A. Pawlicka and J. P. Donoso, Universidade de São Paulo, Brazil Introduction Grafted natural polymer-based solid polymer electrolytes Plasticized natural polymer-based solid polymer electrolytes Other natural polymer-based systems Magnetic resonance spectroscopy of polymer electrolytes obtained from natural polymers Conclusions and future trends References

3.1 3.2 3.3 3.4 3.5 3.6 3.7 4

4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9 4.10 4.11 4.12 4.13 5

5.1 5.2 5.3 5.4

Composite polymer electrolytes for electrochemical devices F. Alloin and C. Iojoiu, LEPMI CNRS, Grenoble Institute of Technology, France Introduction Composite electrolytes for lithium batteries: introduction Solid polymer electrolytes Composite polymer electrolytes based on poly(ethylene oxide) and clays Composite polymer electrolytes based on poly(ethylene oxide) and non-ionic fillers Gel polymer electrolytes Composite electrolytes for proton exchange membrane fuel cells Composite polymer electrolytes based on metal oxides Hygroscopic solid inorganic proton conductor composite polymer electrolytes Self-humidifying composite electrolytes Future trends Sources of further information and advice References Lithium-doped hybrid polymer electrolytes V. de Zea Bermudez, University of Trás-os-Montes e Alto Douro, Portugal, and M. M. Silva, University of Minho, Portugal Introduction Ionic conductivity Thermal properties Electrochemical stability


95

95 97 103 110 112 123 124

129

129 130 131 132 137 148 151 156 160 164 166 167 168 176

176 178 182 184

Contents

vii

5.5 5.6 5.7 5.8

Spectroscopic studies Electrochromic displays Conclusion References

185 207 209 210

6

Hybrid inorganic–organic polymer electrolytes V. di Noto, E. Negro and S. Lavina, University of Padova, Italy, and M. Vittadello, City University of New York, USA Introduction Fundamentals of polymer electrolytes Overview of hybrid inorganic–organic polymer electrolytes Methods The real component of the conductivity spectra in the framework of the jump relaxation model and polymer segmental motion Conclusions Acknowledgements References

219

6.1 6.2 6.3 6.4 6.5

6.6 6.7 6.8 7

7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8 7.9 7.10 7.11 7.12 7.13 8

8.1

Using nuclear magnetic resonance spectroscopy in polymer electrolyte research S. Abbrent, University of South Bohemia, Czech Republic, and S. Greenbaum, Hunter College of the City University of New York, USA Introduction Nuclei possibility Liquid state nuclear magnetic resonance Solid state nuclear magnetic resonance Relaxation processes Diffusion measurements Magic angle spinning Double resonance experiments Two-dimensional methods Exchange nuclear magnetic resonance Electrophoretic nuclear magnetic resonance Conclusions References Molecular dynamics simulations of Li ion and H-conduction in polymer electrolytes D. Brandell, Uppsala University, Sweden Introduction


219 220 221 254

266 272 273 273

278

278 279 281 282 291 295 299 300 304 305 307 308 309

314 314

viii

Contents

8.2 8.3 8.4

Computational chemistry The molecular dynamics methodology Li+-conducting poly(ethylene oxide)-based electrolytes for batteries Polymer electrolytes for fuel cells: perfluorosulphonic acid systems Conclusions and future trends References

8.5 8.6 8.7 9

9.1 9.2 9.3 9.4

9.5 9.6 9.7 9.8

Characterisation and modelling of multivalent polymer electrolytes M. J. C. Plancha, Laboratório Nacional de Energia e Geologia, Portugal Introduction Polymer–complex formation Ionic transport properties Morphological and crystallographic structures: characteristics and influence on ionic transport properties Ionic association: influence on ionic transport properties Phase diagrams: crystallinity and conductivity Conclusions References

Part II Applications 10

10.1 10.2 10.3 10.4 10.5 10.6 10.7 10.8 10.9

Polymer electrolytes for dye-sensitized solar cells J. N. de Freitas, J. E. Benedetti, F. S. Freitas, A. F. Nogueira and M. A. de Paoli, University of Campinas – UNICAMP, Brazil Introduction Polymer electrolytes Plasticized and gel polymer electrolytes Additives in the polymer electrolytes Stability of polymer electrolyte-based dye-sensitized solar cells Up-scaling: towards commercialization of polymer electrolyte-based dye-sensitized solar cells Conclusions and future trends Acknowledgements References © Woodhead Publishing Limited, 2010

315 316 325 329 336 337

340

340 342 344

358 365 367 373 373

379

381

381 384 390 402 417 421 423 424 424

Contents 11

11.1 11.2 11.3 11.4 11.5 11.6 11.7 11.8 11.9 11.10 11.11 11.12 12

12.1 12.2 12.3 12.4 12.5 12.6 12.7 12.8 13

13.1 13.2 13.3 13.4 13.5 14

14.1

ix

Solid polymer electrolytes for supercapacitors A. B. Samui and P. Sivaraman, Naval Materials Research Laboratory, India Introduction Solid electrolytes Conduction in solid electrolytes Solid electrolytes in supercapacitors Conducting polymer electrodes Activated carbon electrodes Cation exchange membrane-based supercapacitors Current research activities Applications Conclusions List of abbreviations References

431

Polymer electrolytes for electrochromic devices X. Fu, College of Chemistry and Chemical Engineering Southwest University, P.R. China Introduction Electrochromic effect and electrochromic devices Electrolytes for electrochromic devices Polymer matrix Classification of polymer electrolytes Proton-conducting polymer electrolytes and alkaline polymer electrolytes New type of polymer electrolyte References

471

Hyperbranched polymer electrolytes for high temperature fuel cells T. Itoh, Mie University, Japan Introduction Hyperbranched polymer electrolytes with a sulfonic acid group at the periphery Hyperbranched polymer electrolyte with a phosphonic acid group at the periphery Conclusions References Polymer electrolytes as solid solvents and their applications L. Ye and Z. Feng, Beijing Institute of Technology, China Introduction © Woodhead Publishing Limited, 2010

431 432 433 439 442 447 451 456 463 463 464 466

471 472 474 475 477 500 506 513

524 524 525 537 547 548

550 550

x

Contents

14.2 14.3

Structure of lithium ion battery Advantages of polymer electrolytes in lithium ion batteries Main properties of polymer electrolytes Solid polymer electrolytes applied in lithium ion batteries Gel polymer electrolytes in lithium ion batteries Composite polymer electrolytes in lithium ion batteries Polymer electrolytes in other battery types Conclusions Acknowledgments References

14.4 14.5 14.6 14.7 14.8 14.9 14.10 14.11 15

15.1 15.2 15.3 15.4 15.5 15.6 15.7

Hybrid polymer electrolytes for electrochemical devices F. L. de Souza, Federal University of ABC, Brazil, and E. R. Leite, Federal University of São Carlos, Brazil Introduction Physicochemical properties of hybrid polyelectrolytes General discussion Applications Conclusions Acknowledgments References

Index

552 552 554 557 566 570 572 577 578 578

583

583 587 595 596 598 599 599

603


Contributor contact details

(* = main contact)

Chapter 2

Editors and Chapter 1 Professor César A. C. Sequeira Department of Chemical and Biological Engineering Instituto Superior Técnico Technical University of Lisbon Av. Rovisco Pais 1049-001 Lisboa Portugal E-mail: [email protected] Dr Diogo M. F. Santos Department of Chemical and Biological Engineering Instituto Superior Técnico Technical University of Lisbon Av. Rovisco Pais 1049-001 Lisboa Portugal E-mail: [email protected]

J. S. Syzdek Warsaw University of Technology Faculty of Chemistry Inorganic Chemistry and Solid State Technology Division Polymer Ionics Research Group ul. Noakowskiego 3 00664 Warszawa Poland E-mail: [email protected] and Université de Picardie Jules Verne Laboratoire de Réactivité et de Chimie des Solides 33 Rue St Leu 80039 Amiens Cedex France

xi © Woodhead Publishing Limited, 2010

xii


Chapter 3 A. Pawlicka* Instituto de Química de São Carlos Universidade de São Paulo PO Box 780 13560-970 São Carlos SP, Brazil E-mail: [email protected] J. P. Donoso Instituto de Física de São Carlos Universidade de São Paulo PO Box 369 13560-970 São Carlos SP, Brazil E-mail: [email protected]

M. M. Silva Department of Chemistry/Center of Chemistry University of Minho 4710-057 Braga Portugal E-mail: [email protected]

Chapter 6 V. Di Noto,* E. Negro and S. Lavina Department of Chemical Sciences University of Padova Via Marzolo 1 I-35131 Padova (PD) Italy E-mail: [email protected]

Chapter 4 F. Alloin* and C. Iojoiu Laboratoire d’Electrochimie et de Physico-chimie des Matériaux et des Interfaces (LEPMI) CNRS-Grenoble-INP-UJF Domaine Universitaire BP 75 F38402 St Martin d’Hères Cedex France E-mail: [email protected]; [email protected]

Chapter 5

V. Di Noto CNR-ISTM Via Marzolo 1 I-35131 Padova (PD) Italy M. Vittadello PECS Department City University of New York/ Medgar Evers College 1650 Bedford Avenue Brooklyn, NY 11225 USA

V. de Zea Bermudez* Department of Chemistry and CQ-VR University of Trás-os-Montes e Alto Douro 5000-911 Vila Real Portugal E-mail: [email protected]



xiii

Chapter 7

Chapter 10

S. Abbrent* University of South Bohemia Branisˇovska 31 370 05 Cˇ eske Budeˇjovice Czech Republic E-mail: [email protected]

J. N. de Freitas, J. E. Benedetti, F. S. Freitas, A. F. Nogueira* and M. A. De Paoli* Institute of Chemistry University of Campinas – UNICAMP PO Box 6154, 13083-970 Campinas SP Brazil E-mail: [email protected]; [email protected]

S. Greenbaum Hunter College of the City University of New York 695 Park Avenue New York NY 10065 USA E-mail: steve.greenbaum@hunter. cuny.edu

Chapter 8 D. Brandell Department of Materials Chemistry Uppsala University PO Box 538 SE-751 21 Uppsala Sweden E-mail: daniel.brandell@mkem. uu.se

Chapter 9 M. J. C. Plancha Unidade de Pilhas de Combustível e de Hidrogénio Laboratório Nacional de Energia e Geologia, I.P. Estrada do Paço do Lumiar, 22 1649-038 Lisboa Portugal E-mail: [email protected]

Chapter 11 A. B. Samui* and P. Sivaraman Polymer Division Naval Materials Research Laboratory Shil Badlapur Road, Ambernath (E) Thane, Maharashtra-421506 India E-mail: [email protected]

Chapter 12 X. Fu College of Chemistry and Chemical Engineering Research Institute of Applied Chemistry Southwest University The Key Laboratory of Applied Chemistry of Chongqing Municipality Chongqing 400715 P.R. China E-mail: [email protected]


xiv


Chapter 13

Chapter 15

T. Itoh Division of Chemistry for Materials Graduate School of Engineering Mie University 1577 Kurima Machiya-cho Tsu City Mie 514-8507 Japan E-mail: [email protected]

F. L. de Souza* Grupo de Materiais e Métodos Avançados – GMAv Centro de Ciências Naturais e Humanas (CCNH) Federal University of ABC Santa Adélia, 166, 09210-170 – Santo André São Paulo Brazil E-mail: [email protected]

Chapter 14 L. Ye and Z. Feng* School of Materials Science and Engineering Beijing Institute of Technology No.5 Zhongguancun South Road, Haidian District Beijing 100081 China E-mail: [email protected]

E. R. Leite Department of Chemistry Federal University of São Carlos, Rod. Washington Luiz, Km 235, 13565-905 São Carlos São Paulo Brazil


Preface

Polymers represent a class of materials currently in use in a broad variety of applications. The first meaningful application occurred in 1935 when Wallace E. Carothers invented nylon, which quickly found application in stockings, fabrics, ropes, combs, tyres and many household items. Along with metals, semiconductors, and ceramics, polymers serve as one of the cornerstones of materials science and engineering. A recent study by the National Research Council (UK) outlined some of the primary applications of polymers. Already, plastics such as styrofoam (polystyrene) are actively competing with paper in the packaging field. High strength plastics, such as polycarbonates and acrylonitrile–butadiene– styrene (ABS), serve as alternatives for house walls and roofing materials. Kitchen surfaces employ such plastics as polyethylene terephthalate (PET). Polymers are used in dentistry and in a variety of implantable drug-delivery mechanisms. In electronics, the high conductivity of certain classes of polymeric materials makes them attractive alternatives to conventional metallic wiring. They are used as photoresists in microlithography processes and are utilised as a packaging material for microelectronic devices. They have been used in optoelectronic devices as non-linear optical elements and as lightemitting diodes. They serve as sensors for chemical and biochemical agents. Polymers may act as hosts for ions in much the same way that liquid solvents do. Suppose that a salt is introduced into a polymer: when the solvation energy of the ions in the polymer host exceeds the ionisation energy of the salt, the individual ions separate and attach themselves to Lewis base sites on the polymer. Conductivity is possible above the glass transition temperature, where the polymer molecules are free to move. The free volume also provides room for the ions to move. The ionic conductivity possesses a diffuse liquid-like behaviour in addition to the hopping-type behaviour characteristic of an electron in a disordered solid. One way of thinking about how the conductivity is brought about relies on the dynamic-percolation model in which one envisages sites on a lattice interconnected by a set of randomly placed conductors, whose distribution can be changed continuously. xv © Woodhead Publishing Limited, 2010

xvi

Preface

One of the goals of this book is to review the types, structure, geometry, topology, conductivity and other fundamental aspects that are unique to these ionic conductors, which were first introduced by P.V. Wright in 1975 and proposed as material of interest for the development of electrochemical devices by M. Armand in 1978. By using polymer electrolytes it is possible to design solid state batteries and related devices. The second goal of this book discusses a set of current applications of polymer electrolytes, namely high energy density batteries, fuel cells, supercapacitors, electrochromic devices, etc. An attempt is made to relate the applications to the underlying science described in the first part of the book. The work is the result of the joint efforts of many colleagues who contributed to it or helped with the editing process. First of all we would like to express our sincere thanks to all of the authors who supplied their high quality contributions within the short time frame available. This turned out to be a challenge for authors, editors and publishers. Furthermore, thanks are due to our families from whom we have stolen a number of weekends. César Sequeira Diogo Santos Lisbon


1 Introduction to polymer electrolyte materials C. A. C. S E Q U E I R A and D. M. F. S A N T O S, Technical University of Lisbon, Portugal

Abstract: Polymer electrolytes are electrolytic materials that offer many advantages in the area of large, high energy density batteries for electric propulsion and in fuel cells for electric vehicle or stationary applications. At the other end of the spectrum, polymer electrolytes are of interest for small, portable electronic devices where the battery represents a significant proportion of the device’s size and weight. This chapter focuses on relevant structural, physical and electrical properties of these materials and on the related sectors where there has been considerable industrial input. Key words: polymer electrolyte classes, polymer electrolyte properties, batteries, fuel cells and capacitors, photoelectro chemical devices, electrochromic applications.

1.1

Introduction

Polymers are at the forefront of the materials revolution; their uses are expected to grow at a faster rate (~15% per year) than that of any other structural compound. This trend may even be accentuated as polymers with new functionalities appear. We shall discuss here polymers having electrolyte properties for which important applications are foreseen. Ionic conductivity is usually restricted to salts in the molten state or their solutions in polar solvents. Despite their ubiquity, the understanding of such complex media is far from complete. On the other hand, a few remarkable solids with high ion mobility have been discovered, ranging from PbF2 by Faraday (1839) to recent newcomers including β-alumina and sulphide glasses. In such materials, one of the sublattices is in a quasi-molten state while still held by a rigid framework. Clays, zeolites, and natural and synthetic ion exchange resins may be considered as intermediate materials between solutions and the above-mentioned solid electrolytes. This is particularly true in the sense that a fixed network bears immobilised electrical charges while the discrete counter-ions are mobile, but only when the coulombic interactions are screened by a polar solvent. Water or alcohols are usually required as both anion (hydrogen bonding) and cation (electron pairs) solvation is needed (Gray, 1997). The motion of ions in polymeric 3 © Woodhead Publishing Limited, 2010

4


matrices in the absence of a solvent is a relatively new phenomenon whose existence and importance have been recognised only in the past decades (Armand, 1986; Bruce, 1995a; Gray, 1997; Linford, 1987; Lipkowski and Ross, 1994; MacCallum and Vincent, 1987; Skotheim, 1988; Takahashi, 1989; Vincent, 1987a; Wright, 1975) and corresponds to a different concept; the macromolecule itself acts as a solvent for a salt which becomes partially dissociated in the matrix, leading to electrolyte behaviour. No small molecules are required for the conducting process, though the polymer–salt complex is routinely made from a solution of both constituents. An interesting parallel can be drawn between electronically and ionically conducting polymers: the latter (colourless) donate electron pairs from heteroatoms included in the macromolecular array from ‘s’ orbitals to Lewis acids (cations), while the former (highly absorbing) release single electrons from their conjugated ‘π’ system upon oxidative doping, with the formation of polarons (radical cations) possibly as dimers (bipolarons). Polymer electrolytes are to solutions what redox polymers (polypyrrole, polythiophene) are to metals. As a rule, electronic conductivity is best in spatially regular systems (crystals) while ionic conductivity prefers disordered matter. The ionic conductivity of polymer electrolytes is typically 100 to 1000 times less than exhibited by a liquid- or ceramic-based electrolyte. Although higher conductivities are preferable, and indeed a great deal of effort has gone into improving the bulk conductivity of polymer electrolytes over the years, 100-fold or 1000-fold increases are not essential, as a thin film electrochemical cell configuration can largely compensate for the lower values. Many polymer electrolyte materials will exhibit to a greater or lesser extent the following properties: • • • • •

adequate ionic conductivity for practical purposes; low electronic conductivity; good mechanical properties; chemical, electrochemical and photochemical stability; ease of processing.

These critical attributes are necessary if the materials are to be considered as practical replacements for their liquid counterparts. In addition, their properties, particularly conductivity and transport properties, should be sufficiently practical to stimulate their development when compared with other highly conducting solid electrolyte materials. Since 1978, when Michel Armand first introduced polyether–alkali–metal salt complexes to the solid state community as potential materials for electrochemical devices, there has been an enormous amount of research carried out on these (particularly high molecular weight poly(ethylene oxide)–lithium salt) systems, to obtain


Introduction to polymer electrolyte materials

5

a better understanding of their fundamental properties and to use this information for the development of new generations of polymer electrolyte materials that are commercially more attractive (Armand, 1983; Plancha et al., 1997a).

1.2

Categories of polymer electrolytes

Rather confusingly, the term polymer electrolyte is widely used in fuel cell technology to refer to ion-selective polymers which, when hydrated, have two intermingled phases – the solid polymer region and a liquid-like region. An example is DuPont’s Nafion which is a fluorinated polymer with sulphonate end groups. The liquid phase is sometimes described as composed of micelles which are channel-like regions filled with liquid. Electrostatic regions in the ionic structure of the polymer backbone, with the mobile charges in water, help to form these regions which contain a liquid phase. There have been attempts to develop similar single ion conductive polymers for lithium batteries but, as yet, all materials reported have a resistivity which is too high for practical use. Finally, the term polymer electrolyte is also used for compositions which are essentially liquids adsorbed into a polymer. An example is found in the Bellcore system (Gozdz et al., 1994). US Patent 5, 296, 318 describes a copolymer of hexafluoride propylene and vinylidene 1,2-difluoride which is wetted with a liquid electrolyte. The term polymer electrolyte has, therefore, come to mean any polymerbased structure with significant ionic conductivity. These are found with a wide range of solid-like character. The solid character of polymers is, in general, related to the molecular weight of the polymer. Low molecular weight polymers are often liquids so the range can be from liquids to very hard and rigid materials. Some polymers can organise at the molecular level in such a fashion as to be crystalline. Since conductivity comes about through molecular motion in the structure, crystalline polymers have low conductivity and are not options for batteries. The ‘dry’ polymer electrolyte is a single phase, non-crystalline material containing dissolved salt, and where the ions of the salt are mobile. Plasticised polymers, a term from the polymer industry, are single phase and contain organic additives which have the effect of softening the polymer. These, in general, have higher conductivity than the dry polymers because of greater freedom for molecular motion. Solventdoped polymer electrolytes may be either single phase or two-phase depending on the degree of doping and the molecular make-up. The twophase materials are frequently described as gels. Both anions and cations are mobile at the molecular level. Finally, ion-selective, two-phase polymer electrolytes are in use in fuel cell technology but have not found application in lithium batteries (Gray, 1997).


6


Among the different types of polymer electrolytes, the (liquid) solventfree system, where the ionically conducting phase is formed by dissolving salts in a high molecular weight polar polymer matrix, constitutes the ‘polymer electrolyte’ in its original sense. This category is the main subject of this chapter, due to its intensive development for lithium secondary batteries, electric propulsion applications, etc. Another category of primary importance is the proton-conducting polymer electrolyte used as the electrolyte in proton-exchange membrane (PEM) fuel cells, also known as solid polymer electrolyte fuel cells (SPEFC). These two categories will be further considered in the following sub-sections.

1.2.1 Lithium polymer electrolytes The history of lithium polymer electrolytes begins in 1973 with work by Fenton et al. (1973) at the University of Sussex (UK), and continues with several works by Armand (1986) at the University of Grenoble. Subsequently, two separate groups – a Harwell-led consortium of companies and a Hydro Quebec-Elf Aquitaine joint venture – set out to explore commercial possibilities of polymer electrolytes for lithium batteries (Araújo and Sequeira, 1994; North et al., 1982; Patrick et al., 1987; Sequeira and Hooper, 1983a). The first model to explain conductivity in polymer electrolytes emphasised the helical structure of poly(ethylene oxide) (PEO) and postulated the lithium cation either moved along the axis of the helix or along the outside of the strand. This model is based on ordered structures. It is not able to explain why the conductivity of PEO (the material can be described as solid if the molecular weight is >100 000) sharply changes at temperatures around 65 °C. By the mid-1980s it was recognised that most of the conduction occurred in the amorphous phase and the reason for low conductivity below about 65 °C was the polymer crystallinity. Subsequently, efforts were applied to reducing the crystallisation temperature of PEO (and other polymers). One method is to attach PEO strands to a polymer backbone. The relatively short strands are unable to self-align into a crystalline structure, thus extending the amorphous phase to much lower temperatures. Another option is to form copolymers which are not able to self-align as readily as the PEO strands. These types of material are what Megahed and Scrosati (1994) call the second generation of polymer electrolytes. Several solid polymer electrolytes (SPEs) based on copolymers have been prepared and complexed with lithium salts to prepare polymer electrolytes for lithium batteries. Incorporation of 10 to 20% poly(propylene oxide) (PPO) units in PEO depresses considerably the crystallinity of the materials. In addition to copolymer containing polyether grafted polyether, polysiloxane and polyphosphazene backbones have been used. Polymer



7

electrolytes prepared with lithium salts and copolymers have shown enhanced transport properties at room temperature but conductivities and transport numbers are not sufficiently high to be attractive for practical lithium batteries (Fauteux et al., 1995). By the mid-1980s, measurements of transport numbers were showing that the anion was, at least, carrying a current equal to the cation and, in many measurements, a current greater than that carried by the cation. This observation made the original conductivity model inadequate. Molecular models have now been developed that describe motion of the ions based on oscillations of the polymer strands. And the free volume model for conductivity has been adopted from polymer chemistry. It is found that simple salts, such as LiCl, do not provide high conductivity. In general, the bulkier the anion, the higher is the conductivity. For example, Armand (1994) synthesised lithium bis(trifluoromethylsulphonyl)imide, LiN(CF3SO2)2, and observed high conductivity when dissolved in PEO-type polymer electrolytes. The high conductivity is the result of a more ‘plasticised’ polymer. Such an approach showed promising results with LiN(CF3SO2)2, through an increase in amorphous phase in the polymer electrolyte and an increase in conductivity. Therefore, new salts with perfluorosulphonate-polyether based single ion conductors have been synthesised. These salts have been incorporated into solvating polymers by co-crosslinking or copolymerisation. The resulting polymer electrolytes have been characterised by conductivity measurements. The room temperature conductivities of lithium polymer electrolytes are slightly better than the conventional polymer electrolytes but they are not high enough for practical applications in lithium batteries. Both the lithium cation and the anion of the dissolved lithium salt are mobile in the amorphous polymer. As a result, polarisation effects are encountered in battery operation. Efforts (e.g. Shriver at Northwestern University) to prepare single mobile ion electrolytes by building the anion into the polymer structure have not produced practical electrolytes. In all cases, the obtained conductivity has been at least an order of magnitude less than the comparable double ion structure (MacCallum and Vincent, 1987). The concentration of salt in the polymer matrix is higher than generally found in liquid solvents. It happens that the oxygen atom in the ethylene oxide backbone of PEO has just the right charge density to solvate the lithium cation but not to bind the lithium so tightly that it is immobile. On average, the lithium cation interacts with four ether oxygen atoms of the polymer (but not with two sequential ether oxygen atoms in the same strand). It has been observed that ion pairing sets in when the salt concentration exceeds a ratio of one lithium cation to eight ether oxygen atoms of the PEO strands and that ion aggregates appear at concentrations of one


8


cation to four ether oxygen atoms. A maximum in conductivity occurs at concentrations around one lithium cation to eight ether oxygen atoms. The third generation of polymer electrolytes is variously described as containing a plasticiser, as dilute solvent or as gelled (Andrieu et al., 1995; Benrabah et al., 1995). Each of these terms may have had precise meaning at one time or another, but recently the terms seem to be used indiscriminately. Towards 1990, a group at the Engineering Research Laboratories in Denmark, with a group at the Mead Paper Company in the US (in what came to be known as the MHB venture) developed a solvent-diluted electrolyte which was polymerised in situ (Lee et al., 1989). This electrolyte is characterised by the incorporation of a liquid matter, such as propylene carbonate in a polymer matrix that is formed by crosslinking a monomer after mixing in the liquid phase. Following this, many methods and recipes for solvent-doped or ‘plasticised’ polymer electrolytes have been developed. (The term plasticised frequently means that the salt anion is an organic which, when dissolved, makes the polymer soft. But it is also used to refer to softening the polymer by addition of an organic liquid. In other ways, the hardness of the electrolyte depends on the molecular size and constitution of the polymer, so either dry or solvent-doped polymer electrolytes can range from essentially liquid-like to hard-brittle materials.) Surprisingly, up to about 80 wt% of some solvents can be incorporated into the matrix and yet it has the appearance of a dry solid. Other compositions have a texture more that of a gel. Even in the solid matrix, if the solvent has liquid-like character this can be observed by thermogravimetric analysis. The principal feature of the analysis is evaporation of the solvent at the temperature of the bulk liquid. Some compositions, however, do not undergo mass loss until temperatures are reached well above the boiling point of the bulk solvent. These compositions appear to be truly homogeneous ‘dry’ materials. Examples of plasticised polymer electrolyte systems with room temperature conductivities in the range of 2.6 × 10−3 S cm−1 are the polyacrylonitrile (PAN), polyvinylpyrrolidone (PVP), poly(methyl methacrylate) (PMMA) and poly(ethylene oxide) diacrylate (PEODA) plasticised with a low molecular weight solvent such as propylene carbonate, gamma butyrolactone or a high molecular weight solvent such as tetraethylene glycol dimethyl ether (tetraglyme) (Abraham and Alamgir, 1990; Cazzanelli et al., 1995; Huq et al., 1992). The main disadvantage with these polymer electrolytes is the reactivity of these solvents plasticised with lithium metal, accompanied by passivation and short cycle-life in lithium batteries. Also, for room temperature operation, molten salts containing 1-butylpyridinium chloride or bromide and aluminium chloride have been used for high conductivity polymer electrolytes. Poly(l-butyl-4-vinyl pyridinium) halides are known to be compati-



9

ble with the molten salts to form polymer electrolytes. The addition of a small amount of the polypyridinium salts provides viscoelastic rubbery properties to the polymer electrolytes. The conductivity of the polymer electrolyte is influenced by the electrolyte composition and temperature (Watanabe et al., 1995). However, these electrolytes containing chloride ions, and particularly aluminium chloride are too corrosive and their processability into thin film form with mechanical stability is yet to be shown. They are not compatible with lithium anode and possibly not with many potential cathode materials to be useful for practical battery operation. Also in 1995, three superacid salts of lithium of the type CnF2n+1SO3Li where n = 4, 8 and 10 have been prepared and used in the preparation of SPEs with PEO (Nagasubramanian et al., 1995). The room temperature conductivity data show that these electrolytes have low conductivities. The conductivity of the polymer electrolytes PEO/C4F9SO3Li and PEO/C10F21SO3Li is of about 10−6 S cm−1, whereas the conductivity of PEO/C8F17SO3Li is < 10−6 S cm−1. These conductivity values are much lower than the conductivity of the polymer electrolyte with lithium bis(trifluoromethylsulphonyl)imide, PEO/(CF3SO2)2NLi, with a conductivity of 10−5 S cm−1. The increasing need for novel electrolytes that have high conductivity, high lithium ion transport number, a wide electrochemical stability window, and are suitable for lithium batteries with safe operation (Armand, 1994) is stimulating the third generation materials to be amorphous to temperatures as low as −40 °C and have conductivity at room temperature of the order of 10−3 S cm−1, approaching that of liquid electrolytes. Chiang et al. (1987) focused their attention on this need and produced interpenetrating polymer networks (IPNs) which maintained good electrical conductivity at room temperature. They synthesised IPNs from PEO– LiClO4 complexes and epoxy (bisphenol A diglycidyl ether) crosslinked with poly(propylene oxide) (PPO) triamine (Texaco Jeffamine T-403). The IPNs could be described as epoxy–PPO networks plasticised with PEG400. In this system, the epoxy and Jeffamine intercombined to form a solid matrix upon which PEO was held in an amorphous conformation at room temperature, as well as at higher temperatures. An analogy that could be drawn to this system is that of a wet sponge in which the water corresponded to the conducting liquid PEO–LiClO4 complex and the sponge corresponded to the solid support matrix of the epoxy and Jeffamine. Quantitatively, it was found that the introduction of the epoxy increased the mechanical properties of the polymer but decreased its conductive properties. When the epoxy content was greater than 20 wt% (w/o) (of the total mixture of lithium perchlorate (LiClO4) and PEO), free-standing films were formed and their mechanical properties improved as the percentage of epoxy increased. The conductivity, however, of the IPN remained nearly


10


constant from pure PEO–LiClO4 to about 30 w/o epoxy. A further increase in the epoxy wt% led to an exponential decrease in the conductivity. Chiang et al. (1987) therefore concluded from their results that improved properties for a solid electrolyte were obtained with a composition of about 30 w/o epoxy in the IPN system. We have extended the study of Chiang et al. and had three main aims (Hudson and Sequeira, 1995): (i) to prepare ‘solid membrane films’ that could be properly tested – this involved finding suitable proportions of epoxy and Jeffamine which would form a suitable solid material; (ii) to investigate how the conductivity was affected by varying the amounts of LiClO4 present in the system; (iii) finally, to investigate the preparation and conductivity of an IPN in which the epoxy was crosslinked with tetraethylenepentamine rather than PPO triamine. This study showed that thin films (~0.015 cm) of IPNs could be prepared by a simple process, and have good mechanical and conductive properties, which are two of the essential requirements of a SPE. The ease of preparation and handling of these films would be vital if the IPNs fulfilled their commercial potential, if the large-scale mixing of different materials to form solutions of the IPNs would not cause the chemical engineer too many problems, and if the casting of the films would also be quite simple on a large scale. Moreover, the work showed that the conductivity of the IPN system was dependent on the concentration of the salt. The conductivity did not increase directly with increasing amounts of added salt because at very high salt concentrations neutral ion pairs were probably formed. Instead, there seemed to be an optimum value for the conductivity, which appeared at Table 1.1 The composition and room temperature conductivity of some solid polymer electrolytes (SPEs) and plasticised polymer films

Lithium salt PEO complex

Li : O ratio

Room temp. conductivity (S cm−1)

[CF3(CF2)3SO2]NLi : PEO [CF3(CF2)3SO2]NLi : PEO [CF3(CF2)3SO2]NLi : PEO [CF3(CF2)3SO2]NLi : PEO Lithium salt plasticised with PAN [CF3(CF2)3SO2NLi2] : PEO [CF3(CF2)3SO2NLi2] : PEO Dilithium salt plasticised with PAN C4F9SO2NLi2 plasticised with PAN C8F17SO2NLi2 : PEO C8F17SO2NLi2 : PEO LiClO4 : PEO in IPN LiClO4 : PEO in IPN

1:6 1:8 1 : 16 1 : 24 – 1:8 1 : 24 – – 1:8 1 : 16 1 : 88 1 : 176

5.31 3.24 3.0 4.77 1.02 1.79 3.03 8.06 1.1 4.24 4.37 6.10 1.80


× × × × × × × × × × × × ×

10−6 10−6 10−6 10−6 10−4 10−6 10−6 10−5 10−5 10−7 10−7 10−5 10−6


11

quite a low salt concentration where LiClO4 probably existed as mobile, free ions. The maximum and minimum conductivities at room temperature were obtained with (PEO)88:(LiClO4) and (PEO)176:(LiClO4), being 6.1 × 10−5 S cm−1 and 1.8 × 10−6 S cm−1, respectively. Compositions and conductivities of some SPE films and plasticised films that have been synthesised and characterised are listed in Table 1.1. Many more lithium polymer electrolytes have been developed for commercial applications (Araújo et al., 1996a; Cao et al., 2006; Dong and Wang, 2005; Hudson and Sequeira, 1993, 1994; Itoh et al., 2008; Kolonitsyn et al., 2005; Liao and Ye, 2004; Magistris et al., 2001; Nakagawa et al., 2003; Niedzicki et al., 2009; Niitani et al., 2005; Othman et al., 2007; Plancha et al., 1997b; Sequeira and Hooper, 1992; Shi et al., 2002; Tao et al., 2007; Wang et al., 2004). However, most of these lithium polymer electrolytes are still not good enough for lithium batteries and other electrochemical devices.

1.2.2 Proton polymer electrolytes The need to operate polymer electrolyte membrane (PEM) fuel cells at temperatures above 100 °C, where the amount of water in the membrane is restricted, has provided much of the motivation for understanding the mechanisms of proton conduction at low degrees of hydration. Although experiments have not provided any direct information, numerous theoretical investigations have begun to provide the basis for understanding the mechanisms of proton conduction in these nano-phase separated materials. Both the hydrated morphology and the nature of the confined water in the hydrophilic domains influence proton dissociation from the acidic sites (i.e. SO3H), transfer to the water environment, and transport through the membrane (Paddison, 2003). Hereinafter we discuss progress on proton membranes in SPEFCs. The Nafion membrane is an example of this type of membrane. Classically, the Nafion membranes are chemically synthesised in four steps according to the DuPont de Nemours process (Grott, 1978): (a) the reaction of tetrafluoroethylene (TFE) with SO3 to form the sulphone cycle; (b) the condensation of these products with sodium carbonate followed by copolymerisation with the TFE to form an insoluble resin; (c) the hydrolysis of this resin to form a persulphonic polymer; and (d) the chemical exchange of the counter ion Na+ with the proton, H+, in an appropriate electrolyte. This membrane was introduced by DuPont in 1966. The development of perfluorinated membranes by DuPont during the 1960s has played a vital role in electrochemical system applications. The Nafion family or perfluorinated ionomer membranes meet the requirements for several electrochemical systems (chloroalkaline, fuel cells and some other non-fuel cell applications). A lifetime of over 60 000 hours has


12

Polymer electrolytes ( CF2 – CF2 )x ( CF – CF2 )y [OCF2CF]z – O(CF2)2 SO3H CF3

1.1 The general structure of Nafion membrane: x = 6 − 10 ; y = z = 1.

been achieved at 80 °C. Their general formula is given in Fig. 1.1 (Savadogo et al., 1995). The values of x, y and z in Fig. 1.1 are x = 6−10, y = z = 1. These values can be varied to produce materials with different equivalent weights (EWs) and pendant chain lengths, the equivalent weight (EW) being the number of grams of polymer per mole of fixed SO3 sites. The high EWs of the first Nafion membranes’ family limited their use in fuel cells and prompted the development of the Dow membrane. This membrane is structurally and morphologically similar to the Nafion membrane, but differs with respect to its EWs, which are typically in the 800 to 850 range, and have shorter size chains (z = 0 for Dow and z = 1 for Nafion) (Utracki and Weiss, 1989). The Dow membranes are a short side chain perfluorinated ionomer whereas the Nafion membranes are the long side chain perfluorinated ionomer. The specific conductance of 800 and 850 EW experimental membranes has been reported as 0.20 and 0.12 Ω−1 cm−1, respectively (Savadogo et al., 1995). It must be pointed out that the Dow monomer is more complicated to elaborate than the DuPont monomer. Therefore, the synthesis of the Dow epoxy is more complicated than that of the Nafion, which is a commercially available material (Savadogo et al., 1995). Testing the Dow membrane, during the period 1987–88, in Ballard fuel cells resulted in a dramatic increase in the SPEFC performance levels (Watkins, 1988). Since this success by Dow, DuPont has been actively developing their membranes with respect to durability and continuous improvement. They have increased power densities by further decreasing the equivalent weight, from 1100 EW (Nafion 117) to 1000 EW (Nafion 105), and the membrane thickness from 178 to 127 µm. Nafion 115 (127 µm) and Nafion 105 (127 µm) are some of the latest materials developed specially for fuel cell applications. Perfluorinated ionomer membranes have also been developed by Asahi Chemical and the Asahi Glass Company and commercialised as Aciplex S and Flemion, respectively. The general properties of the long side chain perfluorinated ionomer membranes (e.g. Nafion, Flemion, Aciplex) and the short side chain perfluorinated ionomer membranes (e.g. Dow) are shown in Table 1.2.



13

Table 1.2 Properties of perfluorinated monomer membranes (e.g. Nafion, Flemion, Aciplex, Dow) EW range

800–1500 −1

Conductivity range (S cm ) Conductance range (S cm−2) Dimensional stability

Proven lifetime Thickness (µm)

0.20–0.05; for example: 1100 EW = 0.10 and 850 EW = 0.15 2–20; for example: 1100 EW (178 and 51 µm membrane) = 5 and 17, respectively 10–30% expansion X and Y direction from 50% RH liquid water (25 °C); for example: 1000 EW = 16% expansion >50 000 hours for Nafion, Flemion and Aciplex; >10 000 hours for Dow 250–50

EW stands for equivalent weight and RH stands for the relative humidity of the membranes.

Nowadays, the long side chain perfluorinated polymer electrolytes have proven to have a prolonged service life under electrolysis and electrosynthesis conditions and in fuel cell systems (Appleby and Yeager, 1986; Blomen and Mugerwa, 1983; Chapiro, 1962; Cohen et al., 1949; GuzmanGarcia et al., 1992; Hodgdon, 1968; Küver et al., 1994; La Conti, 1988; Prater, 1990; Riedinger and Faul, 1988; Samms et al., 1996; Savadogo and Roberge, 1997; Wang et al., 1996a). They are not efficient for high temperature (>150 °C) fuel cell systems because their conductivity decreases significantly due to dehydration. Although some efficiency has been gained by lowering the membrane EWs, the main improvements in fuel cell performance were achieved by simply thinning the membranes (Prater, 1990). The advantages gained with this simple strategy include lower membrane resistance, lower material utilisation (which obviously allows cost savings), and improved hydration of the membrane. However, there is a limit to the extent to which such membranes can be thinned because of difficulties with durability and reactant crossover. W.L. Gore and Associates, Inc. proposed to solve some of these difficulties using perfluorinated monomer membranes reinforced with woven polytetrafluoroethylene (PTFE). Perfluorinated membranes based on weak acid functions have been developed by Asahi Chemicals (Japan). Their EW is similar to that of the above membranes. The exact synthesis process of the membrane is not indicated. They have been developed for the chlor-alkali industry where they are used in bilayer membranes. In these membranes a strongly acidic membrane (–SO3H) (pKa < 1) is coated on one side, and the second side is coated with a thin layer of a weak-acid membrane (–COOH) (pKa = 3).


14


Only the strongly acidic membrane is exposed to the anolyte so that acid may be added while the weak-acid layer is exposed to the catholyte to prevent sodium hydroxide transport from catholyte to anolyte. A comparison between the membranes containing –SO3H and –COOH ion exchange groups has shown that the water content and the ionic conductivity of the –SO3H-based membranes are higher than those of the –COOH-based membranes. Therefore, the latter are less suitable for fuel cell applications. Perfluorinated membranes reinforced with woven PTFE, such as Nafion 324 or Nafion 417, are used in many industrial electrochemical processes. Unfortunately, the relatively coarse weave of the woven PTFE reinforcement results in membranes that are much too thick for high electrochemical performance. Some non-woven PTFE/perfluorinated ionomer composite membranes have been formulated for ion transport studies and for investigation of other properties relevant to chlor-alkali and fuel cell applications. Advancements in materials and processing technology have resulted in the introduction of PTFE/perfluorinated ionomer composite membranes claimed by Gore and Associates, Inc. under the Gore Select trademark. Most types of ion exchange membranes contain some type of reinforcement because they are weak and tend to swell substantially as they incorporate a solvent onto their structure. This reinforcement can be a ‘macro reinforcement’ or a ‘micro reinforcement’. The reinforced membranes are, of course, more hydrophobic than the conventional Nafion membranes. It may be interesting to investigate if the low resistance is due to the high reinforcement of the membrane (which is very thin) or to its water transport ability. It is well established that, for all types of membranes, the low EW enhances the water transport ability. But it does not seem that the reinforcement has a beneficial effect on water transport. Up to now, the Nafion membrane and the Dow membrane have been most effective advanced membranes suitable for use in practical systems. The Dow membrane exhibits superior performance to that of the Nafion 117 membrane. The major disadvantage of these membranes is their high cost. This is due to the long preparation process and to the thickness of these membranes. To achieve direct chemical synthesis of Nafion membranes without these problems, novel methods of preparation were developed. It was found that the ionic conductivities of the membranes elaborated with Nafion, heteropolyacids and/or thiophene are higher than those with just Nafion. The ionic conductivities of these membranes fabricated with Nafion electrolyte and silicotungstic acid (STA), with and without thiophene, are 2.3 times higher than that of the Nafion 117 membrane. Their conductances are also higher than that of the Nafion 117 membrane. The water uptake of the Nafion, Dow, Gore Select and the membranes fabricated with Nafion electrolyte and STA are shown in Table 1.3. As it



15

Table 1.3 Polymer electrolyte membranes water uptake Membrane

EW

Thickness (µm)

Water uptake (%)

Nafion 117 Nafion 112 Dev Dow Gore-Select Gore-Select NASTA NASTHI NASTATHI

1100 1100 800 1100 900 1100 1100 1100

200 60 100 20 12 410 120 150

34 34 56 32 43 48 32 30

may be seen, the water uptake of the Dow membrane is higher than that of the Nafion 117. This is related to the chemical composition of the membrane and not to its thickness. Effectively, Nafion 117 (200 µm) and Nafion 112 (60 µm), which have the same chemical composition but different thicknesses, exhibit the same water uptake. The water uptake of the Nafion solution and STA is 40% higher than that of Nafion. This may indicate that the STA may modify the membrane chemical properties. In the same way, membranes fabricated with Nafion 117 and STA (NASTA) enhance water transport. Similar results were obtained on other heteropolyacids. The improvements on the membranes’ conductivity, conductance and water transport were attributed only to the presence of STA. The preparation of the monomer α,β,β-trifluorostyrene (TFS) was first carried out by Cohen et al. (1949) who reported its total synthesis. Hodgdon (1968) firstly investigated the sulphonation of TFS with applications to structures and cells. Optimising the reaction conditions, Hodgdon was able to prepare a multiplicity of EW or different ion-exchange capacity (IEC) of both linear and crosslinked poly-TFS sulphonic acids. He showed that the extreme difficulty in sulphonating poly-TFS was caused by the betadirecting influence of the perfluorinated polyalkyl group attached to the aromatic ring (Hodgdon, 1968). It was claimed that the high stability of the fluorine atoms attached to the alkali carbon atoms impacts oxidative and thermal stability superior to that exhibited by conventional ion exchange polymers such as polystyrene sulphonic acid (La Conti, 1988). The limited use of these membranes in fuel cells may be related to their limited mechanical properties and/or chemical resistance in real fuel cell conditions. Based on the above work on sulphonated TFS membranes, a novel family of sulphonated copolymers incorporating TFS and a series of substituted TFS monomers provided the group of materials referred to as BAM3G (Ballard Advanced Materials 3rd Generation of membranes) (Savadogo et al., 1995; Savadogo and Roberge, 1997).


16


The most significant earliest work carried out on the radiation grafting of polymers was achieved and reviewed by Chapiro (1962). The cation exchange membrane (CEM) was prepared by graft polymerising TFS onto the inactive fluorine-containing polymer film, and then sulphonating the grafted film in different patterns. The grafted polymer membranes were obtained at a graft rate of 10 to 50%. The membrane containing grafted polystyrene sulphonic acid and two low density polystyrene membranes (PTFE and a copolymer of TFE and hexafluoropropylene) have shown similar conductivities to those of the Nafion and Dow membranes (Savadogo and Roberge, 1997). Many reports (Savadogo et al., 1995) indicate that the oxidation stability of these grafted foils is very limited. This makes them appropriate for fuel cell applications; but only the membranes based on the PTFE backbone showed some significance as candidate materials. It is well established that the conductivity of the perfluorinated ionomer membranes currently used in PEM fuel cells depends on the presence of water to solvate the protons generated by the ionisation of the sulphonate acid groups. When such polymers are subjected to temperatures above 100 °C at atmospheric pressure, dehydration occurs, and conductivity decreases significantly (Samms et al., 1996). It has been shown that for a prototype direct methanol fuel cell (DMFC), the methanol crossing over the membrane was responsible for the decrease in cathode potential by at least 100 mV (Küver et al., 1994). From the investigation of the methanol diffusion in Nafion (Küver et al., 1994) it was concluded that methanol readily transports across perfluorosulphonic acid (PFSA) membranes, and the investigation of a few membrane systems for DMFCs was suggested. The low cost and excellent oxidation and thermal stability of phosphoric acid doped polybenzimidazole (PBI) prompted researchers at Case Western University (Samms et al., 1996; Wang et al., 1996a,b,c) to develop this membrane as a polymer electrolyte for DMFCs. After investigation of the thermal stability of PBI doped with phosphoric acid up to 600 °C, it was concluded that this membrane is adequate for use as PEM in a high temperature fuel cell. These studies may also be extended to other polymers like the polybenzimidazobenzophenanthrolines (PBIPAs) which exhibit excellent thermal and mechanical properties (Zhou and Lu, 1994). The sulphonated polyimide membranes are based on 4,4′-diaminobiphenyl2,2′-disulphonic acid (BDSA), 4,4′-oxydianiline (ODA), 4,4′-oxydiphthalic anhydride (ODPA) and 1,4,5,8-naphthalenetetracarboxylic dianhydride (NTCDA). The study of the surface properties of fluorinated polyimides exposed to vacuum ultraviolet (VUV) radiation and atomic oxygen may help us understand the stability of these polymers. It will also be interesting to apply this sulphonation method to some novel high molar mass polyetherimide, as reported by Hsio and Yang (1997).



17

The poly(benzylsulphonic acid) siloxane (PBSS)-based membrane is based on organic–inorganic protonic polymer electrolyte (ormolyte–organically modified silane electrolyte) (Chambouillet et al., 1988; GauthierLuneau et al., 1992). The synthesis was carried out by the sol–gel process during different steps. It was based on the hydrolysis-condensation of the benzyltriethoxysilane precursor in a methanol solution containing water and trifluoromethanesulphonic acid (CF3SO3H) (Gauthier-Luneau et al., 1992). The final inorganic–organic ion-exchange ormolyte had good thermal stability up to 250 °C, and showed a conductivity of 1.6 × 10−2 Ω−1 cm−1 at room temperature (Gauthier-Luneau et al., 1992). It was claimed that these properties made them useful for DMFCs but, up to now, no fuel cell performance data were given. It would be interesting to study the methanol crossover in this membrane; its chemical stability in practical fuel cell systems should also be addressed. Besides the studies on anhydrous proton polymer electrolytes, various research groups have studied proton conductivity in systems in which aqueous solutions of H3PO4 or H2SO4 were used as additives to the polymeric matrix (Lassagues et al., 1992; Raducha et al., 1996; Wieczorek et al., 1994). It has been demonstrated that thin film proton conducting polymer gels based on polyacrylamide (PAM) matrix and agar with a H3PO4 additive exhibited reasonable mechanical strength and room temperature conductivities in the range of 10−3 to 10−2 S cm−1 (Wieczorek et al., 1995). The conductivities are higher than those obtained from many of the above non-fluorinated membranes. Differential scanning calorimetry (DSC) measurements indicate that these protonic gels are stable up to 120 °C. Their mechanical stability and conductivities are very dependent on the water content. Also, their water uptake is not well determined in various operation conditions. According to the available literature data, the conductivities of these systems are less than those of perfluorinated membranes at room temperature. It would be of interest to evaluate their performance in practical fuel cell systems. In particular, the stability of these proton conductors in actual practical systems must be addressed in detail. In summary, the modifications made to the Nafion membranes, the conceptual design of substitutes for PFSA materials and the modifications made to aromatic membranes to render them suitable for fuel cells and other electrochemical device applications constitute promising avenues for present and future research.

1.3

Structure and its implications

It is of maximum importance to understand the ion transport in polymer electrolytes. It is the basic principle governing the interactions between ions


18


and their polymeric host (and the ions themselves). Of course, the short range dynamics of the polymer host almost invariably control the ratedetermining step for ionic motion in simple polymer electrolytes. Many efforts have been made over the years to address these issues, in particular through Raman, infrared, nuclear magnetic resonance (NMR) and extended X-ray absorption fine structure (EXAFS) studies of amorphous systems (Aihara et al., 2000; Bronstein et al., 2004; Forsyth et al., 2002, Frech and Manning, 1992; Glasse et al., 1992; Jin et al., 2000; Ouyang et al., 2007; Wintersgill et al., 1988). For a salt MX dissolved in a polymer host solvent, the formation of ion pairs, [MX]•, produces neutral species and the concentration of charge carriers drops. Larger aggregates may also exist (e.g. [M2X]+, [MX2]−), whose mobilities will be impaired by their size in comparison to free ions, and conductivity will again be adversely affected. The importance of ion association in a salt solution is largely dependent on the dielectric constant, ε, of the solvent. As the salt concentration is increased, the inter-ionic distance decreases and ion–ion interactions become progressively more significant. The onset of ion-pair formation occurs at lower ion concentrations in solvents of low ε. As ε ~ 5–10 for polyethers (cf. 78.5 for water), extensive ion–ion interactions are expected to be favourable. When a salt is dissolved in a polymer matrix, the conductivity of that polymer increases due to the concentration of charge carriers. Moreover, as the salt concentration, for example for a polyether, is increased above ~0.1 mol dm−3 (M:O ratio of ~ 1:100 to 1:50), the conductivity is found to reach a maximum and then falls. Ion transport is closely coupled to the segmental motion of the polymer. Cations moving between coordinating sites must adopt a transient state, where they are coordinated by both sites, before the motion of the polymer breaks one coordinating bond. The fall in conductivity as the salt concentration increases is thought to be due to the introduction of an ever-increasing number of the transient crosslinks in the system, which causes a reduction in chain mobility. Ion aggregation can still contribute to the observed decrease in conductivity as, if aggregates do exist, they are likely to have retarded diffusion rates, making an additional contribution to the fall in conductivity. Infrared and Raman spectroscopy have been used to study both interactions between the ions and host polymer and those between cations and anions (MacCallum and Vincent, 1989). This relies on monitoring charges in the vibrational modes of anions, namely the trifluoromethanesulphonate (CF3SO3−) and perchlorate (ClO4−) anions (MacCallum and Vincent, 1989). Variation in vibrational frequency and the Raman spectral line width can provide information about the influence of the environment on the molecular ions. Changes in dipole moment arising from an ion-pair vibration are Raman inactive but strong infrared absorptions, usually in the far infrared,



19

are expected. In addition, Raman scattering studies have been used to quantify the degree of ion association and prove its variations with temperature and salt concentration (Ferry et al., 1995; Gray, 1991; Schantz and Torell, 1993). Raman and Fourier transform infrared (FTIR) spectroscopy studies have been carried out in a wide range of metal salt complexes. Many trends, such as increased pair/cluster formation with increased temperature or salt concentration, have been noted (Bernson and Lindgren, 1993; Wendsjö et al., 1992). As a result of the many studies carried out, a wider variety of configurations is possible in the complex systems, and complicated temperature dependence for solvation is expected (Andreev and Bruce, 2000; Papke et al., 1981a; Staunton et al., 2005; Thomson et al., 1996). Ion pairs or aggregates may be defined as contact or solvent separated species. The former is produced when a molecule of the first-neighbour solvent shell around an ion is replaced by an oppositely charged ion. The latter results when there is an overlap of solvent cospheres, linking a pair of ions to form a mechanical unit. The ion centre-to-centre distance is therefore greater in this case, not needing to be limited to one interlinking solvent molecule. In general, it is very easy for contact ion pairs to form in polymer electrolytes, even at low salt concentrations (Bruce and Gray, 1995). While the stoichiometric crystalline complexes do not contribute significantly to the conductivity of the polymer electrolyte, their structure provides valuable information on ion–polymer and ion–ion interactions. There is new evidence from infrared and Raman spectroscopy, EXAFS, thermodynamic and other studies that the local environment in the melt is likely to be very similar to that in the crystal from which it was obtained. Therefore, the crystal structure could provide a useful insight into the structural make-up of the conducting amorphous phases. X-ray powder diffraction has been used extensively to establish the unit cell dimensions of polymer electrolyte crystalline complexes. Determination of crystal structures of polymer electrolytes is not so straightforward since single crystal X-ray diffraction is not suitable. In addition, the large unit cells and low symmetry of polymer electrolyte crystalline phases cause conditioning problems. Success in polymer electrolyte crystal structure determination has come through high quality powder X-ray data of polycrystalline powders. These data can be analysed using Rietveld profile refinement in which an unrefined starting model for the structure is fitted to the observed data by a non-linear least squares technique. Alternatively they can be analysed by using powerful ab initio methods, where no starting model is required. To date, many structures have been elucidated in this way (Andreev et al., 1996; Baskaran et al., 2007; Bruce, 1995b; Chatani and Okamura, 1987; Frech et al., 1997; Fullerton-Shirey and Marenas, 2009; Henderson et al., 2003; Lightfoot et al., 1993; Martin-Litas et al., 2002; Murakami et al., 2002).


20


There are many aspects of the structural features of polymer electrolytes that cannot yet be fully resolved experimentally and significant contributions could be made by undertaking computer simulations. Areas where simulation studies could be beneficial include the following: •

•

•

• •

Determination of the form that polymeric helices take-up above the melting point. Then, the relative importance of inter- and intra-chain crosslinks over a concentration range could be correlated with the glass temperature (Tg) and conductivity measurements. Crystal structure and vibrational spectroscopic data have only provided information on a single stoichiometric phase, that is, on short-range environments, so it would be useful to address this issue through computational methods. Spectroscopic data suggest that there are at least three environments for anions, i.e. different ion associated species, which vary in concentration at moderately high salt concentrations, and computer simulations could assign structures to these ionic clusters. Elucidating the role of various plasticisers in gel-type electrolytes. More structural information from simulation studies could play a role in designing gel electrolytes for optimum transport and dynamic properties.

In addition to building our structural knowledge, computer simulation studies can also play a key role in resolving ion transport processes. Several groups have carried out molecular dynamic simulations of polymer electrolytes. These simulations have been implemented in investigations of the polymer electrolytes’ structural properties and are revealing accurate descriptions of these materials and their potential applications (Capiglia et al., 2000; Catlow and Mills, 1995; Ennari et al., 2000; Gadjourova et al., 2001; Joo et al., 2002; Karlinsey et al., 2004; Neyertz et al., 1995; Payne et al., 1995; Ward et al., 1995; Zheng et al., 2000).

1.4

Conductivity measurements

The conductivity of a polymer electrolyte is an important parameter. This section addresses electrochemical conductivity techniques for the study of SPEs. The different types of conductivity are discussed, followed by an outline of the features, applicability and validity of direct current (DC) and alternating current (AC) conductivity measurements. Techniques for the identification of the individual species responsible for conduction are then briefly reviewed.

1.4.1 Contributions to conductivity The total conductivity, σ, is the sum of the electronic, σel, and ionic, σi, contributions © Woodhead Publishing Limited, 2010

Introduction to polymer electrolyte materials σ = σel + σi

21 [1.1]

Consider a polymer electrolyte of the type Pn:MX (where P is the structural repeat unit of the polymer chain and n is the stoichiometric ratio of structural repeat units to formula units of salt MX). Furthermore, assume that the valences of M and of X are plus one and minus one, respectively. An average value of the electrical conductivity can be obtained by measuring the AC conductivity of Pn:MX between two inert electrodes and with an inert gas passing over the sample. Alternatively, four probes can be used and the IR drop is measured between the two inner probes (see Fig. 1.2). However, both of these configurations leave the chemical potentials of M and X ill defined. It is known that the partial conductivities may be decisively affected by the M/X ratio, which will be established by fixing the chemical potentials. Because of this, it is preferable to fix the M/X ratio by fixing the chemical potential of one of the two constituents, as shown schematically in Fig. 1.3. Figure 1.3(a) involves symmetric electrodes of the parent metal M. These can fix the chemical potential of M. Figure 1.3(b) shows a setup using an alloy, MN, to fix the chemical potential of M at a value lower than that

Argon gas Pn : MX

Pt

Pt

IR drop

1.2 Schematic diagram of electrodes used to measure the AC conductivity of Pn:MX either by the two-probe or four-probe technique. The chemical potentials in the polymer electrolyte are not defined.

Argon gas (a)

Pn : MX

M

M

Argon gas (b)

M–N alloy

Pn : MX

Porous Pt (c) pX2

M–N alloy Porous Pt

Pn : MX

pX2

1.3 Schematic diagram of electrodes used to fix the M/X ratio in a solid polymer electrolyte, Pn:MX: (a) electrodes of the parent metal, M; (b) electrodes of an alloy, MN, in which N is inert; (c) electrodes of porous platinum with a coexisting partial pressure of X2 gas.


22


depicted in Fig. 1.3(a). Here, N is an inert element. Figure 1.3(c) shows a diagram of electrodes to fix the chemical potential of M at its lowest value. In each of these cases, the electrodes are assumed to be reversible. Using such setups, one can obtain a well-defined value of the total conductivity σ as defined above. A useful electrolyte must behave not only as an ionic conductor but also as an electronic insulator. In practice, this means that the electronic conductivity must be so low (~10−9 S cm−1) that the cell would take a very long time, say four years, to self-discharge on standing. In general, the electronic conductivity of an electrolyte should be at least five orders of magnitude lower than the ionic conductivity, which means that the self-discharge current is below 10−5 of the normal current under load. Polymer electrolytes, unlike many other forms of solid electrolytes, have no difficulty in fulfilling this requirement. The electronic conductivity arises from current carried by electrons, σe, and/or electron holes, σh: σel = σe + σh

[1.2]

It is possible to discriminate between these two contributions by using techniques like the Wagner polarisation cell, which is discussed in Section 1.4.4. The disadvantage of such methods is that the test electrodes replace those that would be used in a practical device application and so the behaviour of the electrode–electrolyte interfaces is significantly different. In consequence, the predicted electronic conductivity may be some orders of magnitude higher or lower than that found when the electrolyte is incorporated in a working device. The ionic conductivity of a polymer electrolyte of the type Pn:MX would be expected to arise solely from cationic contributions from M+ and anionic contributions from X−. For polymer electrolytes where P is PEO, for which the relative molar mass of the repeat unit is 44, and n is typically about 4 to 10, the concentration of the ‘solution’ of salt in the polymer is about 2–5 M. This approaches the concentration found in molten salts. A feature of the behaviour of such concentrated systems is the preponderance of ion pairs and triple ions. The ionic transport then includes contributions from cations such as M2X+ and anions such as MX2−. Most techniques for determining the individual conductivities of different species in fact only discriminate between two classes: those that are reversible at a given non-blocking electrode and those that are not. If the test cell is configured so that the contribution of the cations to the total ionic conductivity can be determined, then the cationic transference number, τ+, can be obtained. This may be defined as: τ+ =

σ+ = σi

∑ (σ σi

+

)j

= ∑ ( t+ ) j


[1.3]


23

where (σ+)1, (σ+)2, etc., are the conductivities of the first, second, etc., cationically conducting species. The anionic transference number, τ−, is similarly defined, and is related to τ+ by τ− = 1 − τ+

[1.4]

The term transport number, t, is reserved for the contribution of an individual species. For the polymer electrolyte Pn:MX, there are separate transport numbers t (M+), t (M2X+) (and perhaps, if water is also present within the electrolyte, t (H+) for each of the individual mobile cation species, but it is easier to determine the sum of these t+, than to separate out the speciated values.

1.4.2 Direct current conductivity DC conditions using either blocking or non-blocking electrodes give rise to polarisation problems in conductivity measurements. In spite of these pitfalls, the total DC conductivity has routinely been used to characterise polymer electrolytes and a threshold value of approximately 10−5 S cm−1 has been used as the criterion for possible application purposes. A large number of materials reach such a conductivity value between room temperature and 100 °C (Aoki et al., 2006; Blonsky et al., 1986; Chiang et al., 1985; Dupon et al., 1984; Egashira et al., 2008; Fang et al., 1988; Geiculescu et al., 2004; Gray, 1990; Henderson, 2007; Marzantowicz et al., 2007; Xi et al., 2004; Y. Zhang et al., 2006). Several aspects govern the magnitude of the conductivity, namely the degree of crystallinity, the salt concentration, and the nature of the salt and polymer (Martins and Sequeira, 1990). Temperature and pressure have been used in a great number of studies as variables in investigations of these conductivity aspects. With blocking electrodes, there is neither a source nor a sink for mobile ions. The migration of the ions under the influence of the electric field leads to an enrichment of the mobile species in the electrolyte region adjacent to one electrode and depletion near the other electrode. The ionic motion is then opposed by a chemical potential gradient and when, after a short time, this has increased sufficiently to counterbalance the electric field, the migration stops. The cell is then said to be concentration polarised. Since the calculation of the cell conductance depends on the assumption of ohmic behaviour, which pertains only at the instant of initial application of the applied voltage, it is difficult to determine conductance accurately from DC measurements on test cells with blocking electrodes. The mechanism by which current initially flows across a test cell with blocking electrodes under the influence of an applied voltage can be visualised as follows. Neither the electronic charge carriers within the electrode nor the ionic charge carriers within the electrolyte can cross the electrode–


24

Polymer electrolytes Cg

Cdl (a)

Rb

Cdl

Cg

Rct

Rb (b)

1.4 (a) Equivalent circuit for a test cell with blocking electrodes. (b) Equivalent circuit for a test cell with non-blocking electrodes.

electrolyte interface. The applied voltage causes a build-up of negative charge on the electrode side of one of the interfaces. This induces a corresponding build-up of positively charged ions on the electrolyte side of the same interface. A similar ‘double layer’ of separated charges, but with the signs reversed, is set up at the other interface. The electric field that causes ionic motion within the electrolyte is that caused by the charge concentration gradient from the electrolyte side of the interface. The double layers each have infinite resistance to the direct passage of charge, but have a capacitance term that results from the layers of separated charge. The double layer can be visualised as a parallel-plate capacitor, and since ε = ε0εrA/d

[1.5]

(where ε0 is the permittivity of free space, εr is the relative permittivity, A is the area of the plates and d is the distance between them), the capacitance term, C, is relatively large (of the order of µF) since the separation, d, is very small. The equivalent circuit to the cell is displayed in Fig. 1.4(a). The so-called geometric capacitance, Cg, that appears in parallel is a consequence of the charges on the electrodes themselves. This term would be present even if the electrolyte were to be removed. Its magnitude is small compared with the double-layer capacitance, Cdl, because the intercharge separation is much greater. The electrolyte conductance, G, is the reciprocal of the electrolyte bulk resistance, Rb.



25

In the case of measurements with non-blocking electrodes, there is both a source and a sink for the mobile ionic species. As in the case of blocking electrodes, however, DC measurements do not easily provide accurate values for Rb. This is because, although ions can cross the boundary, electrons and/or holes cannot. Since these are the conducting species within the electrodes, the charge-transporting process has to be transferred to and from electrons and ions at the two interfaces. The kinetics of this charge-transfer process is not infinitely facile and, consequently, there is a resistance, Rct, associated with the charge-transfer process. This results in a build-up of charge on either side of the interface, which produces an associated, and quite large, charge-transfer capacitance, in parallel with the resistance. The equivalent circuit for this situation is similar to that for blocking electrodes, but the double-layer capacitance is shunted by the charge-transfer resistance, as shown in Fig. 1.4(b). A DC measurement, therefore, does not give the value of Rb directly. It has been observed for both hydrated and dehydrated salts that isotherms of log (σ/S cm−1) against stoichiometric number, n, are usually irregular (Patrick et al., 1986; Yang et al., 1986a). Consequently, isopleths of log (σ/S cm−1) against 1000 K/T vary in gradient, and may cross each other. In some systems, e.g. PEOn:Ca(ClO4)2·6H2O, the conductivity maximum appears to lie at concentrations of salt that are more dilute (i.e. n > 18) than those studied. This inverse relationship between conductivity and salt content is also observed (Poulsen et al., 1985) for PEOn:LiClO4 and other polymer electrolytes. In Table 1.4, the compositions for maximum conductivity, nmax, are tabulated as a function of temperature for a range of systems. Inspection of Table 1.4 reveals that nmax varies substantially with the anion, cation and the temperature, from below 8 to above 16. The observed limits of the variation, i.e. 4–24, may be artificially narrow and merely reflect the truncated range chosen by individual authors. Values of nmin are presented in Table 1.5. Although also irregular, it can be seen that, except for some hydrated salt systems near room temperature, nmin < nmax. It is clear that nmax and nmin do not correlate with cationic radius (or charge density) although at certain temperatures it has been noted that the conductivity at fixed n decreases with increase in cationic radius (Patrick et al., 1986). As can be seen in Table 1.6, the conductivity results for PEO:X(ClO4)2·6H2O complexes, where X = Mg2+, Zn2+, Ca2+, Sr2+, are also in accordance with that observation. There appears to be some ambiguity regarding the choice of the anion that will yield maximum conductivity with a given cation, and vice versa. Further significant observations can be reported, however, as discussed below. Patrick and co-workers (1986) have found that the polyether electrolytes of divalent cation perchlorates (ClO4−) are generally better conductors than the corresponding thiocyanates (SCN−). Table 1.6 shows results


26


Table 1.4 PEOn : salt ratios for maximum conductivities Temperature (° C) Salt

20

30

40

50

60

70

80

90

100

110

120

Ca(ClO4)2·6H2O Ca(SCN)2·3H2O Cu(ClO4)2 MgCl2 Mg(ClO4)2·6H2O Mg(SCN)2·4H2O PbBr2 Sr(ClO4)2·6H2O ZnCl2 Zn(ClO4)2·6H2O PbI2 PbCl2

12 9

12 9 8 16 12 18 8 12 8 18 24 16

18 12 8 16 12 18 8 12 8 18 24 16

18 12 8 16 12 18 8 12 8 18 24 16

18 15 8 16 12 18 8 12 8 18 24 16

18 15 8 16 15 18 8 12 8 18 24 16

18 15 8 16 18 12 8 12 4 18 24 16

18 15 8 16 18 12 8 12 4 18 24 16

18 15 8 16 18 12 8 12 4 18 24 16

18 15

18 15

16 15 12 8 12

15 15 12 8 12

18 24 16

18 24 16

12 15 12 18

Table 1.5 PEOn : salt ratios for minimum conductivities Temperature (° C) Salt

20

30

40

50

60

70

80

90

100

110

120

Ca(ClO4)2·6H2O Ca(SCN)2·3H2O Cu(ClO4)2 Mg(ClO4)2·6H2O Mg(SCN)2·4H2O Zn(ClO4)2·6H2O ZnCl2 MgCl2 PbBr2 PbI2

15 6

15 6 4 6 6 9 12 4 16 16

9 6 4 6 6 9 12 4 16 16

9 6 4 6 6 9 12 4 16 16

9 6 4 6 9 9 12 4 16 8

9 6 4 6 6 9 12 4 16 8

6 6 4 6 6 9 12 4 16 8

6 6 4 6 6 9 12 4 16 8

6 6 4 6 6 9 12 4 16 8

6 6 4 6 6 9 12 4 16 8

6 6 4 6 9 9 12 4 16 8

15 6 9

for the Mg and Ca systems confirming this observation, which seems to be also valid for other salts (e.g. the conductivity of PEO:Mg(ClO4)2·6H2O complexes is higher than the conductivity of PEO:MgCl2 complexes). The fact that PEO:Zn(ClO4)2.6H2O complexes (Patrick et al., 1986) present lower maximum conductivities than PEO:ZnCl2 complexes (Abrantes et al., 1986) may be due to the different preparation procedures of the polymers (different water content trapped in the polymer films), apart from the larger size of the ionic carriers. The hypothesis of ion-pairing between ClO4− and Zn2+ can also be advanced, but requires experimental confirmation.



27

Table 1.6 Maximum conductivities of the PEO : salt complexes, given in (−log σ) values Temperature (° C) Salt

20

30

40

50

60

70

80

90

100

110

120

Ca(ClO4)2·6H2O Ca(SCN)2·3H2O Cu(ClO4)2 MgCl2 Mg(ClO4)2·6H2O Mg(SCN)2·4H2O PbBr2 Sr(ClO4)2·6H2O ZnCl2 Zn(ClO4)2·6H2O PbI2 PbCl2

8.0 8.8 4.7 9.4 6.4 9.2 9.0

7.2 8.2 4.3 8.8 5.6 8.3 8.6 7.5 4.7 6.7 9.6 9.1

6.2 7.5 4.0 8.0 5.0 7.0 8.2 6.8 4.4 5.8 9.2 8.7

5.2 6.6 3.4 7.2 4.7 6.2 7.8 6.3 4.1 4.9 8.4 8.2

4.8 5.8 3.2 6.5 4.3 5.9 7.4 5.7 4.0 4.6 7.4 7.6

4.7 5.5 3.0 5.7 4.1 5.6 7.0 5.3 4.0 4.3 6.5 6.7

4.4 5.3 3.0 5.5 4.0 5.4 6.0 5.0 3.7 4.0 6.4 6.5

4.1 5.2 2.9 5.2 3.9 5.1 5.9 4.7 3.6 3.8 6.3 6.4

4.0 5.0 2.8 5.0 3.8 5.0 5.8 4.4 3.4 3.8 6.2 6.3

3.9 4.9

3.8 4.8

4.9 3.7 4.8 5.7 4.2

4.7 3.6 4.7 5.5 3.9

3.7 6.1 6.3

3.6 5.9 6.2

5.3 7.6 9.4

As far as the effect of the different anions on the conductivity is concerned, it has been shown that (Abrantes et al., 1986; Huq et al., 1987; Yang et al., 1986a,b): σ(PEO:PbBr2) > σ(PEO:PbCl2) > σ(PEO:PbI2)

(T < 60 °C)

σ(PEO:PbBr2) = σ(PEO:PbI2) > σ(PEO:PbCl2)

(T = 60 °C)

σ(PEO:PbI2) > σ(PEO:PbCl2) > σ(PEO:PbBr2)

(T = 70 °C)

σ(PEO:PbBr2) > σ(PEO:PbI2) > σ(PEO:PbCl2)

(T > 70 °C)

Comparing these sequences with the values of the ionic radii for the anions Cl−, Br− and I−, it can be concluded that there is no relationship between maximum conductivities and the ionic radii. If we consider a fixed value of n, we can see that the increase in conductivity is a direct function of the increase of the cationic mobility. For example, the ionic radii for Sr2+, Ca2+, Zn2+ and Mg2+ decrease in the order Sr2+ > Ca2+ > Zn2+ > Mg2+ and the opposite should occur with the corresponding mobilities. Indeed, Table 1.5 shows that at 30 °C, for example, the maximum conductivity for the corresponding complexes increases in the order PEO12:Sr(ClO4)2·6H2O < PEO12:Zn(ClO4)2·6H2O < PEO12:Mg(ClO4)2·6H2O. An obvious exception to this rule is shown by complexes such as PEO:Cu(ClO4)2·6H2O and PEO:Mg(ClO4)2·6H2O. The ionic radius of Cu2+ is higher than that of Mg2+, so the σMg complex should be higher than the σCu complex. However, this is not the case, which can be easily understood


28


on the basis of the fact that Mg2+ mobile ions are surrounded by their hydration shells. Moreover, the Mg2+ ions are apparently trapped in strong electrostatic bonds with the ether linkages on the polymer chains. Papke et al. (1982) have found that PEO–alkali metal salt complexes exhibit no detectable crystallisation for the larger cations such as Rb+ and Cs+, while Na+ salt complexes showed a high degree of crystallinity. A similar observation was noticed for PEO:metal salt complexes with larger anions. Furthermore, Watanabe and Ogata (1987) also reported that the larger the cation radius, the higher the conductivity values for PPO:SCN (alkali metals) electrolytes. Accepting that this phenomenon can also be applied to PEO:divalent metal salt, we can assume that large cations and/ or anions favour the production of complexes possessing large proportions of amorphous regions, which explains the higher conductivities obtained experimentally. In fact, the suppression of crystallisation improves markedly the transport properties even at room temperature where linear PEO is most crystalline, as we have shown (Araújo and Sequeira, 1993) by synthesising cell-amorphous interpenetrating polymer networks (Fig. 1.5). For the same complex, at a certain temperature, and variable n, we can observe that the conductivity generally increases as the salt concentration increases (n decreases), then it attains a maximum for intermediate n values, and finally it decreases for higher salt concentrations. This behaviour is typical of, for example, PEOn:Cu(ClO4)2, PEOn:ZnCl2 and PEOn:MgCl2 complexes. High salt concentrations have four main effects, they: (a) favour the formation of the PEO:salt complex, leaving a film depleted in free PEO (Yang et al., 1986b); (b) hinder the complete salt dissociation; (c) facilitate the

1 95 °C 85 °C 75 °C 65 °C 55 °C 45 °C 35 °C 25 °C

Log (σ /S cm–1)

0 –1 –2 –3 –4 –5 –6 0

8

16

24

32

40

n

1.5 Conductivity as a function of n in PEOn:Mg(ClO4)2. Reprinted with permission from Arau´jo and Sequeira, 1993. Copyright 1993, The Electrochemical Society.



29

formation of ionic multiplets, such as triplets (X− M+ X−, M+ X− M+) and quartets (M+ X− M+ X−), which are less mobile than free ions and, in addition, decrease their effective concentration (Ratner, 1987); and (d) diminish the number of vacancies available for ionic transport (Sequeira, 1983a; Sequeira and Hooper, 1983b,c). These four effects explain the low conductivities observed at higher salt concentration. The low conductivities measured at low salt concentrations (n ≥ 12) are to be expected due to the reduced number of available charge carriers. Cation–anion interactions inside the polymer electrolyte, such as the phenomenon of ‘ion-pairing’ (Papke et al., 1981b), can explain conductivity variations over the whole range of temperatures and stoichiometries used. Salts whose anions are large, easily polarisable and weak Lewis bases, i.e. bases whose conjugate acids are strong, have a reduced tendency for ionpairing, and their incorporation in the polymeric material leads to higher conductivity complexes. Thus, the concentration dependence of the conductivity is a very complex function, as it is illustrated in Fig. 1.6, which relates an amorphous interpenetrating polymer network of PEO–LiClO4 (Hudson and Sequeira, 1993, 1995). A number of studies have been carried out where pressure rather than temperature has been used as a variable in conductivity measurements. Archer and Armstrong (1980) found that the bulk conductivity of a PEO4.5:LiCF3SO3 electrolyte decreased with increasing pressure over the range 60 to 300 MPa. When a cell was held at a constant pressure and

+

–4.0

+

Log σ (S cm–1)

+

+

+

–4.5 +

+

+

–5.0

+ –5.5

50

100 n

150

200

1.6 Variation in conductivity with salt concentration for an amorphous interpenetrating polymer network containing PEOn:LiClO4. Reprinted with permission from Hudson and Sequeira, J. Electrochem. Society 142(12), 4013–4017. Copyright 1995, The Electrochemical Society.


30


temperature, the conductivity was again seen to fall with time. These effects were attributed to some form of structural change at higher pressures and may possibly involve increased crystallisation (Vincent, 1987a). Measurements of transport processes in crystalline solids as a function of pressure allow the activation volume, V*, for the conduction process to be evaluated. It is defined by (Flynn, 1972): V* =

( )

δG T δP

[1.6]

where δG represents the free energy change associated with the basic steps in the conduction mechanism. Values of V* can then be used to test theoretical models of the process. This approach has had application in a wide number of solid systems and has been applied to polymer electrolytes by Bridges and Chadwick (1988), Chadwick et al. (1983), Greenbaum et al. (1988) and Wintersgill et al. (1987), among others. Both isobaric and isothermal conductivity studies have been carried out. In determining the most appropriate method of association of the free energy change with the ionic conductivity, two options were discussed, one described by an Arrhenius expression, the other by a free volume expression. Interpreting trends in activation volumes with respect to charge transport is complex. In particular, it has been proposed that factors such as the possession of a permanent dipole by the anion (e.g. CF3SO3− or SCN−) may have a significant effect on its mobility (Lee and Crist, 1986) and thus in its contribution to conductivity.

1.4.3 Alternating current conductivity The problems of concentration polarisation which complicate DC measurements are largely avoided if AC is used instead. AC studies are similar to the DC technique in that the ratio of voltage to current is measured. For DC, this ratio provides the value of the resistance, R, measured in ohms (Ω). For AC the ratio gives an analogous quantity, the impedance, Z, also measured in Ω. The impedance contains four main contributions; these are from resistance, capacitance, constant phase elements (CPE) and inductance. The latter is unimportant for polymer electrolytes although it can play a role in other electrochemical applications of polymers. Measurement of the impedance as a function of frequency is called impedance spectroscopy. The detailed interpretation of this variation provides important information about the contributions to the test-cell electrochemical behaviour of its various components: interfaces, intra-electrolyte boundaries and the bulk electrolyte itself. This information is normally extracted by finding an idealised equivalent circuit, the calculated fre-



31

quency-dependent behaviour which mimics the actual behaviour of the test-cell. The components (resistors, capacitors, etc.) of the equivalent circuit which are networked together in some combination of series and parallel connections can be identified with parts of the test-cell. The frequency dependence can be displayed in several ways, each of which emphasises a different frequency regime or a particular contribution. To appreciate these, it is necessary to recognise the fact which is not particularly palatable to the non-mathematically inclined reader, that impedance is a complex quantity. In practice, this means that it is made up of two parts: one called real, which is labelled Z′, and the other called imaginary and labelled as Z″. The contributions to Z″ are all terms which involve the quantity j, the square root of minus one (√−1). Z′ and Z″ are related to Z by: Z = Z′ − jZ″

[1.7]

The various resistors, capacitors, etc., which are connected together in series and parallel combinations to form the equivalent circuit that corresponds to the test-cell, contribute to the real and imaginary parts in a way that is quite easy to calculate. The frequency-dependent information can be displayed in the form of Z′ or Z″ as functions of the angular frequency, ω, or of log (ω). An alternative display, much favoured in the polymer electrolyte field, is the so-called complex plane representation, in which −Z″ (y-axis) is plotted against Z′ (x-axis). This display is characterised for polymer electrolytes by arcs, which have the form of depressed or flattened semicircles, and straight lines, inclined to the x-axis, which are usually called titled spikes. The advantages of this form of data presentation are that each arc or spike is characteristic of a particular region of the cell and that the bulk resistance can be easily read from the graph. A disadvantage is that the frequency of each measurement point is not immediately apparent. The simplest test-cell that can be envisaged consists of two blocking electrodes in contact with the polymer whose conductance is to be measured. The three-component equivalent circuit is shown in Fig. 1.4(a). The polymer acts as a resistor, Rb, which is in series with the double layer capacitor, Cdl, at the interface, and in parallel with the geometric capacitance, Cg, as discussed in Section 1.4.2. This simple network is dealt with by evaluating the parallel term first to give an impedance Zp, which is then added to the double-layer impedance. The RC parallel combination gives Zp =

R (1 − jωCg R ) 1 + (ωCg R )

2

and the total impedance Z is now


[1.8]

32

Polymer electrolytes Z=−

R (1 − jωCg R ) j + 2 ωCdl 1 + (ωCg R )

[1.9]

so that Z′ =

R 2 1 + (ωCg R )

[1.10]

and ⎡1 + (ωCg R ) (1 + Cdl Cg )⎤⎦ Z ′′ = ⎣ ωCdl ⎡⎣1 + (ωCg R )2 ⎤⎦ 2

[1.11]

This is already a very complicated expression, although the equivalent circuit was the simplest that could be chosen to represent the physical situation of a conducting cell. The impedance plot is what might be intuitively expected from bringing together a series combination, which produces a vertical spike giving a real-axis value of R, and a parallel combination, giving a semicircle tending to the value of R at low frequency. The closer the values of Cdl and Cg (i.e. the thinner the electrolyte), the more the semicircle and the spike tend to merge, as shown in Fig. 1.7. If the electrodes are non-blocking, then the Cdl is now shunted in parallel by a charge-transfer resistance, as shown in Fig. 1.4(b) and discussed in Section 1.4.2. Evaluation of the impedance expression for this equivalent circuit produces two semicircles. The high-frequency semicircle is related to the bulk electrolyte and the low-frequency semicircle, which is more distant from the origin, arises from interfacial processes. The bulk resistance is the Z′ value at the high-frequency end of the interface semicircle.

–Z ˝

Cd/Cg = 25

Cd/Cg = 1000 Z´

1.7 The effect of the Cdl/Cg ratio on the impedance plot.


Introduction to polymer electrolyte materials 2.4 E4 2.4 E5

– Im Z (Ω cm2)

6.25 E3 – Im Z (Ω cm2)

2

– Im Z (Ω cm )

42 °C

0

70 °C

0

(a)

2.0 E5 Re Z (Ω cm2)

51 °C 0 (b)

128.5 °C

0

58 °C

24 °C 0

33

2.0 E4 Re Z (Ω cm2)

114 °C 0

5.00 E3 Re Z (Ω cm2)

(c)

1.8 Measured impedances for the cell Ni/PEO6:NiCl2/Ni at seven different temperatures (from Plancha et al., 1992a). Reprinted from Solid State Ionics, 58, Plancha MJC, Rangel CM and Sequeira CAC, AC conductivity of polymer complexes formed by poly(ethylene oxide) and nickel chloride, page 6, Copyright (1992), with permission from Elsevier.

Complex impedance plots for the Ni/PEO6NiCl2/Ni cell recorded at seven different temperatures are shown in Fig. 1.8. At temperatures up to 50 °C, the system exhibits two semicircles corresponding to the electrolyte impedance (parallel combination of the geometrical capacitance, Cg, and the electrolyte resistance, Rb) and to the interfacial impedance (charge transfer resistance, Rct, and double layer capacitance, Cdl, coupling). For temperatures above 50 °C, the high frequency semicircle does not appear; this is due to the fact that the corresponding characteristic frequency is higher than 65 kHz, the limit of the frequency range of the apparatus used. Rb and Rct values also decrease with increasing temperature. For temperatures from 70 °C up to 130 °C, the second semicircle, characteristic of intermediate frequencies, almost disappears and a diffusion process is evidenced by an arc with a 45° high-frequency branch, curving towards the real axis at lower frequencies (see Fig. 1.8c) (Plancha et al., 1992a). The value of ω at the top of the semicircle related to the bulk electrolyte is given by ω = 1/CgR

[1.12]

and increases substantially as the resistance falls (for example, with the temperature of the polymer sample), modifying the usual aspect of the impedance curves for polymer electrolytes. A feature of the impedance plot that is particularly pronounced for polymer electrolytes is the depression


34


of the semicircle and the tilting of the spike. No combination of resistors and capacitors (or inductors) will produce either the depression or the tilting, but both of these are natural consequences of using CPE. CPE are hybrids between a resistor and a capacitor with an impedance of the form: Zcpe = K [cos(pπ/2) − j sin(pπ/2)]/ωp

[1.13]

where 0 ≤ p ≤ 1 and K is a constant. For a simple circuit consisting of a resistor and a CPE in series, the Z plot is represented by a line, inclined at an angle pπ/2 to the Z′ axis and contacting this axis when Z′ = R, as shown in Fig. 1.9(a). For a resistor and a CPE in parallel, a depressed semicircle is produced, as shown in Fig. 1.9(b). This cuts the Z′ axis at the origin and at Z′ = R. Because all angles subtended within a semicircle are right angles, the point Z′ = R lies vertically above the far end of the diameter of the semicircle. The length of this diameter, d, can be seen from Fig. 1.9(b) to be given by: d = R/sin(pπ/2)

[1.14]

–Z ˝

CPE pπ 2

R Z´

(R,O) (a) –Z ˝

(R,R) CPE

T X d 2

(R,O) Z´

C

R (b)

1.9 Impedance representation of the combination of a resistor and a constant phase element (a) in series and (b) in parallel.



35

and the distance, CT, from the centre to the ‘top’ of the semicircle is d/2. The height of the top of the semicircle, XT, is clearly CT − XC, and since XC = [d cos(pπ/2)]/2

[1.15]

then XT =

d d R [1 − cos ( pπ 2 )] − cos ( pπ 2 ) = 2 2 2 sin ( pπ 2 )

[1.16]

The undepressed semicircle resulting from an RC combination in parallel obviously has a top at the point (R,R) and so the depression is given by R − XT.

1.4.4 Direct current electronic conductivity To determine the transport number of the electrons and electron holes in a polymer electrolyte, which may be a factor of 10−3 or less than that of the ions, the DC polarisation method of Wagner may be used. In this method, the electrolyte Pn:MX is sandwiched between an electrode of the parent metal, M, which acts as reversible electrode, and an ion-blocking electrode such as graphite or platinum. The ion-blocking electrode can exchange electrons with Pn:MX but not ions. A DC potential E, below the decomposition potential Ed, is applied to the symmetric cell. The positive pole is on the ion-blocking electrode (see Fig. 1.10). Because there is no source or sink of M or X at the ion blocking electrode, at steady state only electrons and holes migrate through the sample of Pn:MX. For an electrolyte in which tion = tcation = 1, the relevant equation is: I ss = I − + I + =

+

{

( )

( ) }

ART 0 ⎡ EF ⎤ EF + σ +0 ⎡exp σ − 1 − exp − − 1⎤ ⎢ ⎥ ⎢ ⎥⎦ ⎣ ⎣ LF RT ⎦ RT

Pn : MX

M C+ = C+0 C– = C0–

Graphite

[1.17]

–

C+ = C+0 exp (EF/RT) C– = C0– exp (–EF/RT)

1.10 Wagner’s DC polarisation cell. The negative electrode is a reversible electrode and the activity of M is unity at the M|Pn:MX interface. A DC potential E is applied so that the activity of M at the Pn:MX|ion-blocking electrode interface is exp(−EF/RT) at steady state. Boltzmann statistics for an ideal solid solution are assumed for the electrons and electron holes in Pn:MX. The concentration of electrons and electron holes are denoted by C− and C+, respectively. The superscript zero denotes that the concentration in Pn:MX is at equilibrium with M.


36


Iss

0 0

E

1.11 Schematic diagram of Iss versus E for a polarised cell of the type shown in Fig. 1.10.

where Iss is the steady-state electronic current, I− and I+ are the current due to electrons and holes, respectively, F is the Faraday’s constant, σ0− and σ0+ are the partial conductivities due to electrons and holes, respectively (the superior zero denotes equilibrium with pure metal, M), E is the applied voltage, A is the cross-section area of the sample and L is its thickness. A schematic diagram of the Iss–E curve for such a polarised cell is shown in Fig. 1.11. Generally, the equation for Iss may be simplified, yielding

( )

I ss ART ⎡ 0 EF = σ + exp + σ −0 ⎤ ⎦⎥ 1 − exp ( − EF RT ) LF ⎣⎢ RT

[1.18]

from which a plot of Iss/[1 − exp(−EF/RT)] versus exp(EF/RT) yields, in principle, the values of σ0− and σ0+. However, within each volume element there is equilibrium between the electrons and electron holes; that is, the intrinsic constant, k, is satisfied. Hence, the product C− C+ is very small and σ0−, σ0+ and exp(−EF/RT) are small. Accordingly, the plot from equation [1.18] rarely yields a positive slope and positive intercept within the experimental uncertainty of the data. The use of AC measurements to obtain the total conductivity and the DC polarisation cell to obtain the electronic conductivity allows the separation of ionic and electronic conductivity whose values may differ in many orders of magnitude. A key feature of the Wagner cell methodology is that the activity of the mobile ionic species is controlled at the reversible electrode, and for many polymer electrolytes both anions and cations conduct (Abe et al., 2004; Chandrasekhar, 1998; Hu et al., 2007; Itoh et al., 2009a; Judeinstein et al., 2008; Martins and Sequeira, 1990; Plancha et al., 1992a; Rajendran et al., 2008; Saito et al., 2003; Singh and Bhat, 2004; Stephan, 2006; Xi and Tang, 2004; B. Zhang et al., 2006). It is, therefore, only for polymer electrolytes in which ionic transport is effectively entirely either anionic or cationic that the use of the Wagner technique can be usefully contemplated. An alternative to the Wagner



37

polarisation technique is the use of transient methods. Electrons and holes respond much more rapidly to fluctuations in an applied potential than ions do. A cell of the type M|electrolyte|M has a DC voltage applied across it and the current–time characteristics are monitored as a rapid voltage pulse is superimposed. The response will be rapid for a sample with a high electronic conductivity and slow for a good electronic insulator.

1.4.5 Ionic conductivity There are many AC and DC techniques for probing the transport of charged and neutral species. Among them, we can refer to the electrolysis method, the concentration cell techniques, and the ionic polarisation cell technique. The electrolysis method consists of studying the changes in the amount of material in the anode and cathode compartments of a cell. For liquids, this is done by the Hittorf method in which concentration changes are measured by titration, subsequent to the passage of a known total amount of charge. The technique has been applied to low molar mass polymer electrolytes by Cameron et al. (1989), among others. For solids, this type of technique is known as the Tubandt method, and involves the careful weighing of easily separated, non-adhering discs of electrodes and electrolyte material (Linford and Hackwood, 1981). This type of polymer material for which adhesion between the cell layers has been successfully avoided belongs to the group of network crosslinked materials studied by Cheradame and Niddam-Mercier (1989). In both the Hittorf and Tubandt methods, it is necessary to verify that the electrolyte compartment remains unchanged. The concentration cell techniques involve the use of concentration cells. Bouridah et al. (1986) measured the electromotive force (EMF) of suitable cells under conditions where no current was passed, and obtained the anion transference number from τ− =

−FdEcell RTd ln a

[1.19]

where a is the activity of the salt. Such measurements require an independent determination of the variation of the activity with the concentration. To avoid this, later studies on the voltages of similar cells with and without transference were carried out (Bouridah et al., 1988). The ionic polarisation cell technique involves transport under a chemical potential and electrical gradient and, therefore, includes the contributions of both charged and neutral species. The AC procedure has been suggested by Sorensen and Jacobsen (1982) and is based on the theory of MacDonald (1973, 1974) which involves analysis of the impedance spectra of the symmetric cell M|Pn:MX|M.


38


Rb

Re

Z ˝ (ohms)

Zd

Cb

Ce

f

Rb

Rb+Re

DC limit

Z´ (ohms)

1.12 Idealised equivalent circuit and complex plane plot for a polymer electrolyte cell with electrodes that are non-blocking to the cation. Re = electrode resistance; Ce = electrode capacitance; Rb = bulk (or electrolyte) resistance; Cb = bulk (or electrolyte) capacitance; Zd = diffusion-controlled impedance.

For the idealised case, the equivalent circuit and response of Fig. 1.12 may be anticipated. The high-frequency semicircle is due to cation migration through the bulk and dielectric polarisations. At lower frequencies, a semicircle arises as a result of the charging and discharging of the electrode– electrolyte interface and reactions of the electroactive ion at the electrode interface. At the lowest frequencies, the current is affected by concentration gradients that give rise to diffusion in the electrolyte, and a skewed semicircle that approaches the limiting DC value of the cell impedance is observed. For fully dissociated electrolytes, the cation transference number τ+ can be evaluated by comparison of the width of the skewed semicircle, the diffusion-controlled impedance, Zd, with the value of the bulk resistance, Rb: τ+ = 1/(1 + Zd/Rb)

[1.20]

The AC technique has been by far the most widely used method for obtaining transference numbers on polymer electrolytes (Plancha et al., 1991, 1992b,c, 1993a,b, 1994, 1997c, 1998, 1999; Sorensen and Jacobsen, 1982; Yang et al., 1986a,b), but again the analysis is relevant only when mobile associated species are not present. In addition, electrode phenomena such as finite electrode kinetics or formation of passivating layers are not con-



39

sidered and, as suggested by Fauteux (1985), the diffusion responsible for the low frequency arcs may not be a bulk process but one of the passivating films on the electrode surface, which are generally observed when polymer electrolytes are in contact with metallic lithium. The preceding considerations have dealt with simpler situations. The effects of non-ideality and higher associated charged clusters have yet to be rigorously analysed. The extent to which the preceding techniques will lead to results deviating from the real situation remains unknown.

1.5

Applications in practical devices

The future prospects for polymer electrolytes look promising because it has been appreciated that they form an ideal medium for a wide range of electrochemical processes. Other than primary and secondary batteries, and high and low temperature fuel cells, practical applications for polymer electrolytes that are under consideration include electrochromic devices, modified electrode/sensors, solid-state reference electrode systems, supercapacitors, thermoelectric generators, high-vacuum electrochemistry and electrochemical switching. Device applications that have been the main driving force behind the development of polymer electrolytes are treated hereafter.

1.5.1 Batteries Recently there has been a phenomenal growth in electronic devices and they are all demanding in terms of electricity supply. The only way of achieving this in practice in the foreseeable future is by a rechargeable battery. Batteries will play a key role in the development of new, self-powered medical implants, in ‘hybrid’ vehicles that run alternately on combustion engines and batteries, and in energy storage for renewable power sources such as wind and solar. Each type of battery will require different characteristics in terms of the storage capacity, the number of times the battery can be discharged and recharged, and the cost of the materials. The concept of the rechargeable battery is elegantly simple. Two electrodes are separated by an electrolyte that is an ionic conductor and electronic insulator. Charging the battery involves applying an electrical potential to force cations – lithium, say – up an energy gradient into the anode, where the ions are intercalated within a crystalline lattice. When the two electrodes are connected via an external circuit this pent-up chemical energy is released: the positive ions drift across the electrolyte and fall into an energy well within the cathode; at the same time, electrons flow to the cathode via the external circuit, providing an


40


electric current. Within this system there are several ways to improve the battery, depending upon the priority required for the given application. Most water-based battery systems, such as nickel–cadmium, where the electrolyte is in an aqueous solvent, are limited to around 1.5 to 2 volts, whereas a lithium battery can give 3 to 4 volts per cell. Since lithium batteries can store more energy than other systems, most research effort on rechargeable batteries is focused on lithium technology (Koksbang et al., 1994; Sequeira, 1983b, 1987; Sequeira and Hooper, 1985; Sequeira and Marquis, 1986; Stephan, 2006). Rechargeable lithium ion batteries were introduced commercially by the Sony Corporation in the mid-1990s. A lithium cell that sits in a laptop computer typically consists of a graphite anode and a cathode whose ‘active’ ingredient is a lithium transition metal oxide, such as LiCoO2, with the two electrodes separated by an electrolyte of LiPF6 dissolved in an organic solvent, or an SPE (Aravindan et al., 2008; Itoh et al., 2009b; Jeon and Kwak, 2006; Kang et al., 2009; Koksbang et al., 1994; Meyer, 1998; Reiter et al., 2009a; Seki et al., 2005; Stephan and Nahm, 2006; T. Zhang et al., 2008). There are two ways to increase the amount of energy that such a cell can store. One is to increase the lithium mobility in the system – the more ions that can make their way from the anode to the cathode, the greater the storage capacity of the battery. This means materials that can store more lithium per unit weight and volume. The other approach is to increase the cell voltage. So, materials chemists are looking for new electrode materials that can store more lithium at higher voltage in a way that is inexpensive, safe and not toxic. The system must also be capable of being cycled – charged and discharged – hundreds or thousands of times with no appreciable loss of efficiency. Since the introduction of the lithium cell, most improvements to the technology have relied on engineering rather than chemistry. For example, the positive electrode consists not only of the lithium transition metal oxide, but also of conducting additives and binders. Over the years, new manufacturing processes have allowed the proportion of the inactive components of the electrodes to be lowered, giving way to a greater quantity of the lithium-containing compound. But this has effectively reached its limit and new approaches are needed. In current lithium batteries based on LiCoO2 cathodes, only ‘half’ a lithium ion is available for shuttling between the electrodes. Most recent work worldwide has focused on how to get that extra half lithium into the system. However, this will only double the charge stored and more radical approaches are required to go further. One approach is to remove the intercalation host altogether, and let the lithium react directly with oxygen in the air. The lithium entering the porous electrode reacts directly with gaseous oxygen in the air contained within the pores to form lithium oxide.



41

Early results have shown that such an oxygen electrode can store significantly more charge than conventional systems relying on intercalation chemistry. Storage is expressed in milliampere hours per gram (mA h g−1), and LiCoO2 can typically store 140 mA h g−1, with a theoretical maximum of some 300 mA h g−1. Greater storages can be obtained by combining chemistry and novel manufacturing techniques. For example, researchers at St. Andrews (Scotland) have synthesised LiNi1/3Co1/3Mn1/3O2 using a new route that results in a macroporous material delivering a higher capacity at a higher rate than is currently possible (Shaju and Bruce, 2006). Given that the ions shuttle between one electrode and another, the ions that emanate from the cathode must have a home to go to in the anode. In other words, the anode’s storage capacity must be able to match that of the cathode. So far this has not been a problem: graphite can store around 300 mA h g−1. However, if the cathode’s storage capacity starts to outstrip this, the imbalance will need to be addressed. Sony Corporation has already improved this aspect of the negative electrode by using an alloy of lithium and tin, which can store between 700 and 1000 mA h g−1. Along with the storage capacity, a key characteristic of a rechargeable Li ion battery is just how rechargeable it is, i.e. how many times it can be cycled without appreciable loss of capacity. This is intimately connected with phenomena at the interfaces between the electrodes and the electrolyte (North et al., 1982; Sequeira, 1983c, 1985; Sequeira and Hooper, 1983c,d,e,f; Sequeira et al., 1984). A typical battery system operates by transferring lithium ions from a transition metal oxide host into graphite. During the lifetime of the battery, certain degradation products are collected at the electrode surfaces. These can lead to unwanted reactions, in which the heat generated can in turn cause further degradation, resulting in a process called ‘thermal runaway’. This can bring about potentially hazardous instabilities within the system. Paradoxically, the initial stage of the degradation of the electrolyte is critical in stabilising the electrodes, but excessive subsequent degradation is undesirable. At the surface of the electrodes, lithium ions react with the electrolyte to form a layer of typically lithium fluoride, lithium carbonate or organic lithium compounds. If this layer, called the passivation layer, did not form, the graphite negative electrode would delaminate and eventually disintegrate. The important thing is that this layer must remain permeable to lithium ions during subsequent cycling of the battery. It is a bit like a brick wall with some bricks missing – a supporting structure that still allows lithium to go through. However, the presence of the layer increases the resistance of the battery, which in turn generates heat – something that must be minimised. One way of mitigating thermal runaway is to use additives in the electrolyte so that the initial passivation layer is more stable and less likely to grow upon subsequent cycling. Such additives can include large


42


organic molecules that degrade to form a stable passivation layer; alternatively, complexing agents can be placed at the interface that chemically scavenges ions or molecules within the electrolyte to form a stable film. It is important to minimise the loss of capacity during the lifetime of the battery by improving the efficiency of ion transport in and out of the electrodes. If lithium ions react with products within the electrolyte to cause this layer to grow, some lithium is effectively removed from the process. Ideally, we want the same amount of lithium to move in and out of the electrodes on the thousandth cycle as on the first cycle. If we lose only 1% lithium on each cycle, we would not have much capacity left after a few hundred cycles. Apart from the lithium polymer electrolyte batteries as a whole, other battery technologies have been studied, as reported by several researchers (Abrantes et al., 1985; Araújo et al., 1994a,b; Bruce, 1995a; Gray, 1997; Patrick et al., 1987). Polymer electrolyte batteries are presently produced at a rate of several million units per month and are rapidly replacing the bulkier and less energetic nickel–cadmium and nickel–metal hydride batteries in popular devices, such as cellular phones and computers. In addition, lithium ion batteries are also scaled-up in view of their use in electric vehicles. However, although a commercial reality, polymer electrolyte batteries are still the object of intense research with the aim of further improving their properties and characteristics (Araújo and Sequeira, 1991, 1993, 1995a,b; Araújo et al., 1996b; Bruce, 2005).

1.5.2 Fuel cells Fuel cells constitute an attractive class of renewable and sustainable energy sources that are alternatives to conventional energy sources, such as petroleum, which have finite reserves. Energy generation from petroleum oil and natural gas through combustion in a heat engine being subjected to the Carnot cycle limitation is inherently inefficient and accompanied by environmental pollution. In contrast, a fuel cell is intrinsically energy efficient, non-polluting, silent and reliable. In some embodiments, it is a low temperature device that provides electricity instantly upon demand, and exhibits a long operating life. Energy efficiencies of about 50–70% can be achieved with fuel cells. Fuel cells combine the advantages of both combustion engines and batteries, at the same time eliminating the major drawbacks of both. Similar to a battery, a fuel cell is an electrochemical energy device that converts chemical energy into electricity and, akin to a heat engine, a fuel cell supplies electricity as long as fuel and oxidant are supplied to it. Among the various types of fuel cells developed so far, polymer electrolyte fuel cells (PEFCs) have the advantage of high power densities at relatively low operating temperatures (≤80 °C) (Antolini, 2004) and therefore



43

are considered promising power sources for portable and residential applications. Research and development on PEFCs using hydrogen as the fuel have progressed enormously but their successful commercialisation is restricted due to problems with the safety and storage efficiency of the flammable hydrogen gas (Igarashi et al., 1995; Shukla et al., 2003). In order to overcome these difficulties, liquid methanol can be used instead to fuel PEFCs. Direct use of liquid fuel in a PEFC simplifies the engineering issues, thereby driving down the system complexity and hence the cost. PEFCs that are fed with methanol as the fuel are referred to as direct methanol fuel cells (DMFCs). However, DMFCs possess several limitations in terms of the high cost of the noble-metal electrocatalysts, low open circuit voltage (OCV), and methanol crossover from anode to cathode compartments (Shukla et al., 2003, 2004). Moreover, CO is formed during the inefficient methanol electrooxidation, which leads to poisoning of the platinum anode (Ganesan and Lee, 2005). The problems arising from the use of methanol in DMFCs can be overcome by using other hydrogen carrying materials as the fuel, such as various borohydride (BH4−) compounds. Sodium borohydride (NaBH4), which has a charge capacity value of 5.67 A h g−1 and a hydrogen content of 10.6 wt% (Lee et al., 2006; Santos, 2006, 2009; Santos and Sequeira, 2007) is a good alternative to methanol as a fuel. A PEFC that uses a BH4− compound, usually NaBH4 in aqueous alkaline medium, directly as a fuel is termed as direct borohydride fuel cell (DBFC). Although the concept of the DBFC was first demonstrated by Indig and Snyder (1962) in the early 1960s, Amendola et al. (1999) were the first to report a direct borohydride–air fuel cell that employed an anion exchange membrane (AEM) as electrolyte and exhibited a maximum power density of 60 mW cm−2 at 70 °C. The DBFC reported by Amendola et al. (1999) could use about seven out of a theoretically maximum eight electrons (per BH4− ion), exhibiting high fuel utilisation efficiency. DBFCs are considered attractive energy suppliers, especially for portable applications. DBFC supersedes DMFC in terms of capacity value, electrochemical activity, theoretical OCV (1.64 V for DBFC employing oxygen as oxidant as against 1.21 V for DMFC using oxygen as oxidant) and power performance at room temperature. In addition, use of alkaline electrolytes which feature relatively low corrosion activity, opens up the possibility of applying readily available and lowcost non-precious metal anode catalysts (Santos and Sequeira, 2010; Santos et al., 2007). Both BH4− and its oxidation product, metaborate (BO2−), are relatively inert and non-toxic. BO2− can be recycled to produce BH4− and the techniques involved are under investigation (Fakiogˇlu et al., 2004; Sequeira et al., 2007). One issue that needs further investigation is the development of polymer electrolyte membranes whose structure–property relationships increase their applicability in fuel cell systems, including PEFCs, DMFCs and DBFCs.


44


The evolution of membranes for fuel cell applications started as early as 1959, by General Electric (GE), with the testing of phenolic membranes prepared by polymerisation of phenol-sulphonic acid with formaldehyde. These membranes had low mechanical strength and a short lifetime of 300–1000 hours and showed a power density of 5–10 mW cm−2 (Magnet, 1968). During 1962–1965, GE attempted to improve the power density by developing partially sulphonated polystyrene sulphonic acid membranes (prepared by dissolving polystyrene sulphonic acid in ethanol-stabilised chloroform followed by sulphonation at room temperature). This membrane exhibited a better water uptake and an improved power density of 40–60 mW cm−2, which enabled its application in NASA’s Gemini flights (Bockris and Srinivasan, 1969). Although initial attempts were unfruitful, GE subsequently redesigned their PEFC, and the new model P3, despite malfunctions and poor performance on Gemini 5, served adequately for the subsequent Gemini flights. However, this membrane exhibited brittleness in the dry state. Another approach to improve the mechanical strength and the membrane life was again undertaken by GE in the late 1960s by preparing crosslinked polystyrene–divinylbenzene sulphonic acid membrane/ polymer in an inert matrix. The life of the membrane ranged from 103 to 104 hours and the power density attained 75–80 mW cm−2 (Grott, 1985). The main problem encountered with all the above-mentioned types of membranes was that the proton conductivities were not sufficiently high to reach a power density even as low as 100 mW cm−2 (Wakizoe et al., 1995). In the 1970s, DuPont developed a polymer based on PFSA called Nafion that not only showed a two-fold increase in the specific conductivity of the membrane but also extended its lifetime (104–105 hours). This soon became a standard for PEFC and remains so today. The Dow Chemical Company and Asahi Chemical Company synthesised advanced PFSA membranes with shorter side chains and a higher ratio of SO3H to CF2 groups (Costamagna and Srinivasan, 2001). Table 1.7 (adapted from Smitha et al., 2005) provides a comparison of some commercial cation exchange membranes (CEMs). Many more proton polymer electrolytes have been developed for PEFCs and DMFCs as described in Section 1.2.2 and reported by Chen and Sequeira (2002), Gottesfeld and Fuller (1999), Hartnig et al. (2008), Lee et al. (2006), Neburchilov et al. (2007), etc. Concerning DBFCs, both the AEM and the CEM have served in a majority of fuel cell operating conditions (Atwan et al., 2005; Choudhury et al., 2005; Liu et al., 2003, 2005; Ponce de Leon et al., 2007; Santos and Sequeira, 2009a; Sequeira and Santos, 2009). In particular, Nafion 117 relying on the sodium ion for the ionic conductivity, instead of the hydrogen ion as in the case of the hydrogen PEFC, is often employed (Santos and Sequeira, 2006, 2009b). But development of low-cost polymer membranes with little BH4− crossover continues to be a significant key in bringing DBFC to the level of common usage.



45

Table 1.7 Properties of commercial cation exchange (adapted from Smitha et al., 2005). Reprinted from J Membrane Sci, 259, Smitha B, Sridhar S and Khan AA, Solid polymer electrolyte membranes for fuel cell applications – a review, page 13, Copyright (2005), with permission from Elsevier

Membrane designation

Membrane type

Conductivity (S cm−1) Thickness Gel water at 30 °C and IEC (%) (mequiv g−1) (mm) 100% RH

Asahi Chemical Industry Company Ltd (Japan) K 101 Sulphonated 1.4 0.24 polyarylene

24

0.0114

0.15

25

0.0051

0.15

–

0.0071

0.15

–

–

Ionac Chemical Company, Sybron Corporation (USA) MC 3470 – 1.5 0.6 MC 3412 – 1.1 0.8

35 –

0.0075 0.0114

Ionics Inc. (USA) 61AZL386 – 61AZL389 – 61CZL386 –

2.3 2.6 2.7

0.5 1.2 0.6

46 48 40

0.0081 – 0.0067

DuPont Company (USA) N 117 Perfluorinated N 901 Perfluorinated

0.9 1.1

0.2 0.4

16 5

0.0133 0.01053

Pall RAI Inc. (USA) R-1010 Perfluorinated

1.2

0.1

20

0.0333

Asahi Glass Company Ltd (Japan) CMV Sulphonated 2.4 polyarylene DMV Sulphonated – polyarylene Flemion Perfluorinated –

Note: IEC stands for the membranes’ ion-exchange capacity and RH stands for the relative humidity.

1.5.3 Photoelectrochemical devices A photoelectric energy converter is based on semiconductors as active components. These may be crystalline, polycrystalline or amorphous. They may be inorganic, such as the well-established silicon semiconductor, or organic. Absorption of photons of requisite energy – those with energy exceeding the bandgap of the material – leads to the generation of electron–hole pairs, which can be separated by diffusion or drift, in a sufficiently strong electric field. The electric field is due to the double layer found at the junction of two dissimilar materials. These materials may be the same kind but with different doping, as found for instance in the p–n junction


46


cells constructed from single crystal silicon. They may also be a semiconductor in contact with a metal, as in the Schottky barrier cell. The third class of cells is constructed from the junction of a semiconductor with an electrolyte. These are called photoelectrochemical cells (PECs) (Nozik, 1978). In a PEC, light is incident on the n-type semiconductor/electrolyte junction (photoanode), where light absorption occurs and an electron–hole pair is formed. The pair is separated by the strong electric field found just beneath the semiconductor surface, and the hole is driven towards the interface between semiconductor and electrolyte. Charge transfer to the redox species A contained in the electrolyte results in the oxidation to A+. Conversely, the electron is driven to the metal/electrolyte interface (counter electrode), where the redox species is reduced. No net chemical work is done and we can extract the energy as a current from the cell. The potential advantages of a photoelectrochemical energy converter are: • •

•

the ease of formation; the barrier is formed by contacting the semiconductor with the electrolyte; the possibility of forming high barriers by using a redox couple of extreme potential; these are reflected in high open circuit photovoltages; the possibility of using lower semiconductor materials; polycrystalline materials are easier to use than in solid state photoelectric devices (Hodes et al., 1984).

The first use of SPE containing redox species for use in PECs was in a tandem cell (Skotheim, 1981) and as contacts to silicon (Inganas et al., 1986; Skotheim and Inganas, 1985; Skotheim and Lundstrom, 1982). Other materials have also been used in combination with polymer electrolytes (Cook and Sammels, 1985; Sammels and Ang, 1984; Sammels and Schmidt, 1985). The requirements for application of polymer electrolytes in PECs is, of course, that they can complex desired redox species, that their conductance is high enough (due to high conductivity and/or small thickness), and that their mechanical and electronic properties are such that interfaces with desired electronic properties can be formed between the electrolyte and the semiconductor. We also require the optical absorption to be sufficiently low within the polymer film, so that negligible conversion losses are introduced in this manner. They should furthermore be stable over considerable periods of time, and should not allow deleterious reactions at the interface between the polymer film and the semiconductor. Practical aspects of manufacture and encapsulation strongly recommend the solid state approach to PECs. Polymer electrolytes are easily processed into thin films that can be spun or cast over large areas, which are some requirements for photoelectrochemical materials. They are encapsulated



47

more easily than PECs using liquids, as the electrolyte does not seep out. They might have to be protected from the atmosphere, but this has been handled. Much work is needed for the development of rugged interfaces between the various phases in a PEC, including those between the polymer electrolyte and the semiconductor as well as those between the polymer electrolyte and surface bound electrocatalysts (Skotheim, 1986a). Nafion, PEO, and many other polymer-based electrolytes have been widely used in PECs, which show better performances than their counterparts using liquid electrolytes (Li et al., 2006; Skotheim, 1986b; Wang, 2009).

1.5.4 Electrochemical capacitors In recent years there has been increasing interest in the power capacitors, ultracapacitors or supercapacitors based on electrochemical systems. These include electric double layer capacitor (EDLC) types based on carbon electrodes with suitable electrolyte systems, and electrochemical capacitors with pseudocapacitance (Conway, 1991, 1995). Several types of capacitors are classified in Table 1.8 according to their energy storage mechanisms.

Table 1.8 Classification of capacitors according to the energy storage mechanism Basis of charge or energy storage

Type

Regular dielectric capacitor Electrostatic Vacuum Dielectric Electrolytic capacitor Electrostatic Oxide-film EDLC Electrostatic Aqueous electrolyte Non-aqueous electrolyte Ion-conducting polymer Solid inorganic salt Redox capacitor Redox oxide film Redox polymer film Soluble redox system Underpotential deposition

Example

– – Al, Ta oxide condenser

Carbon, activated carbon electrode, aqueous H2SO4 (Charge separation TEABF4/propylenecarbonate at double layer PAN gel electrolyte electrode interface) RbAg4I5 Faradic charge transfer RuO2·xH2O, IrO2, NiO/Ni, TiS2/Li+ (Pseudocapacitance) Polypyrrole, polythiophene, polyacene [Fe(CN)6]3−/[Fe(CN)6]4− H/Pt, Pb/Au (adsorption)


48


Redox capacitors function by using electrochemical pseudocapacitance based on fast Faradic redox (Conway, 1991), insertion of protons into noble metal oxides (Liu and Anderson, 1996; Zheng et al., 1995), p- and n-doping of conducting polymers (Arbizzani et al., 1995), lithium ion intercalation (Passerini et al., 1995) or underpotential deposition (adsorption) processes (Conway and Tilak, 1992). Among the various materials for redox capacitors, proton insertion into noble metal oxides and p- and n-doping of conducting polymers have been widely investigated. Conway (1991) reported the use of RuO2 and IrO2 as electrodes of redox capacitors. Zheng et al. (1995) found the specific capacitance, 760 F g−1, of amorphous hydrous ruthenium oxide, where the process can be expressed as RuO2·xH2O + γH+ + γe− ↔ RuO2−x−γ(OH)2x+γ

0≤γ≤2

[1.21]

Noble metal oxides, however, are expensive and usually have poor electrical conductivity. The process of charge storage in conducting polymers requires electronic transport through the polymer backbone. Charging commonly used conducting polymers corresponds to one charge unit per 2–3 monomer units at the maximum. Conducting polymers belong to a class of materials with exciting potential in the field of supercapacitors because they can be charged and discharged at high rates. The principle of EDLC operation is very simple and is based on the wellknown electrical or double layer phenomenon. The device operates within a potential range in which no Faradic reactions take place, and thus the behaviour is fully capacitive. Polarisation of the electrodes in opposite directions leads to accumulation of opposite charges at the electrode–solution interfaces. The higher the electrode surface area and the polarity of the electrolyte solution and its ionic concentration, the higher the capacity and energy density of these devices. The capacitance (C) and the accumulated electrostatic energy (E) stored are given by equations [1.22] and [1.23], respectively (Nishino et al., 1985): C=∫ E=

ε dS 4πδ

[1.22]

1 CV 2 2

[1.23]

EDLC with very low internal resistance (no electrochemical reaction resistance), very high capacitance and extremely long cycle life (ca. 106 cycles) can be achieved (Murphy and Kramer, 1994). Thus, the EDLCs have been envisaged as pulse-power sources, an especially useful complement to batteries for electric vehicles. Improving the energy and power density of EDLC requires development of new electrode materials and electrolyte systems as well as new cell



49

design. It appears from our own experience that using thin films of SPE in EDLC may be useful in achieving the advanced goals for ultracapacitors. Ion conducting polymers may be preferable in these devices’ electrolytes because of their flexibility, malleability, easy manufacture and chemical stability (for the same reasons that they have been applied to lithium secondary batteries). The gel electrolyte systems, which consist of a polymeric matrix, organic solvent (plasticiser) and supporting electrolyte, show high ionic conductivity of about 10−3 S cm−1 at room temperature and have sufficient mechanical strength (Reiche et al., 1995). Therefore, the gel electrolyte systems are superior to SPEs and organic solvent-based electrolytes as materials for batteries and capacitors, for room temperature operation (Choudhury et al., 2009; Kalpana et al., 2006; Kumar and Bhat, 2009; Zhang et al., 2009). The application of SPE to EDLC is attractive, but there have been few attempts in this direction for the following reasons: (1) relatively low ionic conductivity of the most relevant polymer electrolyte systems at room temperature; (2) poor contact at the electrode/electrolyte interface; and (3) detrimental influence of crystalline domains of the polymer electrolyte systems on their conductivity. Thus, many researchers focus on improving the ionic conductivity and increasing the electrode’s capacitance by controlling the electrode/electrolyte interface. The proton conducting polymer (Nylon 6-10, 2H3PO4) has been used as an electrolyte in a power capacitor (Lassagues et al., 1995), and this polymer blend associated with activated carbon gives an electrostatic capacity similar to that obtained when using liquid electrolyte solutions. However, the energy density of a capacitor with proton conducting polymer is limited to a voltage of about 1 V (which is the maximum voltage applicable for such systems). Non-proton-conducting SPEs such as PEO or PPO salts have a low dielectric constant (e.g. ε = 5 for PEO) and low ionic conductivity at room temperature ( 0.5 mol/l.79,86–88 A decrease in Tg was also obtained with the incorporation of acid zirconia, SO42−–ZrO2 in PEO.89 However, the acidic sulphate function is very aggressive for PEO and this reactivity may explain such behaviour. Indeed, Chauvin et al.90 have shown that PEO is very sensitive to an acid medium and a strong decrease in its molecular weight was observed. In other studies, the opposite trend has been observed: a small increase in Tg (one or two degrees) was noted with the incorporation of SiO2 in PEO8LiTFSI and PEO8LiClO4,58 while a large increase in Tg, i.e. 20 °C, and a large decrease in the tan δ peak size obtained by dynamic mechanical analysis (DMA), were observed with the incorporation of hydrophilic SiO2 in PEO/LiClO4.76 The tan δ decrease was associated with a small amount of product involved in the glass transition. A net increase in Tg was observed with the addition of Al2O3 in PEO and in polyoxypropylene.91,92 The Tg of a polymer is directly related to the macroscopic viscosity of the amorphous phase. In a semicrystalline polymer, the amorphous phase is constrained by the crystalline phase, thus the Tg measured in a semicrystal-


Composite polymer electrolytes for electrochemical devices

139

line polymer is higher than the Tg of the same polymer when totally amorphous. Thus, a decrease in the amount of crystalline phase induces a decrease in Tg. It is well known that the incorporation of lithium salts leads to an increase in the Tg of a polyether due to the specific interaction between the lithium cation and the oxygen groups of the polyether. Changes in the Tg of PEO in nanocomposite electrolytes may thus be related to the reduced crystalline phase and the modification or establishment of specific interactions between the polymer, the fillers and the salt. In most of the studies,58,71,81– 84 the effect of filler incorporation on Tg is small, which may be related to antagonist and often weak effects, i.e. decrease in Tg due to PEO amorphisation, specific interactions between fillers and PEO that may increase Tg, or specific interactions between ionic species and fillers that may reduce the cation/polyether interaction and thus decrease PEO Tg. These specific interactions can be investigated using IR and Raman spectroscopy, NMR or quasielastic neutron scattering. Several studies have been carried out on nanocomposite polymer electrolytes using these techniques. Restriction of polymer matrix mobility due to the presence of nanofillers, via filler/polymer interactions, was observed by Tsagaropoulos and Eisenberg93 in nanocomposites with silica aerogel and poly(vinylalcohol), poly(styrene), poly(methylmethacrylate) or poly(4-vinylpyridine) as the polymer matrix. The presence of a second Tg was observed when the silica content was higher than 10 wt%. This Tg, which was higher than the bulk polymer Tg, was associated with the formation of tightly bound polymer, due to polymer/filler interactions. Mobility restriction was also observed by quasi-elastic neutron scattering.94 Elastic scattering was affected by filler addition, with the results obtained suggesting that an immobilised layer of the polyether matrix, 3PEG, was present around the TiO2 filler, roughly 5% of the total polymer with 10 wt% TiO2 (size 21 nm), whereas the bulk polymer was unaffected by the presence of fillers. While 7Li NMR has provided information about ion dynamics, 1H and 13 C NMR have contributed to elucidate PEO dynamics in pure and composite materials. Numerous NMR studies on nanocomposite polymer electrolytes have been published.60,95–100 Owing to the complexity of the phenomena occurring in nanocomposite polymer electrolytes, different observations and interpretations are found in the literature. It is generally expected that the NMR linewidth of the mobile nucleus decreases when ionic motion occurs. However, Dai et al.,60 with highly concentrated PEO1.5/ LiI/Al2O3, and Forsyth et al.,95 with 3PEG/LiClO4/TiO2, SiO2 and Al2O3 nanoparticles, observed an increase in the 7Li linewidth with the addition of nanoparticles, although the ionic conductivity increased. Chung et al.96 showed that the filled electrolyte, PEO8/LiClO4/TiO2, and the unfilled one, exhibit comparable linewidths for nuclei, proton and


140


lithium. The 1H diffusion coefficient was slightly decreased with the incorporation of TiO2 at 80 °C. In contrast, enhancement of the 7Li diffusion coefficient was observed with the incorporation of TiO2. Cheung et al.97 also observed an improvement in the lithium diffusion rates without a corresponding increase in segmental motion for PEO/Al2O3/LiCF3SO3. The effect of nanoparticle size was not considered significant, as the highest 7Li diffusion coefficient was obtained with the highest nanoparticle size, 0.05 µm. Mustarelli et al.,98 using PEO8/LiClO4 or LiN(CF3SO2)2/SiO2, found an increase in the activation energy for lithium motion and concluded that 7 Li spin dynamics were influenced by fillers. Using a 7Li–19F decoupling experiment on PEO/LiBF4/α-Al2O3, Bloise et al.99 observed that the Li–F interaction was weaker in the composite with α-Al2O3 than in the composite with γ-Al2O3. The 7Li linewidth decreased significantly when Al2O3 was added to PEG (polyethyleneglycol, Mw = 2000 g/mol)/LiClO4 (O/Li = 26), while there was only a marginal decrease in 1H linewidth.100 Vogel et al.,101 using 2H NMR, found that the presence of TiO2 nanoparticles hardly affected the dynamics or crystallisation behaviour of the polymer in low molecular weight PEO nanocomposites. Bloise et al.102 showed the mobility of Li+ in PEO8LiClO4/TiO2 to be very high. These different behaviours may be associated with the complexity of nanocomposite polymer electrolytes. One possible explanation offered was that nanofillers cause a decrease in Li–H distance and thus an increase in the Li–H heteronuclear dipolar interaction, which is shown to be the main contributor to the linewidth 1H in decoupling experiments.99 Positron annihilation lifetime spectroscopy data suggest that there may be competition between the polymer/Li interaction and the polymer/TiO2 interaction in the composite electrolyte.95 Salt dissociation and PEO conformation depend on the nature and concentration of the salt and the presence of fillers and temperature, and this has been widely studied using IR and Raman spectroscopy.103,104 The D-LAM (disorder-longitudinal acoustic mode) bands between 200 and 300 cm−1 have been assigned to longitudinal backbone motions of the polymer chain.105 The intensity of the PEO D-LAM bands has been shown to decrease significantly with the addition of fillers.106 With the addition of TiO2 to the pure polymer, 3PEG, these modes were suppressed,83 suggesting that there was a restricted mobility of the polymer phase, which agrees with differential scanning calorimetry (DSC) measurements93 and quasi-elastic neutron scattering results.94 In contrast, the addition of Al2O3 did not suppress the D-LAM mode, indicating that the level of interaction between polymer and Al2O3 is not the same as that between polymer and TiO2. The addition of TiO2 did not affect any of the PEO conformation sensitive modes in the CH spectral regions (800–900 cm−1, 2800–3100 cm−1) in 3PEGbased electrolyte.106 In contrast, using a polymer electrolyte based on high


Composite polymer electrolytes for electrochemical devices O

O C M

R

O Unidentate

C

M

M

O

M

O

R

O Bidentate

C

141

R

Bridging

Scheme 4.1 Three modes of bonding structures from Xiong et al.73

molecular weight PEO and LiClO4, Chen-Yang et al.78 observed a sharpening of the peak associated with the stretching vibrations of the C—O—C when silica aerogel was added. They linked this behaviour to the presence of Lewis acid–base interactions between the silica oxygen atom and the Li+, resulting in a weakening of the interaction between the Li+ and the PEO. A large change in the IR spectrum of PEO/LiClO4 was found to occur with the addition of ZnO.73 The C—O—C stretching vibration becomes a single band centred on 1107 cm−1. The authors linked this modification to the formation of crosslinks that further weaken the C—O bonds. Some evidence of amorphisation of the PEO matrix with the addition of ZnO was also observed using the CH2 wagging mode modification. A change in the coordination modes of acetate groups with zinc metal was observed with its incorporation in the PEO/LiClO4 electrolyte.73 The coordination of lithium ions with both PEO segments and ZnO acetate groups changed the bonding structure of the acetate groups, with bidentate and bridging modes73 (Scheme 4.1). Consequently, the filled electrolyte was more amorphous than the unfilled one. Using the symmetric stretching mode of ClO4 or SO3, no significant decrease in ion pairing was observed with the addition of both TiO2 and Al2O3.83 However, the addition of TiO2 to the electrolyte shifted the symmetric stretching vibrations to smaller frequencies for both triflate and perchlorate salts, indicating weak coordination between the anion and the TiO2 surface. Such behaviour was not observed with Al2O3 fillers.83 Johansson et al.71 studied the interactions between fillers and LiTFSI by directly mixing fillers and salt together. The peaks in the 720–770 cm−1 region revealed no major changes in the anion modes, in particular no increase in the amount of free ions. In order to assess the amount of ion pairing, the ratio between the intensity of the 741 cm−1 band (associated with free ions) and the integrated intensity at 726–759 cm−1 was used.107 An increased ratio was observed with the addition of a small amount of SiO2 to PEO9LiTFSI.71 When the filler amount increased, this ratio decreased, with an associated increase in electrolyte viscosity, which should lower anion mobility and induce broadening owing to the slow orientation of the anion. No direct effect of filler addition on anion dissociation was observed. A small increase in the quantity of free ClO4− ions was obtained with the incorporation of silica aerogel in a highly concentrated PEO6LiClO4 electrolyte.78 The highest


142


dissociated electrolyte was obtained with 2% of fillers, subsequently dissociation decreased with an increase in filler amount.78 The dielectric study of nanocomposite polymer electrolytes performed by Jayathilaka et al.108 indicated that the quantity of free ions was not modified significantly with the addition of Al2O3 to PEO/LiTFSI electrolyte. However an increase in ion mobility was noted. In the various studies reported, little or no improvement in ionic dissociation was found, but some specific interactions were observed which induced PEO amorphisation in semicrystalline electrolytes and restricted PEO mobility in amorphous electrolytes. These behaviours were dependent on the nature and amount of filler. Conductivity behaviour Considerable work on improving ionic conduction in nanocomposite polymer electrolytes has been conducted. The best results were obtained with Al2O359,83,86–88,108 and TiO2 ,59,83,96 while improvements with SiO2 were minor.58,109 In PEO electrolytes, conductivity values are very sensitive to the amount of crystalline phase present, which is related to the thermal history of the sample. A large difference in PEO/LiClO4 conductivity at 25 °C was obtained in relation to the sample’s thermal history.73 High conductivity was obtained just after sample cooling, with σ = 10−6 S/cm, then conductivity fell dramatically after the electrolyte had been kept for two weeks at room temperature in a vacuum, σ = 10−9 S/cm. It is difficult to compare conductivity data obtained at room temperature; however, conductivity is generally improved below the PEO melting point after the sample has been previously heated. This improvement is associated with a decrease in the amount of electrolyte in the crystalline phase present59 and a reduction in crystallisation kinetics due to the incorporation of fillers. At room temperature, conductivity was found to increase with the addition of fillers, generally reaching a maximum between 5 and 15 wt% of filler.74,75,97,110–112 For large amounts of filler, conductivity decreased, which was associated with the blocking of conduction pathways by the electrically insulating nanoparticles. The results for conductivity values at high temperature are more conflicting. Some improvement in ionic conductivity has been observed,72,87,97,108,111,112 but many studies show a quasi-invariance in conductivity values at high temperatures,73,106,113 while others report a conductivity decrease.70,71 Croce et al.72 and Jayathilaka et al.108 observed that the nanoparticle surface groups influence improvements in conductivity. The best conductivities were obtained using neutral and acidic Al2O3 nanoparticles (Fig. 4.2). The model developed suggested that enhanced



143

–3.0

Log (σ) (S/cm)

–3.2 –3.4 –3.6 –3.8 –4.0 –4.2

Ceramic-free Acid Neutral Basic

–4.4 2.6

2.7

2.8

2.9

3

3.1

1000/T (K–1)

4.2 Conductivity of PEO20LiCF3SO3 10 wt% Al2O3 nanocomposite electrolytes with different Al2O3 surface groups, extracted from Croce et al.72

Acidic HO Al HO Al

Li+ CF3-SO3–

Al OH Al OH

O Li+ O

CH2 CH2

Neutral Li+

CF3-SO3–

HO Al O Al

Al Al Al

O O

Li+ O

CH2 CH2

Basic Li+

CF3-SO3– O

Al Al

Al Al

Al

Al

O O

Li+

CH2 CH2

O

Scheme 4.2 Model of the surface interactions between nanosized Al2O3 particles and the PEO–LiCF3SO2 electrolyte, extracted from Croce et al.72

ionic conduction is related to the presence of hydrogen bonds between the acidic or neutral ceramic surface and the anion and polymer chain (Scheme 4.2), which improve ionic dissociation. With basic surface nanoparticles, no improvement in ionic conductivity was obtained. A model based on acid– base interactions has been proposed by several authors62,64,72,108 to explain conductivity improvement. However, no evidence of an increase in salt dis-


144


Log (σ) (S/cm)

–2.6 5.8 nm

–2.8

10–20 nm 37 nm

–3.0 –3.2 –3.4 –3.6 –50

2 mol/l. At low salt concentrations, the filled and unfilled electrolytes exhibited the same conductivity behaviour at high temperatures. This behaviour was consistent with the idea that the filler influences ion aggregation in cases where salt aggregation is significant. In contrast, a decrease in ionic conductivity over the entire temperature range was observed with the incorporation of SiO2 in a liquid electrolyte based on PEG dimethylether (Mw = 500 g/mol). This decrease was associated with an increase in the viscosity of the electrolyte. The nanocomposite electrolyte PEO–LiClO4–ZnO exhibits the same conductivity as the unfilled electrolyte at high temperatures,73 whereas at room temperature, nanocomposite electrolytes have a much higher conductivity (Fig. 4.4). The nature and amount of fillers were found to have no effect on conductivity at high temperature in PEO–LiSO3CF3 electrolytes obtained by hot-pressing with the addition of both SiO2 and Al2O3 fillers.113


146


Cationic transference number and interface properties Even if there is no conductivity improvement at high temperature, the incorporation of fillers can be useful if it improves the cationic transference number or interfacial stability. An increase in cationic transference number, T +, is generally observed with the addition of nanofillers.59,72,77,80,111,114,115 Cationic transference numbers were determined using the electrochemical method developed by Bruce et al.116 This method assumes an ideal electrolyte and complete dissociation of the electrolyte, which is not the case in concentrated electrolytes. However, no method adapted to concentrated electrolytes exists, thus the data must be used with caution. An increase in T + was obtained by the addition of Al2O3 to P(EO/ MEEGE)–LiAl(HFIP)4 (MEEGE = 2-(2-methoxyethoxy)ethyl glycidyl ether, LiAl(HFIP)4 = lithium tetra(1,1,1,3,3,3-hexafluoro-2-propyl) aluminate).114 The highest T + improvement was obtained with the largest amount of Al2O3, even though T + remained low, i.e. 0.12 for 20 wt% Al2O3. An increase in T + was obtained in PEO–LiClO4 electrolytes with the addition of TiO2,59 and mesoporous SiO2.111 In PEO–LiCF3SO3, an improvement was obtained with ZrO2,115 and with acidic and neutral Al2O3 but not with basic Al2O3.72 A moderate decrease in interfacial impedance was observed with the addition of fillers by Appetecchi et al.117 Careful attention must be paid to the elaboration process in order to evaluate the real effect of nanofillers. Xie et al.92 observed no change in interfacial resistance with the addition of SiO2 in high molecular weight PEO, whereas a decrease was observed when a viscous polyether was used. The nanocomposite PEO/LiCF3SO3/γ-LiAlO2 presented good interfacial stability versus metallic lithium.80 However, the TiO2 particles were found to present poor stability versus metallic lithium,50 so that preference must be given to other nanoparticles. Indeed, elaboration conditions, and especially water content, have a major impact on how filler addition affects interfacial properties. With lithium electrodes, cycle duration was much longer with the addition of 20 wt% γ-LiAlO2 to PEO20LiBF4.80 The fillers did not have a direct effect on dendrite growth but the improved mechanical properties of the electrolyte may explain this longer cycle duration.92 Good cyclability in a Li/ PEO35LiCF3SO3 + 5 wt% SiO2/LiFePO4 battery was obtained113 with a capacity fade of only 0.17% per cycle at 100 °C.

Mechanical properties Some authors92,114,118 observed improved mechanical properties in polymer electrolytes with added nanofillers. The rheological behaviour of filled and unfilled PEO–LiTFSI electrolytes was studied at 85 °C.92 A significant



147

increase in modulus was obtained with the incorporation of 10 wt% of fumed silica, which may be due to hydrogen bonding between the PEO matrix and the nanoparticle surface. The filled matrix exhibited longer relaxation times, associated with a slower segmental motion. The mechanical properties were evaluated by the dimensional stability of the electrolyte in relation to creep. The latter was indirectly determined by impedance spectroscopy. Improved mechanical stability was reported for α-Al2O3,15 alumina whisker reinforced PEO8–LiClO4118 and hyper-branched polymer electrolyte filled with α and γ-LiAlO2 particles.119 An improvement in tensile modulus above the matrix melting point was observed with the addition of 10 wt% Al2O3.114 Above 50 °C, the unfilled electrolyte P(EO/MEEGE)– LiAl(HFIP)4 flowed, whereas the filled electrolyte exhibited a storage modulus higher than 10 MPa. The storage modulus was stabilised by the incorporation of 10 wt% TiO2 in PEO30LiTFSI, with a storage modulus equal to 4 MPa.120 Croce et al.59 reported a significant increase in both Young’s modulus and yield stress for Al2O3 and TiO2 reinforced PEO8LiClO4. Unfortunately, no modulus value was reported in their paper.

4.5.2 Organic fillers Cellulose whiskers were added to PEO–LiTFSI electrolytes to improve the mechanical properties of the polymer electrolytes without compromising ionic conduction.120–124 The use of cellulose whiskers in nanocomposite polymers was shown to result in a high reinforcing effect at T > Tg and an increase in thermal stability, even at low loading levels. This effect was attributed to the formation of a percolating network within the matrix, resulting from strong hydrogen bonds between nanoparticles. The average aspect ratio, L/d, of these whiskers was estimated to be close to 70. The incorporation of cellulose whiskers in PEO was found to have no effect on PEO Tg, and a decrease in PEO melting temperature was observed only when the whiskers exceeded 10 wt%.122 The same behaviour was obtained with PEO–LiTFSI electrolytes.120 At high temperatures, the storage modulus of the filled electrolyte was found to be greater than that of the unfilled polymer electrolyte by more than a factor of 100.120 Tunicin whiskers were found to have a much greater mechanical reinforcing effect than TiO2 (Fig. 4.5), owing to whisker/whisker interactions. The incorporation of tunicin whiskers induced an approximately threefold decrease in conductivity. Possible explanations include (i) the low dielectric constant of cellulosic fillers, (ii) interactions between cellulose and PEO, and (iii) the effect of whiskers on salt dissociation and ion mobility. Indeed, the relaxation time of 1H protons was significantly reduced with the addition of whiskers, and the diffusion coefficients for both cation and anion decreased by almost a factor of 3.120 The decrease in ionic mobility is in


148

Polymer electrolytes 10

Log (E´)(Pa)

9 8 7 6 –100

–50

0 50 Temperature (°C)

100

150

4.5 Storage tensile modulus E′ vs. temperature at 1 Hz for the unfilled POE30–LiTFSI polymer electrolyte (䊉) and related composites filled with 10 wt% TiO2 (䊊) and tunicin whiskers (ⵧ).

good agreement with the reduction in ionic conduction. Thus the main effect seems to be the decrease in chain dynamics. Cationic transference numbers, T +, in the composites were calculated from the diffusion coefficient values. These values were almost constant for the composites, and were found to be temperature-independent. They ranged between 0.22 and 0.29 and are comparable to those reported for POE-based unfilled polymer electrolytes.125 Since the mechanical performance of the tunicin whisker/PEO–LiTFSI polymer electrolyte was ascribed to the percolating filler network, experiments were performed with plasticised polymer.124 PEO–LiTFSI-based polymer electrolyte with 30 wt% tetraethylene glycol dimethyl ether, TEGDME, as a plasticiser and tunicin whiskers as the nanometric reinforcing phase were processed. A thermal stabilisation effect above Tm was observed. The high temperature modulus, around 100 °C, was about 200 MPa for the composite PEO/TEGDME 30 wt% filled with 6 wt% whiskers and 500 MPa when filled with 10 wt% whiskers. This is an indication that plasticisation of the matrix did not interfere with the formation of the cellulosic network within the matrix. The conductivity of two different POE-based electrolytes was compared.124 It was observed that POE12–LiTFSI plasticised with 30 wt% TEGDME and filled with 6 wt% tunicin whiskers displayed a similar conductivity to the unplasticised and unfilled POE12–LiTFSI polymer electrolyte, but with much better mechanical properties.

4.6

Gel polymer electrolytes

A gel polymer electrolyte is a polymer swollen by a liquid electrolyte. These electrolytes exhibit high ionic conductivity at room temperature but present



149

some reactivity with the lithium electrode owing to the presence of organic solvents; they also have poor mechanical properties. The challenges with these systems are therefore to improve their stability towards the lithium electrode, decrease gas evolution and enhance their mechanical properties. One solution might be to incorporate nanoparticles that would improve electrochemical stability and mechanical properties without decreasing ionic conductivity. The most common polymers used as the polymer matrix in gel polymer electrolytes are poly(vinylidene fluoride), PVDF, and polymethymethacrylate, PMMA. The nanoparticles used in gel polymer electrolytes are the same as those used in solid polymer electrolytes, i.e. organic clay126,127,128 and metal oxide fillers.129,130

4.6.1 Composite gel polymer electrolytes based on clays MMT is naturally hydrophilic and must be rendered hydrophobic or organophilic in order for it to be compatible with the polymer matrix. Gel nanocomposite polymer electrolytes were obtained by dispersing organophilic MMT in a PVDF or PVDF-hexafluoropropylene (PVDF-HFP) solution. The composite membranes were obtained by film-casting;128,131,132 the film was dried and then the liquid electrolyte was added by swelling.131,132 The MMT was exfoliated and, in order to improve MMT delamination, the MMT was sonicated in dimethylformamide (DMF).131 The filled PVDF membrane shown excellent wettability with the electrolyte solution. Solvent and electrolyte uptake by the filled membrane was considerably greater than that of the unfilled membrane.131 This increase in electrolyte uptake could explain the improvement in ionic conductivity observed. A similar study was performed by Jacob et al.126 using PVDF and PVDF– HFP as the polymer matrix, LiSO3CF3 as the salt and PC or EC as the solvent. A decrease in ionic conductivity was observed with the addition of clays, with conductivity equal to 4.4 × 10−4 S/cm for the filled gel electrolyte compared with 1.9 × 10−3 S/cm for the unfilled one, at room temperature. However, the conductivity of the filled electrolyte remained unchanged after several temperature cycles, while that of the unfilled electrolyte decreased owing to solvent evaporation. The nanoparticles may act as a physical and chemical barrier limiting solvent evaporation. The inconsistent results may be related to the difficulty in determining whether the added nanoparticles affect ionic conductivity directly, since the addition of nanoparticles can modify the amount of solvent and salt, which in turn could modify conductivity. The mechanical properties of nanogel electrolytes were also significantly improved. The nanogels exhibited a rubbery behaviour with a storage modulus of about 1 GPa at room temperature (one order of magnitude


150


higher than the corresponding gels).126 MMT was incorporated in PMMA by dispersion in the monomer, and polymerisation was performed in situ.127,133 The gel PMMA–clay nanocomposites presented an intercalated structure and a slight increase in ionic conductivity was observed with 1.5 wt% clay.127,133 The PMMA nanocomposite gel electrolyte showed a stable lithium interfacial resistance, similar to PMMA gel. The mechanical properties of the pure PMMA were improved with the incorporation of 10 wt% MMT. The problem associated with this process is the need to purify the polymer by eliminating unreacted monomers.

4.6.2 Composite gel polymer electrolytes based on metal oxide fillers Nanocomposites based on PMMA were synthesised via sol–gel transformation and in situ free radical polymerisation of MMA. Gel nanocomposite polymer electrolytes were obtained by the addition of PMMA–TiO2 or PMMA–SiO2 to the PC-based liquid electrolyte. After 2 h at 55 °C, transparent films were obtained. The gel composite based on PMMA–TiO2 exhibited higher viscosity than the unfilled gel electrolyte and the PMMA– SiO2-based electrolyte. The conductivity values of filled and unfilled gel electrolytes were very similar, but the nanocomposites exhibited better mechanical properties than the unfilled polymer electrolytes. Nanocomposites based on PMMA were also obtained by SiO2 dispersion and PMMA solubilisation in PC/LiCF3SO3 liquid electrolyte.129 Improved ionic conductivity was obtained with 2 wt% SiO2 for temperatures higher than 50 °C, but at room temperature the addition had no effect on ionic conductivity. No change in Tg following incorporation of SiO2 was observed. The incorporation of fumed silica resulted in more physical linkages in the gel electrolyte in the form of siloxane linkages, which caused a substantial increase in the gel modulus. Wu et al.134 studied the effect of several nanocharges, TiO2, ZnO and MgO, on gel electrolytes based on PVDF–HFP. The gel electrolytes were obtained by mixing the nanocharges and the polymer in acetone. After solvent removal, the liquid electrolyte was added. A decrease in the porosity of the PVDF–HFP and an increase in solvent uptake were obtained with the addition of nanocharges. Owing to the higher amount of solvent introduced, an improvement in ionic conductivity was obtained. A PEO-based nanocomposite gel was studied by Nan et al.130 A solvent mixture EC/PC was added to mesoporous SiO2. The nanoparticles thus obtained were mixed with PEO–LiClO4 electrolyte. A decrease in PEO crystallisation was observed with the incorporation of EC/PC–SiO2. This decrease was greater than that observed with the addition of SiO2, which may be related to the presence of organic solvents. A net improvement in



151

ionic conductivity was obtained with the incorporation of EC/PC–SiO2, even at high temperature. However, this increase may be related more to the incorporation of solvents than the direct effect of nanoparticles. The EC/PC–SiO2 filled PEO exhibited a very high Young’s modulus at room temperature, higher than the unfilled PEO electrolyte and EC/PC plasticised PEO electrolyte. This improvement in mechanical properties may be related to a specific interaction between SiO2 and PEO. Gel nanocomposite polymer electrolytes based on PVDF-HFP were studied by NMR.135 In this type of electrolyte, BaTiO3 and clay were used as nanoparticles and PC + LiCF3SO3 as the liquid electrolyte. The filled gel electrolytes presented better mechanical properties than the gels without the filler. The fillers lowered the conductivity by a small amount. The diffusion coefficient measured by NMR indicated that both anion and cation mobilities were reduced by the presence of the filler, but the effect was generally greater for the anions.

4.7

Composite electrolytes for proton exchange membrane fuel cells

Proton conducting membranes for proton exchange membrane fvel cells (PEMFCs) are usually based on thin ionomer films. The membrane separates the electrodes, allows proton transportation from anode to cathode and creates a barrier against the passage of gases or fluids (e.g. methanol). However, the ionomer itself does not provide any appreciable conductivity. The membrane must therefore be swollen by molecules, e.g. water, to ensure proton conductivity. Water uptake by the membrane is therefore one of the essential parameters for obtaining high conductivity. Indeed, owing to its high dielectric constant, water favours dissociation of the ion pair, increasing the concentration of charge carriers and the solvation of both anion and proton. Generally, two principal mechanisms are used to describe proton diffusion through the membranes.136–138 One is the vehicle mechanism, whereby a proton combines with vehicles such as H3O+ or CH3OH2+ and also with unprotonated vehicles (H2O), thus allowing the net transport of protons.136 In this mechanism, conductivity is directly dependent on the rate of vehicle diffusion. The other is the Grotthus mechanism (hopping), whereby protons are transferred from one proton acceptor site to another by hydrogen bonds. In this mechanism, additional reorganisation of the proton environment, consisting of reorientation of individual species or even more extended ensembles, leads to the formation of an uninterrupted trajectory for proton migration.136 Recently Marechal et al.139 found that four water molecules are involved in ion-pair solvation of Nafion® membranes. In Nafion®, which is the most advanced proton-conducting membrane, the drastic decrease in proton


152


conductivity at low relative humidity limits the operating temperature to about 80 °C.140–149 An important focus for PEMFC research is developing a PEM (proton exchange membrane) which has satisfactory proton conductivity at higher temperatures (above 100 °C) and low relative humidity,140–142 because cells that can operate in such conditions have a number of advantages over lower temperature cells: •

Considerable improvement in the platinum tolerance of the electrocatalysts to carbon monoxide poisoning. Carbon monoxide is a major problem because trace amounts of CO in the H2 feed gas (more than 10 ppm) will poison the Pt anode electrocatalyst in PEMFCs operating at 80 °C. A quantitative analysis of the free energy for H2 and CO adsorption as a function of temperature150 suggests that, by elevating the operating temperature of the cell to 145 °C, CO tolerance at the anode should increase by a factor of 20 (from 5–10 to 100–200 ppm). • Significant improvement in PEMFC thermal management. By elevating the temperature of the fuel cell stack, thermal management can be simplified due to more efficient waste heat removal. • The amount of noble metal catalyst can potentially be reduced due to fast reaction kinetics at high temperature. • Water management is improved. At the moment, reactant gases need to be humidified before entering the cell to prevent the membrane from drying out. Membrane dehydration also causes the membrane to shrink, reducing contact between electrode and membrane and possibly leading to pinholes, which permit the crossover of reactant gases. To benefit from these advantages, it is now becoming clear that pure ionomer membranes are not suitable, and new types of proton-conducting membranes that work at temperatures higher than 100 °C have to be developed. There are currently three principal polymer electrolyte types proposed for high temperature PEMFCs: • •

•

Anhydrous proton-conducting membranes, in which proton conductivity is assured by molecules other than water. Hygroscopic inorganic/organic composite membranes, in which the ionomer or polymer is blended with hydroscopic inorganic particles so that the membranes will remain hydrated at higher operating temperatures. Self-humidifying membranes, in which the water is produced directly in the membrane.

Thermally stable anhydrous proton-conducting membranes involve membranes based on (i) acid–base composite membranes, i.e. acid blended with polymers containing Lewis base moieties, namely H3PO4/PBI (polybenz-



153

imidazole),151–153 and (ii) proton-conducting ionic liquid composite membranes, i.e. a functional polymer blended with a proton-conducting ionic liquid.154–156 For the latter, an effective way to achieve low humidity and high temperature operation is to disperse particles of hygroscopic oxides or hygroscopic solid inorganic proton conductors with low solubility (e.g. heteropolyacids) inside the polymer matrix. In the case of self-humidifying membranes, research has focused mainly on incorporating Pt or Pt/C catalysts in the membrane to combine the permeable oxygen and hydrogen to produce water and humidify the membrane. These two types of polymer electrolytes are discussed below.

4.7.1 Components of composite electrolytes Polymer matrix Inorganic/Nafion® composites have been proposed for use at high temperatures in fuel cells.140–142,157–167 In addition to Nafion®, sulphonated hydrocarbon polymers such as (Scheme 4.3) sulphonated polystyrene (SPSt),168 sulphonated polysulphone (SPSF),169,170 sulphonated polyetheretherketone (SPEEK),171–175 and others without functional groups, such as PVDF,176–179 PEO180–183 and PBI,184–186 have been widely used as the host matrix for preparing inorganic/organic composites to be used at high operating temperatures. There are a number of reasons for the interest in developing inorganic– organic composite membranes: (i) to improve the self-humidification of the membrane on the anode side by dispersing a hydrophilic inorganic component homogeneously in the polymer; (ii) to reduce electro-osmotic drag and therefore prevent the membrane from drying out on the anode side; (iii) to eliminate fuel crossover, e.g. methanol in direct methanol fuel cell (DMFC); and (iv) to improve the mechanical strength of the membranes without sacrificing proton conductivity. In sulphonated polymers, for example, a high degree of sulphonation is desirable for high conductivity. However, this will be accompanied by undesirable swelling (or even solubility in water) of the membrane and therefore reduced mechanical strength. Introduction of an inorganic component into the polymer will compensate for the mechanical behaviour by helping to improve thermal stability and enhancing proton conductivity when solid inorganic proton conductors are used. This may permit the use of sulphonated polymers with a low degree of sulphonation. Inorganic particles Solid inorganic particles include oxides such as amorphous silica, inorganic proton conductors, and in many cases silica-supported inorganic proton


154


PBI

N C

N

N N

H2 C

SPSt

CH2 n CH

CH

n

m

SO3H SPEEK

(

O

O

)n (O

CO

O

)m

CO

SO3H SPSF

O

R

O

SO2

n*

SO3H

O

R

O

SO2

p

m

p

R = –H or –C(CH3)2 SPAEEKK

O

O

O

O

CF3

C

C

C

70

O

O

O

C

C

30

CF3

SO3H SDFF

F

F F

F

CF3 O

F

F F

C O

O

CF3

F

n

SO3H SPPEK N N C=O

C

O n

O

SO3H

HO3S

Scheme 4.3 Polymer structures.

conductors. These compounds are crystalline and therefore mechanically poor when used alone in the form of membranes. The combination of these inorganic proton conductors with a polymer component provides flexibility. Composite polymer electrolytes The properties of these composite membranes depend not only on the nature of the ionomer and the solid used, but also on the amount, homo-



155

geneous dispersion, size, and orientation of the solid particles dispersed in the polymer matrix.

4.7.2 Composite polymer electrolyte elaboration Composite membranes can be prepared according to two main procedures: • •

dispersion of inorganic particles in an ionomer solution followed by casting; and growth of the filler particles within a preformed membrane or in an ionomer solution.

The first attempts to disperse proton-conducting inorganic particles in an ionomer solution followed by film casting were performed over 25 years ago.187–188 In this procedure, the solids are first ground until a fine powder is obtained and then dispersed under strong stirring in an organic solution of the polymer. The membrane is obtained by film casting and solvent evaporation. The method is very simple, but it is usually difficult to prevent the formation of particle agglomerates inside the polymer matrix, even by the use of ultrasound, and thus membranes containing non-homogeneous dispersions of micro-sized particles are usually obtained. In order to reduce particle sizes and obtain more homogeneous membranes, Rosiere and co-workers175,189 developed colloidal dispersions of hydrous oxides. The advantage of colloidal solutions is that the solids are powdered into very small particles, which generally give homogeneous dispersions. Colloidal dispersions of hydrous oxides in aqueous solution are easily obtained by hydrolysis of metal alkoxides. Transfer of nanometresized silica particles from commercial aqueous silica suspensions to higher boiling point solvents such as N-methyl-2-pyrrolidinone (NMP) has been described.175,189 Such small and well-dispersed particles, preferentially localised in the hydrophilic domain of polymers, maximise the interface region between the inorganic and organic components. Mauritz and co-workers190 proposed a sol–gel approach involving the addition of silica to a Nafion® membrane, which might provide better performance than particles because the silica is homogeneously dispersed on the angstrom scale as opposed to submicrometre-length scales for particle composites. This method was developed for all types of ionomers. From a practical point of view, it is a very interesting way to modify commercial membranes. The filler precursors are incorporated in the polymer matrix by simple impregnation (if the filler is neutral) or by ion-exchange reaction (if the filler is a cationic species), and then the membrane is treated with the necessary reactants to convert the precursor into the final inorganic filler.191,192


156


Another interesting family of PEMFCs is that of class II inorganic– organic hybrids, in which the inorganic part, generally silica, is covalently bonded to the polymer. These systems are more stable from a structural and mechanical point of view than other hybrids in which the inorganic units are only physically dispersed in the polymer matrix. Generally speaking, hybrid membranes have been prepared by sol–gel processes, starting with 3-isocyanatopropyltriethoxysilane and several polyethers. In this kind of system, the proton carriers are supplied by incorporating various types of acids into the membrane, in particular heteropolyacids, e.g. phosphotungstic (PWA) and silicotungstic acid (SWA) or H3PO4.180–183,193–196

4.8

Composite polymer electrolytes based on metal oxides

An effective way of achieving Nafion® membranes that operate under low humidity and high temperature conditions is to recast Nafion® membranes with mixed hygroscopic oxides (e.g. SiO2, TiO2, ZnO, Al2O3). Watanabe et al.197 presented the first report on this in 1994. It has been shown that water uptake by the oxide-containing membrane is higher than that of the pristine Nafion®. For recast Nafion® membranes pre-dried at 80 °C, the water absorbing ability under humidification with water vapour at 60 °C was found to be 17 wt%, whereas for membranes containing 3 wt% SiO2 of 7 nm size, the water absorbing ability was increased to as much as 43 wt%.197 The characteristics of composite membranes depend not only on the type of metal oxide, but also on surface area, size, concentration and surface chemical modification of the metal oxide, and on the type of polymer matrix. The primary method for characterising membranes is to compare the current–voltage response of the cell as a function of membrane chemical characteristics and cell humidity/temperature.

4.8.1 Electrochemical performance Antonucci et al.198 attributed the enhanced performance of the composite membrane to the hygroscopic properties of silica. Watanabe et al.197 investigated silica- and titania-impregnated Nafion® composite membranes in H2/O2 proton-exchange membrane fuel cells operating at 100 °C with minimal external humidification. They reported that the silica particles were superior to titania particles in terms of water retention qualities within the Nafion® membrane, and attributed this to the higher water sorption properties of silica. Adjemian et al.199,200 reported that the performances of silica and titania composites at 130 °C and 75% RH are comparable (0.33 Ω cm2 and 0.35 Ω cm2 cell resistivities) while recast Nafion®-based composites present high cell resistivity (0.9 Ω cm2).



157

Some authors 146,199,200 have suggested that improved cell behaviour is not related to water retention or the acidity of the metal oxide component, but rather to the effect of the metal oxide particles on the temperaturedependent structure of the polymer matrix. To investigate the potential role of a chemical interaction between the metal oxide interface and the Nafion® sulphonate groups, Adjemian et al.199 constructed composite membranes containing alumina and zirconia particles. They reported that these systems behaved in a very similar manner, with both alumina and zirconia-based composites generally producing high resistance cells that were not tolerant of low humidity conditions. One possible reason given for the observed decrease in performance is the potential release of Al3+ or the production of zirconium oxysulphate in the acidic Nafion® environment, disturbing some of the sulphonate sites and leading to modification of the membrane topology, and thus potentially limiting proton diffusion.199 Park et al.201 compared the impact of physically mixed SiO2 + ZnO2 and binary SiO2–ZnO2 oxide fillers on Nafion® membrane properties. They concluded that the binary oxide improved proton conductivity and water uptake at high temperatures and low humidity. Cell performances improved at 120 °C with increasing silica content in the composite membrane containing the binary oxide. The binary oxide SiO2–ZnO2 (1–1) (10% w/w in Nafion®) gave the best performance. It is believed that the synergistic effects are heightened when both the oxides coexist in a single particle. The performance of binary oxide SiO2–ZnO2 (10%) was better than that of the physical mixture SiO2 + ZnO2 (10%) and higher than that of a membrane containing only one oxide in the same concentration (10%).

4.8.2 Interfacial characteristics Adjemian et al.199 examined the physicochemical characteristics of the various simple metal oxide composites, in particular the interfacial chemistry that occurs between the metal oxide particles and the Nafion® membrane. To investigate this, they processed titania samples using different treatments, degreasing or acidification in order to increase the surface hydroxyl content, and silylation of hydroxyl groups in order to reduce the surface hydroxyl content. In the first case, the treatments produced a membrane with improved PEMFC performance (resistivity value of 0.33 Ω cm2 and cell potential (at 600 mA/cm2) of 590 mV vs. 0.58 Ω cm2 and 582 mV for the untreated particles at 75% RH and 130 °C), suggesting surface contamination of the titania particles and that such surface contamination was detrimental to membrane performance. In the second case, the silylation led to an increase in resistivity (0.36 Ω cm2) and a decrease in cell potential (at 600 mA/cm2) to 400 mV under the same operating conditions, indicating


158

Polymer electrolytes SO3H

Si-OH

Water molecules Proton

Proton transport

Scheme 4.4 Schematic representation of the model involved in proton-conduction mechanism and the state of water in the membrane of SPAEEKK/silica hybrid membrane.

that surface hydroxyl groups are the reactive sites needed to provide the requisite interaction between the titania particles and the Nafion® polymer. These results clearly show that the OH of the oxide surfaces are involved in the conduction mechanism. Mechanism Guiver and co-workers202 studied sulphonated poly (arylene ether ether ketone ketone) (SPAEEKK)/silica (Scheme 4.3) and proposed a mechanism for hybrid membranes (Scheme 4.4). They assumed that the nonfreezing and the freezing bound water participates by the Grotthus mechanism, and the free water participates by both vehicle and Grotthus mechanisms. The free water is evaporated with increasing temperature and consequently the contribution of the vehicle mechanism decreases. To maintain proton conductivity with increasing temperature, the decrease in the contribution of the vehicle mechanism must be compensated by an increase in the contribution of the Grotthus mechanism. Proton conductivity increased with silica content in all temperature conditions. This suggests that the silica doped into the membrane enhanced the formation of proton conduction pathways due to molecular water absorption. Thus, the probable mechanism for new proton conduction due to the presence of silica is proton hopping between Si-OH and water molecules.202,203 Nogami et al.203 reported that the dissociated proton moves to a water molecule bound with an SiOH bond, forming the activated H2O:H+ state



159

(SiOH—H2O ⇒ SiO2 + H+:H2O). The proton from the activated H2O:H+ state dissociates to form a new activated state with a neighbouring H2O. They proposed that protons attached to protogenic surface functional groups of SiOH participate in the proton conduction process via the Grotthus mechanism.204

4.8.3 Physical/chemical characteristics of the matrix In addition to nanoparticle characteristics, the physical/chemical characteristics of the membranes have a strong impact on composite membrane properties. It was reported200 that the resistivity of the cell decreases as the equivalent weight and thickness of the Nafion® decreases in 10 wt% SiO2 Nafion® composites. For example, Nafion® 112 and 115 have the same equivalent weight of 1100 g of polymer per mol of sulphonic acid, while Nafion® 112 is 50 µm and Nafion® 115 is 125 µm thinner. The thinner Nafion® composite membrane (Nafion® 112) exhibits a lower resistivity (R = 0.53 Ω cm2 vs. 0.78 Ω cm2, 130 °C). Nafion® 105 and 115 have the same thickness of 125 µm, while Nafion® 105 has a higher sulphonate density (1000 equivalent weight versus 1100 equivalent weight). The lower equivalent weight Nafion® membrane (Nafion® 105) has a lower resistivity (R = 0.45 Ω cm2 vs. 1.3 Ω cm2). At a cell temperature of 130 °C, Nafion® 105 exhibits the lowest resistivity of the extruded Nafion® membranes, indicating that equivalent weight plays a larger role than membrane thickness in maintaining proton conductivity at 130 °C.

4.8.4 Metal oxide concentration Most publications193–196 reported that proton conductivity increased with metal oxide content under all temperature conditions. This suggests that the metal oxide doped into the membrane improves the formation of proton conduction pathways due to molecular water absorption. However, some authors observed146,205 an improvement in the mechanical properties of polymer electrolytes with the addition of nanofillers at optimal concentrations. The optimal concentration depended not only on the polymer and inorganic oxide structure, but also on the elaboration procedure. Adjemian et al.146 showed that 6% SiO2 by weight was the optimal percentage of silicon oxide for extruded Nafion®, while 10% by weight was optimal for recast Nafion®. An excess of the metal oxide phase causes not only low proton conductivity but also poor physical properties, which suggests that an excess of the metal oxide phase prevents the organic–inorganic hybrid membranes from forming an interpenetrated network. A composite membrane with sulphonated poly(decafluorobiphenyl-4-4′(hexafluoroisopropylidene) diphenol) (SDFF)206 (Scheme 4.3) + 4 wt% by


160


Tensile strength (MPa)

30 25 20

∗ Nafion

15 10 5 0

0

2

4 6 8 SiO2 content (%)

10

12

4.6 Tensile strength of membranes as a function of SiO2 content: (䊏) polymer membrane and composite membrane; (*) Nafion® film, adapted from Kim et al.206

Table 4.1 Proton conductivity of membrane containing various SiO2

Conductivity (S/cm)

SDFF

SDFF–SiO2 (4%)

SDFF–SiO2 (7%)

SDFF–SiO2 (10%)

SDFF–SiO2 (18%)

1.05 × 10−5

3.9 × 10−5

3.7 × 10−5

3 × 10−5

Very brittle

weight SiO2 shows significant improvement, especially at high temperature. A higher silica concentration reduces conductivity and mechanical properties (Fig. 4.6, Table 4.1).

4.9

Hygroscopic solid inorganic proton conductor composite polymer electrolytes

4.9.1 Hygroscopic solid inorganic proton conductor Hygroscopic solid inorganic proton conductors as membrane fillers increase the number of protonic carriers and thus improve the hydrophilic character of the membranes. Among the inorganic solid proton conductors,137,200,207,208 zirconium phosphates, heteropolyacids, metal hydrogen sulphate and a few others are of special interest for their potential role in developing high temperature composite membranes for PEMFCs.



161

Zirconium phosphates (ZP) have long been recognised as interesting inorganic ion exchangers,209 especially after their crystalline form Zr(HPO4)2 · 2H2O was prepared in the 1960s.210 This group of compounds can be expressed as MIV(RXO3) · nS, where M is a tetravalent metal such as Zr, Ti, Ce, Th or Sn; R is an inorganic or organic group such as –H, –OH, –CH3OH or (CH2)n; X is P or As; and S is a solvent, i.e. H2O. They form two types of layered structures, known as α and γ.137,211 These compounds exhibit good proton conductivity in a temperature range up to 300 °C. In the form of glassy plates or films, a conductivity of 10−2 S/cm at room temperature has been reported.212 The conductivity of Zr(HPO4) · 2H2O and H3O–Zr2(PO4)3 at 100 °C and 100% RH was determined to be about 10−1 S/cm213,214 and 7 × 10−7 S/cm,215 respectively. Heteropolyacids (HPA) possess strong acidity and high proton conductivity in their hydrated form,139 and their molecules (about 1 nm in diameter) can be regarded as nanoparticles. HPAs typically exist in hydrated phases, with the degree of hydration varying from 30 to 6 molecules of water (waters of hydration) per HPA molecule.216–218 The proton conductivities of these molecules drops respectively from approximately 0.06 to 2 × 10−5 S/ cm.214 The exact number of waters of hydration depends on the temperature and relative humidity of the environment. At 120 °C and 35% relative humidity, 12-phosphotungstic acid ((H3PW12O40 · nH2O) PWA) has approximately 6 waters of hydration.214 Detailed information about various HPA structures can be found in the literature.218–221

4.9.2 Composite polymer electrolytes based on zirconium phosphate Chalkova et al.222 studied the effects of ZP particle size, structure and distribution in a membrane on the performance of Nafion®/ZP composite membranes in PEMFCs at high temperatures and low RH. They synthesised Nafion®/ZP composite membranes with different structures containing in situ precipitated ZP, crystalline layer-structured α-zirconium hydrogen phosphate (Zr(HPO4) · 2H2O) (α-ZPL), and three-dimensional-network zirconium hydrogen phosphate H3O–Zr2(PO4)3 (ZPTD). The difference in performance of composite Nafion®/ZP membranes containing ZP in different structures was more pronounced at 120 °C than at 80 °C, particularly at 13 and 26% RH. The Nafion®/α-ZPL composite membranes demonstrated the best performance for all RH values. The advantage of these membranes over Nafion®/ZPTD and Nafion®/ZP in situ membranes was especially obvious at 13% RH. The current density obtained at a cell voltage of 0.6 V for the Nafion®/α-ZPL membrane at 13% RH was 10 times higher than that of an unmodified recast Nafion® membrane at 26% RH. Membranes with 10%


162


in situ precipitated ZP demonstrated much lower performance in a PEMFC than membranes with 10% α-ZPL or ZPTD. To enhance the performance of the Nafion®/ZP in situ precipitated composite membrane, the concentration of ZP nanoparticles has to be substantially increased. Improved performance of Nafion®/ZP membranes in a fuel cell were reported for membranes containing 23 and 36 mass% of ZP.219,223 These studies suggested that, first, the effect of pore constriction219,222 and/or, second, scaffold formation222,224 were determining factors for improving membrane performance. In the former case, it was possible that 10% ZP was not enough to fill all the available pores and induce a capillary condensation effect, which enhances water retention in the membrane. In the latter, 10% ZP could also be insufficient to form internal rigid scaffolding within the whole membrane, which would help to resist compression due to the sealing pressure in a fuel cell and facilitate water uptake by the membrane at a reduced RH.219 The difference in performance of Nafion®/ZP composite membranes containing α-ZPL and ZPTD can be attributed to the structure and surface properties of the inorganic particles and membrane morphology. Different structural modifications of ZP have different types of proton diffusion path, which affect proton conductivity through the inorganic material framework.225 In layered structures, the bulk diffusion of the hydronium ions occurs through the open cavities parallel to the layer,226 but the mobility of interior hydronium ions is about 104 times less than the mobility of surface (adsorbed) hydronium ions227 because the free water molecules on a crystal surface can easily rotate and contribute to ionic conduction by the Grotthus mechanism. It is difficult for hydronium ions to rotate in the interlayer spaces of ZP crystals due to strong hydrogen bonding with layered phosphate groups. As a result, proton transport in α-ZPL is dominated by surface transport.222,228 In a three-dimensional structure, conduction at room temperature is believed to occur via the Grotthus mechanism, mainly through grain boundaries. As the temperature increases, the surface water is lost, and grain boundary conduction decreases, being replaced by classical H+-ion hopping through the cavities inside the crystal structure.229 Particle distribution in composite membranes can also be regarded as a factor affecting membrane performance in a fuel cell. Chalkova et al.222 reported improved performance for α-ZPL, explained by the negative zeta potential of this filler, which provides electrostatic dispersion of particles in the acidic suspension during membrane casting and prevents aggregation and settling. The fact that the zeta potential of α-ZPL was more negative than that of ZPTD suggests that its surface phosphate groups undergo more complete dissociation (–POH = −PO− + H+) in an acidic environment. Composite Nafion®–Teflon®-α-ZPL membranes, prepared by incorporating Nafion® and α-ZPL into the Teflon® matrix, have been developed



163

and tested in an H2/O2 fuel cell at 120 °C, ambient pressure and 35% RH by Si et al.230 Improved performance was due to both the incorporation of α-ZPL filler and reduced membrane thickness. Hydrogen crossover for the composite membrane was about 2.5 times higher than for a commercial Nafion® 1135 membrane (2 mA/cm2 with Nafion®-Teflon®-α-ZPL vs. 0.8 mA/cm2 Nafion® 1135) at 120 °C, which is still low enough not to cause significant performance loss. Kim et al.231 prepared Nafion®/zirconium sulphophenyl phosphate (ZSPP) composite membranes. ZSPP was precipitated by the reaction of Zr4+ ion and m-sulphophenyl phosphonic acid (SPP) with a stoichiometric ratio P/Zr = 2. They found the conductivity of ZSPP to be about 10−2 S/cm, much higher than ZPL.216 The conductivity increased with the crystallinity of ZSPP. The increased conductivity may be related to more conduction channels through the interlayer of crystalline ZSPP. When 12.5 wt% ZSPP particles was incorporated in Nafion®, it showed quite high conductivity, 0.07 S/cm at 140 °C.

4.9.3 Composite polymer electrolytes based on heteropolyacids Nafion® recast membranes filled with silica-immobilised phosphotungstic acid, PWA/SiO2 (30/70 weight ratio) and silicotungstic acid, SWA/SiO2 (45/65 weight ratio), so that the silica content was in all cases 3 wt%, were prepared by bulk mixing a 5% Nafion® solution with the powdered filler and then thermally treating it at 160 °C.232 Tests in direct methanol fuel cells at 145 °C showed that the PWA/SiO2 membrane had better electrochemical characteristics at high current densities compared with membranes loaded with SWA/SiO2 or only 3 wt% SiO2. Peaks of power density of 400 and 250 mW/cm2 were reached by feeding the cell with oxygen and air. For the same fillers, Tian and Savadogo233 reported better improvements for Nafion® +PWA/SiO2 membranes obtained by casting from DMF (0.3 A/ cm2 at 0.6 V) solution than for those obtained by aqueous solution (0.15 A/ cm2 at 0.6 V). Hongwei et al.234 found that sulphonated poly(phthalazinone ether ketone) (SPPEK)/12-PWA composite membranes showed improved conducting properties, and the SPPEK/ZrP composite membrane showed similar conducting properties to the pure SPPEK membrane. The SPPEK/ PWA composite membrane showed the highest proton conductivity, 0.17 S/ cm at 80 °C. The difference in conducting properties between the SPPEK/ PWA composite membrane and the SPPEK/ZrP composite membrane can be explained by their conduction mechanism. Kim et al.235 obtained hybrid PEMs based on polytetramethylene oxide (PTMO) and polydimethysiloxane (PDMS) using the sol–gel technique.


164


The organic–inorganic interfaces were formed by sol–gel processing the end groups (OH) of organic parts with reactive inorganic parts, leading to hybrids at the molecular scale. The PWA was incorporated using sol–gel processes and the effects of PWA doping concentration on the thermal stability and protonic conductivity were discussed.236–239 The hybrid membranes showed superior thermal and mechanical properties because of the temperature-tolerant organic–inorganic moieties and the nanostructure of the hybrid matrix. The maximum proton conductivities for zirconium phosphate–PTMO membranes and titania–PTMO–40 wt% PWA hybrids were 2 × 10−3 and 8 × 10−2 S/cm, respectively, under saturated humidity conditions such that PWA or phosphate are bonded with hydrophilic metal phases. The cell performance of the membrane electrode assembly (MEA) with zirconium phosphate–PTMO membrane showed a maximum power density of 13 mW/cm2 and the titania–PTMO–PWA composite membrane with Nafion® supporter showed a maximum power density of 30 mW/cm2. Moreover, the organic–inorganic hybrid membrane showed stable performance up to 130 °C under saturated humidity conditions. Smitha et al.240 prepared solution cast films of composite membranes containing 40 wt% PWA and 60 wt% polysulphone (PSF) and SPSF. The PWA–40/SPSF composite membranes showed an increase in Tg, suggesting an intermolecular interaction between PWA and SPSF, while a reduction in the Tg of the PWA–40/PSF composites containing 40% PWA was found, indicating the absence of physicochemical interactions. Fourier transform infrared (FTIR) band shifts of this composite showed that sulphonic acid groups interact with both bridging tungstic oxide and terminal tungstic oxide, implying that the hydrogen bonding is attributed solely to miscibility and physicochemical properties. Incorporation of PWA into the sulphonated polymer significantly reduced the water swelling behaviour of the composite membrane. The membranes exhibited high conductivity (0.09 S/ cm) at room temperature, and a maximum of about 0.14 S/cm at 120 °C. Another sulphonated polysulphone composite membrane, obtained by bulk mixing of powdered H3Sb3P2O14 with the polymer solution in dichloroethane-isopropanol, exhibited higher conductivity, lower water swelling and lower gas permeability than unmodified SPSF. The conductivity of these membranes showed a maximum for 8 wt% filler. For this loading, the conductivity of SPSF with an ion exchange capacity of 1.1 mequivalent/g was enhanced by a factor of 10 and reached 2 × 10−2 S/cm at 80 °C and 100% RH.169,170

4.10

Self-humidifying composite electrolytes

Another class of inorganic/organic composite membranes, called ‘selfhumidifying’ polymer electrolyte membranes, have been developed.241–244



165

These membranes contain highly dispersed nanosized Pt, which stimulates catalytic recombination of H2 and O2 inside the membranes, and a hygroscopic metal oxide, which adsorbs the water produced at the Pt particles and at the cathode. Self-humidifying membranes, in particular Nafion®–Pt– SiO2, Nafion®–Pt–TiO2 and Nafion®–Pt–zeolite,245,246 exhibited significantly higher water uptake and lower fuel crossover than unmodified Nafion® membranes. However, it is difficult to have uniformly distributed Pt particles in hygroscopic materials made by conventional methods such as electroless plating, sputtering or Pt black doping. There is also the possibility of a short circuit forming in the membrane because of non-uniformly dispersed Pt particles. In order to avoid a short circuit after incorporating Pt or Pt/C particles, another type of self-humidifying membrane has recently been developed. This is a thin double-layer composite membrane consisting of one layer of Pt/C catalyst-dispersed Nafion® and another layer of recast Nafion®. Researchers have also developed three-layered membrane structures to resolve this problem.247–249 Yang et al.249 reported that, with dry H2 and O2, the composite membrane showed about 90% of the performance obtained with the humidified reactant under similar conditions. However, the fabrication processes for these membranes are too complex for widespread use. To improve PEMFCs with dry reactants, an inorganic/organic self-humidifying membrane based on a SPEEK hybrid with Cs2.5H0.5PW12O40 supported Pt catalyst (Pt-Cs2.5 catalyst) was investigated.250 The Pt-Cs2.5 catalysts incorporated in the SPEEK matrix provide sites for catalytic recombination of permeable H2 and O2 to form water, and meanwhile avoid short circuits through the whole membrane thanks to the insulating properties of the Cs2.5H0.5PW12O40 support. Furthermore, the Pt-Cs2.5 catalyst can adsorb the water and transfer the proton inside the membrane. The physicochemical and electrochemical properties, such as ion exchange capacity (IEC), water uptake and proton conductivity, between the plain SPEEK and the SPEEK/ Pt-Cs2.5 composite membrane were compared. Additive stability measurements indicated that the Pt-Cs2.5 catalyst showed improved stability in the SPEEK matrix compared to the PWA particles in the SPEEK matrix. Single cell tests employing the SPEEK/Pt-Cs2.5 self-humidifying membrane and the plain SPEEK membrane under wet or dry operating conditions and primary 100 H fuel cell stability measurements were also conducted by these authors. The performance of cells using self-humidifying composite membranes has improved considerably compared with that of composite membranes without self-assembled Pt nanoparticles. However, the performance of selfhumidifying composite membranes without external humidification is still inferior to that of cells with external humidification. At a current density of 900 mA/cm2, cell potential with external humidification at 80 °C was found


166


to be 0.603 V, higher than for the cell without external humidification (0.402 V). Similar improvements in cell performance were also observed at an operating temperature of 60 °C. This indicates that the self-humidifying effect of the self-assembled Pt nanoparticle layer may not be sufficient to humidify the membrane under the conditions investigated in the literature. As a result, the advantages of self-humidifying PEMs lie in their ability to reduce the need for external humidification when excellent performance is required.

4.11

Future trends

For polymer electrolytes based on polyether used in lithium batteries, it is clear that the addition of nanofillers has a positive impact on ionic conductivity at temperatures below the membrane composite melting temperature. At temperatures above the membrane melting temperature, the improvement in ionic conductivity is questionable. The use of nanocharges with high dielectric constant or/and which induce specific interactions with the anion might be worth investigating. Basic research is needed to confirm the conduction mechanism in nanocomposite electrolytes. When studying the effects of fillers, the nanofillers themselves and the composite electrolytes must be dried carefully to ensure that any observed changes are not due to the effects of water trapped by the fillers. Mechanical improvement can be obtained by incorporating large aspect ratio fillers, which may be an advantage if they allow the thickness of the electrolyte to be reduced, thus lowering internal resistance. However, more research is needed in this field in order to get a good compromise between mechanical properties and ionic conductivity. In gel polymer electrolytes, the addition of fillers can have beneficial effects on both liquid electrolyte uptake and mechanical properties. However, as for SPEs, composite gel polymer electrolytes must be performed carefully in order to remove water trapped in the fillers. The current challenge for PEMFC is to raise the working temperature above 80 °C. Composite membranes are a potential solution. The addition of inorganic fillers induces important improvements in water retention at high temperature, conductivity, cell resistivity, mechanical properties, etc. These improvements are related to filler concentration, structure and size, interfaces, polymer matrix and membrane characteristics. It is difficult to compare these ionomer/filler composites because their performances depend on the electrolyte preparation and testing conditions (RH, temperature, etc.). H2/O2(air) cells based on composite polymer electrolytes have been successfully operated at temperatures up to 120 °C under ambient pressure, and up to 150 °C under pressures of 3–5 atm, but more research



167

is needed into their application in high temperature PEMFC in order to establish how long these composite electrolytes continue to exhibit high performance at low relative humidity and high temperature. An interesting idea is to produce water within a membrane by catalytic recombination of H2 and O2 on a platinum catalyst dispersed in the membrane. One of the difficulties in developing such electrolytes is preventing short circuits, which requires a uniform dispersion of Pt inside the membrane, another is the prohibitive price. Since the water produced within the membrane is insufficient to ensure good performance, external humidification is essential. The advantage of self-humidifying PEMs lies in their ability to reduce the humidity of the reactant gases. Anhydrous proton conducting electrolytes, such as electrolytes based on proton conducting ionic liquids, might be better adapted for PEMFC working at temperatures higher than 120 °C.

4.12

Sources of further information and advice

Books on composite electrolytes and polymer composites can be consulted for complementary information. Indeed, the properties of composite materials are associated with both components, the matrix and the charge, and also with interfacial properties. Some interesting data and information on experimental investigations in the composite polymer field can be found in the following books: •

•

•

•

•

Hydrocarbon Polymer Electrolytes for Fuel Cell Application (J. Qiao and T. Okada, 2008, Nova Science Publishers) focuses on composite membranes for direct methanol fuel cells. Polymer Electrolyte Fuel Cell Durability (F.N. Büchi, M. Inaba, and T.J. Schmidt, (Eds), 2009, Springer) covers one of the most important aspects of fuel cells, durability, and gives pertinent information on polymer membrane degradation, which may be an important orientation for future studies on composite membranes. Composite Polymeric Electrolytes (W. Wieczorek and M. Siekierski, 2007, Springer) reviews the considerable experience of W. Wieczorek and his team in work on composite electrolytes for lithium batteries. Polymer Composite Materials Interface Phenomena and Processes (Y. Ivanov, V. Cheshkov and M. Natova, 2001, Kluwer Academic) evaluates the properties of polymer composites in relation to acid–base properties. Handbook of Polymer Blends and Composites (C. Vasilen and A.K. Kulshrehtha, 2003, Woodhead Publishing) provides an overwiew of the theory and pratice of polymer blends and composites.


168


4.13

References

1. armand, m., chabagno j.m., duclot, m. 1979, in Fast Transport in Solids (Edited by P.Vashishta), vol. 131, North-Holland, New York. 2. ballard, d.g.h., cheshire, p., mann, t.s., przeworski, j.e. 1990, Macromolecules, 23, 1256. 3. allcock, h.r., austin, p.e., neenan, t.x., sisko, j.t., blonsky, p.m., shriver, d.f. 1986, Macromolecules, 19, 1508. 4. hamaide, t., goux, a., llauro, m.f., spitz, r., guyot, a.a. 1996, Makromol. Chem., 237, 55. 5. cowie, j.m.g., martin, a.c.s., firth, a.m. 1988, Br. Polym., 20, 247. 6. abraham, k.m., alamgir, m., perrotti, s.j. 1988, J. Electrochem. Soc., 135, 535. 7. hall, p.g., davis, g.r., macintyre, j.e., ward, i.m., bannister, d.j., le brocq, k.m.f. 1986, Polym. Commun., 27, 98. 8. allcock, h.r., olmeijer, d.l., oconnor, s.j.m. 1998, Macromolecules, 31, 753. 9. fujinami, t., buzoujima, y. 2003, J. Power Sources, 119–121, 438. 10. sadoway, d.r. 2004, J. Power Sources, 129, 1. 11. killis, a., lenest, j.-f., gandini, a. cheradame, h. 1981, J. Polym. Sci., 19, 1073. 12. alloin f., sanchez j-y. 1995, Electrochim. Acta, 40, 13–14, 2269. 13. nishimoto, a., watanabe, m., ikeda, y., kohjiya, s. 1998, Electrochim. Acta, 1177. 14. zhang, z., sherlock, d., west, r.k., amine, l., lyons, j. 2003, Macromolecules, 36, 9176. 15. weston, j.e., steele, b.c.h. 1982, Solid State Ionics, 7, 81. 16. yang, l., heather, f., blease, t.g., thompson, r.i.g. 1996, Eur. Polym. J., 32, 535. 17. morlat, s., gardette, j-l. 2001, Polymer, 42, 6071. 18. chen, k.q., shen y., li h.i., xu, x. 1985, J. Macromolecular Sci. Chem., A22, 455. 19. usuki, a., kojima, y., kawasumi, m., okoda, a., fukushima, y., kurauchi, t. 1993, J. Mater. Res., 8, 1179. 20. lebaron, p.c., wang, z., pinnavaia, t.j. 1999, Appl. Clay Sci., 15, 11. 21. ratna, d., manoj, n.r., singh raman, r.k., varley, r., simon, g.p. 2003, Polym. Int., 52,1403. 22. kawasumi, m. 2004, J. Polym. Sci. A: Polym. Chem., 42, 819. 23. ruiz-hitzky, e., aranda, p. 1990, Adv. Mater, 2, 545. 24. shen, z., simon, g.p., cheng, y.b. 2002, Polym. Eng. Sci., 42, 2369. 25. wong, s., vaia, r.a., giannelis, e.p., zax, d.b. 1996, Solid State Ionics, 86, 547. 26. vaia, r.a., giannelis, e.p. 1997, Macromolecules, 30, 8000. 27. kurian, m., galvin, m.e., trapa, p.e., sadoway d.r., mayes, a.m. 2005, Electrochim. Acta, 50, 2125. 28. fan, l., nan, c.w., dang, z. 2002, Electrochim. Acta, 47, 3541. 29. aranda, p., mosqueda, y., perez-cappe, e., ruiz-hitzky, e. 2003, J. Polym. Sci. B: Polym. Phys., 41, 3249. 30. doeff, m.m., reed, j.s. 1998, Solid State Ionics, 113–115, 109. 31. aranda, p., ruiz-hitzky, e. 1994, Acta Polym., 45, 59. 32. vaia, r.a., vasudevan, s., kraweic, w., scanlon, l.g., giannelis, e.p. 1995, Adv. Mater., 7, 154. 33. chen, w., xu, q., yuan, r.z. 2000, Mater. Sci. Eng. B, 77, 15. 34. ratna, d., divekar, s., patchaiappan, s., samui, a.b., chakraborty, b.c. 2007, Polym. Int., 56, 900. 35. loyens, w., jannasch, p., maurer, f. 2005, Polymer, 46, 903.


Composite polymer electrolytes for electrochemical devices 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72.

169

chen, b., evans, j.r.g. 2004, J. Phys. Chem. B, 108, 14986. loyens, w., maurer, f.h.j., jannasch, p. 2005, Polymer, 46, 7334. strawhecker, k.e., manias, e. 2003, Chem. Mater., 15, 844. ogata, n., kawakage, s., ogihara, t. 1997, Polymer, 38, 5115. loyens, w., jannasch, p., maurer, f. 2005, Polymer, 46, 915. loyens, w., maurer, f.h.j., jannasch, p. 2005, Polymer, 46, 7334. chen, h.w., chang, f.c. 2001, Polymer, 42, 9763. liao, c.s., ye, w.b. 2004, Mater. Chem. Phys., 88, 84. hackett, e., manias, e., giannelis, e.p. 2000, Chem. Mater., 12, 2161. yang, d.-k., zax, d.b. 1999, J. Chem. Phys., 110, 5325. leroux, f., besse, j.p. 2001, Chem. Mater., 13, 3507. yang, q.z., sun, d.j., zhang, c.g., wang, x.j., zhao, w.a. 2003, Langmuir, 19, 5570. liao, c.s.,ye, w.b., 2004, Electrochim. Acta, 49, 4993. leroux, f., aranda, p., besse, j-p., ruiz-hitzky, e. 2003, Eur. J. Inorg. Chem., 1242. nagasubramanian, g., attia, a.i., halpert, g., peled, e. 1993, Solid State Ionics, 67, 51. capuano, f., croce, f., scrosati, b. 1991, J. Electrochem. Soc., 138, 1918. croce, f., scrosati, b. 1992, Chem. Mater., 4, 1134. borghini, m.c., mastragostino, m., passerini, s., scrosati, b. 1995, J. Electrochem. Soc., 142, 2118. gang, w., roos, j., brinkmann, d., capuano, f., croce, f., scrosati, b. 1992, Solid State Ionics, 53–56, 1102. khan, s.a., baker, g.l., colson, s. 1994, Chem. Mater., 6, 2359. fan, j., fedkiw, p.s. 1997, J. Electrochem. Soc., 144, 399. matsuo, y., kuwano, j. 1995, Solid State Ionics, 79, 295. capiglia, c., mustarelli, p., quartarone, e., tomasi, c., magistris, a. 1999, Solid State Ionics, 118, 73. croce, f., curini, r., martinelli, a., persi, l., ronci, f., scrosati b., caminiti, r. 1999, J. Phys. Chem. B, 103, 10632. dai, y., wang, y., greenbaum, s.g., bajue, s.a., golodnitsky, d., ardel, g., straus, e., peled, e. 1998, Electrochim. Acta, 43, 1557. croce, f., appetecchi, g.b., persi, l., scrosati, b. 1998, Nature, 394, 456. wieczorek, w., such, k., chung, s.h., stevens, j.r. 1994, J. Phys. Chem., 98, 9047. siekierski, m., wieczorek, w., przyluski, j. 1998, Electrochim. Acta, 43, 1339. wieczorek, w., zalewska, a., raducha, d., florjanczyk, z., stevens, j.r. 1998, J. Phys. Chem. B, 102, 352. golodnitsky, d., ardel, g., peled, e. 1996, Solid State Ionics, 85, 231. golodnitsky, d., ardel, g., strauss, e., peled, e., lareah, y., rosenberg, y. 1997, J. Electrochem. Soc., 144, 3484. dai, y., greenbaum, s.g., golodnitsky, d., ardel, g., strauss, e., peled, e., rosenberg, y. 1998, Solid State Ionics, 106, 25. best, a.s., ferry, a., macfarlane, d.r., forsyth, m. 1999, Solid State Ionics, 126, 269. kumar, b., scanlon, l.g. 2000, J. Electrochem. Soc., 5, 127. walls, h.j., zhou, j., yerian, j.a., fedkiw, p.s., khan s.a., stowe, m.k., baker, g.l. 2000, J. Power Sources, 89, 156. johansson, p., ratner, m.a., shriver d.f. 2001, J. Phys. Chem. B, 105, 9016. croce, f., persi, l., scrosati, b., serraino-fiory, f., plichta, e., hendrickson, m.a. 2001, Electrochim. Acta, 46, 2457.


170


73. xiong, h-m., zhao, x., chen, j-s. 2001, J. Phys. Chem. B, 105, 10169. 74. dissanayake, m.a.k.l., jayathilaka, p.a.r.d., bokalawala, r.s.p., albinsson, i., mellander, b.e. 2003, J. Power Sources, 119–121, 409. 75. liu, y., lee, j.y., hong, l. 2003, J. Appl. Polym. Sci., 89, 2815. 76. kerr, j.b., han, y.b., liu, g., reeder, c., xie, j., sun, x. 2004, Electrochim. Acta, 50, 235. 77. xiong, h-m., wang, z-d., xie, d-p., cheng l., xia, y-y. 2006, J. Mater. Chem., 16, 1345. 78. chen-yang, y.w., wang, y.l., chen, y.t., li, y.k., chen, h.c., chiu, h.y. 2008, J. Power Sources, 182, 340–348. 79. xiong, h-m., wang, z-d., liu, d-p., chen, j-s., wang, y-g., xia, y-y. 2005, Adv. Funct. Mater, 15, 1751. 80. croce, f., persi, l., ronci, f., scrosati, b. 2000, Solid State Ionics, 135, 47. 81. krawiec, w., scanlon, l.g., fellner, j-p., vaia, r.a., vasudevan, s., giannelis, e.p. 1995, J. Power Sources, 54, 310. 82. nookala, m., kumar, b., rodrigues s. 2002, J. Power Sources, 111, 165. 83. best, a.s., adebahr, j., jacobsson, p., macfarlane, d.r., forsyth, m. 2001, Macromolecules, 34, 4549. 84. kim, y.w., lee, w., choi, b.k. 2000, Electrochim. Acta, 45, 1473. 85. choi, b.k., shin, k.h. 1997, J. Appl. Electrochem., 27, 365. 86. marcinec, m., bac, a., lipka, p., zeleska, a., zukowska, g., borskowska, r., wieczorek, w. 2000, J. Phys. Chem. B, 104, 11088. 87. wieczorek, w., lipka, p., zukowska, g., wycislik, h. 1998, J . Phys. Chem. B, 102, 6968. 88. marcinek, m., zalewska, a., zukowska, g., wieczorek w. 2000, Solid State Ionics, 136–137, 1175. 89. xi , j., qiu, x., zheng, s., tang, x. 2005, Polymer, 46, 5702. 90. chauvin, c., alloin, f., iojoiu, c., sanchez, j-y. 2006, Electrochim. Acta, 51, 5876. 91. wieczorek, w., florjanczyk, z., stevens, j.r. 1995, Electrochimica ACM, 40, 2251. 92. xie, j., duan, r.g., han, y., kerr, j.b. 2004, Solid State Ionics, 175, 755. 93. tsagaropoulos, g., eisenberg, a. 1995, Macromolecules, 28, 6067. 94. karlsson, c., best, a.s., swenson, j., howells, w.s., börjesson l. 2003, J. Chem. Phys., 9, 4206. 95. forsyth, m., macfarlane, d.r., best, a., adebahr, j., jacobsson, p., hill, a.j. 2002, Solid State Ionics, 147, 203. 96. chung, s.h., wang, y., greenbaum, s.g., marcinek, m., persi, l., croce, f., wieczorek, w., scrosati, b. 2001, J. Phys. Condens. Mater., 13, 11763. 97. cheung, i.w., chin, k.b., greene, e.r., smart, m.c., abbrent, s., greenbaum, s.g., prakash, g.k.s., surampudi, s. 2003, Electrochim. Acta, 48, 2149. 98. mustarelli, p., capiglia, c., quartarone, e., tomasi, c., ferloni, p. linati, l. 1999, Phys. Rev. B, 10, 7228. 99. bloise, a.c., tambelli, c.c., franco, r.w.a., donoso, j.p., magon, c.j., souza, m.f., rosario, a.v., pereira, e.c. 2001, Electrochim. Acta, 46, 1571. 100. singh, th.j., bhat, s.v. 2004, J. Power Sources, 129, 280. 101. vogel, m., herbers, c., koch, b. 2008, J. Phys. Chem. B, 112, 11217. 102. bloise, a.c., donoso, j.p., magon, c.j., rosario, a.v., pereira, e.c. 2003, Electrochim. Acta, 48, 2239. 103. frech, r., huang, w. 1995, Macromolecules, 28, 1246.



171

104. peterson, g., jacobsson, p., torell, l.m. 1992, Electrochim. Acta, 37, 1495. 105. schantz, s., torell, l.m., stevens, j.r. 1988, J. Appl. Phys., 64, 2038. 106. best, a.s., ferry, a., macfarlane, d.r., forsyth, m. 1999, Solid State Ionics, 126, 269. 107. rey, i., lassegues, j.c., grondin, j., servant, l. 1998, Electrochim. Acta, 43, 1505. 108. jayathilaka, p.a.r.d., dissanayake, m.a.k.l., albinsson, i., mellander, b-e. 2002, Electrochim. Acta, 47, 3257. 109. swierczynski, d., zalewska, a., wieczorek, w. 2001, Chem. Mater., 13, 1560. 110. singh, th.j., bhat, s.v. 2004, J. Power Sources, 129, 280. 111. xi, j., qiu, x., zhu, w., tang, x. 2006, Micropor. Mesopor. Mater., 88, 1. 112. kim, s., park, s-j. 2007, Electrochim. Acta, 52, 3477. 113. appetecchi, g.b, croce, f., hasson, j., scosati, b., salomon, m., cassel, f. 2003, J. Power Sources, 114, 105. 114. tokuda, h., watanabe, m. 2003, Electrochim. Acta, 48, 2085. 115. d’epifanio, a., serraino fiory, f., licoccia, s., traversa, e., scrosati, b., croce f. 2004, J. Appl. Electrochem., 403–408, 403. 116. bruce, p.g., evans, j.e., vincent, c.a. 1987, Polymer, 28, 2324. 117. appetecchi, g.b., scaccia, s., passerini, s. 2000, J. Electrochem. Soc., 147, 4448. 118. wen, z., wu, m., itoh, t., kudo, m. 2002, Solid State Ionics, 148, 185. 119. wen, z.y., itho, t., hirata, n., ikeda, m., kudo, m., yamamoto, o. 2000, J. Power Sources, 90, 20. 120. azizi samir, m.a.s., alloin, f., gorecki, w., sanchez, j-y., dufresne, a. 2004, J. Phys. Chem. Part B, 108, 10845. 121. azizi samir, m.a.s., alloin, f., sanchez, j-y., dufresne, a. 2004, Macromolecules, 37, 4839. 122. azizi samir, m.a.s., alloin, f., sanchez, j-y., dufresne, a. 2004, Polymer, 45, 4149. 123. azizi samir, m.a.s., alloin, f., sanchez, j-y., el kissi n., dufresne a. 2004, Macromolecules, 37, 1386. 124. samir, m.a.s., montero mateos, a., alloin, f., sanchez j-y., dufresne, a. 2004, Electrochim. Acta, 49, 4667. 125. gorecki, w., jeannin, m., belorizki, e., roux, c., armand, m. 1995, J. Phys. Condens. Mater., 7, 6823. 126. jacob, m.m.e., hackett, e., giannelis, e.p. 2003, J. Mater. Chem., 13, 1. 127. meneghetti, p., qutubuddin, s., webber, a. 2004, Electrochim. Acta, 49, 4923. 128. wang, m., zhao, f., guo, z., dong, s. 2004, Electrochim. Acta, 49, 3595. 129. ahmad, s., bohidar, h.b., ahmad, s., agnihotry, s.a. 2006, Polymer, 47, 3583. 130. nan, c.w., fan, l., lin, y., cai, q. 2003, Phys. Rev. Lett, 91, 266104. 131. wang, m., dong, s. 2007, J. Power Sources, 170, 425. 132. liu, h.j., hwang, j.j., chen-yang, y.w.j. 2002, Polymer Sci.: Part A: Polymer Chem., 40, 3873. 133. meneghetti, p., qutubuddin, s. 2006, Thermochim. Acta, 442, 74. 134. wu, c-g., lua, m-i., tsai, c-c., chuang, h-j. 2006, J. Power Sources, 159, 295. 135. abbrent, s., chung, s.h., greenbaum, s.g., muthu, j., giannelis, e.p. 2003, Electrochim. Acta, 48, 2113. 136. kreuer, d. 2000, Solid State Ionics, 136–137, 149. 137. kreuer, k.d. 1996, Chem. Mater., 8, 610. 138. pourcelly, g., gavach, c. 1992, in Proton Conductors, Solids, Membranes, and Gels – Materials and Devices (Edited by P. Colomban), Cambridge University Press, New York.


172


139. marechal, m., souquet, j.-l., guindet, j., sanchez, j.-y. 2007, Electrochem. Comm., 9(5), 1023. 140. alberti, g., casciola, m. 2003, Annu. Rev. Mater. Res., 33, 129. 141. kumara, b., fellner, j.p. 2003, J. Power Sources, 123, 132. 142. li, q., he, r., jensen, j.o., bjerrum, n.j. 2003, Chem. Mater., 15, 4896. 143. watanabe, m., uchida, h., seki, y., emori, m. 1996, J. Electrochem. Soc., 143, 3847. 144. antonucci, p.l., arico a.s., creti p., ramunni e., antonucci v. 1999, Solid State Ionics, 125, 431. 145. miyake, n., wainright j.s., savinell r.f. 2001, J. Electrochem. Soc., 148, A898. 146. adjemian, k.t., lee s.j., srinivasan, s., benziger, j., bocarsly, a.b. 2002, J. Electrochem. Soc., 149, A256. 147. yang, c., costamagna, p., srinivasan, s., benziger, j., bocarsly, a.b. 2001, J. Power Sources, 103, 1. 148. costamagna, p., yang, c., bocarsly, a.b., srinivasan, s. 2002, Electrochim. Acta, 47,1023. 149. boysen, d.a., uda, t., chisholm, c.r.i., haile, s.m. 2004, Science, 303, 68. 150. fermandez, r.e. 1999, in Polymer Data Handbook (Edited by J.E. Mark), Oxford University Press, New York, 233. 151. fontanella j.j., wintersgill, m.c., wainright, j.s., savinell, r.f., litt, m. 1998, Electrochim. Acta, 43, 1289. 152. bruijn, d., frank a., makkus, r.c., mallant, r.k.a.m., janssen g.j.m. 2007, Adv. in Fuel Cells, 1, 235. 153. li, q., he, r., jensen, j.o., bjerrum, n.j. 2004, Fuel Cells, 4, 147. 154. greaves, t.l., drummond, c. 2008, J. Chem. Rev., 108, 206. 155. fernicola, a., panero, s., scrosati, b. 2008, J. Power Sources, 178, 591. 156. iojoiu, c., martinez, m., hanna, m., molmeret y., cointeaux, l., leprêtre, j-c., el kissi, n., guindet, j., judeinstein, p., sanchez , j-y., cointeaux, l. 2008, Polym. Adv.Tech., 19, 1406. 157. miyake, n., wainright, j.s., savinell, r.f. 2001, J. Electrochem. Soc., 148, A898. 158. miyake, n., wainright, j.s., savinell, r.f. 2001, J. Electrochem. Soc., 148, A905. 159. tung, s-p., hwang, b-j. 2007, Fuel Cells, 7, 32. 160. liang, z.x., zhao, t.s., prabhuram, j. 2006, J. Membr. Sci, 283, 219. 161. chen, z., holmberg, b., li, w., wang, x., deng, w., munoz, r., yan, y. 2006, Chem. Mater., 18, 5669. 162. baglio, v., di blasi, a., aricò, a.s., antonucci, v., antonucci, p.l., nannetti, f., tricoli, v. 2005, Electrochim. Acta, 50, 5181. 163. thomassin, j.-m., pagnoulle, c., caldarella, g., germani, a., jerome, r. 2005, Polymer, 46, 11389. 164. baglio, v., di blasi, a., aricò, a., antonucci, v., antonucci, p.l., trakanprapai, c., esposito, v., licoccia, s., traversa, e. 2005, J. Electrochem. Soc., 152, A1373. 165. rhee, c.h., kim, y., lee, j.s., kim, h.k., chang, h. 2006, J. Power Sources, 159, 1015. 166. satterfield, m.b., majsztrik, p.w., ota, h., benziger, j.b., bocarsly, a.b. 2006, J. Polym. Sci. B, 44, 2327. 167. bauer, f., willert-porada, m. 2006, Fuel Cells, 6, 261. 168. amarilla, j.m., rojas, r.m., rojo, j.m., cubillo, m.j., linares, a., acosta, j.l. 2000, Solid State Ionics, 127, 133. 169. genova-dimitrova, p., baradie, b., foscallo, d., poinsignon, c., sanchez, j.y. 2001, J. Membr. Sci., 185, 59.



173

170. baradie, b., poinsignon, c., sanchez, j.y., piffard, y., vitter, g., bestaoui, n., foscallo, d., denoyelle, a., delabouglise, d., vaujany, m. 1998, J. Power Sources, 74, 8. 171. bonnet, b., jones, d.j., roziere, j., tchicaya, l., alberti, g., casciola, m., massinelli, l., baner, b., peraio, a., ramunni, e. 2000, J. New Mat. Electrochem. Syst., 3, 87. 172. zaidi, s.m.j., mikhailenko, s.d., robertson, g.p., guiver, m.d., kaliaguine, s. 2000, J. Membr. Sci., 173, 17. 173. nunes, s.p., ruffmann, b., rikowski, e., vetter, s., richau, k. 2002, J. Membr. Sci., 203, 215. 174. mikhailenko, s.d., zaidi, s.m.j., kaliaguine, s. 2001, Catal. Today, 67, 225. 175. roziere, j., jones, d.j., tchicaya-bouckary, l., bauer, b. 2000, Int. Patent No. WO 02/05370. 176. navarra, m.a., fernicola, a., panero, s., scrosati, b. 2006, J. Electrochem. Soc., 153, A1284. 177. croce, f., hassoun, j., tizzoni, c., scrosati, b. 2006, Electrochem. Commun., 8, 1125. 178. peled, e., duvdevani, t., melman, a. 1998, Electrochem. Solid-State Lett., 1, 210. 179. shen, j., xi, j., zhu, w., chen, l., qiu, x. 2006, J. Power Sources, 159, 894. 180. honma, i., takeda, y., bae, j.m. 1999, Solid State Ionics, 120, 255. 181. nakajima, h., honma, i. 2002, Solid State Ionics, 148, 607. 182. honma, i., nomura, s., nakajima, h. 2001, J. Membr. Sci., 185, 83. 183. nakajima, h., nomura, s., sugimoto, t., nishikawa, s., honma, i. 2002, J. Electrochem. Soc., 149, A953. 184. staiti, p., minutoli, m., hocevar, s. 2000, J. Power Sources, 90, 231. 185. staiti, p., minutoli, m. 2001, J. Power Sources, 94, 9. 186. staiti, p. 2001, Mater. Lett., 47, 241. 187. alberti, g., casciola, m., costantino, u., levi, g. 1978, J. Membr. Sci., 3, 179. 188. alberti, g., casciola, m., costantino u. 1983, J. Membr. Sci., 16, 137. 189. tchicaya-bouckary, l., jones, d.j., roziere, j. 2002, Fuel Cell, 2, 40. 190. deng, q., moore, r.b., mauritz, k.a. 1995, Chem. Mater., 7, 12, 2259. 191. grot, w., rajendran, g. 1996. Int. Patent No. WO 96/29752. 192. mauritz, k.a., warren, r.m. 1989, Macromolecules, 22, 1730. 193. nakamura, o., kodama, t., ogino, i., miyake, y. 1979, Chem. Lett., 1, 17. 194. honma, i., nakajima, h., nomura, s. 2002, Solid State Ionics, 154–155, 707. 195. nakajima, h., nomura, s., sugimoto, t., nishikawa, s., honma, i. 2002, J. Electrochem. Soc., 149, A953. 196. lavrencic, s., stangar, u., orel, b., vince, j., jovanovski, v., spreizer, h., surca vuk, a., hocevar, s. 2005, J. Solid State Electrochem., 9, 106. 197. watanabe, m., uchida, h., seki, y., emori paul stonehart m. 1996, J. Electrochem. Soc., 143, 3847. 198. antonucci, p.l., arico, a.s., creti, p., ramunni, e., antonucci, v. 1999, Solid State Ionics, 125, 431. 199. adjemian, k.t., krishnan, l.d., ota, h., majsztrik, p.t., zhang, p., mann, j., kirby, b., gatto, l., velo-simpson, m., leahy, j., srinivasan, s., benziger, j.b., bocarsly, a.b. 2006, Chem. Mater., 18, 2238. 200. adjemian, k.t., srinivasan, s., benziger, j., bocarsly, a.b. 2002, J. Power Sources, 109, 356.


174


201. park, k.t., jung, u.h., choi, d.w., chun, k., lee, h.m., kim, s.h. 2008, J. Power Sources, 177, 247. 202. kim, d.s., michael, b.l., guiver, d. 2006, Polymer, 47, 7871. 203. nogami, m., nagao, r., wong, c. 1998, J. Phys. Chem. B, 102, 5772. 204. krishnan, p., park, j.s., kim, c.s. 2006, J. Membr. Sci., 279, 220. 205. mustarelli, p., carollo, a., grandi, s., quartarone, e., tomasi, c., leonardo, s., magistris, a. 2007, Fuel Cells, 6, 441. 206. kim, y.m., choi, s.h., lee, h.c., hong, m.z., kim, k., lee, h-i. 2004, Electrochim. Acta, 49, 4787. 207. reyna-valencia, a., kaliaguine, s., bousmina, m. 2005, J. Appl. Polym. Sci., 98, 2380. 208. alberti, g., casciola, m. 2001, Solid State Ionics, 145, 3. 209. alberti, g., casciola, m., costantino, u., vivani, r. 1996, Adv. Mater., 8, 291. 210. clearfield, a., stynes, j.a. 1964, J. Inorg. Nucl. Chem., 26, 117. 211. alberti, g., casciola, m. 1997, Solid State Ionics, 97, 177. 212. abe, y., li, g., nogami, m., kasuga, t., hench, l.l. 1996, J. Electrochem. Soc., 143, 144. 213. arico, a.s., baglio, v., di blasi, a., antonucci, v. 2003, Electrochem. Commun., 5, 862. 214. hogarth, w.h.j., diniz da costa, j.c., lu, g.q. 2005, J. Power Sources, 142, 223. 215. subramanian, m.a., roberts, b.d., clearfield, a. 1984, Mater. Res. Bull., 19, 1471. 216. ramani, v., kunz, h.r., fenton, j.m. 2004, J. Membr. Sci., 232, 31. 217. colomban p. (Ed.) 1992, Chemistry of Solid State Materials. 2. Proton Conductors – Solids, Membranes and Gels – Materials and Devices, Cambridge University Press, Cambridge. 218. nakamura, o., ogino, i. , kodama, t. 1981, Solid State Ionics, 3–4, 347. 219. pope, m.t. 1983, Heteropoly and Isopoly Oxometalates, Springer-Verlag, Berlin. 220. moffat, j.b. 1986, Polyhedron, 5, 261. 221. brown, g.m., noe-spirlet, m.r., busing, w.r., levy, h.a. 1977, Acta Cryst., B33, 1038. 222. chalkova, e., fedkin, m.v., komarneni, s., lvova, s.n. 2007, J. Electrochem. Soc., 154, B288. 223. yang, c., srinivasan, s., arico, a.s., creti, p., baglio, v., antonucci, v. 2001, Electrochem. Solid-State Lett., 4, A31. 224. yang, c., srinivasan, s., bocarsly, a.b., tulyani, s., benziger, j.b. 2004, J. Membr. Sci., 237, 145. 225. clearfield, a., roberts, b.d., subramanian, m.a. 1984, Mater. Res. Bull., 19, 219. 226. ticianelli, a., derouin, c.r., redondo, a., srinivasan, s. 1998, J. Electrochem. Soc., 135, 2209. 227. alberti, g., casciola, m., costantino, u., levi, g., ricciardi, g. 1978, J. Inorg. Nucl. Chem., 40, 533. 228. casciola m., bianchi, d. 1985, Solid State Ionics, 17, 287. 229. trakanprapai, c., esposito, v., licoccia, s., traversa, e., baglio, v., di blasi, a, arico, a.s., antonucci, v., antonucci, p.l. 2003, Abstract 1080, The Electrochemical Society Meeting Abstracts, Vol. 2003-2, Orlando, FL, Oct. 12. 230. si, y., kunz, h.r., fenton, j.m. 2004, J. Electrochem. Soc., 151, A623. 231. kim, y-t., song, m-k., kim, k-h., park, s-b., min, s-k., rhee, h-w. 2004, Electrochim. Acta, 50, 645.



175

232. staiti, p., aricio, a.s., baglio,v., lufrano, f., passalacqua, e., antonucci, v. 2001, Solid State Ionics, 145, 101. 233. tian, h., savadogo, o. 2005, Fuel Cell, 5, 375. 234. hongwei, z., baoku, z., youyi, x. 2006, J. Appl. Polym. Sci., 102, 3972. 235. kim, j.-d., mori, t., honma, i. 2006, J. Electrochem. Soc., 153, A508. 236. coronado, e., garcia, c.j.g. 1998, Chem. Rev., 98, 273. 237. nakamura, o., kodama, t., ogino, i., miyake, y. 1979, Chem. Lett., 17. 238. honma, i., nakajima, h., nomura, s. 2002, Solid State Ionics, 154/155, 707. 239. honma, i., nakajima, h., nishikawa, o., sugimoto, t., nomura, s. 2002, J. Electrochem.Soc., 150, 616. 240. smitha, b., sridhar, s., khan, a.a. 2005, J. Polym. Sci.: Part B: Poly. Phys., 43, 1538. 241. watanabe, m., uchida, h., seki, y., emori, m. 1996, J. Electrochem. Soc., 143, 3847. 242. yang, t.h., yoon, y.g., kim, c.s., kwak, s.h., yoon, k.h. 2002, J. Power Sources, 106, 328. 243. kwak, s.h., yang, t.h., kim, c.s., yoon, k.h. 2003, J. Power Sources, 118, 200. 244. uchida, h., ueno, y., hagihara, h., watanabe, m. 2003, J. Electrochem. Soc., 150, A57. 245. wang, l., xing, d.m., liu, y.h., cai, y.h., shao, z.g., zhai, y.f., zhong, h.x., yi, b.l., zhang, h.m. 2006, J. Power Sources, 161, 61. 246. son, d.h., sharma, r.k., shul, y.g., kim, h.s. 2007, J. Power Sources, 165, 733. 247. zhu, x., zhang, h., liang, y., zhang, y. 2006, Electrochem. Solid-State Lett., 9, A49. 248. zhang, y., zhang, h.m., zhu, x.b., liu, g., bi, c., liang, y.m. 2007, J. Power Sources, 165, 786. 249. yang, y.z., fu, a., manthiram, r. 2005, J. Power Sources, 139, 170. 250. zhang, y., zhang, h., bi, c., zhu, x. 2008, Electrochim. Acta, 53, 4096.


5 Lithium-doped hybrid polymer electrolytes V. de ZEA B E R M U D E Z, University of Trás-os-Montes e Alto Douro, Portugal, and M. M. S I LVA, University of Minho, Portugal

Abstract: The interest in sol–gel derived polymer/siloxane hybrid electrolytes has grown considerably during the last decade because of their potential application in solid state electrochemical devices, in particular batteries and electrochromic devices (ECDs). This review intends to provide an overview of the latest advances in the investigation of the structure, morphology, thermal properties, electrochemical behavior and spectroscopic features of Li+-containing di-urea crosslinked polyoxyethylene (POE)/siloxane (di-ureasil) electrolytes. Applications of these materials in ECDs will be addressed. Key words: hybrid electrolytes, sol–gel, lithium salts, characterization, electrochromic devices.

5.1

Introduction

The interest in polymer electrolytes (PEs) has not ceased to increase since the first report of salt/polymer complexes by Wright1 and the subsequent recognition by Armand et al.2 of the potential use of these materials as electrolytes in solid state electrochemical devices.3 Highly conducting, thin, solvent-free PE films may be produced simply by dissolving ionic salts in high molecular weight polymers carrying appropriate coordinating species, such as ether oxygen atoms. Owing to its extraordinary solvating ability towards salts, polyoxyethylene (POE) has been the archetypal host polyether employed to synthesize PEs. Within this category of electrolytes, those incorporating Li+ ions have been the most extensively investigated to date.4 These materials fulfil the prerequisites of environmental compatibility, offer no hazard to the consumer, are of low density and can be produced as thin films. They are particularly suitable for application in advanced batteries (primary and secondary) and ECDs. The practical application of POE-based electrolytes in solid state electrochemical devices has been delayed by two main drawbacks: their poor processability and marked tendency to crystallize. The latter aspect is particularly limiting. POE/salt complexes are usually crystalline, but pristine POE is itself already semicrystalline. For nearly two decades the view that 176 © Woodhead Publishing Limited, 2010

Lithium-doped hybrid polymer electrolytes

177

prevailed in the PE community, based on the pioneer work of Berthier et al.,5 was that in semicrystalline PEs ionic transport was confined to amorphous phases. It was accepted that high ionic conductivity could be expected in systems composed of host polymers with low glass transition temperatures (Tg). Understandably, until very recently the elucidation of the phase diagram of semicrystalline electrolytes was considered to be indispensable to determine the ranges of salt concentration and temperature for which the conductivity attained a maximum value. In 2001 Gadjourova et al.6 showed for the first time that ionic conductivity in the crystalline domains of the polymer matrix can be significantly higher than in the corresponding amorphous regions. However, because of the low conductivity exhibited by the complexes proposed by these authors6 current research remains focused on the development of entirely amorphous systems. To overcome the disadvantages associated with conventional POE-based PEs, three strategies have been basically attempted in the last two decades: the modification of the polymer architecture, the modification of the anion design and the addition of plasticizers.3 The class of organically modified silicate electrolytes (ormolytes) that emerged from these efforts has attracted considerable attention. The ormolyte strategy relies on the production in mild synthetic conditions of organic/ inorganic hybrid host frameworks7 by means of the sol–gel method.8 The incorporation of POE segments in the organic component endows these siloxane-based hybrid structures with the ability to accommodate ionic species. Interestingly, in these hybrid electrolytes the proportion of crystalline regions is in general very low or even null. Moreover, extremely high contents of salt may be added to the hybrid matrices, while avoiding the inconvenient ‘salting-out’ often found in classical PEs. Xerogel samples of these easy-to-process materials may be readily produced as thin films and a remarkable improvement in the mechanical properties and chemical/ thermal stability results. The organic/inorganic hybrid concept has been successfully employed in the last decade for the development of Li+-doped ormolytes.9–24 The present work reviews the studies that have been carried out by our group for more than a decade to elucidate the structure, morphology, thermal properties and electrochemical behavior of a series of POE/ siloxane ormolytes doped with lithium salts. In an attempt to understand the ionic conductivity mechanism of the Li+-doped di-ureasils, we have used extensively vibrational spectroscopy to examine the cation/polymer, cation/anion and cation/crosslink interactions. The host matrix of the materials analysed belongs to the class of di-ureasils,25–27 a family of hybrids in which the organic and inorganic components are bonded through urea groups. The terminology adopted to represent these materials is the following:


178 •

•


The non-doped di-ureasils are designated as d-U(Y), where d indicates di, U denotes the urea group and Y = 2000, 900 and 600 corresponds to the average molecular weight (in g mol−1) of the starting diamine precursor (corresponding to approximately 40.5, 15.5 and 8.5 (OCH2CH2) repeat units, respectively). The doped di-ureasils are described by the d-U(Y)nMXm notation, where n (salt composition) represents the number of ether oxygen atoms per Mm+ cation and X is the anion. For instance, the sample d-U(2000)20LiBF4 is a long chain di-ureasil incorporating an amount of lithium tetrafluoroborate such that the OPOE/Li+ ratio is equal to 20.

5.2

Ionic conductivity

The total ionic conductivity is an important characterization parameter which has been used as the criterion for possible application in devices. A PE intended for use in diverse electrochemical applications must have adequate ionic conductivity, together with negligible electronic conductivity if self-discharge on standing is to be avoided. A PE is considered to be a promising candidate for commercial application if its ionic conductivity is as high as 10−5 S cm−1 at room temperature.28–30 In general, the ionic conductivity of the electrolytes is measured as a function of salt composition and temperature. The objective of this characterization is to identify the electrolyte with the most favourable behavior for use as a component of the practical device. In general, salts with a polarizing cation and a large anion with a welldelocalized charge, and therefore also with low lattice energy, are the most suitable for use in PEs.31,32 In spite of the dangers associated with the anion, lithium perchlorate (LiClO4) is a salt that satisfies the conditions mentioned above. Lithium trifluoromethanesulfonate (or triflate) (LiCF3SO3) and LiBF4 have also been extensively employed in this context.3 Lithium bis(trifluoromethanesulfonyl)imide (LiTFSI) is particularly interesting as a guest species in solid PEs and also one of the best choices. In common with other salts that contain large polarizable anions, LiTFSI has low lattice energy and a low tendency to form ion-pairs, leading to enhanced ionic mobility. This salt also performs as a plasticizer in polyether electrolytes by creating free-volume. This is a significant advantage in polymer hosts that have an inherent tendency to crystallize. Figures 5.1–5.3 are included to demonstrate the effect of the choice of the di-ureasil network on the ionic conductivity. The Arrhenius plots show the variation of ionic conductivity with temperature of selected compositions of the d-U(2000)-, d-U(900)- and d-U(600)-based di-ureasil systems doped with the four lithium salts mentioned above.20,22,23,33



112

97

84

T (°C) 60 49

72

39

30

179

21

–2.5 log conductivity (S cm–1)

–3.0 –3.5 –4.0 –4.5 –5.0 –5.5 –6.0

d-U (2000)10LiTFSI d-U (2000)20LiClO4

–6.5

d-U (2000)20LiCF3SO3 d-U (2000)10LiBF4

–7.0 2.6

2.7

2.8

2.9 3.0 3.1 1000/T (K)

3.2

3.3

3.4

5.1 Arrhenius conductivity plots of the d-U(2000)nLiX di-ureasil systems. Adapted from Silva et al.,20,21 Gomes Correia et al.23 and Barbosa et al.33

–2.5

112

97

84

72

T (°C) 60 49

39

30

21

3.2

3.3

3.4

log conductivity (S cm–1)

–3.0 –3.5 –4.0 –4.5 –5.0 –5.5

d-U (900)35LiBF4

–6.0

d-U (900)8LiClO4

–6.5

d-U (900)15LiTFSI

–7.0 2.6

2.7

2.8

2.9 3.0 3.1 1000/T (K)

5.2 Arrhenius conductivity plots of d-U(900)nLiX di-ureasil systems. Adapted from Barbosa et al.33–35

Typically the total ionic conductivity of the di-ureasils discussed here was measured by placing the sample between gold blocking electrodes, along the so-called electrode/di-ureasil/electrode assembly, which was secured in a suitable constant-volume support, to form a symmetrical cell.


180


112

97

84

72

T (°C) 60 49

39

30

21

3.2

3.3

3.4

–2.5 log conductivity (S cm–1)

–3.0 –3.5 –4.0 –4.5 –5.0 –5.5 d-U (600)35LiBF4

–6.0

d-U (600)15LiTFSI

–6.5

d-U (600)10LiClO4

–7.0 2.6

2.7

2.8

2.9

3.0

3.1

1000/T (K)

5.3 Arrhenius conductivity plots of d-U(600)nLiX di-ureasil systems. Adapted from Barbosa et al.33,35

Low-amplitude alternating potentials at frequencies between 65 kHz and 0.5 Hz were applied over a range of temperatures from 25 to 100 °C. It is clear from the analysis of Figs 5.1–5.3 that the materials characterized in this study show a non-linear variation of log conductivity with 1/T in the range between 25 and 100 °C. This behavior is characteristic of amorphous polymer electrolytes. The highest room temperature conductivity of the electrolyte system d-U(2000)nLiTFSI is 3.2 × 10−5 S cm−1, registered with the d-U(2000)35LiTFSI composition. This value is similar to that reported by Armand et al.36 for the POEnLiTFSI electrolyte system. The highest conductivity over the temperature range studied was found with the d-U(2000)10LiTFSI ormolyte, and was recorded as 1.2 × 10−3 S cm−1 at 95 °C (Fig. 5.1). The conductivity isotherms derived from conductivity measurements revealed the presence of a maximum conductivity for the electrolyte system located in the composition interval 8 ≤ n ≤ 10. Ionic conductivity in this hybrid electrolyte increases gradually with salt content for compositions with n ≥ 10. This parameter is generally expected to increase with the number of charge carriers, up to a limit found at about n = 8. Further increase in salt content beyond this concentration resulted in a decrease in total ionic conductivity. This observation can be explained by the formation of associated ionic species (e.g. contact ion pairs, triplets or higher multiplets or ultimately ion aggregates) and an increased tendency for ions to form bridging interactions between adjacent polymer chains a process



181

designated as ‘ionic crosslinking’. Associated ionic species may be expected to show lower mobility than dissociated ions and the restriction of host matrix segment mobility also contributes to a reduction in the rate of ion transport and electrolyte conductivity. We will return to the aspect of ionic association in Section 5.5. For the hybrid system d-U(900)nLiTFSI, the most conducting composition is n = 15 which reaches 7.6 × 10−5 S cm−1 at 95 °C.34 At room temperature the sample with the highest conductivity is d-U(900)25LiTFSI (2.9 × 10−6 S cm−1). The conductivity isotherms of this electrolyte system confirmed that ionic transport in this hybrid electrolyte system is almost independent of the salt content over a wide range of composition. The broad conductivity maxima is located between n = 15 and 25. Further increase in salt content beyond about n = 15 led to a decrease in total ionic conductivity. This observation is consistent with a decrease in the flexibility of the host polymer chain segments, confirmed by an increase in the Tg of the electrolyte matrix. The Tg can be defined as the temperature at which the polymer chains in the amorphous phase begin segmental motion.3 According to the classical study of Berthier et al.5 it has been recognized that ionic mobility is closely correlated with the viscoelastic properties of the host polymer and so the relative magnitude of Tg in these polymer/salt complexes can be of considerable importance. As the mechanism of ion transport in amorphous PEs is dependent on the motion of polymer chain segments, the ionic conductivity is expected to decrease with an increase in Tg (Fig 5.4). The observed increase in Tg at lower values of n can be explained by the increase in number of bridging interactions which take place between adjacent polymer chains and guest salt species, generally designated as ‘ionic crosslinking’. Figure 5.2 compares the ionic conductivity of di-ureasil systems composed of the same host matrix. While at temperatures higher than 60 °C the d-U(900)nLiClO4 ormolytes are more conducting than comparable d-U(900)nLiTFSI materials, at lower temperatures the difference is negligible and LiTFSI is a better choice for safety reasons. The most conducting electrolyte of the hybrid d-U(900)nLiBF4 systems is the d-U(900)35LiBF4 composition (1.70 × 10−4 and 2.09 × 10−5 S cm−1 at about 95 °C and room temperature, respectively) (Fig. 5.2). These results are comparable with those reported by Chiodelli et al.37 and Armand et al.38 for the POEnLiBF4 system. The results show that for hybrid electrolyte networks conductivity increases with the salt content as the number of charged species available to support ionic conduction is increased and ionic transport is almost independent of the salt content over a wide range of composition of 40 ≥ n ≥ 200. In the case of the d-U(900)nLiBF4 system, isotherm curves reveal two conductivity maxima at n = 35 and at 20 ≥ n ≥ 10 at higher temperatures.


182

Polymer electrolytes Tg d-U (600)

log conductivity d-U (600)

Tg d-U (900)


Tg d-U (2000)


10

–4.0

0

–5.0

Tg (°C)

–6.0

–20 –30

–7.0

–40 –8.0 –50

log conductivity (S cm–1)

–10

–9.0

–60 0

20

40

100

150

200

n (d-U (900, 600 or 2000) : LiBF4)

5.4 Comparison of the variation of Tg and of the ionic conductivity of selected compositions of three di-ureasil systems doped with LiBF4 at 30 °C as a function of composition. Adapted from Barbosa et al.33

Figures 5.1–5.3 demonstrate that the di-ureasils doped with LiBF4 are the best conducting electrolytes. We may also infer from these data that the di-ureasils based on higher molecular weight POE chains are more conducting than ormolytes based on the di-ureasil structure d-U(600). These results were expected, since it is generally accepted that local segment motion plays an important role in the ionic migration in polymer electrolytes. The oxyethylene segments of d-U(600) are very short, causing restrictions of polymer segment mobility and a decrease in the transport of the guest ions. The higher molecular weight chains of the doped d-U(900) and d-U(2000) di-ureasils were found to support higher conductivity.20,33,35

5.3

Thermal properties

Thermal analysis techniques, in particular differential scanning calorimetry (DSC) and thermogravimetric analysis (TGA), are valuable tools to study the thermal behavior of PEs. DSC allows us to calculate the proportion of crystallinity, to detect the formation of new crystalline phases, free guest salt or uncomplexed polymer chains, to monitor the loss of solvent(s) (e.g. occluded water, alcohol), to determine the Tg value and to distinguish between endo- and exothermic events. TGA provides rich information



183

n = 200 n = 80

2 W/g

n = 35 n = 15 n=8 Exo

n=5 100

50

150

200

250

200

250

Temperature (°C) (a) n = 200 n = 80

1 W/g

n = 40 n = 25 n = 15

Exo

n=5 n = 0.5

50

100

150 Temperature (°C)

(b)

5.5 DSC curves of (a) d-U(2000)nLiTFSI (adapted from Barbosa et al.22) and (b) d-U(900)nLiTFSI (adapted from Barbosa et al.34) di-ureasil systems.

about the thermal degradation of the samples and their thermal stability domain. For the DSC measurements of di-ureasil ormolyte samples, sections have been usually removed from dry films and subjected to thermal analysis under a flowing inert atmosphere between 25 and 300 °C and at a heating rate of 5 °C min−1. Samples are then transferred to aluminium cans.


184


Figure 5.5 reproduces the DSC curves obtained for the d-U(2000)nLiTFSI and d-U(900)nLiTFSI ormolytes, respectively.22,34 The absence of crystallinity in samples of both networks was expected from conductivity measurements. No endotherms associated with either the fusion of oxyethylene chain segments or the evaporation of water or ethanol are observed, leading us to conclude that the samples with 200 ≥ n ≥ 5 are completely amorphous. At high salt concentration (n = 0.5) an endothermic peak is observed at approximately 150 °C in the d-U(900)nLiTFSI system. This endothermic event is probably associated with the fusion of a salt/polymer complex, as reported in other di-ureasil systems based on other lithium salts.34 The onset of thermal decomposition was estimated by extrapolation from TGA curves (not shown). The lowest decomposition temperature observed with the d-U(900)nLiTFSI series was registered at n = 40 (358 °C), a finding that indicates that this di-ureasil system is more stable than the analogous d-U(900)nLiClO4 ormolyte. It is evident that the presence of the guest salt destabilizes the polymer matrix. However there is apparently no direct relationship between the onset of weight loss and the amount of salt present in the electrolyte. This situation is similar to that observed with electrolytes of other di-ureasil systems.20,35 Figure 5.4 illustrates the variation of Tg for selected compositions of the LiBF4-doped d-U(600)-, d-U(900)- and d-U(2000)-based electrolytes.33 The incorporation of guest salt leads to an increase in Tg, which means that the electrolyte sample becomes less flexible. Compositions with the highest salt content present poor mechanical properties, which in practical devices limits seriously the usefulness of the electrolyte in electrochemical applications. As expected from previous studies of similar POE/siloxane ormolytes,33 an increase in salt concentration causes a corresponding increase in the Tg of the di-ureasils (Fig. 5.4). Another aspect worth noting is that the increase in Tg with salt concentration is much more pronounced in the d-U(600)-based electrolytes than in the d-U(900)- or d-(2000)-based xerogels, which may be explained by the restrictions of segmental motion of the low molecular weight polyether chains present in the d-U(600) network. Figure 5.4 also reveals that the decrease in conductivity is consistent with an increasing Tg and significant levels of ionic conductivity exist only above this temperature.

5.4

Electrochemical stability

The application of the ormolytes in electrochemical applications depends on their stability window. To evaluate the electrochemical stability window of di-ureasil ormolyte compositions we have typically used a two-electrode cell configuration involving the use of a 25 µm-diameter gold microelectrode surface. Cell assembly was initiated by locating a freshly cleaned



185

0.70 × 10–6 0.05 × 10–7

0.60 × 10–6

0 –0.05 × 10–7 I/A

0.50 × 10–6

I/A

0.40 × 10–6 –6

0.30 × 10

–6

0.20 × 10

–0.10 × 10–7 –0.15 × 10–7 –0.20 × 10–7 –0.25 × 10–7 –0.30 × 10–7 –5.00 –4.00 –3.00 –2.00 –1.00 0

1.00 2.00 3.00

E/V versus Li/Li+

0.10 × 10–6 0 –0.10 × 10–6 –2.00

–1.00

0

1.00

2.00

3.00

4.00

5.00

6.00

7.00

E/V versus Li/Li+

5.6 Voltammogram of d-U(2000)15LiTFSI di-ureasil electrolyte at a 25 µm diameter gold microelectrode vs. Li/Li+. Initial sweep direction is anodic and sweep rate is 100 mV s−1. Reproduced from Barbosa et al.22

lithium disk counter electrode on a stainless steel current collector. A thinfilm sample of ormolyte was centered over the counter electrode and the cell assembly completed by locating and supporting the microelectrode in the center of the electrolyte disk. Measurements were conducted at room temperature within a Faraday cage. The electrochemical stability range of the lithium-doped di-ureasils was determined by microelectrode cyclic voltammetry over the potential range between −1.5 and 6.5 V.20,22,33 In the anodic region, all ormolytes are stable up to 4.5 V vs. Li/Li+. Lithium deposition begins in the cathodic region at about −0.5 V vs. Li/Li+ (Fig. 5.6). These results show that the electrochemical stability of the di-ureasils is acceptable for application in practical devices.

5.5

Spectroscopic studies

Infrared and Raman spectroscopy are powerful tools in the elucidation of the ionic conductivity/ionic association relationship in PE systems. This sort of spectroscopic analysis usually involves the examination of diagnostic bands of the host polymer and of the anion. In the case of the organic/ inorganic hybrid matrices incorporating organic cross-links (e.g. diureasils), this study must also include the analysis of the bands characteristic of these functional groups (i.e. urea groups).


186


The mid-infrared Fourier transform infrared (FT-IR) spectra (4000– 400 cm−1 range) should be acquired using a spectrometer placed inside a glove box with a dry argon atmosphere. If this is not possible, it is advisable to perform the measurement in vacuum, by averaging at least 150 scans at a high resolution (1–2 cm−1). The solid samples should be finely ground and mixed with dried spectroscopic grade potassium bromide. Prior to recording the spectra, it is highly recommended to keep the resulting pellets in an oven under vacuum at 90 °C for several days in order to reduce the levels of adsorbed water. The Fourier transform Raman (FT-Raman) spectra should be collected over the same wavenumber range, also at a high resolution (2–1 cm−1). An accumulation time for each spectrum of 4 hours is adequate. To evaluate complex band envelopes and to identify underlying component bands of the spectra, it is often necessary to perform classical curvefitting procedures, using for instance an iterative least-squares curve-fitting procedure such as that proposed in the Peakfit (San Rafael, CA) software. The best fit of the experimental data may be obtained by varying the frequency, bandwidth and intensity of the bands. In most cases it is appropriate to employ Gaussian band shapes.

5.5.1 Cation/polymer interactions Several vibrational modes of POE are sensitive to the interaction of the polymer backbone with cations and can be thus employed as diagnostic tools to monitor the changes undergone by the polyether chains upon addition of the guest salt. Two spectral regions have been widely used to probe the coordination of the cations to host polyethers: Skeleton COC stretching, νCOC, region. In this frequency interval (1200–1000 cm−1) the complexation of the cations by the heteroatoms of the polyether chains produces a shift of the strong νCOC band to lower wavenumbers. The magnitude of this shift depends on the strength of the interaction. • CH2 rocking, rCH2, region. Between 1000 and 800 cm−1 the bands ascribed to a mixture of CH2 rocking and CC stretching modes (1000– 900 cm−1) and the bands originated from CH2 rocking modes coupled with CO stretching modes (900–800 cm−1) appear.39–41 Their frequency strongly depends on the –O–C–C–O– angle and consequently on polymer conformation. •

A typical procedure adopted to monitor the coordination of cations to POE chains, based on the inspection of the skeleton νCOC modes in the FT-IR spectra, follows. The example given deals with a study previously carried out by our group on d-U(2000)nLiCF3SO3 di-ureasil ormolytes.17 In the


1090

187

n

1137

1163

1111


5

Absorbance (a.u.)

10

40 60 80 100 200

1080

1137

1148

Absorbance (a.u.)

1175

1090

1111

1200 1180 1160 1140 1120 1100 1080 1060 Wavenumbers (cm–1) 1162

(a)

n

5

10 1220 1200 1180 1160 1140 1120 1100 1080 1060 1040 Wavenumbers (cm–1) (b)

5.7 FT-IR νCOC region of selected d-U(2000)nLiCF3SO3 di-ureasils (a) and results of the curve-fitting results for the most concentrated samples (b). Adapted from Nunes et al.17


188


νCOC region the complexation of the Li+ ions by the oxygen atoms of the POE chains is reported to induce a shift of the prominent νCOC band to 1095–1089 cm−1.42,43 The room temperature FT-IR spectra of selected d-U(2000)nLiCF3SO3 samples in the νCOC region and the results of the curve-fitting performed with two representative samples are reproduced in Fig. 5.7. The FT-IR spectrum of d-U(2000) displays in this spectral region a prominent broad band around 1111 cm−1 and a shoulder at about 1148 cm−1 (Fig. 5.7), due to the νCOC vibration mode and to the coupled vibration of the νCO and rCH2 modes, respectively.39,40 The analysis of the spectra of the samples with 200 ≥ n ≥ 20 allowed us to conclude that the intensity and frequency of both features, characteristic of non-coordinated, disordered POE chains,39,40 remains essentially unchanged (Fig. 5.7). These results led us to conclude that in di-ureasil samples with n ≥ 20 the ether oxygen atoms of the polymer segments do not in principle bond to the Li+ ions. Their presence also made us suggest that the POE chains remain amorphous over the same composition range. Upon further addition of LiCF3SO3 to d-U(2000) (n = 10) the polymer features at 1148 and 1111 cm−1 persist, three new components emerge at 1162, 1137 and 1090 cm−1 and two new shoulders are seen at 1175 and 1080 cm−1 (Fig. 5.7). In the case of the d-U(2000)5LiCF3SO3 sample the νCOC band profile becomes considerably better resolved (Fig. 5.7). The 1175 and 1162 cm−1 bands – attributed to the asymmetric stretching vibration of the CF3 group – and the 1137, 1090 and 1080 cm−1 features coincide with those produced by the POE3LiCF3SO3 crystalline complex,44 indicating the formation of this compound in the diureasils with n = 10 and 5. The presence of the 1111 cm−1 band in the spectra of both samples supports that this crystalline compound coexists with noncoordinated and amorphous POE chains at n = 10 and 5. It must be emphasized that the determination of the composition that corresponds to the beginning of cation complexation by amorphous POE chains is often difficult. For instance, in the present case, although the first spectroscopic signs of Li+/POE bonding were evident in the νCOC region at n = 10, the salt concentration at which the Li+ ions start to effectively interact with the ether oxygen atoms of the POE chains may be lower. At n > 10 it is probable that the broad and strong νCOC envelope centered at 1111 cm−1 masks the characteristic band of Li+-coordinated amorphous POE chains (Fig. 5.7). To get more insight into the latter aspect it was considered of interest to examine the rCH2 region of the d-U(2000)nLiCF3SO3 ormolytes.45 The examination of the FT-IR spectra reproduced in Fig. 5.8 allowed us to immediately conclude that the rCH2 region of the Li+-based hybrids with 200 ≥ n ≥ 10 is very similar to that of d-U(2000). The main features of these spectra over the 980–900 cm−1 and 900–800 cm−1 ranges were a medium


189

895

n

833 859

1

848

Absorbance (a.u.)

969

953

843

937


5 10 20 40 60 80 200 ∞ 980 960 940 920 900 880 860 840 820 800 780 Wavenumbers (cm–1)

5.8 Room temperature FT-IR spectra of d-U(2000)nLiCF3SO3 di-ureasils in the rCH2 region. Reproduced from Nunes et al.45

intensity band at 950 cm−1 and a weak, broad and ill-defined band at 848 cm−1 (Fig. 5.8). Both features support that the conformation of the POE chains of d-U(2000) remains unaffected upon addition of increasing amounts of salt and therefore that the POE chains do not complex the Li+ ions at 200 ≥ n > 10. The spectra of the samples with n = 10, 5 and 1 differ markedly from those of the less concentrated materials. At n = 10 a feature at 969 cm−1, two medium intensity bands at 843 and 833 cm−1 and a weaker, broad event at 859 cm−1 emerge (Fig. 5.8). At higher salt concentration, the intensity of the bands increases significantly. At n = 1 the 937 and 843 cm−1 bands became the strongest events of the rCH2 region and a new band appears at 895 cm−1 (Fig. 5.8).


190


Our interpretation of the spectral data obtained in the rCH2 region relied on early works of the infrared spectrum of high molecular weight poly(ethylene glycol)dimethyl ether (PEGDME).39–41 According to these studies, crystalline PEGDME exhibits a characteristic band at 844 cm−1,41 which undergoes a shift to 855 cm−1 upon melting,39,40 and a characteristic band at 937 cm−1.41 Consequently, we considered that the changes detected in the 1000–800 cm−1 interval of the FT-IR spectra of the d-U(2000)n LiCF3SO3 di-ureasils with 10 ≥ n ≥ 1 are a direct consequence of the alterations undergone by the POE chains to coordinate to the Li+ ions, since the POE segments of these three ormolytes adopt gauche conformations with a –O–C–C–O– torsional angle similar to that of the non-coordinated crystalline POE. In other words, in compounds with n > 10 the POE chains are amorphous, whereas in the samples with n = 10, 5 and 1 the chains adopt a helical-like structure. Comparison of the FT-IR spectrum of the POE3 LiCF3SO3 complex reported by Dissanayake and Frech44 with those of d-U(2000)5LiCF3SO3 and d-U(2000)1LiCF3SO3 revealed that the bands exhibited by the two hybrids at 969, 953, 937 and 833 cm−1 coincide with those produced by the crystalline POE3LiCF3SO3 complex, an indication of its formation in both salt-rich hybrid xerogels. A straightforward way of detecting the coordination of guest cations to the ether oxygen atoms of POE implies the analysis of the FT-Raman spectra in the region where the ‘oxygen breathing’ mode (νM–O) is expected (925–775 cm−1).46 In this spectral range the FT-Raman spectra of the d-U(2000)nLiCF3SO3 hybrids with n > 20 resemble that of d-U(2000) (Fig. 5.9). At n = 20 a shoulder emerges at 865 cm−1 and becomes a strong event in samples with n ≤ 7. Simultaneously the intensity of the 809 cm−1 band is reduced with the increase of salt concentration (Fig. 5.9). The 865 cm−1 band is attributed to a stretching vibration mode related with wrapping of the POE chains around the Li+ ion during complexation42,47,48 and therefore with the presence of gauche conformations of the O–C–C–O bonds,42 whereas the 809 cm−1 band indicates the presence of trans –O–C–C–O– conformations in the POE chains.49,50 The occurrence of the 865 cm−1 band in the FT-Raman spectra of the d-U(2000)nLiCF3SO3 with n ≤ 20 corroborates the claim that POE/Li+ interactions take place at high salt concentration. The conclusions derived from the νCOC17 and rCH245 regions allowed us to place the beginning of Li+/ether oxygen bonding near n = 10.

5.5.2 Cation/anion interactions In PEs several types of charge carriers may participate in the conduction mechanism: (a) ‘free’ or weakly coordinated ions with relatively high mobility; (b) cations interacting strongly with the host polymer and thus with low


191

809

843

865


885

n

Intensity (a.u.)

1

5

7 20 40 60 80 200 ∞ 920

900

880

860

840

820

800

780

Raman shift (cm–1)

5.9 Room temperature FT-Raman spectra of d-U(2000)nLiCF3SO3 di-ureasils in the rCH2 region. Reproduced from Nunes et al.45.

mobility; (c) charged aggregates with moderate-low mobility. Papke et al.51 stated that in POE-based electrolytes containing monovalent ions an increase in the number of uncharged ion pairs is accompanied by a decrease in the ionic conductivity. To determine the extent of ionic association in PEs on the basis of vibrational spectroscopic studies, it is essential to introduce the cations as salts including appropriate anions. To be able to play the role of ion probes, the latter species must have vibration modes easily detected both in the infrared and Raman spectra, the attribution of which should be perfectly established and well documented in the literature.


192


Three examples of these analyses will be given as follows. The first case mentioned here will be the spectroscopic study of the d-U(2000)-based diureasil system doped with LiCF3SO3.17,45 We will then refer a spectroscopic study of the d-U(2000)-based di-ureasil system doped with LiTFSI.52 Our third example is very recent and deals with d-U(2000)-based di-ureasil networks doped with LiBF4.53 Di-ureasils doped with LiCF3SO3 It is accepted that a ‘free’ CF3SO3− ion in a staggered configuration with a C3v point group symmetry has 18 normal modes with the symmetry representations 5A1+A2+6E. The A1 and E modes are infrared and Raman active. The A2 mode, associated with the internal torsion of the anion, is inactive both in infrared and Raman. Because all these modes are more or less sensitive to coordination effects they can be used to look into the anion local chemical surrounding.42,43,47,54–102 We will focus the following discussion on the non-degenerate symmetric stretching vibration mode of the SO3 group (νsSO3) of the CF3SO3− ion. Typically the νsSO3 band of a ‘free’ triflate ion appears at 1032 cm−1.103 Upon coordination to the Li+ ion, this feature shifts to higher wavenumbers.43,55,57,65,69,73,95–98,101,104,105 We will discuss first the νsSO3 region of the FT-IR spectra of the d-U(2000)nLiCF3SO3 compounds45 and then we will examine the information retrieved from the corresponding FT-Raman spectra.17 The FT-IR νsSO3 envelope and the results of the curve-fitting carried out are reproduced in Fig. 5.10(a) and (b), respectively. The presence of the 1031 cm−1 band in the FT-IR spectra of the doped samples indicates the presence of ‘free’ anions over the whole range of compositions studied. ‘Crosslink separated ion pairs’106 may also contribute to this feature. The shoulders at 1038 and 1025 cm−1 are attributed to weakly coordinated triflate ions located in two different anionic environments:106 (1) CF3SO3− species weakly bonded to Li+ ions, which simultaneously interact with the carbonyl oxygen atoms of the urea groups; (2) CF3SO3− ions hydrogenbonded to the urea N–H groups.106 The feature at 1043 cm−1 in samples with n ≤ 40 is associated with monodentate Li+CF3SO3− ions pairs or negatively charged triplets [Li(CF3SO3)2]− 57,104,105 The 1053 cm−1 component found at n = 5 and 1 is associated to a bidentate bridging aggregate (aggregate I)57 (the positively charged triplet [Li2(CF3SO3)]+).57,104,105 The event at 1063 cm−1 observed only in the case of the most concentrated di-ureasil is ascribed to a tridentate bridging aggregate (aggregate II)57 (the divalent positively charged multiplet [Li3(CF3SO3)]2+). The FT-Raman spectra of selected di-ureasils in the νsSO3 region and the results of the curve-fitting performed in the νsSO3 envelope are represented


193

1031

1038

1043

1054


n

Absorbance (a.u.)

1

5

10

20 40 80 200 1080 (a)

1070

1060

1050

1040

1030

1020

1010

Wavenumbers (cm–1)

5.10 Room temperature FT-IR spectra (a) (adapted from Nunes et al.45), FT-Raman spectra (c) (adapted from Nunes et al.17) and corresponding curve-fitting results ((b) and (d), respectively) of selected d-U(2000)nLiCF3SO3 di-ureasils in the νsSO3 region. In order to examine exclusively the contribution of the νsSO3 mode, the FT-Raman spectrum of the pure polymer had to be first subtracted from those of the hybrids with n ≥ 10.

in Fig. 5.10(c) and (d), respectively. The detection of the 1032 cm−1 band in the FT-Raman spectra of all the doped di-ureasils represents more evidence of the existence of ‘free’ anions and very likely of the ‘crosslink separated ion pairs’106 over the whole range of salt concentration. This would explain the regular increase of the 1032 cm−1 feature as LiCF3SO3 concentration is increased and specially its marked intensity at n = 1. As noted above, the


n

1031

1043 1043

1038

1053 1053


1062

194

Absorbance (a.u.)

1

1025

5

10

20

40 1080

1070

(b)

1060 1050 1040 1030 Wavenumbers (cm–1)

1020

1010

5.10 (cont.)

shoulders at 1039 and 1025 cm−1 are associated with weakly coordinated triflate ions located in two different sites. The component at 1045 cm−1 in samples with n ≤ 40 is associated with the formation of Li+CF3SO3− ion pairs or [Li(CF3SO3)2]− triplets.57,104,105 The component at 1052 cm−1 in the spectra of the di-ureasils with n = 5 and 1 is ascribed to the occurrence of the [Li2(CF3SO3)]+ triplet57,104,105 (aggregate I).52,57 This result is in perfect agreement with the identification of the POE3LiCF3SO3 crystalline complex65,69,73 at n = 5 and 1. In this compound the triflate ion vibrates essentially as the


195

1032

1042

1049


n

Intensity (a.u.)

1

5

7 20 40 200 1080 1070 1060 1050 1040 1030 1020 1010 1000 (c)


5.10 (cont.)

[Li2(CF3SO3)]+ entity because two oxygen atoms of the CF3SO3− ion bridge, in a monodentate fashion, two Li+ ions that lie adjacent to each other within the polymer helix.65,69,73 The event at 1063 cm−1 in the νsSO3 region of the most concentrated di-ureasil is associated with the formation of the [Li3(CF3SO3)]2+ multiplet (aggregate II).57 The anionic configurations detected in the FT-Raman νsSO3 region are in perfect agreement with those found in the FT-IR spectra. Comparison of the FT-IR45 and FT-Raman data17 of the d-U(2000)n LiCF3SO3 system allowed us to conclude that, although the number of components and their frequency coincide in samples with the same composition, their relative intensity differs considerably. For instance, while the 1031 cm−1 feature is the strongest event of the FT-Raman νsSO3 band of d-U(2000)5LiCF3SO3 (Fig. 5.10c), in the FT-IR spectrum of the same di-ureasil the most intense band is the 1043 cm−1 peak (Fig. 5.10a). We


1032

1039

1063

1045


196

n

Intensity (a.u.)

1

1025

5

7

20

40 1070 (d)

1060

1050

1040

1030

1020

–1

Raman shift (cm )

5.10 (cont.)

speculated that this mode is IR active, but weak in Raman scattering, meaning that the vibration associated with it may imply changes in the dipole moment but not in the polarizability of the species in question. Di-ureasils doped with LiTFSI The fundamental vibrations of the TFSI− ion are situated below 1400 cm−1.107 If the anion adopts C2 symmetry, its 39 internal vibrations can be classified into 20 A and 19 B modes, the former being polarized in Raman.107 In addition, for C2 symmetry, each of these modes will split into in-phase (A) and



197

out-of-phase (B) components because of the coupling between the two SO2 groups.107 In our investigation of the d-U(2000)nLiTFSI materials we centered our attention on the region of the FT-Raman spectra characteristic of the symmetric deformation mode of the CF3 group (δsCF3) (also considered by some authors as a symmetric stretching mode of the SNS group (νsSNS)). We also examined the FT-IR spectra in the range of wavenumbers where the asymmetric stretching mode of the SNS group (or νaSNS), the asymmetric deformation mode of the CF3 group (δaCF3) and the δsCF3/νsSNS mode are expected. The Raman δsCF3 mode of the ‘free’ imide ion is found at 739 cm−1.107 This ion-pairing sensitive band shifts to higher wavenumbers upon coordination of the anion to a cation. We note that the assignment of this band has been a matter of debate. Bakker et al.108 reported that it has a marked νS–N character and therefore many authors prefer to associate it predominantly to the νsSNS.109,110 Henceforth this mode will be indicated as δsCF3/ νsSNS. Figure 5.11 reproduces the results of the curve-fitting carried out in the FT-Raman δsCF3/νsSNS band of LiTFSI-doped di-ureasils and the composition dependence of the area of the resolved components, respectively. At n ≥ 40 the band was decomposed into a single component at 739 cm−1 due to ‘free’ TFSI− ions,107 whereas at n = 20 it was resolved into two peaks at 743 and 739 cm−1. The former component is ascribed to the occurrence of contact ion pairs.107,108 Figure 5.11(b) shows that, while the concentrations of ‘free’ anions and coordinated anions remain essentially constant at 40 < n ≤ 8, at n = 5 the fraction of ‘free’ anions decreases significantly and the proportion of associated species increases concomitantly. The FT-IR spectra of the LiTFSI-doped di-ureasils with n ≥ 20 in the νaSNS, δaCF3 and δsCF3/νsSNS regions are shown in Fig. 5.12(a). The results of the curve-fitting performed in the νaSNS and δsCF3/νsSNS envelopes of selected samples are shown in Figs. 5.12(b) and (c), respectively. The FT-IR νaSNS region was decomposed into two components at 791 and 784 cm−1 for samples with 40 ≥ n ≥ 8 (Fig. 5.12b). The band at 784 cm−1 supports the presence of ‘free’ TFSI− ions,109 whereas that at 791 cm−1 is indicative of the occurrence of contact ion pairs. The FT-IR δsCF3/νsSNS band of the LiTFSIdoped di-ureasils with 40 ≥ n ≥ 8 was resolved into two components at 742 and 738 cm−1 (Fig. 5.12c). The 738 cm−1 band indicates the existence of ‘free’ anions,107–110 whereas the 742 cm−1 band is associated with the formation of contact ion pairs. The main conclusion derived from the study of the TFSI− characteristic modes of the d-U(2000)-based di-ureasils was that, although the Li+ ions interact extensively with the anions for compositions n ≤ 40, this process leads exclusively to the occurrence of contact ion pairs.


739

743

n

Intensity (a.u.)

5

8

20 760

750

(a)

740 Raman shift (cm–1)

730

720

100 Integral area fraction (%)

741 cm–1 739 cm–1

80 60 40 20 0 40

(b)

20

15 n

10

5

5.11 Curve-fitting results of the FT-Raman δsCF3 region of selected d-U(2000)nLiTFSI di-ureasils (a) and salt concentration dependence of the integral intensity of the different spectral resolved components (b). The lines drawn are just guides for the eyes. Adapted from Nunes et al.17


199

741 739

789 787


Absorbance (a.u.)

762

n

8

20 40

800 (a)

775

750

725


5.12 FT-IR spectra of selected d-U(2000)nLiTFSI of di-ureasils in the 800–700 cm−1 interval (a) and results of the curve-fitting of the νaSNS band (b) and δsCF3/νsSNS band (c) of representative samples. Adapted from Nunes et al.52

Di-ureasils doped with LiBF4 The non-bonded (‘free’) tetrahedral BF4− ion (point group Td) has nine vibration normal modes: the non-degenerate mode ν1(A1); the doubly degenerate mode ν2(E); the triply degenerate modes ν3(T2) and ν4(T2). All the modes are Raman active, but only ν3(T2) and ν4(T2) are infrared active. The lowering of the local symmetry around BF4− resulting from coordination of this anion to a metal cation has the following consequences: (i) band splitting of degenerate vibrations; (ii) frequency shifts of non-degenerate vibrations; (iii) activation of the infrared-forbidden ν1 and ν2 modes. The magnitude of the band splittings and of the frequency shifts depends on the strength of ion association. The BF4− ion may adopt three possible coordinating configurations: monodentate, tridentate or bidentate. •

Monodentate (or tridentate) configuration. The association of one (or three) of the BF4− fluorine atoms with a cation lowers the anion Td


738


200

n

Absorbance (a.u.)

8

20

40 750 (b)

745

740 Wavenumbers (cm–1)

735

730

5.12 (cont.)

symmetry of the anion to C3v. Both configurations are manifested through the splitting of the threefold degenerate ν3 and ν4 vibrations into two components (A1 and E) both in the infrared and Raman spectra. • Bidentate configuration. The association of a cation with two of the BF4− fluorine atoms reduces the anion symmetry to C2v. This configuration leads to the splitting of the twofold degenerate ν2 vibration into two components (A1 and A2) in the Raman spectrum and to the splitting of the threefold degenerate ν3 and ν4 vibrations into three components (A1, B1 and B2) both in the infrared and Raman spectra. Theoretical quantum chemical calculations suggested that the bidentate structure is the most favourable for ion pairs of BF4− with Li+.111–113


201

784

791


n

Absorbance (a.u.)

8

20

40 815 810 805 800 795 790 785 780 775 770 765 (c)


5.12 (cont.)

In the Raman spectrum of molten/aqueous solution of sodium tetrafluoroborate (NaBF4) the ν1(A1), ν2(E), ν3(T2) and ν4(T2) modes appear at 777/773 cm−1, 360/357 cm−1, 1070/1080 cm−1 and 533/528 cm−1, respectively.114 The appearance of the IR inactive ν1 mode at 770 cm−1 in the FT-IR spectra of the LiBF4-doped di-ureasils shown in Fig. 5.13 revealed the formation of ion pairs over the whole range of salt composition studied. The three features seen at 574, 534 and 522 cm−1 in the same spectra were tentatively attributed to the lifting of the degeneracy of the ν4BF4 mode resulting from the formation of ion pairs in a presumably bidentate coordination. This spectroscopic analysis enabled us to conclude that in all the d-U(2000)nLiBF4 samples examined there was a considerable tendency for ionic association.


Polymer electrolytes 534 522

202

770

574

n

2.5

Absorbance (a.u.)

5 20

60 200 ∞

800

750

700

650

600

550

500


5.13 FT-IR spectra of selected d-U(2000)nLiBF4 di-ureasils in the ν1BF4− and ν4BF4− regions. Reproduced from Barbosa et al.35

5.5.3 Cation/crosslink interactions In di-ureasil-based ormolytes doped with a lithium salt three coordinating sites are available for the Li+ ions: the urea carbonyl oxygen atoms (–C=O) and the polymer ether oxygen atoms (–COC–) of the host di-ureasil matrix and the coordinating atoms of the guest anion (e.g. the triflate oxygen atoms (CF3SO3−) in the case of LiCF3SO3, the perchlorate oxygen atoms (ClO4−) in the case of LiClO4, etc.). To illustrate the spectroscopic analysis usually adopted to study Li+/urea interactions in di-ureasil systems, we will refer in this section to the analysis carried out with the d-U(2000)-based xerogels doped with LiCF3SO3.45



203

Data retrieved from other spectral regions had demonstrated that in d-U(2000)nLiCF3SO3 samples with n > 20: (1) the POE chains do not complex the Li+ ions and (2) the concentration of coordinated anionic species is quite low. Therefore it seemed obvious to assume that at n > 20 the coordination of the Li+ ions in the d-U(2000)-based framework is entirely ensured by the C=O groups. At n > 20 two sorts of interactions could be then envisaged for the Li+ ions after being incorporated into the di-ureasil framework: they could either bond to ‘free’ C=O groups or coordinate to hydrogen-bonded C=O groups. The latter situation is more complex since the Li+ ions need to disrupt the hydrogen-bonded array formed throughout the d-U(2000) framework via urea groups. The ability of urea groups to participate in highly specific hydrogen bonding interactions is well known. In bis-ureido compounds, these interactions are responsible for unique gelling properties.115,116 This behavior relies essentially on geometrical considerations, since the C=O group of each urea group may form two hydrogen bonds with the N–H moieties of a neighbor urea group, leading to the formation planar bifurcated hydrogen bonds.117 In non-doped di-ureasils disordered POE/urea hydrogen-bonded associations and highly ordered self-associated urea/urea hydrogen-bonded structures were identified.118 To examine not only the Li+/urea interactions but also the strength and extent of hydrogen bonding in the d-U(2000)nLiCF3SO3 xerogels, we decided to inspect the ‘amide I’ and ‘amide II’ bands. The FT-IR spectra of selected samples in both spectral regions are reproduced in Fig. 5.14(a). ‘Amide I’ region (1800–1600 cm−1). The ‘amide I’ region of the di-ureasils corresponds to the amide I119 region of polyamides.120 The amide I mode (or simply carbonyl stretching mode) is a very complex vibration that receives a major contribution from the C=O stretching vibration.120 As it is sensitive to the specific type and magnitude of hydrogen bonding, the amide I band consists of several distinct components which correspond to different C=O environments, usually designated as associations, aggregates or structures. Since the absorption coefficients of C=O groups belonging to different aggregates may be different, only the changes undergone by each mode provide information of concentration variations of each aggregate.120,121 Figure 5.14(a) shows that the addition of Li+ cations to d-U(2000) perturbs deeply the ‘amide I’ region, an indication that the Li+ ions coordinate to the carbonyl oxygen atoms of the urea crosslinks over the whole range of salt concentration examined. The intensity maximum of the ‘amide I’ envelope, situated around 1721 cm−1 in samples with n ≥ 40, is displaced to 1662 cm−1 for n < 40. This result suggests that the hydrogen-bonded



n

1

1720 1710

1752

1632

‘Amide II’ 1662

‘Amide I’

1572

204

Absorbance (a.u.)

5 10 20

40

60 80 200 ∞ 1850 (a)

1800

1750 1700 1650 1600 Wavenumbers (cm–1)

1550

1500

5.14 (a) Room temperature FT-IR spectra of selected d-U(2000)n LiCF3SO3 di-ureasils in the ‘amide I’ and ‘amide II’ regions. (b) Curve-fitting results in the ‘amide I’ region. (c) Salt concentration dependence of the integral intensity of the different spectral components resolved in the ‘amide I’ region. Adapted from Nunes et al.45

associations formed at high salt concentration are more ordered (i.e. stronger) than those found in the more dilute xerogels. To get additional insight into the modifications that take place in the ‘amide I’ region of the d-U(2000)nLiCF3SO3 materials at increasing salt concentration, we performed curve-fitting in the 1800–1600 cm−1 envelope. The results obtained are illustrated in Fig. 5.14(b). The composition dependence of the area of the resolved components is shown in Fig. 5.14(c).


205

1632

1685

1710

1662


n

1

1642

1720

1752

Absorbance (a.u.)

5

20

80

∞ 1850 (b)

1800

1750 1700 1650 Wavenumbers (cm–1)

1600

1550

5.14 (cont.)

The ‘amide I’ band of d-U(2000) was decomposed into five individual components situated at 1751, 1720, 1686, 1662 and 1640 cm−1 (Fig. 5.14b). The band at 1751 cm−1 is associated with the absorption of urea groups in which the N–H or C=O groups are free from any interactions.118 The bands at 1720, 1686 and 1662 cm−1 are ascribed to hydrogen-bonded C=O groups of disordered hydrogen-bonded POE/urea associations of increasing strength.118 The 1640 cm−1 feature is assigned to the absorption of C=O groups included in more ordered hydrogen-bonded urea/urea associations.118


206

Polymer electrolytes 1752 cm–1 1720 cm–1 1710 cm–1 1685 cm–1 1662 cm–1 1642 cm–1 1632 cm–1

80

Integral area fraction (%)

70 60 50 40 30 20 10 0 ∞ (c)

200 80

n

20

10

0

5.14 (cont.)

Globally, the increase of LiCF3SO3 content has three major consequences (Figs. 5.14b and c): The 1752, 1720 and 1642 cm−1 features, ascribed to free C=O groups, disordered POE/urea aggregates and highly ordered urea/urea associations, respectively, suffer a considerable loss of intensity and vanish at n = 5. • The intensity of the 1662 cm−1 band, due to ordered POE/urea structures, is significantly enhanced. • Two new events emerge at 1710 and 1632 cm−1 in the case of the salt-rich materials with n = 5 and 1. •

The most relevant effects, observed at n = 5 (i.e. the growth of new components at 1710 and 1632 cm−1 and the concomitant disappearance of the 1752, 1720 and 1642 cm−1 events), led us to conclude, not only that the saturation of the urea crosslinks is attained at this composition (i.e. no C=O groups remain free), but also that new disordered POE/urea aggregates (1710 cm−1) and highly ordered hydrogen-bonded aggregates (1631 cm−1) are formed at the expense of the disruption of the more disordered POE/urea structures of d-U(2000) (1720 cm−1) and of the highly ordered urea/urea aggregates that give rise to the 1642 cm−1 component. At n = 1 a marked increase of the area of the 1631 cm−1 feature and a sharp reduction of the proportion of the POE/urea aggregates (1662 cm−1) occur. The latter effect must be correlated with the fact that at this salt content the POE segments are



207

massively required to complex the Li+ ions to form the POE3LiCF3SO3 compound and can thus no longer interact with the N–H groups of the urea linkages through hydrogen-bonding to form such hydrogen-bonded aggregates. ‘Amide II’ region The ‘amide II’ mode of the di-ureasils corresponds to the amide II119 feature of polyamides.120 The amide II (or N–H in-plane bending) band is a mixed mode containing a major contribution from the N–H in-plane bending vibration.120 It is sensitive to both chain conformation and intermolecular hydrogen bonding and provides rich information about the distribution of hydrogen bond strengths.120 In the spectrum of d-U(2000) the ‘amide II’ mode appears as a medium intensity, broad band centered at about 1560 cm−1 (Fig. 5.14a),118 which was decomposed into three components at 1571, 1555 and 1529 cm−1 (Fig. 5.14b). This suggests that in the host hybrid hydrogen-bonded aggregates with at least three distinct degrees of order exist. The inclusion of LiCF3SO3 affects the ‘amide II’ envelope. At n = 20 the band profile is significantly modified, because of the loss of the 1529 cm−1component and the growth of a new component at 1586 cm−1, which shifts the intensity maximum of the band envelope to 1571 cm−1 (Fig. 5.14b). This finding indicates that, in agreement with the conclusions retrieved from the analysis of the ‘amide I’ region, in the samples with n ≤ 20 the hydrogen bonds are stronger than in the less concentrated ones.118

5.6

Electrochromic displays

Electrochromic (EC) technology and EC devices are drawing attention for their high potential with respect to solar control and display applications. EC materials have received a great deal of attention due to the important applications stemming from such systems, which include smart windows, displays, anti-glare mirrors, eye-glasses or solar-attenuated windows.122,123 Electrochromism may be defined as the phenomenon that gives rise to a reversible color change by the electrochemical insertion and extraction of electrons and ions into inorganic materials, for example, tungsten oxide (WO3). The difference in the transmittance between the colored (written) state and the fully bleached (unwritten) state indicates the light modulation capability of ECDs. A PE component in an optical device must meet various prerequisites. One of these is that the film must exhibit high transmission. If this condition is not fulfilled, the electrolyte will reduce the color contrast of the optical device.124


208


The central part of an ECD is the electrolyte or ion conductor. It can be made of a thin (inorganic) film, a polymeric layer or a liquid capable of ion transport. PEs are of great and increasing interest for ECDs and allow convenient design incorporating resilient and adhesive layers. In contrast with the liquid electrolyte, there are no obvious problems with leakage and fluid-mechanics-induced distortions.125 In this section we describe the assembly of a prototype solid-state ECD based on a four-layer sandwich structure incorporating materials based on di-ureasils doped with controlled quantities of lithium salts.22,33,35,126 The device assembly was carried out by direct application of a small volume of the di-ureasils to the surface of a glass plate onto which an indium zinc oxide (IZO)/WO3 coating had been previously deposited. The gel was dried in a vacuum oven for a period of seven days. The thickness used for each layer were 170 nm for IZO and 400 nm for WO3. The entire assembly procedure described was carried out under atmospheric condition. A second IZO-coated glass plate was placed on top of the dry electrolyte layer and the two plates were pressed together to spread the electrolyte in a thin film between the EC surfaces. The entire assembly procedure described was carried out under atmospheric conditions. Figure 5.15 reports the optical transmittance in the 300–900 nm wavelength range for the ECDs based on d-U(900)5LiTFSI and d-U(600)25LiBF4 di-ureasils. These images show the compositions with good optical performance and stability. The average transmittance in the visible region of the spectrum was above 69 and 74% for devices based on d-U(900)5LiTFSI and d-U(600)25LiBF4 di-ureasils, respectively. After coloration, while the device based on d-U(900)5LiTFSI presented a good color contrast (above 24%) and an optical density above 0.49, that based on d-U(600)25LiBF4 exhibited an average transmittance of 48% and an optical density above 0.17, providing a good performance in the coloring/bleaching process (Fig. 5.16). These results are very satisfactory, especially when compared to those obtained with the incorporation of d-U(900)nLiClO4 electrolytes.35 The reasons for the variation of optical density with electrolyte composition are not yet clear and require further study. These findings are consistent with mechanical interfacial effects or the supply of ions due to higher ionic conductivity, both aspects of electrolyte performance that vary significantly with composition. The solid-state ECD tested exhibit good electrochemical stability, showed memory effect and may therefore be of interest for application in ‘smart windows’. The evaluation of devices based on di-ureasils confirmed that hybrid electrolytes provide adequate optical density, good stability and enhanced optical memory. Therefore PEs based on di-ureasil composites incorporating lithium guest salts represent an attractive alternative to liquid electrolytes in ECDs.126



209

90

Transmittance (%)

80 70 60 50 U(900)5LiTFSI as-deposited U(900)5LiTFSI colored

40 30 20 400

500

(a)

600

700

800

Wavelength (nm)

Transmittance (%)

80

60

40

d-U(600)25LiBF4 bleached d-U(600)25LiBF4 colored

20 300

400

500

600

700

800

900

Wavelength (nm)

(b)

5.15 Optical transmittance as a function of wavelength for the electrochromic device in bleached and colored state using (a) d-U(900)5LiTFSI and (b) d-U(600)25LiBF4 di-ureasils. Adapted from Rodrigues et al.126

5.7

Conclusion

The present chapter, devoted to a series of attractive lithium ion conductors based on di-urea crosslinked POE/siloxane host networks (di-ureasils) prepared by the sol–gel process, demonstrates that the combination of the hybrid strategy with the host–guest concept derived from inorganic chemistry is extremely interesting. We have shown that the ormolytes discussed here exhibit good thermal, mechanical and electrochemical properties. In particular, we have shown that these materials can be employed as the ion conducting layer of practical


210


(a)

(b)

5.16 ECD in bleached and colored states for selected di-ureasils: (a) d-U(900)5LiTFSI; (b) d-U(600)25LiBF4. Adapted from Rodrigues et al.126

EDCs. The prototype solid state ECDs assembled present encouraging optical contrast and appropriate response times. The advantages of using solid PEs instead of conventional liquid-based electrolytes in ECDs are related to the capacity of preventing leaking, eliminating the need for sealing, since no solvent evaporation takes place. Moreover, ormolytes are flexible materials and thus able to provide good contact between the electrolytic component and the active EC layer. In addition, film deposition of large surface areas is possible, while ensuring that the device is mechanically robust. Organic–inorganic hybrid materials are being increasingly explored and more developments are expected within the next years. It is worth noting that many of these hybrids satisfy socio-economical demands in energy and the environment.

5.8

References

1. wright p v, ‘Electrical conductivity in ionic complexes of poly(ethylene oxide)’, Br Polym J, 1975, 7, 319–327 2. armand m, duclot m t and chabagno j m, ‘Polymeric solid electrolytes’, in Proceedings of the Second International Meeting on Solid State Electrolytes, St. Andrews, UK, Extended Abstract 6.5, 1978



211

3. gray f m, Polymer Electrolytes, RSC Materials Monographs, The Royal Society of Chemistry, London, 1997 4. tarascon j-m and armand m, ‘Issues and challenges facing rechargeable lithium batteries’, Nature, 2001, 414, 359–367 5. berthier c, gorecki w, minier m, armand m b, chabagno j m and rigaud p, ‘Microscopic investigation of ionic conductivity in alkali metal saltspoly(ethylene oxide) adducts’, Solid State Ionics, 1983, 11, 91–95 6. gadjourova z, andreev y g, tunstall d p and bruce p g, ‘Ionic conductivity in crystalline polymer electrolytes’, Nature, 2001, 412, 520–523 7. gomez-romero p and sanchez c, Editors, Functional Hybrid Materials, Wiley Interscience, New York, 2003 8. brinker c j and scherer g w, Sol–Gel Science: The Physics and Chemistry of Sol–Gel Processing, Academic Press, San Diego, CA, 1990 9. ravaine d, seminel a, charbouillot y and vincens m, ‘A new family of organically modified silicates prepared from gels’, J Non-Cryst Solids, 1986, 82, 210–219 10. popall m, andrei m, kappel j, kron j, olma k and olsowski b, ‘ORMOCERs as inorganic–organic electrolytes for new solid state lithium batteries and supercapacitors’, Electrochim Acta, 1998, 43 (10–11), 1155–1161 11. judeinstein p, titman j, stamm m and schmidt h, ‘Investigation of ion-conducting ormolytes: structure-properties relationship’, Chem Mater, 1994, 6, 127–134 12. dahmouche k, atik m, mello n c, bonagamba t j, panepucci h, aegerter m a and judeinstein p, ‘Investigation of new ion-conducting ORMOLYTES: Structure and properties’, J Sol-Gel Sci Technol, 1997, 8, 711–715 13. de zea bermudez v, alcácer l, acosta j l and morales e, ‘Synthesis and characterization of novel urethane cross-linked ormolytes for solid state lithium batteries’, Solid State Ionics, 1999, 116, 197–209 14. wang c, wei y, ferment g r, li w and li t, ‘Poly(ethylene oxide)–silica hybrid materials for lithium battery application’, Mater Lett, 1999, 39, 206–210 15. maccallum j r and seth s, ‘Conductivity of poly(ethylene oxide)/silica composite films containing lithium trifluorosulphonate’, Eur Polym J, 2000, 36, 2337–2341 16. nishio k and tsuchiya t, Sol Energy Mater Sol Cells, 2001, 68, 295–306 17. nunes s c, de zea bermudez v, ostrovskii d, silva m m, barros s, smith m j, sá ferreira r a, carlos l d, rocha j and morales e, ‘Di-urea cross-linked poly(oxyethylene)/siloxane ormolytes for lithium batteries’, J Electrochem Soc, 2005, 152, A429–A438 18. nunes s c, de zea bermudez v, silva m m, smith m j, carlos l d, sá ferreira, r a and rocha j, ‘Sol-gel derived Li+-doped poly(ε-caprolactone)/siloxane biohybrid electrolytes’, J J Solid State Electrochem, 2006, 10(4), 203–210 19. nunes s c, de zea bermudez v, ostrovskii d and martins n v, ‘Vibrational spectra and microstructure of poly(ε-caprolactone)/siloxane biohybrids doped with lithium triflate’, J Molec Struct, 2008, 879, 72–80 20. silva m m, nunes s c, barbosa p c, evans a, de zea bermudez v, smith m j and ostrovskii d, ‘Sol–gel preparation of a di-ureasil electrolyte doped with lithium perchlorate’, Electrochim Acta, 2006, 52, 1542–1548 21. nunes s c, de zea bermudez v, ostrovskii d, tavares p b, barbosa p c, silva m m and smith m j, ‘Spectroscopic and structural study of novel di-ureasils based on lithium perchlorate’, Electrochim Acta, 2007, 53, 1466–1475


212


22. barbosa p c, silva m m, smith, m j, gonçalves a, fortunato e, nunes s c and de zea bermudez v, ‘Di-ureasil xerogels containing lithium bis(trifluoromethanesulfonyl)imide for application in solid-state electrochromic devices’, Electrochim Acta, 2009, 54, 1002–1009 23. gomes correia s, de zea bermudez v, silva m m, barros s, sá ferreira r a, carlos l d, passos de almeida a p and smith m j, ‘Morphological and conductivity studies of di-ureasil xerogels containing lithium triflate’, Electrochim Acta, 2002, 47, 2421–2428 24. sanchez c, julián b, belleville p and popall m, ‘Applications of hybrid organic–inorganic nanocomposites’, J Mater Chem, 2005, 15, 3559–3592 25. armand m, poinsignon c, sanchez j-y and de zea bermudez v, ‘Proton conducting polymer and application thereof as electrolyte in electrochemical devices’, U.S. Patent 5,283,310, 1993 26. de zea bermudez v, poinsignon c and armand m b, ‘Chemistry and physical properties of sulfamide and its derivatives: proton conducting materials’, J Mater Chem, 1997, 7(9), 1677–1692 27. de zea bermudez v, baril d, sanchez j-y, armand m and poinsignon c, ‘Protonvacancy conducting polymers based on anion-grafted Ormosils synthetized by the sol-gel process’, Hugot-Le Goff A, Granqvist C-G and Lampert C M, Editors, Proc SPIE, 1992, 1728, 180–191 28. catlow c r a, chadwick a v, greaves g n, moroney l m and worboys m r, Solid State Ionics, 1983, 9/10, 1107–1113 29. bruce p g, krok f, evans j and vincent c a, ‘Preliminary results on a new polymer electrolyte PEO-Hg(ClO4)2’, Br Polym J, 1988, 20, 193–194 30. armstrong r d and clarke m d, ‘Lithium ion conducting polymeric electrolytes based on poly(ethylene) adipate’ Electrochim Acta, 1984, 29, 1443–1446 31. vincent c a, ‘Polymer electrolytes’, Prog Solid State Chem, 1987, 17, 145– 261 32. watanabe m, rikuwa m, sanui k and ogata n, ‘Effects of polymer structure and incorporated salt species on ionic conductivity of polymer complexes formed by aliphatic polyester and alkali metal thiocyanate’, Macromolecules, 1986, 19, 188–192 33. barbosa p, rodrigues l, silva m m, smith m, gonçalves a and fortunato e, ‘Application of di-ureasil ormolytes based on lithium tetrafluoroborate in solid-state electrochromic displays’, J Mater Chem, 2010, 20, 723–730; Barbosa P, Silva M M and Smith M J, unpublished results 34. barbosa p c, rodrigues l c, silva m m and smith m j, ‘Preparation of hybrid organic–inorganic materials based on a di-ureasil matrix doped with lithium bis(trifluoromethanesulfonyl)imide’ J Power Sources, 2008, 180, 607–611 35. barbosa p c, silva m m, smith m j, gonçalves a and fortunato e, ‘Studies of solid-state electrochromic devices based on PEO/siliceous hybrids doped with lithium perchlorate’, Electrochim Acta, 2007, 52, 2938–2943 36. armand m, gorecki w and andrani r, ‘Perfluorosulphonimide salts as solute for polymer electrolytes’, in 2nd International Symposium on Polymer Electrolytes, Scrosati B, Editor, Elsevier, 1990, 91–97 37. chiodelli g, ferloni p, magistris a and sanesi m, ‘Ionic conduction and thermal properties of poly (ethylene oxide)-lithium tetrafluoroborate films’, Solid State Ionics 1988, 28–30, 1009–1013



213

38. armand m b, chabagno j m and duclot m j (1979), ‘Poly-ethers as solid electrolytes’, in Fast Ion Transport in Solids, Vashita P, Mundy J N and Shenoy G K, Editors, North-Holland, Amsterdam, 131–136 39. machida k and miyazawa t, ‘Infrared and Raman spectra of polyethyleneglycol dimethylethers in the liquid state’, Spectrochim Acta, 1964, 20, 1865– 1873 40. matsuura h and miyazawa t, ‘Vibrational analysis of molten poly(ethylene glycol)’, J Polym Sci, 1969, 7(A-2), 1735–1744 41. matsuura h, miyazawa t and machida k, ‘Infrared spectra of poly(ethylene glycol) dimethyl ethers in the crystalline state’, Spectrochim Acta, 1973, 29A, 771–779 42. papke b l, ratner m a and shriver d f, ‘Vibrational spectroscopic determination of structure and ion pairing in complexes of poly(ethylene oxide) with lithium salts’, J Electrochem Soc, 1982, 129(7), 1434–1438 43. bernson a, lindgren j, huang w and frech r, ‘Coordination and conformation in PEO, PEGM and PEG systems containing lithium or lanthanum triflate’, Polymer, 1995, 36(23), 4471–4478 44. dissanayake m a k l and frech r, ‘Infrared spectroscopic study of the phases and phase transitions in poly(ethylene oxide) and poly(ethylene oxide)-lithium trifluoromethanesulfonate complexes’, Macromolecules, 1995, 28(15), 5312– 5319 45. nunes s c, de zea bermudez v, ostrovskii d and carlos l d, ‘Ionic environment and hydrogen bonding in di-ureasil ormolytes doped with lithium triflate’, J Mol Struct, 2004, 702/1–3, 39–48 46. sato h and kusumoto y, ‘Highly symmetric structure of alkali metal complexes of 18-crown-6 and 15-crown-5 in liquid and crystalline states as revealed by Raman spectra’, Chem Lett, 1978, 7(6), 635–638 47. ericson h, mattson b, torrel l m, rinne h and sundholm f, ‘A Raman spectroscopic investigation of methoxyl end capped PPO doped with NaCF3SO3’, Electrochim Acta, 1998, 43(10–11), 1401–1405 48. brodin a, mattson b, nilsson k, torrel l m and hamara j, ‘Ionic configurations in PEO based electrolytes end-capped by CH3-, OH- and SO3-groups’, Solid State Ionics, 1996, 85, 111–120 49. maxfield j and shepherd w, ‘Conformation of poly(ethylene oxide) in the solid state, melt and solution measured by Raman scattering’, Polymer, 1975, 16(7), 505–509 50. matsuura h and fukuhara k, ‘Conformational analysis of poly(oxyethylene) chain in aqueous solution as a hydrophilic moiety of nonionic surfactants’, J Mol Struct, 1985, 126, 251–260 51. papke b l, dupon r, ratner m a and shriver d f, ‘Ion-pairing in polyether solid electrolytes and its influence on ion transport’, Solid State Ionics, 1981, 5, 685–688 52. nunes s c, de zea bermudez v, ostrovskii d, barbosa p c, silva m m and smith m j, ‘Cationic and anionic environments in LiTFSI-doped di-ureasils with application in solid-state electrochromic devices’, Chem Phys, 2008, 345, 32–40 53. fernandes m and de zea bermudez v, unpublished results 54. wendsjö å, lindgren j and paluszkiewicz c, ‘Structure, dynamics and morphology in the system M(CF3SO3)2PEOn for M = Zn and Pb’, Electrochim Acta, 1992, 37(9), 1689–1693


214


55. johnston h and shriver d f, ‘Vibrational study of the trifluoromethanesulfonate anion: unambiguous assignment of the asymmetric stretching modes’, Inorg Chem, 1993, 32, 1045–1047 56. manning j and frech r, ‘The structure of associated ionic species in poly(propylene oxide)–alkali metal trifluoromethanesulphonate complexes’, Polymer, 1992, 33(16), 3487–3494 57. huang w, frech r and wheeler r a, ‘Molecular structures and normal vibrations of trifluoromethane sulfonate (CF3SO3−) and its lithium ion pairs and aggregates’, J Phys Chem, 1994, 98, 100–110 58. huang w and frech r, ‘Dependence of ionic association on polymer chain length in poly(ethylene oxide)–lithium triflate complexes’, Polymer, 1994, 35(2), 235–242 59. huang w, frech r, johansson p and lindgren j, ‘Cation–polymer interaction and ionic association in diglyme-LiCF3SO3 and diglyme–propylene carbonate–LiCF3SO3 complexes’, Electrochim Acta, 1995, 40(13–14), 2147–2151 60. bergström p-å and frech r, ‘Raman and infrared study of the polymer electrolytes formed by poly(propylene oxide) with Sn(CF3SO3)2 and Pb(CF3SO3)2’, J Phys Chem, 1995, 99, 12603–12611 61. frech r and chintapalli s, ‘Effect of propylene carbonate as a plasticizer in high molecular weight PEO–LiCF3SO3 electrolytes’, Solid State Ionics, 1996, 85, 61–66 62. chintapalli s, zea c and frech r, ‘Characterization studies on high molecular weight PEO-ammonium triflate complexes’, Solid State Ionics, 1996, 92, 205–212 63. chintapalli s and frech r, ‘Effect of plasticizers on ionic association and conductivity in the (PEO)9LiCF3SO3 system’, Macromolecules, 1996, 29, 3499–3506 64. chintapalli s, quinton c, frech r and vincent c a, ‘Compound formation and ionic association in the poly(ethylene oxide)–potassium triflate system’, Macromolecules, 1997, 30, 7472–7477 65. frech r, chintapalli s, bruce p and vincent c a, ‘Structure of an amorphous polymer electrolyte, poly(ethylene oxide)3:LiCF3SO3’, Chem Commun, 1997, 157–158 66. chintapalli s and frech r, ‘Kinetic effects in the ionic association of poly(ethylene oxide)–lithium triflate complexes’, Electrochim Acta, 1998, 43(10–11), 1395–1400 67. rhodes c p and frech r, ‘Cation–anion and cation–polymer interactions in (PEO)nNaCF3SO3 (n = 1–80)’, Solid State Ionics, 1999, 121, 91–99 68. rhodes c p, kiassen b, frech r, dai y and greenbaum s g, ‘Studies of cation– anion and cation–polymer association in poly(ethylene oxide): Pb(CF3SO3)2 complexes’, Solid State Ionics, 1999, 126, 251–257 69. frech r, chintapalli s, bruce p g and vincent c a, ‘Crystalline and amorphous phases in the poly(ethylene oxide)–LiCF3SO3 system’, Macromolecules, 1999, 32, 808–813 70. rhodes c p and frech r, ‘A symmetry-based analysis of Raman and infrared spectra of the compounds (poly(ethylene oxide))3LiCF3SO3 and (poly(ethylene oxide))NaCF3SO3’, Solid State Ionics, 2000, 136–137, 1131–1137 71. york s, frech r, snow a and glatzhofer d, ‘A comparative vibrational spectroscopic study of lithium triflate and sodium triflate in linear poly(ethylenimine)’, Electrochim Acta, 2001, 46, 1533–1537



215

72. york s, kellam iii e c, allcock h r and frech r, ‘A vibrational spectroscopic study of lithium triflate in polyphosphazenes with linear oligoethyleneoxy sidechains of different lengths’, Electrochim Acta, 2001, 46, 1553–1557 73. rhodes c p and frech r, ‘Local structures in crystalline and amorphous phases of diglyme–LiCF3SO3 and poly(ethylene oxide)–LiCF3SO3 systems: implications for the mechanism of ionic transport’, Macromolecules, 2001, 34, 2660–2666 74. papke b l, ratner m a and shriver d f, ‘Vibrational spectroscopy and structure of polymer electrolytes, poly(ethylene oxide) complexes of alkali metal salts’, J Phys Chem Solids, 1981, 42, 493–500 75. macfarlane d r, meakin p, bishop a, mcnaughton d, rosalie j m and forsyth m, ‘FTIR study of ion-pairing effects in plasticized polymer electrolytes’, Electrochim Acta, 1995, 40(13–14), 2333–2337 76. bishop a, macfarlane d r, mcnaughton d and forsyth m, ‘FT-IR investigation of ion association in plasticized solid polymer electrolytes’, J Phys Chem, 1996, 100, 2237–2243 77. bishop a, macfarlane d r and mcnaughton d, ‘Triflate ion association in plasticized polymer electrolytes’, Solid State Ionics, 1996, 85, 129–135 78. ferry a, jacobsson p and torell l m, ‘The molar conductivity behavior in polymer electrolytes at low salt concentrations; a Raman study of poly(propylene glycol) complexed with LiCF3SO3’, Electrochim Acta, 1995, 40(13–14), 2369–2373 79. ferry a, jacobsson p and stevens j r, ‘Studies of ionic interactions in poly(propylene glycol)4000 complexed with triflate salts’, J Phys Chem, 1996, 100, 12574–12582 80. ferry a, orädd g and jacobsson p, ‘A pfg-NMR investigation of the diffusion coefficients of anionic species in the system PPG4000–LiCF3SO3’, Macromolecules, 1997, 30, 7329–7331 81. ferry a and tian m, ‘Influence of hydroxyl terminal groups on the ionic speciation and ionic conductivity in complexes of poly(propylene glycol)(4000) and LiCF3SO3 salt’, Macromolecules, 1997, 30, 1214–1215 82. ferry a, ‘Effects of dynamic spatial disorder on ionic transport properties in polymer electrolytes based on poly(propylene glycol)(4000)’, J Chem Phys, 1997, 107(21), 9168–9175 83. ferry a, ‘Ionic interactions and transport properties in methyl terminated poly(propylene glycol)(4000) complexed with LiCF3SO3’, J Phys Chem B, 1997, 101, 150–157 84. ferry a, doeff m m and dejonghe l c, ‘Transport property and Raman spectroscopic studies of the polymer electrolyte system P(EO)n-NaTFSI’, J Electrochem Soc, 1998, 145, 1586–1592 85. ferry a, orädd g and jacobsson p, ‘A Raman, ac impedance and pfg-NMR investigation of poly(ethylene oxide) dimethyl ether (400) complexed with LiCF3SO3’, Electrochim Acta, 1998, 43(10–11), 1471–1476 86. ferry a, edman l, forsyth m, macfarlane d r and sun j, ‘Connectivity, ionic interactions, and migration in a fast-ion-conducting polymer-in-salt electrolyte based on poly(acrylonitrile) and LiCF3SO3’, J Appl Phys, 1999, 86(4), 2346–2348 87. edman l, ferry a and jacobsson p, ‘Effect of C60 as a filler on the morphology of polymer–salt complexes based on poly(ethylene oxide) and LiCF3SO3’, Macromolecules, 1999, 32, 4130–4133


216


88. stevens j r and shantz s, ‘Variation of the glass transition temperature of poly(propylene glycol)4000 complexed with lithium and sodium salts’, Polym Commun, 1988, 29, 330–331 89. shantz s, sandahl j, börjesson l, torell l m and stevens j r, ‘Ion pairing in polymer electrolytes; a comparative Raman study of NaCF3SO3 complexed in poly(propylene-glycol) and dissolved in acetonitrile’, Solid State Ionics, 1988, 28–30, 1047–1053 90. kakihana m, schantz s and torell l m, ‘Raman spectroscopic study of ion–ion interaction and its temperature dependence in a poly(propylene-oxide)-based NaCF3SO3–polymer electrolyte’, J Chem Phys, 1990, 92(10), 6271–6277 91. kakihana m, schantz s, torell l m and stevens j r, ‘Dissociated ions and ion–ion interactions in poly(ethylene oxide) based NaCF3SO3 complexes’, Solid State Ionics, 1990, 40–41, 641–644 92. shantz s, torell l m and stevens j r, ‘Ion pairing effects in poly(propylene glycol)–salt complexes as a function of molecular weight and temperature: a raman scattering study using NaCF3SO3 and LiClO4’, J Chem Phys, 1991, 94(10), 6862–6867 93. yang x q, lee h s, hanson l, mcbreen j and okamoto y, ‘Development of a new plasticizer for poly(ethylene oxide)-based polymer electrolyte and the investigation of their ion-pair dissociation effect’, J Power Sources, 1995, 54, 198–204 94. johnston s f, ward i m, cruickshank j and davies g r, ‘Spectroscopic studies of triflate ion association in polymer gel electrolytes and their constituents’, Solid State Ionics, 1996, 90, 39–48 95. torell l m, jacobsson p and petersen g, ‘A Raman study of ion solvation and association in polymer electrolytes’, Polym Adv Tech, 1992, 4, 152–163 96. petersen g, torell l m, panero s, scrosati b, silva c j and smith m j, ‘Ionic interactions in MCF3SO3-polyether complexes containing mono-, di- and trivalent cations’, Solid State Ionics, 1993, 60, 55–60 97. brodin a, mattson b, nilsson k, torell l m and hamara j, ‘Ionic configurations in PEO based electrolytes endcapped by CH3-, OH- and SO3-groups’, Solid State Ionics, 1996, 85, 111–120 98. petersen g, brodin a, torell l m and smith m j, ‘Light scattering and luminescence studies of M(CF3SO3)x–polyether complexes containing trivalent cations’, Solid State Ionics, 1994, 72, 165–171 99. brodin a, mattsson b and torell l m, ‘A site selective luminescence excitation and Raman scattering study of Eu(CF3SO3)3-PPO; ion association and ionic exchange rate’, J Chem Phys, 1994, 101(6), 4621–4627 100. bernson a and lindgren j, ‘Ion aggregation and morphology for poly (ethylene oxide)-based polymer electrolytes containing rare earth metal salts’, Solid State Ionics, 1993, 60, 31–36 101. bernson a and lindgren j, ‘Free ions and ion pairing/clustering in the system LiCF3SO3–PPOn’, Solid State Ionics, 1993, 60, 37–41 102. gejji s p, hermansson k, tegenfeldt j and lindgren j, ‘Geometry and vibrational frequencies of the lithium triflate ion pair: an ab initio study’, J Phys Chem, 1993, 97, 11402–11407 103. wendsjö å, lindgren j, thomas j o and farrington g c, ‘The effect of temperature and concentration on the local environment in the system M(CF3SO3)2PEOn for M = Ni, Zn and Pb’, Solid State Ionics, 1992, 53–56, 1077



217

104. stevens j r and jacobsson p, ‘A comparison of acetone and poly(propy1ene glycol) as solvents for lithium triflate and lithium perchlorate’, Can J Chem, 1991, 69, 1980–1984 105. jacobsson p, albisson i, mellander b-e and stevens j r, ‘Ion association effects and phase separation in poly(propylene oxide) modified poly(dimethylsiloxane) complexed with triflate salts’, Polymer, 1992, 33(13), 2778–2783 106. de zea bermudez v, ostrovskii d, lavoryk s, gonçalves m c and carlos l d, ‘Urethane cross-linked poly(oxyethylene)/siliceous nanohybrids doped with Eu3+ ions. Part II: Ion association’, Phys Chem Chem Phys, 2004, 6(3), 649–658 107. rey i, johansson p, lindgren j, lassègues j c, grondin j and servant l, ‘Spectroscopic and theoretical study of (CF3SO2)2N− (TFSI−) and (CF3SO2)2NH (HTFSI)’, J Phys Chem A, 1998, 102(19), 3249–3258 108. bakker a, gejji s, lindgren j, hermansson k and probst m m, ‘Contact ion pair formation and ether oxygen coordination in the polymer electrolytes M[N(CF3SO2)2]2PEOn for M = Mg, Ca, Sr and Ba’, Polymer, 1995, 36, 4371–4378 109. ferry a, furlani m, franke a, jacobsson p and mellander b-e, ‘Spectroscopic studies of luminescent and ionically conducting Eu[N(CF3SO2)2]3-PPG4000 complexes’, J Chem Phys, 1998, 109(7), 2921–2928 110. furlani m, ferry a, franke a, jacobsson p and mellander be, ‘Time resolved luminescence and vibrational spectroscopic studies on complexes of poly(ethylene oxide) oligomers and EuTFSI3 salt’, Solid State Ionics, 1998, 113–115, 129–138 111. francisco j s and williams i h, ‘Structural and spectral consequences of ion pairing. 4. Theoretical study of alkali metal tetrafluoroborates (M+BF4−; M = lithium, sodium, potassium, and rubidium)’, J Phys Chem, 1990, 94, 8522–8529 112. suresh c h, gadre s r and gejji s p, ‘Theoretical studies on the structure of M+ BF4− ion pairs M = Li+, NH4+: the role of electrostatics and electron correlation’, Theor Chem Acc, 1998, 99, 151–157 113. dhumal n r. and gejji s p, ‘Theoretical studies on blue versus red shifts in diglyme–M+–X– (M = Li, Na, K and X = BF4, ClO4, SCN)’, J Molec Struct (THEOCHEM), 2006, 758, 233–240 114. quist a s, bates j b and boyd g h, ‘Raman spectra of molten NaBF4 to 606 °C and 8% NaF-92% to 503 °C’, J Chem Phys, 1971, 54, 4896–4901 115. de loos m, van esch j, stokroos i, kellogg r m and feringa b l, ‘Remarkable stabilization of self-assembled organogels by polymerization’, J Am Chem Soc, 1997, 119, 12675–12676 116. van esch j, kellogg r m and feringa b l, ‘Di-urea compounds as gelators for organic solvents’, Tetrahedron Lett, 1997, 38(2), 281–284 117. born l and hespe h, ‘On the physical crosslinking of amine-extended polyurethane urea elastomers: a crystallographic analysis of bis-urea from diphenyl methane-4-isocyanate and 1,4-butane diamine’, Colloid Polym Sci, 1985, 263, 335–341 118. de zea bermudez v, carlos l d and alcácer l, ‘Sol–gel derived urea crosslinked organically modified silicates. 1. Mid-infrared spectra’, Chem Mater, 1999, 11, 569–580 119. miyazawa t, shimanouchi t and mizushima s-i, ‘Characteristic infrared bands of monosubstituted amides’, J Chem Phys, 1956, 24(2), 408–418


218


120. skrovanek d j, howe s e, painter p c and coleman m m, ‘Hydrogen bonding in polymers: infrared temperature studies of an amorphous polyamide’, Macromolecules, 1985, 18, 1676–1683 121. coleman m m, lee k h, skrovanek d j, and painter p c, ‘Hydrogen bonding in polymers. 4. Infrared temperature studies of a simple polyurethane’, Macromolecules, 1986, 19, 2149–2157 122. wang x j, lau w m and wong k y, ‘Display device with dual emissive and reflective modes’, Appl Phys Lett, 2005, 87, 113502-1–113502-3 123. mortimer r g, ‘Electrochromic materials’, Chem Soc Rev, 1997, 26, 147–156 124. cui m-h, guo j-s, xie h-q, wu z-h and qiu s-c, ‘All-solid-state complementary electrochromic windows based on the oxymethylene-linked polyoxyethylene complexed with LiClO4’, J Appl Polym Sci, 1997, 65, 1739–1744 125. granqvist c g, azens a, hjelm a, kullman l, niklasson g a, rönnow d, strømme mattsson m, veszelei m and vaivars g, ‘Recent advances in electrochromics for smart windows applications’, Solar Energy, 1998, 63, 199–216 126. rodrigues l c, barbosa p c, silva m m, smith m j, gonçalves a and fortunato e, ‘Application of hybrid materials in solid-state electrochromic devices’, Opt Mater, 2009, 31, 1467–1471


6 Hybrid inorganic–organic polymer electrolytes V. D I N O T O, E. N E G R O and S. L AV I N A, University of Padova, Italy, and M. V I T TA D E L L O, City University of New York, USA

Abstract: Polymer electrolytes (PEs) are macromolecular systems capable of transporting charged species such as ions or protons. The main application of PEs is in energy conversion and storage devices such as batteries and fuel cells. The chapter overviews the synthesis, structure, physical and electrical properties of three classes of hybrid inorganic– organic PEs: three-dimensional hybrid inorganic–organic networks as polymer electrolytes (3D-HION-APE), zeolitic inorganic–organic polymer electrolytes (Z-IOPEs) and hybrid gel electrolytes (HGEs). The basic structure of the materials consists of organic macromolecules bridging inorganic clusters or species. The chapter also includes an overview of the methods used in the characterization of the structure and of the electrical conductivity of PEs, with a particular reference to the jump relaxation model. Key words: inorganic–organic polymer electrolytes, Z-IOPE, electrical spectroscopy, vibrational spectroscopy, conductivity spectra.

6.1

Introduction

One of the most active research areas in solid state electrochemistry concerns the development of ion-conducting materials for application in the conversion and storage of energy (e.g. high energy density batteries and fuel cells). Polymer electrolytes (PEs) are a class of materials which are particularly promising in this sense, as witnessed by the massive research efforts spent in the last 20 years to obtain systems characterized by a good conductivity and a high chemical, thermal and electrochemical stability. This chapter overviews hybrid inorganic–organic PEs, which are characterized by inorganic species included in the chemical composition of the materials and/or in their structure at a molecular or nanometric level. Three families of hybrid inorganic–organic PEs are reviewed: (a) three-dimensional hybrid inorganic–organic networks as polymer electrolytes (3D-HION-APEs); (b) zeolitic inorganic–organic polymer electrolytes (Z-IOPEs); and (c) hybrid gel electrolytes (HGEs). It is shown that these materials are able to transport either monovalent species (e.g. lithium ions or protons) or bivalent species (e.g. magnesium ions), showing a remarkable conductivity (ca. 10−5 S cm−1 or more at room temperature). In addition, nanocomposite 219 © Woodhead Publishing Limited, 2010

220


inorganic–organic proton-conducting PEs based on perfluorinated ionomers and various ceramic oxoclusters are briefly overviewed. The chapter includes a discussion about the methods used in the structural characterization and in the study of the conductivity of these families of hybrid inorganic–organic PEs. In conclusion, the jump relaxation model is discussed as a theoretical framework for the interpretation of the electrical response of PEs in terms of the real component of their conductivity spectra.

6.2

Fundamentals of polymer electrolytes

PEs are crucial materials for the development of modern electric devices such as high energy density batteries, sensors, fuel cells and electrochromic displays (Scrosati, 1993, p182; Gray, 1997; Wright, 1998; Dias et al., 2000). After more than 20 years of research in the field of PEs, the distinct desirability of an entirely solid state energy storage system is still driving the quest to enhance the ionic conductivity of ‘dry’ solid PEs (higher than 10−3 S cm−1 at room temperature), as described by Scrosati (1993, p182), Gray (1997) and Armand et al. (2002). This is still true even after the recent introduction of the lithium–polymer technology for the reversible storage of energy. The latter systems are based on polymer gels consisting of liquid electrolytes immobilized in various polymer matrices, resulting in a conductivity of 10−2–10−3 S cm−1 at room temperature (Feuillade and Perche, 1975; Abraham and Alamgir, 1990; Croce et al., 1993; Stallworth et al., 1995). Unfortunately, these latter systems have the following drawbacks: (a) leakage of liquid upon squeezing and (b) loss of specific power and energy due to the inert host polymer. Up to now, however, the early expectation to obtain electrolytic complexes with conductivities comparable to those of super-ionic conductors has not been met, as discussed by Gadjourova et al. (2001). The research in the field of PEs is now facing a major dualism regarding the best way to improve ionic conductivity in conventional systems. On one hand, many highly amorphous electrolytic complexes have been synthesized with the goal of reducing the crystalline order as much as possible. This approach included the use of inorganic–organic polymer hosts with a low glass transition temperature (Tg), as discussed by Bouridah et al. (1985) and Blonsky et al. (1986), and the addition of plasticizers to polyether electrolytes (Huq et al., 1992). On the other hand, more recently it was suggested that the design of preferentially ordered materials could provide faster ion-conducting pathways. This has been achieved in liquid crystalline PEs (Wright et al., 1998) and in stretched PEs above the Tg (Chung et al., 1999). Finally, proton-conducting PEs play a crucial role in the development of advanced energy conversion devices such as proton exchange membrane fuel cells (PEMFCs). These systems are particularly attractive


Hybrid inorganic–organic polymer electrolytes

221

as sources of power in portable applications and light-duty vehicles because of their high energy density (Cleghorn et al., 1997; Appleby, 1999; Larminie and Dicks, 2000; Ogden, 2003; Theisen, 2003). Perfluorinated polymer electrolytes such as Nafion, Aciplex, Flemion and Dow membranes are some of the most promising electrolyte membranes for PEMFCs (Mauritz and Moore, 2004). Nevertheless, the major drawbacks to their large-scale commercial use involve the high costs, the large crossover of reagents and a low proton conductivity at a low hydration degree and at temperatures higher than 100 °C (Mauritz and Moore, 2004; Kocha, 2003; Neergat et al., 2003).

6.3

Overview of hybrid inorganic–organic polymer electrolytes

In inorganic–organic polymer electrolytes, inorganic atoms are introduced in the polymer matrix of hybrid PEs. Three classes of inorganic–organic hybrids endowed with promising ionic conductivities will be presented as follows. The first class (Class I) consists of materials prepared by copolymerization of organic macromolecules with metal and non-metal alkoxides and are known as 3D-HION-APE (3-dimensional hybrid inorganic organic networks as polymer electrolytes), discussed by Di Noto et al., 1996a,b, 2002a, 2003a, 2004a; Münchow et al., 2000; Biscazzo et al., 2002; Di Noto, 2002; Vittadello et al., 2002; Di Noto and Zago, 2004. These materials are threedimensional networks where organic macromolecules are bridged by inorganic atoms like Si, Ti, Zr, Al. Their conductivity depends on the doping salts and on the size of the coordination ‘nests’ present in the host material, and is up to 10−5 S cm−1 at room temperature. It is also possible to include in Class I materials nanocomposite systems obtained by dispersing nanopowders of ceramic oxoclusters in perfluorinated polymer electrolytes such as Nafion (Di Noto et al., 2006, 2007, 2008, 2009, 2010a,b; Vittadello et al., 2008; Thayumanasundaram et al., 2010). These hybrid inorganic–organic materials are devised to overcome the drawbacks of conventional perfluorinated proton-conducting PEs. Indeed, with respect to the pristine PE, they are characterized by a higher proton conductivity at higher temperatures (up to ca. 5.9 × 10−2 S cm−1 vs. 3.3 × 10−2 S cm−1 at 115 °C, as reported by Di Noto et al., 2010b). Class I materials are completed by organosiloxanes featuring dangling —SO3H groups. The latter systems are devised as proton-conducting materials and belong to the 3D-HION-APPE family as described by Di Noto and Vittadello (2005) and Di Noto et al. (2005). The second class of inorganic–organic hybrids (Class II) is that consisting of the so-called zeolitic inorganic–organic polymer electrolytes (Z-IOPE), described by Di Noto (1997, 2000), Di Noto et al. (2000, 2001, 2002b, 2003b)


222


and Vittadello et al. (2003). This class exhibits the following properties: (a) the organic macromolecules are linked to one another by bridging inorganic clusters; (b) the inorganic clusters are formed by the aggregation of two or more inorganic coordination complexes; and (c) the inorganic clusters can be either positively or negatively charged. These systems are generally prepared by sol → gel → plastic transitions and show a high ionic conductivity (up to 10−5 S cm−1 at room temperature). Hybrid inorganic–organic gels (HGEs) make up the third class of materials (Class III, described by Di Noto et al., 2004b). They are obtained starting from hard and soft metal precursors in a sol → gel process using a lowmolecular weight organic ligand. HGEs can be considered as an upgraded version of Z-IOPEs.

6.3.1 Class I: three-dimensional hybrid inorganic–organic networks as polymer electrolytes (3D-HION-APE) 3D-HION-APEs are reviewed presenting the following examples: • poly[(oligo ethyleneoxide) ethoxysilane] (I) and poly[(oligo ethyleneoxide) ethoxysilane]/(EuCl3)0.67 (II) (Di Noto et al., 1996a,b); • poly[PEG400-alt-DEOS]/(MgCl2)x (6.28 × 10−2 ≤ x ≤ 13.16, PEG = polyethylene glycol, DEOS = diethoxydimethylsilane) (Biscazzo et al., 2002; Di Noto, 2002; Di Noto et al., 2002a; Vittadello et al., 2002); • two poly[(oligo ethyleneglycol) dihydroxytitanate] electrolytic systems (Münchow et al., 2000); • {Zr[(CH2CH2O)8.7]ρ/(LiClO4)z}n (1.80 ≤ ρ ≤ 1.99, 0 ≤ z ≤ 0.90) (Di Noto et al., 2003a) and {Al[(CH2CH2O)8.7]ρ/(LiClO4)z}n (1.85 ≤ ρ ≤ 2.24, 0 ≤ z ≤ 1.06) (Di Noto and Zago, 2004; Di Noto et al., 2004a); • two siloxanic proton conducting membranes (Di Noto and Vittadello, 2005; Di Noto et al., 2005); • proton conducting membranes based on Nafion and [(TiO2)·(WO3)0.148] ‘core–shell’ ceramic oxoclusters (Di Noto et al., 2010b). Poly[(oligo ethyleneoxide) ethoxysilane], i.e. {Si(OEt)1.45 [–O(CH2CH2O)8.7]2.55}n [I] and poly[(oligo ethyleneoxide) ethoxysilane]/ (EuCl3)0.67 i.e. {Si(OEt)1.62 [–O(CH2CH2O)8.7]2.38}n/(EuCl3)0.67 [II] were synthesized by reacting tetraethoxysilane with oligo(ethylene glycol) of molecular weight 400 and oligo(ethylene glycol)400/(EuCl3)0.317, respectively, as shown in Fig. 6.1. The resulting products were very transparent and rubbery; Fourier transform infrared (FT-IR) and Raman studies allowed clarification of the structure of these materials. The latter are cross-linked macromolecular systems where the Si atom is bonded to one –OEt group and to three poly(ethylene oxide) 400 chains. Laser luminescence investigations on [II] showed that Eu3+ ion in the polymer host is located in two different types



223

H–O–(CH2–CH2–O)P–H · (EuCl3)w

H–O–(CH2–CH2–O)P–H + w EuCl3 OEt x

Si OEt

+ y H–O–(CH2–CH2–O)P–H · (EuCl3)w

OEt OEt

1. Refluxed at 160–170 °C for 2 h 2. Reduced pressure at 0.1 atm

OEt –(O–CH2–CH2)P–O Si –(O–CH2–CH2)P–O

O–

• (EuCl3)z

(C

H

2–

CH

2–

O–(CH2–CH2–O)P–

O)

P

Si O–(CH2–CH2–O)P– OEt n OEt = –O–CH2–CH3 p=8.68 z=0.67

6.1 Preparation procedure of poly[(oligo ethyleneoxide) ethoxysilane]/ (EuCl3)0.67 [II]. Copyright Wiley-VCH Verlag GmbH & Co. Reproduced with permission from Di Noto et al. (1996a).

of sites having a distorted C2v symmetry; the structure of one site is shown in Fig. 6.2. The conductivity of these systems showed a Vogel–Tamman– Fulcher (VTF)-type dependence on temperature. At 70 °C the conductivities of [I] and [II] were 9 × 10−6 S cm−1 and 14.3 × 10−6 S cm−1 respectively. The preparation of poly[PEG400-alt-DEOS]/(MgCl2)x complexes was performed by gradually adding δ-MgCl2 to poly[PEG400-alt-DEOS]. Seven complexes with the formula poly[PEG400-alt-DEOS]/(MgCl2)x with 6.28 × 10−2 ≤ x ≤ 13.16 were prepared by diluting the mother solution 1 : 2n (n = 0, 1, 2, 3, 4, 5, 6) with the pure copolymer poly[PEG400-alt-DEOS]. To obtain a pure polymer host suitable for the preparation of the desired material, the polycondensation reaction between PEG400 and DEOS was performed


224

Polymer electrolytes (a)

Eu O C Cl

+

Cl(1) O(1) O(2)

Cl(2)

Cl–

O(4)

O(3) (b)

Cl(1)

C2 Cl(2) Eu

O(4) O(3)

O(2)

O(1)

72

6.2 Drawing of the TGT conformational model of a PEO/EuCl3 complex along (a) and parallel (b) to the 72 helix. Copyright Wiley-VCH Verlag GmbH & Co. Reproduced with permission from Di Noto et al. (1996b).

without any catalyst. Analytical data and spectroscopic investigations (1H-, 13C-, and 29Si-NMR (nuclear magnetic resonance), FT-IR) showed that this material is based on host macromolecules whose chains correspond to α-hydro-ω-oligo (oxyethylene) hydroxypoly[oligo (oxyethylene) oxydimethyl sililene], as shown in Fig. 6.3(a). Thus, the copolymer poly[PEG400alt-DEOS] consists of polyethereal moieties linked together by dimethylsililenic units. Solvent-free poly[PEG400-alt-DEOS]/(MgCl2)x electrolytic materials are particularly suitable for the fundamental study of ion–polymer interactions due to the fact that the traces of solvent which can influence the ionic transport properties are completely absent, as described by Ratner and Shriver (1988) and Sun et al. (1996). The polymer poly[PEG400-alt-DEOS] is a highly flexible linear chain based on the PEG400 moieties linked by the dimethylsiloxane bridges. The flexibility is due to the high torsional mobility of the dimethyl sililenic units, which is typical of the polysiloxanic derivatives. Owing to this latter property, it is expected that the typical conformation of PEG400 is maintained in the polyethereal moieties of poly[PEG400-alt-DEOS]. The semiquantitative analysis of the OH and CO vibrational modes provided a deep insight into the structure and the ion–polymer interactions in these systems. Results indicated that: (a) the polyethereal moieties assume



80

75

70

β

65

60

ω

α

α

55

50

45

35

ppm

40

30

25

20 15

10

5

0

–5 –10

ε

HO(CH2CH2O) Si OCH2CH2O(CH2CH2O)CH2CH2O H β β ω 8.671 CH 6.671 ω 3 20.739 ε

(b)

C2

+1

C2

PPD-MgCl2

PPD-[MgCl]+

PPD-Mg2+

6.3 (a) 13C-NMR spectrum of poly[PEG400-alt-DEOS] (PPD) in CDCl3 solution; reproduced from Biscazzo et al. (2002) by permission of Elsevier. (b) Possible coordinations of the Mg2+ cation along the polyethereal chains of PPD; reproduced from Vittadello et al. (2002) by permission of Elsevier.

(a)

85

α

ε

CH3

+2

C2

226


Fibre period 19.3 Å (a)

Oxygen

Carbon

60° (gauche)

(c) (b)

191.5° (trans)

1Å

6.4 Scheme of the typical 72 helical conformation, factor group D(4π/7), of PEO fragments in polymer electrolytes: (a) lateral view; (b) axial view; and (c) distinction between the trans (T) and the gauche (G) spatial configuration.

a TGT (T = trans, G = gauche) type conformation, as highlighted in Fig. 6.4(c); (b) the doping MgCl2 salt substantially modifies the hydrogen bonding structures of the pure polymer and does not change the chain conformation significantly; (c) the Mg2+ cations are coordinated in a distorted C2v symmetry by the ethereal oxygen atoms of the polyether units. The exact coordination of Mg depends strongly on the ratio between the number of Mg atoms and the number of oxygen atoms, nMg/nO, as shown in Fig. 6.3(b). When nMg/nO ≤ 1.23 × 10−3, the Mg2+ cation is coordinated in a pseudo-tetrahedral coordination geometry (PPD–Mg2+; PPD = poly [PEG400-alt-DEOS]). When nMg/nO = 3.04 × 10−3, the monovalent cationic [MgCl]+ species is present along the polyether chains with a pseudo-trigonal bipyramidal coordination geometry (PPD–[MgCl]+). When 3.04 × 10−3 < nMg/nO ≤ 5.55 × 10−2, the monovalent cationic [MgCl]+ and the neutral MgCl2 species are present with the pseudo-trigonal bipyramidal and pseudooctahedral coordination geometries (PPD–[MgCl]+ and PPD–MgCl2), respectively. When nMg/nO ≤ 1.23 × 10−3, Cl− anions preferentially form hydrogen bonding clusters with the OH groups; as nMg/nO > 1.23 × 10−3, each OH group is hydrogen-bonded to a single Cl− anion. The proposed mecha-



227

+2 +1 Cl Mg2+ Intra-CH hopping

+2 [Mg-Cl]+ Inter-CH hopping +1

Cl +1

[Mg-Cl]+ Intra-CH hopping

+1 (a)

(b) Oxygen

Carbon

6.5 Models of charge migration in poly[PEG400-alt-DEOS]/(MgCl2)x: (a) inter-chain migration; (b) intra-chain migration.

nisms for the inter- and intra-chain charge migration in these materials are shown in Fig. 6.5. Two 3D-HION-APE composed from poly(oligoethylene glycol) moieties linked together by Ti atoms are prepared by a condensation reaction between anhydrous PEG 400 (or a solution of LiCl in PEG) with Ti(OEt)4 (Münchow et al., 2000). The poly[(oligoethylene glycol) dihydroxytitanate] (I) and poly[(oligoethylene glycol) dihydroxytitanate]/LiCl complex (II) were obtained; the reaction scheme is reported in Fig. 6.6, while the chemical structures of (I) and (II) are reported in Fig. 6.7. In the synthesis reaction two distinct nucleophilic substitutions to Ti(OEt)4 are observed: (a) the terminal hydroxyls of PEG 400 chains coordinate Ti with the subsequent elimination of ethanol and (b) H2O molecules coordinate Ti(IV) atoms with the elimination of ethanol. Vibrational studies show that polyether chains are present in the TGT conformation and that extended hydrogen bond crosslinks occur in these polymers. In polymer (I), these inter-chain interactions occur between hydroxyl groups belonging to titanium atoms and ethereal oxygens, while polymer (II) exhibits both these interactions, together with a second type of hydrogen bonding interaction due to the formation of hydrogen bonding clusters around Cl− ions. The analysis of the


228


(I) n Ti(OEt)4 + 2n H2O + n HO(CH2CH2O)8.671H

Ti(OH)2O(CH2CH2O)8.671

n

+ 4n EtOH

(II) 1.2 LiCl + HO(CH2CH2O)8.671H

OH(CH2CH2O)8.671H/(LiCl)1.2 Ti(OH)2O(CH2CH2O)8.671

4n Ti(OEt)4 + 8n H2O + 5n HO(CH2CH2O)8.671H/(LiCl)1.2

2 HO(CH2CH2O)8.671H·LiCl

Ti(OH)2O(CH2CH2O)8.671

2

· (LiCl)2 ·(LiCl) · (LiCl)2

+ 16n EtOH n

6.6 Reaction procedures to obtain poly[(oligoethylene glycol) dihydroxytitanate] (I) and poly[(oligoethylene glycol) dihydroxytitanate]/LiCl complex (II); reproduced from Münchow et al. (2000) by permission of Elsevier.

conductivity profiles obtained for these materials evidenced two conductivity regions for polymer (I) and three for polymer (II). All the regions were fitted very well by the empirical VTF equation, suggesting that both polymers conduct ionically by two distinct mechanisms, which are strongly influenced by the segmental motions of the polymer chains. The first conductivity mechanism involves the migration of anionic species from the cathode to the anode due to the presence of reduced titanium species in –[Ti(OH)2–O(CH2–CH2–O)8.671–]n– chains, as outlined in Fig. 6.8. The second conductivity mechanism occurs due to the classical hopping migration of Li+ and Cl− ions. Figure 6.9 reports the VTF graphs for (I) and (II). The thermal stability of these polymers and the possibility of using them to produce thin films are features that make these materials very promising polymer electrolytes. Furthermore, (I) and (II) are rare examples of polymer electrolytes having transition metal atoms in their backbone chains. Eleven network complexes with general formula {Zr[O(CH2CH2O)8.7]ρ / (LiClO4)z}n were prepared by a substitution reaction starting from PEG400, Zr(O(CH2)3CH3)4 and LiClO4 (Di Noto et al., 2003a), as outlined in Fig. 6.10. The resulting materials were transparent and rubbery; 1.80 ≤ ρ ≤ 1.99 and 0 ≤ z ≤ 0.90. FT-Raman analyses showed that: (a) the polyether chains in the bulk materials are present in a TGT conformation; (b) the polymer electrolytes are inorganic–organic network materials with zirconium atoms bonded together by polyether bridges; (c) an unexpected anion-trapping ability toward ClO4− in the hybrid inorganic–organic host matrix is present. The 3D structure of the {Zr[O(CH2CH2O)8.7]ρ/(LiClO4)z}n materials is outlined in Fig. 6.11, while the tetrahedral coordination geometry around the cross-linking Zr atoms is depicted in Fig. 6.12. Impedance spectroscopy studies demonstrated that the proposed systems conduct ionically by a mechanism mediated by the segmental motion and by the concentration of different ionic species distributed in the bulk materials. In particular, two



H

Ti

H

O

O

O

O

O

Ti

H

H

O

= (CH2CH2O)8.671

O

O

H

Ti

H

O

O

O

O

O

O

H

Ti

O

O

H

Ti

H

O

O

O

O H

Ti

H

H

Ti

H

O

O

O

H

O

Ti

O

O

O

O

H

O H

O

O

H

Ti

[Ti(OH)2O(CH2CH2O)8.671]n

H

(I)

H–O

O H

O

O

O

O

O

H

H

Ti

H

O

O

O

Ti

H

Cl–

O

O

Li

H

H

+

O

O

Ti

H

= (CH2CH2O)8.671

O–H = HO(CH2CH2O)8.671H

Ti

H

H

O

Ti

H

H

Ti O

O

O

Li+

H

H

H

Cl–

H

O

O

O

Li+

H O

O

Ti

H

[Ti(OH)2O(CH2CH2O)8.671]

2

HO(CH2CH2O)8.671H·LiCl

2

O

O

Cl– H

O

· (LiCl)2

· (LiCl)2

Ti

O

O

H

H

O

H

Ti

[Ti(OH)2O(CH2CH2O)8·671]

O

H

O

O

O

Li+

Ti

Cl–

H

O

O

O

(II)

O

O

·(LiCl)

H

Ti

H

n

6.7 Structures proposed for poly[(oligoethylene glycol) dihydroxytitanate] (I) and poly[(oligoethylene glycol) dihydroxytitanate]/LiCl complex (II); reproduced from Münchow et al. (2000) by permission of Elsevier.

H

O

Ti

H

H

O

Ti

H

Polymer electrolytes R

R

+ n e–

O (II/III) n– Ti(OH)2

O (IV) 0 Ti(OH)2

O

O – n e–

R

n = 1, 2

R

6.8 Redox mechanism for the conductivity mechanism involving titanium species in –[Ti(OH)2–O(CH2–CH2–O–)8.671–]n– chains of poly[(oligoethylene glycol) dihydroxytitanate] (I) and poly[(oligoethylene glycol) dihydroxytitanate]/LiCl complex (II); reproduced from Münchow et al. (2000) by permission of Elsevier.

(IIC) –8

(IB)

–9

–10 In [σ (T ) / S cm–1]

230

(IIB) –11 (II)' (IIA)

–12 (I) (II) –13

(II)'

–14 (IA)

2.8

2.9

3.0

3.1

3.2

3.3

3.4

T –1 (K–1) × 103

6.9 VTF graphs for poly[(oligoethylene glycol) dihydroxytitanate] (I) and poly[(oligoethylene glycol) dihydroxytitanate]/LiCl complex (II). Conductivity regions of polymers (I) and (II) are indicated; reproduced from Münchow et al. (2000) by permission of Elsevier.


Hybrid inorganic–organic polymer electrolytes (I) PEG400 + y LiClO4

231

PEG400/(LiClO4)y 0≤ y ≤0.49

mα PEG400/(LiClO4)y + m Zr[OCH2CH2CH3]4 4m CH3CH2CH2OH

(II)

{Zr[O(CH2CH2O)8.7]α/(LiClO4)β}m Wash three times with boiling anhydrous toluene {Zr[O(CH2CH2O)8.7]r /(LiClO4)z}n 1.80≤ r ≤1.99

0≤ z ≤0.90

6.10 Reaction steps (I and II) to obtain {Zr[O(CH2CH2O)8.7]ρ/(LiClO4)z}n complexes with 1.80 ≤ ρ ≤ 1.99 and 0 ≤ z ≤ 0.90.

conductivity regions were detected: region I was detected for c1/2 Li ≤ 0.4 −1 1/2 (mol kg−1)1/2, and region II for c1/2 ≥ 0.4 (mol kg ) . Finally, the hybrid Li network with a nLi/nO molar ratio of 0.0223 exhibits a conductivity of ca. 1 × 10−5 S cm−1 at 40 °C, thus suggesting that this solid polymer electrolyte is a good hybrid inorganic–organic ionic conductor. Eleven new hybrid inorganic–organic networks, with the general formula {Al[O(CH2CH2O)8.7]ρ/(LiClO4)z}n, where 1.85 ≤ ρ ≤ 2.24 and 0 ≤ z ≤ 1.06, were prepared by a polycondensation reaction starting from aluminium isopropoxide and polyethylene glycol (PEG400)/(LiClO4)y liquid polymer electrolytes with 0 ≤ y ≤ 0.49 (Di Noto and Zago, 2004; Di Noto et al., 2004a). These materials, which present a glassy rubbery consistency, consist of Al atoms bonded together by PEG400 bridges, as outlined in Fig. 6.13. Thermogravimetric investigations indicated that they are thermally stable up to 260 °C. Medium FT-IR and FT-Raman studies showed that the polyether moieties exhibit a TGT conformation with a helical geometry and detected the presence of Li+. . .ClO4−. . .Al[O(CH2CH2O)8.7]3 neutral species at c1/2 Li ≥ 0.4 (mol kg−1)1/2. Impedance spectroscopy studies, yielding results exemplified in Fig. 6.14, showed that the {Al[O(CH2CH2O)8.7]ρ/(LiClO4)z}n materials conduct ionically by a charge transfer mechanism mainly regulated by the segmental motion and fast ion-hopping processes between equivalent coordination sites distributed along the polyether chains. It was evidenced that in bulk {Al[O(CH2CH2O)8.7]ρ/(LiClO4)z}n complexes the acid aluminium crosslinking sites, which are characterized by two different coordination geometries as reported in Fig. 6.15, exhibit the anion trapping phenomenon


232

Polymer electrolytes OH Zr

Zr O– +

Zr Zr

OH

Zr

Zr

Zr Zr

O– Zr + Zr

Zr

Zr

Zr OH

O–

Zr

Zr

Zr

Zr

Zr

Zr Zr

O– Zr

Zr

+ Zr

Zr Zr

Zr

Zr

O–

Zr

Zr

Zr

+ Zr

Zr

Zr

Zr

O–

–

O

OH Zr

Zr + Zr

Zr

Zr

Zr

Zr Zr

–

O + Zr

Zr OH

OH

OH

O– Zr

+ Zr

Zr Zr

Zr Zr

Zr

Zr

Zr

Zr Zr

Zr

+ Zr

Zr

Zr Zr

O–

Zr Zr

Zr

+ Zr

Zr

+ Zr

O–

Zr +

+ Zr

OH

Zr+ Zr

Zr

O– Zr + Zr

OH

Zr

Zr Zr

O–

Zr

Zr + Zr

O–

Zr

Zr

OH

Zr Zr

Zr

Zr

Zr

Zr Zr

Zr Zr

Zr

+ Zr

Zr

Zr Zr

Zr

O–

Zr

O–

+ +Zr OH Zr OH Zr+

Zr

OH

Zr

Zr

Zr

Zr Zr

Zr

OH = CIO4–

= Li+

= –O(CH2CH2O)7.671CH2CH2O–

6.11 3D structure of {Zr[(CH2CH2O)8.7]ρ/(LiClO4)z}n materials.



233

PEG O Zr O

PEG

PEG

O O

PEG

6.12 Tetrahedral coordination geometry around the crosslinking Zr atoms in {Zr[(CH2CH2O)8.7]ρ/(LiClO4)z}n materials. OH Al

Al Al

HO

Al Al

HO Al

O–

HO Al

Al

Al Al HO

H O

H

Al

O–H

Al HO

Al

Al Al

H O Al Al OH

OH

Al

OH O H + Al

H O Al Al Al

Al O H

Al

Al

OH H Al O

HO

Al

Al Al

OH Al

Al HO

OH

Al Al

H O

Al Al

= ClO4–

Al +

HO O

OH Al

H O

HO

Al

Al

Al + Al

HO

Al + O–H

Al

O

O

O H HO

Al

O–H + Al

O–H

H

Al

Al

OH

Al Al

Al

Al

Al+

Al HO

O H

Al Al

Al

HO

Al

OH

+ Al

HO

HO

OH

HO

= Li+

= –O(CH2CH2O)7.671CH2CH2O–

6.13 Structural model proposed for {Al[O(CH2CH2O)8.7]ρ/(LiClO4)z}n materials; reproduced with permission from Di Noto and Zago (2004). Copyright 2004, The Electrochemical Society. © Woodhead Publishing Limited, 2010

234


18 °C

15 10 5 0 0

50

100 150 200 r´ (Ω cm) × 103

250

1.0 – r˝ (Ω cm) × 106

20

80 °C

0

(a)

– r˝ (Ω cm) × 103

ω

140 120 100 80 60 40 20 0

ω

0.8

18°C 0.6 80°C

0.4

5

300

– r˝ (Ω cm) × 103

– r˝ (Ω cm) × 103

25

0.2

0 0.0

0.5

1.0 1.5 r´ (Ω cm) × 106

30

160 140 120 100 80 60 40 20 0

0.0 (b)

10 15 20 25 r´ (Ω cm) × 103

2.0

50 100 150 r´ (Ω cm) × 103

200

6.14 Nyquist plots for selected {Al[O(CH2CH2O)8.7]ρ/(LiClO4)z}n materials. (a) {Al[O(CH2CH2O)8.7]2.24/(LiClO4)1.06}n and (b) {Al[O(CH2CH2O)8.7]1.89/ (LiClO4)0.01}n. The measurements were carried out from 20 Hz to 1 MHz; the data near the origin of the plots on the left are magnified in the right panels. Reproduced with permission from Di Noto et al. (2004a). Copyright 2004, The Electrochemical Society.

PEG PEG

H

O

O PEG

AI PEG (a)

PEG

O

AI

O

O

O

PEG

O PEG

(b)

6.15 Coordination geometries around the crosslinking Al atoms in {Al[O(CH2CH2O)8.7]ρ/(LiClO4)z}n materials: (a) Four-coordinate Al sites (more abundant); (b) three-coordinate Al sites.



235

toward perchlorate anions. The best conductivity observed in these materials is 1.66 × 10−5 S cm−1 at 25 °C. 3D-HION-APPE proton conducting materials are presented as an important development to replace the ubiquitous and costly perfluorinated ionomers currently in use. These materials can be synthesized by the copolymerization of metal alkoxides and organic-substituted silicon alkoxides or suitable oligomers. In particular, two new siloxanic proton-conducting membranes are described (Di Noto and Vittadello, 2005; Di Noto et al., 2005): [I] {Si(CH3)3O[Si(CH3)HO]21.26[Si(CH3)((CH2)3SO3H)O]1.8 [Si(CH3)((CH2)3Si(CH3)2O)O]14Si(CH3)3}n; [II] {Si(CH3)3O[Si(CH3)HO]21.26[Si(CH3)((CH2)3SO3H)O]1.8 [Si(CH3)((CH2)3(Si(CH3)2O)w)O]v[Si(CH3)((CH2)3Si(CH3)2O–) O]14−vSi(CH3)3}n with w = 20.31. [I] is synthesized by a two step procedure: at first a precursor is prepared by hydrosilylation of allylsulphonylchloride and allyldimethylchlorosilane using polymethylhydrosiloxane (Fig. 6.16); afterwards, the hydrolysis of this precursor produced a cross-linked silicone polymer where pendant silicone chains are endowed with –SO3H acid groups (Fig. 6.17). [II] is prepared by the polycondensation of the same precursor used in the preparation of [I] with α,ω-dihydroxy(polydimethylsiloxane), obtaining another crosslinked silicone polymer (Fig. 6.18). The two proton-conducting polymer electrolytes [I] and [II] show a remarkable chemical stability and do not decompose until ca. 200 °C. [I] and [II] can be easily transformed into polymer films by sintering processes, and are characterized by Tg values typical of crosslinked silicone materials at −44 and −60 °C, respectively. [I] and [II] have ionic exchange capacities of 0.33 and 0.15 meq g−1, respectively. In [I] and [II] the charge transport takes place owing to an inter-cluster migration of protons through a vehicular mechanism mediated by rotational relaxations of silicone side groups bearing sulphonic acid groups. In this case, the clusters comprise water molecules solvating the sulphonic acid groups. The plot of ln σDC vs. 1/T for [I] and [II] is shown in Fig. 6.19. The σDC at 125 °C is ca. 1.9 × 10−3 S cm−1 and 1.8 × 10−4 S cm−1 for [I] and [II], respectively, at the highest hydration. These values classify these materials as good proton conductors. NMR proton diffusion data on a non-activated [II] membrane showed a two-region behaviour: below 60 °C the sample shows good diffusion properties but at higher temperatures water is quickly released from the sample. A drastic treatment with sulphuric acid reduces the sample into a gummy powder with very high diffusion coefficients all the way up to 110 °C. Taken together, these results are encouraging with respect to the


236

Polymer electrolytes SO3– Na+ POCI3

Na2SO3

Br

SO2CI

(a)

CH3 H3C Si (b)

O

CH3

CH3

Si O

Si

H

CH3

35.42

SO2CI

1/2 CH3

+ Si(CH3)2CI

CH3 1/2

RhCI[(C6H5)3P]3 in 1,2-dimethoxyethane

H3C Si CH3

CH3

CH3

CH3 O

Si H

O

Si x

CH2 H2C CH2 CIO2S

CH3

CH3 O

Si y

CH2 H2C CH2

O

Si z

CH3 (P)

CH3

H3C Si CH3 CI

6.16 Synthesis of the precursor necessary for the preparation of the siloxanic proton-conducting materials [I] and [II]. (a) Preparation of allylsulphonylchloride; (b) hydrosilylation of allylsulphonylchloride and allyldimethylchlorosilane using polymethylhydrosiloxane; reproduced from Di Noto and Vittadello (2005) by permission of Elsevier.

possibility of preparing membranes with performances comparable to those of Nafion. Promising membranes were obtained by doping Nafion with ‘core–shell’ [(TiO2)·(WO3)0.148] ceramic nanoparticles (Di Noto et al., 2010b). This research has been carried out in the framework of a series of efforts aimed at elucidating the effect of ceramic oxoclusters on the structure and the conductivity of hybrid nanocomposite systems (Di Noto et al., 2006, 2007, 2008, 2009, 2010a,b; Vittadello et al., 2008; Thayumanasundaram et al., 2010). The [(TiO2)·(WO3)0.148] nanofiller was obtained by grinding together a dimethylformamide (DMF) suspension of TiO2 and WO3 in the desired ratio (Fig. 6.20). Seven homogeneous membranes with the formula {[Nafion/ [(TiO2)·(WO3)0.148]} where 0 ≤ wt%TiO2 ≤ 15 were prepared by solventcasting. The membranes had a thickness lower than 350 µm, and were stable up to 170 °C. Vibrational spectroscopy investigations allowed us to distin-


Hybrid inorganic–organic polymer electrolytes CH3 n • CH3 Si

CH3

CH3 O

O

Si

O

x

y

CH2

O

Si

CH2

H

CH3

CH3

CH3

Si

CH2 SO2

237

Si

CH2 z CH2 CH2 CH3 Si CH3

CH3

CH3

CI

CI

Hydrolysis in 0.1 M H2SO4 at 60 °C for 12 h

CH3 CH3 Si

CH3

CH3

CH3 O

Si H

O

Si

O

Si

CH2 x

CH2 CH2 SO3H

CH3

CH3

y

O

CH2 CH2

Si z

CH3

CH3 n

CH2 CH3 Si

CH3

O

6.17 Synthesis of the siloxanic proton-conducting polymer electrolyte [I]; reproduced from Di Noto and Vittadello (2005) by permission of Elsevier.

guish hydrophobic and hydrophilic domains at the nano-scale. In particular, it was shown that the fluorocarbon backbone chains are present with two distinct helical conformations, 157 and 103 as outlined in Fig. 6.21; four different species of water domains were also detected. Both groups of vibrational modes are modulated by the amount of [(TiO2)·(WO3)0.148] present in the material. Modulated differential scanning calorimetry (MDSC) analysis revealed that the membranes undergo four thermal transitions at temperatures lower than 300 °C. As the temperature is increased, the transitions are assigned to the melting of small and imperfect fluorocarbon domains in Nafion (100 < T < 150 °C), to the decomposition of the –SO3H groups of Nafion (170 < T < 230 °C) and to the melting of different fluorocarbon domains of Nafion, either not stabilized (230 < T < 270 °C ) or stabilized by the [(TiO2)·(WO3)0.148] nanofiller (T ≈ 300 °C). The temperature of these transitions depends on the membrane composition, as the nanofiller interacts with the sulphonic acid groups of Nafion, improving their thermal stability.


238

Polymer electrolytes CH3 H2O

CI

CI

Si

CH3 CH3 CH3 Si CH3

CH3

CH3 O

Si H

O

O Si CH2 x

CH2 CH2 SO2

Si

Si z

CH3 + Zi• OH

CH3

CH3

Si H

Si

O x

CH2 CH2 CH2 SO3H

H 19.21

Hydrolysis in 0.1 M H2SO4 at 60 °C for 12 h

CH3

CH3

CH3 O

O

Si

CH3

CI

CH3

CH3

O

CH2 y CH2 CH2 CH3 Si

CI

CH3 Si

CH3

CH3

CH3

Si

O y

CH3

CH3 O

CH2 CH2 CH2

z–w

CH3 Si CH3 O

Si O CH2 CH2

Si CH3 CH3 w

CH2 CH3 Si O CH3

n

20.31

6.18 Synthesis of the siloxanic proton-conducting polymer electrolyte [II]; reproduced from Di Noto and Vittadello (2005) by permission of Elsevier.

Dynamic mechanical analysis (DMA) investigations revealed four distinct mechanical relaxation events. α and α′ modes are detected at T > 100 °C and were assigned to the long-range motions of both the backbone and the side chains facilitated by the weakening of electrostatic interactions within the ionic aggregates. β and β′ relaxation events, measured at 35 and 65 °C, were attributed to the 136 → 157 and to the order–disorder conformational transitions occurring in hydrophobic polytetrafluoroethylene (PTFE)-like domains of Nafion, respectively. The electrical response of the membranes was investigated by broadband dielectric spectroscopy (BDS). Figure 6.22 reports the plot of ln σDC vs. 1/T for {Nafion/[(TiO2)·(WO3)0.148]} nanocomposite membranes with 0 ≤ wt%TiO2 ≤ 15. It is highlighted that the nanocomposite materials show a σDC higher by 25–60% with respect to pristine Nafion at 5 < T < 155 °C. σDC reaches a maximum value equal to 5.9 × 10−2 S cm−1 at T = 115 °C for the membrane doped with 5 wt% of TiO2. In the same conditions, the conductivity of pristine Nafion is equal to 3.3 × 10−2 S cm−1. In addition, it is observed that nanocomposite membranes are characterized by a wider Stability Range of Conductivity (SRC) with respect to pristine Nafion. SRC is defined as the temperature range where equation 6.1 holds true: ∂σ DC ≤0 ∂ (1 T )

[6.1]



239

10–2

log (σ / S cm–1)

10–3

10–4

10–5

Nafion 105, 29.78 –wt% of H2O [I] 33.12 –wt% of H2O [II] 27.76 –wt% of H2O fit

10–6 2.6

2.8

3.0

3.2

3.4

3.6

3.8

1/T (K–1) x 10–3

6.19 Temperature dependence of direct current conductivity, σDC, of membranes [I], [II] and Nafion 105 at maximum hydration; reproduced from Di Noto et al. (2005) by permission of Elsevier.

In detail, 0 < SRC < 105 °C for pristine Nafion and 0 < SRC < 145 °C for the {[Nafion/[(TiO2)·(WO3)0.148]} material with 15 wt% of TiO2. A detailed study of the conductivity mechanism has been carried out for other similar materials (Di Noto et al., 2006), and it is suggested that proton migration along interconnecting channels and polar hydrophilic clusters takes place owing to proton exchange processes modulated by the amount and types of interstitial water domains, the density of hybrid nanofiller–(HSO3)–crosslinks, and the segmental motions of the fluorocarbon backbone of Nafion polymer. In turn, this leads to the hypothesis that the exchange of protons between different fluctuating water species domains occurs through hopping processes, and that this mechanism is strongly regulated by the molecular relaxation events in the materials.

6.3.2 Class II: zeolitic inorganic–organic polymer electrolytes (Z-IOPE) Z-IOPEs are hybrid inorganic–organic polymer electrolytes characterized by a structure that can be described as a network of inorganic clusters, © Woodhead Publishing Limited, 2010

240

Polymer electrolytes 1.5 g (≈70%)

0.650 g (≈30%)

DMF

‘Hard’ TiO2

‘Soft’ WO3

Planetary ball mill

5h 500 rpm

with grinding jars of tungsten carbide 0.148 = Dispersion A Mohs Mohs hardness hardness 3 6

molWO3 molTiO2

[(TiO2)(WO3)0.148] nano-clusters Nanometric ‘core–shell’ oxoclusters of a ‘hard’ TiO2 core homogeneously covered by a layer of ‘soft’ WO3.

Ball WO3 milling TiO2 oxocluster oxocluster

The size of ‘core–shell’ oxoclusters ranges from 20 to 60 nm.

6.20 Solid state synthesis of [(TiO2)·(WO3)0.148] ‘core–shell’ nanofiller; reproduced from Di Noto et al. (2010b) by permission of Elsevier.

(a)

(b)

6.21 Axial view of Nafion model compound with different conformations of the backbone: (a) 157 helix; (b) 103 helix.

formed by the aggregation of two or more inorganic coordination complexes either positively or negatively charged, bridged by organic macromolecules. The general structure of Z-IOPEs is shown in Fig. 6.23. Z-IOPEs are prepared starting from two different solutions. The first contains a hard


Hybrid inorganic–organic polymer electrolytes –1.2

241

(II)

–1.3

log (σDC / S cm–1)

–1.4 –1.5 –1.6 –1.7

(I)

wt%TiO2 0 3 5 9 11 13 15

–1.8 –1.9 –2.0 –2.1 –2.2

2.3 2.4 2.5 2.6 2.7 2.8 2.9 3.0 3.1 3.2 3.3 3.4 3.5 3.6 × 10–3 1/T (K–1)

6.22 Dependence of ln σDC on 1/T for {Nafion/[(TiO2)·(WO3)0.148]} nanocomposite membranes with 0 ≤ wt%TiO2 ≤ 15. I and II indicate the conductivity regions. Dotted lines show VTF fitting curves reproduced from Di Noto et al. (2010b) by permission of Elsevier.

n–

n–

n–

n–

= Inorganic cluster n–

n–

n–

= Organic macromolecule

n–

6.23 General structure of Z-IOPEs.

transition metal (Fe, Co) cyanometallate and the organic macromolecule; the second includes a soft metal (Pd, Sn) chloride and the same organic macromolecule. After mixing the two solutions a sol → gel → plastic transition occurs, resulting in the hybrid network. This general synthesis protocol was used to prepare the following Z-IOPEs: [I] [Fe0.922Pd0.663(CN)5.407Cl1.511(C2nH4n+2On+1)K2.826] (Di Noto, 1997); [II] [Fe0.153Pd0.277(CN)0.876Cl0.940(C2nH4n+2On+1)K0.803] (Di Noto, 1997);


242

Polymer electrolytes [III] [Fe0.048Pd0.077(CN)0.103Cl0.44(C2nH4n+2On+1)K0.245] (Di Noto, 1997); [IV] [Co0.110Pd0.273(CN)0.591Cl1.295(C2nH4n+2On+1)K1.01] (Di Noto, 2000); [V] [Fe0.127Sn0.164(CN)0.617Cl0.616(C2nH4n+2On+1)K0.323] (Di Noto et al., 2000); [VI] [Fe0.088Sn0.187(CH3)0.374(CN)0.578Cl0.444(C2nH4n+2On+1)K0.472] (Di Noto et al., 2001, 2002b); [VII] [Fe0.082Pd0.149(CN)0.492Clv(C2nH4n+2On+1)Li0.523] (Di Noto et al., 2003b; Vittadello et al., 2003).

[I], [II], [III] and [IV] were obtained from a first solution composed of K3Fe(CN)6 (or K3Co(CN)6 in the case of [IV]), H2O and PEG600, and a second solution composed of K2PdCl4, H2O and PEG600. By modulating the stoichiometry of the reactants different networks, [I], [II] and [III] were produced. These Z-IOPEs are thermally stable up to ca. 210 °C. The mechanism of the sol → gel → plastic transition was clarified by rheological studies, yielding results exemplified in Fig. 6.24. After mixing the two starting solutions, a direct substitution of the Cl− in PdCl42− by H2O or by the nitrogen atoms of the cyanometallate ligands is likely to predominate, giving rise to a viscoelastic solution according to the reactions shown in Fig. 6.25. Afterwards, H2O is displaced by the hydroxyls groups of PEG600 and the nitrogen atoms of the cyanometallate ligands, giving rise to the transition viscoelastic solution → hard gel, which occurs according to the reactions shown in Fig. 6.26. This reaction mechanism is strongly supported by these observations: (a) if the starting solutions include PEG600 in the absence of water, no gel is formed as the solutions are mixed; (b) the volume of the gel decreases vs. time; at the same time, the gel releases a transparent liquid composed mainly of H2O and traces of PEG600, K, Cl and Pd; (c) the use of water/methanol or water/ethanol solvent mixtures leads to the formation of unstable materials. As a consequence, it is reasonable to assume that the gel-plastic transition occurs mainly owing to a slow, indirect substitution reaction. The conductivity of Z-IOPE [IV] at room temperature is ~3 × 10−5 S cm−1; since the material is very stable and easy to prepare, it appears to be a promising candidate for the development of a new class of electric energy storage systems. The Z-IOPEs [I], [II] and [III], whose structures are reported in Fig. 6.27, show conductivities higher than electrolytic systems based on PEG, poly(ethylene oxide) (PEO) and poly(propylene oxide) (PPO) doped with inorganic salts such as LiClO4, LiCF3SO3, LiSCN, etc. (MacCallum and Vincent, 1989, p285; Scrosati, 1993, p182). In particular, [I] exhibits a conductivity of ≈1 × 10−3 S cm−1 at T = 290 K, which is approximately two orders of magnitude higher than that observed for the electrolyte polymers



243

(a) 30.3 °C 40.1 °C 49.3 °C 59.7 °C 70.1 °C

G´ (Pa)

1000

100

10

1

0

10

20 Time ×103 (s)

30

(b) 30.3 °C 40.1 °C 49.3 °C 59.7 °C 70.1 °C

G˝ (Pa)

1.5

1.0

0.5

0

10

20 Time ×103 (s)

30

6.24 Typical plots of the elastic G′ (a) and viscous modulus G″ (b) versus time for the sol–gel transition occurring during the preparation of a Z-IOPE at various temperatures; reproduced from Di Noto (2000) by permission of the American Chemical Society.

currently used in electric energy storage devices (Scrosati, 1993, p182; Guyomard and Tarascon, 1994). The temperature dependence of the conductivities was of the VTF type. VTF fitting parameters help us to understand that: (a) [I], [II], and [III] conduct ionically; (b) the increase in the configurational entropy of these inorganic–organic networks is an important factor for the mobility of free ions; and (c) as the concentration of PEG600 in these networks increases, their conductivity decreases. Z-IOPEs [V] and [VI] were obtained applying the general synthesis protocol mentioned above, using SnCl4 and SnCl2(CH3)2 as the soft metal reagent for [V] and [VI], respectively.


244

Polymer electrolytes CI

OH–R´–OH

CI

OH–R´–OH

CI

N

M

M

CI

N

–CI– +HO–R´–OH CI

CI

–CI–

(–)

CI

M

M

CI

OH–R´–OH

N

N N CI

CI

–CI– (2–)

M = Pd X = Fe or Co

M

HO–R´–OH = PEG600 CI

CI

N = [X(CN)6]3–

–CI– +H2O CI

(–)

–H2O CI +HO–R´–OH

OH–R´–OH

CI

–CI +HO–R´–OH CI

OH–R´–OH

–H2O +HO–R´–OH

M

N

CI

N –CI–

OH2

CI

CI

OH2

CI

–H2O +HO–R´–OH

N M N –H2O

–H2O

OH2

CI

OH––R´–OH

CI

OH2 M

M CI

CI

M

M

CI

N

CI

–CI– +H2O OH–R´–OH

CI

N

(–)

OH2

–

CI

N (–)

CI M

M CI

N

CI

–CI– +HO–R´–OH

CI

(–)

CI

N N

6.25 Reactions giving rise to a viscoelastic solution in the preparation of a Z-IOPE.


Hybrid inorganic–organic polymer electrolytes CI

N

CI

N +

Pd CI

N

CI

OH–R–OH

N Pd

Pd

N

CI

N

N

N

OH–R–OH Pd

CI

N

HO–––R–HO –2n(CI–)

Pd

CI

N Pd

OH–R–OH

OH–R–OH

N

CI –2CI–

N

N

HO–––R–HO

OH–R–OH

N

Pd

OH–R–OH

245

OH–––R–OH Pd

N

N

OH–––R–OH n

HO–R–OH PEG600 N

N

M = Fe, Co [M(CN)6]3–

N

N

6.26 Typical reactions occurring during the gelification process in the preparation of Z-IOPEs.

[I] NC

[II] 7–

CN CN Fe

NC CN

H

CN

O

Pd CI

NC

CI

R

O H

NC NC NC

CN

CN

•7n K+

CN NC

CN

CN

H

O

O Pd

CI

CN

CI

CI

•2n K+ 4 PEG

Fe NC

CN Fe

R

Pd

Fe

Pd

H

O

O

R

CN

2–

H

H

R

CN CN n

CN

CN

n

[III] –

H

H O

O R

R O

O

O Pd

Pd

H

R

H R

R

CN CN

NC Fe NC

CN

CN

H •n K+ 20 PEG

O H n

6.27 Chemical structure of the Z-IOPEs [I], [II] and [III]. Reproduced from Di Noto (1997) by permission of the Materials Research Society.


246

Polymer electrolytes 5–

N Me OH

C C

Me C

N Sn

HO

Fe

Sn

N C

Me

C

N

OH

N +

* 5K * [PEG600*(KCl)0.254]

C

HO Me

N Me CI

Sn

OH

Me

n

OH

(a)

= (CH2CH2O)12·211 CH2CH2 Temperature (K) 310 320 330 340

300

350 1.5

–13

(I)

In (τ / s)

–9.0

–9.5

(II)

–14

τ1

–15

VTF fit

–16

1.3

p

1.2

–17 In σ (T)

1.1

–18

1.0

VTF fit

–10.0

1.4

d

In (σ (T) / S cm–1)

–8.5

–19 0.9 2.9

(b)

3.0

3.1

3.2

1/T (K–1) ×103

3.3

2.9

3.0

3.1

3.2

3.3

1/T (K–1) ×103

6.28 (a) Structural model proposed for the [VI] Z-IOPE; reproduced from Di Noto et al. (2001) by permission of Elsevier; (b) selected electric properties of [VI]: I) ln[σ(T)] vs. reciprocal absolute temperature; II) ln(τ1) and p as a function of reciprocal absolute temperature. Copyright Wiley-VCH Verlag GmbH & Co. Reproduced with permission from Di Noto et al. (2002b).

The structural hypothesis for [VI] Z-IOPE is shown in Fig. 6.28(a). The TGT conformation of polyether chains is maintained in the final materials. Both [V] and [VI] conduct ionically by a mechanism mainly regulated by the segmental motion of the host material; however, charge migration by ion hopping between equivalent coordination sites in the host network is not to be completely excluded. The conductivities at 25 °C of [V] and [VI] are ca. 3.7 × 10−5 S cm−1 and 4.77 × 10−5 S cm−1, respectively. The electrical properties of [VI] were studied in detail by fitting the data with a universal power law function 6.2; the main results are shown in Fig. 6.28(b):


Hybrid inorganic–organic polymer electrolytes σ ′(ω ) = σ ′( 0 ) ⎡⎣1 + (ωτ1 ) ⎤⎦ p

247 [6.2]

τ1 is a time related to the initial site relaxation rate τ2, while p = τ2/τ* can be expressed in terms of the ratio of the initial back-hopping rate (1/τ*) and the initial site-relaxation rate (1/τ2) in accordance with the jump-relaxation model (see Sections 6.4 and 6.5). [VII] is the first lithium-based Z-IOPE. This material is closely related to its potassium analogues based on K2PdCl4, K3[Fe(CN)6] and PEG600 (Di Noto, 1997). As in the case of the parent potassium Z-IOPEs, the synthesis protocol required the combination of two separate solutions (A and B). Solution A consisted of Li3Fe(CN)6 dissolved in water and PEG600; solution B was obtained by dissolving Li2PdCl4 in water and adding PEG600. The change of cation from potassium to lithium cannot be a simple translation of the chemistry of potassium precursors. Indeed, it is well known that lithium salts are often deliquescent materials and exhibit a high solubility even in organic solvents (Cotton et al., 1999, p92). Therefore, a different behaviour of the reaction mixture used to obtain the lithium Z-IOPE with respect to the potassium Z-IOPE material was expected. First, it was necessary to optimize the preparation of Li2PdCl4 and Li3Fe(CN)6 dry precursors starting from the commercially available compounds. The palladium complex was obtained by reacting LiCl and PdCl2 stoichiometrically in a single step reaction as discussed by Parker and O’Fee (1983): PdCl2 + 2LiCl → Li2PdCl4 The synthesis of the lithium ferricyanide was carried out via a two-step reaction, as described by Chadwick et al. (1985): K3Fe(CN)6 + 3AgNO3 → Ag3Fe(CN)6↓ + 3KNO3 Ag3Fe(CN)6 + 3LiCl → Li3Fe(CN)6 + 3AgCl↓ A possible pathway for the formation of the final product involves the reaction of the two [PdCl4]2− derivatives [PdCl2PEG2] and [PdCl2(H2O)(Fe(CN)6)]3− with each other; their condensation can take place through chloride elimination to produce inorganic–organic clusters that can cross-link, giving rise to a three-dimensional network with the formula [FexPdy(CN)zClv(C2nH4n−2On−1) Lil]. At first, the result of this complexation reaction is the formation of a gel. When the forced release of water under vacuum is completed, a zeolitic polymer electrolyte complex is obtained. The actual final product was a solidplastic material, whose structural model is proposed in Fig. 6.29(a). Elemental analyses and spectroscopic investigations determined that this lithium Z-IOPE is a mixed inorganic–organic network where Fe and Pd are linked by cyanide bridges and Pd atoms are linked by PEG600 bridges. PEG600 assumes the TGT helix conformation. The electrical conduction in this material is mainly due to the displacement of Li+. 1H and 7Li NMR linewidth, spin-lattice


248

Polymer electrolytes 2– N R

C

O R

C

N O

Pd

CI

C

H

Fe

N C

O

C

N

2Li+ · [11 PEG600 · 5LiCl]

H

C CI

H

CI

Pd N

R

N

n

(a) –2 I

8

–3 II

6

–4

–5

4 fion s DC

–6

log (sDC / S m–1)

log (fR/Hz)

fseg

2 –7 0 –8

3.0

(b)

3.2

3.4

3.6

3.8

4.0

1/T (K–1) × 10–3

6.29 (a) Structural model proposed for the Z-IOPE [VII]; (b) dependence on temperature of fion, fseg and σDC for Z-IOPE [VII]; reproduced from Vittadello et al. (2003) by permission of Elsevier. fion, fseg and σDC are the frequency of the ion-mode relaxation, the frequency of the relaxation event attributed to the segmental motion and the conductivity of the material, respectively.

relaxation and pulsed field gradient diffusion measurements suggest that the lithium ion transport is correlated with the mobility of the polymer, as in the case of ‘conventional’ polymer electrolytes. The electrical spectra for frequencies higher than 15 kHz evidenced the presence of relaxation events associated to local ion motion dynamics and long-range diffusion. Results are shown



249

in Fig. 6.29(b). These two phenomena were interpreted in terms of: (a) ion hopping processes; (b) site relaxations; and (c) host medium reorganization processes. Taken together, these studies demonstrated that at temperatures higher than 35 °C, this Z-IOPE conducts ionically by charge transfer mechanisms mainly regulated by ion hopping between equivalent polyether coordination sites followed by correlated host medium reorganizations. Finally, a conductivity of 5.3 × 10−5 S cm−1 at 35 °C classifies this hybrid inorganic–organic network as a good lithium ion conductor.

6.3.3 Class III: hybrid gel electrolytes (HGEs) The HGEs are prepared through a sol → gel process, starting from a combination of hard and soft metal precursors; a low molecular weight organic ligand is used instead of a macromolecule such as PEG600. Two products obtained with glycerol as the molecular ligand are described as examples of this class of materials. Glycerol is a very well-known glass former (Angell, 1997) that has been thoroughly investigated in BDS (Lunkenheimer et al., 2000) and in thermodynamic studies (Ito et al., 1999; Martinez and Angell, 2001). On the basis of the ‘strong–fragile’ classification scheme (Angell, 1997; Ito et al., 1999; Martinez and Angell, 2001), glycerol is considered to be a strong glass former and therefore is characterized by a molecular relaxation that increases with an almost VTF-like (simple activated) profile (Lunkenheimer et al., 2000). The analysis of the frequency-dependent dielectric loss of glycerol reveals that it is a ‘Type A’ glass former (Angell, 1997; Kudlik et al., 1999), since it exhibits a well-defined α-relaxation followed by an excess wing, without any slow β-relaxation as in the case of ‘type B’ glass formers (Kudlik et al., 1999). Therefore, glycerol is a suitable organic molecule to investigate the effect of inorganic clusters in an electrolytic gel structure. Glycerol also represents a promising precursor for the preparation of gel lithium ion conductors able to operate at low temperatures owing to the absence of thermal transitions over a wide temperature range (Angell, 1997; Kudlik et al., 1999). Two systems are presented: [I] [FexPdy(CN)zClv(C3H8O3)Lil] (Di Noto et al., 2004b); [II] [FexSny(CH3)2y(CN)zClv(C3H8O3)Lil] (Di Noto et al., 2004b). The synthesis of these new HGE electrolytic complexes required the combination of two solutions (A and B). [I] was prepared by mixing solutions A and B under an argon atmosphere. Solutions A and B consisted of glycerol dissolving Li2PdCl4 and Li3Fe(CN)6, respectively. Solutions A and B were mixed, and a gel was obtained quickly. The latter expelled a small amount of glycerol over 1 week. The resulting material was filtered under an argon atmosphere and left under vacuum for 10 days. The final product


250

Polymer electrolytes a Li3Fe(CN)6 + b glycerol

a K + b glycerol with K = Li2PdCI4; Sn(CH3)2CI2 A

B

+

sol–gel transition

[FexMy(CN)zCIv(CH2OHCHOHCH2OH)Li5]gel M = Pd; Sn(CH3)2

6.30 Synthesis of hybrid inorganic–organic gels based on Li2PdCl4 (I) or Sn(CH3)2Cl2 (II), glycerol and Li3Fe(CN)6; reproduced from Di Noto et al. (2004b) by permission of the American Chemical Society.

(a)

(b)

6.31 Morphology of HGEs determined by scanning electron microscopy (SEM): (a) HGE [I]; (b) HGE [II]; reproduced from Di Noto et al. (2004b) by permission of the American Chemical Society.

was paste-like and dark-brown with a greenish cast. [II] was synthesized by mixing A and B solutions in open air. Solutions A and B contain glycerol dissolving (CH3)2SnCl2 and Li3Fe(CN)6, respectively. As A and B were stirred after the mixing, a yellow-orange gel was immediately obtained. The material was filtered and left under vacuum for 10 days, yielding a yelloworange gel with a slightly green cast. A summary of the synthesis procedure is shown in Fig. 6.30. It is expected that the sol → gel transition occurring in both HGE systems takes place with the same mechanism outlined for conventional Z-IOPEs (Di Noto et al., 2001, 2002b, 2003b). The morphology of [I] and [II] is reported in Fig. 6.31. © Woodhead Publishing Limited, 2010


251

Spectroscopic investigations concluded that these materials consist of mixed inorganic–organic networks containing cavities flooded with glycerol glass-former lithium electrolyte materials. The hybrid inorganic–organic networks in [I] are constituted of Pd atoms bound to Fe by CN bridges, and Pd atoms linked together by glycerol bridges. Likewise, the networks in [II] are composed of Sn atoms bound to Fe through CN bridges and Sn atoms linked together by glycerol. The structural models for [I] and [II] HGEs are shown in Fig. 6.32. The glycerol molecules in lithium electrolyte glass formers

(–) N C

CN

CN Fe

O

H

HO

HO

H

OH

CN

CN

O

Pd

∙ Li+ OH

HO OH

N C

OH

C N

HO OH

OH OH

H

NC

OH

O H

HO

Fe

O

O

H

CN

O

CN

CI

Pd

Pd

OH

OH

CN 244 ∙ HOCH2CHOHCH2OH 12∙LiCI

C N

(a) HO

HO OH OH

OH

n

HOCH2CHOHCH2OH

OH =

CH3

HO Sn

N C

CH3 HO HO

Sn

OH

OH

CN

HO

CN

CN

H3C

CN CH3

Sn

CH3 CH3

CN Fe

H3C

CN

CN

OH

CN H O

O H

OH CN

OH

Sn C N

(b)

CI

CI

O

5 LiCI 272 ∙ HOCH2CHOHCH2OH

–CH3

OH OH

n

OH

H O HO

O Sn

H O

HO

CH3

OH

CI

Fe

N C

HO

HO

OH

OH

O H

OH

HO

OH

HO HO

OH HO

CI

O H

H3C

HO

OH =

HOCH2CHOHCH2OH

6.32 (a) Structural model proposed for HGE [I]; (b) structural model proposed for HGE [II] reproduced from Di Noto et al. (2004b) by permission of the American Chemical Society.


252


αα conformer

αγ conformer

γγ conformer

6.33 Typical conformers of glycerol molecules.

(a)

HGE[I]

HGE[II]

100

50 80 40

Permittivity

20

Permittivity

80 °C 80 °C 30

60 40 20

10 –60 °C

0

0 100

102

104

106

108

100

1010

–60 °C 102

Frequency (Hz)

104

106

108

1010

Frequency (Hz)

(b) 20

40 –60 °C

–60 °C 30

80 °C

Permittivity

Permittivity

15

10

5

0 100

80 °C

20

10

0 102

104

106

Frequency (Hz)

108

1010

100

102

104

106

108

Frequency (Hz)

6.34 Imaginary component of the dielectric spectra of HGEs: (a) dielectric loss spectra and (b) difference dielectric loss spectra obtained by subtracting the spectral contribution of electrode polarization and σDC from dielectric spectra shown in (a). Measurements were carried out from −60 to +80 °C in 10 °C increments; reproduced from Di Noto et al. (2004b) by permission of the American Chemical Society.


1010


253 10–5

0.16

10–6

0.14

10–7 I II

10–8 10–9

0.10

10–10

0.08

D (cm2 s–1)

λ (nm)

0.12

10–11 0.06

I II

10–12

0.04

10–13

0.02

10–14 220

240

260

280

300

320

340

360

Temperature (K)

6.35 Temperature dependence of migration step length λ and diffusion coefficient D in HGE[I] and HGE[II] reproduced from Di Noto et al. (2004b) by permission of the American Chemical Society.

flooding the HGEs’ bulk cavities are predominantly a mixture of αα, αγ and γγ conformers, depicted in Fig. 6.33. Variable-temperature 1H and 7Li NMR linewidth spin-lattice relaxation and pulsed field gradient diffusion measurements suggested that lithium ion transport in the HGEs differs from that in common polyether-based PEs. Indeed, charge transport in HGEs occurs via short-range motions within glycerol solvation clusters. BDS investigations (Fig. 6.34) were in accordance with NMR results and revealed that the conductivity of the two HGEs is modulated by the α and the slow β mode relaxations of glycerol glass-forming molecules. Furthermore, accurate analyses of the temperature dependence of the mode relaxation parameters and of σDC values indicated that ion diffusion in electrolytes flooding the cavities of the HGEs takes place over distances in the order of magnitude of glycerol intermolecular hydrogen bond lengths, as shown in Fig. 6.35. These results permit the conclusion that the two HGEs probably conduct ionically through a charge migration mechanism, which involves short-range exchange of lithium ions between glycerol coordination cages, as shown in Fig. 6.36. The lithium coordination cages are generated owing to intermolecular hydrogen bonds between glycerol molecules forming the solvation layer of the hybrid inorganic–organic network backbone. Finally, [I] and [II] HGEs exhibit room temperature conductivities equal to 5.0 ×


254

Polymer electrolytes GC1

GC2

GC1Li+ + GC2 ↔ GC1 + GC2Li+

6.36 Model of the exchange of lithium ions between glycerol coordination cages.

10−5 S cm−1 and 8.5 × 10−5 S cm−1, respectively. Thus, HGEs can be classified as good lithium electrolyte materials.

6.4

Methods

6.4.1 Structural characterization of hybrid inorganic–organic polymer electrolytes The structural characterization of hybrid inorganic–organic polymer electrolytes and the study of their conductivity requires a holistic approach, integrating data obtained by several techniques. A useful guiding principle for a fruitful investigation of these materials is to build up knowledge of corresponding conventional polymer electrolytes seen as the components of a modular structure. Typically, the structural study of the materials starts from the chemical analysis, carried out by elemental microanalysis and inductively coupled plasma atomic emission spectroscopy (ICP-AES). On the basis of the compositional information it is possible to calculate the ratio between the organic molecules (e.g. PEG and glycerol) and the inorganic component. The comparison between the expected and the experimental composition of the investigated material is helpful to understand the mechanism of the chemical reactions involved in the synthesis. A viscosimetric study in terms of G′(ω) and G″(ω) is useful to monitor the gel formation in Z-IOPEs and HGEs, as shown in Fig. 6.37. Scanning electron microscopy (SEM) and energy-dispersive X-ray spectroscopy (EDX) are used to confirm the morphological and compositional homogeneity of the synthesized materials. Far-infrared (FIR) spectroscopy provides


1

100

0

80

–1

60

–2

40

δ (°)

log (G′/Pa); log (G″/Pa)

(a)

log (G′/Pa) log (G″/Pa) δ (°)

–3

20 0

–4 –1.0

–0.5

0.0

0.5

1.0

1.5

log (ω / rad s–1) 30

3

25

2

20 1 15 0

log (G′/Pa) log (G″/Pa) δ (°)

–1

δ (°)


(b)

10 5 0

–2 –1.0

–0.5

0.0 0.5 log (ω / rad s–1)

1.0

1.5


(c) 3.0 2.5 2.0 log (G′/Pa) log (G″/Pa)

1.5 1.0 0.5 0.0 –1

0

1 log (τ / Pa)

2

3

6.37 G′, G″ and δ (deg) vs. frequency plots measured after a reaction time of 35 min (a) or 6 h (b). (c) G′, G″ as functions of the logarithm of the stress (τ/Pa) measured after a reaction time of 6 h. The time when the solutions A and B leading to the Z-IOPE were mixed was taken as t = 0; reproduced from Di Noto (2000) by permission of the American Chemical Society.


256


a qualitative indication about the metal–ligand interactions in the system. The compositional data and the metal–ligand vibrations allow a preliminary structural hypothesis to be made on the inorganic clusters or on the metal/ semi-metal atoms distributed in the material and the organic molecules bridging the inorganic components. Detailed structural information is obtained from medium infrared (MIR) and Raman spectra. Vibrational modes are identified by a correlative analysis based on the literature. The full assignment of the spectra is carried out by identifying the spectral features of the organic component and those of the inorganic component. It is particularly useful to determine the conformation of the macromolecules (e.g. PEG) which are included in the structure of the hybrid inorganic– organic polymer electroytes. An example of FT-IR absorption spectra of a Z-IOPE is shown in Fig. 6.38. In general, it is observed that the macromo(a)

1.2

Absorbance

1.0 0.8 0.6 0.4 0.2 0.0 500

1000

1500

2000

2500

3000

3500

4000

Wavenumber (cm–1)

(b)

Absorbance

1.2 1.1 1.0 0.9 0.8 0.7

100

200

300

400

500

600

Wavenumber (cm–1)

6.38 FT-IR absorption spectra of the Z-IOPE [VII]: (a) MIR; (b) FIR. Reproduced from Di Noto et al. (2003b) by permission of Elsevier.


1000

Residuals

Residuals


0

–1000 × 10–6

0.004

0.000

–0.004

60 × 10–3

HGE [I]

Absorbance (a.u.)

Absorbance (a.u.)

30 ×10 –3

20

10 Current

257

HGE [II]

40

20

pesk

0

0 1900

2000

2100

2200

Wavenumber (cm–1)

2300

2000

2100

2200

Wavenumber (cm–1)

6.39 Decomposition by Gaussian functions of the MIR FT-IR spectral range from 1900 to 2500 cm−1 of HGE [I] and HGE [II]; reproduced from Di Noto et al. (2004b) by permission of the American Chemical Society.

lecular conformation of PEG in conventional PEs such as PEG/(MX)n and in the hybrids overviewed in this chapter is the same. PEG is characterized by a TGT conformation that can be interpreted using the symmetry classes of the D(4π/7) group (Di Noto et al., 1996b). PTFE fluorocarbon chains of Nafion show either a 157 or a 103 helical conformation (Fig. 6.21). The relative contribution of the relevant bands in the MIR spectra is determined by Gaussian decomposition of the appropriate spectral regions (Di Noto et al., 2006, 2007, 2008, 2010b; Vittadello et al., 2008). In the case of the Z-IOPE and HGE materials the structural hypothesis is refined using a semi-quantitative vibrational analysis carried out on the stretching CN and the stretching OH vibrations. The distinctions between bridging and terminal cyanide groups (CNb and CNt, respectively) allows the determination of the structural connectivity of the inorganic clusters, as shown in Fig. 6.39 and Table 6.1. The distinction between bridging hydroxyl groups (OHb) and hydroxyl groups involved in hydrogen bonding cages with other hydroxyl groups (OHHy) allows us to determine the proportion of organic moieties directly linked to the inorganic clusters. In the case of 3D-HION-APE materials and Nafion-based composites, the MIR and Raman spectra are used not only to confirm the structural hypotheses but also to determine ion–ion and ion–polymer interactions by using difference spectroscopy techniques; an © Woodhead Publishing Limited, 2010

258


Table 6.1 Band parameters of the ν(CN) FT-IR vibrational modes for HGE [I] and [II]. Reproduced from Di Noto et al. (2004b) by permission of the American Chemical Society HGE

Band

[I]

ν(–CN)b ν(–CN)t

[II]

ν(–CN)b ν(–CN)t

νi (cm−1)

{ { { {

2218 2185 2158 2126 2063 2025 2144 2103 2073 2038

AiCN a

fwhmia

0.45 ± 0.01 0.21 ± 0.01 0.05 ± 0.01 0.92 ± 0.02 1.97 ± 0.03 0.09 ± 0.01 0.50 ± 0.02 1.02 ± 0.70 2.3 ± 0.3 0.32 ± 0.07

30.4 28.1 17.2 46.4 63.6 24.3 23.3 33.4 34.7 29.6

± ± ± ± ± ± ± ± ± ±

0.5 1.7 2.0 0.5 1.7 2.0 0.7 4.0 3 3

RiCN (%)b

RiCN c,d

RteoCN

12.19 5.69 1.35 24.93 53.39 2.44 12.08 24.63 55.56 7.73

RbCN = 0.19

≈1/6

RtCN = 0.81

≈5/6

RbCN = 0.37

≈2/6

RtCN = 0.63

≈4/6

a

Ai and fwhmi are the band area and the full-width at half-maximum of the peak centred at νi, respectively. b RiCN (%) = (AiCN × 100/ΣAiCN). c RbCN = (ΣAib/ΣAiCN); Aib is the band area of bridging CN groups. d RtCN = (ΣAit/ΣAiCN); Ait is the band area of the terminal CN groups.

example is shown in Fig. 6.40 (Di Noto et al., 2003a). The application of these methods allows the salt–polymer interactions to be determined as the concentration of salt is varied. Further structural indications are obtained by thermal analysis (thermogravimetric (TG) and MDSC). The thermal stability is estimated, the Tg and the phase transitions are measured. The mechanical relaxations of the structure can be measured by dynamic mechanical analysis (DMA). Solid state NMR is particularly useful to determine the structural characteristics of siloxanic systems using 29Si and to probe the environment of Li ions using 7Li. The 1H and 13C spectra are useful to determine the correlation between the Li ions and the organic host on the basis of the spin-lattice relaxation time T1 and the lineshape.

6.40 (a) Difference FT-Raman spectra of {Zr[(CH2CH2O)8.7]ρ/(LiClO4)z}n complexes with 1.8 ≤ ρ ≤ 1.99 and 0 ≤ z ≤ 0.90. The difference spectra were determined by subtracting the spectrum of the pristine hybrid inorganic organic network (cLi = 0) from the spectra of the {Zr[(CH2CH2O)8.7]ρ/(LiClO4)z}n complexes. Dotted line shows the spectrum of pure LiClO4 salt; (b) decomposition by Gaussian functions of difference FT-Raman spectrum of the {Zr[(CH2CH2O)8.7]1.99/(LiClO4)0.90} n complex. The inset shows the dependence of ip% on the lithium concentration, cLi1/2 (mol kg−1)1/2. ip% is the percentage of the ClO4− anions involved in the formation of ion pairs. ip% = A940 × 100/(A940 / A930), where A940 and A930 are the band areas of the peaks at 940 and 930 cm−1, respectively. Reproduced from Di Noto et al. (2003a) by permission from Elsevier.


Intensity (a.u.)


cLi mol kg–1 0.92104 0.41700 0.21758 0.10706 0.05807 0.03055 0.01484

LiCIO4

200

259

400

600

(a)

1200 800 1000 Wavenumber (cm–1)

1.0

1400

1600

25

ip (%)

20

Intensity (a.u.)

0.8

15 10 5

0.6

0 0.2

0.4

0.4 0.6 0.8 cLi½ (mol½ kg–½)

1.0

0.2

0.0 200 (b)

400

600

800 1000 1200 Wavenumber (cm–1)


1400

1600

260


6.4.2 Conductivity studies on hybrid inorganic–organic polymer electrolytes Conductivity is studied mainly by BDS. This technique is better known as impedance spectroscopy at frequencies lower than 1 MHz. Impedance spectroscopy yields Nyquist plots (−Z″(ω) vs. Z′(ω)), which can be analysed in terms of an equivalent circuit (EC). Figure 6.41 shows a typical example of these plots. Typically, the Nyquist plot of a polymer electrolyte at a given temperature and salt concentration can be simulated by a parallel resistorcapacitor (RC) in series with a constant phase element, as shown in Fig. 6.42.

ω

4000

–Z˝ (ohm)

3000

21.3 °C 31.6 °C 41.0 °C 51.8 °C 61.3 °C 70.7 °C 79.9 °C

2000

1000

0 2000 3000 Z´ (ohm)

1000

4000

6.41 Nyquist plots obtained for the Z-IOPE [VI] at various temperatures and in the frequency ranges from 20 Hz to 1 MHz. Copyright Wiley-VCH Verlag GmbH & Co. Reproduced with permission from Di Noto et al. (2002b).

Zb CPE Cb

6.42 Equivalent circuit used to model Nyquist plots shown in Fig. 6.41.



261

The capacitance and the constant phase element can be distinguished by using two distinct constant-phase elements (CPE) in the simulation and by determining the corresponding value of nCPE. The parallel RC is represented by a semicircle at higher frequencies, while the Warburg element appears as a linear branch at lower frequencies. The value of R determined by EC analysis is used to calculate the direct current conductivity (σDC = 1/ρ, where ρ is the resistivity of the material). A VTF equation (equation 6.3) can be used to fit the plot of σDC vs. 1/T: σDC (T ) = Aσ ⋅T

−

1 2

⋅ exp [ − EVTF R (T − T0 )]

[6.3]

The VTF equation models a viscous material whose conductivity is regulated by segmental motion. More than one VTF branch is observed in some samples in different temperature regions (Münchov et al., 2000; Di Noto et al., 2010b). The VTF equation has three parameters: Aσ is a pre-exponential factor proportional to the concentration of carriers, EVTF is a pseudoactivation energy for the conduction and T0 is the ideal thermodynamic glass transition temperature [(Tg − 55) ≤ T0 ≤ (Tg − 40)] (Di Noto, 2002). These parameters can be analysed vs. the molar ratio between the concentration of cations and the concentration of oxygen atoms in the backbone or vs. the square root of the concentration of cations to identify regions with different conductivity mechanisms. It was found that the σDC value determined at low frequencies (below 10 kHz) from EC analysis matches the real conductivity at zero frequency σ′(0) determined above 10 kHz with the universal power law (UPL, Equation 6.2). Figure 6.43 shows a typical analysis carried out on the spectra of σ′(ω) in the framework of the UPL. The result σDC = σ′(0) is important since the EC analysis fails to predict σ′(ω) at high frequencies (above 10 kHz). The other two parameters of the UPL are also interesting. τ1 is a time related to the initial site relaxation rate τ2, while p = τ2/τ* can be expressed in terms of the ratio of the initial backhopping rate (1/τ*) and the initial site-relaxation rate (1/τ2) following the jump relaxation model (see Section 6.5). Depending on the value of p it is possible to estimate the effectiveness of forward hopping. On the basis of the value shown by p (either larger or smaller than 1) and knowing what ionic species are present in the system it is possible to distinguish between mostly cationic or mostly anionic transport mechanisms, within different materials in different temperature regions. In the case of Z-IOPEs it was shown that p is decreased as T is raised, as shown in the right panel of Fig. 6.28(b) (Di Noto et al., 2002b). This evidence implies that as T is raised the initial site relaxation time τ2 decreases more quickly than the initial backhopping time τ*. The ionic species are determined by plotting the equivalent conductivity ⌳ vs. the square root of the cationic concentration, as shown in Fig. 6.44. By extrapolating ⌳ to zero concentration it is also


262

Polymer electrolytes [II]

400

80° C

300 200 100 18° C

0 2

(a)

3

5

4

σ´ (ω) (S cm–1) × 10–6

σ´ (ω) (S cm–1) × 10–6

[I] 16

80° C

14 12 10 8 6 4 2

18° C

0

6

3

2

4

5

6

log (f/Hz)

log (f/Hz) 2.4

1.4

2.2

σ´ (ω) / σ´(0)

σ´ (ω) / σ´(0)

2.6 1.6 1.2 1.0 0.8 0.6 0.4

1.6 1.4 1.0

2

3

4

5

6

2

3

In[σ´ (ω) / σ´(0)–1]

0

18 °C

–1 –2 –3 –4 –5 –6

80 °C 5.4

5.5

5.6

5.7

4

5

6

log (f/Hz)

log (f/Hz)

(b) In[σ´ (ω) / σ´(0)–1]

1.8

1.2

0.2

(c)

2.0

5.8

5.9

1 18 °C

0 –1 –2 –3 –4

80 °C 5.2

6.0

5.4

log (f/Hz)

5.6

5.8

6.0

log (f/Hz)

6.43 Typical analysis of the real component of the complex conductivity σ′(ω) in the framework of the UPL. (a) Plots of the raw data fitted with the equivalent circuit shown in Fig. 6.42; (b) plots of the normalized figure σ′(ω)/σ′(0); and (c) plots of ln[σ′(ω)/σ′(0)-1]; the angular coefficient is the exponent p of the UPL. [I] and [II] are the materials {Al[O(CH2CH2O)8.7]2.24/(LiClO4)1.06} and {Al[O(CH2CH2O)8.7]1.89/ (LiClO4)0.01}n, respectively. Reproduced with permission from Di Noto et al. (2004a). Copyright 2004, The Electrochemical Society.

possible to determine ⌳0, which is correlated to the diffusion coefficient. A typical ⌳0 vs. T example is detailed on Fig. 6.45. The UPL equation can be generalized in the GUPL equation 6.4 (Di Noto, 2002): ⎡ p ⎤ σ ′(ω ) = σ ′( 0 ) ⎢1 + ∑ fi ( τ1,i ω ) i ⎥ ; ∑ fi = 1 ⎣ i =1 ⎦ i =1 N

N


[6.4]

Hybrid inorganic–organic polymer electrolytes 293.15 K 303.15 313.15 323.15 333.15 343.15 353.15 fit

1.0 Λ [S kg (cm mol)–1] × 10–3

263

0.8 0.6 0.4 0.2 0.0 0.2

0.4

0.6

1.0

0.8

cLi½ (mol/kg)½

6.44 Equivalent conductivity (⌳) as a function of salt concentration in {Al[O(CH2CH2O)8.7]r/(LiClO4)z}n complexes. Reproduced with permission from Di Noto et al. (2004a). Copyright 2004, The Electrochemical Society.

–3.2 log (Λo) fit

log [Λo / S kg (cm mol)–1]

–3.4 –3.6 –3.8 –4.0 –4.2 –4.4

2.9

3.0

3.1 3.2 1/T (K–1) × –3

3.3

3.4

6.45 Equivalent conductivity at infinite dilution (⌳0) vs. the reciprocal of temperature in {Al[O(CH2CH2O)8.7]r/(LiClO4)z}n complexes. Reproduced with permission from Di Noto et al. (2004a). Copyright 2004, The Electrochemical Society.


264


6 –2

log e /e 0

eel e´

2 0 –2

e˝

Dielectric relaxation

Electrode polarization –3

0

tion 3 log f [Hz]

6

tseg 9

log σ [S/m]

DC conduction 4

Electrode polarization

–4

tel DC conduction

–6

s´

Dielectric relaxation

sDC s˝

–8 –3

0

3 log f [Hz]

6

9

6.46 Electrical response of materials in terms of dielectric (left) and conductive spectra (right).

The GUPL equation includes the possibility of multiple relaxations characterized by a different fractional contribution fi. The application of BDS methods up to 1 GHz allows us to measure directly the ionic and segmental relaxation phenomena. These measurements yield the complex permittivity ε*(ω) and the complex conductivity σ*(ω). The latter two functions are complementary, as shown in equations 6.5 (Furukawa et al., 1997; Di Noto, 2002): σ′(ω) = ω · ε″(ω) σ″(ω) = ω · ε′(ω)

[6.5]

ε*(ω) and σ*(ω) can be used to emphasize different information, as highlighted in Fig. 6.46. ε*(ω) is particularly suitable to visualize the ionic and segmental relaxation phenomena (τion and τseg) after subtracting the contribution arising from σDC. By comparing σDC, τion and τseg as a function of the reciprocal temperature 1/T it is possible to determine whether the mechanism dominating the conductivity mechanism is the ionic phenomenon or the segmental motion (Vittadello et al., 2003). The measurement of ε″(ω) by BDS also provides a method to distinguish between different types of molecular motions in the polymer host, i.e. the segmental mode, the local mode β and the normal mode n (taking place between end groups), as shown in Fig. 6.47. The distinction is possible on the basis of the different dependence of the logarithm of the relaxation time (τα, τβ and τn) as a function of the reciprocal temperature: ln(τβ) is linear, while ln(τα) and ln(τn) show a Vogel–Tamman–Fulcher–Hesse (VTFH) behaviour, shown in equation 6.6 (Furukawa et al., 1997; Di Noto, 2002): ln ( τ ) = ln ( τ 0 ) −

A (T − T0 )


[6.6]

Hybrid inorganic–organic polymer electrolytes Segmental and ionic motions

10

265

0

8

–2

fseg

O 6

X–

log f (Hz)

O

O

X

O

O

–

M+ O

–4 sDC

4

–6

2

–8

0

–10

–2

–12

Tg

O

–100

log s (S/m)

O M+

–50

0

50

100

Temperature (°C)

1010

10–4

108

10–6 fseg sDC

106

10–8

104

10–10

102

10–12

100

10–14 Tg

10–16

10–2 0

1

2

3

4

sdc (S m–1)

fseg (Hz)

6.47 Example of the dependence vs. temperature of σDC and segmental mode.

5

–3 –1

1/T (10 K )

6.48 VTFH simulations of conductivity vs. T profiles. The conduction mechanism is regulated by the segmental motion.

A good fit of data with a VTFH equation indicates that the conductivity mechanism is regulated by segmental motion, as exemplified in Fig. 6.48. The segmental mode and the n normal mode can be distinguished on the basis of their position on the frequency axis; indeed, mode n is characterized


266


by lower frequencies. The measurements of the molecular motions in siloxanic proton conductors was particularly insightful in the identification of the role of the sulphonic groups. The latter give rise to the β mode and promote proton mobility among water domains. Solid state NMR can be used to determine: (a) the diffusion coefficients of protons and metal cations and; (b) the activation energies for conduction. As a consequence, this technique provides an independent method to validate the results of the electrical analysis (Di Noto et al., 2005).

6.5

The real component of the conductivity spectra in the framework of the jump relaxation model and polymer segmental motion

The discussion of the conductivity mechanism of hybrid inorganic–organic materials can benefit by taking into account the real component of conductivity spectra determined by equation 6.7 (Di Noto et al., 2002b): σ ′(ω ) =

Z′ {k ⋅[Z ′(ω )]2 + [Z ′′(ω )]2 }

[6.7]

Z is the complex impedance of the system and k = l/S is the cell constant; l and S are the length and the cross-section of the ion conductor. This relationship is easily determined by considering that: σ* (ω ) =

1 1 = ρ* (ω ) [ k ⋅ Z* (ω )]

[6.8]

where σ*(ω), ρ*(ω) and Z*(ω) are the complex conductivity, the complex resistivity and the complex impedance, respectively. The experimental profile of σ′(ω) vs. ω can be interpolated by using the UPL phenomenological equation 6.2, provided that it shows a single relaxation event. The UPL equation 6.2 can be interpreted on the basis of the jump relaxation model. This model was originally developed to describe inorganic solid ionic conductors such as RbAg4I5, Na-β-alumina and AgBr below the limit of 100 GHz (Funke et al., 1992; Cramer et al., 1995). The UPL equation cannot be simply explained by hypothesizing a random hopping of the charge carriers in a lattice of mostly empty sites. Random hopping would imply that the charge carriers do not affect one another and it would be possible to write equation 6.9 for a single mobile species: ⎧⎪ σ* (ω ) ∝ FT ⎨ ⎩⎪

N

∑ ν ( 0 ) ⋅ ν (t ) i

i, j

j

⎫⎪ ⎬ ≈ FT {N ⋅ δ (t )} ⎪ hops ⎭

[6.9]

In this case, the complex conductivity σ*(ω) is proportional to the Fourier transform (FT) of the correlation function of the velocity vectors of the



267

In[σ´(ω)] In σ´(∞)

In σ´(0) ω ω2 = t2–1

ω1 = t1–1

6.49 Typical profile of the real part of the complex conductivity vs. ω in solid electrolytes characterized by a disordered structure.

carriers involved in inter-site hopping (hops) (Funke, 1991). Equation 6.9 predicts a linear correlation between ln[(σ′(ω)] and ω, which is not observed. The jump relaxation model includes the possibility of an effective correlation and memory of the charge carriers; then, it is possible to write equation 6.10: ⎧⎪ σ* (ω ) ∝ FT ⎨ ⎩⎪

N

∑ ν ( 0 ) ⋅ ν (t ) i

j

i, j

⎫⎪ ⎬ ≈ FT νi ( 0 ) ⋅ ν j (t ) ⎪ hops ⎭

{

hops

}

[6.10]

Equation 6.10 predicts correctly that the profile of ln[(σ′(ω)] vs. ω is sigmoidal, as shown in Fig. 6.49. The potential V(x) of the ionic carrier is the sum of the periodic lattice potential Vp(x), which is sinusoidal, and the coulombic potential VC(x), which is approximated as parabolic as shown by Funke and Riess (1984) (see Fig. 6.50a): V ( x ) = Vp( x ) + VC ( x ) z2 ⋅ e 2 1⎞ ⎛ 1 VC ( x ) = − ⎟ ∑ ⎜ 4π ⋅ ε j ⎝ rj − r rj ⎠

= VC*( x ) ⋅ r 2 + L ( j = 1, . . , N ions ) configurat ions

[6.11] z · e is the charge of the carriers and L indicates higher order terms which can be ignored (here and hereafter). From time to time, the charge carrier in the filled state A is thermally activated and hops into a vacant neighbouring site B. The charge carrier may now either hop back relatively soon (this occurrence is energetically possible) or stay at its new position. If the charge carrier settles in, the medium reorganizes and the coulomb potential will move towards site B, turning the empty site B into a filled state A. These events are summarized on Fig. 6.50(b). If the charge carrier is in A at t = 0, the probability to find it in an empty site B at the time t is: WB(t) = W0(t) + W2(t) + L


[6.12]

268

Polymer electrolytes (a) Vp(x)

VC(x)

V(x) δ B

Δ

A (b)

δ Δ

t1

δ(t)B→C

δ(t)B→A

x(t)

x0 A

x0 B

C

6.50 Components of the overall potential (a) and hopping dynamics (b) in the jump relaxation model.



269

and the probability to find it in A is WA(t) = W1(t) + L

[6.13]

W0(t) is the probability that the charge carrier is still in B at the time t, W1(t) is the probability of finding the charge carrier in A at the time t after a back-hopping from B and W2(t) is the probability of finding the charge carrier in B after a forward-hopping followed by a back-hopping. On the basis of the linear response theory of Kubo and the theorem of fluctuationdissipation it is possible to write equation 6.14 (Kubo, 1957; Funke and Riess, 1984): ∞

σ ′(ω ) = σ ′( ∞ ) ⋅ ∫ {δ ( t ) + WB′(t )} ⋅ cos (ωt ) dt

[6.14]

0

The term WB(t) can be calculated directly using the probability theory or can be obtained by the correlation function of the velocity vectors of the ions. If the cross terms are neglected as in the approximation underlying the Nernst–Einstein (NE) equation, it is possible to write equation 6.15 (Funke, 1991):

(1 N ) ⋅

N

∑ νi ( 0 ) ⋅ ν j ( t ) i, j

≈ ν ( 0 ) ν (t )

NE hop

=

hop

1 d2 2 ⋅ r (t ) 2 dt 2

hop

[6.15]

The mean squared displacement 〈r2(t)〉hop of each charge carrier can be related to WB(t) by the following equation 6.16: r 2(t )

t

hop

x0 2

= Γ 0 ⋅ ∫ WB(t ′ ) dt ′

[6.16]

0

where x0 is the A–B distance between two neighbouring sites and (Funke and Riess, 1984): Γ0 = ν0C · exp[−(δ/kBT)]

[6.17]

is the back-hopping rate through the energy barrier δ with the attempt frequency ν0C. Therefore equation 6.18 is obtained (Funke, 1987) where W′B(t) indicates the first derivative of WB(t): v ( 0 ) ⋅ v (t )

hop

=

x0 2 Γ 0{δ ( t ) + WB′( t )} 2

[6.18]

By substituting 6.18 in 6.10, it is found that: σ ′(ω ) = N

∞

x02 Γ 0 ⋅ ∫ {δ (t ) + WB′(t )} ⋅ cos (ωt ) dt 2 0

[6.19]

where σ′(∞) is equal to the factor outside of the integral. A successful jump forward of a charge carrier from a site A to a site B requires a site relaxation


270


following the hopping event before the back-hopping event can occur. The initial back-hopping rate is defined as 1/τ* while the initial site-relaxation rate is 1/τ2. The times τ2 and τ1 are the reciprocal of the frequency points read on the sigmoid ln[(σ′(ω)] vs. ω before and after the flexus point, respectively (see Fig. 6.49). If we indicate with WA(t) the probability of backhopping it is possible to write (Cramer et al., 1995):

(

dWA (t ) 1 =− τ* dt

)

[6.20]

t =0

In addition,

(

1 1 dδ back(t ) = τ 2 kBT dt

)

[6.21]

t =0

where δback(t) is the energy barrier preventing back-hopping. It is demonstrated that, in the UPL equation 6.2, p = τ2/τ*. For small values of t the site-relaxation is not complete and it is possible to write: δ (t ) −WA′ (t ) = WA ( t ) ⋅ ν0 B ⋅ exp ⎡⎢ − back ⎤⎥ ⎣ kBT ⎦

[6.22]

In addition, −

(δ (t ) − δ ) ⎤ WA′ ( t ) 1 = exp ⎡⎢ − back ⎥⎦ WA ( t ) τ* kBT ⎣

[6.23]

where −δ ⎤ 1 = ν0 B exp ⎡⎢ − τ* ⎣ kBT ⎥⎦

[6.24]

The initial equation 6.23 for 1/τ* is obtained as WA(0) = 1 and δback(t) = δ. The rate of back-hopping is regulated by the temporal dependence of δback(t). If there is no site-relaxation, δback(t) = δ for every t and the integration of equation 6.23 yields equation 6.25: t WA ( t ) = exp ⎡ − ⎤ ⎢⎣ τ* ⎥⎦

[6.25]

However, numerical simulations have shown that for short values of t (high frequencies) the rate of site-relaxation is well described by the following relationship (Maass et al., 1991): ⎡ − −δ back(t ) ⎤ = 1 ⎢⎣ kBT ⎥⎦ τ 2 + t

[6.26]

This equation coincides with equation 6.21 as t → 0. The integration of equation 6.26 yields:


Hybrid inorganic–organic polymer electrolytes t δ back(t ) − δ = kBT ⋅ ln ⎡⎢1 + ⎤⎥ ⎣ τ2 ⎦

271 [6.27]

As equation 6.27 is substituted in equation 6.23, the following result is obtained: ⎡ WA′ ( t ) 1 ⎢ 1 − = ⎢ WA ( t ) τ* ⎢ 1 + t τ2 ⎣

⎤ ⎥ ⎥ ⎥ ⎦

[6.28]

The integration of equation 6.28 yields: t WA ( t ) = ⎛ 1 + ⎞ ⎝ τ2 ⎠

τ − 2 τ*

[6.29]

If τ2 → ∞, the original exponential form is obtained: t WA ( t ) = lim ⎛ 1 + ⎞ τ 2 →∞ ⎝ τ2 ⎠

τ − 2 τ*

t = exp ⎡ − ⎤ ⎢⎣ τ* ⎥⎦

[6.30]

By using the formula of Kubo: ∞

⎡ ⎤ σ ′(ω ) − σ ′( 0 ) = [σ ′( ∞ ) − σ ′( 0 )]⋅ ⎢1 + ∫ WA ( t ) ⋅ cos (ωt ) dt ⎥ ⎣ 0 ⎦

[6.31]

and using equation 6.30 it is found that, at high frequencies, equation 6.32 holds true: −p

σ ′(ω ) − σ ′( 0) = [ σ ′( ∞) − σ ′( 0)]⋅ [1 + (ωτ 2 )−1 ]

[6.32]

with p = τ2/τ*. The value of p can vary between 0 and 2, as required by the transformation conditions. This equation is satisfactory for the simulation of the plateau at high frequencies. At the lower frequencies (Fig. 6.49), it is necessary to consider the case ωτ2 1/2) such as 6Li, 7Li, 59Co, and 27Al. Each local environment in a solid gives rise to its own distinct resonance, and as long as 278 © Woodhead Publishing Limited, 2010

Using nuclear magnetic resonance spectroscopy

279

the environments are sufficiently different, these local environments can be resolved and quantified. The resolution of the spectrum will depend on the nature of the interactions that control, shift, and broaden the resonances. In a diamagnetic material, the nature of the coordination environment for lithium (6Li or 7Li) will give rise to a distinct and characteristic ‘chemical shift.’ Unfortunately, Li is a light element, and as in the case of 1H NMR, its chemical shift range is small, which can result in overlapping resonances. NMR spectroscopy may be used to probe processes with a wide range of correlation times. Spin–lattice relaxation (T1) measurements are sensitive to motional processes occurring at time scales of the reciprocal of the NMR frequency (several to tens of nanoseconds), and other relaxation times (T1ρ and T2) and sample millisecond scale correlation times. Similarly, the observation of spectral line-narrowing signifies motion comparable to the reciprocal of the line width (i.e. milliseconds for kilohertz linewidths). In addition, self-diffusion coefficients (D) of mobile species (D > 10−9 cm2/s) can be determined by pulsed gradient spin-echo (PGSE) methods, as originally described by Stejskal and Tanner, to be described in detail later. This is an important tool in studies of ion transport in gel and polymer electrolytes, in that both cation (i.e. 7Li) and anion (for anions containing 19F) diffusion coefficients can be obtained.

7.1.1 Polymer electrolytes The need for an efficient, compact and environmentally friendly energy storage source has led to the evolution of all solid battery systems that utilize electrolytes consisting of solid or at least non-leaking polymer materials. Properties of these are often compromises achieved by optimization of mechanical, dynamic and compatibility issues with addition of economic and environmental impact aspects. There have been many different approaches to this problem described in the literature,4–8 where the research moved from solid polymer (i.e. high molecular weight or cross linked electrolytes9 to polymer matrix hosting liquid electrolytes, gel electrolytes10 and later composites containing filler additives.11–16 In the past years the use of nanoscopic additives in polymer electrolytes has brought many interesting results, which will be described in more detail later.17–24

7.2

Nuclei possibility

Essentially every element in the periodic table has at least one isotope with a nuclear magnetic moment. Not all can be easily observed by the NMR


280


technique as the receptivity of different nuclei varies greatly, depending on both the nucleus’s gyromagnetic ratio, its natural abundance, and other factors such as quadrupole moment. The frequency at which a response from a certain nucleus can be observed is also dependent on the strength of the magnetic field. With optimized equipment and experimental settings a wide range of different species can be studied. Furthermore, it is often possible to purchase chemicals enriched with isotopes that do not occur naturally in high enough abundance, common examples being 2H and 13C. Most easily measured is the highly abundant, sensitive, and stable nucleus of hydrogen 1H, followed by only slightly less receptive fluorine 19F. Deuterium 2D is used as an alternative to proton measurements, and deuterated samples are often used as standards in organic analysis. Considering the importance of carbon in organic and polymer chemistry, one of the most often used nuclei is 13C. It is much less sensitive but has other advantages compared with the proton, such as a much larger chemical shift range and, because of the much lower abundance as compared to the proton (1.1%), practically no observed spin–spin couplings, interactions between nuclei of the same type. The resulting spectrum is therefore greatly simplified and much less crowded than a proton one, yielding sharp signals in a wide chemical shift range, which means that the individual bands are easily identified. Lithium, often used as cation in polymer electrolytes, has two useful nuclei. 6Li yields sharp signals but has low sensitivity due to its low natural abundance (6%) and lower (than 7Li) gyromagnetic ratio. On the other hand, 7Li is highly sensitive but has a larger quadrupole moment so its signals are broader. Both nuclei exhibit only a modest chemical shift range. Other isotopes often studied in polymer electrolyte research are those of different metal cations. In the literature, measurements on sodium 23Na, magnesium 25Mg, potassium 39,40K, copper 63,65Cu, nickel 61Ni, aluminum 27Al and many others can be found. Besides fluorine many other measurable species occur in polymer electrolyte materials as anions, for example 31P, 11B or 35,37Cl. Finally, important nuclei for biochemical research, but also to lesser extent for this field are 17O and 14,15N. Owing to this richness of measurable nuclei, it is often possible to examine several species in one single sample, and usually, if choosing carefully, many if not all the different components of the polymer electrolyte mixture can be observed separately and mutually compared.25,26 What follows is a nonexhaustive review of previous NMR investigations of polymer electrolytes representing a broad range of approaches and materials systems. This chapter is restricted to lithium-containing polymers and does not discuss proton-conducting polymer electrolytes for fuel cell applications.



7.3

281

Liquid state nuclear magnetic resonance

5.0

4.5

4.0

3.5

3.0

2.5

1.030

1.341 1.234

2.089

3.358 3.099 2.396 2.736 2.607

3.641

4.484

In liquid samples, because of the interactions that are averaged to zero by the rapid tumbling of individual molecules on the picosecond timescale, the resulting spectra show sharp, narrow bands. Their position, the so-called chemical shift, is determined by the local magnetic field arising from the chemical bonds in the vicinity of the nuclei and their interaction with the laboratory field. The resulting spectra can therefore be used for identification of unknown chemical species, or for clarifying unknown chemical structures. This was performed on a new class of difluoroalkoxyborane compounds later used as additives in a polymer electrolyte material.26 Zheng et al.27 used 1H NMR to solve molecular structure of a newly synthesized co-polymer. Wang and Tang28 compared 1H NMR spectra of a polymer in solution (Fig. 7.1) with Fourier transform infrared (FTIR) spectroscopic results, when characterizing novel gel polymer electrolyte material based on a blend of several polymeric species. Besides the use of novel materials reported by many research groups, polymer electrolytes consisting of even well-known chemical structures present significant scientific challenges because interactions between the components are often not entirely clear. It is these interactions that facilitate or impede the desired ionic mobility within the materials, and knowledge and understanding of these is therefore of major importance. The great sensitvity of high resolution NMR to the exact surrounding of the observed species makes it possible to explore molecular interactions in a unique way. As an example, the interactions that exist between added plasticizer and a new polymer matrix, examined by high resolution NMR, can be elucidated29 (see Fig. 7.2). Polymer electrolytes are solid or semi-solid systems not easily examined properly at high resolution. For the purpose of revealing interactions

2.0

1.5

1.0

0.5

ppm

7.1 1H NMR spectrum of PMAML. From: Z-l. Wang, Z-Y. Tang, Electrochimica Acta 2004, 49, 1063–1068.28


282


(c)

(b)

(a)

200 180 160 140 120 100 80 d (ppm)

60

40

20

0

7.2 13C -NMR spectra of (a) PC (liquid), (b) VA–PSVE–Na (solid), (c) PC/ VA–PSVE–Na = 1.5/3.0 (gel). From: L-Y. Tian, X-B. Huang, X-Z. Tang, European Polymer Journal 2004, 40, 735–42.29

between individual species, model systems are often used. The classical long-chained polymers are replaced by polymers with short chains or even oligomers and liquid electrolytes are obtained. High molecular weight poly(ethylene oxide) (PEO) has often been replaced by so-called glymes, basically short chain PEO molecules,30–32 as they behave similarly to amorphous-phase PEO when end-capped with alkyl groups, shown both experimentally33–35 and by theoretical calculations.36–38 Ion-pair formation constants for different salts with fluorine-containing anions have been estimated by 19 F NMR for several glymes and successfully compared to data obtained from conductivity measurements.39

7.4

Solid state nuclear magnetic resonance

To study condensed and relatively rigid systems, solid state NMR cannot give as clear a picture of the composition of an examined sample due to resolution limitations compared to the liquid state. Nevertheless, magic angle spinning (MAS) methods do succeed in offering moderately high resolution and this will be discussed more fully later. Although the identification of individual species is more limited, the dynamics of the system as a whole can be probed from line width, line shape and relaxation behavior of the resulting peaks as described by Bloembergen et al.,40 and many others since that seminal work. As mobility within an electrolyte is of major



283

importance, solid state NMR experiments are of great use for the assessment of dynamic properties of these materials.

7.4.1 Morphology Polymer electrolytes are heterogeneous materials with complicated phase diagrams usually consisting of both crystalline and amorphous components.41,42 Spectra of pure PEO and dilute salt PEO complexes typically show a broad and a narrow NMR signal that are superimposed. For pure PEO, these components have been assigned to the amorphous and the crystalline domains, but the assignments have varied. Connor and Hartland43 originally used proton spin–lattice relaxation to study the heterogeneous nature of pure PEO. The broad component, with a fast relaxation rate, was assigned to the amorphous and the narrow component, with longer relaxation rates, to the crystalline component. This assignment was later reversed by Dechter44 who assigned the broad component to the crystalline and the narrow component to the amorphous based on a reanalysis of the proton and 13C rotating frame spin–lattice relaxation data (T1ρ, see below). Johansson and Tegenfeldt45 reanalyzed and agreed with the results found by Dechter. All NMR studies of PEO electrolytes since have used this assignment. However, pure PEO and PEO electrolytes are fundamentally different samples and NMR assignments of their bands are not always straightforward.46 High salt concentrations have been shown to retain only the narrow component, while the more dilute PEO–salt samples contain both the narrow and broad components.47–49 In dilute PEO:salt electrolytes, three regions have been shown to exist: (1) pure crystalline PEO, (2) amorphous PEO, (3) amorphous PEO–Li salt complex, and (4) crystalline PEO–Li salt complex.42,50

7.4.2 Thermal properties Variable temperature study with a subsequent line shape analysis enables a clear insight of the temperature-dependent dynamics of the system, where the glass transition temperature (Tg) of the amorphous phase, melting point or a crystallization onset of a present crystalline phase, or any other changes caused by altered dynamics of the involved species can be assessed. This was recognized by early investigations of polymer electrolytes and will be discussed in greater detail later. Linewidth changes dependent on salt amount were examined by Liao et al.51 (see Fig. 7.3), and the line shape was also examined on heating and cooling, as shown in Fig. 7.4. From these graphs it is possible to conclude that spectral lines are narrow in the liquid crystal phase, while they broaden


284


14 000

Tm

(d)

Linewidth (Hz)

12 000

–90 °C

10 000

(a) Cryst. phase

8000 (c)

(b)

6000

Liquid cryst. phase

–20 °C 25 °C

4000 2000 –150

40

0 –100

–50

0

50

100

20 0 –20 Frequency (kHz)

–40

150

Temperature (°C)

Linewidth (Hz)

7.3 7Li and 19F NMR linewidths versus temperature for C18O5: LiBF4 (1 : x) second heating cycle data, Tm denotes the side chain melting temperature. Filled symbols are for 7Li and open symbols are for 19F. (a) 䉬, 䉫, x = 0.7, ‘salt-deficient’ sample; (b) 䊏, x = 1, ‘equimolar’ sample; (c) 䉱, 䉭, x = 1.3 ‘excess salt’ sample; (d) selected 7Li NMR signals from (b). From: Y-P. Liao, D. C. Apperley, J. Liu, Y. Zheng, P. V. Wright, Electrochimica Acta 2007, 53, 1444–54.51

2000 1800 1600 1400 1200 1000 800 600 400 200 0 –80

Tm (c)

(a)

(d) Cryst. phase (b)

Liquid cryst. phase

4

(a)

2

0

–2

(e)

(b) (f) –60

–40

–20 0 20 Temperature (°C)

40

60

80

10 8 6 4 2 0 –2 –4 Frequency (kHz)

7.4 (a) C18O1 : LiBF4 (1 : 0.6); 䉱, 䉭, 7Li and 19F NMR linewidths for the second heating and cooling cycles. Filled symbols are for 7Li and open symbols are for 19F. (b) C16O1O5(21%) : LiBF4 (1 : 0.7); 䊏, 䊐, 7Li and 19F NMR linewidths of the signals in C16O1O5(21%) : LiBF4 (1 : 0.7) for the second heating and cooling cycles. (c) 7Li lineshape at −7 °C; (d) 7Li lineshape at −20 °C; (e) 19F line shape at −70 °C. Tm denotes the side chain melting temperature. (f) 19F line shape at −20 °C. From: Y-P. Liao, D. C. Apperley, J. Liu, Y. Zheng, P.V. Wright, Electrochimica Acta 2007, 53, 1444–54.51



285

considerably as the material crystallizes, indicating a decrease in dynamics, in accordance with both thermal analysis and conductivity measurements. The influence of salt concentration is also easily observed in the curvature of the graph lines, where the onset of increase of the linewidths moves to lower temperatures, indicating higher dynamics of the material at low temperatures. The difference between 7Li and 19F spectral linewidths is partly due to their different gyromagnetic ratios, but also indicates differences between the ionic species’ mobilities. Often, sets of samples are studied where one parameter, concentration or relative amount of one of the components, or one component itself, is varied. The observed changes in dynamics and morphology caused by the altered composition of the sample, such as relative amount of crystallinity, altered coordination of the involved species, or different degree of ionpairing, are of paramount importance in the quest for an optimized electrolyte system.

7.4.3 Composition The complexity of polymer electrolyte systems and the influence of addition of different nanoparticles were discussed by Singh and Bhat52 based on 7Li linewidth observation. They state that while it is generally expected that when ionic motion occurs, the NMR linewidth of the mobile nucleus decreases, several discrepancies have been observed. For example, Dai et al.17 with (PEO)1.5 + LiI + nanoscale Al2O3 or MgO, Chung et al.21 with (PEO)8 + LiClO4 + TiO2 nanoparticles and Forsyth et al.18 with 3PEG + LiClO4 + TiO2, SiO2 and Al2O3 nanoparticles observed an increase in the 7 Li linewidth on addition of nanoparticles, although the ionic conductivity increased. One possible explanation offered was that nanoparticles cause the Li–H distance to decrease which causes an increase in the Li–H heteronuclear dipolar interaction as the main contributor to the linewidth, observed through 1H decoupling experiments22 that will be described later. While this explanation is reasonable, Mustarelli et al.20 with PEO8 + LiClO4 or LiN(CF3SO2)2 + nanoscale SiO2 found a decrease in the 7Li linewidth of nanocomposites. Bloise et al.19 with PEO + LiClO4 or LiBF4 + α-Al2O3 or γ-Al2O3 observed mixed behavior where γ-Al2O3 caused an increase in the linewidth while α-Al2O3 resulted in a decrease in the linewidth. Thus the behavior of nanoparticles clearly varies in different systems. Further linewidth measurements of γ-Al2O3 doped and undoped systems were undertaken for both 1H and 7Li species by Singh and Bhat52 combined with conductivity and thermal measurements (Figs 7.5 and 7.6). There was no significant change in Tg or in 1H NMR linewidths, indicating no effect of the filler on segmental motion. From the combination of the investigative techniques it was finally concluded that the increase in ionic motion upon



Doped

Undoped

5000 Δn (Hz)

–4000 –2000

4000

0 Hz

2000

4000

3000 2000 1000 0 240

260

280

300

320

340

Temperature (K)

7.5 1H linewidth vs. temperature plot for (PEG)46LiClO4 + γ-Al2O3 (10 mol%); circles and triangles represent the linewidths for the undoped and doped samples respectively. Inset shows 1H signals at room temperature. From: Th. J. Singh, S. V. Bhat, Journal of Power Sources 2004, 129, 280–7.52

(a)

6000 Doped

5000 4000

–1000 –500

Undoped

0 Hz

500 1000

1200

3000

Δn (Hz)

Δn (Hz)

286

2000

(b)

800 400 0 240 260 280 300 320 340 Temperature (K)

1000 0 220

240

260

280

300

320

340

Temperature (K)

7.6 7Li linewidth vs. temperature plot for (PEG)46LiClO4 + γ-Al2O3 (10 mol%); circles and triangle represent the linewidths for the undoped and doped samples respectively. Inset (a) shows 7Li signals at room temperature. Inset (b) is an expanded plot of main figure to show sharpness of the transition for doped sample. From: Th. J. Singh, S. V. Bhat, Journal of Power Sources 2004, 129, 280–7.52



287

(a) LiIn

Intensity

LiIon pair

LiOut

(b)

LiPEO LiIn

2

LiOut 0 –2 –4 Chemical shift (ppm)

–6

7.7 Least square peak fits of 7Li NMR spectra of (a) SBA-15:LiClO4 (1:1 wt%); and (b) 8 wt% of SBA-15 in PEO:LiClO4. From: M. J. Reddy, P. P. Chu, U. V. Subba Rao, Journal of Power Sources 2006, 158, 614–19.55

addition of filler causes a decrease in degree of crystallinity of the polymer which is furthermore highly dependent on the amount of the filler added. Interactions in composite PEO systems with added mesoporous silica were explored by Reddy and co-workers53–55 and three possible environments of Li ions were identified by NMR studies, as shown in Fig. 7.7. In Fig. 7(a), the NMR spectrum of a Li salt solution containing silica is fitted by several peaks, a downfield peak near 0.0 ppm assigned to the Li species of an undissociated salt (ion pairs), and two upfield peaks, the middle ‘LiIn’ (0.5–0.7 ppm) and higher ‘LiOut’ (1.1–1.5 ppm), which were assigned to lithium ions coordinated on the interior channel surface and onto the outer surface of the silica particles (SBA-15), respectively. The second spectrum (Fig. 7.7b) is that of a polymer electrolyte with added silica. Here the same peaks of lithium coordinated to the silica particles are found (‘LiOut’ and ‘LiIn’) in the same positions, and while the peak representing ion pairs has disappeared a new peak designed ‘LiPEO’ emerges at (0.8–1.0 ppm) and is thought to represent lithium species coordinated to amorphous PEO. It was also shown here, that when the amount of silica SBA-15 increases the components of ‘LiOut’ and ‘LiIn’ increase dramatically while LiPEO decreases, which indicates that lithium favors the association with silica over PEO.

7.4.4 Dynamics It is very important to study the motion of the mobile ionic species in polymer electrolytes as these are dynamic environments created by the


288


polymer motion in the amorphous phase and the ionic transport is closely connected to the polymer segmental dynamics. In order to investigate the dynamics of an electrolyte system, the line width of the central 7Li transition as a function of P(EO-EC)/LiCF3SO3 electrolyte content (P(EO-EC) = poly(ethylene oxide-co-ethylene carbonate)) as well as temperature were measured by Jeon and Kwak56 on solventfree polymer electrolyte materials based on poly(vinylidene fluoride) (PVdF) membranes filled with P(EO-EC). As seen in Fig. 7.8 a typical temperature dependence is observed for this system with broad spectral lines below Tg and narrow linewidths above Tg, indicating Li ion mobilities closely associated with segmental motion of P(EO–EC) polymer. In a subsequent study57 it was demonstrated how some important parameters can be further calculated from the linewidth measurements. Based on the Bloembergen–Purcell–Pound (BPP) theory40 NMR motional narrowing of the linewidth occurs when the rate of fluctuations of the local magnetic fields or electric field gradients, generally described as the correlation time, τc, is of the order of the rigid lattice line width, δωo, as described by following equation, where the measured linewidth, δυ, obeys the relation

( Δυ ) 2 =

(π )δω 2

0

2

tan −1(τ c Δυ )

[7.1]

It is generally assumed that temperature dependence of τc follows Arrhenius behavior, and then the activation energy Ea for Li ion mobility is determined by

τ c = τ 0 e( Ea / RT )

[7.2]

where τ0 is the prefactor (dwell time) corresponding to the reciprocal frequency attempt of cation jumps. From this equation both τ0 and Ea can be obtained from plotting ln τc as a function of inverse temperature, resulting in a graph such as in Fig. 7.9. The plots show two different temperature regions with widely different activation energies, separated by Tsc (temperature of slope change) that correspond to an abrupt change in ionic mobility, together with change in correlation time, activation energy, and ionic conductivity. These results are comparable to those of other polymer electrolytes such as poly(vinyl alcohol)/LiCF3SO3-based solid polymer electrolytes (SPEs)58 and polyurethane/poly-(dimethylsiloxane)/LiClO4-based SPEs.59 Similar observations have been made on conductivity measurements, where Tsc corresponds to Tg, the glass transition temperature of the amorphous polymer phase. The activation energies obtained by conductivity measurements corresponding to an energy barrier of Li ion motion are here approximately twice larger than those from 7Li NMR data. Bishop and Bray60 have also observed differences in NMR and conductivity activation energies in



340 K 320 K 300 K 280 K 260 K 245 K 230 K

330 K 310 K 290 K 270 K 250 K 240 K (a)

100

50

0

289

–50

–100

ppm

104

E-V6E4

–12 ln tc (s)

103

7

Li FWHM (Hz)

E-V10E0 E-V8E2 E-V6E4

–13 –14 Tsc

–15 3.0

3.0 (b)

3.5

4.0

3.5 4.0 4.5 1000/T (K–1)

4.5

1000/T (K–1)

7.8 (a) Temperature dependence of the 7Li NMR spectrum of E-V6E4 for a salt concentration of 1.5 mmol LiCF3SO3/g-P(EO-EC) as a function of temperature. (b) Linewidths of the 7Li NMR spectra of the three polymer electrolytes as functions of temperature. The graph in the inset shows the temperature dependence of ln τc determined from the linewidth data for E-V6E4. From: J-D. Jeon, S-Y. Kwak, Journal of Membrane Science 2006, 286, 15–21.56

lithium borate glasses because the 7Li NMR and conductivity can detect ion motions on a local scale and a more global scale, respectively. That is, the energy barrier for the Li ion to exceed in order to undergo long-range transport is higher than that of local motions.


290



–12

ln tc (s)

–13

–14 Tsc –15 3.0

3.5

4.0

4.5

–1

1000/T (K )

7.9 Temperature dependence of the corresponding correlation time, τc, calculated from the 7Li NMR line width data by using the BPP equation for polymer electrolytes. From: J-D. Jeon and S-Y. Kwak, Macromolecules, 2006, 39, 8027–34.57

7.4.5 Amorphous phase transport One of the important contributions to electrolyte research coming from the NMR measurements is the recognition of ion conduction taking place predominantly in the amorphous phase of the electrolyte and only at temperatures above the glass transition of this phase. These results were obtained by motional narrowing studies, where the linewidth of resulting spectral bands is studied as a function of temperature.61–64 The results indicate that the ionic mobility is closely connected to the mobility of the polymer chains and a segmental motion-assisted ion-transport mechanism has been proposed and is now broadly accepted in the polymer electrolyte community.

7.4.6 Oriented systems This transport mechanism has been further studied on oriented systems, polymer electrolyte materials that had been mechanically65or magnetically66 oriented in a certain direction. NMR studies have revealed important facts about lithium–polymer complexes within these anisotropic systems. It was established that the re-ordering induced by the stretching extends down to the immediate environment of the cation, as evidenced by obtained angular-dependent linewidths. At the same time the diffusion coefficients were shown to increase along the stretched direction67 indicating that other factors contribute to the ion-transport mechanism, such as the alignment of the structural units of the polymer chains. Additional evidence for this was



291

provided by the observation that polymer electrolytes of certain composition can be prepared in either amorphous or crystalline forms and that the latter exhibits higher conductivity in some cases.68 These results suggest a possible re-examination of Armand’s early model of ion transport in PEO in which cations move within the helical polymer structure.26 Although this mechanism has been discounted for some years, it may in fact yield a substantial contribution to the cation conductivity, but only when the helices are aligned. For randomly oriented polymer chains and helices, ionic transport assisted by amorphous polymer segmental motion remains as the dominant mechanism. However, over two decades of research on suppressing the crystalline phase and lowering the Tg of the amorphous phase have produced only incremental improvements in Li+ ion conductivity.

7.5

Relaxation processes

As a practical matter in NMR spectroscopy, relaxation times determine the length of the delay between signal-averaged acquisitions, but at the same time they are a property of the nucleus that reflects on its dynamic environment within a sample. Any excited magnetic moment relaxes back to equilibrium. There are generally two components of this relaxation for isotropic systems in the absence of chemical exchange: longitudinal or spin–lattice (T1) and transverse or spin–spin (T2) relaxations. The often measured rotating-frame relaxation time (T1ρ) is a modified T1 relaxation which is sensitive to dynamic processes of a frequency range intermediate between T1 and T2.

7.5.1 Spin–lattice relaxation (T1) Relaxation in the applied static field direction is characterized by spin– lattice or longitudinal relaxation time T1. It is the process of energy transfer from the excited nucleus to the surroundings or the lattice, usually meaning the neighboring molecules, that returns the nuclear spin system to thermal equilibrium. Since the first comprehensive treatise of relaxation published over 60 years ago,40 measurements of T1 as a function of temperature have provided a useful molecular level probe of dynamics because the most effective relaxation mechanisms, whether mediated by nuclear dipole– dipole, quadrupole, or electron spins (in samples containing paramagnetic centers), must have significant spectral density components close to the NMR frequency. Thus T1 is very sensitive to motional processes with a timescale of ~10−10 s. This can be seen from the original formulation of the relaxation rate involving a single process via:40 1 τc 4τ c ⎤ = K ⎡⎢ + T1 ⎣ 1 + ω 02τ c2 1 + 4ω 02τ c2 ⎦⎥


[7.3]

292


where K depends on the nuclear interaction, ω0 is the NMR frequency, and τc is the correlation time. For measuring T1 a so-called inversion-recovery technique is often used, where the sample is subjected to sequences of 180° and 90° pulses separated by variable time τ. Jeon and Kwak56 have employed T1 measurements using 1 H solid state NMR to determine miscibility of P(VdF–HFP)/P(EO–EC)based porous membranes (HFP = hexafluoropropylene), where it was assumed that a single T1 relaxation complies with the condition that M − Mτ ⎤ τ ln ⎡⎢ e =− ⎥ T1 ⎣ 2 Me ⎦

[7.4]

where Mτ and Me are the intensities of the resonance at a given delay time, τ, and equilibrium state (τ ≥ 5T1), respectively. By plotting ln[(Me − Mτ)/ (2Me)] against τ, straight lines with different slopes were obtained corresponding to the pure and blend membranes, indicating a single component T1 relaxation behavior (Fig. 7.10). From the slopes of the plots, T1 values can be obtained. The fact that a single T1 was observed indicates that the spin diffusion process is sufficiently fast to equilibrate the relaxation times for all protons among the chemically different constituents and it was concluded that the blended membranes are completely homogeneous on the time and length scales probed by T1.

In[(Me – Mτ)/(2Me)]

0

M-V10E0 M-V8E2 M-V6E4

–1

–2

–3

–4 0.0

0.5

1.0 1.5 Delay time (s)

2.0

2.5

7.10 Logarithmic plots of resonance intensity vs. delay time for porous membranes. The slope yields the proton spin–lattice relaxation time in the laboratory frame, T1. From: J-D Jeon and S-Y Kwak, Macromolecules 2006, 39, 8027–34.57



293

7.5.2 Spin–spin relaxation (T2) Relaxation of the nuclei perpendicular to the static field direction is characterized by spin–spin or transverse relaxation time T2. This is an entropic process, where the coherence of spins is gradually lost without any energy exchange with the lattice. The main significance of this relaxation lies in its relationship with the homogeneous linewidths of the resulting spectral bands, Δω =

1 πT2

[7.5]

Δω being the spectral width at half-height. Inhomogeneous broadening mechanisms, for example from the nuclear quadrupole interaction, give addition contributions to the linewidth.

7.5.3 Spin–lattice relaxation in the rotating frame (T1ρ) In solid state systems the perturbation from the ideal radiofrequency (rf) pulses often causes random fluctuations of dipole–dipole interactions that hinder accurate measurements of the relaxation time. Undertaking measurements in a rotating frame, achieved by a spin-locking process, an unperturbed value of the relaxation time of T1ρ is obtained. Another benefit is that T1ρ is sensitive to motions on a somewhat longer timescale than T1. In contrast to the other relaxation processes T1ρ magnetization decay is often non-exponential, owing to complex interactions or a multitude of phases within the solid samples. In the case of quadrupolar nuclei, such as 7Li (I = 3/2), the spin–lattice relaxation is often described by a distribution of two or more exponential functions.69 Because of the relatively small quadrupole moment of 7Li, however, deviations from a single-exponential function are often hard to detect,59,63,70 in which case only one exponential function is necessary when evaluating 7Li relaxation measurements. Such was also the procedure of Jeon and Kwak57 when performing analysis of Li-ion transport behavior. Figure 7.11 (the lower graphs) shows the temperature dependence of the 7Li spin–lattice relaxation rates in the laboratory frame, T1−1, of P(VdF–HFP)/P(EO–EC)-based porous membranes, mentioned above. From the graph it is clear that above 240 K, T1−1 increase significantly with increasing temperature up to a maximum value of ca. 3 s−1 at T ≥ 360 K in all of the polymer electrolytes investigated. One of the samples, E-V6E4, displays well-defined T1−1 maximum at 360 K, which is an indication of only one motional process of the Li ions. In the case of the other samples, however, T1−1 maxima were not observed in the investigated temperature range. To gain better understanding of the Li ion mobility in polymer


294

Polymer electrolytes Tmax

1/T1ρ (s–1)

103

102


101

1/T1 (s–1)

100

10–1 2.7

3.0

3.3

3.6

3.9

4.2

–1

1000/T (K )

7.11 Temperature dependence of the 7Li spin–lattice relaxation rates, T1−1 and T1ρ−1, for polymer electrolytes measured at the Larmor frequency 155.4 MHz. The uncertainties in the 7Li spin-lattice relaxation rates are approximately 5%. From: J-D. Jeon and S-Y. Kwak, Macromolecules 2006, 39, 8027–34.57

electrolytes, 7Li spin–lattice relaxation time in the rotating frame, T1ρ, measurements were undertaken in the temperature range of 240–340 K. Figure 7.11 (the upper graphs) displays the Arrhenius plots of the 7Li T1ρ−1 for the same polymer electrolyte samples as before. For all, the T1ρ−1 plots versus reciprocal temperature exhibit clear maxima which indicate the most efficient relaxation; the T1ρ−1 values to the left of the maximum imply fast motion. The T1ρ−1 maxima (also called Tmax) of the polymer electrolytes appear to have a trend to shift to lower temperatures with increase of the amount of the P(EO-EC)/LiCF3SO3 electrolyte, indicating that the Li ions are more mobile. This can be correlated with the increased conductivity observed for these polymer electrolytes. As above, the correlation time for the relaxation process can be calculated, here expressed by the equation:

τc 1 3 4 2 ⎡ 2.5τ c 1.5τ c ⎤ = γ h⎢ + + T1ρ 2r 6 ⎣ 1 + ω 0 2τ c 2 1 + 4ω 0 2τ c 2 1 + 4ω 0 2τ c 2 ⎥⎦


[7.6]


295

Activation energy Ea for the relaxation process, corresponding to the barrier height for the potential hindered motion, can also be obtained. For these systems the energy values are consistent with results obtained from ionic conductivity measurements and linewidth analysis.

7.6

Diffusion measurements

Another use of NMR technique in polymer electrolyte research is the measurement of self-diffusion coefficients. Here the possibility of measurements of different nuclei in one sample is fully used, and the diffusion coefficients of cationic and anionic species and the polymer chain can be obtained, together with the diffusion constant of the sometimes present additives, such as fillers or plasticizers. Diffusion experiments, first proposed by Carr and Purcell,71 are based on the fact that nuclei in a sample placed into a non-uniform magnetic field will experience different local magnetic fields depending on their spatial location. The sample is then subjected to a sequence of rf pulses. A 90° pulse is followed at time τ by a refocusing 180° pulse. This will produce a spinecho at time 2τ. The resonant spins will completely refocus only if the precession frequency is constant during the experiment. If the nuclei move along the magnetic gradient, the refocusing will be incomplete. The reduction in echo intensity due to incomplete refocusing will give the diffusion coefficient, expressed by the equation:72 2τ ⎤ ⎡ 2 δH ⎞ M ( 2τ ) = M ( 0) exp ⎢⎢ ⎥⎥ exp ⎢ − ⎛ γ Dτ a ⎥ ⎝ ⎠ δ 3 z ⎣ T2 ⎦ ⎦ ⎣ 2

[7.7]

T2 being the spin–spin relaxation time as discussed previously, γ the gyromagnetic ratio, δH/δz the magnetic field gradient, and D the diffusion coefficient. If τ is much smaller than T2, the decay of the spin-echo as a function of τ is dominated by the diffusion process and the effect of spin–spin relaxation needs not be taken into account. However in the case of a short T2, it is difficult to separate the two processes, relaxation and diffusion from each other.

7.6.1 Pulse gradient spin-echo (PGSE) measurements In this technique, the magnetic field gradient is switched off during the observation of the echo signal, first described by Stejskal and Tanner.73 This requires special equipment that allows both variation of gradient strength and variable duration and interval of gradient pulses. The contribution of T2 relaxation is here kept constant by the use of fixed delay τ between the rf pulses. The only unknown variable in the above equation is thus the diffusion coefficient. Also, it is possible to obtain high resolution spectra, which


296


allow measurements of diffusion coefficients of different molecular species separately. However, this technique has its restrictions, namely in the strength of the gradient that can be used, typically ranging between 10 and 70 mT cm–1.

7.6.2 Fringe-field measurements Unlike PGSE, very strong gradients can be used in the fringe field method74,75 in which the expression previously given for static gradients applies. Compared with the pulse field gradient technique, the spectral line width is here broadened, causing the loss of sensitivity, but this is compensated for by the fact that very small diffusion coefficients can be measured (below 10−12 m2/s), in the strong fringe fields characteristic of superconducting NMR magnets. The experimental set-up uses the fact that the gradient field outside the center of the magnet decreases linearly. Thus placing the probe-head in an exactly chosen position outside the center of the magnet, the sample is exposed to a large static magnetic gradient in the order of 10 T/m. The transverse relaxation time T2 needs to be determined independently (in a uniform field experiment), as it cannot be otherwise separated from the diffusion echo decay process.

7.6.3 Using diffusion experiments Conductivity and ionic diffusion are related by the modified Nernst– Einstein equation:

σ calc =

Nq2( DLi + Danion ) α kT

[7.8]

where σcalc is the calculated ionic conductivity, N the number of ions per unit volume, q the charge of the ions, α the degree of dissociation, and DLi and Danion are the diffusion coefficients of cation and anion, respectively. It has been generally observed76–81 that the calculated conductivity with the measured self-diffusion coefficients of fluorine containing anions DF and the cations DLi (obtained from pulsed field gradient (PFG)-NMR) is larger than experimentally observed conductivity for reasons that will be discussed later. The effect of molecular orientation on ionic mobility in liquid crystal electrolytes was investigated by Saito et al.82 The liquid crystals in this study undergo a phase transition with change in molecular ordering. Conductivity measurements did not show any differences between the isotropic and ordered phases; for example, no change in slope in the conductivity plots. Therefore diffusion measurements on 7Li, 19F and 1H were conducted in order to determine individual mobilities of the involved species.



297

–6.2

Log D (cm2 s–1)

–6.4 –6.6 –6.8 –7.0 –7.2 –7.4 2.7

25-50 °C Li 28 F 27 H 69 2.8

2.9

50-80 °C 43 kJ mol–1 42 kJ mol–1 44 kJ mol–1 3.0

3.1

3.2

3.3

3.4

1000 K / T

7.12 Temperature dependence of the diffusion coefficient of the cation (O), anion (Δ), and solvent species (䊐) from the nuclear species of 7Li, 19 F, and 1H, respectively, of LCE-B with 5 mol % Li-TFSI. Activation energies of diffusion are incorporated in the figure. From: Y. Saito, K. Hirai, S. Murata, Y. Kishii, K. Kii, M. Yoshio, T. Kato, J. Phys. Chem. B, 2005, 109, 11563–71.82

From the temperature dependence of the diffusion coefficients DLi and DF which follow closely together, it was observed that the two ionic species remain coupled, suggesting a low degree of salt dissociation (Fig. 7.12). The plot shows a change in slope at the phase transition temperature, indicating a change of the ion migration mechanism and therefore dependence on the degree of molecular orientation. It is characteristic that the activation energy of diffusion (Ea) above and below the phase transition point is reversed between the ionic species and the liquid crystal molecules (slope of DH). That is, Ea of the cation and anion diffusion in the nematic phase is lower than those of the isotropic phase. However, Ea of the diffusion of the liquid crystal molecules increases upon going to the nematic phase. This difference could be attributed to a fundamental difference in the migration mechanism of the dissolved small species and large molecules of the medium. The disagreement often observed between the ionic conductivity values measured directly and calculated from diffusion coefficient measurements via the Nernst–Einstein equation originates from several factors. The main one is attributed to the difference in the contributing species between the measurements. Electrical conductivity detects the motion of charged ions regardless of the type. However, the diffusion coefficient measures all mobile species, whether they are dissociated ions, charged multiplets, or


298


electrically neutral ion pairs as long as they coexist and contain the same probed nucleus such as 7Li or 19F.83 Another factor is the possible difference in migration distance probed by self-diffusion and DC conductivity, often occurring in inhomogeneous media. Kao et al.,84 in polymer electrolytes containing LiClO4, compared DLi behavior with DPC, diffusion coefficient of propylene carbonate, a typical solvent in plasticized polymer electrolytes. A significant difference in temperature dependence was found and it could be concluded that there is very little interaction between lithium ions and the solvent. Correlation between ionic conductivity measurements and self-diffusion coefficients of lithium was also shown by Nicotera et al.85 on a group of poly(methylmethacrylate) (PMMA)–PVdF-based gel electrolytes. The mobility of lithium ion has been shown to strongly depend on the morphology of individual samples. Corain et al.88 have used measurements of diffusion coefficients in combination with electrochemical spectroscopy to investigate the degree of crosslinking inside a gel polymer electrolyte. Saunier et al.25 tried to compare diffusion coefficients of liquid electrolyte and macroporous PVdF matrix swollen with the same electrolyte, which was found to be too low for significant values to be obtained. Conductivity measurements showed a decrease by a factor of 3.6. Diffusion measurements on similar systems were conducted by several other groups.87–90 Diffusion measurements in connection with T1 relaxation experiments can provide an estimate of the diffusive pathlength and hence the sizes of blend heterogeneities,57 especially when it is not possible to measure diffusion coefficients directly. A useful approximation of the upper limit to the domain size in the case of proton NMR can be calculated by the equation where = (6DT)1/2

[7.9]

where is the average diffusive path length for the effective spin diffusion, and D is the effective spin diffusion coefficient determined by the average proton–proton distance which determines the strength of the dipolar interaction. The relaxation time T can be represented by T1 or T1ρ according to the type of relaxation measurements.

7.6.4 Measurements of transport numbers In polymer electrolyte research one of the desired properties is high selective mobility of the cations (i.e. Li+). A multitude of approaches for increasing this selective mobility has been published,91–93 including decreasing crystalline content, inhibition of the anion mobility, solvating of cation in a mobile medium or introduction of inorganic fillers as referenced above.



299

One way of measuring selective mobility in a material is calculating transference numbers of individual species as shown by Kalita et al.,14 where diffusion coefficients of measurable electrolyte species were obtained and transference numbers were calculated using the relation: t+ =

D+ D+ + D−

[7.10]

where t+ is the lithium transference number, and D+, D− are the cation and diffusion coefficients, respectively. Obtaining transference numbers for free lithium ions is not straight forward in any of the techniques commonly used. While impedance spectroscopy measures the combined mobility of any charged species, including H+ and charged ion multiplets, diffusion measurements of lithium will show diffusion of all lithium species, including neutral ion pairs, as discussed previously. Usually, the comparison between NMR and electrochemical data will give additional information about the examined material, its conduction mechanism, and aggregates formed.94

7.7

Magic angle spinning

The broadness of the spectral lines in solid state NMR is caused by strong interactions occurring in condensed phases, mainly represented by dipolar and quadrupolar interactions and chemical shift anisotropy. However, these interactions have the same geometrical scale factor, [3 cos2θ − 1], where θ is the angle between the principal interaction direction and the external magnetic field, and can be averaged by spinning the solid sample at the so-called magic angle of 54.7°, for which 3 cos2θ − 1 = 0. With increasing spinning speed a central band emerges at the isotropic chemical shift surrounded by sidebands, whose intensities can be used to calculate the chemical shift tensor. When the spinning speed exceeds the span of the chemical shift interaction, the sidebands become very small and a liquid-like spectrum is obtained. As the spinning speeds are extremely high, on the order of several times 104 rotations per second, the sample needs to be homogeneous and therefore it is possible to measure mainly finely pulverized samples. Many polymer electrolytes which are mechanically soft at room temperature require special handling procedures such as cryo-grinding in order to perform MAS measurements. A typical example of such an approach is investigation of nanoparticle surfaces where the formation and dynamics of lithium species is studied. The resulting high resolution spectra measured at different temperatures show a multitude of environments for the measured species, corresponding to different complexes, in some cases between which ionic exchange can occur.95–97


300


7.8

Double resonance experiments

For better elucidation of the measured MAS spectra, several techniques well known to the polymer science community are often added to the classical MAS measurements.

7.8.1 Proton decoupling One well-known and often used technique employed in conjunction with MAS is proton decoupling, where the influence of protons surrounding the measured species, usually carbon or lithium, is inactivated or saturated by a 90° pulse train at the proton frequency, a so-called broad band decoupling. The resulting spectral line is then free from proton dipolar interactions and is considerably narrowed. Another advantage of this technique is that the heteronucleus directly bonded to a proton will be affected much more than other sites that are not, therefore facilitating spectral assignments. Decoupling protons from the 7Li MAS spectrum has for example enabled Liang et al.98 to study interaction of lithium ions with other constituents of a gel polymer electrolyte, namely with plasticizer propylene-carbonate (PC), acrylonitrile (AN) segments and PEO side chains (PEGMEN). Figure 7.13

Site 2

Site 1

Site 3

(d)

(c) (b) (a) 6

4

2

0 ppm

–2

–4

–6

7.13 Deconvolution of 7Li proton-decoupled MAS NMR spectra at 223 K for AP2/PC 50:50 wt% doped with (a) 0.25 mmol LiClO4 g−1, (b) 0.5 mmol LiClO4 g−1, (c) 1.0 mmol LiClO4 g−1, and (d) 2.0 mmol LiClO4 g−1 polymer. From: Y-H. Liang, C-C. Wang, C-Y. Chen, Journal of Power Sources 2008, 176, 340–6.98



301

shows deconvolution of the resulting spectral lines for various compositions of the electrolyte. According to older works on similar systems,99–102 site 1 is associated with coordination between PC and Li+ ions, site 2 is attributed to coordination between oxygen atom on the ether groups and Li+ ions, while site 3 is related to coordination between C≡N groups and Li+. At a doping level of 0.25 mmol LiClO4 g−1 polymer (Fig. 7.13a), site 2 showed the highest intensity resonance peak in the system, indicating that Li+ ions are preferentially coordinated to PEO, because of the higher donor number of its oxygen atoms. Also, peak intensity for site 1 increases as a function of salt concentration. Therefore, the excess Li+ ions are shown to coordinate to PC. It was further shown that these different coordination situations correlate with both thermal properties and ionic conduction of the electrolytes.

7.8.2 Cross-polarization MAS is often used in combination with another double resonance technique called cross-polarization (CP). This technique makes use of the large magnetization of the protons (or in some cases 19F) that can be imparted to the heteronucleus, most commonly 13C, thus enhancing the sensitivity of the heteronuclear signal. This tool has made 13C NMR spectroscopy a routine endeavor without the need for isotopic enrichment. An additional benefit is that the longitudinal relaxation time of the protons, which is usually shorter than that of 13C is the parameter that determines the cycle time of the experiment. CP is achieved by the so-called Hartmann–Hahn condition for example, for 1H → 13C, which is γ1B1 = γ13B13, where γ is the gyromagnetic ratio, B is the amplitude of the rf field and the subscript refers to either 1H or 13C. The signal is significantly enhanced and the spectra are then further resolved by proton decoupling during detection.103 13 C CP/MAS NMR spectra were carried out at different LiClO4 concentrations to analyze the effect of the salt uptake on the polymeric chain of a polyether–siloxane hybrid and to study the chemical process involved.104 Figure 7.14 shows the 13C CP/MAS NMR spectra of the polymer doped with various amounts of LiClO4. Three major peaks at ca. 70 ppm, which are associated with the polyether backbone moieties are shown. The individual peaks are assigned to the differently bonded carbon atoms of the polymer. Upon addition of LiClO4 salt line broadening and the displacement to lower ppm values are observed for the band corresponding to methylene carbons in the PEO block (see the figure detail). This indicates that the presence of lithium salts causes a more broad distribution of the PEO-segment environments and/or reduces the segmental motion of the polymer chains; the latter results from the electronic interaction between


302


10

20

30

50 78 76 74 72 70 ppm

80 25 20 15 10 ppm

Salt-free 90 80 70 60 50 40 30 20

ppm

7.14 13C CP/MAS spectra of EDS-LiClO4 complexes doped with various salt concentrations (O/Li+). The ether carbon (ca. 70 ppm) and methyl carbon (18 ppm) regions are enlarged as shown in the inset. From: W-J. Liang, Y-P. Chen, C-P. Wu, P-L. Kuo, J. Phys. Chem., 2005, 109, 24311–18.104

the Li+ cations and the ether oxygens in PEO segments. However, this interaction is less pronounced for the peaks of methylene carbons adjacent to ether oxygens (76 and 74 ppm) and methyl groups (18 ppm) on the polypropylene (PPO) segments until the salt concentration of O/Li+ = 10. Such behavior coincides with the earlier differential scanning calorimetry (DSC) results that the abrupt rise of the rate of Tg1 increases at the salt concentration of O/Li+ = 10. A similar experiment was used to characterize interactions between polymer and salt in LiClO4–PEO–poly(acrylonitrile) (PAN) hybrid polymer electrolyte.105 A broadening of the corresponding band in the observed spectra upon PAN addition was determined to indicate that a more complicated distribution of the PEO segment environments or a reduction of the segmental motion of the PEO chains results.



303

7.8.3 Rotational echo double resonance nuclear magnetic resonance Despite the successes of the different NMR experiments described above in which many general structural and dynamic aspects of polymer electrolytes have been elucidated, so far none of these have been able to provide detailed information on internal structures of the electrolyte materials. Information about the short-range order on a length-scale of 1–2 Å is usually extracted from the chemical shift values obtained from static or MAS NMR. For a detailed account of the structural motifs on intermediate length scales (~4–8 Å) another approach is to use dipolar double resonance techniques such as REDOR (rotational echo double resonance) and its variants. With REDOR, the structural features are evaluated via the determination of internuclear distances between two nuclei I and S, measuring the heteronuclear dipolar coupling between these. Recently, such NMR experiments106–108 have been reported showing detailed local structure information of lithium in crystalline polymer electrolyte materials that were successfully compared with previously published X-ray diffraction results, for example 6 : 1 PEO : LiPF6 or 3 : 1 PEO : LiTf (LiTf = Lithium trifluoromethane sulfonate) samples.46 REDOR was also used to examine the microcrystalline domains of the 20 : 1 PEO : LiTf sample and it was revealed that these domains have the same local lithium environment as the completely crystalline 3 : 1 PEO : LiTf sample.46 Also, dipolar coupling constants were determined using a method, in which computer simulations with the so-called SIMPSON program109 were used to create a theoretical REDOR dephasing curve. Simulations involving one 13C and one 7Li were compared with the first four data points to determine the primary lithium–carbon dipolar-coupling constant. From this dipolar-coupling constant the average primary Li–C distances were calculated using equation D = (µ0 n) / 4π) [(γ Iγ S) / r 3]

[7.11]

where D is the dipolar-coupling constant, µ0 is the permittivity of free space, γ I the magnetogyric ratio of the dephasing nuclei (7Li), γ S the magnetogyric ratio of the observed nuclei (13C), and r the internuclear distance. 13 C–{7Li} REDOR experiments applied to polymer electrolytes have been conducted by Reichert et al.,110 who were able to determine carbon– lithium distances in crystalline PEO6LiPF6. In another study, Wickham et al.111 have used the REDOR approach to investigate the microscopic structure of the crystalline domains in PEO20LiTf. The interatomic distances obtained from solid state NMR were reported to be in excellent agreement with the values from X-ray diffraction.


304


However, in the presence of motion, the efficacy of REDOR is severely limited: the dynamic processes average the dipolar interaction, consequently leading to biased, overestimated interatomic distances. Thus, when studying amorphous polymer electrolytes or the amorphous parts of a heterogeneous system, REDOR NMR spectroscopy must be performed at temperatures below the Tg, where all the mobility is suppressed.112

7.9

Two-dimensional methods

Two-dimensional (2D) spectra show either chemical shifts along both x- and y-axes, the so-called 2D correlated spectra, or J-coupling constants, for example, can be plotted against chemical shifts (2D J-resolved spectrum). The spectra can show the same nucleus, usually protons, or they can be heteronuclear, the second axis showing 13C, in most cases. 2D methods are based on couplings between nuclear dipoles or transfer of magnetization by chemical exchange. A wide variety of associated techniques exists and is readily used. Only a small number of these will be mentioned. Judeinstein et al.113 have conducted direct measurement of through-space NMR interactions that provide definitive evidence for spatial proximity of different species. Dipole–dipole interactions can be measured in principle between any NMR active nuclei with heteronuclear correlation experiments in the liquid or solid state.114 The dipole–dipole interactions decay quickly with the internuclear distances (r−3), and are difficult to evaluate for long-range distances and even more difficult when exchange, conformation, or motion phenomena are present. However, the measurement of the nuclear Overhauser method115 based on the dipole–dipole-induced crossrelaxation, was proposed to successfully measure intermolecular interactions116 and the formation of ion pairs.117,118 In agreement with recent studies, the pulsed field gradient enhanced inverse HOESY (heteronuclear Overhauser enhancement spectroscopy) sequence is usually preferred because it is more sensitive for isotope pairs 1H–7Li and also improves the digital resolution in the 1H crowded spectrum.119 Figure 7.15 presents the 1H–7Li HOESY 2D spectra for lithium triflate salt dissolved inside a block oligomer C5H11NHCONH(CH2CH2)11NHCO NHC5H11 which can be considered as a liquid model compound of the PEO/ SiO2 nanocomposites described in previous studies.120–122 The major correlation peak corresponds to the proximity of Li+ cations with the ethylene oxide segments of the heterogeneous solvent, which is obviously the most solvating moiety of these molecules. However, the small size of these segments allows some spatial proximity of the alkyl chains. Similar studies could also be used in heterogeneous and gelled SPE to observe the preferential solvating of the different constituents.123 This HOESY methodology was also used to study the ion-pairing between Li+



305

OCH2 CH2

NH 1

H

ppm

7

Li

8.5 9.0 9.5 10.0 10.5 11.0 11.5 12.0 7

6

5

4

3

2

1 ppm

7.15 Heteronuclear Overhauser spectroscopy. 2D 1H-7Li correlation between the Li+ cation and the copolymer evidencing the preferential solvatation by ethylene oxide segments. From: P. Judeinstein, D. Reichert, E. R. deAzevedo, T. J. Bonagamba, Acta Chim. Slov., 2005, 52, 349–60.113

cation and fluorinated anions, tetrafluoroborate or hexafluoroborate (BF4− or PF6−) in molten PEO.

7.10

Exchange nuclear magnetic resonance

The exchange NMR method124 is another class of experiments, described by Judeinstein et al.,113 that can be used to study molecular motions which has the advantage of being much more model independent. The principles of these experiments rely on the orientation dependence of the NMR interactions for solid materials. The chemical shift in a solid is described by a 3 × 3 tensor, thus the NMR frequency of a 13C nucleus (for example) in a specific molecular site depends on the orientation of the molecule. This is known as chemical shift anisotropy (CSA) and, for amorphous or polycrystalline samples, the NMR spectrum defined by such interaction has a characteristic shape known as NMR powder patterns (Fig. 7.16a) when all orientations are present in the sample.125 The basic idea of an exchange


H

H

B0

O O H 13C H H H O

(b) –10 Siloxane/PEO 10 hybrid

HH 13

C O H O H

110 90

70

50

Isotropic motions PEG (c) B A Siloxane structures Restrict dynamics

30

H 13

C H O

A

50

–55 °C

70 B

90

10 ppm

Simulation

30

ω1 (ppm)

(a)


tm = 200 ms

110 110 90 70 50 30 10 –10 w 2 (ppm)

90 70 50 30 10 –10 w 2 (ppm)

(d) tm = 200 ms

0.4 E (tm,δtCSA)

306

0.3

tm = 200 ms –55 °C

–75 °C

0.2 Isotropic motion

0.1 0.0

Isotropic motion

Small angle rotations (s = 7°)

Small angle rotations (s = 7°)

0 0.5 1.0 1.5 2.0 2.5 0 0.5 1.0 1.5 2.0 2.5 3.0 2tCSA (ms) 2tCSA (ms)

7.16 13C NMR of SiO2-PEO hybrid materials near the glass transition temperature: (a) powder pattern solid-state spectrum; (b) 13C 2D-exchange: off-diagonal signals correspond to carbon atoms experiencing changes of their orientation during the mixing time tm; (c) 1D-PUREX : bimodal curves evidence the simultaneous presence of small and high angle motions; (d) multiphasic dynamical character of PEO. Reproduced from E. R. deAzevedo, D. Reichert, E. L. G. Vidoto, K. Dahmouche, P. Judeinstein, T. J. Bonagamba, Chem. Mater., 2003, 15, 2070–8, by permission of The American Chemical Society.

NMR experiment is to take advantage of this orientation dependence for obtaining information about possible molecular reorientations. This is achieved by correlating the chemical shift frequencies of a nuclear spin observed at two different periods of time (t1 and t2) separated by a long mixing time tm (generally in the range 50–500 ms), during which motioninducing changes in the NMR frequency can occur. The usual way of correlating molecular motion with chemical shifts is by acquiring a 2D spectrum that contains the NMR frequencies before and after tm along the dimensions (ω1) and (ω2), respectively. Then, if no change in the molecular orientation occurs during tm (no molecular motions), the 2D spectrum is fully diagonal (ω1 = ω2). Changes in the molecular orientations will produce off-diagonal spectral intensities, and the final shape of the 2D spectrum will depend directly on the geometry of the molecular motion, as shown in Fig. 7.16(b) (left). In more complicated cases, the analysis of these 2D spectra can be performed using simple numerical simulations, allowing the extraction of the motion geometry, as in Figure 7.16(b) (right).



307

As an example, 13C exchange NMR spectroscopy experiments have been performed on siloxane /PEO hybrid materials to get more accurate information about the polymer segmental dynamics.126 The 2D exchange spectrum shown in Fig. 7.16(b) (left) has two clearly distinct regions. In region A (see Fig. 7.16c), attributed to the carbons located in segments near the silica clusters, the 2D spectrum is purely diagonal, indicating that these segments are immobile or execute very small amplitude motion, within milliseconds on the second timescale. In contrast, off-diagonal intensities are observed in region B, which accounts for carbons in the PEO units, showing that a significant fraction of these segments execute large amplitude motions. Further experiments, named 1D Pure Exchange (PUREX),127 were also performed in the same study, allowing measurement of the amplitude of these motions and the fraction of carbons involved. In these experiments, instead of correlating the NMR frequencies before and after tm in a 2D spectrum, the pulse sequence produces a modulation in the spectral intensity in such a way that depends on the difference between the frequencies before and after tm, (ω1 − ω2). Because the increase of signal intensity as a function of the total evolution time under CSA, τCSA, depends on the magnitude of (ω1 − ω2), which is proportional to the reorientation angle, information about the reorientation angles can be obtained in this 1D method. This technique is more sensitive to small amplitude motions than the 2D exchange experiment. Figure 7.16(d) shows this dependence for the 13C belonging to PEO groups in siloxane /PEO hybrid at −75 °C and −55 °C. The curves present a bimodal behavior, the fast increase associated with the large amplitude and angular motion (isotropic motion) and the slow increase with restricted motion. This shows that, besides the large angle motions observed in the 2D exchange, some of the PEO units also execute small angle motions at temperatures around the PEO glass transition temperature. Therefore, exchange NMR clearly evidences the short reorientational behavior of the alkyl spacer and the larger motion amplitude of the PEO segments.

7.11

Electrophoretic nuclear magnetic resonance

A novel application of NMR spectroscopy is electrophoretic NMR (ENMR), where a PGSE diffusion experiment is undertaken on a sample in an electrochemical cell in which electric current is flowing, i.e. there is also an electric field present.128 The NMR pulse sequence illustrating this method is depicted in Fig. 7.17. The resulting echo decay follows a relationship given by S ( 2τ ) = M0 e

−

2τ T2

2 2γ 2 ( t + 2 δ 1

e − Dg δ

3) igγδ vt1

e


[7.12]

308

Polymer electrolytes 90°x

180°y Δ rf and gradient pulses

t1

gg

t2 t1

d

d

Time

Electric field pulse

EDC

FID

Echo

t

t

7.17 Pulse sequence for ENMR experiment (FID = free induction decay).

where the three exponential terms are respectively due to spin–spin relaxation, self-diffusion, and electromigration. The third term contains information about the mobility of the ion, which is µ = v/E, where v is the velocity and E is the electric field strength. Because the mobility (and velocity) is charge dependent, via its sign, this allows unambiguous evaluation of ion transport numbers when there is more than one charge carrier. This technique has been used to determine transference numbers in gel electrolytes in a manner generally considered more reliable and less ambiguous than electrochemical methods, yielding net transference numbers regardless of the extent of ion association and is thus well suited for salt concentration studies as shown by Dai and Zawodzinski.129

7.12

Conclusions

Many research groups within the polymer electrolyte field have benefited from the great power and vast number of possible applications of the NMR technique. As a stand-alone method which yields both structural and dynamical information at the ionic and molecular level, or more commonly when used in conjunction with electrical and thermal characterization,14,46,57,98,112,113,130–133 the contribution of NMR spectroscopy to this field is firmly established and will provide continued understanding as new polymer electrolytes are developed.



7.13

309

References

1. a. e. derome, Modern NMR Techniques for Chemistry Research, Pergamon Press, Oxford, 1993 2. h. friebolin, Basic One- and Two-Dimensional NMR Spectroscopy, WileyVCH Verlag GmbH, Wienheim, 1998 3. k. j. d. mackenzie, m. e. smith (eds), Multinuclear Solid-state NMR of Inorganic Materials, Pergamon Materials Series, Pergamon, Oxford, 2002 4. b. scrosati, Nature 1995, 373 557–8 5. p. v. wright, Br. Polymer J. 1975, 7, 319 6. m. b. armand, Ann Rev. Mater. Sci. 1986, 16, 245–61 7. c. a. vincent, Prog. Solid St. Chem. 1987, 17, 145–261 8. p. g. bruce (ed), Solid State Chemistry, Cambridge University Press, Cambridge, 1995 9. f. m. gray, Solid Polymer Electrolytes: Fundamentals and Electrochemical Applications, VCH Publishers, Inc., New York, 1991 10. b. scrosati (ed), Applications of Electroactive Polymers, Chapman & Hall, London, 1993 11. z. florjanczyk, e. zygadło-monikowska, w. bzducha, Electrochim. Acta 2000, 45, 1203 12. z. florjanczyk, e. zygadło-monikowska, e. rogalska-jonska, f. krok, j. r. dygas, b. misztal-faraj, Solid State Ionics 2002, 152/153, 227 13. a. błazejczyk, w. wieczorek, r. kovarsky, d. golodnitsky, e. peled, l. g. scanlon, g. b. appetecchi, b. scrosati, J. Electrochem. Soc. 2004, 151, A1762 14. m. kalita, m. bukat, m. ciosek, m. siekierski, s. h. chung, t. rodriguez, s. g. greenbaum, r. kovarsky, d. golodnitsky, e. peled, d. zane, b. scrosati, w. wieczorek, Electrochim. Acta 2005, 50, 3942 15. p. periasamy, k. tatsumi, m. shikano, t. fujieda, y. saito, J. Power Sources, 2000, 88, 269–73 16. a. m. christie, l. christie, c. a. vincent, J. Power Source 1998, 74, 77–86 17. y. dai, y. wang, s. g. greenbaum, s. a. bajue, d. golodnitsky, g. ardel, e. straus, e. peled, Electrochim. Acta 1998, 43, 1557 18. m. forsyth, d. r. macfarlane, a. best, j. adebahr, p. jacobsson, a. j. hill, Solid State Ionics 2002, 147, 203 19. a. c. bloise, c. c. tambelli, r. w. a. franco, j. p. donoso, c. j. magon, m. f. souza, a. v. rosario, e. c. rereira, Electrochim. Acta 2001, 46, 1571 20. p. mustarelli, c. capiglia, e. quartarone, c. tomasi, p. ferloni, Phys. Rev. B 1999, 60, 7228 21. s. h. chung, y. wang, s. g. greenbaum, m. marcinek, l. persi, f. croce, w. wieczorek, b. scrosati, J. Phys.: Condens. Matter 2001, 13, 11763 22. c. c. tambelli, a. c. bloise, a. v. rosário, e. c. pereira, c. j. magon, j. p. donoso, Electrochim. Acta 2002, 47, 1677 23. y. masuda, m. seki, m. nakayama, m. wakihara, h. mita, Solid State Ionics 2006, 177, 843 24. a. błazejczyk, m. szczupak, p. cmoch, w. wieczorek, r. kovarsky, d.golodnitsky, e. peled, l. g. scanlon, g. b. appetecchi, b. scrosati, Chem. Mater. 2005, 17, 1535 25. j. saunier, w. gorecki, f. alloin, j. y. sanchez, J. Phys. Chem. B 2005, 109, 2487–92


310


26. e. zygadło-monikowska, z. florjanczyk, a. tomaszewska, m. pawlicka, n. langwald, r. kovarsky, h. mazor, d. golodnitsky, e. peled, Electrochim. Acta 2007, 53, 1481–9 27. y. zheng, j. liu, y.-p. liao, g. ungar, p. v. wright, J. Power Sources 2005, 146, 418–22 28. z-l. wang, z-y. tang, Electrochim. Acta 2004, 49, 1063–8 29. l-y. tian, x-b. huang, x-z. tang, European Polym. J. 2004, 40, 735–42 30. j. stygar, a. biernat, a. kwiatkowska, p. lewandowski, a. rusiecka, a. zalewska, w. wieczorek, J. Phys. Chem. B 2004, 108, 4263 31. p. g. hall, g. r. davies, i. m. ward, j. e. mclntyre, Polym. Commun. 1986, 27, 100 32. j. stygar, g. zukowska, w. wieczorek, Solid State Ionics 2005, 176, 2645 33. s. panero, b. scrosati, s. g. greenbaum, Electrochim. Acta 1992, 37, 1533 34. w. wieczorek, p. lipka, g. zukowska, h. wycislik, J. Phys. Chem. B 1998, 102, 6968 35. m. marcinek, m. ciosek, g. zukowska, w. wieczorek, k. r. jeffrey, j. r. stevens, Solid State Ionics 2004, 171, 69 36. d. brandell, a. liivat, a. aabloo, j. o. thomas, J. Mater. Chem. 2005, 15, 4338 37. p. johansson, p. jacobsson, Solid State Ionics 2004, 170, 73 38. d. brandell, a. liivat, h. kasemagi, a. aabloo, j. o. thomas, J. Mater. Chem. 2005, 15, 1422 39. a. plewa, m. kalita, m. siekierski, Electrochim. Acta 2007, 53, 1527–34 40. n. bloembergen, e. m. purcell, r. v. pound, Phys. Rev. 1948, 73(7), 679–712 41. å. wendsjö, j. lindgren, b. e. mellander, Solid State Ionics 1993, 59, 271–7 42. a. vallee, s. besner, j. prud’homme, Electrochim. Acta 1992, 37, 1579 43. t. m. connor, a. hartland, J. Polym. Sci: Polym. Phys. Ed. 1969, 7, 1005 44. j. j. dechter, J. Polym. Sci.: Polym. Lett. Ed. 1985, 23, 261 45. a. johansson, j. tegenfeldt, Macromolecules 1992, 25, 4712–15 46. j. r. wickham, r. n. mason, c. v. rice, Solid State Nuclear Magnetic Resonance 2007, 31, 184–92 47. j. spevacek, Macromol. Rapid Commun. 1999, 20, 435 48. s. l. maunu, k. soljamo, m. laantera, f. sundholm, Macromol. Chem. Phys. 1994, 195, 723 49. j. spevacek, j. brus, Macromol. Symp. 2003, 203, 111 50. r. frech, s. chintapalli, p. g. bruce, c. a. vincent, Macromolecules 1999, 32, 808 51. y-p. liao, d. c. apperley, j. liu, y. zheng, p. v. wright, Electrochim. Acta 2007, 53, 1444–54 52. th. j. singh, s. v. bhat, J. Power Sources 2004, 129, 280–7 53. m. j. reddy, p. p. chu, J. Power Sources 2002, 109, 340 54. m. j. reddy, p. p. chu, J. Power Sources 2004, 135, 1 55. m. j. reddy, p. p. chu, u. v. subba rao, J. Power Sources 2006, 158, 614–19 56. j-d. jeon, s-y. kwak, J. Membrane Science 2006, 286, 15–21 57. j-d. jeon, s-y. kwak, Macromolecules 2006, 39, 8027–34 58. h. a. every, f. zhou, m. forsyth, d. r. macfarlane, Electrochim. Acta 1998, 43, 1465–9 59. c-l. lin, h-m. kao, r-r. wu, p-l. kuo, Macromolecules 2002, 35, 3083–96 60. s. g. bishop, p. j. bray, J. Chem. Phys. 1968, 48, 1709–17 61. c. berthier, w. gorecki, m. minier, m. b. armand, j. m. chabagno, p. rigaud, Solid State Ionics 1983, 11, 91 62. s. h. chung, k. r. jeffrey, j. r. stevens, J. Chem. Phys. 1991, 94, 1803



311

63. j. p. donoso, t. j. bonagamba, h. c. panepucci, l. n. oliveira, w. gorecki, c. berthier, m. armand, J. Chem. Phys. 1993, 98, 10026 64. a. johansson, j. tegenfeldt, J. Chem.Phys. 1996, 104, 5317 65. s. h. chung, y. wang, s. g. greenbaum, d. golodnitsky, e. peled, Electrochem. Solid-State Lett. 1999, 2, 553 66. d. golodnitsky, e. livshits, r. kovarsky, e. peled, s. h. chung, s. suarez, s. g. greenbaum, Electrochem. Solid-State Lett. 2004, 7–11, A412–A415 67. d. golodnitsky, e. livshits, a. ulus, z. barkay, i. lapides, e. peled, s. h. chung, s. greenbaum, J. Phys. Chem. 2001, A 105, 10098 68. z. gadjourova, y. g. andreev, d. p. tunstall, p. g. bruce, Nature 2001, 412, 520 69. a. abragam, Principles of Nuclear Magnetism, Oxford Science Publications, Oxford, 1996 70. n. c. mello, t. j. bonagamba, h. panepucci, k. dahmouche, p. judeinstein, m. a. aegerter, Macromolecules 2000, 33, 1280–1288 71. h. y. carr, e. m. purcell, Phys. Rev. 1954, 94(3), 630–8 72. p. c. slichter, Principles of Magnetic Resonance, Springer-Verlag, Berlin, 1990 73. e. o. stejskal, j. e. tanner, J. Phys. Chem. 1965, 42, 288 74. r. kimmich, w. uncath, g. schnur, e. rommel, J. Magn. Reson. 1991, 91, 136 75. d. e. demco, a. johansson, j. tegenfeldt, J. Magn. Reson, 1994, 110, 183 76. j. adebahr, m. forsyth, p. gavelin, p. jacobsson, g. oradd, J. Phys. Chem. B 2002, 106, 12119 77. a. johansson, a. gogoll, j. tegenfeldt, Polymer 1996, 37, 1387 78. w. gorecki, m. jeannin, e. belorizky, c. roux, m. armand, J. Phys.: Condens. Matter 1995, 7, 6823 79. k. hayamizu, y. aihara, w. s. price, J. Chem. Phys. 2000, 113, 4785 80. k. hayamizu, y. aihara, w. s. price, Electrochim. Acta 2001, 46, 1475 81. w. gorecki, c. roux, m. clemancey, m. armand, e. belorizky, Chem. Phys. Chem. 2002, 7, 620 82. y. saito, k. hirai, s. murata, y. kishii, k. kii, m. yoshio, t. kato, J. Phys. Chem. B 2005, 109, 11563–71 83. h. kataoka, y. saito, t. sakai, s. deki, t. ikeda, J. Phys. Chem. B. 2001, 105, 2546 84. h-m. kao, p-c. chang, s-w. chao, c-h. lee, Electrochimica Acta 2006, 52, 1015–27 85. i. nicotera, l. coppola, c. oliviero, m. castriota, e. cazzanelli, Solid State Ionics 2006, 177, 581–8 86. m. j. williamson, h. v. st. a. hubbard, i. m. ward, Polymer 1999, 40, 7177 87. c. capiglia, y. saito, h. kataoka, t. kodama, e. quartarone, p. mustarelli, Solid State Ionics 2000, 131, 291 88. b. corain, a. a. d’archivio, l. galantini, s. lora, a. a. isse, f. maran, Chem. Eur. J. 2007, 13, 2392–401 89. c. capiglia, y. saito, h. yamamoto, h. kageyama, p. mustarelli, Electrochim. Acta. 2000, 45, 1341 90. i. m. ward, m. j. williamson, h. v. hubbard, j. p southall, g. r. j. davies, Power Sources 1999, 81–82, 700 91. f. croce, g. b. appetecchi, l. perci, b. scrosati, Nature 1998, 394, 456 92. f. zhou, d. r. macfarlane, m. forsyth, Electrochim. Acta 2003, 48(12), 1749 93. k. hayamizu, y. aihara, s. arai, w. s. price, Electrochim. Acta 2000, 45, 1313 94. c. chauvin, f. alloin, p. judeinstein, c. p. del valle, j-y. sanchez, Chem. Phys. Chem. 2006, 7, 1921–9 95. s. wong, d. b. zax, Electrochim. Acta 1998, 42, 3513


312


96. j. zwanziger, s. m. depaul, r. ulrich, u. wiesner, h. w. spiess, J. Am. Chem. Soc. 1999, 121, 5727 97. g. ardel, d. golodnitsky, e. peled, y. wang, g. wang, s. bajue, s. greenbaum, Solid State Ionics 1998, 113–115, 477 98. y-h. liang, c-c. wang, c-y. chen, J. Power Sources 2008, 176, 340–6 99. w. h. hou, c. c. wang, c. y. chen, Polymer 2003, 44, 2983–91 100. b. huang, z.wang, l. chen, r. xue, f. wang, Solid-State Ionics 1996, 91, 279–84 101. y. h. liang, c. c. wang, c. y. chen, J. Power Sources 2005, 148, 55–65 102. w-j. liang, p-l. kuo, Macromolecules 2004, 37, 840–5 103. d. marks, s. vega, J. Magnetic Resonance, Series A 1996, 118(2), 157–72 104. w-j. liang, y-p. chen, c-p. wu, p-l. kuo, J. Phys. Chem. 2005, 109, 24311–18 105. c.-g. wu, c.-h. wu, m-1. lu, h-j. chuang, J. Appl. Polym. Sci. 2006, 99, 1530–40 106. t. gullion, Concepts Magn. Reson. 1998, 10, 277 107. t. gullion, j. s. schäfer, J. Magn. Reson. 1989, 81, 196 108. d. reichert, o. pascui, p. judeinstein, t. gullion, Chem. Phys. Lett. 2005, 402, 43 109. m. bak, j. t. rasmussen, n. c. nielsen, J. Magn. Reson. 2000, 147, 296 110. d. reichert, o. pascui, p. judeinstein, t. gullion, Chem. Phys. Lett. 2005, 402, 43 111. j. r. wickham, s. s. york, n. m. rocher, c. v. rice, J. Phys.Chem. 2006, B110, 4538 112. t. k.-j. köster, l. van wüllen, Solid State Ionics 2008, 178, 1879–89 113. p. judeinstein, d. reichert, e. r. deazevedo, t. j. bonagamba, Acta Chim. Slov. 2005, 52, 349–60 114. m. h. levitt, in: Spin Dynamics; Basics of Nuclear Magnetic Resonance, J. Wiley & Sons, Chichester, 2001 115. d. neuhaus, m. williamson, in: The Nuclear Overhauser Effect in Structural and Conformational Analysis, VCH, New York, 1989 116. h. mo, t. c. pochapsky, Prog. Nucl. Magn. Reson. Spectrosc. 1997, 30, 1–38 117. a. macchioni, Eur. J. Inorg. Chem. 2003, 195–205 118. c. hofstetter, t. c. pochapsky, Magn. Reson. Chem. 2000, 38, 90–4 119. w. bauer, Magn. Reson. Chem. 1996, 34, 532–7 120. j. a. chaker, k. dahmouche, c. v. santilli, s. h. pulcinelli, v. briois, a.-m. flank, p. judeinstein, J. Non-Cryst. Solids 2002, 304, 109–15 121. n. c. mello, t. j. bonagamba, h. panepucci, k. dahmouche, p. judeinstein, m. a. aegerter, Macromolecules 2000, 33, 1280–8 122. p. h. desouza, r. f. bianchi, k. dahmouche, p. judeinstein, r. m. faria, t. j. bonagamba, Chem. Mater. 2001, 13, 3685–92 123. j. krˇ íž, s. abbrent, j. dybal, d. kurková, j. lindgren, j. tegenfeldt, a. wendsjö, J. Phys. Chem. A 1999, 103, 8505–15 124. j. jeener, b. h. meier, p. bachmann, r. r. ernst, J. Chem. Phys. 1979, 71, 4546–53 125. m. h. levitt, in: Spin Dynamics; Basics of Nuclear Magnetic Resonance, J. Wiley & Sons, Chichester, 2001 126. e. r. deazevedo, d. reichert, e. l. g. vidoto, k. dahmouche, p. judeinstein, t. j. bonagamba, Chem. Mater. 2003, 15, 2070–8 127. e. r. deazevedo, t. j. bonagamba, k. schmidt-rohr, J. Mag. Reson. 2000, 142, 86–96 128. p. c. griffiths, a. paul, n. hirst, Chem. Soc. Rev. 2006, 35, 134 129. h. dai, t. a. zawodzinski, J. Electroanal. Chem. 1998, 459, 111



313

130. a. l. b. s. bathista, e. r. deazevedo, a. c. bloise, k. dahmouche, p. judeinstein, t. j. bonaga, Chem. Mater. 2007, 19, 1780–9 131. e. zygadło-monikowska, z. florjanczyk, e. rogalska-jonska, a. werbanowska, a. tomaszewska, n. langwald, d. golodnitsky, e. peled, r. kovarsky, s. h. chung, s. g. greenbaum, J. Power Sources 2007, 173, 734–42 132. j. spevacek, j. brus, j. dybal, y. s. kang, Macromolecules 2005, 38, 5083–7 133. h-m. kao, t-t. hung, g. t. k. fey, Macromolecules 2007, 40, 8673–83


8 Molecular dynamics simulations of Li ion and H-conduction in polymer electrolytes D. B R A N D E L L, Uppsala University, Sweden

Abstract: This chapter describes the use of molecular dynamics (MD) simulations to understand structure and transport processes in polymer electrolytes for energy storage and conversion applications. For batteries, the polymer electrolytes studied with MD techniques have generally been of poly(ethylene oxide) (PEO)-type, while the fuel cell polymer electrolytes have been perfluorosulphonic acid (PFSA) membrane materials. The MD methodology, its benefits and its limitations are explained in the chapter, together with a review of some significant MD studies of polymer electrolytes. Key words: computational chemistry, molecular dynamics, poly(ethylene oxide), Li mobility mechanism, perfluorosulphonic acid membranes, Nafion, proton transport, battery, fuel cell.

8.1

Introduction

With advances in computer technology leading to ever faster computers, computational chemistry has become an increasingly reliable tool for investigating systems where experimental techniques still provide too little information. Ultra-fast spectroscopy can be used to follow fast reactions but only at a molecular level. A variety of diffraction techniques can also give detailed information about crystalline structure, but have difficulties monitoring changes at a molecular level. This is why the exponential growth in computer power has led to a corresponding growth in the number of computational chemists and to a variety of different computational techniques available for solving chemical problems: ab initio quantum mechanics (QM), semi-empirical methods, density functional theory (DFT), Monte Carlo (MC) simulations, molecular mechanics (MM), molecular dynamics (MD), etc.1 MD can be considered an especially helpful computational tool for studying polymer electrolytes. Polymers are usually too large for more complex simulation techniques using less approximation. MD gives insight into both structural and dynamical properties, and how these are related to 314 © Woodhead Publishing Limited, 2010

Molecular dynamics simulations of Li ion and H-conduction

315

each other, which is of high significance when analysing ionic transport in these systems. Furthermore, the transport mechanisms usually do not involve redox reactions, which make the approximations of classical MD justifiable.

8.2

Computational chemistry

Since the development of the first computers in the early 1950s, scientists have tried to explore how these machines might be used in chemistry. From the very beginning, the field of computational chemistry focused on either solving complex mathematical problems, typically quantum mechanical, or trying to model the dynamical behaviour of atomic and molecular systems. The boundaries between these two areas have never been well defined and during the last decades, we have seen a convergence between quantum chemistry and dynamical simulations when studying chemical reactions.2 The field of computational chemistry has always been inter- or multidisciplinary: mathematicians have constructed numerical models and algorithms which computer scientists implemented into computer programs; chemists have addressed the atomic and molecular details of the systems; physicists have expanded the view to the electronic level and probed underlying forces; and biologists have used the techniques for describing cellular components. Four main branches can perhaps be distinguished within the computational chemistry community: the computationally expensive methods which try to explore the electronic structure of small systems or systems with fixed crystal structures by quantum mechanical methods (QM, DFT); methods which focus on the atomic structure and dynamics of larger systems but using less complex calculations (MM, MC, MD); methods where larger molecular aggregates are treated as separate units, so-called coarse-grain simulations (such as dissipative particle dynamics, DPD); or – for macroscopic systems – the solution of the physical-chemistry equations of state using solver techniques such as finite element analysis (FEA). Figure 8.1 displays a schematic over these domains. In this chapter, the focus is on one of the second approaches: simulating atomic and molecular interaction with the mathematics of classical mechanics. At least two recent trends within computational chemistry, depending on the increasing computational strength and on algorithm development, can be identified: first, the exploration of the domain between electronic structure calculations and molecular level simulations using methods such as QM/MM, Car-Parrinello, reactive force fields, etc.; second, multiscale modelling, where results from complex level calculations are used as input in more macroscopic approaches in a coupled model.3


316


Time (seconds)

108

Finite element analysis Coarsegrain methods Molecular dynamics

10–9

10–15

Quantum chemistry 10–10

10–8

103 Length (m)

8.1 Different computational chemistry techniques.

8.3

The molecular dynamics methodology

8.3.1 The simulation methodology In reality, atoms and molecules in solid materials are far from static unless the temperature is low; but even at 0 K, vibrational motion remains. For ionically conductive materials, atomic movement is the subject of major interest. MD4,5 allows us to simulate the dynamics of the particles in a welldefined system to gain greater insights into local structure and local dynamics – such as ion transport in solid materials. In an MD simulation, atomic motion in a chemical system is described in classical mechanics terms by solving Newton’s equations of motion: Fi = mi ai for each atom i in a system of N atoms; mi is their respective atomic mass; ai = d2ri/dt2 is their acceleration; and Fi is the force acting upon atom i due to interactions with all other particles in the system. The forces are generated from a universal energy potential E: −dE d 2 ri = Fi = mi 2 dri dt The basic idea of MD goes back to a classical Newtonian idea in physics – that if one knows the location of all the particles in the universe, and the forces acting between them, one is able to predict the entire future. In a normal MD simulation, this universe comprises only some thousands of atoms – or in extreme cases, millions – gathered in an MD simulation box.



317

With Newton’s equations, it is possible to calculate sequentially the locations and velocities of all particles in the system. This generates a sequence of snapshots which constitutes a ‘movie’ of the simulated system on the atomic scale. Owing to the massive computer time necessary to solve these equations for a large number of particles, the movies are generally fairly short – usually within the nanosecond regime. All that is needed to solve the equations of motion are the masses of the particles and a description of the potentials, E. The solution of this set of equations is managed by a computer algorithm. The most common is the so-called ‘leapfrog’, which works stepwise by: • •

calculating the acceleration of all particles at time t from the forces Fi; updating the velocity vi at t + Δt/2 using vi(t + Δt/2) = vi(t − Δt/2) + ai(t)Δt

where Δt is the time-step between two snapshots; normally 0.1–1.0 fs; • calculating the atom position in the snapshot using: ri(t + Δt) = ri(t) + vi(t + Δt/2)Δt The MD simulation method is very straightforward, but one must bear in mind that it is based on some severe approximations. At the highest level, the Born–Oppenheimer approximation is made, separating the wavefunction for the electrons from those of the nuclei. The Schrödinger equation can then be solved for every fixed nuclear arrangement, giving the electronic energy contribution. Together with the nuclear–nuclear repulsion, this energy determines the potential energy surface, E. At the next level of approximation, all nuclei are treated as classical particles moving on the potential energy surface, and the Schrödinger equation is replaced by Newton’s equations of motion. This means that no covalent bonds can be formed or broken in a classical MD simulation, but must be represented as springs. At the lowest level of approximation, the potential energy surface is approximated to an analytical potential energy function which gives the potential energy and interatomic forces as a function of atomic coordinates.

8.3.2 Interaction potentials Since the analytical description of the potential energy surface, the force field, strictly determines the outcome of any MD simulation, it is necessary that this description is as precise as possible. The common methodology is thus to generate specific potentials for the simulated system. These can be generated and fine-tuned in two different ways: empirically or non-empirically.



2Li+/SiF62– complex energy (kcal/mol)

318

R

–160 –200 –240 –280 –320 –360 –400

D2h, DFT D3d, DFT D4h, DFT D2h, FF D3d, FF D4h, FF

1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 Li+–Si distance (Å)

8.2 Development of potentials for Li+–SiF62− interactions. Single-point energies have been obtained by DFT calculations in three different symmetries: D2h, D3d and D4h.

Empirical potentials are derived by fitting the potential expression to macroscopic experimental observables, such as bond length, lattice parameters, bond vibrations, density, pressure, temperature. Such potentials thus reproduce the properties they are modelled on extremely well, but can fail when it comes to other properties. Non-empirical potentials are derived from high-level ab initio calculations. The structural and thermodynamical properties of the system are thus not intrinsically dependent on any experimental quantity, which make comparison with such data a good test for the validity of the model. Single-point energy calculations are used to map the potential energy surface, and the analytical expressions for the potentials are then fitted to reproduce the surface. To get a good description of the energy surface, the analytical expression is usually generated from calculations made on several different geometries and local configurations, and the analytical expression is fitted to all these situations. An example of this procedure for Li–SiF62− interactions can be seen in Fig. 8.2. Polymers, which consist of many covalently bonded fragments, are usually modelled with a set of intramolecular potentials describing direct bonds, bond angles and torsional or dihedral bond angles. Common analytical functions are: Vbond (r ) =

k1 (r − r0 )2 2

Vbend (θ ) =

k2 (θ − θ 0 )2 2 6

Vtorsional (τ ) = ∑ an( −1) cos n τ n

n= 0



319

The intermolecular potentials, on the other hand, are commonly described by electrostatic and two-body interactions in, for example, the Born–Mayer– Huggins form: VB− M − H(r ) =

q1q2 C D + Ae − r B − 6 − 4 4πε 0 r r r

or the Lennard–Jones form: VL − J(r ) =

A C q1q2 − + r 12 r 6 4πε 0 r

where k1, r0, k2, θ0, an, A, B, C and D are constants depending on the interacting atom types involved. The total force field acting on an atom i in the simulation is then the sum of interactions with all other particles in the box: V = ∑ Vbond(r ) + ∑ Vbend (θ ) + ∑ Vtors(τ ) + ∑ VB− M − H(r ) + ∑ VL − J(r ) i ,r

θ ,i

τ ,i

i, j

i ,k

There are also several standard force fields to use which are applicable to almost any chemical system: AMBER, UFF, CHARMM, OPLS-AA, etc. However, force fields developed for specific systems are usually considered to be of higher precision, and the MD results therefore of higher quality. Furthermore, in order to correct for polarization effects some force fields use a core–shell model, where one analytical function describes the forces for the nuclei, and another the forces of the electron sphere surrounding it.

8.3.3 Periodic boundary conditions and thermodynamical ensembles Since the computation time required to calculate the trajectories of all N particles in a simulation box increases with N2, the simulated system can often not be made large enough to accurately represent the bulk properties of an actual crystal or amorphous material, and there is thus a risk that surface effects will be present at the edges of the simulation box. This problem is solved by implementing periodic boundary conditions, in which the simulation box is replicated through space in all directions; see Fig. 8.3. The set of atoms present in the box is thus surrounded by exact replicas of itself, i.e. periodic images. If an atom moves though a boundary on one side of the simulation box, so will its replica on the other side. This keeps the number of atoms in one box constant, and if the box has constant volume the simulation then preserves the density of the system. An MD simulation must also follow the laws of thermodynamics. At equilibrium, it should have a specific temperature, volume, energy, density, pressure, heat capacity, etc. In statistical thermodynamics, this constitutes the state of the system; its ensemble. Since MD is a statistical mechanics


320


8.3 Periodic boundary conditions in two dimensions.

method, an evaluation of these physical quantities can be made from the velocities and masses of the particles in the system, and MD can serve as a link between these atomic-level quantities and macroscopic properties. When performing an MD simulation, one chooses a specific ensemble in which the simulation model is retained. This ensemble then scales the velocities of the particles. There are three commonly used ensembles: •

The microcanonical ensemble (NVE), which maintains the system under constant energy (E) and with constant number of particles (N) in a well-defined box with volume (V). This ensemble is usually appropriate during the initial equilibration phase of a simulation. • The isothermal–isobaric ensemble (NPT), where temperature (T) and pressure (P) are kept constant. This is normally the best model of the experimental conditions for polymer electrolytes. • The canonical ensemble (NVT), where volume and temperature are kept constant. This ensemble facilitates comparisons with experimental data from structures with fixed dimensions.

8.3.4 Equilibration and non-equilibrium molecular dynamics When generating an MD simulation box, the conformation is rarely in a stable position at equilibrium. The box content must therefore go through an equilibration phase before sampling of the statistics can begin. This is especially important for complex glassy systems such as polymers, which can often get entangled if they are randomly generated. The equilibration



321

is often monitored from the energy value of the system, which should be stable before the data-sampling period starts. In order to increase the efficiency of the equilibration, the MD box is sometimes heated up or expanded, before slowly decreasing the temperature or volume back to its normal values. Chemical equilibrium is characterized thermodynamically in terms of uniform pressure, temperature and chemical potentials. Non-equilibrium is characterized by gradients in these variables, leading to a flux in the system which transports mass, momentum and charge along the gradient. This flux serves to destroy the gradient and bring the system to equilibrium. Nonequilibrium systems are thus characterized by mixing and dissipation processes. Such processes arise for example in the discharge of a battery: electronic and ionic motion try to compensate for the difference in electrochemical potential between cathode and anode. This gives rise to an electric field acting over the polymer electrolyte, and non-equilibrium MD, with an electric field gradient applied over the simulation box, can thus be suitable for studying such systems. In non-equilibrium MD, a perturbation is switched on at time t = 0 and is held constant thereafter. The long-term steady state response then yields transport coefficients. Problems occur though, since non-equilibrium systems dissipate heat, leading to an increase in the temperature of the system. The field also gives rise to an undesirable material drift across the whole simulation box. These problems can be overcome by constantly rescaling the velocities to maintain some desired temperature, or by fixing certain atoms to hinder this drift. However, it is hard to justify the physicality of these constraints. In a system where polymer-chain relaxation and ionic motion are critical to the properties of interest, it is better to allow some drift in the system than to keep parts of it fixed. Moreover, high field values give rise to a non-linear response in different properties in the systems, e.g. the conductivity, making it difficult to compare properties calculated from these simulations with experiment.

8.3.5 Structural properties After the MD simulation has been performed, the statistics provided are used to calculate different properties relating to structural and dynamical behaviour: this, and its chemical interpretation, constitute the analysis phase of an MD study. One of the most important properties extracted from the MD simulation is the pair radial distribution function (RDF). It is a function, usually written ga. . .b(r), which presents the probability of finding a particle of type b at a distance r from particle of type a. In a perfect crystal without thermal motion, the RDF would appear as periodically sharp peaks, which gives information about both short-range and long-range order in


322


Infinite RDF Infinite CNF Nematic RDF Nematic CNF Smectic RDF Smectic CNF

gLi...O (r)

20

15

10

5

0

0

1

2

3

4

5

r/A

8.4 Radial distribution function (RDF) and coordination number function (CNF) for Li–O interactions in LiPF6PEO6 generated by MD simulations. ‘Smectic’ and ‘nematic’ refer to different polymer endgroup arrangements, while ‘infinite’ represent polymers without endgroups. It is seen from the graphs that the preferred Li–O distance is ~2 Å, and that the Li–O CN is 5 or 6 depending on chain configurations.

the system. In liquids or glassy systems however, only short-range order exists and after some initial peaks at short distance – corresponding to the first coordination shells – the RDF fluctuates around 1. An example is given in Fig. 8.4. The RDF can be calculated by counting the number of atom pairs within some distance range, and averaging this over a number of time-steps and particle pairs: M

∑ N (r k

gab(r ) =

, Δr )

ab

k =1

M

( )ρ 1 N 2

V (rab, Δr )

where Nk is the number of atoms found at time k in a spherical shell of radius r and thickness Δr; and ρ is the average system density, N/V, of a given atom type. Integrating this RDF over r gives the coordination number function (CNF), which is the average coordination number of particle type a to particle type b at distance r (see Fig. 8.4). The RDF can be compared directly with experimental data from X-ray or neutron diffraction, and can thus be used as a check on the reliability of the potentials in many systems. In chemically complex systems such as polymer electrolytes, average atomic distances calculated as RDFs can be too rough a measurement to



323

8.5 The operation of ‘folding’ the content of the polymer electrolyte MD box back onto a crystallographic asymmetric unit.

capture all the structural information available. The spatial arrangement of atoms can also be of major interest. These can be captured in either the spatial distribution function (SDF), which also takes the angular component of the atomic distribution into account, or be obtained by calculating bond angle and dihedral angle distributions in the polymers. For polymeric systems, the radius of gyration (Rg) and end-to-end distance are easily calculated from MD data, and can contribute to vital information on the system’s chemistry. Many polymeric blends or polymers in solvent (for example fuel cell membranes) experience a certain degree of phase separation, which can also be monitored from MD data. By dividing the MD box in a fine subgrid, investigating the content in every part of the grid, and comparing the occurrence frequency of different atomic species with a random distribution, a quantitative measurement of the phase separation can be obtained. For crystalline systems, one way to study the simulated structure and to compare it with experiment is to ‘fold’ the atom positions back onto the crystallographic asymmetric unit (Fig. 8.5). This is done by applying the symmetry operations of the specific space group in combination with translations. Doing this for several time-steps generates a distribution of atomic positions within the asymmetric unit, which can then be compared with crystallographic displacement parameters. The isotropic mean-square thermal displacement parameter (Uiso) for a given atom in the asymmetric unit is calculated from its mean-square displacements, σα2, using:

σ α2 =

1 N ∑ rα ,k − rα N k =1

2


324

Polymer electrolytes 3000 Water molecules Hydronium ions

MSD (Å2)

2500 2000 1500 1000 500 0

0

500

1000 Time (ps)

1500

2000

8.6 MSD functions for H2O and H3O+ in a Nafion membrane. The slope of the curves is proportional to D, and thus the mobility of H2O is larger than for H3O+.

for α = x, y and z for atom position k. This gives: U iso =

1 2 (σ x + σ y2 + σ z2 ) 3

8.3.6 Dynamical properties The diffusion coefficient D for an atom type in a material can generally be calculated from an MD simulation via the time evolution of its displacement vector: D=

1 N ∑ xi 2 + yi 2 + zi 2 6 Nt i =1

xi, yi and zi are direct results from any MD simulation, and D is thus proportional to the slope of the mean-square displacement (MSD) function’s time-development (see Fig. 8.6). From the diffusion coefficient, the molar conductivity (⌳°m) can be calculated by the Nernst–Einstein equation: Λ m =

F2 (v+ z+2 D+ + v− z−2 D− ) RT

Here, ν+ and ν− are the number of positive and negative ions per formula unit and z+ and z− their respective charges, while F is Faraday’s constant. This is perhaps the most straightforward way of evaluating mobility in a simulated system. Transport numbers can also subsequently be obtained: t+ =

D+ D+ + D−



325

Time correlation functions are perhaps the most convenient tool to study dynamical properties. These relate some property B at a time t to some property A at t0: CAB(t) = 〈A(t0)B(t0 + t)〉t

0

Here, A(t) and B(t) are dynamical variables of the system, for example a specific coordination around a cation and an ion jump. If CAB grows towards unity, there is a maximum correlation between these properties. Correlation functions can thus be used to monitor conduction mechanisms. Correlation can also be calculated independent of time (i.e. at t = 0) – the straightforward probability that B occurs if A occurs can be readily measured from the MD statistics. Besides MSDs and correlation functions, which are based on the MD statistics, several more qualitative approaches can give information on the system behaviour. By simply pinpointing the most mobile ions or polymer segments, many conclusions can be made. One example is time-evolution plots, which show how the coordinating atoms around one specific atom change over the simulation. By correlating these with the changing mobility of ions during a simulation, the transport mechanisms in the systems can be understood.

8.4

Li+-conducting poly(ethylene oxide)-based electrolytes for batteries

The Li ion battery polymer electrolyte systems by far studied most with MD techniques are the poly(ethylene oxide) (PEO)-based, although with many modifications. Only a limited number of MD simulations of other Li ion battery polymer electrolytes have been published. For PEO however, several force fields have been developed. The early force fields were nonpolarizable, i.e. they did not include many-body polarizable interactions. The torsional parameters were usually obtained by fitting analytical expressions to quantum calculations, while the non-bonded parameters were taken from generic force fields. One example is the force field by Neyertz et al.,6 based on MP2/D95+(2df,p)//HF/D95** calculations on diglymes, and validated against crystallographic data.7 de Leeuw and co-workers8 also compared this force field towards neutron spin echo experiments, and found good agreement. Neyertz and Brown also managed to perform early MD simulations of NaI in PEO for both the crystalline and amorphous phases, while at the same time adjusting the force field to reproduce experimental data.9 Müller-Plathe used a standard force field to simulate LiI in PEO;10 however, since there was virtually no ionic movement, the dielectric constant was raised to 3 in order to decrease the strength of the ionic bonds.


326


In recent years, there have been significant improvements in the PEO force field. In several studies, Borodin and Smith developed a polarizable force field for PEO and PEO-based Li ion electrolytes,11–13 using conventional anions: BF4− and TFSI−. After finding that nearly 40% of the Li+ cation/ether oxygen complex energy is due to polarization,14 it was concluded that neglect of this contribution to the complex energy will result in significant reduction of the interaction. The polarizable force field – constructed by different DFT and QC calculations at a high degree of sophistication – was developed on the basis of gas-phase dimer energies, liquid densities, heats of vaporization and alkane self-diffusion coefficients, and validated against experimental data: structure factors, dielectric loss, 13C nuclear magnetic resonance (NMR) spin–lattice relaxation times, and incoherent intermediate structure factors. A substantial number of MD simulations have been performed on amorphous linear PEO of different molecular weight, at different temperatures and with different Li salts: LiCl, LiI, LiPF6, LiBF4, LiCF3SO3, LiClO4 and LiTFSI, at different concentrations.15–23 Generally, it has been found that Li+ is coordinated by consecutive monomers forming a loop around the cations, and that Li+ is perturbing PEO helical formation and slowing down the molecular motion of the polymer chain. Li+ also have a strong tendency to cross-link the polymer; i.e. to coordinate to different parts of the polymer matrix. The cations are rather transported along the chains than between the chains, while the anions often play a significant role in the transfer of cations from one polymer chain to another. Ion pairing is not simply inhibiting ionic transport as sometimes suggested in literature; neutral ion pairs can instead often be seen travelling together. The ion pairs also show a lot of dynamics, with frequent complexation and decomplexation. In a recent study,24 Borodin and Smith put several of these findings into perspective when they discovered that Li+ motion is subdiffusive in amorphous PEO up to 30–40 ns at ambient temperatures. This means that very long MD simulation times are necessary to accurately extract the diffusion coefficients in these systems. Since branched PEO chains have shown promising properties for battery applications,25,26 a few studies have MD simulated PEO with ethylene oxide side-chains. Hektor et al.27 developed a force field for such a system, which was later used by Karo and co-workers.28,29 PEO systems with side-chains of length 3–15 EO units, separated by 5–50 EO units, were simulated for different concentrations of LiPF6. The result showed that at higher concentrations, the Li+ ion transport was promoted by many but short side-chains, while the mobile properties at lower concentrations was highest for intermediate side-chain lengths at intermediate separation – thus pointing towards an ‘optimal’ polymer configuration. The studies also showed that the side-chains play a significant role in the Li ion transport mechanism, and not merely suppressed crystallinity. Borodin and Smith30 also investi-



327

gated a branched PEO system, but with LiTFSI and at elevated temperatures. They showed that the Li+ ions preferred to be located around the polymer backbone, but that the most mobile ions were transferred from side-chain to side-chain. Brandell, Liivat and co-authors were first to make an in-depth MD study of a ionically conductive crystalline polymer system; LiPF6·PEO6.31–35 The crystalline phases of the polymer electrolytes were for a long time regarded as insulators. This view was, however, overturned in 2001 by the demonstration of ionic conductivity in the complexes LiXF6·PEO6 (X = P, As or Sb).36 Although the conductivity was relatively low in these materials, they still showed 10 times higher ionic conductivity than their amorphous counterparts, and it was also shown that the conductivity could be enhanced by doping.37 Furthermore, the transference number for Li, t+, was very high. The LiXF6·PEO6 complexes also display fascinating structures38,39 – the materials are composed of coaxial hemi-helices of PEO, which pairwise form cylindrical channels containing the lithium ions coordinated to ether– oxygen chains; the anions lie outside the hemi-helical pairs, with no direct contact to the lithium ions (see Fig. 8.7).

C Oet H Li P F

Y

Z

8.7 The MD simulated LiPF6·PEO6 structure.


328


The MD simulations displayed some deviations from the experimentally determined structures. It was notable that the Li–O coordination was predominantly 6-fold in the simulations, while a Li coordination number of 5 was found experimentally. The MD-derived Li+ ion coordination corresponded more closely to that found in the equivalent short-chain polymer system (CH3(OCH2CH2OCH3)2·LiSbF6.40 The Li+ ions inside the polymer channels were further equi-spaced, while the experimental geometry involved two Li–Li distances (of 7.4 and 4.4 Å) along the chain. However, despite these differences, the simulated infinite polymer chain model was generally in agreement with the structure suggested from neutron diffraction: the hemi-helical structure and the ion separation were retained. The differences could perhaps arise because the structure determination neglected the high concentration of chain-end groups, which was frequent due to usage of rather short oligomer chains. When chain ends were introduced into the MD simulations, the coordination changed,33 generating a ‘liquid-like’ peak broadening in the RDFs. On plotting the dihedral angles along the entire PEO backbone of the simulation box for the chain-end free system, an obvious pattern emerged, suggesting the existence of an asymmetric unit of the same size as the experimentally determined structure. Some discrepancies appeared, however, between the simulated crystalline LiPF6·PEO6 system and the experimentally determined structure. Firstly, the bond angles had a considerably smaller spread in the simulated system. Some extreme values in the experimental model, e.g. an –OCC– bond angle as low as 85°, lie far from the minimum in the bond-angle force field: the angular bond energy contribution dropped from 217 to 16 kJ/mol during the simulation, implying that the experimentally determined structure may contain some unphysical detail (such as a bond angle at very high energy). The MD studies32 also revealed that inserting a divalent SiF62− anion or a neutral SF6 molecule (with appropriate Li+ compensation) is seen to destabilize the local environment near the dopant. The more highly charged SiF62− anion also repels the neighbouring PF6− ions; an effect which is compensated by the extraction of a Li+ ion from within the polymer hemihelices, to form an Li+–SiF62− ion pair with a net charge of −1. Either form of doping was shown to have a significant effect of both the cation and anion conduction, which clearly benefited from interstitial ions or ion vacancies. In order to stimulate the ionic movement, some simulations were conducted under the influence of an external electric field, i.e. under non-equilibrium conditions. The conductivity was seen as ion jumps parallel to the hemihelical axes. In the infinite-chain systems, the jumps were easily distinguished as long, simultaneous anion jumps, whereby a PF6− ion migrates into the position of its neighbour along an inter-helical anion column, or as shorter lithium ion jumps inside the helices. In the short-chain systems, long and short jumps occur for both ion types. © Woodhead Publishing Limited, 2010


329

MD studies of PEO containing ionic liquids – based on 1-alkyl-3-methylimidazolium cations – was pioneered by Costa and Ribeiro,41,42 where the influence of ionic liquid concentration, temperature and the 1-alkyl-chain length was systematically investigated. The authors showed that the ionic liquid is dispersed in the polymeric matrix, but with ionic pairs remaining in the polymer electrolyte. However, both polymer and anions were clearly able to dissolve the cations. The imidazolium cations were coordinated by both the anions and the oxygen atoms of PEO chains, and the calculated diffusion coefficients showed that cations were more mobile than both anions and PEO chain atoms. The ionic mobility was significantly higher in the systems with longer 1-alkyl-chains, well in agreement with experiment. Time correlation functions also confirmed that the ion pairs in ionic liquid polymer electrolytes are relatively weakly bonded. The promising experimental results of polymer/nanoparticle composites regarding both ionic conductivity and mechanical strengths have triggered some authors to use MD to simulate similar systems in the context of battery electrolytes. For example, Borodin et al. studied a thin layer of PEO sandwiched between TiO2,43 and Aabloo et al. studied the behaviour of PEO near a V2O5 surface.44 The most comprehensive study was perhaps performed by Kasemägi et al.,45–47 who used MD to simulate an entire Al2O3 nanoparticle embedded in a PEO matrix with different Li salts. The simulations showed that both PEO and the salts were stabilized close to the particle surface, but that mobility was promoted further out in the polymer matrix, leading to an overall enhanced ion transport.

8.5

Polymer electrolytes for fuel cells: perfluorosulphonic acid systems

The polymer electrolytes used for low-temperature proton exchange membrane fuel cells (PEMFCs) are fundamentally different from the polymer electrolytes used in batteries. Here, the polymer is a medium for a solvent, normally water, and it is mainly in the solvent that ion transport occurs. The polymer serves several functions, of which the most important is to provide mechanical stability and electrode separation in the fuel cell application. Since the fuel cell needs proton transport from the anode to the cathode, the polymer also contains proton donating groups, often sulphonic acid (–SO3H). The prototype PEMFC membrane materials have been perfluorosulphonic acids (PFSAs), of which the most established membrane material is Nafion® (Fig. 8.8). These consist of hydrophobic teflon –CF2–CF2– backbones, with fluorinated hydrophilic and acidic sidechains; for Nafion –OCF2CF(CF3)OCF2CF2SO3H. The inhomogeneous multiphase system that a PEMFC membrane such as Nafion comprises makes the system very hard to analyse experimentally.48 © Woodhead Publishing Limited, 2010

330


–S –H3O+ –H2O

8.8 A projection of a 76×76×5 Å slice through a Nafion® MD box. Empty encircled regions represent the polymer backbone, while the points represent accumulated coordinates of sulphur and water- and hydronium-oxygen atoms during the simulation. The ‘clouds’ thus represent the dynamic motion of atomic species during the MD simulation. The hydrophobic/hydrophilic interface is highlighted by a thick line.

The size and nature of these regions are matters of continual discussion. Several models for Nafion have been suggested based on diffraction studies: small-angle X-ray scattering (SAXS), small-angle neutron scattering (SANS) and wide-angle X-ray diffraction (WAXD). Some models of the hydrophilic domains involve dispersed spherical objects with characteristic repeat distances between the particle centres, while others maintain that the dominant structural features are individual local clusters. The shape and morphology of the ionic domains has also been disputed: spherical clusters,49 lamellae,50 rods51 and ribbon-like52 structures have been proposed. Perhaps the most widely accepted nanoscale structural model has been suggested by Hsu and Gierke49 (the ‘cluster-network model’), where spherical ionic clusters (diameter: 4–5 nm) are interconnected by narrow water channels (diameter: 1 nm; length: 4–5 nm). This model has recently been challenged, however, by Schmidt-Rohr and Chen,53 who claim that elongated parallel but otherwise randomly packed water channels are surrounded by partially hydrophilic side-chains to form inverted-micelle cylinders. Structural mod-



331

elling, such as MD, can therefore make a vital contribution in this discussion by shedding light upon these issues. A further complexity arises from the proton transport mechanism in the materials. Protons are transferred across the membrane by both vehicular diffusion of H3O+ ions and by the Grotthuss mechanism. The latter involve constant breaking and reformation of hydrogen bonds, which cannot be accurately described by a classical MD force field. QM methods could therefore be considered more suitable for studying proton jumping. Indeed, QM methodology can accurately describe charge transfer processes and the vital hydrogen bond networks. On the other hand, they can only take account of local structure around the side-groups, and at best model small channels and pores. Although outside the scope of this chapter, QM-MM methods have been extensively used for PFSA membranes, in particular the semi-empirical empirical valence bond (EVB) method.54–57 Some other recent reviews on MD simulations of PFSA membranes can be found in Elliott and Paddison,58 Devanathan59 and Elliott.60 Early MD studies of PFSA membranes were based on partially atomistic models, consisting of atomistic side-groups in some specific local geometry.61,62 This allowed conclusions on ion transport within these systems, but could not provide any structural or structure–property information. Some early all-atom MD simulations were not based on PFSA models per se, but focused on less complex analogues: Ennari et al. for example studied sulphonic acid groups attached to PEO.63 Using the DREIDING force field,64 Elliott and co-workers65 simulated sulphonic acid anions in a hydrated system (without the fluorinated backbone), generating the archetype phaseseparated morphology with water-channels, however of somewhat small dimensions. The simulations showed the H3O+ being transported by a jumping mechanism of the ions between the sulphonate groups, but this was neglecting the Grotthuss transport mechanism. Vishnyakov and Neimark simulated an oligomeric Nafion-resembling PFSA system at 298 K, consisting of 10 monomers solvated in water and methanol.66 They found clear differences between the geometries of the backbones in the two different solvents, with the fluorocarbon chain being more folded in water. The side-chains of the polymer were in comparison rather stiff, with only few conformational transitions. No significant clustering into larger water domains could however be observed, probably due to the limited system size. The authors later published simulations of Nafion oligomers with four side-chains, using a careful equilibration technique in both NPT and NVT ensembles and a more sophisticated force field model,67 consisting of a combination of experimental data and ab initio and DFT results. Na+ was used as counterion to the sulphonic groups, and the solvent was a mix of water and methanol. A high mobility in the fluorocarbon backbone was found due to the solvent – simulation in pure water gave rise


332


to only a third of the transitions of dihedral angles. This is well in line with the increased swelling observed experimentally for mixed solvents,68 which has been suggested to originate in an increased flexibility of the polymer backbone chain.69 These simulations were truly a significant step forward, but the extremely short polymer chains made it hard to draw conclusions on the structure–property relationships in the material. Vishnyakov and Neimark thereafter simulated polymeric Nafion,70 with similar force field as in previous studies67 but with K+ as counterion and using a united atom (UA) approach for the –CF2– units. UA models are less rigorous then all-atom force fields, but allow larger systems to be simulated. The MD box consisted of 15 polymers with 10 side-chains each, which were simulated for three different levels of hydration: 5.0 wt%, 12.5 wt% and 17.0 wt% H2O. Despite still having Mw an order of magnitude smaller than real PFSA materials, this is a good approximation at the short timescales used in MD simulations. In these simulations, a micro-scale phase-separation could be observed, with specific hydrophilic (water) and hydrophobic (polymer backbone) domains. According to the RDFs, the phase separation became more pronounced at higher water content. The water clusters, containing up to 100 water molecules, were separated from each other, which made the authors suggest that any connectivity in the PFSA system is very short-lived (~100 ps). However, considering the small size of the MD box and the limited simulation time, it was difficult to reflect the transport processes in real Nafion. Urata and co-workers also used a UA coarse grain model for a MD simulation of Nafion, but developed an all-atom model for the polymer side-chains.71–76 The polymeric configuration was similar to that used by Vishnyakov and Niemark,70 but here hydronium was used as counterion. Again, a nanophase-separated structure was observed. Small clusters of – SO3− groups with bonded water displayed typical spacings between 4.6 and 7.7 Å. An interesting observation was that the side-chains were clearly oriented perpendicular to the hydrophobic–hydrophilic interface, thus with limited side-chain entanglement. The side-chain orientations were also correlated at higher water content. The MSDs showed that the water dynamics was very limited at lower level of hydration, but increased upon higher water content when a smaller proportion of the water was bound by the sulphonate groups. Urata et al. were also first75 with calculating the structure factors S(q) from their RDFs, which can be directly compared with experimental scattering data. These results showed that the ionic clusters were smaller in the MD simulations than in the real material, which the authors explained by the lower Mw. However, this might well be an effect of the small-scale MD box, which later studies would suggest. Jang et al. were the first77,78 to investigate the side-chain sequence on the polymer chain using classical MD. Considering the synthesis methodology



333

of Nafion, it is assumed – yet not proven – that the side-chains are randomly distributed along the backbone, which would result in a statistical distribution of the spacings. The authors investigated two extreme cases of Nafion side-chain distribution on a backbone: one with randomly placed sidechains, and one ‘blocky’ with the side-chains concentrated in two backbone regions. This reflected the potential for improving the membrane transport properties by altering the molecular design. The systems consisted of four such polymer chains with 560 water molecules of the F3C model79 and 40 hydronium ions, resulting in 15 H2O per SO3 group (~20 wt% water), and the simulations were performed using a modified DREIDING force field for Nafion.64 The results showed interesting aspects of the nanostructure of the membranes. The polymers with aggregated side-chains showed a stronger tendency to cluster formation, and thereby with larger hydrophilic domains than the random copolymer, while the S–S distance distribution was similar in both simulated systems. Calculation of structure factors based on the simulations also confirmed that the ‘blocky’ Nafion had a more segregated morphology. The calculated S(q) profiles contained an ionomer peak which corresponded to characteristic distances of 50 Å for the ‘blocky’ polymer, which is in agreement with experimental values for similar water content. However, the random copolymer displayed a peak at 30 Å, which made the authors conclude that there is a significant degree of ‘blockiness’ in real Nafion. Jang et al.77 also analysed the transport properties of protons in their MD boxes. Both calculated diffusion coefficients and activation energies were generally consistent with experimental data, but the proton diffusion is naturally underestimated due to the absence of the Grotthuss transport mechanism. Interestingly, the results showed that while water transport is more restricted in the random copolymer, H3O+ diffusion is more or less independent of monomeric sequence. It was therefore concluded that ‘proton hopping’ plays a significant role in the transport process. There has been some criticism of the study by Jang et al., which address the importance of the equilibration procedure for polymeric systems. Elliott and Paddison58 have argued that the difference in density between the MD simulated system and the experimentally measured values are due to incomplete chain relaxation or an erratic force field. Since later MD simulations of Nafion using similar force field provided good agreement in density,80,81 the latter can probably be ruled out. Jang et al. used a rather complex equilibration process, were the MD boxes were heated, and expanded and compressed several times in the NPT ensemble. The resulting densities (1.60 and 1.67 g/cm3) are 5–10% lower than experimental values, which can partly be explained by the lack of semicrystalline domains in the MD box, which do exist in real membranes and contribute to a somewhat increased density.


334


Cui et al.82 used a UA force field similar to that of Urata et al.,75 but using significantly shorter Nafion chains, while varying the level of hydration (λ; the number of H2O molecules per SO3 group). The simulations showed, similar to experimental observations, that the level of hydration controlled the clustering of water molecules. At low λ, H2O molecules formed small and isolated clusters, which became connected as λ increased, finally forming one single cluster-network. The solvation of H3O+ ions also increased with λ, forming Zundel and Eigen ions which are of importance for proton transport via the Grotthuss mechanism. The poor conductivity found experimentally for Nafion at lower levels of hydration could therefore be explained by the bad connectivity of the water cluster-network. Recently, Devanathan and co-workers80,81,83 made an MD survey of the hydrated Nafion system using different all-atom force fields (including DRIEDING different λ and different temperatures, which is probably the most complete MD investigation of the Nafion system to date. The MD simulation box size was however not of significantly larger size than previous simulations, but consisted of four Nafion chains of 10 monomers each, using equidistant spacing of the side-chains. For a sampling period of 2 ns each, 11 levels of hydration from λ = 1 to λ = 20 were systematically investigated. A detailed analysis of the RDFs and CNFs showed that H3O+–SO3− hydrogen bonds played an important role in determining the phase separation of the nanostructure, and forming the hydrophilic/hydrophobic interface. For λ less than 7, hydrogen bonds bridged several SO3 groups into tight and constrained clusters. Since most water molecules and hydronium ions are strongly bonded to the sulphonic groups, proton transport – still modeled only as vehicular transport – is severely restricted. Furthermore, the strong bonding of the hydronium ions to the sulphonic groups sterically hinders their solvation in the water domain. However, there are still vital dynamics in the system also at low level of hydration. The H3O+ ions and H2O molecules sometimes release their bonding to the sulphonic groups and travel a distance in the water channel, before eventually getting bonded again to another sulphonate group from a different part of the polymer matrix. The water channel connectivity is thus under constant change. From the weak interactions between the water molecules and the non-sulphonic group side-chain atoms, Devanathan et al. concluded that the side-chains rather are a part of the hydrophobic backbone matrix, although they contain hydrophilic ether oxygen atoms. As the level of hydration increases in the membrane, the MD simulations showed that fewer of the H3O+ ions participated in bridging several SO3 groups. This was also shown to be in very good agreement with experimental quasi-elastic neutron scattering (QENS) results.84 The major structural changes take place at λ = 5–7, were the number of non-bonded H3O+



335

increase significantly and the SO3 groups separate from each other. At a high level of hydration, the distance of the water molecules and hydronium ions to the nearest SO3 group stretched to ~1 nm, which signals that this is the maximum radius of any water cluster or channel. However, considering the limited MD box sizes, it is doubtful if this model can actually predict the nanostructure, since the boxes contain at most two or three hydrophilic clusters. It is unlikely that they are true representatives of the distribution of cluster sizes and shapes found in real PFSA membranes. Devanathan et al. also analysed the mobility of the H2O and H3O+ ions in Nafion, by calculating mean residence time (MRT) and diffusion coefficients from MSD data. Raising the temperature from 300 to 350 K or raising the level of hydration was shown to have profound impact on the dynamics of the system. The MRT for the H3O+ ions around the sulphonic acid groups decreased from almost 1 ns at λ = 3 to 0.2–0.5 ns in fully hydrated Nafion. Based on the MRT values and RDFs, the authors could divide water in Nafion into three distinct classes: free (bulk-like) water, weakly bonded water and bound water, with the latter two increasing at increasing levels of hydration. Brandell et al.85 were first to expand the Nafion MD simulation to other PFSA systems; the longer Aciplex membrane side-chain and the shorter Dow membrane (now known as Hyflon Ion) side-chain. The authors based their force field on modifications of the DREIDING force field,64 and simulated all three different membranes at λ = 5 and 15. Not surprisingly, the results for Nafion were similar to Devenathan et al.80,81 and Jang et al.77 Using a qualitative approach, the authors concluded that the shorter sidechains gave rise to a larger but more poorly connected cluster, while longer side-chains resulted in a more non-uniform network. This is seemingly contradictory to the conclusion that the side-chain is merely a part of the hydrophobic matrix. On the other hand, the authors also found that only the very short Dow membrane side-chain was fully hydrated. For the fully hydrated systems, the diffusion coefficient resulting from the MSD functions was highest for H3O+ in Nafion, while H2O displayed highest mobility in Dow. The sulphonic acid group dynamics was also highest for Nafion, signalling a correlation between the side-chain and proton mobility. The authors concluded that the intermediate side-chain length gave rise to both an ideal water-channel network and an ideal sidechain mobility, promoting proton transport. This is in conflict with experimental results, where proton transport has been found to be somewhat higher for short side-chains.86 However, it could well be argued that also this box size is too small to adequately represent all complexities of the nanostructure, and therefore the conclusions about the total vehicular proton transport are questionable.


336


8.6

Conclusions and future trends

MD has indeed provided significant contributions to our understanding of polymer electrolytes, and will continue to do so. The increasing computer power will provide computational chemists with tools for simulating both larger structures and run longer simulations. For polymer electrolytes, longer simulation times will be particularly important for analysing diffusion and migration processes at ambient temperatures especially in all-solid state systems, like PEO-based electrolytes.24 The microsecond regime for moderate system sizes and all-atom force fields are not too far away. Similarly, larger systems will help us analyse the heterogeneous nano-structures of fuel cell membranes; so far the simulated systems have been confined to MD boxes with cubic side length of ~20–30 Å – this is slightly less than the very typical distances found in experimental investigations. Simulations of boxes with about 10 times this side length will contribute to many more insights regarding the membrane topology. We can also anticipate new polymeric systems being simulated. For battery electrolytes, gels and ionic liquid systems have been very interesting for many applications;87 these have not yet been greatly studied with MD methodology. Considering fuel cell polymers, almost all simulations have focused on PFSA membranes. However, there is a large group of alternative aromatic polymer systems that have shown promising properties,88 and where MD can reveal many details of their structure and dynamic behaviour. New and more user-friendly software will also help in expanding the number of polymeric systems simulated. On the other hand, these computer programs will most certainly provide standard force fields which have not been bench-marked for the particular systems researchers would like to study. This implies additional dangers. It is also very likely that we will see more multi-scale modelling and use of mixed methods. It has been argued that computational chemistry is both ‘theory’ and ‘experiment’: ‘theory’ since clearly no measurements are made on a real system, and ‘experiment’ since the potentials used are often based on experimental data on simple systems and since the simulations display limited reproducibility. MD is indeed often referred to as a ‘computer experiment’. Today, most computational chemists would probably say that computation is neither theory nor experiment, but rather a third leg on the chemical body – both to test theory and to interpret experiment; alternatively, to perform ‘experiments’ on systems inaccessible to normal experimental techniques. This discussion puts focus on the relationship between MD and experiment. Experimentalists also interpret their data using theories and models – they do not anticipate reality. Experimental data can often be interpreted in several ways, sometimes even within the same theoretical context. Not



337

rarely are data interpreted on the basis of incorrect or inappropriate theory for the system under study. The interpretation of experimental results is therefore not a search for biblical ‘truth’. Just like the computational chemist, the experimenter uses models to make their interpretation, thereby creating a gap between themselves and reality. MD can indeed sometimes be as good (or bad) a method as experiment for modelling this reality.

8.7 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31.

References

a. leach, Molecular Modelling: Principles and Applications, Prentice Hall, 2001. t. schlick, Molecular Modeling and Simulation, Springer, 2002. q.h. zeng, a.b. yu, g.q. lu, Prog. Polym. Sci., 33 (2008) 191. m.p. allen, d.j. tildesley, Computer Simulation of Liquids, Oxford University Press, 1989. j.m. haile, Molecular Dynamics Simulations, Wiley, 1997. s. neyertz, d. brown, j.o. thomas, J. Chem. Phys., 101 (1994) 10064. s. neyertz, d. brown, j.o. thomas, Computational Polym. Sci., 5 (1995) 107. b. mos, p. verkerk, s. pouget, a. van zon, g.j. bel, s.w. de leeuw, c.d. eisenbach, J. Chem. Phys., 113 (2000) 4. s. neyertz, d. brown, J. Chem. Phys., 104 (1996) 3797. f. müller-plathe, Acta Polymer, 45 (1994) 259. o. borodin, g.d. smith, J. Phys. Chem. B, 110 (2006) 6279. o. borodin, g.d. smith, J. Phys. Chem. B, 110 (2006) 6293. o. borodin, g.d. smith in l. curtiss, m. gordon (eds.), Methods and Applications in Computational Materials Chemistry, Kluwer Academic Publishers, 2004. g.d. smith, r.l. jaffe, d.y. yoon, J. Chem. Phys., 106 (1997) 3798. o. borodin, g.d. smith, Macromolecules, 33 (2000) 2273. f. müller-plathe, w.f. van gunsteren, J. Chem. Phys., 103 (1995) 4745. l.j.a. siqueira, m.c.c. ribeiro, J. Chem Phys., 122 (2005) 194911. y.h. duan, j.w. halley, l. curtiss, p. redfern, J. Chem. Phys., 122 (2005) 054702. o. borodin, g.d. smith, Macromolecules, 31 (1998) 8396. o. borodin, g.d. smith, J. Chem Phys., 104 (2000) 8017. a. palma, a. pasquarello, g. ciccotti, r. car, J. Chem. Phys., 108 (1998) 9933. b.g. dixon, r.s. morris, J. Power Sources, 119–121 (2003) 856. o. borodin, g.d. smith, r.l. jaffe, J. Comput. Chem., 22 (2001) 641. o. borodin, g.d. smith, Macromolecules, 39 (2006) 1620. a. nishimoto, k. agehara, n. furuya, t. watanabe, m. watanabe, Macromolecules, 32 (1999) 1541. y. ikeda, y. wada, y. matoba, s. murakami, s. kohjiya, Electrochim. Acta, 45 (2000) 1167. a. hektor, m.k. klintenberg, a. aabloo, j.o. thomas, J. Mater. Chem., 13 (2003) 214. j. karo, a. aabloo, j.o. thomas, Solid State Ionics, 176 (2005) 3041. j. karo, d. brandell, Solid State Ionics, 180 (2009) 1272. o. borodin, g.d. smith, Macromolecules, 40 (2007) 1252. d. brandell, a. liivat, h. kasemägi, a. aabloo, j.o. thomas, J. Mater. Chem., 15 (2005) 1422.


338 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69.


d. brandell, a. liivat, a. aabloo, j.o. thomas, Chem. Mater., 17 (2005) 3673. d. brandell, a. liivat, a. aabloo, j.o. thomas, J. Mater. Chem., 15 (2005) 4338. a. liivat, d. brandell, j.o. thomas, J. Mater. Chem., 17 (2007) 3938. a. liivat, d. brandell, a. aabloo, j.o. thomas, Polymer, 48 (2007), 6448. z. gadjourova, y.g. andreev, p.g. bruce, Nature, 412 (2001) 520. a.m. christie, s.j. lilley, e. staunton, y.g. andreev, p.g. bruce, Nature, 433 (2005) 50. g.s. macglashan, y.g. andreev, p.g. bruce, Nature, 398 (1999) 792. z. gadjourova, d. martín y marero, k.h. andersen, y.g. andreev, p.g. bruce, Chem. Mater., 13 (2001) 1282. r. frech, v. seneviratne, z. gadjourova, p.g. bruce, J. Phys. Chem., 107 (2003) 11255. l.t. costa, m.c.c. ribeiro, J. Chem. Phys., 124 (2006) 184902. l.t. costa, m.c.c. ribeiro, J. Chem. Phys., 127 (2007) 164901. o. borodin, g.d. smith, r. bandyopadhyaya, o. byutner, Macromolecules, 36 (2003) 7873. a. aabloo, m. klintenberg, j.o. thomas, Electrochim. Acta, 45 (2000) 1425. h. kasemägi, m. klintenberg, a. aabloo, j.o. thomas, J. Mater. Chem., 11 (2001) 3191. h. kasemägi, m. klintenberg, a. aabloo, j.o. thomas, Solid State Ionics, 147 (2002) 367. h. kasemägi, m. klintenberg, a. aabloo, j.o. thomas, Electrochim. Acta, 48 (2003) 2273. k.a. mauritz, r.b. moore, Chem. Rev., 104 (2004) 4535. w.y. hsu, t.d. gierke, J. Membr. Sci., 13 (1983) 307. m.h. litt, Polymer Preprints, 38 (1997) 80. b. loppinet, g. gebel, Langmuir, 14 (1998) 1977. l. rubatat, a.l. rollet, g. gebel, o. diat, Macromolecules, 35 (2002) 4050. k. schmidt-rohr, q. chen, Nat. Mater., 7 (2008) 75. e. spohr, Molecular Simulation, 30 (2004) 107. m. petersen, g. voth, J. Phys. Chem. B, 110 (2006) 18594. m.k. petersen, f. wang, n.p. blake, h. metiu, g.a. voth, J. Phys. Chem. B, 109 (2005) 3727. hoffman, kuleshova, d’aguanno, J. Mol. Model., 14 (2008) 225. j.a. elliott, s.j. paddison, Phys. Chem. Chem. Phys., 9 (2007) 2602. r. devanathan, Energy Environ. Sci., 1 (2008) 101. j.a. elliott in s.j paddison, k.s. promislow (eds.), Device and Materials Modelling in PEM Fuel Cells, Springer, 2009. y.a. dyakov, y.k. tovbin, Russian Chem. Bull., 44 (1995) 1186. x.d. din, e.e. michaelides, Aiche J., 44 (1998) 35. j. ennari, m. elomaa, f. sundholm, Polymer, 40 (1999) 5035. s.l. mayo, b.d. olafson, w.a. goddard, J. Chem. Phys., 94 (1990) 8897. j.a. elliott, s. hanna, a.m.s. elliott, g.e. cooley, Phys. Chem. Chem. Phys., 1 (1999) 4855. a. vishnyakov, a.v. niemark, J. Phys. Chem. B, 104 (2000) 4471. a. vishnyakov, a.v. niemark, J. Phys. Chem. B, 105 (2001) 7830. j.a. elliott, s. hanna, a.m.s. elliott, g.e. cooley, Polymer, 42 (2001) 2251. x. gong, a bandis, a.tao, g. meresi, y. wang, p.t. inglefield, a.a. jones, w.y. wen, Polymer, 42 (2001) 6485.



339

70. a. vishnyakov, a.v. niemark, J. Phys. Chem. B, 105 (2001) 9586. 71. s. tsuzuki, t. uchimaru, m. mikami, s. urata, J. Chem. Phys, 121 (2004) 9917. 72. s. urata, s. tsuzuki, m. mikami, a. takada, t. uchimaru, a. sekiya, J. Comput. Chem., 23 (2002) 1472. 73. s. urata, j. irisawa, a. takada, s. tsuzuki, w. shinoda, m. mikami, J. Fluorine Chem., 126 (2005) 1312. 74. s. urata, j. irisawa, a. takada, w. shinoda, s. tsuzuki, m. mikami, J. Phys. Chem. B, 109 (2005) 17274. 75. s. urata, j. irisawa, a. takada, w. shinoda, s. tsuzuki, m. mikami, J. Phys. Chem. B, 109 (2005) 4269. 76. s. urata, j. irisawa, a. takada, s. tsuzuki, w. shinod, m. mikami, Phys. Chem. Chem. Phys., 6 (2004) 3325. 77. s.s. jang, v. molinero, t. çag˘in, w.a. goddard, J. Phys. Chem. B, 108 (2004) 3149. 78. s.s. jang, s.-t. lin, t. çag˘in, v. molinero, w.a. goddard, J. Phys. Chem. B, 109 (2005) 10154. 79. m. levitt, m. hirshberg, r. sharon, k.e. laidig, v. daggett, J. Phys. Chem. B, 101 (1997) 5051. 80. r. devanathan, a. venkatnathan, m. dupuis, J. Phys. Chem. B, 111 (2007) 8069. 81. r. devanathan, a. venkatnathan, m. dupuis, J. Phys. Chem. B, 111 (2007) 13006. 82. s. cui, j. liu, m.e. selvan, d.j. keffer, b.j. edwards, w.v. steele, J. Phys. Chem. B, 111 (2007) 2208. 83. a. venkatnathan, r. devanathan, m. dupuis, J. Phys. Chem. B, 111 (2007) 7234. 84. a.m. pivovar, b.s. pivovar, J. Phys. Chem. B, 109 (2005) 785. 85. d. brandell, j. karo, a. liivat, j.o. thomas, J. Mol. Model., 13 (2007) 1039. 86. k. kreuer, m. schuster, b. obliers, o. diat, u. traub, a. fuchs, u. klock, s.j. paddison, j. maier, J. Power Sources, 178 (2008) 499. 87. m. armand, f. endres, d.r. macfarlane, h. ohno, b. scrosati, Nature Mater., 8 (2009) 621. 88. m.a. hickner, h. ghassemi, y.s. kim, b.r. einsla, j.e. mcgrath, Chem. Rev., 104 (2004) 4587.


9 Characterisation and modelling of multivalent polymer electrolytes M. J. C. P L A N C H A, Laboratório Nacional de Energia e Geologia, Portugal

Abstract: Multivalent polymer electrolytes are among the diverse and numerous ion-conducting solid systems used in electrochemical devices such as high specific energy batteries. This chapter describes the importance of the morphology and crystallographic structures of multivalent polymer electrolytes, and their effects on critical ionic transport properties such as electrolyte ionic conductivity and cationic transference number. The mechanisms for ionic conduction are described and discussed, and examples that illustrate the complexity of this scientific area are briefly reviewed. The factors upon which a multivalent polymer electrolyte’s crystallinity depend are described, and the relationship between this feature of polymer systems and their ionic conductivity is highlighted by discussing the equilibrium phase diagrams that have been established for a number of multivalent polymer electrolyte systems. Key words: multivalent polymer electrolytes, ionic conductivity, cationic transference number, crystallinity in polymer electrolytes, equilibrium phase diagrams.

9.1

Introduction

Interest in polymer electrolytes, as originally described by Armand and colleagues,1 mainly arose due to their technological applications, for instance as solid electrolytes in high specific energy batteries and other electrochemical devices. Many types of solid polymer electrolyte based on organic polymer matrixes with dissolved inorganic salts have been proposed, in particular alkali metal salts such as LiX (X = SCN, ClO4, CF3SO3, etc.).2–4 Subsequent attention has also focused on understanding the properties of polymer electrolytes with multivalent cation salts, and their possible application in devices such as batteries, electrochromic devices, sensors, lasers and photoluminiscent materials.5–8 Multivalent cations are easier to handle than alkali metal ions and the neutral respective elements are less expensive than lithium, the most commonly used alkali metal. 340 © Woodhead Publishing Limited, 2010

Characterisation and modelling of multivalent polymer electrolytes

341

While the existence of polymer electrolytes has been known since 1964,9 based on the multivalent salts HgCl2 and CdCl2, it was only in the 1970s that their ionic nature was recognised.10,11 The ease of preparation for these compounds permitted rapid development of numerous complexes using a large variety of salts at different concentrations. The majority were based on poly(ethylene oxide), PEO, of high molecular weight. The criterion for forming these electrolytes was clearly dependent on the delocalised charge of the anion, a condition related to the lattice energy of the corresponding salt.12 This chapter begins by highlighting the general theory behind the formation of stable multivalent polymer electrolytes, and by providing an overview of the kind of complexes formed between the polymer, PEO, and divalent or trivalent cation salts. Details of the associated studies are described briefly. The transport properties of these materials and the mechanisms for ionic conduction are discussed in detail in Section 9.3. Since the conductivity of a polymer electrolyte containing monovalent or multivalent salts is affected by the movements of polymeric chains,13 emphasis is placed on the ionic charge carrier’s mobility and concentration. Since PEO is a medium with very low dielectric permittivity, interactions between salt ions are likely to be significant in these systems, and there is a high possibility of the existence of long-lived charged and uncharged ion aggregates, such as ion pairs or triple ions. The complexity of the relationship between salt concentration and conductivity is illustrated with various studies. However, conductivity is not sufficient to describe the behaviour of multivalent polymer electrolytes in which both the anions and cations are mobile. Another parameter that must be taken into account is the cationic transport number, a critical property for batteries based on polymer electrolytes.14 For partially associated salts, this parameter is actually the cationic transference number. The difficulty of determining this property correctly in multivalent polymer electrolytes is discussed in the light of the characteristic features of each system. Multivalent polymer electrolytes that use PEO as the polymer base are usually multiphase systems consisting of salt-rich crystalline phases, another crystalline phase of pure polymer, and amorphous phases with dissolved salts. Conductivity is thus often affected by factors such as slow crystallisation and salt redistribution processes between the phases, which result in values dependent on the thermal history, preparation methods, etc. The morphological and crystallographic structures of these polymeric systems are presented, and the influence of factors such as crystallisation of the pure polymer on ionic conduction, and the factors that influence them in turn, are highlighted. Discussion of conduction takes into account the two possible mechanisms of mobility of charged ionic clusters and exchange of


342


cations and anions of ion pairs and larger clusters. The existence of these associated ionic species within a multivalent polymer electrolyte is described in several studies, which are summarised briefly. Finally, the thermal behaviour of several multivalent polymer electrolyte systems is reviewed. The importance of establishing equilibrium phase diagrams is discussed in some detail in the last section of this chapter, concluding that the thermodynamic interactions that exist in multivalent polymer electrolytes are, even qualitatively, very useful in understanding the mechanical properties, conduction and stability of these compounds.

9.2

Polymer–complex formation

9.2.1 Ion–polymer interactions For polymer systems, the solvation process is generally dominated by enthalpy changes. Salt dissolution only takes place if the exothermic ion– polymer interactions compensate for the lattice energy. Anionic solvation results from hydrogen bonding, while the solvation of cations is mainly due to electrostatic interactions. Since most polymer hosts used in polymer electrolytes cannot form hydrogen bonds, the enthalpy of solvation is mainly related to electrostatic interactions between the positively charged ion and the negative ends of dipolar groups within the polymer, or to partial sharing of a lone pair of electrons from the polymer coordinating atom, which leads to the formation of a coordinating bond.15 For multivalent cation polymer electrolytes, the hard and soft acids and bases (HSAB) principle16 (relating to ion–solvent forces) accounts for and predicts the stability of the complexes formed between salt ions and polymer. Molecules such as the ethers and certain amines, which have donor atoms with high electronegativity but low polarisability, are ‘hard’ bases. Thioethers are examples of ‘soft’ bases because they retain their electrons less strongly and are highly polarisable. The cations characterised as ‘hard’ acids are small, without unshared electron pairs in the valence shell, and have generally low polarisability. In contrast, large cations that are polarisable and have unshared pairs of electrons act as ‘soft’ acids. The most stable complexes are formed between either ‘hard’ acids and ‘hard’ bases or ‘soft’ acids and ‘soft’ bases. Of the complexes that have been studied in most detail, the constituent cations can be roughly ordered as follows, going from ‘hard’ acids to ‘soft’: Mg2+, Ca2+ > Li+, Na+ > Ni2+, Cu2+, Zn2+ > Cd2+, Ag+, Hg2+ Most studies conducted on polymer electrolytes have been based on commercial PEO since it is easily available and manageable as a solvent polymer for salts.



343

PEO is a ‘hard’ polybasic molecule, so it is expected that Mg2+ ions will form strong complexes with the polymer, whereas Hg2+ should only interact weakly with the ‘hard’ ethereal oxygen atoms. Complexes of PEO with both these cations are formed readily, but transference number measurements have demonstrated that, in this medium, Mg2+ ions are immobile17 and Hg2+ ions are mobile,18 which suggests a possible inverse relationship between promoting complex formation and the consequential effects on cation mobility. Although the HSAB principle gives a good indication of the stability of a complex once it is formed, it does not predict which salts are likely to form polymer–ion complexes. For example, mercury halides are too covalent to allow free Hg2+ ion to form in solution, and therefore complexes in this case are between polymer and molecular salt. Crystallographic studies of PEO–HgCl2 complexes show the linear salt molecule to be slightly bent in the complex because of unsymmetrical interactions.19,20

9.2.2 Poly(ethylene oxide) as polymer base The ability of PEO to dissolve salts of divalent cations and monovalent anions was first demonstrated by Blumberg and colleagues, who synthesised and studied PEO–salt solutions formed with HgCl2 and CdCl2.9 However, only since the mid-1980s have studies been conducted on complexes formed between PEO and other salts, such as Ca(SCN)2·4H2O and Ba(SCN)2·3H2O,21 PbBr2, PbCl2 and PbI2,22 MgCl2,17 Cu(ClO4)2, CuCl2, ZnCl2 and ZnI2,23 Sr(ClO4)2·6H2O and Mg(SCN)2·4H2O,24 NiBr225 and Cu(CF3SO3)2.26 Salts of trivalent metals have been also successfully incorporated into PEO, including La(ClO4)3,27,28 La(CF3SO3)3,29 EuCl3 and EuBr3,30–32 Eu(pic)3,33 Nd(CF3SO3)334 and Tm(CF3SO3)3.35 Ionic conductor complexes with mixtures of salts and PEO have also been studied. Examples of these systems include: PEO–CaBr2–CaI2 (30 : 1 : 1),36 PEO–Na+/Hg2+,18 PEO16–CoBr2/LiBr,37 PEO–PEGDME–ZnBr2/ LiBr38 (PEGDME-poly(ethyleneglycol) dimethyl ether) and PEO9– KCuxI1+x with x = Cu/K ≤ 1.39 In the latter case, the KI electrolyte acts as a medium, favouring the formation of the complex CuI2−, which transports the charge of Cu+ cations by a vehicular mechanism. Another example with multivalent cations that shows higher ionic conductivity of the mixed salt system than that containing only lithium salts, is the crosslinked polymer matrix PEO–PMMA(polymethyl methacrylate) with the mixture of metal salts MXn: M = Li, La, Yb, X = CF3SO3.40 Many of these studies were designed to understand the charge transport process,41 because the structure of these materials is much more complex than that of traditional crystalline and glassy solids. Berthier et al.42 used nuclear magnetic resonance (NMR) to demonstrate that ionic conductivity


344


in polymer electrolytes was achieved by the movement of ions along and between polymer chains, which occur primarily in the amorphous region. Electrolytes such as PEO–Hg(ClO4)2, with compositions up to at least 4 : 1, are completely amorphous and stable at room temperature for months, thus showing no evidence of the formation of any crystalline complex.18 Complexes synthesised by Atchia and colleagues involving PEO and salts of M(CF3SO2–N–CF3SO2)2 with M = Pb, Ni and Cu, showed that, in general, they are completely amorphous for compositions n ≤ 12 at temperatures above room temperature.43 The perfluorosulphonimide anion, (CF3SO2– N–CF3SO2)−, in addition to a large charge delocalisation (from the two CF3SO2− electronegative groups), also has a molecular structure that allows great mechanical flexibility, which modifies the dynamics of polymer chain segmental motions near the ions. The low glass transition temperature (Tg), with little or no variation with salt concentration in the various electrolytes, was indicative of the plasticising effect of the salt on the polymer chains. The rise in amorphousness associated with an increase in salt concentration in complexes formed by the addition of Ba(SCN)2 to PEO was attributed to a weak type of crosslinking between polymer chains through the Ba2+ ion, resulting in an increase in the degree of distortion of the polymer backbone.44 A similar result was obtained for complexes formed between PEO and Fe(SCN)3. James et al.45 found that divalent halides of Zn, Co, Fe and Cu form single phase amorphous complexes with poly(propylene oxide) (PPO). None of these authors reported any studies of ion transport. Fontanella et al.21 prepared PEO complexes with Ca(SCN)2 and Ba(SCN)2 and found them to have high Tg values and low ionic conductivities. Yang et al.46 studied the preparation and properties of PEO complexes with MgCl2 and PbCl2 salts and concluded that these materials, in certain salt compositions, have a moderated level of ionic conductivity. Many divalent cation polymer electrolytes have, however, a high level of ionic conduction. The ionic conductivity of PEO16–MgCl2, for example, is comparable to that of PEO9–LiCF3SO3 in the 80–150 °C temperature range.

9.3

Ionic transport properties

9.3.1 Conductivity and ion mobility In general, PEO–salt complexes, including those containing multivalent cations, present one of three types of behaviour regarding the dependence of conductivity on temperature:1 •

Type I. The conductivity of these compounds follows a free volume law in all experimental temperature ranges, reflected by a VTF (Vogel– Tamman–Fulcher) equation:47,48


Characterisation and modelling of multivalent polymer electrolytes σ = A0T−1/2exp[−EA/k(T − T0)]

345 [9.1]

or a Williams–Landel–Ferry (WLF) equation (extension of the latter). The A0 parameter is a pre-factor slightly dependent on temperature and related to the concentration of charge carriers.49–51 EA takes the form of an activation energy, but not in this case associated with a simple process of activation but rather related to the expansivity of polymer complexes50,52,53 or, in other words, with the segmental motion of polymer chains,54 and k is the Boltzmann constant. T0 is a temperature which can be associated with the glass transition temperature of the electrolyte. Amorphous electrolytes typically belong to this category. • Type II. These compounds have free volume behaviour at high temperatures and Arrhenius behaviour at low temperatures, with lower values of ionic conductivity. The Arrhenius law is expressed as follows: σ = σ0e−Ea/RT

[9.2]

where σ0 is the pre-exponential factor and Ea the activation energy of the system. This equation assumes the existence of an activated transport mechanism of ionic jumps, where the movement of charge conductors is uncoupled from the segmental motion of polymer chains. The transition between these two regimes corresponds to the melting temperature of the complex or to the melting temperature of the eutectic, depending on the salt concentration of the electrolyte. This temperature is the transition temperature, Tt. • Type III. For electrolytes with a high degree of crystallinity, the dependence σ = f(T) is described by the Arrhenius relationship in all ranges of experimental temperatures, although a discontinuity (at the transition temperature) is observed in the log σ versus 1/T curve, in which case equation 9.2 applies to each segment with changes in the activation energy values. Below Tt, the increase in ionic conduction is interpreted as a mechanism of jumps between coordination sites, local structural relaxations and segmental movements of polymer. As the amorphous region progressively increases (for T > Tt), the polymer acquires faster internal modes, where bond rotations produce segmental motion, which in turn facilitates ionic jumps inter- and intra-chains, and polymer electrolyte ionic conductivity increases for a higher range of values.55 For T < Tt, the regime is characterised by a higher activation energy, decreasing for T > Tt. The conductivity behaviour of an electrolyte according to temperature ascribes it to one of the classes described above, and depends significantly on factors such as presence or absence of solvent and/or water traces, and on the thermal history of the compound, namely the recrystallisation kinetics of the complexes and of the pure PEO, which often exist simultaneously.


346


The composition and volume fraction of the amorphous phase, the main phase that ensures conductivity in these materials as demonstrated first by Minier and colleagues,56 strongly depend on the thermal history. The existence of water/solvent in systems affects not only the degree of amorphousness, but also the existence of plasticising effects. Thus it is natural that the conductivity of the electrolyte will be affected. Ionescu-Vasii et al. characterised the multivalent polymeric system, which consists of a triblock–copolymer system poly(iminoethylene)-b-poly(oxyethylene)-bpoly(iminoethylene) complexed with perfluorosulphonimide salts of Cu2+ or of both Cu2+ and Li+, and showed the effect that existing domains with different degrees of crystallinity have on the conductivity behaviour of the polymer electrolyte.57 A great number of multivalent polymer electrolytes are semicrystalline in nature and belong to type II or III. One example of type II behaviour is shown by the PEOn–ZnI2 system, with the variation of ionic conductivity with temperature behaving differently below and above the transition temperature (Tt).58 For T < Tt, the data gave a linear variation due to the existence of crystalline compounds below Tt and the Arrhenius law was applied. For T > Tt, the data produced a convex variation, in accordance with the amorphous regime observed above Tt and the VTF law was applied. Examples of type I PEO–MXn (M = multivalent cation, X = anion) include complexes of PEO-based calcium and barium salts,21 PEO-based zinc chloride (in composition O/Zn = 4 and 8),23,59 PEO–Cu(ClO4)2 (certain compositions)23 and PEO-Tm(SO3CF3)3.35 When the polymeric system is predominantly amorphous, conductivity–temperature behaviour is sometimes better described by the VTF law, for instance for the gel polymer electrolyte studied by Pandey et al., where the addition of liquid electrolyte provokes substantial conformational changes in the crystalline texture of the host polymer due to immobilisation of the liquid electrolyte in the gel system.60 The polymer crystallinity almost disappears and the VTF equation applies to the σ–T relationship. Some polymer electrolytes show conductivity temperature dependence that falls outside the three types described above, with neither the Arrhenius law nor the VTF (or WLF) law being followed in the temperature ranges studied.46,61 Here, if there are no phase changes, effects associated with ionic aggregate equilibria are likely, superimposed on the simple variation in ionic mobility. In all cases, it is important to consider not only this parameter, but also the number and types of charge carriers,15 which are influenced by the ionic association that probably exists in ionic transport.62 One curious example comes from recent work by Jeong et al., who showed ionic transport decoupled from polymeric segmental motion in poly(vinyl



347

alcohol) (PVA)-based Mg-conducting solid polymer electrolytes, with Mg cations moving between adjacent ionic aggregates.63 This single activated behaviour, mainly observed at high salt concentrations, did not follow the VTF law, in spite of no crystalline complex forming between PVA and the salt, Mg(CF3SO3)2, present in the system. Investigating the conductivity behaviour of a series of electrolyte membranes based on ethylene carbonate (EC), propylene carbonate (PC), PEGDME and their mixtures, dimensionally stabilised by blending with poly(vinylidene fluoride-cohexafluoropropylene) (PVDF–HFP), showed64 a linear relationship between log σ versus T−1 above 40 °C and a downward curvature at lower temperatures, which is described by the VTF equation. Another behaviour different from that of the typical three types was shown by a gel polymer electrolyte nanocomposite consisting of PVDF– HFP with a mixture of EC and PC as the liquid electrolyte, the ion salt Mg(ClO4)2, and the nanosized filler magnesium oxide, MgO.60 The conductivity temperature dependence followed the relation log σ vs. T−1, with curves that were linear up to ~60 °C, and thereafter no conductivity enhancement was observed. This is typical conductivity behaviour for gel polymer electrolyte nanocomposite films.60 The behaviour of the PEO–MgCl217 and PEO–Mg(ClO4)265 systems was found to be similar to that of the PEOx–ZnCl2 (x = 4–16) system, which has been studied in great detail by Plancha.59 With the exception of PEO4– ZnCl2, the other, more diluted electrolytes followed the Arrhenius law for two different temperature regions, separated at the transition temperature. This corresponds to the melting temperature of the eutectic (mixture of pure PEO and intermediate complex), predicted by the phase diagram. The conductivity variations observed at the transition temperatures are correlated with the amount of amorphous phase formed when the melting event occurs. For x = 16 electrolytes, with a higher mass fraction in pure PEO, the variation in conductivity values at this temperature is larger, and for x = 8 electrolytes, this variation is lower. Furthermore, the amount of available sites for cation coordination in the amorphous phase increases when the PEO melts, with the greater the amount of melted material, the greater the number of available sites. Polymeric chain segmental motions also increase, increasing the free volume66,67 and facilitating ionic movements. Before the PEO melts, the interphase between the crystallites and the amorphous region acts as a trap for ionic migration, since the segmental motion linked to crystallites is significantly suppressed.68 The different activation energy values (Ea) obtained for PEO–ZnCl2 electrolytes are in the same range as those calculated for other divalent salt systems with the same conductivity behaviour, such as PEO–PbX2 (X = Br, I, Cl),22 PEO–Mg(ClO4)265 and PEO–EtMgBr–THF.69 These values are about 65 kJ mol−1 for T < Tt and about 34 kJ mol−1 for T > Tt.


348


The VTF conductivity behaviour exhibited by n = 4 electrolytes through the whole temperature range studied, with no transition temperature, is in agreement with results obtained from thermal analysis, and may reveal a relatively high degree of amorphousness. In the PEO–Tm(SO3CF3)3 system studied by Silva et al., the more concentrated electrolytes showed non-linear behaviour for the temperature dependence of conductivity, unlike the Arrhenius behaviour observed for the more diluted electrolytes.35 Like PEO4–ZnCl2, PEO3–Tm(SO3CF3)3 gives very high values of σ (10−3 S cm−1 at 90 °C). In these cases, Ea values are about 3 kJ mol−1, much lower than that for Arrhenius behaviour.

Influence of salt concentration The variation of conductivity with concentration is a complex problem, even for completely amorphous systems. For polymer electrolytes, the number of charge carriers per unit volume depends on the concentration, but not in a simple and direct way such as that for ‘strong electrolytes’ in a medium with relatively high permittivity values, like water. In this case, it is necessary to take into account the formation and dissociation of ionic aggregates (especially at low concentration), and the effect of an increase in the dielectric constant of the medium on those species as the salt concentration of the electrolyte increases. Ionic mobility is correlated to the relaxation modes of the polymer, which can be observed through the increase of the Tg value of polymeric systems with salt concentration. A rise in viscosity is also observed. The reduction in polymer segmental motion is usually interpreted as being due to an increase in the number of transient crosslinks formed by intra- and intermolecular bonds between ions and coordination sites belonging to the same or to different polymer chains.54,70–73 Ion–dipole interactions between cations and the ethereal oxygen atoms in the case of PEO are often considered to cause the hardening of polymer chains.74 Besides the decrease in ionic conduction caused by this matrix stiffening, the availability of vacant sites for the formation of transition states is also quite limited at high concentrations. Strong ion–ion interactions are also likely to exist in systems of low permittivity such as polyethers, and therefore ion migration is likely to involve the cooperative motion of several ions. The cationic transport of charge must involve dissociative steps, where the cations solvated by the polymer are transferred between adjacent coordination sites, combined with the migration and diffusion of ionic aggregates weakly coordinated with the polymer solvent. The type of charge carriers referred to in the equation for conductivity dependence on charge carrier number (ni), the charge of each (qi), and the respective mobility (µi):



σ (T ) = ∑ ni qi μi

349 [9.3]

i

thus includes mobile charge aggregates in addition to simple ions. Ionic conduction in systems where the salt is present mainly in the form of ionic aggregates is believed to be due to a combination of two processes:2,15 • •

motion of charged aggregates; transfer of ions between the associated species.

Di Noto and Vittadello estimated the ionic conductivity of the PEG400– (MgCl2)x system and inferred that it was due to two distinct phenomena: ‘jumping’ of the Mg2+ and [MgCl]+ cations between coordination sites along the polyether chains (intra-CH ‘jumps’), and migration of these ions between different chains of poly(ethylene glycol) PEG (inter-CH ‘jumps’). This last movement results in conduction, if followed by correlation of movements and geometric relaxation of polymer coordination sites.75 The dependence of conductivity on concentration is a very complex function. Briefly, at low salt content, the mobility of ions is relatively unaffected by concentration as the transient density of crosslinking will be low, and therefore conductivity will be controlled by the number of charge carriers. As the salt concentration increases, ion pairs and mobile higher aggregates form,76 which may then set down bigger, less mobile clusters and may also act as transient crosslinking species.77 At the highest salt concentrations, typically for an EO : salt molar ratio below 10 : 1, the system may be best thought of as a continuous ‘Coulomb fluid’, where long-range interactions are important and where the material may be considered to have much in common with a molten salt hydrate. Therefore, the number of charge carriers and their mobility vary with salt concentration in opposite directions and, for a given temperature, the conductivity in relation to salt concentration must pass through a maximum. Contrary to this are the results for the PVA–Mg(CF3SO3)2 system conduction mechanism, which indicated no maximum for the dependence of conductivity on concentration, showing an increase of this property with an increase in the proportion salt : polymer.63 The authors explained this fact by describing a transport mechanism independent of the polymer segmental motion. In a relatively recent study of the PEO–ZnI2 system, the profile of conductivity against electrolyte concentration at 100 °C decreased between n = 20 and n = 12, then rose to n = 8, with a maximum at n = 20.78 Other zinc systems containing zinc triflate and plasticisers, studied recently by Kumar and Sampath, also showed maximums in the evolution of conductivity with molar concentration.79,80 A study by Liebenow of a PEO-based system with ethyl magnesium bromide containing a plasticiser showed an increase in conductivity with Mg content, from n = 10 to n = 4 for T = 40 °C,69 which


350


is comparable with the results of the Zn systems at low temperatures. Zinc triflate-doped sol–gel-derived di-urea crosslinked PEO/siloxane ormolytes were studied, and the conductivity maxima were located at n = 60 and n = 20, at 30 and 100 °C, respectively.81 Another very recent result describes the conductivity behaviour of polymer electrolyte systems based on PVA as the polymer matrix with mixed cation salts, KOH and zinc salts (zinc acetate or zinc triflate).82 Conductivity values at 27 °C decreased from 1.57 × 10−2 S cm−1 (with zinc acetate) to 5.37 × 10−3 S cm−1 with an increase in concentration of the zinc salt from 25 to 35% (with 40% of PVA). A similar result was obtained for zinc triflate (ZnTr), with a decrease in σ from 2.54 × 10−2 to 5.70 × 10−3 S cm−1 as ZnTr concentration increased from 25 to 35% (also with 40% PVA). The formation of ion pairs as zinc salt concentration increased was suggested as a probable reason for this result. The formation of ion pairs was first observed in polymeric systems containing monovalent salts.51,83–87 Later, the formation of these aggregates was also shown by Fourier transform infrared (FTIR) studies on systems such as PEO–M(CF3SO3)2 with M = Zn, Pb88 and PEO–M(CF3SO3)3 with M = rare earth metal.7 In a recent work, a novel magnesium ion conducting gel polymeric system, Mg(Tf)2/EMITf/PVDF–HFP (EMITf = 1-ethyl-3-methylimidazolium trifluoro-methane sulphonate) polymer electrolyte, was characterised and an IR band from FTIR spectroscopic studies conducted on the system was assigned to ion pairing between the cation Mg2+ and the anion CF3SO3−.89 Raman spectroscopy studies carried out on (PPO)20 (ZnBr2)x(LiBr)1−x and (PEO)20(ZnBr2)x(LiBr)1−x polymer electrolytes90 showed that, even for very low concentration ranges, ZnBr2 salt was present. The existence of a neutral triplet has also been described for electrolytes based on other zinc salts with lower salt concentrations, such as triplets ZnI2 in PEO30–ZnI291 and ZnTf2 in PEO24–ZnTf2.88,92 FT-Raman spectroscopy conducted on the gel polymer electrolyte (GPE) containing polyacrylonitrile (PAN), EC, PC and ZnTr, with variation in salt concentration, allowed the changes in zinc ion coordination to be followed.79 The presence of ion pairs and higher aggregates was revealed, and a correlation between ionic conductivity values and their variation with respect to increasing salt concentration was found to match the spectroscopy results. These are represented schematically in Fig. 9.1. The percentage variation of free ions, ion pairs and higher aggregates found for a specified composition of GPE system constituents with increase in salt concentration is shown in Fig. 9.1(a). Figure 9.1(b) shows schematically the variation in ionic conductivity of the GPE system with increase in metal salt concentration. It can be seen that the initial rising trend for ionic conductivity is coincident, at the same range of salt concentrations, with the initial decrease in aggregates (ion pairs and others) and increase in free ions. In addition, the composition at which maximum conductivity of the system is reached corresponds to ion



351

% of components

(a)

Free ions Ion pairs Higher aggregates

Salt concentration

Ionic conductivity

(b)

Salt concentration

9.1 (a) Schematic percentage variation of free ions, ion pairs and higher aggregates with increasing salt concentration for a determined composition of the GPE system studied by Kumar and Sampath.79 (b) Schematic ionic conductivity as a function of ZnTr concentration, in the referred GPE system with the same composition.


352


pairs and higher aggregates being at their lowest and free ions at their highest percentage. This composition is about 1.56 M for ZnTr.79 Based on the presence of plasticisers and the results of found ion association in a study of the PEO–ZnBr2 electrolyte, Chintapalli and colleagues93 postulated the existence of equilibria 9.6 and 9.7, in competition with equilibria 9.4 and 9.5. They found evidence for the possibility of charge being transported by ZnBr3− or ZnBr42− ions, or by both, as these equilibria are more or less displaced in one or the other direction. R

O----- ZnBr2+ Br–

R R

O----- ZnBr3–+ Br–

H

[9.4]

R

O----- ZnBr42–

[9.5]

R O----- ZnBr2+ Br–

H R

O----- ZnBr3–

R

R R

R

R

O----- ZnBr3–

[9.6]

H O----- ZnBr3–+ Br–

R

O----- ZnBr42–

[9.7]

H

Unlike the pure system without plasticisers, where only the ZnBr3− anion was observed, in the plasticising system the ZnBr42− ion cluster was also observed. By analogy, the existence of the two cluster anions ZnX3− and ZnX42− (X = Cl or I) was reported as being probable in PEO–ZnCl2 and PEO–ZnI2 systems,59 with water traces which act as a plasticiser, competing with ether oxygen for complexation with the cation. The increase in overall ion mobility with salt concentration through a combination of two effects was suggested for the PEO–ZnCl2 system. The two effects are: (a) saturation of polymer matrix coordination sites, increasing the population of more mobile ions, and (b) an increase in the number of additional sites (Cl− free anions, ZnCl+ ionic pairs, the triplet, ZnCl2 and larger aggregates of the type ZnCl3− and ZnCl42−) that become available for fast ion exchange (vehicular mechanism)94 with increasing salt concentration. In a study of the PPO–Eu(CF3SO3)3 system by Raman spectroscopy and luminescence of the Eu3+ ion, the existence of rapid ion exchange between cations coordinated to the polyether sites and cations coordinated to anions was shown.95 The existence of ion aggregation has also been demonstrated in studies with different nickel salt systems, particularly PEOn–Ni(CF3SO3)2, with n = 3–24,92 in PEO8–NiBr296 and in PEOn–NiBr2, with n = 8–64,97 with NiBr42− clusters being identified in this last work. In a recent study of nonaqueous magnesium conductor systems, the MgCl+ ion pair and the Mg2Cl3+



353

aggregate were indicated as Mg species formed for the conduction of charge.98 In a gel polymer electrolyte based on PEO–ZnSO4–nanoclay, an increase in salt concentration resulted in the formation of the ion pair [M2X]+[MX2]− with restricted mobility.77 Several authors have observed that the ion association in multivalent systems increases with temperature.7,78,92,99–103 Xie and Farrington showed the appearance of Zn2Br3+ aggregates when the ‘temperature’ parameter was increased in molecular dynamic and mechanical simulations of a PEO– ZnBr2 system.103 The study of cation–anion and cation–polymer associations in PEO–Pb(CF3SO3)2 complexes showed that the concentration of free ions decreased more rapidly between 60 and 70 °C.102 With temperature increase, precipitation of the polymer salt may even occur.104,105 The variation in ionic conductivity of the PEO–ZnI2 system with salt concentration at 120 and 150 °C was revealed to be relatively small.59 The reasons may be that, at these high temperatures, crystals are absent and ion aggregation is proportionately large. This leads to competition between the effects of increased vehicular mechanism, with the increase of ion mobility on the one hand (the ionic conductivity, σ, increases) and the effect of increased local viscosity due to the increase of ionic crosslinking between the cations and the ethereal PEO oxygen atoms, which decreases the polymer segmental motion (σ decreases). The free ion fraction will be lower due to aggregation, leading to a reduction in transport of these ions, which are more mobile than the aggregates. Other studies also demonstrated a small variation in conductivity with salt concentration for high temperature ranges.78,106,107 Examples include the behaviour of the PEO–NiI2 and PEO– NiCl2 systems.59 The reinforcement of cation–anion bonding and the favouring of ionic aggregate formation with temperature increase were explained84,87 according to entropy: the release of cations by an increase in temperature reduces crosslinking, increasing the disorder.

9.3.2 Cationic transference numbers If a particular salt forms phases that are electrically conductive when dissolved in a coordinating polymer, determining which species are responsible for transporting electric charge is essential for characterising the polymer electrolyte. Most studies have assumed that the only mobile species are simple anions and cations, and experimental data have been interpreted to give values for transport numbers. However, there is increasing evidence for ionic association in an extremely high concentration range. It is very difficult to determine which species are formed, and impossible to establish from conductivity data which of these species contribute to ionic transport


354


in these materials. Thus, what have really been obtained experimentally in these studies are the transference numbers of the cation or anion constituent and not the transport numbers of the ions. These two quantities are equal only when the electrolyte dissociates into two ionic species.108–110 Another important consideration is the fact that even uncharged associated species may be mobile. If this is the case, then in methods such as DC polarisation and pulsed field gradient techniques, which are affected by neutral species,2 the value obtained is neither the ionic transference number nor the ionic transport number. The Hittorf method applied to a zinc trifluoroethane sulphonate polymer electrolyte permitted the true zinc transference number to be calculated and, by comparing it with the value obtained by DC polarisation of a symmetric cell with Zn electrodes, the existence of mobile uncharged ZnX2 triplets was postulated.111 Whatever the state of the electrolyte, the value obtained is of great significance for characterising ion transport. In a battery, for example, it is not important which species is responsible for transporting the current if the total flow of the electroactive constituent M from the anode to the cathode is high. In circumstances in which the cations are immobile, neutral species are the only way to transport M through the cell. The mobility of ions in polymeric systems, particularly those that contain multivalent cations, is mainly based on breaking and re-forming ion-chain bonds. For example, in PEO, as in other aprotic solvents, cations are more strongly solvated than anions.112 In the transitions between different conformational states coupled with chain mobility, ether oxygen–cation bonds will be broken and restored again to produce a new chain conformation.113 Cation motion and segmental polymer motion are closely coupled in this manner. Cation–chain interaction can involve single polymer chains and also create pseudo-crosslinks between chains. Divalent cation salt–PEO complexes display a wide variation in cationic transference number values. While some polymeric systems with salts of Mg2+,17,114 Ca2+ 27 and Co2+ 115 are essentially pure anion conductors (with cationic transference number values generally lower than 0.05), PEO solutions with salts of Cu2+,17,23,116,117 Cd2+,115 Hg2+ 18 and Pb2+ 22,115 conduct both anions and cations, with cationic transference numbers between 0.18 and 0.70, depending on the system, the temperature and method of determination used. For example, the transference number for Pb2+ determined in PEOn–PbBr2 was about 0.7 at 140 °C.22 In another work, PEO electrolytes formed with nickel chloride, nickel iodide, zinc chloride and zinc iodide were characterised and cationic transference number data reported.118 The authors found low cationic transference number values for the Ni systems, indicating that the possible anionic species are the main carriers responsible for conduction. The higher cationic transference numbers obtained in the Zn systems were due to simple cation Zn2+ and/or larger associated charged



355

zinc species being weakly bonded to the polymer, thus allowing conductivity contributions by positive and negative charge carriers. For example, in PEO–ZnCl2 system, the following reaction equilibria of autocomplexation may exist: 2ZnCl 2 ZnCl −3 + ZnCl +

[9.8]

ZnCl 2 + ZnCl + ZnCl −3 + Zn 2+ −

ZnCl 3 + ZnCl 2 ZnCl

2− 4

+ ZnCl

[9.9] +

[9.10]

ZnCl+ species contribute to an increase in cationic transference number values, while ZnCl3− and ZnCl42− species are responsible for a decrease. The low cationic transference number values in the Ni systems referred to above did not differ too much from those found by Huq and Farrington25 for PEO8–NiBr2 as a cast/dried electrolyte. Results obtained for two PEO– Zn-based electrolytes118 did not agree with those reported by Yang and Farrington.107 Their cationic transference number values, obtained by AC and DC polarisation measurements, indicated negligible cation mobility. However, this was probably masked by unusually high interfacial impedance. Moreover, scanning electron microscopy (SEM) and energy-dispersive analysis of X-rays (EDAX) observations at high temperatures have shown that Zn species are reasonably mobile. In a more recent work, which introduced thermodynamic models and estimated diffusion coefficient values from conductivity data and phase diagram parameters for some polymer systems, including PEOn–ZnI2 and PEOn–ZnBr2, values over 0.5 were obtained for cationic transference number, again indicating the mobility of existing cations.119 In a sample of composition (0.5 PEGDME + 0.5 PEO)16–ZnBr2, where the PEGDME acts as a plasticiser, a transport number of 0.4 was obtained for Zn2+ at 53 °C.120 The contribution of both cationic and anionic charge carriers to conductivity was also observed in polymeric systems with ZnTf2 salts.80,111,121,122 Gina and colleagues obtained a cationic transference number (T+) of 0.22 for a PEO9–Zn(CF3SO3)2 system, measured in the temperature range 90 to 100 °C.117 In another work, a PEO16–ZnBr2 system gave a T+ value of 0.18 at 82 °C.120 The cationic transference number for a gel polymer electrolyte system studied by Sivaraman et al., which contained zinc sulphate as the source of mobile cations, was found to be 0.48. This value included the contribution of complex ions such as triple and/or quadruple ions.77 Huq and Farrington115 reported that the properties of PEO-based electrolytes vary significantly, particularly in relation to cationic transference number. These authors studied several electrolytes formed by PEO and divalent cations such as Cd2+, Ni2+, Co2+, Zn2+ and Pb2+. To determine the cationic transference number, Huq and Farrington used an electrochemical technique previously developed by Sørensen and Jacobsen123 for the PEOn–


356


LiSCN monovalent system. The authors115 proposed that the electrolytes formed by divalent cations dissolved in PEO can be grouped into three categories: •

Essentially pure anion conductors. This category includes polymer electrolytes with salts of small cations, which are highly polarisable, such as Mg2+ and Ca2+. In these materials, the cationic transference number is generally lower than 0.05 in the range 100–150 °C. The ether oxygens in the polymer chains trap the cations electrostatically, and ionic conduction is due to the anions. • Conductors with a significant cationic transference number. This category includes larger cations with low polarisability, such as Pb2+ and Cd2+. These cations have much weaker interactions with the polymer chains than Mg2+ and Ca2+ ions. The electrolytes formed with these cations conduct both anions and cations, being predominantly cation conductors at temperatures higher than 130 °C. • Conductors with cationic transference number controlled by the degree of hydration/dehydration. A good example of this class is the PEOn– NiBr2 system, in which the transference number of the Ni2+ ion is dependent on the state of hydration of the electrolyte.25 This system is essentially an anionic conductor at 120 °C in anhydrous conditions, but in a controlled hydration/dehydration process, Ni2+ mobility apparently increases substantially. The process of hydration/dehydration used by Huq and Farrington25 consisted of heating the samples to 140 °C, cooling them to room temperature, hydrating in moist gas and then dehydrating by heating to 140 °C in a flow of dry argon or nitrogen.

Although soft–hard interactions appear to be a requirement for high conductivity, considering the ion transport mechanism suggests that lability is of greater importance than the strength of the cation–polymer bond per se. Labile bonds are necessary for cation mobility. The similarity between the electron donicity of oxygen atoms in molecules of water and in molecules of a polyether can be used to predict the lability of the cation–polymer bond, by analysing the exchange rates of the hydrated cations with water in aqueous solution (Fig. 9.2).124 Although no systematic studies have been conducted, evidence from several experiments provides an adequate depiction of PEO multivalent complexes. For example, the alkaline earth ions Be2+ and Mg2+ have low rates of water exchange and are immobile in PEO. On the other hand, the Hg2+ ion has the fastest exchange rate with water of all the divalent cations, and it was predicted18 that Hg2+ might therefore be mobile in PEO. The mixed ionic charge carrier character obtained for PEO–ZnCl2 and PEO–ZnI2 systems118 are in agreement with the position of the Zn2+ ion in Fig. 9.2. This cation is close to the limit where its solvation by the polymer allows some mobility. There is thus



357

+ + Na+ K Cs+ Rb Li +

Be2+

Cr3+ 3+ Ru

Fe3+Ga3+ V3+

Al3+

V2+

Ru2+

10–4

Ti3+

Ni2+

Pt2+

10–6

Mg2+

Ca2+ Sr2+ Ba2+

In3+

2+ Co2+ Fe 2+ Cu2+ Cr2+ Mn

2+ 2+ Zn2+ Cd Hg

Pd2+

10–2

100

102

Yb3+ –Ho3+ –Gd3+

104

106

108

1010

9.2 Characteristic rate constants (s−1) for substitution of inner-sphere water molecules on various metal ions. Adapted from the work of Eigen124 with added data from other sources.125 Dashed line: estimated threshold for cation mobility.126

some lability in Zn2+–PEO bonds, as also expected due to the medium soft cation characteristics of zinc (from the R.G. Pearson principle),16 which lead to the Zn2+–ethereal oxygen bonds being slightly weak since PEO is considered a strong base. In contrast, measurements of the exchange rates of bound water molecules made by Eigen,124 located the Ni2+ ion in Fig. 9.2 in the area where the cations are considered immobile.126 The low lability of Ni2+–PEO and Ni2+–H2O bonds set by this diagram is consistent with the low T+ values estimated for PEO–NiCl2 and PEO–NiI2 systems.118,127 Trivalent cations of the lanthanide group have water exchange rates of about 107–108 s−1, and on this basis, it may be expected that they exhibit mobility in an anhydrous polyether electrolyte. However, conductivity and transference number measurements have shown that, although Hg2+ is indeed mobile18 and Mg2+ immobile17 in accordance with their water exchange rates, Ca2+ and La3+ are immobile,27,128 which does not agree with their position on the Eigen diagram. Nevertheless, taking into account the hard–soft acid–base principle, as La3+ is one of the hardest of the lanthanide ions, the complex formed between PEO and La3+ is strong, reducing the mobility of this cation. The factors affecting the kinetic lability of transition metal and other ion complexes are complicated. Clearly, electronegativity and ionic radius are


358


important. There are, however, more subtle effects. For example, the dominant factor in the case of the transition metals is the change in d-orbital energy on going from, say, a ground state octahedral coordination to some favourable five-coordinated transition state.129 Cu2+ has fast rates of exchange with water (and is effectively mobile in PEO)116 because of the Jahn–Teller weakening of the axial groups, which makes attainment of the transition state an easier process. More importantly, the mobile species are likely to be associated and the nature of these species and degree of association will depend on the anion, cation and salt concentration. The majority of studies on divalent cation polymer electrolytes have involved divalent ions as the salt. These series of salts have an associated problem in that results may not be directly comparable because there is considerable variation in the character of the anion–cation bond. For example, the chlorides of zinc, cadmium and mercury show a sharp transition from ionic to covalent character. Likewise, the tendency of, for example, zinc halides to form complexes, decreases on going from the chloride to the iodide. Mercury halides are too covalent to allow free Hg2+ ion to form in solution. Thus, HgCl2 dissociates only slightly to give HgCl+ and Cl−. These factors are important considerations when interpreting data, as the mobile species may vary from electrolyte to electrolyte. In addition to the species, the morphology of the polymer electrolyte is important when determining transference number data. Many of the electrolytes containing divalent or trivalent cations comprise complexes with high melting temperatures (180 °C and upward).27,115 In many instances, measurements may have been made on heterogeneous systems.

9.4

Morphological and crystallographic structures: characteristics and influence on ionic transport properties

The behaviour of PEO-based electrolytes with multivalent cation salts is strongly related to the nature and concentration of ionic species involved, preparation conditions, thermal history130 and state of hydration of the sample. Martins and Sequeira analysed the effect of these factors on the observed performance of PEO-based systems with salts of divalent metals.131 One of the properties reviewed was the composition of maximum conductivity, and a huge variation was observed. In a number of studies of the PEOn–ZnI2 system, different preparation conditions for the polymeric films were shown to affect the crystallinity of the electrolyte and displace the conductivity maximum (as a function of salt concentration) for different compositions. Examples include the compositions n = 30, for the range



359

40–120 °C (except to 60 °C where the conductivity maximum occurred for n = 15);106 n = 24 for 60 °C and 100 °C;78 n = 20 for the range 60 to 140 °C107 and n = 16 for 140 °C.132 Most studies on polymer electrolytes so far have been based on the polymer PEO with mono and divalent cation salts. A small number of studies have been carried out on trivalent cation-containing polymer electrolytes, mostly lanthanum systems. The morphological, electrical and structural properties of polymer electrolytes based on lanthanide salts are still poorly understood compared with what is known about other cations. Some experimental difficulties associated with the achievement of salts with the appropriated properties have contributed to this. Petersen et al.101 and Bernson and Lindgren7 using, respectively, Raman and FTIR spectroscopy, showed the influence of lanthanide cations in the structural organisation and properties of polymer electrolytes. These studies tried to correlate the characteristics of possible interactions between cations and the oxygen atoms of the polymer. Bernson and Lindgren studied PEO-based electrolytes, and evaluated the influence of parameters such as temperature and nature of the trivalent cation (La3+, Nd3+, Eu3+, Dy3+ and Yb3+) and the effect of possible interactions between cations and different groups of triflate anions. This study showed that, with the exception of ytterbium, the characteristics of the interactions between each anion and each cation were similar in all studied salts. The authors observed that the Yb3+ cation participates in a complex, involving almost exclusively the anion and where the polymer is apparently excluded. This difference in behaviour was attributed to the electronic structure of the cation. The results for electrolytes based on the cation La3+ revealed, for a composition of n = 9, that a complex of well-defined stoichiometry was formed. In this complex, the cation appeared to be coordinated by nine oxygen atoms and by the three anions oriented by means of the group SO3. It was also observed that, in the composition range between n = 9 and n = 16, the electrolytes were entirely amorphous, irrespective of the cation. Studies done by Gray et al.27 on electrolytes based on lanthanum perchlorate focused in particular on evaluating the concentration effect on morphology and conductivity properties. Provisionally, a eutectic was identified with the ratio O/La near 10 at a temperature of 50 °C. Conductivity measurements showed that the electrolyte was predominantly amorphous for values of O/La less than 12, and that the more diluted electrolytes were semicrystalline at temperatures below 60 °C. The low conductivity of electrolytes with amorphous properties was attributed to the reduced mobility of the La3+ cation, a consequence of the high charge/radius ratio. Important features of PEO-based electrolytes with lanthanide cation salts are their thermal stability and morphology. Huq and Farrington30 studied the influence of the anion (Cl− or Br−) on the morphology of


360


polymer electrolytes based on a europium cation. It was observed that the anion was particularly significant in the morphology of the crystalline structures formed. For example, PEO spherulites obtained for PEO16–EuX3 (X = Cl or Br) electrolytes were larger and more well defined in the bromide electrolyte. On the other hand, the influence of the anion on ionic conductivity values was small, and the materials presented a relatively high conductivity of the order of 10−4 S cm−1. An additional observation was the extension of the polymer electrolytes’ thermal stability to a temperature of about 240 °C (under nitrogen atmosphere). The observed luminescence of these materials, an orange-yellow colour in the case of chloride and yellowgreen in the case of bromide, was apparently independent of morphology (amorphous or semicrystalline) and of the presence of solvation water molecules. Puga et al.133 also studied the stability and morphology of polymer electrolytes based on europium bromide with PEO. According to these authors, the morphology of this system at room temperature is a mixture of spherulites immersed in an amorphous region. For low salt concentration regions, the spherulites are large and well defined, and their size decreases with increasing salt concentration. Studies of the emission spectrum of Eu3+ in these electrolytes showed that, regardless of whether the cation is in a crystalline or amorphous phase, the geometry of the cation coordination sphere presents a coordination number of eight.134,135 Machado and Alcácer worked on polymer electrolytes based on PEOn–Nd(CF3SO3)3,34 focusing on the effects of composition and temperature on conductivity, electrochemical stability and thermal stability. They identified the existence of a complex between the salt and the polymer for a composition close to n = 36 and a eutectic around n = 23, with a melting point in the range 25–45 °C. For this composition, the ionic conductivity (2 × 10−7 S cm−1 at 21 °C) coincided with the maximum value found in the range of compositions studied. Investigating the electrochemical stability showed that the anodic and cathodic limits of this material, on platinum electrodes, decrease with increasing salt concentration. Mehta136 studied the electrochemical properties of the Eu3+/Eu2+ pair (as triflates) in PEO-based polymer electrolytes. Mehta’s main contribution came from cyclic voltammetry studies at a temperature close to 75 °C, which demonstrated that the observed ohmic drop had a large contribution from the low conductivity of the polymer electrolyte. The author concluded, from the characteristics of the peaks corresponding to the Eu3+ reduction and the Eu2+ oxidation, that these were situated at electrical potential values close to theoretical values, but the processes did not show reversibility. Impedance studies revealed that the behaviour of the system was controlled by mass transport processes and the respective main role was assigned to the Eu3+ ion, with a diffusion coefficient of 3.66 × 10−16 cm2 s−1.



361

As mentioned for polymeric systems with divalent cation salts, for PEObased electrolytes with lanthanide cation salts, variations in preparation conditions influence the properties of the electrolytes significantly. The distinctions produced in electrolytes with multivalent cations when the films were synthesised by the deposition method or by the pressing method were studied.29 The differences were related to the presence of solvent in the polymer electrolyte samples. For high salt concentrations, the differences in conductivity values are more considerable for lanthanum triflate-based electrolytes than for europium or neodymium systems. The films obtained by deposition showed lower conductivity than those produced by pressing, at ranges of higher salt concentration. A possible interpretation for this fact may be that ion–solvent interactions immobilise the ions, making the material less able to conduct a charge. The high dependence of important properties (e.g. ionic conductivity) on factors such as the process used for film formation, thermal history, low molecular weight solvent and water residues, in the case of polymer electrolytes based on multivalent cations,2 means that the morphology and ion mobility are not directly comparable, because the conditions are often difficult to reproduce. For samples prepared by the usual method of solvent evaporation (solvent casting), besides the water content, the drying conditions (thermal history) play an important role. Previous work has discussed the use of ionic salts, the polymer and the solvent after a preliminary drying in order to eliminate interference from water. Alternatively, solutions can be prepared with the constituents as acquired, namely using salts in hydrated form, with subsequent drying of the films after the solvent has been evaporated by submitting them to a vacuum under heating. The water and solvents generally participate in the coordination sphere of ions, or act as plasticisers. Polar solvents, on the other hand, seem to facilitate the solvation of salt by the polymer.2 This was suggested by results obtained in a study27 of PEO-based polymer electrolytes containing multivalent cations salts, where different morphologies were found when PEO–Cu(CF3SO3)2 films were prepared using solvent evaporation or by hot pressing, an alternative technique developed by Gray et al.137 in which any solvent is used. Wendsjö and Yang also noted the formation of different complexes in PEO16–PbI2 electrolytes when using dimethylformamide (DMF) or dimethylsulphoxide (DMSO) to dissolve the PbI2 salt. The first solvent inhibited the formation of crystalline PEO, in contrast to DMSO.138 In addition, the use of water or of a mixture of acetonitrile and methanol as cosolvents for the formation of PEO8–ZnCl2 electrolyte, led to different degrees of crystallinity with a consequence for the properties of conduction. It was also found that the effects of the solvent were not felt in films obtained after prolonged heating above the higher melting temperature of the crystalline material.106


362


In a recent investigation of a PVDF–HFP-based Mg2+ ion-conducting gel polymer electrolyte nanocomposite dispersed with nanosized MgO particles, significant changes in the crystallinity of the polymer were observed when the liquid electrolyte, a mixture of EC and PC, was added.60 The semicrystalline microstructure of the PVDF–HFP almost vanished due to the blending of the liquid electrolyte with the polymer at the molecular level, functioning as a plasticiser for the polymer and making the gel polymer electrolyte predominantly amorphous. With the addition of MgO nanoparticles, the morphology of the polymeric system changes. From a surface with small pores at the microscopic level, the surface becomes flat and uniform, without pores or any phase separation. This indicates the high affinity of the polymer host, PVDF–HFP, with the liquid electrolyte.60 The presence of impurities, such as traces of residual solvent, affect the morphological behaviour of polymer systems. Low molecular weight impurities may have a different number of effects, and possibly these are competitive. On the one hand, impurities may act as plasticisers, changing the structure by breaking the coordination bonds between the ethereal oxygen atoms of the polymer chain and the metal cation, causing a decoupling between ion movement and that of the polymer chain. Ion mobility increases due to the decrease in viscosity in the vicinity of the ions, i.e., there is an increase in the mobility of polymer chains, resulting in an increase of conductivity. On the other hand, they may help to increase polymer crystallisation speed and thereby change the distribution and relationship between the amorphous and crystalline phases, reducing the total conductivity if the amount of this latter phase increases. If the impurity is water, highly polar, there is a high probability of a strong connection to the ions (especially to small cations and those with high charge), which will interfere with the structure of the transition state through which the ion is transferred between coordination sites. The oxygen atoms of water molecules provide additional possibilities for the coordination of cations along the passage of one conformation to another, effectively reducing the energy barrier between the two conformational states, thus assisting the transition between conformations. This is one of the possible mechanisms for the enhanced mobility of polymer chains. According to Wendsjö and his colleagues, coordination between PEO oxygen atoms and the cations is weakened in the presence of water, which provokes the production of a greater number of free mobile ions, which in turn increases conduction in polymeric complexes.88 The conductivity increase with the amount of water in the system is higher for more concentrated electrolytes than for more diluted ones.91,113 Another possible mechanism is based on the assumption that the ions can migrate through a liquid solution layer, created within the polymer film by the absorption of water.139 This is essentially the mechanism behind the



363

type of conduction found mainly in gel-type polymer electrolytes. For example, Lee et al. suggested that the polymeric chain was of little importance to the ionic conductivity behaviour of the system compared with the liquid electrolyte.140 A polymeric gel electrolyte consisting of poly(ethylene oxide)-modified poly(methacrylate) (PEO–PMA) with magnesium imide as the electrolyte salt and EC and dimethyl carbonate (DMC) mixture as the plasticiser, prepared by photo-induced radical polymerisation, showed an increase in ionic conductivity at both 20 and 60 °C with an increase in plasticiser content. In addition, the activation energy for ionic conduction became lower for gels with higher EC + DMC contents. These results suggest that the increase in the liquid plasticising component causes a decrease in the interaction between Mg2+ ion and EO units in the polymer chain, so that most Mg2+ ions are solvated by the alkyl carbonates in gel systems that contain larger amounts of the liquid component.141 Another example is the interaction of the anion to enhance the conductivity of a polymeric system. The addition of nanosized grains of TiO2 to PEO8–ZnCl2 polymer electrolytes, prepared using PEO γ-irradiated from a C-60 source, provoked an increase in conductivity at room temperature of up to two orders of magnitude. It was suggested that this was not just because of changes in crystallinity, but also possibly due to anion interactions with the nanograins of added TiO2.142 The addition of nanoparticles increases the solvating power and flexibility of PEO chains, which also results in higher conductivity values, as obtained for a PEO–ZnSO4–nanoclay–H2O gel polymer electrolyte.77 The effect of hydration by exposure to atmospheric humidity on polymer electrolytes has also been studied, and a loss of crystallinity was observed in most cases106,138,143 with a consequent increase in ionic conductivity.22,106,113 However, in PEO-based electrolytes containing Pb halides, the crystallinity is not affected by water,138 in contrast to those using another lead salt, PEO–Pb(CF3SO3)2. Using NMR spectroscopy, Lauenstein and colleagues established that, for the latter system, the relative amounts of crystalline phase when the electrolyte was left in contact with atmospheres of different moisture content were lower than those of a PEO–Zn(CF3SO3)2 system.113 A possible explanation for the crystallinity reduction with hydration given by these authors is the ability of the hydrated cation to coordinate a higher amount of PEO than the free ion, similar to the fact that PEOn–Pb(CF3SO3)2 electrolytes are less crystalline than PEOn–Zn(CF3SO3)2 electrolytes, due to the larger coordination number of Pb2+ compared to Zn2+.92 In systems based on PEO and salts with divalent cations, the presence of water causes drastic changes in the mechanism of electrolyte ionic transport. This might lead to confusion regarding the comparison of published results. For example, in a study by Abrantes et al.,23 the authors concluded that the Zn2+ transference number in PEO4–ZnCl2 was 0.9, while other


364


authors112 claim that this is a pure anionic electrolyte. The zinc transference number, at a temperature of 120 °C, was 0.03 in studies made by Shi144 for a PEOn–Zn(CF3SO3)2 system composition n = 50. The difference between the values obtained by the two groups is quite large and may be due to the preparation of the electrolyte or to the method used to evaluate cationic transference number. As already mentioned, the microstructure of a polymer electrolyte is complex and depends on the composition, temperature and thermal history. Usually, in divalent PEO electrolytes, four characteristic microstructural regions can be identified.112 Region I occurs between the Tg and the melting temperature of the PEO. Tg varies with the composition of the electrolyte, but the melting temperature of the PEO is always around 65 °C. This region is perhaps the most complexed one because it consists of a mixture of crystalline PEO, amorphous solution of salt in PEO and one or more crystalline PEO–salt complex(es). The conductivity in this region is generally low and the energy of activation is high. Region II starts at the melting temperature of PEO and extends up to temperatures at which a high melting crystalline PEO–salt complex begins to form. Not all divalent PEO electrolyte compositions form a complex with a high melting temperature, but many do. One example is the PEO16–PbBr2 electrolyte, in which the complex starts to form at 85 °C and melts above 135 °C. This second region is characterised by high ionic conductivity with an activation energy lower than Region I. Region III is defined as the region where the high melting complex is beginning to form up to its melt. The formation of this complex tends to stiffen the electrolyte, lowering conductivity, although this effect may be only slight. Region IV is the zone above the melting point of the high temperature complex and apparently is a reasonably homogeneous region and of high conductivity. The electrolytes usually decompose above 200 °C. Drying with heat is often avoided when preparing polymer electrolyte films, because it induces the formation of spherulites, not only of the pure polymer but also of salt–polymer complexes. Using optical microscopy and infrared spectroscopy, Wendsjö and colleagues found that the process of annealing applied to PEOn–M(CF3SO3)2 electrolyte films with n = 9 and M = Zn or Pb did not cause any change in results compared with the same electrolytes without any heat treatment. But they also found that, for the n = 16 composition, this treatment led to an acceleration in the growth of spherulites associated with pure PEO.88 Any process that involves heating can also introduce at least partial melting of existing spherulites. The extension of this fusion or any change in the subsequent cooling may affect the recrystallisation process and change the amount and form of the residual amorphous phase where the conduction occurs. Other examples exist. In a systematic study of the degree of crystallisation in a system family with halides of divalent cations dissolved in PEO, it



365

was found that the PEO4–ZnCl2 electrolyte prepared at a drying temperature of 70 °C contained a greater amount of crystalline complex relative to amorphous material than when prepared at 25 °C. Therefore, the higher temperature accelerated the crystallisation process in this case as well.138 In addition to pure PEO crystallites, which melt at ~65 °C, and salt-rich complex crystallites that melt above 110 °C, crystalline material with an intermediate melting temperature may also exist, depending on the salt used.145,146 Glasse and his colleagues studied PEO–ZnX2 systems with X = I, Br, Cl, obtained from aqueous solutions and dried at room temperature, and found spherulites with melting temperatures of around 75–85 °C for electrolytes with zinc bromide and zinc iodide. The higher ionic conductivity observed for PEO–ZnCl2 electrolytes was explained by the absence of spherulites in these films.106

9.5

Ionic association: influence on ionic transport properties

One of the most important factors underlying the mechanism of ionic transport in polymer electrolytes is the degree and nature of the existing ionic association. A knowledge of the local structure surrounding the potentially mobile ion is important for understanding the conductivity mechanism. Studies of ionic association in polyethers have been performed with a number of divalent cations, including Cu2+,101 Ca2+,101,147 Mg2+, Cd2+,147 Sn2+,100 Ni2+,92,148 Zn2+ 88,92,147,148 and Pb2+.88,92,100,102,148 One of the problems of using X-ray diffraction methods to obtain the local structure at an atomic scale is that a substantial part of the information belongs to sample areas not directly related to the conducting species. It is difficult to obtain structural information regarding the amorphous phase by these methods. Therefore, the EXAFS (extended X-ray absorption fine structure) technique is used for completely amorphous materials. In a semicrystalline electrolyte, the structures determined in the crystalline and amorphous phases are effectively the same.149 Although the crystalline phase also contains minor quantities of amorphous material, there is no indication that two spectra are superimposed. The determination of crystal structures may give useful insights into the structural form of the ion-conducting amorphous phase. With this new technique, it is possible to obtain information about the local structure around a particular atom whether the material is glassy, amorphous, liquid or crystalline. We may be able to determine whether any ion pairing between cation and anion exists within a 4 Å separation.150 For technical reasons, light elements such as Li cannot be analysed by this technique. However, it is excellent for the range of divalent cations from Mg2+ to Pb2+,112 enabling us to study the structure/property relationships that influence ion transport in materials of this type.


366


Andrews and colleagues made the first use of this technique in a divalent polymer electrolyte, studying the PEO4–CaI2 electrolyte. They found that 10 oxygen atoms in the first coordination layer150 surrounded Ca2+ ions. However, it was not possible to determine whether the contributions of oxygen atoms derived from the PEO matrix or from a possible hydration layer, since the conditions did not exclude the presence of H2O. The study indicated the absence of ion pairs between the first two layers of the neighbouring cation. Studies by EXAFS on anhydrous PEOn–ZnX2 complexes (X = I, Br and n = 8–30), suggested that the Zn cation is coordinated to two halide ions and to four or six oxygen atoms when X = I or X = Br, respectively.91 Such a local structure indicates the formation of ion pairs. The PEOn–M(CF3SO3)2 electrolyte system with M = Zn, Pb and n = 9, 16 and 24 also presented evidence, in a FTIR study, for the formation of ion pairs and larger clusters, to a greater extent in the more concentrated samples.88 An EXAFS study of the PEOn–ZnBr2 system showed that, for PEO6–ZnBr2, the coordination number of the Zn in relation to the oxygen atoms was 3.5 and for the electrolytes with n = 10 and n = 15, rose to approximately 6.151 Caminiti and colleagues found an octahedral coordination to the Zn2+ cation in the PEG20–ZnTf2 solution, obtained with six oxygen atoms of the PEG polymer,152 indicating a stoichiometry of n = 6 for the intermediate compound. The results of a study on the local vicinity of the zinc ion in PEO4–ZnBr2 and PEO8–ZnBr2 electrolytes suggested that, on average, each zinc ion seemed to have about two oxygen atoms,153 thus indicating a composition of approximately n = 2 for a possible complex existing in these electrolytes. In a structural study of the PEO8–NiBr2 electrolyte conducted by Cai and colleagues, the local atomic structure of the anion was determined by AXS (anomalous (resonance) X-ray scattering), a X-ray diffraction technique that leads to more precise information regarding the coordination number and the structure in the middle distance than EXAFS.96 It was found that, on average, the Br− ion is close to one Ni2+ ion and to four carbon atoms, with distances consistent with those found by EXAFS between the Ni2+ ion and the nearest ligands, four oxygen atoms and two bromide ions. The results indicate that the Br− species in PEO8–NiBr2 may be in two different environments: one associated with the Ni2+ ions and the other with the carbon atoms of the PEO chains. This structural configuration clearly shows the formation of ion pairs and indicates that the Br− ions have weaker bonds in their local vicinity compared to Ni2+ ions. Therefore, it is not surprising that the conductivity data indicate that PEO8–NiBr2 is a better conductor of anions than of cations.25 In all the above results, an association is always observed between cations and anions of the electrolyte salts. However, there are cases where this does



367

not happen, for instance in a polymeric electrolyte system consisting of a crosslinked polymer matrix, poly(ethylene oxide)-grafted poly(methacrylate) (PEO–PMA) with a plasticising agent, PEGDE, and the magnesium salt, Mg(CF3SO3)2. Here, most Mg2+ ions were found by Raman spectroscopy to be isolated from the counter anion.154 The magnesium salt (MgX2) dissolved in (PEO–PMA)/PEGDE is fully dissociated to Mg2+, coordinated by EO units and to X−. In a molecular dynamic model proposed by Lin and colleagues, the amorphous regions of PEO used to study ionic transport mechanisms were represented. The structural characteristics and dynamics of the model were compared with those obtained experimentally.155 However, the model did not take into account the non-equilibrium condition of the amorphous regions, as the interesting temperatures for polymer electrolytes are between glass transition and melting.156 Although it is not yet possible to include amorphous and crystalline regions in a realistic form in the simulation, it is possible to develop methods that can subtract the crystalline contributions from the data, in order to extract a signal attributable to the amorphous regions.156 Aabloo and Thomas demonstrated that this method of simulation can be used to model ion behaviour on the PEO surface.157 A simulation carried out on the PEO–NdCl3 electrolyte predicted the gradual appearance of an amorphous region near the polymer surface, and a tendency for the Nd3+ ion to coordinate to a Cl− ion near the surface and to two Cl− ions when deeper into the PEO structure.

9.6

Phase diagrams: crystallinity and conductivity

For the same type of salt, the relative proportions of the phases generally observed in PEO-based electrolytes strongly depend on temperature and salt concentration. At low temperatures, an amorphous ‘intracrystalline’ phase may dominate relative to the amorphous ‘intercrystalline’ phase. The first phase is formed during the crystallisation process of both the pure PEO and the PEO–salt complex, when salt concentration changes and/or crosslinking of polymer chains prevent the further growth of polymeric crystals. This phase generally exists as microdomains embedded in spherulites or in the limits of polymeric crystals, not being in thermodynamic equilibrium with the crystalline phase, thus its proportion and salt concentration vary significantly, depending on the mechanism and crystallisation kinetics. It is assumed that ionic conductivity at low temperatures is related to ionic mobility in the ‘intracrystalline’ amorphous phase. The ‘intercrystalline’ amorphous phase is present in a significant proportion only at temperatures above the melting temperature of pure PEO or in the eutectic, and is in


368


thermodynamic equilibrium with the crystalline PEO–salt complex. In this case, the proportion of ‘intercrystalline’ amorphous phase and its salt concentration vary, depending on the temperature according to thermodynamic equilibrium. It is in the presence of this phase that ionic conductivity increases significantly. The establishment of phase diagrams for polymer electrolyte systems where the stable regions of the various crystalline phases and the ‘intercrystalline’ amorphous phase are defined, determining the relationship between these and their characterisation as a function of temperature and composition of the electrolyte, is therefore extremely useful, e.g. in predicting the ionic conduction behaviour of electrolytes. These phase diagram studies were initiated because ionic conductivity studies showed considerable variations as a function of several experimental parameters, such as thermal history.112 The relationship between the levels of conductivity obtained for a given electrolyte and, for example, the volume fraction of coexisting crystalline and amorphous phases, is not totally direct or simple. There are, simultaneously, many other factors involved, including nature, concentration and mobility of charge carriers. However, phase diagrams contribute descriptively, at a certain level, to a better understanding and explanation for the dependence of a particular property on salt concentration and temperature. Nevertheless, it is necessary to take certain issues into account: •

•

•

Phases that are in a non-equilibrium thermodynamic state may be present in polymer electrolytes. Because the ionic transport mechanisms and crystallisation kinetics can be very slow (especially at low temperatures), the system may be far from equilibrium. An electrolyte formed at a precise composition corresponding to an identified crystalline complex normally contains crystalline and amorphous material at that composition and at all temperatures below the melting temperature of the complex. In this case, there may be a divergence between the result expected for a certain property based on the phase diagram and that which is actually obtained in practice. The crystalline and liquid states are not as well defined as those in systems of small molecules, atoms or ions. In polymers, the presence of a crystalline phase can have a significant effect on the properties of the adjacent liquid (amorphous) phase, because the polymer chains in this latter phase may be ‘fixed’ as part of the crystalline phase. The behaviour of an entirely amorphous electrolyte is different from an amorphous phase with the same composition, in contact with a crystalline phase. What is usually considered pure crystalline polymer also contains amorphous regions that can dissolve salts, up to a determined limit of solubility. In completely equilibrated systems, below the eutectic melting



369

temperature, salts dissolved in the polymer ‘intracrystalline’ amorphous phase can be responsible for the conductivity of polymer electrolytes. Phase diagrams have been constructed based on information derived from different techniques, such as thermal analysis differential scanning calorimetry (DSC), differential thermal analysis (DTA), thermogravimetric analysis (TGA), X-ray diffraction, NMR, optical microscopy and conductivity studies. As stated above, we must take into account that the possibly slow nature of ionic transport and crystallisation kinetics, especially at low temperatures, may mean that the system is far from equilibrium. In most cases, hystereses observed in measurements of the physical properties of polymer electrolytes occur because equilibrium phases have not yet been established. The thermal behaviours of PEO-based polymer electrolytes with salts of multivalent cations have also been studied, although few corresponding phase diagrams have been constructed. One of the few phase diagrams completed has been established for the PEOn–ZnCl2 system.158 The proposed phase diagram agrees with results obtained by other authors.106,107,138,159 For an electrolyte with n = 4, Wendsjö and Yang found a single crystalline phase due to a complex being formed between ZnCl2 and PEO.138 Yang and Farrington obtained the same result.107 In a study by Kim and Bae, the phase diagram for this system was constructed by applying a thermodynamic molecular model that considered both ion–polymer and ion–ion interactions. The liquidus line, which separates the completely amorphous region from the region that contains the crystalline complex in equilibrium with the ‘intercrystalline’ amorphous phase (molten PEO), was slightly shifted to higher temperatures, i.e. the crystalline complex melted at a slightly higher temperature (about 20 °C) for polymer electrolyte compositions of n = 8 and n = 12.159 In the same study, the transition temperatures of samples of each electrolyte and of the pure PEO were determined by optical microscopy using TOA (thermooptical analysis). The electrolyte was pre-dried at 90 °C, which may have caused changes in the microstructure, specifically increasing the crystallinity of the samples and increasing the melting temperature of the crystalline intermediate complex. The study found a eutectic with a melting temperature close to 50 °C and with a composition corresponding to a mass fraction of ZnCl2 of 0.14159 and 0.16.158 Bermudez and colleagues also constructed a phase diagram for the same system, but their proposal was different.160 They found a eutectic mixture of two intermediate compounds and free crystalline zinc chloride in the PEO4–ZnCl2 electrolyte. As for the PEOn–ZnCl2 system, Yang and Farrington found (by DSC analysis) an intermediate complex for n > 4 compositions, melting in the


370


temperature range 75–175 °C for zinc bromide and zinc iodide PEO-based systems.107 These authors attributed the shape of this melting peak to the possible presence of a poorly chemically defined complex or to the gradual dissolution of the stoichiometric phase. The observed melting temperature of a complex with n = 4 stoichiometry formed in the PEO–ZnBr2 system was 230 °C for PEO4–ZnBr2. Other authors have also constructed phase diagrams of polymer electrolytes with divalent cations116,119,161–163 and trivalent cations164,165 with the majority also observing only one intermediate complex at room temperature and coexisting with the crystalline phase of PEO. The melting temperature of this complex is generally relatively high. For the divalent system studied by Nunes et al., consisting of Zn(CF3SO3)2doped di-ureas ormolytes with ∞ > n ≥ 1, materials obtained with n ≥ 5 were completely amorphous. Only in the n = 1 compound was a crystalline PEO– Zn(CF3SO3)2 complex formed. The stoichiometry of this phase is not known.81 For the PEOn–ZnI2 system, X-ray diffraction analysis revealed the existence of two intermediate compounds, unlike the analogous PEOn–ZnCl2 system.58 One of the complexes (designated as C1), with a concentration in the range n = 4–8, was present in all electrolytes studied and the other, C2, was only present in the PEO4–ZnI2 electrolyte with a composition below n = 4. The equilibrium phase diagram of the PEOn–ZnI2 system was also established.59 The result obtained for the PEO6–ZnI2 electrolyte,106 together with results for different salt concentrations,59 indicates that the composition of the C1 complex lies in the range between 4 and 6. The eutectic value was defined to be in the range n < 6. Figure 9.3 illustrates the phase diagram constructed for the PEO–ZnI2 system. The results obtained by Glasse and his colleagues106 agree with the constructed equilibrium phase diagram. Other authors who have constructed a phase diagram for system compositions with n ≥ 8 obtained melting temperatures for the C1 complex119,159,166 higher than those of the above studies.59,106 However, during synthesis, the electrolytes were heated to high temperatures (90 °C), which may have caused a change in the crystallinity of the samples and an increase in the melting temperature of the C1 complex crystals. This complex has also been defined with a n < 8 composition. Kim and Bae, studying the phases of this system over the concentration range 0.05– 0.5 (mass fraction of salt), considered the existence of a semicrystalline complex with n < 8 stoichiometry,159 which agrees with the above results. Ko and colleagues also found an intermediate compound with n < 8, in a study of a hyperbranched polymeric system based on polyols complexed with zinc iodide.166 In a more recent study, for more diluted electrolytes of the PEO–ZnI2 system, with n = 12 and n = 16, a crystalline complex was also found to exist alongside the pure PEO.119



371

n (O/Zn) 40

16

12

4

8

100

T / °C

75

L

L C1 + L

PEO + L

C1 + L

C2 + L

50

C1 + C2

PEO + C1 25

0 0

0.2

0.4

0.6

XZnI2(p/p)

9.3 Equilibrium phase diagram constructed for the PEOn–ZnI2 polymeric system. The transition temperatures were observed by: DTA59 䉬; DSC106 ⵧ.

Polymeric systems based on PEO with salts of divalent cations such as Ca2+ 27 and Pb2+ 161 also showed equilibrium phase diagrams with two crystalline complexes present in a certain composition range and two different eutectics. For the PEO–Ca(ClO4)2 system, the eutectics revealed by thermal analyses were one between 20 : 1 and 12 : 1, and the other close to 8 : 1, with the melting temperature below 25 °C.27 In addition to the pure PEO, which exists in more diluted samples, two more crystalline phases, with melting points above 200 °C, were found. PEO–Cu(CF3SO3)2 also showed two intermediate compounds with n = 11 and n = 6 stoichiometries. The partial phase diagram constructed by Passerini and colleagues suggested the existence of a eutectic with the composition O : Cu of 9 : 1, formed between pure PEO and a crystalline complex.116 Studies of thermal and X-ray diffraction analyses carried out on this system showed the presence of a crystalline complex with a melting temperature between 80 and 90 °C in an electrolyte with molar ratio composition of 5 : 1. Unlike compositions of 50 : 1 and 10 : 1, this


372


more concentrated electrolyte showed no evidence for the existence of pure PEO.27 The PEO–La(ClO4)3 system has more straightforward features than the above. Only a crystalline complex is shown below a molar ratio of O : La = 12 : 1, and the depression observed in the DSC endothermic small peak associated with the PEO indicated a eutectic point near a composition O : La of 10 : 1, with a melting temperature of around 50 °C.27 Singh and colleagues cooled PEO–Cu(ClO4)2 electrolytes that had been taken to a temperature above the melting temperature of the crystalline phases. Using thermomicroscopy, they found that only the 49 : 1 composition showed reproducibility in the fusion and recrystallisation of the crystalline phase.167 The most concentrated samples (8 : 1, 12 : 1 and 16 : 1) remained completely amorphous and spherulitic growth at room temperature was observed, only after a few days, thus this system showed a slow crystallisation kinetic. Slow recrystallisation kinetics have also been demonstrated in PEOn– EuBr3 and PEOn–EuCl3 systems.30 Electrolytes with n = 16 formed eutectics between the PEO and crystalline complexes, with melting points around 38–40 °C, being completely amorphous over 40 °C. A salt-rich complex formed in the PEOn–MgCl2 system with a stoichiometry estimated between 2.7 and 3.3, recrystallised easily after rapid cooling of the molten material, and more easily than the pure PEO, even for compositions with low salt concentration.17 As in systems with salts of monovalent ions, it is apparent that the recrystallisation of pure PEO decreases considerably due to the dissolution of salt in the elastomeric phase. For PEO12–MgCl2, this process occurs at 10 °C, while for electrolytes with n = 16 and n = 24, the temperature of recrystallisation of PEO is around 30–40 °C. Owing to the high concentration of MgCl2, the recrystallisation of the small amount of the PEO phase that exists in salt-rich compositions (n = 4 and n = 8) does not occur. The melting temperature of pure PEO, which coexists with the PEO–salt complex and an elastomeric phase with different MgCl2 concentrations at ambient temperature, is 55 °C. Above this temperature, the complex gradually dissolves and thermal analysis showed an endothermic broad peak associated with that process, with the maximum in the range 148–160 °C. Similarly to monovalent cation systems, multivalent cation polymeric electrolytes sometimes show salt precipitation at high temperatures, which results from the crystalline complex with high melting point. Examples are the PEO–NiBr2,25 PEO–PbBr2138 and PEO–MnBr2112 systems. Equilibrium phase diagrams help to interpret the dependence of ionic conductivity on salt concentration and temperature in polymeric systems with crystalline phases. They indicate the regions where ionic conduction is expected to be higher due to the absence of crystalline phases. However,



373

they should not be considered as true equilibrium phase diagrams because these systems are not thermodynamically stable.

9.7

Conclusions

This chapter was not intended to cover all the areas of science concerning polymer electrolytes, but rather to contribute towards a better understanding of some of the important aspects of systems containing multivalent cation salts and their behaviour. Examples of work carried out in the field of ionic transport have given a general overview of the influence of crystallinity and thermal behaviour of polymeric systems on the relevant electric properties of multivalent polymer electrolytes.

9.8

References

1. armand m b, chabagno j m and duclot m j, in Vashishta P, Mundy J N and Shenoy G K, Fast Ion Transport in Solids, Amsterdam, Elsevier, 1979, p.131. 2. gray f m, Solid Polymer Electrolytes – Fundamentals and Technological Applications, New York, VCH, 1991. 3. angell c a, xu k, zhang s-s and videa m, Solid State Ionics, 1996, 86–88, 17. 4. croce f, appetecchi g b, persi l and scrosati b, Nature, 1998, 394, 456. 5. vincent c a, Chemistry in Britain, 1989, 25, 391. 6. bruce p g and vincent c a, J. Chem. Soc., Faraday Trans., 1993, 89, 3187. 7. bernson a and lindgren j, Solid State Ionics, 1993, 60, 31. 8. jenkins a l and murray g m, J. Chem. Ed., 1998, 75, 227. 9. blumberg a a, pollack s s and hoeve c a j, J. Polym. Sci.: Part A, 1964, 2, 2499. 10. fenton d e, parker j m and wright p v, Polymer, 1973, 14, 589. 11. wright p v, Br. Polym. J., 1975, 7, 319. 12. papke b l, ratner m a and shriver d f, . J. Electrochem. Soc., 1982, 129, 1694. 13. tominaga y, ito k and ohno h, Polymer, 1997, 38, 1949. 14. doyle m, fuller t f and newman j, Electrochim. Acta, 1994, 39, 2073. 15. vincent c a, Prog. Solid State Chem., 1987, 17, 145. 16. pearson r g, J. Am. Chem. Soc., 1963, 85, 3533. 17. yang l l, mcghie a r and farrington g c, J. Electrochem. Soc., 1986, 133, 1380. 18. bruce p g, krok f and vincent c a, Solid State Ionics, 1988, 27, 81. 19. iwamoto r, saito y, isihara h and tadokoro h, J. Polym. Sci., 1968, 6, 1509. 20. yokoyama m, ishihara h, iwamoto r and tadokoro h, Macromolecules, 1969, 2, 184. 21. fontanella j j, wintersgill m c, calame j p and andeen c g, J. Polym. Sci.: Polym. Phys. Ed., 1985, 23, 113. 22. huq r, chiodelli g, ferloni p, magistris a and farrington g c, J. Electrochem. Soc., 1987, 134, 364. 23. abrantes t m a, alcácer l j and sequeira c a c, Solid State Ionics, 1986, 18–19, 315. 24. patrick c a j, glasse m d, latham r j and linford r g, Solid State Ionics, 1986, 18–19, 1063. 25. huq r and farrington g c, J. Electrochem. Soc., 1988, 135, 524.


374


26. bonino f, pantaloni s, passerini s and scrosati b, J. Electrochem. Soc., 1988, 135, 1961. 27. gray f m, vincent c a, bruce p g and nowinski j, Second International Symposium on Polymer Electrolytes, Amsterdam, Elsevier, 1990, p.299. 28. falcão e, silva m m and smith m j, Port. Electrochim. Acta, 1997, 15, 233. 29. silva c j r and smith m j, Electrochim. Acta, 1995, 40, 2389. 30. huq r and farrington g c, Second International Symposium on Polymer Electrolytes, Amsterdam, Elsevier, 1990, p.273. 31. carlos l d and videira a l l, Chem. Phys. Lett., 1997, 264, 57. 32. videira a l l and carlos l d, J. Chem. Phys., 1996, 105, 8878. 33. silva m m and smith m j, Electrochim. Acta, 2000, 45, 1463. 34. machado a s r and alcácer l, Second International Symposium on Polymer Electrolytes, Amsterdam, Elsevier, 1990, p.283. 35. silva m m, gonçalves n, smith m j and lightfoot p, Electrochim. Acta, 1998, 43, 1511. 36. vincent c a, in Linford R G, Electrochemical Science and Technology of Polymers – 2, New York, Elsevier, 1990. 37. chowdari b v r, huq r and farrington g c, Solid State Ionics, 1992, 57, 49. 38. chowdari b v r, huq r and farrington g c, Electrochim. Acta, 1992, 37, 1667. 39. atchia s, petit j-p, sanchez j-y, armand m and deroo d, Electrochim. Acta, 1992, 37, 1599. 40. morita m, araki f, yoshimoto n, ishikawa m and tsutsumi h, Solid State Ionics, 2000, 136–137, 1167. 41. armand m, Faraday Discuss. Chem. Soc., 1989, 88, 65. 42. berthier c, gorecki w, minier m, armand m b, chabagno j m and rigaud p, Solid State Ionics, 1983, 11, 91. 43. atchia s, gorecki w, armand m and deroo d, Electrochim. Acta, 1992, 37, 1743. 44. bartolotta a, di marco g, lanza m and carini g, J. Polym. Sci.: Part B, 1995, 33, 93. 45. james d b, wetton r e and brown d s, Polymer, 1979, 20, 187. 46. yang l-l, huq r, farrington g c and chiodelli g, Solid State Ionics, 1986, 18–19, 291. 47. vogel h, Phys. Z., 1921, 22, 645. 48. fulcher g s, J. Am. Ceram. Soc., 1925, 8, 339. 49. larsson r and andersson o, Electrochim. Acta, 2003, 48, 3481. 50. souquet j-l, duclot m and levy m, Solid State Ionics, 1996, 85, 149. 51. ferry a, J. Phys. Chem. B, 1997, 101, 150. 52. lee y c, ratner m a and shriver d f, Solid State Ionics, 2001, 138, 273. 53. furtado c a, silva g g, machado j c, pimenta m a and silva r a, J. Phys. Chem. B, 1999, 103, 7102. 54. hou w h and chen c y, Electrochim. Acta, 2004, 49, 2105. 55. reddy m j and rao u v s, J. Mater. Sci. Lett., 1998, 17, 1613. 56. minier m, berthier c and gorecki w, J. Phys., 1984, 45, 739. 57. ionescu-vasii l l, abu-lebdeh y and armand m, Solid State Ionics, 2005, 176, 2769. 58. plancha m j c, rangel c m and sequeira c a c, Solid State Ionics, 1999, 116, 293. 59. plancha m j c, PhD Thesis, IST, Technical University of Lisbon, 2008. 60. pandey g p, agrawal r c and hashmi s a, J. Power Sources, 2009, 190, 563.



375

61. huang f, wang x and li s, J. Macromol. Sci. – Chem. A, 1991, 28, 175. 62. lascaud s, perrier m, armand m, prud’homme j, kapfler b, vallée a and gauthier m, Electrochim. Acta, 1998, 43, 1407. 63. jeong s-k, jo y-k and jo n-j, Electrochim. Acta, 2006, 52, 1549. 64. ye h and xu j j, J. Power Sources, 2007, 165, 500. 65. reddy m j and chu p p, Solid State Ionics, 2002, 149, 115. 66. miyamoto t and shibamaya k, J. Appl. Phys., 1973, 44, 5372. 67. druger s d, nitzan a and ratner m a, J. Chem. Phys., 1983, 79, 3133. 68. ogata n, J. Macromol. Sci., Part C – Polymer Reviews, 2002, 42, 399. 69. liebenow c, Electrochim. Acta, 1998, 43, 1253. 70. silva m m, barros s c, smith m j and maccallum j r, Electrochim. Acta, 2004, 49, 1887. 71. siqueira l j a and ribeiro m c c, J. Chem. Phys., 2006, 125, 214903. 72. sanders r a, snow a g, frech r and glatzhofer d t, Electrochim. Acta, 2003, 48, 2247. 73. hayashi a, wang l and angell c a, Electrochim. Acta, 2003, 48, 2003. 74. liang w-j, chen y-p, wu c-p and kuo p-l, J. Phys. Chem. B, 2005, 109, 24311. 75. di noto v and vittadello m, Solid State Ionics, 2002, 147, 309. 76. maccallum j r, tomlin a s and vincent c a, Eur. Polymer J., 1986, 22, 787. 77. sivaraman p, thakur a p, kushwaha r k, ratna d and samui a b, J. Appl. Electrochem., 2008, 38, 189 78. kim j y and bae y c, Fluid Phase Equilibria, 1999, 163, 291. 79. kumar g g and sampath s, Polymer, 2004, 45, 2889. 80. kumar g g and sampath s, Solid State Ionics, 2005, 176, 773. 81. nunes s c, zea de bermudez v, silva m m, smith m j, morales e, sá ferreira r a, carlos l d and rocha j, Solid State Sciences, 2006, 8, 1484. 82. vatsalarani j, kalaiselvi n and karthikeyan r, Ionics, 2009, 15, 97. 83. cameron g g, ingram m d and sorrie g a, J. Chem. Soc., Faraday Trans. 1, 1987, 83, 3345. 84. schantz s, J. Chem. Phys., 1991, 94, 6296. 85. albinsson i, mellander b e and stevens j r, Solid State Ionics, 1993, 60, 63. 86. cameron g g, harvie j l, ingram m d and sorrie g a, Br. Polym. J., 1988, 20, 199. 87. kakihana m, schantz s and torell l m, J. Chem. Phys., 1990, 92, 6271. 88. wendsjö å, lindgren j and paluszkiewicz c, Electrochim. Acta, 1992, 37, 1689. 89. pandey g p and hashmi s a, J. Power Sources, 2009, 187, 627. 90. chintapalli s and frech r, Electrochim. Acta, 1995, 40, 2093. 91. cole m, sheldon m h, glasse m d, latham r j and linford r g, Appl. Phys. A, 1989, 49, 249. 92. wendsjö å, lindgren j, thomas j o and farrington g c, Solid State Ionics, 1992, 53–56, 1077. 93. chintapalli s, frech r and grady b, Polymer, 1997, 38, 6189. 94. cameron g g, harvie j l and ingram m d, Solid State Ionics, 1989, 34, 65. 95. brodin a, mattson b and torell l m, J. Chem. Phys., 1994, 101, 4621. 96. cai h, hu r, egami t, farrington g c, schlindwein w s, latham r j, linford r g and pynenburg r, Electrochim. Acta, 1992, 37, 1663. 97. cai h and farrington g c, J. Electrochem. Soc., 1992, 139, 744. 98. amir n, vestfrid y, chusid o, gofer y and aurbach d, J. Power Sources, 2007, 174, 1234.


376


99. latham r j, linford r g, pynenburg r a j, schlindwein w s and farrington g c, J. Chem. Soc., Faraday Trans., 1993, 89, 349. 100. bergstrom p a and frech r, J. Phys. Chem., 1995, 99, 12603. 101. petersen g, torell l m, panero s, scrosati b, silva c j and smith m, Solid State Ionics, 1993, 60, 55. 102. rhodes c p, kiassen b, frech r, dai y and greenbaum s g, Solid State Ionics, 1999, 126, 251. 103. xie l and farrington g c, Solid State Ionics, 1992, 53–56, 1054. 104. teeters d and frech r, Solid State Ionics, 1986, 18–19, 271. 105. wintersgill m c, fontanella j j, greenbaum s g and ademic k j, Br. Polym. J., 1988, 20, 195. 106. glasse m d, linford r g and schlindwein w s, Second International Symposium on Polymer Electrolytes, Amsterdam, Elsevier, 1990, p.203. 107. yang h and farrington g c, J. Electrochem. Soc., 1992, 139, 1646. 108. bruce p g and vincent c a, Faraday Discuss. Chem. Soc., 1989, 88, 43. 109. evans j, vincent c a and bruce p g, Polymer, 1987, 28, 2324. 110. bruce p g, hardgrave m t and vincent c a, Electrochim. Acta, 1992, 37, 1517. 111. shi j, bruce p g and vincent c a, Electrochim. Acta, 1992, 37, 1707. 112. farrington g c and linford r g, in MacCallum J R and Vincent C A, Polymer Electrolyte Reviews – 2, London, Elsevier, 1989, p.256. 113. lauenstein å, johansson a and tegenfeldt j, J. Electrochem. Soc., 1994, 141, 1819. 114. dissanayake m a k l, bandara l r a k, karaliyadda l h, jayathilaka p a r d and bokalawala r s p, Solid State Ionics, 2006, 177, 343. 115. huq r and farrington g c, Solid State Ionics, 1988, 28–30, 990. 116. passerini s, curini r and scrosati b, Appl. Phys. A., 1989, 49, 425. 117. gina m, panero s, scrosati b, cao x and greenbaum s g, Solid State Ionics, 1996, 83, 73. 118. plancha m j, rangel c m and sequeira c a c, J. Electroanal. Chem., 1998, 442, 91. 119. pai s j, bae y c and sun y k, J. Electrochem. Soc., 2005, 152, A864. 120. yang h and farrington g c, Second International Symposium on Polymer Electrolytes, Amsterdam, Elsevier, 1990, p.265. 121. kumar g g and sampath s, Solid State Ionics, 2003, 160, 289. 122. kumar g g and sampath s, J. Electrochem. Soc., 2003, 150, A608. 123. sørensen p r and jacobsen t, Electrochim. Acta, 1982, 27, 1671. 124. eigen m, Pure Appl. Chem., 1963, 6, 97. 125. cotton f a and wilkinson g, Advanced Inorganic Chemistry, New York, WileyInterscience Pub, 1988. 126. baril d, michot c and armand m, Solid State Ionics, 1997, 94, 35. 127. plancha m j c, rangel c m and sequeira c a c, Solid State Ionics, 1992, 58, 3. 128. bruce p g, nowinski j, gray f m and vincent c a, Solid State Ionics, 1990, 38, 231. 129. basolo f and pearson r g, Mechanisms of Inorganic Reactions, New York, Wiley, 1967. 130. johansson a and tegenfeldt j, Solid State Ionics, 1993, 67, 115. 131. martins m a g and sequeira c a c, J. Power Sources, 1990, 32, 107. 132. farrington g c, yang h and huq r, Mat. Res. Soc. Symp. Proc., 1989, 135, 319. 133. puga m m s, carlos l d, abrantes t m a and alcácer l, Electrochim. Acta, 1995, 40, 2383.


Characterisation and modelling of multivalent polymer electrolytes 134. 135. 136. 137. 138. 139. 140. 141. 142. 143. 144. 145.

146. 147. 148. 149. 150. 151. 152. 153. 154. 155. 156. 157. 158. 159. 160. 161. 162. 163. 164. 165. 166. 167.

377

carlos l d and videira a l l, Phys. Rev. B, 1994, 49, 11721. carlos l d, assunção m and alcácer l, J. Mater. Res., 1995, 10, 202. mehta m a, PhD Thesis, University of St Andrews, UK, 1992. gray f m, maccallum j r and vincent c a, Solid State Ionics, 1986, 18–19, 242. wendsjö å and yang h, Second International Symposium on Polymer Electrolytes, Elsevier, Amsterdam, 1990, p.225. casalbore-miceli g, yang m j, camaioni n, mari c m, li y, sun h and ling m, Solid State Ionics, 2000, 131, 311. lee k-y, chung w-s and jo n-j, Electrochim. Acta, 2004, 50, 295. yoshimoto n, yakushiji s, ishikawa m and morita m, Electrochim. Acta, 2003, 48, 2317. turkovic´ a, pavlovic´ m, dubcˇek p, lucˇ ic´-lavcˇ evic´ m, etlinger b and bernstorff s, J. Electrochem. Soc., 2007, 154, A554. wendsjö å, lindgren j and thomas j o, Proceedings of the Conference on Polymer Ionics, Gohemburg, Sweden, 1992, p.15. Shi J, PhD Thesis, University of St Andrews, UK, 1993. neat r j, hooper a, glasse m d and linford r g, in Poulsen F W, Andersen N H, Clausen K, Skaarup S and Sørensen O T, 6th Risø Intern. Symp. on Metallurgy and Materials Science, Roskilde, 1985, p.341. neat r, glasse m, linford r and hooper a, Solid State Ionics, 1986, 18–19, 1088. frech r and huang w, Solid State Ionics, 1993, 66, 183. wendsjö å, thomas j o and lindgren j, Polymer, 1993, 34, 2243. linford r g, Faraday Discuss. Chem. Soc., 1989, 88, 133. andrews k c, cole m, latham r j, linford r g, williams h m and dobson b r, Solid State Ionics, 1988, 28–30, 929. einset a g, schlindwein w s, latham r j, linford r g and pynenburg r, J. Electrochem. Soc., 1991, 138, 1569. caminiti r, carbone m, panero s and sadun c, J. Phys. Chem. B, 1999, 103, 10348. glasse m d, latham r j, linford r g and pynenburg r a j, Solid State Ionics, 1992, 53–56, 1111. yoshimoto n, tomonaga y, ishikawa m and morita m, Electrochim. Acta, 2001, 46, 1195. lin b, boinske p t and halley j w, J. Chem. Phys., 1996, 105, 1668. johnson j a, saboungi m l, price d l, ansell s, russell t p, halley j w and nielsen b, J. Chem. Phys., 1998, 109, 7005. aabloo a and thomas j o, Electrochim. Acta, 1998, 43, 1361. plancha m j, rangel c m, sequeira c a c and hudson m j, J. Appl. Electrochem., 1997, 27, 1290. kim j y and bae y c, Polymer, 1999, 40, 1979. bermudez v c z, morgado j, abrantes t m a and alcácer l, Second International Symposium on Polymer Electrolytes, Amsterdam, Elsevier, 1990, p.251. wendsjö å, lindgren j and mellander b e, Solid State Ionics, 1993, 59, 271. mehta m a, lightfoot p and bruce p g, Chem. Mater., 1993, 5, 1338. kim j y, noh s t and bae y c, Polymer, 1998, 39, 3473. smith m j and silva c j r, Solid State Ionics, 1992, 58, 269. smith m j, silva c j and silva m m, Solid State Ionics, 1993, 60, 73. ko s j, bae y c, kong s-h and ryu s-o, J. Polym. Sci.: Part B, 2004, 42, 2254. singh k, chiodelli g, magistris a and ferloni p, Second International Symposium on Polymer Electrolytes, Amsterdam, Elsevier, 1990, p.291.


10 Polymer electrolytes for dye-sensitized solar cells J. N. DE F R E I TA S, J. E. B E N E D E T T I, F. S. F R E I TA S, A. F. N O G U E I R A and M. A. DE PAO L I, University of Campinas – UNICAMP, Brazil

Abstract: Polymer electrolytes or gel-type polymer electrolytes are interesting alternatives to substitute liquid electrolytes in dye-sensitized solar cells (DSSC). The interest in this research field is growing, reflected in the increased number of papers published each year concerning these materials. This chapter presents a brief review of the history and development of polymer electrolytes aiming at the application in DSSC. Recent improvements achieved by modifications of the composition and by introduction of additives such as inorganic nanofillers, organic molecules and ionic liquids are described. The stability of DSSC assembled with these materials, and scaling-up of such devices are also discussed. Key words: dye-sensitized solar cell, polymer electrolyte, gel electrolyte, inorganic nanofiller, ionic liquid, plasticizer.

10.1

Introduction

Since the report by O’Regan and Grätzel1 in 1991, dye-sensitized solar cells (DSSC) have been under intensive investigation as a low cost alternative to explore solar energy. Usually these cells consist of a transparent electrode coated with a mesoporous film of nanocrystalline particles of TiO2 sensitized with a dye, an electrolyte containing a suitable redox-couple (usually I−/I3− in acetonitrile) and a Pt-coated counter-electrode. A schematic representation of a DSSC and the processes that occur during cell operation are depicted in Fig. 10.1. Since the DSSC consist of combinations of several materials, the properties of each component directly influence the kinetics and reactions. Thus, device performance depends on the morphology, optical and electrical properties of the porous semiconductor film; the chemical, electrochemical, photophysical and photochemical properties of the dye; the electrochemical and optical properties of the redox couple and solvent in the electrolyte; and the electrochemical properties of the counter-electrode.2 381 © Woodhead Publishing Limited, 2010

382

Polymer electrolytes e–

hv e–

e–

TiO2 D + hv

D*/D*+ e–

TiO2 D*

TiO2 D* TiO2 D+ + ecb

+

TiO2 D + ecb I3–

e–

I– D/D+ Electrolyte TiO2 electrode Dye Pt electrode

[10.1] [10.2]

TiO2 D

[10.3]

TiO2 D+ + 3/2 I– TiO2 D + 1/2 I3–

[10.4]

1/2 I3– + ePt I3– + 2ecb

3/2 I– 3 I–

[10.5] [10.6]

10.1 Representation of a DSSC and the processes involved in energy conversion (D represents the dye and I−/I3− are the charge mediators).

The efficiency of a DSSC in the process for energy conversion depends on the relative energy levels and the kinetics of electron transfer processes at the sensitized semiconductor|electrolyte interface. For efficient operation, the rate of electron injection (Fig. 10.1, equation 10.2) must be faster than the decay of the dye excited state. Also, the rate of re-reduction of the oxidized sensitizer (or dye cation) by the electron donor in the electrolyte (equation 10.4) must be faster than the rate of back reaction (recombination) of the injected electrons with the dye cation (equation 10.3), as well as the rate of reaction of injected electrons with the electron acceptor in the electrolyte (equation 10.6). This reaction, also called ‘dark current’, is the main loss mechanism for the DSSC.3 Finally, the kinetics of the reaction at the counter-electrode must also guarantee the fast regeneration of the charge mediator (equation 10.5), or this reaction could also become rate limiting in the overall cell performance.1,3,4 For the devices based on ruthenium complexes dyes, liquid electrolytes and nanocrystalline TiO2, the charge injection is a very fast process, usually in the femtosecond time domain. On the other hand, the recombination is much slower, and occurs over a much longer timescale (several microseconds or longer). This difference of several orders of magnitude for the forward and reverse electron transfer rates allows the efficient processing of the charge separated products, i.e. the reduction of the dye cation by iodide and the percolation of the injected electrons in the TiO2 film to arrive at the back contact.3,5 The rate for dye regeneration reaction (Fig. 10.1, equation 10.4) is also very important for the efficiency of the cell, since it affects the electron collection efficiency, i.e. the relative amount of electrons that leave the semiconductor and contribute to the photocurrent. Thus, for DSSC based on TiO2 nanoporous electrodes, Ru-bipyridyl complexes,


Polymer electrolytes for dye-sensitized solar cells

OH

383

O C O

R

C OH

N N

N Ru

N C

N

N

S C

R

S

glass

FTO

TiO2

R = COOH (N3) or COOTBA (N719)

10.2 Structure of the N3 and N719 dyes usually employed as sensitizers in DSSC (where TBA = tert-butylammonium); and SEM images of the TiO2 nanoporous film deposited on a SnO2 : F coated glass (FTO-glass).

such as the cis-bis(isothiocyanato) bis(2,2′-bipyridyl-4,4′-dicarboxylato)ruthenium (II) dye (N3, Solaronix), and liquid electrolytes containing the I−/I3− redox couple can have efficiencies as high as 11%.2,6 Figure 10.2 presents some of the materials usually employed in efficent DSSC. Despite the good performances obtained for DSSC with liquid electrolytes, there are still some questions about the presence of the liquid component, which requires perfect sealing in order to prevent leakage, and also limits the shape and stability of the cells. Many groups have focused on the substitution of the liquid electrolyte by solid or gel electrolytes, which could aid in reducing costs and make assembly of dye-sensitized solar cells easier. The main alternatives are inorganic or organic hole conductors, gel electrolytes prepared with ionic liquids or by the solidification of liquids, and polymer electrolytes. In this chapter, we present an overview of some recent developments in dye-sensitized solar cells assembled with polymer and gel electrolytes, especially concerning the modifications introduced to improve the ionic conductivity and mechanical stability of such materials, and how such modifications affect the performance of polymer-based DSSC. Recent


384


reviews on DSSC assembled with polymer and gel electrolytes can also be found elsewhere.7–12

10.2


The investigation of polymer electrolytes began in the 1970s, after the pioneering measurements of ionic conductivity in polymer–salt mixtures done by Wright and co-workers13 and the proposal of Armand and co-workers14 that these systems could be used in secondary batteries. After Wright’s work,13 the polyethers, such as poly(ethyelene oxide) (PEO) coordinated with a range of inorganic salts, such as LiI, NaI, LiClO4, LiCF3SO3, LiSCN, NaSCN, NaClO4 or LiPF6, became the classical systems.15 In PEO, the repeating unit (–CH2–CH2–O–) presents a favorable arrangement for effective interaction of the free electron pair on the oxygen with the alkali metal cations. This occurs because the PEO chains are arranged in a helical conformation with a cavity that presents ideal distances for oxygen–cation interactions. PEO presents a low glass transition temperature (Tg = −50 °C), but the regular structure favors a high crystallinity degree (80%), with melting point of Tm ~ 65 °C. For polymer electrolytes, ionic mobility is closely associated with local structural relaxations which occur in the amorphous phase. Thus solvent-free PEO–salt complexes usually exhibit conductivity in the range from 10−8 to 10−4 S cm−1 at temperatures between 40 and 100 °C, limiting practical applications at room temperature. The solid state nature of polymer electrolytes is an advantage; however, the ionic conductivity which occurs in the amorphous phase for the majority of the polymer electrolytes is too low for application in photoelectrochemical cells or batteries. To decrease the crystallinity degree of the polymer at ambient temperature, and thus increase the ionic mobility, it is necessary to introduce a certain degree of disorder in the structure. This can be achieved by using blends of different polymers, copolymers or crosslinked networks, which can either reduce the crystallinity of the polymer or lower the Tg. Also, it is possible to introduce a third component in the system, which can act in a similar fashion as a plasticizer, increasing the ionic conductivity, as will be discussed in the next sections.

10.2.1 Copolymers of poly(ethylene oxide) The use of PEO copolymers in electrolytes for solar cells began in 1999. Although PEO polymer electrolytes and their derivatives had been used before in combination with conducting polymers for the assembly of phototelectrochemical cells,16,17 the first DSSC assembled with a polymer electrolyte was reported by De Paoli and co-workers.18 The device was assembled using poly(o-methoxyaniline) as sensitizer and a copolymer of



385

poly(epichlorohydrin-co-ethylene oxide) containing NaI/I2 as electrolyte. The monochromatic photon-to-current conversion efficiency was 1.3% with 410 nm and 0.1% with 600 nm irradiation.18 One year later, the same electrolyte was applied in a device using a ruthenium complex as sensitizer, leading to an open circuit voltage (Voc) of 0.71 V, a short-circuit current (Jsc) of 0.46 mA cm−2 and overall conversion efficiency (η) of 0.22% under 120 mW cm−2 of white-light illumination.19 But it was only in 2001, 10 years after Grätzel’s first announcement of an efficient liquid-electrolyte based DSSC, that devices assembled with polymer electrolytes with efficiencies superior to 1% were reported.20,21 Since then, many efforts have concentrated on the search for new polymer electrolytes that could increase the efficiency of the device. The use of PEO copolymers was motivated by the work of De Paoli and co-workers,22 which showed a systematic investigation of the ionic conductivity and thermal properties of three copolymers of ethylene oxide (EO) and epichlorohydrin (EPI) with different monomer ratios. The ethylene oxide/epichlorohydrin ratios in the copolymers were 84/16, 60/40 and 50/50, and they were designated as P(EO–EPI)84–16, P(EO–EPI)60–40 and P(EO–EPI)50–50, respectively. Measurements of the ionic conductivity of copolymers containing LiClO4 showed that this parameter was dependent both on the concentration of salt and on the molar ratio between the comonomers. For example, the higher the content of ethylene oxide units in the copolymer, the higher was the ionic conductivity of the electrolyte, for the same salt concentration. The best ionic conductivity was obtained for P(EO–EPI)84–16 copolymer mixed with ~5.5 wt% of LiClO4 (4.1 × 10−5 S cm−1 at 30 °C). Cyclic voltammetry studies showed an electrochemical stability window in the range of 4.0 V. Temperature-dependence of the ionic conductivity for all copolymer–salt complexes was studied by the empirical Vogel–Tamman–Fulcher relation and it was found that the ionic conductivity in these complexes was strongly coupled to the flow behavior of the matrix, as also observed by other groups.23 The strong decrease of crystallinity observed for the copolymer P(EO–EPI)84–16 after addition of a small content of salt, together with the conductivity data, indicated that Li+ cations interact more strongly with the oxygen atoms from the ethylene oxide units than those from the epichlorohydrin units. The ionic conductivity attained in this work was considered sufficient to motivate the use of the P(EO– EPI)84–16 complexes as polymer electrolytes.22 Aiming at applications in solar cells, the thermal and ionic conductivity properties of the elastomer P(EO–EPI)84–16 filled with NaI or LiI and I2 were explored.19,24 Figure 10.3 shows the plots of the ionic conductivity of the polymer electrolyte as a function of salt concentration, at 26 °C. The ionic conductivity initially increases with the increase in salt concentration, owing to the increasing number of charge carriers. However, after reaching


386


s (S cm–1)

10–5

10–6 0

5

10

15

20

25

Salt content (wt%)

10.3 Effect of salt concentration on conductivity for polymer electrolytes based on P(EO–EPI), I2 and: (-䊏-) LiI and (-䊉-) NaI.

a maximum value, the conductivity decreases because higher amounts of salt lead to the formation of ion pairing and crosslinking sites that hinder the segmental motion of the polymer chains and, as a consequence, decrease ionic mobility.25 For this system, the highest conductivity was 1.5 × 10−5 S cm−1, for the sample containing 9 wt% of NaI.24 The Na+ ions interact with the EO repeating units of the polymer chains by means of Lewis-type acid–base interactions. The empirical Vogel–Tamman–Fulcher equation was used to model the conductivity and temperature relationships, indicating that conduction occurs in the amorphous phase of the copolymer.24 Surprisingly, the ionic conductivity was higher for the electrolyte prepared with NaI, when compared to the electrolytes prepared with LiI. At first, it was expected that Li+ cations would lead to a better conductivity, since they are smaller than Na+ cations. This behavior was attributed to the high energy that is necessary to dissolve LiI in the polymer matrix, in comparison to the energy necessary to dissolve NaI.26 DSSC were assembled using the polymer electrolyte consisting of P(EO– EPI)84–16, 9 wt% NaI and 0.9 wt% I2. The film of the polymer electrolyte was deposited onto the sensitized TiO2 electrodes by casting a solution of P(EO–EPI)84–16 with NaI and I2 in acetone, as presented in Fig. 10.4. The assembly of the cells was completed by pressing the Pt counter-electrode against the sensitized electrode coated with the polymer electrolyte. The electrolyte was also used as a kind of adhesive between the working and counter-electrodes, and no sealing step was necessary. The active area of the



387

(a)

Io

alt er

S Polym

di ne ne

Polymer electrolyte

Ac

Polymer electrolyte TiO2/sensitizer

Counterelectrode

e to

Electrode

Electrolyte deposition (casting)

Closing

DSSC

(b)

10.4 Schematic representation of the steps involved in assembling a ‘solid state’ DSSC using a polymer electrolyte.

cells was typically 1 cm2. Unsealed devices employing this polymer electrolyte achieved solar to electrical energy conversion efficiencies of 1.6% under 100 mW cm−2, and 2.6% under 10 mW cm−2.20 This was the first report of a dye-sensitized solar cell assembled with a pure polymer electrolyte with efficiency superior to 1%. Nevertheless, the efficiency of such ‘solid’ solar cells was much lower than the ~10% obtained for solar cells prepared with liquid electrolytes. The main drawback was attributed to the ionic conductivity of the polymer electrolyte, which was found to be about two orders of magnitude lower than the conductivity usually observed with liquid electrolytes. The open circuit voltage of these cells was surprisingly high (0.82 V). Analyses of liquid electrolyte cells have indicated that this voltage is primarily limited by recombination losses from injected electrons interacting with oxidized redox carriers in solution. These recombination losses accelerate as a function of cell voltage due to an increase in the electron density in the titania film. The high Voc value was assigned to the basic character of the polymer in the elecrolyte.27 Protons have been shown to increase the density of electrons in titania films at a fixed applied potential,28,29 most probably associated with the ability of these ions to intercalate into titania. The basic nature of the polymer used in the electrolyte is likely


388


to shift the voltage dependence of the electron density in the titania to more negative voltages, and therefore minimize the voltage dependence of the recombination losses within the cell. Comparable shifts in the voltage dependence of the recombination dynamics of dye-sensitized TiO2 films have been observed following the addition of a base to a liquid electrolyte.30 Nowadays, many polymers and copolymers with different structures are used to make polymer and gel-type polymer electrolytes for application in DSSC. Some structures of these materials are presented in Table 10.1. From Table 10.1 it can be seen that, for polymers with higher content of EO units, the crystallinity degree is higher than polymers with lower EO contents. The decrease of the linearity of polymer chains also contribuites to the lowest crystallinity degree. This occurs because the structural regularity and flexibility of the polymer allow the packing of chains and increase the crystallinity degree. Besides, in the cases where the chains possess large ramifications, the polymer appears to be amorphous, while for short ramifications only a decrease of the crystallinity degree is observed. Pure PEO has also been used in an electrolyte for DSSC, by combining this polymer with different amounts of KI and I2.31 Using Raman spectroscopy the authors showed the formation of polyiodide species in the electrolyte upon addition of different salt and iodine concentrations. The highest ionic conductivity achieved at room temperature was 8.4 × 10−5 S cm−1 for the electrolyte composition PEO : KI : I2; 12 : 1 : 0.1. Fourier transform infrared (FTIR) spectroscopy was carried out to show that the K+ ions can coordinate to the ether oxygen in PEO chains and a linear Arrhenius-type behavior was observed. DSSC assembled with this electrolyte presented Jsc = 6.1 mA cm−2, Voc = 0.59 V, fill factor FF = 0.56 and η = 2.0% under irradiation of 100 mW cm−2. Polymer electrolytes containing different salts, such as quaternary ammonium iodides,32 or different polymers, such as poly(butylacrylate)33 and poly(dimethylsiloxane),34 have also been used in DSSC. The major problems associated to the use of polymer electrolytes in DSSC arise from the low ionic diffusion in a more viscous medium, low penetration of the polymer inside the nanostructured TiO2 electrode, and an increase in the interfacial charge-transfer resistance between the electrodes and the electrolyte.35 In order to enhance the overall conversion efficiency and the transport properties, the nature/composition of the polymer systems must be improved. Further improvements in device performance are readily achievable through optimization of the ionic conductivity. In this context, the addition of inorganic nanofillers, ionic liquids, oligomers based on EO, plasticizers and other additives has become a common route to elaborate polymer (or gel) electrolytes with improved ionic conductivity properties.



2.2

1.3

1.3 1

0.41

−69

−62

−43

−53 −56

−68

−68

P(EO–EM)78–22

P(EO–EM)95–5

P(EO–EPI)50–50

P(EO–EPI)84–16 P(EO–EPI)87–13

P(EO–PO)56–44

P(EO–PO)79–21

0.27

1.6

Tg (°C)

Copolymer

Mw (×106 g mol−1)

0

31

8

12 14

–

0

Crystallinity degree (%)

2

( CH

2

( CH

2

( CH

CH2

CH2

CH2

O

O

O 2

2

(n( CH

2

(n( CH

(n( CH

Chemical structure

Table 10.1 Structure and properties of some PEO copolymers usually employed in electrolytes for DSSC

CH2

CH3

CH

Cl

2

O

O

O

( CH CH

O

CH2

CH

(m

(m

CH2

(m O

3

(2 CH

390


10.3

Plasticized and gel polymer electrolytes

Adding small molecules or oligomers with coordinating/solvating ability to the polymer is an interesting alternative to overcome the limitations inherent in a mixture of only polymer and salt. Such compounds are often called or referred to as ‘plasticizers’, although it is not necessarily true that these additives will act as a real plasticizer (i.e. by definition, a plasticizer must change the Tg and reduce the crystallinity degree of the polymer). These compounds usually possess low molar mass, high boiling points and are routinely added to highly crystalline polymer matrixes to increase the flexibility of the polymer chains. For polymer electrolytes, this additive contributes to increase the ionic conductivity by several orders of magnitude.36 In fact, DSSC assembled with polymer electrolytes containing plasticizers exhibit much higher efficiencies. However, an increase in the plasticizer content can also be followed by a loss in mechanical properties. It is also important to note that there is a tenuous line between plasticized polymer electrolytes and gel electrolytes. Classical plasticized polymers are routinely used in industry, when the plasticizers are added (in amounts up to 40 wt%) to a polymer, usually aiming at changing the mechanical properties, but always maintaining the solid state characteristics (phthalic acid esters added to poly(vinyl chloride) (PVC, for example). In a classical gel electrolyte, a tridimensional polymer network (not necessarily a coordinating polymer, in fact most gel electrolytes are made of ‘inert’ matrices) holds an organic solution of a salt. For an ‘inert’ gel electrolyte, (polyacrylonitrile, poly(methyl methacrylate) or poly(vinylidene fluoride) derivatives), the ionic transport occurs in the solution phase. However, if the polymer employed possesses the ability to strongly solvate the cation, then both phases can be responsible for ionic transport. As gel electrolytes usually contain a high fraction of liquid components, their mechanical properties are poorer than those observed for pure polymer electrolyte systems, or plasticized polymer electrolytes. The following section contain a brief review about significant improvements made in DSSC assembled with both plasticized and gel electrolytes, without distinguishing between them. In 2001, a dye-sensitized solar cell assembled with a gel network polymer electrolyte based on polysiloxane and PEO, containing 20 wt% of LiI, 5 wt% of I2 and 150 wt% of the mixture ethylene carbonate (EC)/propylene carbonate (PC) (3 : 1 v/v) was reported.21 EC is a high viscosity solvent with high dielectric constant, which is favorable for salt dissociation. However, this material has a tendency to crystallize at low temperature, causing phase separation between the plasticizer and the polymer matrix. Thus, organic solvents such as PC are used to form binary organic solvents with EC and homogeneous gel polymer electrolytes can be obtained. The fully crosslinked electrolyte presented ambient conductivity of 1.1 × 10−3 S cm−1 and



391

the solar cells showed Voc = 0.69 V and Jsc = 1.7 mA cm−2 under white light irradiation (28 mW cm−2), yielding an efficiency of 2.9%.21 Other carbonate molecules, such as diethyl carbonate, can be used in combination with EC or PC as plasticizers in electrolytes.37 The addition of 50 wt% of the plasticizer poly(ethylene glycol)methyl ether, P(EGME), with molar mass of ~350 g mol−1 to the polymer electrolyte based on P(EO–EPI)84–16, NaI and I2 increased the ionic conductivity without compromising the electrochemical, thermal and dimensional stabilities. This material was chosen as plasticizer for P(EO–EPI) owing to the similarity of its chemical structure with the copolymer matrix. In the P(EPI– EO)/NaI/I2 system the ionic conductivity of the sample containing 11 wt% of NaI reached a maximum value of 1.9 × 10−5 S cm−1. After addition of P(EGME), an overall increase in the ionic conductivity was observed for the system at all salt concentrations, reaching a maximum value of 1.7 × 10−4 S cm−1 at 13 wt% of NaI. The addition of the plasticizer to the polymer electrolyte even allowed the dissolution of higher amounts of salt without significantly changing the conductivity of the system, which remained on the plateau of 10−4 S cm−1.38 Figure 10.5 shows the Nyquist diagrams obtained for electrolytes based on the copolymer poly(ethylene oxide-co-2-(2-methoxyethoxy)ethyl glycidyl ether) containing 78% of EO units (P(EO–EM)) mixed with

20 000

O C

O

(

O CH2CH2

Z ˝ (Ω)

15 000

(n

O

C

5% Lil 10% Lil 15% Lil 20% Lil 25% Lil

10 000

5000

0 0

5000

10 000

15 000

20 000

Z´ (Ω)

10.5 Effect of salt concentration on impedance of the system for polymer gel electrolytes based on P(EO–EM), 30% of DIB, [I−]/[I3−] = 10. The structure of DIB is shown in the inset.


392


polyethylene glycol dibenzoate (DIB) and LiI/I2. This figure shows the effect of salt concentration on impedance of the ‘plasticized’ electrolytes, where it can be seen that the overall impedance of the system decreases after addition of salt, indicating lower values of resistance and capacitance in the electrolyte, in agreement with the values of ionic conductivity. Interesting results have also been obtained using γ-butyrolactone (GBL) as plasticizer for P(EO–EPI) copolymer. For the electrolyte prepared with GBL, P(EO–EPI)87–13, 15 wt% of NaI and I2 for example, the maximum ionic conductivity changed from 3 × 10−5 S cm−1 to 1 × 10−4 S cm−1 after addition of 50 wt% of GBL.39 The apparent diffusion coefficient of ionic species in the electrolyte with and without plasticizer was estimated using complex electrochemical impedance spectroscopy (EIS) and the equivalent circuit Rs[Q1(R1O)] (Fig. 10.6) was used to fit the data according to the method proposed by De Paoli and co-workers.40 The Nyquist diagrams are shown in Fig. 10.6. The EIS technique allows the evaluation of the processes involved in electrolytes or photoelectrochemical devices. In this technique, a signal proportional to the frequency of a small pertubation over zero potential (open circuit potential) is generated in the system.41 The response at high frequencies can be attributed to kinetic processes occuring at the counterelectrode/electrolyte interface such as resistance of electrolyte, resistance of charge transfer and capacitance of double layer, while the response at low frequencies can be associated with the transport mechanisms (diffusion) in the electrolyte.40 The impedance data are interpreted by assigning an equivalent circuit showing similar current to that produced by the pertubation in the system,42 considering the simplest electrode processes possible.

100

Z˝ (Ω)

Q1 50 10 kHz 0 0

1 Hz 50

100 Z´ (Ω)

0.001 Hz

150

RS R1

ZD

200

10.6 Nyquist diagrams of the impedance spectra obtained for polymer electrolyte films sandwiched between planar Pt electrodes. Experimental data are represented by symbols and solid lines correspond to the fitting using the equivalent circuit shown in the inset. The polymer electrolytes are based on P(EO–EPI), NaI/I2 (䊊) with plasticizer (50 wt% GBL) and (䊐) without plasticizer.



393

Figure 10.6 shows that the overall impedance of the system decreases after addition of plasticizer. The data are in agreement with the increase observed in ionic conductivity.39 From the parameters obtained by fitting the experimental data shown in Fig. 10.6, the apparent diffusion coefficient can be estimated using equation 10.7,43 where le is the thickness of the electrolyte film and B is a parameter related to the element O in the equivalent circuit proposed, which accounts for a finite-length Warburg diffusion (ZD), which represents a kind of resistance to mass transfer. D = le 2 B 2

[10.7]

s (S cm–1)

The apparent diffusion coefficient of ionic species in the electrolyte containing GBL was estimated to be 3 × 10−6 cm2 s−1, one order of magnitude higher than the diffusion coefficient estimated for the polymer electrolyte without plasticizer (4 × 10−7 cm2 s−1). This value is very close to the diffusion coefficient for I3− species in highly viscous solvents, such as N-methyl oxazolidinone (2.8 × 10−6 cm2 s−1),44 in gels (3 × 10−6 cm2 s−1),45 gellified ionic liquids (1.4 × 10−6 cm2 s−1)46 and in a TiO2 membrane soaked in acetonitrile (3.4 × 10−6 cm2 s−1),47 but is still one order of magnitude lower than the diffusion in acetonitrile (~2 × 10−5 cm2 s−1).48 GBL has been used as plasticizer for electrolytes for application in batteries, and it is well known to be able to coordinate Li+ ions, contributing to the dissolution of lithium salts in polymer systems.49,50 Therefore, GBL also made possible the substitution of NaI for LiI in the polymer electrolye, as shown in the plots of ionic conductivity in Fig. 10.7. As opposed to what

10–4

10–5 0

5

10

15

20

25

Salt content (wt%)

10.7 Effect of salt concentration on conductivity for plasticized polymer electrolytes based on P(EO–EPI), 50 % of GBL, I2 and: (-䊏-) LiI and (-䊉-) NaI.


394


was observed for the non-plasticized polymer electrolyte (Fig. 10.3), in the presence of GBL, electrolytes prepared with LiI present higher conductivities in comparison to the ones prepared with NaI. The addition of GBL leads to an increase in conductivity of two orders of magnitude, when compared with the non-plasticized polymer electrolyte. Also, the ionic conductivity dependence on salt concentration changes remarkably when 50 wt% of GBL is incorporated in the polymer electrolyte. As discussed before, GBL has an ion solvating ability, which allows the dissolution of higher amounts of salt. Thus, increasing the salt concentration leads to a further increase in the conductivity, reaching 5 × 10−4 S cm−1 for the sample prepared with 20 wt% of LiI. Such behavior is the opposite of that usually observed for polymer electrolytes at high salt concentration.22,24 This conductivity value approaches the value for a liquid electrolyte based on organic solvents. The conductivity for the sample containing 30 wt% of LiI was estimated as 6 × 10−4 S cm−4, which is very close to that exhibited by the sample containing 20 wt%, indicating saturation in salt dissolution. Solar cells assembled with P(EO–EPI)87–13 containing 50 wt% of GBL, 20 wt% of LiI and I2 presented efficiencies of 3.3 and 3.5% at 100 and 10 mW cm−2 of irradiation, respectively.51 Other authors also reported recently that the conductivities and diffusion coefficients of ionic species in gel electrolytes can be changed and improved by varying the salt concentration52 or by changing the salt composition, i.e. using different cations.53–55 Figure 10.8 shows the conductivity behavior for electrolytes of P(EO– EM)/GBL/LiI/I2 with different concentrations of GBL. Upon increasing the amount of GBL, an overall increase in the ionic conductivity is observed.

s (S cm–1)

10–2 1.9 × 10–3 –3

10

Increase GBL concentration

10–4

3.1 × 10–5

10–5 0

5

10 15 20 Salt content (wt%)

25

10.8 Ionic conductivity as a function of LiI concentration for the polymer electrolyte P(EO–EM)/GBL/LiI/I2, upon addition of (䊉) 30% or (䉱) 70% of GBL. The figure also shows a picture of liquid electrolyte (left) and a gel polymer electrolyte containing 70 wt% GBL (right). As can be seen, almost no fluidity is observed at room temperature, even after addition of 70 wt% of GBL. This is an important feature for application as electrolyte in a solar cell.



395

Intensity (a.u.)

For the electrolyte containing 70 w% of GBL the ionic conductivity was 2 × 10−3 S cm−1. The high values of ionic conductivity measured were rationalized to originate from a contribution of both ionic transport and electronic conductivity, introduced by the formation of polyiodides in the electrolytes with high salt concentration. The conversion of iodine to polyiodide species when the LiI salt concentration in the electrolyte exceeds 10 wt% was confirmed by Raman spectroscopy measurements,51 and a similar effect was also reported by Yanagida and co-workers56 for gel electrolytes. A similar behavior was also observed using Raman spectroscopy (Fig. 10.9) for the

20% Lil wt

15% Lil wt 10% Lil wt

50

100

150

200

250


(a)

Intensity (a.u.)

20% Lil wt

15% Lil wt 10% Lil wt

50 (b)

100

150

200

250


10.9 Raman spectra as a function of LiI concentration for the polymer electrolyte P(EO–EM)/GBL/LiI/I2 (laser excitation wavelength 632.8 nm) with (a) 30 wt% and (b) 70 wt% of GBL.


396


systems based on the copolymer (P(EO–EM) with high salt concentration.57 This polymer was mixed with LiI, I2 and the plasticizer GBL, where the amount of GBL incorporated into the electrolyte was changed from 30% to 70%. All the P(EO–EM)/GBL/LiI/I2 samples present a band around 110 cm−1, which can be assigned to the symmetric stretch of I3− species.58,59 Another band was observed at ~142 cm−1, assigned to the vibration mode of higher polyiodide species, such as I5−. This band was reported to appear around 150 cm−1 for other electrolytes.56 The polyiodides species have a lower limiting molar conductivity than monoiodide species due to their larger ionic radius. Therefore, considering only the diffusion of ionic species as liable for the conductivity of the electrolytes investigated in this system, a decrease in the conductivity for samples containing large concentrations of iodide/iodine would be expected. However, it is well known that electron exchange might occur between polyiodides and, under high salt concentration conditions, the Grothuss-type charge-transfer mechanism might contribute to the effective conductance of the electrolyte.31,60 In Grothuss-type charge-transfer mechanism, the electron hopping and the polyiodide bond exchange are coupled, and both paths should contribute to the effective conductance of the polymer electrolyte.56 These results suggest that the polymer electrolytes based on P(EO/EM)/GBL/LiI/I2 might act as a mixed conductor, presenting both ionic and electronic conductivity with high LiI concentration. This effect can be better observed through the ratio of polyiodides Im− and I3− (Im−/I3− value), calculated from the intensities of the peaks in the Raman spectra. These values were calculated for the polymer electrolytes prepared with different GBL concentrations, as a function of salt concentration, as presented in Fig. 10.10. The Im−/I3− ratio increases with the addition of salt, and GBL, revealing an increased contribution of electronic conductive pathways for electrolytes with higher GLB and LiI concentrations. This could be a consequence of an increase in ionic mobility provided by the conduction in the less viscous, liquid GBL medium. This could lead to a faster diffusion dynamics of the iodide/iodine species, provided by the presence of more GBL molecules, which can facilitate the collision of these species to form polyiodides. Considering the value of the conductivity reached in the present work (~10−3 S cm−1) it is possible to infer that in the P(EO–EM)/GBL/LiI/I2 system, the electronic transport contribution is small, but still significant. The performance of DSSC using polymer electrolytes based on P(EO– EM)/GBL/LiI/I2 was investigated and the current–voltage characteristics (J–V curves) of these devices are shown in Fig. 10.11. Jsc increases when the amount of GBL is increased from 30 to 70%, and this effect was related to the increase of ionic conductivity of the electrolyte. On the other hand, Voc



397

0.70

Im–/I3– ratio

0.65 0.60 0.55 Increasing GBL concentration

0.50 0.45 15

16

17

18

19

20

21

Salt content (wt%)

10.10 Ratio of polyiodides (Im−) and I3− (Im−/ I3−) as a function of salt concentration for the polymer electrolyte prepared with (䊏) 30 wt% or (䉱) 70 wt% of GBL. The iodide : iodine molar ratio was kept at 10 : 1 for all samples.

Photocurrent (mA cm–2)

16 Jsc = 14.5 mA cm–2 Voc = 0.64 V FF = 0.56 h = 4.4%

12

8

4

0.0

Jsc = 8.5 mA cm–2 Voc = 0.76 V FF = 0.64 h = 3.5%

0.2

0.4

0.6

0.8

Voltage (V)

10.11 J–V curves for the DSSC assembled with P(EO–EM)/GBL/LiI/I2 systems with different GBL contents (䊏) 30% or (䉱) 70%, and fixed salt concentration (20 wt%), under irradiation of 100 mW cm−2.

increases in the opposite direction, i.e. with the increase in polymer content. The loss in Voc values as the amount of GBL is increased was also observed for other PEO-based copolymers, as shown in Fig. 10.12. This effect can be attributed to the loss in the basic character of the electrolyte as more GBL is added. It is believed that the polyether units have more donor capability


398

Polymer electrolytes 0.9

Voc (V)

0.8

0.7

0.6

0.5 0

20

40

60

80

100

GBL (wt%)

10.12 Effect of GBL concentration on the open circuit voltage of DSSC assembled with polymer electrolyte containing (PEO) copolymers and MI/I2 (M = Li or Na).

and thus interact more effectively with the Ti(IV) acid sites, inhibiting charge recombination. The electron lifetime in DSSC assembled with electrolytes containing different amounts of GBL was estimated from the voltage decay transients of the solar cells. A decrease in electron lifetime from the electrolyte containing 30% of GBL to the electrolyte containing 70 wt% of GBL was observed, in agreement with the trend observed in Voc values. These results support the role of the polymer passivating layer in minimizing the charge recombination at the TiO2/electrolyte interface. It seems that the recombination losses (photoinjected electrons return to the dye cation and/or electrolyte) are accelerated when replacing a very basic polymer by the addition of such additives with lower basicity. Another important feature of the system poly(ethylene oxide) copolymer/GBL is that the ionic conductivity of the electrolyte is still significantly dependent on the composition of the copolymer.39 It was known from previous works that, for pure polymer electrolytes (without plasticizer), although the crystallinity degree increased with the increase in EO unit content in the copolymer P(EO–EPI), the conductivity also increased due to having more sites available for cation coordination, once the EPI units do not contribute to ionic transport.22 The same trend was observed for electrolytes prepared with NaI and P(EO–EPI) even after addition of 50 wt% of GBL to the electrolyte, indicating that the oxygen atoms from the polymer chains are probably still contributing to ionic transport, even in the presence of plasticizer.39



399

Table 10.2 Performance of DSSC assembled with polymeric electrolytes with different compositions, under irradiation of 100 mW cm−2. Upon improvements in the ionic conductivity of the electrolyte, Jsc and η are increased, although Voc is decreased Electrolyte composition

Jsc (mA cm−2)

Voc (V)

η (%)

P(EO–EPI)/NaI(9 wt%)/I2 P(EO–EPI)/GBL/NaI(9 wt%)/I2 P(EO–EPI)/GBL/LiI(7.5 wt%)/I2 P(EO–EPI)/GBL(50%)/LiI(20 wt%)/I2 P(EO–EM)/GBL(70%)/LiI(20 wt%)/I2 P(EO–EM)/GBL(70%)/LiI(20 wt%)/I2/CE(1 : 1)

4.2 6.1 6.3 9.1 10.2 11.4

0.82 0.78 0.71 0.67 0.66 0.78

1.6 1.7 2.1 3.3 3.4 3.7

Table 10.2 summarizes the improvements in performance of DSSC parameters achieved by the different modifications introduced in the polymer electrolyte composition. Transient absorption spectroscopy was employed to study electron-transfer dynamics in solar cells incorporating the polymer electrolyte based on EO copolymers with and without plasticizer. Electron-transfer kinetics were collected as a function of electrolyte composition, white light illumination, and device voltage. The results were further correlated with the current/ voltage characteristics of the solar cells. There are two main recombination pathways which can cause loss in DSSC efficiency: electrons injected into the TiO2 conduction band can recombine with either dye cations or with the redox electrolyte (equations 10.8 and 10.9, respectively). TiO2( e − ) + dye + → dye

[10.8]

TiO2(e

[10.9]

−

) + 1 2 I2 → I

−

dye + + 2I − → dye + I 2 −

[10.10]

In a liquid electrolyte DSSC, rapid re-reduction of dye cations by the redox electrolyte (equation 10.10) competes effectively with equation 10.8, and therefore charge recombination to the redox electrolyte, equation 10.9, is the primary recombination loss pathway limiting device efficiency.30,61 In polymer electrolyte-based DSSC, however, the low ionic conductivity of the polymer electrolyte introduces the possibility that dye cation rereduction by the electrolyte may no longer compete effectively with the recombination pathway described in equation 10.8. As a consequence, charge recombination with dye cations may become critical in limiting device efficiency. For the electrolyte containing P(EO–EPI)84–16, NaI and I2, without plasticizer, regeneration of the dye ground state by electron transfer from


400


I− ions (equation 10.10) exhibited half-times (τ1/2) of 4–200 µs, depending upon the concentration of NaI in the polymer electrolyte.27 At low NaI concentrations, kinetic competition was observed between equation 10.10 and equation 10.8. The decay kinetics of the dye cation and the yield of I2− were found to be unchanged by illumination of the cell. From these observations, it was concluded that the charge recombination dynamics in this cell are not strongly dependent upon the TiO2 Fermi level over the investigated voltage range and this observation is probably related to the Lewis base nature of the polymer employed, as discussed previously.27 Platicized polymer electrolytes of P(EO–EPI)87–13 with 50 wt% of GBL and different amounts of LiI/I2 were also investigated using this technique.51 For a sample with a moderate concentration of iodide (7.5 wt% LiI), the dye cation signal exhibited a decay with τ1/2 = 20 µs, which was only marginally faster than charge recombination of the oxidized dye with electrons injected into the semiconductor, suggesting that kinetic competition between equation 10.8 and equation 10.10 might also be significant for this electrolyte composition. Increasing the amount of LiI to 20 and 30 wt%, regeneration (equation 10.10) became clearly predominant with τ1/2~1 µs in both cases,51 similar to the value reported for acetonitrile-based electrolytes.62 For 20 and 30 wt% LiI plasticized polymer electrolytes, the transients were less dependent upon white light illumination. The transient data for the plasticized electrolyte containing 7.5 and 20 wt% of LiI are shown in Fig. 10.13. The data obtained for a non-plasticized polymer electrolyte are also shown for

4.0 × 10 –4

ΔOD

(Without GBL)

(7.5% wt% Lil)

2.0 × 10 –4

(20% wt% Lil) 0.0

10 –6

10 –5

10 –4 Time (s)

10 –3

10 –2

10.13 Transient absorption spectra for dye-sensitized TiO2 films covered with polymer electrolytes based on P(EO–EPI), MI/I2 (M = Li or Na) with GBL (black line) and 7.5 wt% or 20 wt% of LiI, or without plasticizer (gray) (OD = optical density).



401

comparison. The results obtained indicate that for polymer electrolytes with high iodide concentration and high ionic conductivity, equation 10.10 competes more effectively with equation 10.8, indicating a more efficient regeneration. Although the high concentration of iodide/iodine in the electrolyte contributed to accelerate dye cation regeneration, it also increased the dark current of the cell by one order of magnitude. In these cells, unlike what was observed previously for a DSSC assembled with a polymer electrolyte (polymer and salt only), it is not the competition between equation 10.8 and equation 10.10 that limits the efficiency; instead, equation 10.9 or dark current, together with the low Voc caused by the excess Li+ cations on the TiO2 surface, seems to play the main role in determination of efficiency.51 Further support for the kinetic competition between equation 10.8 and equation 10.10, and the equation 10.9 contribution can be obtained by fitting the J–V curves using a two-diode model (equation 10.11),27,63 where IL is the light intensity dependent short-circuit current, kB is the Boltzmann’s constant, T is the temperature and I0, k, m1 and m2 are fitting constants related to the dark current and the ideality factor. I = I L − I 0[exp (qVj m1kBT ) − 1] − kI L[exp (qVj m2 kBT ) − 1] [10.11] The bias drop across the internal junction, Vj can be related to the externally applied bias, V, through equation 10.12, where Rs is the series resistance of the system. I0, m and Rs are assumed to be light intensity independent: Vj = V + IRs

[10.12]

Deviations from this model can be interpreted in terms of a voltage dependent loss of charge separation yield due to either lower electron injection yields or kinetic competition between charge recombination (equation 10.8 and 10.9) and equation 10.10.27,62 The first two terms on the right of equation 10.11 compose the usual non-ideal one diode current–voltage characteristic of a solar cell. The final term in equation 10.11 is a light-dependent recombination current, and is required to describe adequately the observed behavior for the cells assembled with the polymer electrolyte with and without plasticizer.20,27 For the non-plasticized polymer electrolyte, the light-dependent recombination term was introduced because of the high rate observed for equation 10.8, owing to the low ionic conductivity of the electrolyte.27 For the electrolytes plasticized with GBL, however, it would be expected that the J–V curves could be fitted using the simple one diode model, at least for samples containing higher amounts of salt, due to the high ionic conductivity of these electrolytes, and the data from transient absorption spectroscopy presented earlier. In Fig. 10.14 typical J–V curves for DSSC containing polymer electrolytes with and without plasticizer are presented. Fits using the two-diode model


402



10 8 6 4 2

Jsc = 9.1 mA cm–2 Voc = 0.67 V FF = 0.54 h = 3.3% I0 = 100 nA Jsc = 4.2 mA cm–2 Voc = 0.82 V FF = 0.47 h = 1.6% I0 = 1 nA

0 0.0

0.2

0.4

0.6

0.8

Voltage (V)

10.14 J–V characteristics for the DSSC assembled with the polymer electrolyte containing P(EO–EPI), MI/I2 (M = Li or Na), (䊊) with and (䊐) without GBL as plasticizer, under 100 mW cm−2. The fitting using the two diode model (Equation 10.11) is also shown as solid lines.

are also shown. For the plasticizer electrolyte, the need for the second term could be a consequence of the increase in dark current values, owing to higher concentrations of iodine/iodide employed, unlike the previous case, for the system based on the polymer electrolyte without plasticizer. In other words, the performance at high concentrations of iodide/iodine is no longer expected to be dominated by the conductivity of the medium or by the low rate of dye cation regeneration. These results are in agreement with a report by Kang and co-workers,64 which shows that, up to a certain limit, further improvements in the ionic conductivity of the electrolyte do not guarantee any improvements in DSSC performance. Here, the efficiency of this polymer electrolyte DSSC might be dominated by the recombination reactions at the interfaces (mainly equation 10.9 or dark current). Thus, it is expected that the efficiency of these devices can be enhanced even more by improving Voc of the cells containing large amounts of plasticizer and salt. This can be achieved by the incorporation of different additives in the electrolyte, as will be discussed in the following section.

10.4

Additives in the polymer electrolytes

Besides the plasticizers, other additives can be explored to improve the characteristics of polymer or gel electrolytes. These materials can be incorporated for different purposes, such as the improvement of mechanical and thermal stability, enhancement of the charge transport, open circuit voltage, etc.



403

10.4.1 Inorganic nanofillers An interesting approach consists of the addition of nanoscale inorganic fillers, to improve the mechanical, interfacial and conductivity properties of the (gel) polymer electrolytes. Since the pioneering work by Scrosati and co-workers,65 addition of TiO2 and other nanoparticles has been extensively employed to improve the ionic conductivity of polymer electrolytes. It is well known that the presence of such nanoparticles changes the conduction mechanisms assigned to the ions introduced in the polymer; however, how these nanoparticles actually act is still unknown. These materials can also improve the mechanical properties of gel electrolytes and ionic liquid-based electrolytes. However, their effect on the mechanical stability can result in a loss in electrolyte penetration. The most used approach is the addition of TiO2 nanoparticles to the polymer matrix66–73. Falaras and co-workers66,67 investigated the addition of commercially available TiO2 nanoparticles (P25, Degussa) to the polymer electrolytes of PEO, LiI and I2. The filler particles, because of their large surface area, prevented recrystallization, decreasing the crystallinity degree of PEO. The increase in the Tg of the polymer indicated that the polymer electrolyte incorporated a significant quantity of the available inorganic oxide filler. Besides, the large quantity of the filler increased the dissolution of the LiI salt and the system remained mainly amorphous, as confirmed by atomic force microscopy (AFM) measurements. The polymer sub-units are held together in a parallel orientation, forming long straight chains of about 500 nm in width, along with TiO2 spherical particles of about 20–25 nm in diameter. The polymer chains separated by the titania particles are arranged in a three-dimensional, mechanically stable network that creates free space and voids into which the iodide/triodide anions can easily migrate.66 Dye-sensitized solar cells prepared with this nanocomposite polymer electrolyte exhibited Jsc = 7.2 mA cm−2, Voc = 0.66 V, FF = 0.58 and η = 4.2% (65.6 mW cm−2).67 Other reports also describe the effects of TiO2 nanoparticles on gel polymer electrolytes.71 Kang and co-workers71 showed that these nanoparticles additionally lead to a light-scattering effect. The DSSC with the ternary component polymer–gel electrolyte exhibited an energy conversion efficiency of 7.2% (100 mW cm−2). Yang and co-workers74 investigated the use of TiO2 nanotubes as nanofillers. Electrolytes containing polyethylene glycol and 10% of nanotubes showed high penetration and complete filling of the pores of the TiO2 film. Using the X-ray photoelectron spectroscopy(XPS) technique the authors showed that there is an interaction between the titanium atoms of the nanotubes and the polymer network. The ionic conductivity was found to be 2.4 × 10−3 S cm−1, which was achieved through the decrease in the crystallinity degree of the polymer after introduction of the nanotubes. DSSC


404


fabricated with this composite electrolyte showed the maximum overall conversion efficiency of 4.4%, Jsc = 9.4 mA cm−2, Voc = 0.73 V and FF = 0.65 under 100 mW cm−2 of irradiation.74 Nanoparticles with different compositions, such as ZnO, SiO2 and Al2O3, have also been used as nanofillers in polymer electrolytes.75–78 Caruso and co-workers75 investigated solar cells assembled with a composite polymer electrolyte based on PEO, poly(vinylidene fluoride) (PVDF) and SiO2 nanoparticles. To fabricate the solid state DSSC, the composite polymer electrolyte solution was injected into the dye-sensitized TiO2 electrode using a vacuum technique. For a given TiO2 film thickness, devices prepared employing the vacuum method exhibited a better performance than those prepared via the conventional drop-casting method. Besides, the differences became more pronounced with increasing TiO2 film thickness. These results are remarkable since they show that not only is the optimization of the composition of the electrolyte an important issue, but it is also important to guarantee the full filling of the photoelectrode with the solid state or gel electrolyte. Zhao and co-workers76 prepared an electrolyte based on PEO, poly(vinylidenefluoride-co-hexafluoropropylene) (PVDF–HFP), SiO2 and conductive carbon nanoparticles. The conductivity mechanism was analyzed by AC impedance and DC voltage–current measurements. A change in the conduction mechanism was obtained by adding different amounts of carbon nanoparticles. Small amounts of carbon nanoparticles improved the ionic conductivity and a DSSC with 5 wt% of carbon nanoparticles in the electrolyte presented η = 4.3% compared with the original DSSC performance of 3.9%. When the content of nanoparticles was increased to 15 wt%, the efficiency decreased to 3.6%, as a consequence of a decrease in the ionic conductivity and an increase in interface recombination with the electrolyte, because of the electronic conductive path formed by the aggregated carbon nanoparticles. Xia and co-workers77 used an interesting approach to make a composite polymeric electrolyte. First, poly(ethylene glycol methyl ether) with molar mass of 350 g mol−1 was grafted onto the surface of ZnO nanoparticles through covalent bond formation. The electrolyte was composed of KI and I2 dissolved in a low molar mass poly(ethylene glycol methyl ether), and 24 wt% of the polymer-grafted ZnO nanoparticles were used to solidify the electrolyte. For this system, the ionic conductivity increased as the salt concentration increased, reaching a maximum value of 3.3 × 10−4 S cm−1, and then decreased, behaving like a classical polymer electrolyte system. The authors showed that the Voc increased by 0.13 V after the polymergrafted nanoparticles were added to the electrolyte, but the Jsc decreased, probably because of the high viscosity of the gel formed. As a result, the efficiency of the solar cell decreased to 3.1% after addition of the polymer-



405

grafted nanoparticles, when compared with the efficiency of 4.0% observed for the device with a liquid electrolyte. Al2O3 particles with different sizes were incorporated into electrolytes based on mixtures of ionic liquids, a PVDF derivative and polyacrylonitrile.78 It was observed that the added nanoparticles influenced the diffusion coefficient of I3− ions and also the charge transfer rate, and this effect depended on the Al2O3 particle size. The authors suggested that the imidazolium cations might adsorb on the nanoparticle surface, and then the counter-anions I−/I3− gather around them. A few papers have recently reported the addition of clay-like nanoparticles in polymer electrolytes.79,80 Nogueira and co-workers79 investigated the incorporation of a montmorillonite (MMT) derivative into a polymer electrolyte based on a PEO copolymer, the plasticizer GBL and LiI/I2. The initial increase of ionic conductivity after addition of MMT was attributed to the large number of charge carriers introduced into the complex as the clay concentration increased (see inset in Fig. 10.15). The X-ray diffraction (XRD) data suggested that the clay was not exfoliated in the nanocomposite electrolyte, but rather kept its lamellar structure. Figure 10.15 presents the thermomechanical characterization of the polymer electrolytes prepared with and without the addition of 5 wt% clay. According to this data, the presence of the MMT promotes an increase in the mechanical stability of the entire system. This can be viewed considering the force applied to the nanocomposite polymer electrolyte film, which promotes a lower deformation comparing to the film without any clay. The difference in the mechanical properties is also illustrated as an inset in Fig. 10.15, where a picture of spherical shaped samples of both the composite and the plasticized polymer electrolyte prepared without clay are shown. In the case of the system without clay, there is a flow with time owing to the action of gravity, leading to the disruption of the spherical shape. For the electrolyte with clay, the spherical shape remains unchanged. These results showed that the addition of MMT clay to the plasticized polymer electrolyte led not only to an increase in the ionic conductivity, but also to the solidification of the electrolyte, reflected as an improvement in the mechanical stability of the films. The solar cell devices containing the nanocomposite polymer electrolyte presented efficiencies of 1.6% and 3.2% at 100 and 10 mW cm−2, respectively. The FF values were very poor, only 40% under 100 mW cm−2, which was attributed to the low penetration of the composite electrolyte inside the pores of the TiO2 film.79 Lin and co-workers80 prepared a nanocomposite of poly(n-isopropylacrylamide) with MMT clay and applied it to a liquid electrolyte system as gellator. The DSSC assembled with the polymer nanocomposite electrolyte presented Jsc = 12.6 mA cm−2, Voc = 0.73 V, FF = 0.59 and η = 5.4% while the DSSC prepared with the electrolyte gelled with the pure polymer


406


Deformation (a.u.)

(a)

0.0

0.2

0.4 0.6 Force (N)

0.8

1.0

(b) 10–3

s (S cm–1)

CPE

PE

10–4 0

2 4 6 MMT clay (wt%)

8

10.15 Thermomechanical analysis (a) (compression mode) for electrolytes based on P(EO–EPI), GBL and LiI/I2: (dotted line) without clay and (black line) with 5 wt% of MMT clay; and ionic conductivity (b) as a function of clay content. The inset shows a photograph of spherically shaped samples of the polymer electrolyte prepared with (CPE) and without the clay (PE), after some period of time at ambient conditions.

presented Jsc = 7.28 mA cm−2, Voc = 0.72 V, FF = 0.60, and η = 3.2% (100 mW cm−2). Electrochemical impedance spectroscopy of the DSSC revealed that the nanocomposite-gelled electrolyte presented a significant decrease in impedance values. The resistance due to the electrolyte and electric contacts, the impedance across the electrolytes/dye-coated TiO2 interface, and the Nernstian diffusion within the electrolytes were reduced. An increase in the molar conductivity of the nanocomposite-gelled electrolytes was also reported.80

10.4.2 Organic molecules Pyridine derivatives, such as 4-tert-butylpyridine (TBP), are frequently added to the electrolytes as additives to enhance the open-circuit photovolt© Woodhead Publishing Limited, 2010


407

age, and thus the efficiency of DSSC. The Voc and FF of the DSSC are affected by TBP owing to the suppression of dark current at the semiconductor/electrolyte junction, since TBP blocks the surface states that are active in the charge transfer.81 Wu and co-workers82 investigated DSSC with a polymer electrolyte based on a blend of poly(acrylamide) and poly(ethylene glycol). Using poly(ethylene glycol) as both reactant and plasticizer, TBP as additive and the mixture EC/PC or GBL as solvent, a gel polymer electrolyte with quasi-solid state was prepared. The ionic conductivity of the electrolyte was enhanced because of the complexation of the carbonyl, amine and hydroxyl groups on poly(acrylamide) and poly(ethylene glycol) chains to the K+ ions present in the electrolyte. The maximum ionic conductivity (at 30 °C) was ~2.0 × 10−3 S cm−1. DSSC were fabricated using the gel polymer electrolyte with different amounts of TBP. On the other hand, Jsc decreased with the increase in the TBP concentration. Considering the influence of the amount of TBP on both Jsc and Voc, the optimized quantity of TBP was determined to be ~2 wt%. Under irradiation of 60 mW cm−2, the optimized DSSC presented Voc = 0.69 V and Jsc = 4.6 mA cm−2 while the overall energy conversion efficiency was ~3%. Xia and co-workers77 added TBP to the electrolyte composed of poly(ethylene glycol methyl ether), KI, I2 and 24 wt% of polymer-grafted ZnO nanoparticles. The addition of TBP into the electrolyte resulted in a dramatic improvement in Jsc and overall efficiency, from 3.1% to 5.0%. However, the Voc and FF remained insensitive to the presence of this additive. The same phenomenon was also observed in a solar cell based on a room temperature molten salt reported by Kloo and co-workers.83 Although the detailed mechanism was not given, the improvement was attributed to a lower viscosity and better interfacial contact of the electrolyte after addition of TBP. The authors suggested that the effect of TBP on the electrode surface is less significant than the effect on the viscosity, for solar cells that utilize highly viscous electrolytes.77 These reports suggest that the TBP effect on DSSC with polymer electrolytes is still under debate and that the results depend strongly on type and composition of the electrolyte. Recently it was shown that the incorporation of the additive 2,6-bis(Npyrazolyl) pyridine to the electrolyte of PEO, KI and I2 improved both the interfacial contact and the ionic conductivity in DSSC. X-ray diffraction and impedance spectroscopy results have shown that this additive forms a stable complex with the polymer matrix and decreases the crystallinity of PEO. Owing to the coordinating and plasticizing effect of this additive, the ionic conductivity of the polymer electrolyte was enhanced. The DSSC assembled presented Jsc = 21 mA cm−2, Voc = 0.70 V and η = 8.8% under direct solar irradiation of 80 mW cm−2.84 A few reports are found describing the use of pyridine polymers. Polyvinylpyridine (PVP) and its derivatives are the only pyridine polymers which have been used so far to treat the dye-sensitized TiO2 electrode surface.85–89 © Woodhead Publishing Limited, 2010

408


For PVP, a 150 mV enhancement in Voc was observed.85 Lin and coworkers88 investigated the addition of poly(vinylpyridine-co-acrylonitrile) to an electrolyte containing EC/PC, KI and I2. The copolymer was used instead of pure PVP because of the incompatibility of PVP with electrolytes containing alkali iodides. The authors observed that Voc increased ~100 mV, and the efficiency for the quasi-solid DSSC achieved 6.7%, which is higher than the 6.0% obtained for the initial EC/PC/KI/I2 liquid electrolyte. Different additives containing functional end-groups can be used to chemically modify the electrolyte and its interface with the photoelectrode and counter-electrode. Zhao and co-workers90 employed a NH2-terminated functional silane (3-aminopropyltriethoxysilane) to functionalize the electrolyte composed of PEO and PVDF–HFP. The newly formed Si–O–Si network and interactions influenced the ionic conductivity of the modified polymer electrolyte and also enhanced the connection of the polymer electrolyte with the electrodes. The additive deprotonated the TiO2 photoelectrode surface, causing a change in the Fermi level energy that effectively reduced the interface recombination in the DSSC and improved the open circuit voltage. With a moderate concentration of the additive, they achieved an efficiency of 5.1% compared with 3.7% of the original DSSC. The endgroup-functionalized silicone coupling agent dodecyl-trimethoxysilane was also used to modify the PEO/PVDF–HFP/SiO2 nanocomposite polymer electrolytes. The introduction of optimized contents of this additive improved the ionic conductivity and the connections with the photoanode and counter electrode. Optimal efficiencies of 6.4% and 4.9% under 30 and 100 mW cm−2, respectively, were obtained.91 Alternatively, oligomers or low molecular weight polymers containing functional terminal groups can be used in electrolytes to achieve high penetration into the TiO2 film. After deposition, chemical reactions between the functional terminal groups of the oligomer can be used to promote the solidification of the electrolyte, originating solid DSSC with improved efficiency.92,93 Kang and co-workers94 used amorphous oligomers low molecular weight poly(propylene glycol) as additives to the system PEO/KI/I2. They observed that the oligomer not only increased the ionic conductivity of the electrolyte, but also improved the penetration of the electrolytes into the nanopores of the TiO2 film. The DSSC presented better FF (from 0.26 to 0.48) after addition of the oligomer. In a similar approach, Park and co-workers95 added a small amount of liquid electrolyte to a polymer electrolyte based on poly(ethylene-co-methyl acrylate), obtaining improvements in both ionic conductivity and penetration of the electrolyte into the TiO2 nanopores. Recently, an interesting work showed the incorporation of an electrochromic molecule (benzidine) in a polymer electrolyte based on



409

poly(epichlorohydrine). The authors suggested that, since benzidine is a powerful electron donor, it may hasten the rate of re-reduction of the oxidized dye by the electrolyte and also reduce the back reactions (recombination reactions) which cause loss of efficiency in DSSC.96 In another work, a heteropolyacid (H3PW12O40) was mixed with PEO. This acid may act as electron acceptor, since it can react with electrons injected into the TiO2 conduction band, thus preventing the reduction of I− ions through this pathway (recombination reaction). Also, it may act in the same way as inorganic nanofillers, decreasing the crystallinity degree of the polymer.97 Another possibility to improve DSSC performance is the addition of crown-ethers (CE) to the electrolyte. Polymer electrolytes containing CE are usually employed in secondary lithium batteries, but some papers have shown the use of this material in electrolytes for DSSC. For example, Dai and co-workers98 added 18-crown-6 ether to an electrolyte containing the ionic liquid 1,2-dimethyl-3-propylimidazolium iodide. The devices containing CE exhibited a small enhancement in the short-circuit photocurrent. On the other hand, the addition of CE to polymer electrolytes has been shown to decrease the overall ionic conductivity.99 Nogueira and coworkers100 investigated the influence of the addition of 12-crown-4 ether to a polymer electrolyte consisting of P(EO–EM), GBL, LiI and I2, and its application in dye-sensitized solar cells. The copolymer/plasticizer weight ratio was 3 : 7. The 12-crown-4 ether molecules exhibit high size selectivity for Li+ ions, resulting in a very strongly coordinating system, as reported previously.101 Figure 10.16(a) exhibits the Nyquist plot for the gel polymer electrolytes with different 12-crown-4 ether contents. For these systems, the plots did not exhibit a semicircle related to the electrolyte capacitance in the high frequency region, as normally observed for polymer electrolyte samples. A straight line was observed instead. The disappearance of this semicircle is due to the high ionic conductivity of the samples. The intersection of this straight line with the real axis provides the resistance value usually employed to calculate the ionic conductivity values, presented in Fig. 10.16(b). The addition of CE and LiI, in a 1 : 1 proportion, increases the resistance of the electrolyte, decreasing the ionic conductivity from 3.0 to 1.0 × 10−3 S cm−1. Two factors might be contributing to this result. Morita and co-workers99 assigned the decrease in the conductivity to an increase in the viscosity of the entire polymeric system after a large amount of crown ether addition. However, the trapping of the Li+ ions by the CE can also contribute to the observed decrease, since Li+ species are much more mobile than the large polyiodide ions, and the conductivity values measured with EIS correspond to the sum of the conductivities of all ionic species in the electrolyte. From Fig 10.16, it is clear that addition of crown ethers compromises


410

Polymer electrolytes (a) 4000

1:1 CE/Li+ 1:0.5 CE/Li+ without CE

10–2 Hz

3000

Z˝ (Ω)

10–2 Hz

High frequency

2000 90

60 103 Hz

1000 30 105 Hz 0

30

60

90

0 0

1000

2000

3000

4000

Z´ (Ω) (b)

O

O Li+

s (S cm–1)

O

O

12-Crown-4

10–3

0.0

0.2

0.4

0.6

0.8

1.0

CE content (wt%)

10.16 Analysis of electrolytes based on P(EO–EM) + 20 wt% LiI, 2 wt% I2 and 70 wt% GBL with different CE contents: (a) Nyquist plots of the electrochemical impedance spectroscopy and (b) ionic conductivity values. The crown ether structure is shown in the inset.

the ionic conductivity of the electrolyte; therefore a lower device performance would be expected. On the other hand, the addition of CE to the electrolyte increased the steady state current associated with the diffusion of the iodide/triiodide species. The steady state voltammograms for the redox reaction of triiodide/



411

12

Current (μA)

9

6

3I– – 2e

With crown-ether

I3–

3

0

I3– + 2e

3I– Without

–0.6

–0.4

–0.2 0.0 0.2 Potential (V) vs I3– / I–

0.4

0.6

10.17 J–V steady state voltammograms for a gel polymer electrolyte with and without the addition of 12-crown-4 ether, using a Pt microdisk electrode with 30 µm diameter (scan rate = 10 mV s−1).

iodide in the polymer electrolyte were obtained using Pt disk microelectrodes, and are depicted representatively in Fig. 10.17. The apparent diffusion coefficients of the iodide species increased when compared to the system without the crown ether. This suggests that the incorporation of 12-crown-4 ether to P(EO–EM) copolymer makes Li+ trapping possible and contributes to increase the transport number associated to the iodide species, despite a decrease in the overall conductivity of the system. DSSC were assembled with the polymer electrolyte containing CE : Li+ in the proportion 1 : 1, and without CE. The J–V curves are shown in Fig. 10.18. The device containing CE presented Jsc = 11.4 mA cm−2, Voc = 0.78 V and η = 3.7% under 100 mW cm−2. The same device presented Jsc = 10.2 mA cm−2, Voc = 0.66 V and η = 3.4% when prepared without crown ether. As expected, trapping Li+ ions had a positive effect, increasing Voc. Even more remarkable was the increase in the photocurrent values after CE addition, since the overall conductivity of the electrolyte decreased, which is in agreement with an increase of the transport number of iodide species, making the reduction of the dye more efficient in the cells with CE.

10.4.3 Ionic liquid Room temperature ionic liquids (IL) have the attractive properties of chemical and thermal stability, non-volatility and high ionic conductivity at room temperature. These materials were extensively studied for electrochemical device applications102 and efficiencies over 7% for DSSC


412



12 Jsc = 11.4 mA cm–2 Voc = 0.78 V FF = 0.42 h = 3.7%

10 8 6 4

Jsc = 10.2 mA cm–2 Voc = 0.66 V FF = 0.50 h = 3.4%

2 0 0.0

0.2

0.4 Voltage (V)

0.6

0.8

10.18 J–V curves for DSSC assembled with gel polymer electrolyte with (䉱) 1 : 1 CE/Li+ content and (䊊) without CE, under irradiation of 100 mW cm−2.

assembled with pure ionic liquid electrolytes were recently reported.103,104 Furthermore, there is much interest in preparing gel electrolytes by solidifying ionic liquids with polymers or nanoparticles.46 A complete review about the general application of ionic liquids in dye-sensitized solar cells has been published recently by Gorlov and Kloo.105 In this section, we present a brief discussion of the combination of ionic liquids and polymers to form electrolytes and their application to DSSC, focusing in the role of the polymer in the system. Considering DSSC with polymers and ionic liquids, at least two different approaches have been reported. First, the use of a combination of polymers and other molecules that can react and act as gelling agents, or the addition of a pure polymer or copolymer molecule for the ionic liquid-based electrolyte, aiming at the improvement of the mechanical characteristics of the electrolyte. The second approach is the polymerization of ionic liquids or the insertion of ionic liquid molecules in the polymer chain, in order to make an ionic conducting polymer. Figure 10.19 shows some ionic liquids usually employed in electrolytes. In the first approach, PVP was used as gelling agent in electrolytes with ionic liquid.106–109 PVP was combined with HOOC(CH2)nCOOH (n = 4, 7, 10, 14), and the gelation was given by the reaction between the polymer and the dicarboxylic acid.106 Before gelation, the electrolyte based on the ionic liquid provided an efficiency of 5.2%. The gel electrolytes lead to solar cells with efficiencies of 3.4% to 4.4%, for carboxylic acid with different chain sizes, and the decrease in performance was associated with a decrease in the diffusion properties of the redox-active species. This decrease in



N

+

I–

N

+

N

N

N

413

+

I–

N

+

A–= F3

A–

O

O

S

–

F3

O

n C O

N I–

N +

N

I–

O

O N

O

S

N

–

O

+

N

I–

S

– F3

N

+ N I–

+

N Cl–

C

N

C

N

N

O

O

N

N

O

n

PEO (X) X=150, 350, 550, 750

10.19 Chemical structure of some ionic liquids usually employed in electrolytes for DSSC.

efficiency was not observed for electrolytes using silica nanoparticles as gelator. PVP was also combined with 1,2,4,5-tetra(bromomethyl)benzene to act as gelling agent.107,108 The solidification was carried out by heating the cells after the gel electrolyte precursors were inserted into the nanoporous TiO2 electrode. In this case, the authors observed an increase in Jsc when the amount of gelator was increased, which was explained by a decrease of interfacial resistance between the gel and the electrodes.107 Kang and co-workers92 reported the use of a supramolecular electrolyte consisting of 1-methyl-3-propylimidazolium iodide (MPII), I2 and poly(ethylene glycol) (PEG, Mw = 1000 g mol−1) with terminal functional groups [2-(6-isocyanatohexylaminocarbonylamino)-6-methyl-4[1H]pyrimidinone], which were used to solidify the electrolyte through the formation of hydrogen bonds. The electrolyte originated an absolute solid state film, with ionic conductivity of 5.28 × 10−5 S cm−1 for a 10 : 1 [O] : [MPII] ratio. The DSSC showed an overall energy conversion efficiency of 3.34% at 100 mW cm−2 and 4.59% at 10 mW cm−2. Grätzel and co-workers110 reported the preparation of an electrolyte composed of 10 wt % of PVDF–HFP and the ionic liquid MPII. The PVDF– HFP copolymer is used due to its high dielectric constant, which assists in the dissociation of salts, good electrochemical stability and non-flammability. The HFP units are introduced to reduce the crystallinity degree and increase the solubility of this material in organic solvents.111 The diffusion coefficients of I3− and I− in the gel electrolyte were calculated using steady state voltammetry to be 1.9 × 10−7 cm2 s−1 and 3.1 × 10−7 cm2 s−1, respectively.


414


This electrolyte was used to assemble a DSSC, in combination with the Z-907 dye, showing Jsc = 11.3 mA cm−2, Voc = 0.66 V and FF = 0.71, respectively, yielding an overall energy conversion efficiency of 5.3%.110 Comparing this electrolyte with the one prepared without polymer, the authors observed that the presence of polymer had no adverse effect on the conversion efficiency. Since the main component of the gel is the ionic liquid, liquid channels are formed in the polymeric phase and the diffusion of iodide and triiodide occurs in these channels.110 A similar phenomenon was also observed in other polymer gel systems.112,113 Similar electrolytes based on 1-propyl-2,3-dimethylimidazolium iodide containing 5 wt% of PVDF–HFP (5 wt%) were also investigated in combination with the Z-907 dye. The overall conversion efficiency of 6.1% at 100 mW cm−2 illumination was reported.114 PEO has also been used as polymer matrix for the preparation of electrolytes containing KI, I2 and the ionic liquid 1-ethyl 3-methyl imidazolium thiocyanate.115–118 The final room-temperature conductivity of PEO/ionic liquid/KI/I2 solid film was 2.3 × 10−5 S cm−1 at 50% relative humidity. The DSSC prepared with the electrolyte containing 80% of ionic liquid showed Jsc = 1.9 mA cm−1, Voc = 0.65 V, FF = 0.52 and η = 0.6%.115 In a recent work, Nogueira and co-workers119 investigated polymer electrolytes based on the mixture of PMII and poly(ethylene oxide-copropylene oxide) (P(EO–PO)). The thermal properties of these electrolytes were investigated by thermogravimetry (TGA) and differential scaning calorimetry (DSC), as presented in Fig. 10.20. The DSC data shows that the pure copolymer has Tg = −66 °C and a crystallization peak followed by a broad melting transition (between −30 and 50 °C), corresponding to a crystallinity degree of 6%. For the electrolyte containing 20 wt% of MPII, the crystallization and melting transitions can still be seen, but the crystallinity degree is estimated to be only 2%. For the electrolytes containing 40 and 70 wt% of MPII, no crystallization and melting transition are observed. For all the binary mixtures (MPII + P(EO–PO)) the Tg is observed at −56 ± 1 °C. This result suggests that there is a loss in the polymer chain mobility after addition of the ionic liquid, possibly caused by an ion–dipole interaction between the oxygen heteroatoms from P(EO–PO) and the imidazolium cations from MPII. The presence of this interaction was confirmed with nuclear magnetic resonance and FTIR spectroscopy. This interaction is expected to help dissociate the ionic species in the polymer matrix, i.e. the iodide ions from the ionic liquid become more mobile, leading to an increase in the ionic conductivity and diffusion coefficient, as was observed using EIS measurements.119 The TGA data for the electrolyte containing 70 wt% of MPII showed similar characteristics with those of the pure MPII sample. Thus, for the combined samples, the thermal stability can be limited by the stability



415

(a) MPII 70% MPII Endo

40% MPII 20% MPII P(EO–PO)

–100

–50

0 Temperature (°C)

50

100

(b) 100

Weight loss (%)

80 60 40 20 0 100

200 300 Temperature (°C)

400

500

10.20 Thermal analysis (a) DSC of P(EO–PO), MPII and mixtures containing different ratios of these two materials, and (b) TGA curves of (䊐) P(EO–PO), (䉭) MPII and (䊉) mixture containing 70 wt% of MPII. Both analyses were performed under nitrogen flow of 100 mL min−1 and heating rate of 10 °C min−1.

properties of the ionic liquid, which is less stable than the polymer. However, the thermal stability up to 200 °C is still satisfactory enough to allow the application of these electrolytes in DSSC. Figure 10.21 shows J–V curves obtained for DSSC assembled with the gel electrolytes containing P(EO–PO), 40 wt% of MPII and I2, with and without the addition of 20% of LiI. The DSSC presents a significant increase in the photocurrent after incorporation of LiI, which was attributed to the increase in the number of free ionic species in the electrolyte. On the


416

Polymer electrolytes 14 With Lil Jsc = 12.6 mA cm–2 Voc = 0.59 V FF = 0.50 h = 3.7%


12 10 8 6 4

Without Lil Jsc = 10.5 mA cm–2 Voc = 0.76 V FF = 0.53 h = 4.2%

2 0 0.0

0.2

0.4

0.6

0.8

Voltage (V)

10.21 J–V curves of DSSC assembled with polymer electrolytes of P(EO-PO) containing 40 wt% of MPII, I2, with and without addition of LiI, under irradiation of 100 mW cm−2.

other hand, Voc is reduced after the addition of salt, which is related to the intercalation of Li+ ions into the TiO2 structure, shifting the Fermi level of this semiconductor. The loss in Voc is more pronounced than the increase in Jsc, thus the DSSC prepared with LiI presented lower efficiency than those prepared only with polymer and ionic liquid. For a DSSC containing 70 wt% of MPII, the effect of addition of LiI is negligible, and no changes in Jsc or Voc are observed. In this case, the high concentration of ionic species from MPII is expect to saturate the system, therefore the LiI added may not be dissociated effectively into free ions, making the addition of more salt to this electrolyte unnecessary. These results show that ionic liquid and polymers can be successfully combined and used as solid state or gel electrolytes for DSSC, without further addition of salt or other additives. The second approach involving the use of ionic liquids in polymer electrolyte-based DSSC consists of the use of ionic polymers.120–126 The ionic liquid-type polymer may have the desirable properties (similar to that of pure ionic liquids) and be employed in the fabrication of DSSC with high efficiency, without the problems of evaporation or leakage of the electrolyte during long-term operation. A polymer material based on PEO and imidazolium chloride, 1-oligo(ethylene oxide)-3-methylimidazolium chloride was used as a matrix for electrolytes for DSSC.122 This ionic liquid polymer exhibited ionic conductivity of the order of 10−4 S cm−1 at room temperature and also offered high solubility for inorganic salts, owing to the large polarity at the ether oxygen of PEO and good miscibility with the organic salts. Ionic liquid © Woodhead Publishing Limited, 2010


417

oligomers with different molecular weights of PEO oligomer were studied. The ionic conductivity improved as the PEO molecular weight increased, owing to an increase of the degree of freedom for longer PEO chain segments, leading to better ionic dissociation. The calculated apparent diffusion coefficients of triiodide for different ionic liquid oligomer electrolytes were the same, despite the fact that the viscosity of the electrolyte becomes higher when increasing the PEO molecular weight of the ionic liquid oligomers. A maximum performance of 3.1% was reached at 100 mW cm−2 irradiation using the electrolyte with ionic polymer, higher than the 1.9% obtained for the pure polymer electrolyte based on PEO. The larger Voc observed was attributed to the suppression of the dark current in the cell with the ionic polymer electrolyte, owing to the effect of adsorption of imidazolium cations at the TiO2 electrode, blocking the surface states, which act as the mediators of the charge transfer from the conduction band electrons of TiO2 to triiodide.122 In another work, poly(1-oligo(ethylene glycol) methacrylate-3methylimidazolium chloride) was synthesized as the ionic liquid polymer, exhibiting ionic conductivity of 1.0 × 10−4 S cm−1 at room temperature.123 Ionic liquid-based electrolytes were prepared by the addition of the ionic polymer poly(1-oligo(ethylene glycol) methacrylate-3-methylimidazolium chloride) to a binary ionic liquid mixture and LiI/I2. It was shown that the ionic conductivity decreases with increasing polymer content, and the maximum conductivity achieved (2.0 × 10−3 S cm−1) was obtained at a polymer content of 10 wt%. DSSC assembled with the optimized electrolyte presented overall efficiency conversion of 6.1%.124 A new imidazole-based ionic polymer was synthesized by co-polymerization of alkyl-bis(imidazole)s and diiodoalkyls. The resultant polymer chains consisted of alkyl-imidazolium salts. The polymerization can be carried out in situ, and a DSSC prepared with this material presented efficiency of 1.3% under 100 mW cm−2 of irradiation.125 Different ionic polymers containing the imidazolium cation in the main polymer chain,126 or as ramifications,127 have also been synthesized. Other ionic polymers based on different materials which can also be prepared by in situ polymerization have been applied in DSSC, presenting interesting results.128,129

10.5

Stability of polymer electrolyte-based dyesensitized solar cells

The stability of DSSC with polymer electrolytes is very important, although few studies involving this issue can be found. One of the motivations for the substitution of the liquid electrolyte is that the use of a solid electrolyte can minimize leakage and solvent evaporation problems and provide longterm stability, thus extending the life of the devices. At the same time, many © Woodhead Publishing Limited, 2010

418


questions concerning the stability of such organic materials when used in the conditions of device operation have already been raised. De Paoli and co-workers130 investigated solar cells assembled with the polymer electrolyte of P(EO–EPI)84–16, NaI and I2, using flexible and rigid glass substrates. These cells were irradiated under a Xe lamp with UV and IR filters for long periods, alternating with dark periods. An initial decay in the performance was observed in the first 15 days, followed by a plateau of stability during at least 30 days. In another study, Nogueira26 investigated solid state solar cells assembled with a similar polymer electrolyte, continuously irradiated using the same light source. A similar decay profile was observed in the first 600 h of irradiation, followed by 1080 h of stability. De Paoli and co-workers131 also investigated DSSC with a plasticized polymer electrolyte, based on the copolymer P(EO–EPI)87–13, NaI, I2 and the plasticizer GBL. The devices were irradiated under direct sunlight, and present a similar decay profile during the first 40 days after device assembly, reaching a constant performance after this period. Durrant and co-workers132 also investigated the stability of unsealed DSSC containing plasticized polymer electrolytes based on P(EO–EPI), EC/PC NaI and I2. With continuous illumination for 80 h at 20 mW cm−2 the devices lost only 15% of the initial performance. All these results suggest that the initial decay in performance followed by a plateau of stability might be an intrinsic property of DSSC assembled with polymer electrolytes and appears to be independent of the cell size, type of substrate employed and exposure time. Figure 10.22 presents a typical behavior of the variation of DSSC parameters with time, under direct sunlight exposure. De Paoli and co-workers38

0.8 1.0

0.5

0.6

6

0 0

Voc (V)

0.7

12

h (%)

Isc (mA)

18

10

20

30 40 50 Time (days)

60

0.0

0.5

0

10 20 30 40 50 60 70 Time (days)

10.22 Variation of (䊉) Isc, (䊏) Voc and (∗) η as a function of time for a DSSC assembled with a polymer electrolyte, exposed to direct sunlight for 1 h each day (at noon). Voc values stabilize within 20 days, while Isc keeps decreasing for 40 days. The efficiency behavior seems to be determined by the Isc characteristics.



419

demonstrated that no degradation of the plasticized polymer electrolyte occurs with time or during the operation of the solar cell under ambient conditions. The deposition of the polymer electrolyte under heating leaves almost no residual solvent in the electrolyte. Therefore, the loss of performance observed cannot be attributed to residual solvent evaporation.38 On the other hand, Park and co-workers133 demonstrated recently the self-degradation of DSSC containing PEO/PEG-based electrolytes due to shrinkage of the polymer electrolyte at room temperature and high humidity atmosphere (≥60%). According to the authors, this effect can be minimized by reducing the crystallinity degree of the polymer electrolyte, i.e. by adding large molecules such as imidazolium-based ionic liquids, or plasticizers. Some controversy in the literature concerning the ruthenium dye still persists. The N3 dye usually employed is believed to be able to sustain 108 redox cycles without noticeable loss of performance, corresponding to 20 years of operation under sunlight.2 However, when this dye is maintained in the oxidized state for long periods, it degrades through loss of the –NCS ligand. Therefore, regeneration of the dye in the photovoltaic cell should occur rapidly to avoid this unwanted side reaction, as the lack of adequate conditions for regeneration of the dye may lead to dye degradation.134 This is especially true for devices comprising polymer electrolytes, because of such factors as low ionic mobility that can lead to slow reageneration or the incomplete filling of the TiO2 sensitized electrode by the electrolyte. Also, upon exposure for prolonged periods of time at higher temperatures, such as 80–85 °C, degradation of performance in DSSC is frequently observed. Considering this, a few years ago Grätzel and co-workers114 synthesized a new amphiphilic ruthenium dye (Z-907), which was applied in a DSSC in combination with a heat resistant quasi-solid state electrolyte based on a mixture of imidazolium iodide, methoxypropionitrile and a fluorinated polymer. This device showed stability for 1000 h at 80 °C for the first time. Some authors reported the stability of DSSC assembled with polymer electrolytes with different compositions and related this behavior to the properties of the electrolyte. For example, solar cells were assembled with electrolytes combining the ionic liquid 1-methyl-3-propylimidazolium iodine and PEO, poly(propylene oxide) or the copolymer poly(ethylene oxide)-block-poly(propylene oxide)-block-poly(ethylene oxide).135 After introducing less than 10 wt% of the polyether into the liquid electrolyte, the parameters of these quasi-solid state solar cells were still comparable to those of the liquid photochemical cells. It was believed that the polymer can contribute to the maintenance of the efficiency of the solar cells by holding the organic solvent. All devices were sealed with paraffin wax. The cell assembled with liquid electrolyte lost half of the conversion efficiency after 5 days, while the cells fabricated with the polyether-electrolytes had


420


only 10–15% losses of efficiency. AFM images were obtained after one week, showing that for the liquid electrolytes the morphology presented holes, indicating a possible evaporation of the solvent, while the morphology for polymer electrolyte films showed more homogenous surfaces. Using poly(methyl methacrylate) as polymer host, EC, 1,2-propanediol carbonate and dimethyl carbonate as organic solvents and NaI/I2, Wu and co-workers136 prepared a gel electrolyte with high conductivity (6.9 × 10−3 S cm−1). The long-term stability of DSSC with the polymer gel electrolyte was compared with that of a device prepared with liquid electrolyte, where both cells were fabricated using the same methods, employing a cyanoacrylate adhesive and epoxy resin as sealants. After 5 days, the efficiency of the DSSC with the polymer gel electrolyte decreased by 8%, while the DSSC with liquid electrolyte lost 40% of the initial efficiency. After 40 days, the DSSC with the polymer gel electrolyte had maintained 83% of the original energy conversion efficiency, and the DSSC with liquid electrolyte had only 27%. Recently, Xia and co-workers137 also observed that a polymer gel electrolyte-based DSSC using a PVDF–HFP membrane maintained 77% of its initial efficiency in the same period, indicating that the liquid retaining ability of the porous polymer framework is extremely high. Addition of TiO2 nanoparticles can also extend the thermostability of the composite electrolyte in comparison to regular gel electrolyte and liquid electrolyte devices. The results of accelerated aging tests showed that the composite electrolyte-based devices containing PVDF–HFP and TiO2 could maintain 90% of their initial value after heating at 60 °C for 1000 h.138 Two different kinds of composite electrolyte, solidified with pure ZnO nanoparticles or with polymer-grafted ZnO nanoparticles, were applied to DSSC. These cells were sealed with thermal plastic tape for long-term test. The devices were stored at 55 °C and their efficiencies were measured once a weak. The solar cell containing the nanocomposite electrolyte with polymer-grafted nanoparticles kept 93% of its initial efficiency value, even under heating at 55 °C for 34 days, while the efficiency of a solar cell solidified with pure ZnO nanoparticles decreased to 60% of its initial value, which was attributed to a phase separation that occurs when the nonmodified particles are used.77 The stability of DSSC assembled with thermosetting gel electrolytes (TSGE) was also investigated. Wu and co-workers139 reported a TSGE based on poly(acrylic acid)–(ethylene glycol). In this copolymer, the liquid electrolyte absorbed is kept in the networks of the copolymer through chemical reactions, minimizing leakage or volatilization over an extended time. The solar cells assembled with this material presented an initial increase of performance during the first 18 days of measurement, which was attributed to a better penetration of the electrolyte into the pores of the TiO2 photoelectrode. The DSSC did not lose performance significantly for 50 days. Wu and co-workers140 also inves-



421

tigated a thermoplastic electrolyte prepared by heating a mixture of poly(ethylene glycol), PC, KI and I2. This electrolyte was applied to a DSSC and presented similar stability behavior as that observed for the TSGE electrolyte. The stability of other polymer and gel electrolytes based on PEO and PVDF derivatives have also been investigated.141–146 So far, most studies involving the stability of polymer electrolyte-based DSSC show the benefit of the replacement of the liquid component. The positive effect after longterm operation is more evident in the Jsc parameter. Besides, the use of the polymer electrolyte allows for easier assembly while different types of materials can be used as sealants.

10.6

Up-scaling: towards commercialization of polymer electrolyte-based dye-sensitized solar cells

A few years ago, Motohiro and co-workers147 reported a comparison between solar modules assembled with DSSC, with modules made of crystalline silicon solar cells commercialized by Siemens. During 6 months of outdoor exposure in the rooftop of a building located in the southern part of Kariya, Japan, the authors observed that the DSSC module generated 10–20% more energy than the commercial silicon-based module. These results are very promising and stimulate the research to bring DSSC into the market. The up-scaling of DSSC is not a straightforward process. When enlarging the active area of such devices, the resistivity of the FTO-glass substrate employed becomes very significant, and limits the current that is collected in the back contact (i.e. there is a loss in the maximum current that can flow through the device). Therefore, higher photocurrent values are obtained for devices with larger active areas; however, the increase in the current is not proportional to the area enlargement. This increase in the internal resistance is a drawback when upscaling these devices. The same effect is also observed for DSSC with liquid electrolytes, and such limitations in device performance were previously reported by Okada and co-workers.148 These authors showed that the lack of grid collectors in the substrates significantly reduces the performance of dye-sensitized solar cells assembled with liquid electrolytes. Therefore, the major limiting step for scaling-up DSSC lies in the modification of the FTO-glass substrate (by introduction of metallic grids or substitution of this material for metal electrodes, for example). Considering this, it should be pointed that, although the ionic conductivity of the polymer electrolyte is still one order of magnitude lower than the liquid electrolyte, the resistance in the solid electrolyte is not as crucial in determining the FF and η values for large-area DSSC as the FTO layer


422


10.23 Modules assembled with 16 series connected DSSC.

resistance is. Also, the use of the solid polymer or gel electrolyte might allow an easier assembly of the device, which is especially interesting when considering scaling-up. Nevertheless, there are only a few reports in the literature presenting attempts to scale-up solid state or gel DSSC,39,131,149 unlike the case of DSSC modules assembled with liquid electrolytes.150–167 A review of all different possible designs for DSSC module assembly was published by Tulloch.168 Biancardo and co-workers149 assembled a semitransparent quasi-solid state dye-sensitized solar module connecting in series solar cells assembled with a gel electrolyte. Each individual cell presented an efficiency of 1% under 100 mW cm−2. The efficiency of the whole module composed 23 cells (active area of 625 cm2) reached 0.3%. The authors claimed that, although the efficiency was higher for modules assembled with liquid electrolytes, the use of the gel electrolyte clearly improved the performance of the cells after 1 month, leading to devices with extended lifetime.149 Figure 10.23 shows a picture of two modules composed of 16 solar cells each connected in series, assembled in the LPCR, at the University of Campinas-UNICAMP, in Campinas, Brazil.131 The modules were irradiated with a commercial fluorescent lamp of 50 W, positioned 40 cm from the module. Using this light source, which is very diffuse (irradiation of 10 W m−2), the module presented Voc = 7.8 V. Under direct solar radiation, the same module presented Isc = 8.5 mA and Voc = 10.6 V at 800 W m−2, and Isc = 2.5 mA and Voc = 9.6 V at 100 W m−2. The series connection of solar cells can be made using different designs. In this work, the module was built using a design reported as a Z-series interconnected design, which comprises two opposing electrodes with the connection between cells consisting of a conducting medium (copper, in this case). The advantage of this design is high-voltage output, with relatively



423

Power (W m–2)

3.0

2.0

1.0

0.0 06:00

09:00

12:00 15:00 Time (hour)

18:00

10.24 Variation of the maximum generated power produced by a solar module assembled with 13 series connected DSSC during one day of outdoor exposure.

small ohmic losses, and facility of assembly. Besides, it makes it easy to identify and exchange cells that present a limited performance after module assembly. The disadvantage of this design is the low fill factor resulting from the series resistance of the interconnections, which can act as a barrier between cells. The total short-circuit current of the module corresponds to the average of the current generated by each cell and the open-circuit potential corresponds to the sum of the potential generated by each cell. The maximum power generated by the same module, operating with 13 series-connected solar cells, was ~3 W m−2, at 1 p.m. The integrated power generated during the whole period of irradiation, i.e. from 6 a.m. to 7 p.m., was 18 W m−2, as presented in Fig. 10.24. The efficiency of the module was evaluated through J–V curves obtained under solar irradiation of 1000 W cm−2. The values obtained were 0.5% and 0.3%, considering the active area (~60 cm2) and the total area (~100 cm2) of the module, respectively.131 These values are low in comparison to modules assembled with liquid electrolytes, but it should be considered that no modifications of the FTO-glass substrate were made to improve the current collection. Considering this, these results are very promising and might stimulate more research and efforts to scale-up DSSC modules based on polymer or gel polymer electrolytes.

10.7

Conclusions and future trends

The effective application of polymer electrolytes in dye-sensitized solar cells began only 10 years after the first announcement of efficient devices


424


containing liquid electrolytes. Since then, interest in this kind of technology has increased considerably, which is evident through the increasing number of papers published every year describing the assembly of dye-sensitized solar cells with polymer or gel electrolytes. A direct comparison between results reported by different groups, however, is not straightforward, since many parameters other than the electrolyte composition and conductivity can affect device performance, such as the TiO2 film thickness. Also, as is the case for DSSC with liquid electrolytes, the active area of the device still has a major role in the overall efficiency. Many techniques have already been used to understand and explain the charge transport mechanisms and reactions occurring inside the solid or gel electrolyte, such as transient absorption spectroscopy, complex electrochemical impedance spectroscopy, microelectrode voltammetry, Raman spectroscopy and nuclear magnetic resonance, providing powerful tools for the improvement of the so-called ‘solid state’ DSSC. Efficiencies can be further improved, although, considering the solid state nature of the materials and the diffusion limited photocurrent, it is unlikely to reach the 11% efficiencies obtained for liquid solar cells. Nevertheless, the 7–8% efficiencies reached so far can be considered very promising and, along with the stability results, might motivate the research in this field to bring these devices to the market. Efforts should now be directed to scaling-up such devices and also to the assembly of ‘all-flexible’ devices, combining the application of polymer electrolytes with flexible conductive substrates in substitution of the FTO-glass routinely employed.

10.8

Acknowledgements

The authors thank Daiso Co. Ltd, from Osaka, Japan, for supplying the copolymers used in this work, LNLS for the SEM images, and CNPq, Capes, Renami and Fapesp (fellowships 05/56924-0 and 06/58998-3) for financial support. We also thank Prof. Carol Collins for English revision.

10.9

References

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

b. o’regan and m. grätzel, Nature, 1991, 353, 737. m. grätzel, J. Photochem. Photobiol. A, 2004, 164, 3. k. kalyanasundaram and m. grätzel, Coord. Chem. Rev., 1998, 177, 347. a. hagfeldt and m. grätzel, Chem. Rev., 1995, 95, 49. c. a. kelly and g. j. meyer, Chem. Rev., 2001, 211, 295. m. grätzel, C. R. Chim., 2006, 9, 578. a. f. nogueira, c. longo and m.-a. de paoli, Coord. Chem. Rev., 2004, 248, 1455. b. li, l. wang, b. kang, p. wang and y. qiu, Sol. Energy Mater. Sol. Cells, 2006, 90, 549. 9. l. guo, x. pan and s. y. dai, Progr. Chem., 2008, 20, 1595.



425

10. j. h. wu, z. lan, s. c. hao, p. j. li, j. m. lin, m. l. huang, l. q. fang and y. f. huang, Pure Appl. Chem., 2008, 80, 2241. 11. d. m. li, d. qin, m. h. deng, y. h. luo and q. b. meng, Energy Environ. Sci., 2009, 2, 283. 12. y. wang, Sol. Energy Mater. Sol. Cells, 2009, 93, 1167. 13. d. e. fenton, j. m. parker and p. v. wright, Polymer, 1973, 14, 589. 14. m. b. armand, j. m. chabagno and m. j. duclot, in Fast Ion Transport in Solids: Solid State Batteries and Devices, ed. P. Vashisha, J. N. Mundy and G. K. Shenoy, North-Holland, New York, 1979. 15. c. a. vincent, Prog. Solid. State Chem., 1987, 17, 167. 16. t. yohannes, t. solomon and o. inganäs, Synthetic Met., 1996, 82, 215. 17. t. yohannes and o. inganäs, Sol. Energy Mater. Sol. Cells, 1998, 51, 193. 18. a. f. nogueira, n. alonso-vante and m.-a. de paoli, Synthetic Met., 1999, 105, 23. 19. a. f. nogueira and m.-a. de paoli, Sol. Energy Mater. Sol. Cells, 2000, 61, 135. 20. a. f. nogueira, j. r. durrant and m.-a. de paoli, Adv. Mater., 2001, 13, 826. 21. y. ren, z. zhang, e. gao, s. fang and s. cai, J. Appl. Electrochem., 2001, 31, 445. 22. w. a. gazotti, m. a. s. spinacé, e. m. girotto and m.-a. de paoli, Solid State Ionics, 2000, 130, 281. 23. m. a. ratner, in Polymer Electrolyte Reviews, ed. J. R. MacCallum and C. A. Vincent, Elsevier Applied Science, London, 1987, vol. 1, ch. 7, p. 173. 24. a. f. nogueira, m. a. s. spinace, w. a. gazotti, e. m. girotto and m.-a. de paoli, Solid State Ionics, 2001, 140, 327. 25. g. g. silva, n. h. t. lemes, c. n. p. fonseca and m.-a. de paoli, Solid State Ionics, 1997, 93, 105. 26. a. f. nogueira, Ph.D. Thesis, Universidade Estadual de Campinas, 2001. 27. a. f. nogueira, m.-a. de paoli, i. montanari, r. monkhouse, j. nelson and j. r. durrant, J. Phys. Chem. B, 2001, 105, 7517. 28. g. rothenberger, d. fitzmaurice and m. grätzel, J. Phys. Chem., 1992, 96, 5983. 29. b. enright, g. redmond and d. fitzmaurice, J. Phys. Chem., 1994, 98, 6195. 30. s. a. haque, y. tachibana, r. l. willis, j. e. moser, m. grätzel, d. r. klug and j. r. durrant, J. Phys. Chem. B, 2000, 104, 538. 31. g. p. kalaignan, m. s. kang and y. s. kang, Solid State Ionics, 2006, 177, 1091. 32. j. kang, w. li, x. wang, y. lin, x. xiao and s. fang, Electrochim. Acta, 2003, 48, 2487. 33. j. h. kim, m. s. kang, y. j. kim, j. won and y. s. kang, Solid State Ionics, 2005, 176, 579. 34. j. y. lee, b. bhattacharya, d.-w. kim and j.-k. park, J. Phys. Chem. C, 2008, 112, 12576. 35. t. kato, a. okazaki and s. hayase, Chem. Commun., 2002, 363. 36. f. m. gray, in Polymer Electrolyte Reviews, ed. J. R. MacCallum and C. A. Vincent, Elsevier Applied Science, London, 1987, vol. 1, ch. 6, p. 139. 37. d. saikia, c. c. han and y. w. chen-yang, J. Power Sources, 2008, 185, 570. 38. v. c. nogueira, c. longo, a. f. nogueira, m. a. soto-oviedo and m.-a. de paoli, J. Photochem. Photobiol. A, 2006, 181, 226. 39. j. n. de freitas, v. c. nogueira, b. i. ito, m. a. soto-oviedo, c. longo, m.-a. de paoli and a. f. nogueira, Int. J. Photoenergy, 2006, 75483. 40. c. longo, a. f. nogueira, m.-a. de paoli and h. cachet, J. Phys. Chem. B, 2002, 106, 5925.


426


41. j. bisquert, J. Phys. Chem. B, 2002, 106, 325. 42. b. a. boukamp, Equivalent Circuit Users Manual, University of Twente, The Netherlands, 1989. 43. j. r. macdonald, Impedance Spectroscopy: Emphasizing Solid Materials and Systems, Wiley, New York, 1987. 44. n. papageorgiou, m. grätzel and p. p. infelta, Sol. Energy Mater. Sol. Cells, 1996, 44, 405. 45. m. dürr, g. kron, u. rau, j. h. werner, a. yasuda and g. nelles, J. Chem. Phys., 2004, 121, 11374. 46. p. wang, s. m. zakeeruddin, p. comte, i. exnar and m. grätzel, J. Am. Chem. Soc., 2003, 125, 1166. 47. z. zebede and s. e. lindquist, Sol. Energy Mater. Sol. Cells, 1998, 51, 291. 48. g. oskam, b. v. bergeron, g. j. meyer and p. c. searson, J. Phys. Chem. B, 2001, 105, 6867. 49. k. hayamizu, y. aihara, s. araia and w. s. price, Electrochim. Acta, 2000, 45, 1313. 50. k. hayamizu, y. auharam, s. arai and w. s. price, Solid State Ionics, 1998, 107, 1. 51. j. n. de freitas, a. de s. gonçalves, m.-a. de paoli, j. r. durrant and a. f. nogueira, Electrochim. Acta, 2008, 53, 7166. 52. z. lan, j. h. wu, h. m. lin, m. l. huang, p. j. li and q. h. li, Electrochim. Acta, 2008, 53, 2296. 53. e. chatzivasiloglou, t. stergiopoulos, a. g. kontos, n. alexis, m. prodromidis and p. falaras, J. Photochem. Photobiol. A, 2007, 192, 49. 54. j. h. park, j. h. yum, s. y. kim, m. s. kang, y. g. lee, s. s. lee and y. s. kang, J. Photochem. Photobiol. A, 2008, 194, 148. 55. b. bhattacharya, j. y. lee, j. geng, h.-t. jung and j.-k. park, Langmuir, 2009, 25, 3276. 56. w. kubo, k. murakoshi, t. kitamura, s. yoshida, m. haruki, k. hanabusa, h. shirai, y. wada and s. yanagida, J. Phys. Chem. B, 2001, 105, 12809. 57. j. e. benedetti, a. s. gonçalves, a. l. b. formiga, m. a. de pauli, x. li, j. r. durrant and a. f. nogueira, J. Power Sources, 2010, 195, 1246. 58. l. andrews, e. s. prochaska and a. loewenschuss, Inorg. Chem., 1980, 19, 463. 59. m. a. tadayyoni, p. gao and m. j. weaver, J. Electroanal. Chem., 1986, 198, 125. 60. n. papageorgiou, t. athanassov, m. armand, p. bonhote, h. pettersson, a. azam and m. grätzel, J. Electrochem. Soc., 1996, 143, 3099. 61. i. montanari, j. nelson and j. r. durrant, J. Phys. Chem. B, 2002, 106, 12203. 62. a. n. m. green, e. palomares, s. a. haque, j. m. kroon and j. r. durrant, J. Phys. Chem. B, 2005, 109, 12525. 63. a. kassis and m. saad, Sol. Energy Mater. Sol. Cells, 2003, 80, 491. 64. m.-s. kang, j. h. kim, j. won and y. s. kang, J. Phys. Chem. C, 2007, 111, 5222. 65. f. croce, g. b. appetecchi, l. persi and b. scrosati, Nature, 1998, 394, 456. 66. g. katsaros, t. stergiopoulos, i. m. arabatzis, k. g. papadokostaki and p. falaras, J. Photochem. Photobiol. A, 2002, 149, 191. 67. t. stergiopoulos, i. m. arabatzis, g. katsaros and p. falaras, Nano Lett., 2002, 2, 1259. 68. e. chatzivasiloglou, t. stergiopoulos, n. spyrellis and p. falaras, J. Mater. Process. Tech., 2005, 161, 234. 69. h. w. han, w. liu, j. zhang and x.-j. zhao, Adv. Funct. Mater., 2005, 15, 1940.



427

70. j. zhang, h. han, s. wu, s. xu, c. zhou, y. yang and x. zhao, Nanotechnology, 2007, 18, 295606. 71. m.-s. kang, k.-s. ahn and j.-w. lee, J. Power Sources, 2008, 180, 896. 72. s. anandan, Curr. Appl. Phys., 2008, 8, 99. 73. j. zhang, h. w. han, s. xu, s. j. wu, c. h. zhou, y. yang and x. z. zhao, J. Appl. Pol. Sci., 2008, 109, 1369. 74. m. s. akhtar, j.-m. chun and o.-b. yang, Electrochem. Commun., 2007, 9, 2833. 75. h. w. han, u. bach, y.-b. cheng and r. a. caruso, Appl. Phys. Lett., 2007, 90, 213510. 76. j. zhang, h. han, s. wu, s. xu, y. yang, c. zhou and x. zhao, Solid State Ionics, 2007, 178, 1595. 77. x. zhang, h. yang, h.-m. xiong, f.-y. li and y.-y. xia, J. Power Sources, 2006, 160, 1541. 78. g. namutdinova, s. sensfuss, m. schrödner, a. hinsch, r. sastrawan, d. gerhard, s. himmler and p. wasserscheid, Solid State Ionics, 2006, 177, 3141. 79. b. i. ito, j. n. de freitas, m.-a. de paoli and a. f. nogueira, J. Braz. Chem. Soc., 2008, 19, 688. 80. c. w. tu, k. y. liu, a. t. chien, m. h. yen, t. h. weng, k. c. ho and k. f. lin, J. Polym. Sci. A, 2008, 46, 47. 81. m. k. nazeeruddin, a. kay, i. rodicio and m. grätzel, J. Am. Chem. Soc., 1993, 115, 6382. 82. j. wu, z. lan, d. wang, s. hao, j. lin, y. wei, s. yin and t. sato, J. Photochem. Photobiol. A, 2006, 181, 333. 83. h. paulsson, a. hagfeldt and l. kloo, J. Phys. Chem. B, 2003, 107, 13665. 84. s. ganesan, b. muthuraaman, j. madhavan, v. mathew, p. maruthamuthu and s. austin suthanthiraraj, Electrochim. Acta, 2008, 53, 7903. 85. s. y. huang, g. schlichthörl, a. j. nozik, m. grätzel and a. j. frank, J. Phys. Chem. B, 1997, 101, 2576. 86. s. mikoshiba, s. murai, h. sumino and s. hayase, J. Photochem. Photobiol. A, 2002, 148, 33. 87. s. murai, s. mikoshiba, h. sumino, t. kato and s. hayase, Chem. Commun., 2003, 1534. 88. m. y. li, s. j. feng, s. b. fang, x. r. xiao, x. p. li, x. m. zhou and y. lin, Electrochim. Acta, 2007, 52, 4858. 89. j. h. wu, s. hao, z. lan, j. m. lin, m. l. huang, y. f. huang, p. j. li, s. yin and t. sato, J. Am. Chem. Soc., 2008, 130, 11568. 90. j. zhang, y. yang, s. j. wu, s. xu, c. h. zhou, h. hu, b. l. chen, x. d. xiong, b. sebo, h. w. han and x. z. zhao, Electrochim. Acta, 2008, 53, 5415. 91. j. zhang, y. yang, s. j. wu, s. xu, c. h. zhou, h. hu, b. l. chen, x. d. xiong, b. sebo, h. w. han and x. z. zhao, Nanotechnology, 2008, 19, 245202. 92. y. j. kim, j. h. kim, m.-s. kang, m. j. le, j. won, j. c. lee and y. s. kang, Adv. Mater., 2004, 16, 1753. 93. m.-s. kang, j. h. kim, j. won and y. s. kang, J. Photochem. Photobiol. A, 2006, 183, 15. 94. m.-s. kang, j. h. kim, y. j. kim, j. won, n.-g. park and y. s. kang, Chem. Commun., 2005, 889. 95. k.-h. park, d.-w. park, m. dhayal and h.-b. gu, Electrochem. Commun., 2008, 10, 1098. 96. s. anandan, r. sivakumar and r. tharani, Synthetic Met., 2008, 158, 1067.


428


97. m. s. akhtar, k. k. cheralathan, j. m. chun and o. b. yang, Electrochim. Acta, 2008, 53, 6623. 98. c. shi, s. dai, k. wang, x. pan, l. z. l. hu, f. kong and l. guo, Electrochim. Acta, 2005, 50, 2597. 99. m. morita, h. tanaka, m. ishikawa and y. matsuda, Solid State Ionics, 1996, 86, 401. 100. j. e. benedetti, m.-a. de paoli and a. f. nogueira, Chem. Commun., 2008, 1121. 101. r. e. a. dillon and d. f. shriver, Chem. Mater., 1999, 11, 3296. 102. r. hagiwara and j. s. lee, Electrochemistry, 2007, 75, 23. 103. p. wang, s. m. zakeeruddin, j.-e. moser, r. h. baker and m. grätzel, J. Am. Chem. Soc., 2004, 126, 7164. 104. p. wang, b. wenger, r. h. baker, j.-e. moser, j. teuscher, w. kantlehner, j. mezger, e. v. stoyanov, s. m. zakeeruddin and m. grätzel, J. Am. Chem. Soc., 2005, 127, 6850. 105. m. gorlov and l. kloo, Dalton Trans., 2008, 2655. 106. t. kato, a. okazaki and s. hayase, J. Photochem. Photobiol. A, 2006, 179, 42. 107. s. mikoshiba, s. murai, h. sumino and s. hayase, Chem. Lett., 2002, 31, 918. 108. t. kato, m. fujimoto, t. kado, s. sakaguchi, d. kosugi, r. shiratuchi, w. takashima, k. kaneto and s. hayase, J. Electrochem. Soc., 2005, 152, A1105. 109. t. kato, a. ozaki and s. hayase, Chem. Commun., 2005, 363. 110. p. wang, s. m. zakeeruddin, i. exnar and m. grätzel, Chem.Commun., 2002, 2972. 111. g. c. li, z. h. li, p. zhang, h. p. zhang and y. p. wu, Pure Appl. Chem., 2008, 80, 2553. 112. a. r. howells, p. j. zambrano and m. m. collinson, Anal. Chem., 2000, 72, 5265. 113. w. zhang, i. gaberman and m. ciszkowska, Anal. Chem., 2002, 74, 1343. 114. p. wang, s. m. zakeeruddin, j. e. moser, m. k. nazeeruddin, t. sekiguchi and m. grätzel, Nat. Mater., 2003, 2, 402. 115. b. bhattacharya, s. k. tomar and j.-k. park, Nanotechnology, 2007, 18, 485711. 116. p. k. singh, k. w. kim, j. lee and h. w. rhee, Phys. Status Solidi A, 2006, 203, R88. 117. p. k. singh, k. w. kim, n. park and h. w. rhee, Macromol. Symp., 2007, 249/250, 162. 118. p. k. singh, k. w. kim and h. w. rhee, Pol. Eng. Sci., 2009, 49, 862. 119. f. s. freitas, j. n. de freitas, b. i. ito, m.-a. de paoli and a. f. nogueira, ACS Appl. Mater. Interfaces, 2009, 1, 2870. 120. h. j. shim, d. w. kim, c. j. lee, y. g. kang and d. h. suh, Macromol. Res., 2008, 16, 424. 121. m. wang, y. lin, x. w. zhou, x. r. xiao, l. yang, s. j. feng and x. p. li, Mater. Chem. Phys., 2008, 107, 61. 122. m. wang, x. xiao, x. zhou, x. li and y. lin, Sol. Energy Mater. Sol. Cells, 2007, 91, 785. 123. h. ohno, Electrochim. Acta, 2001, 46, 1407. 124. m. wang, x. yin, x. r. xiao, x. w. zhou, z. z. yang, x. p. li and y. lin, J. Photochem. Photobiol. A, 2008, 194, 20. 125. k. suzuki, m. yamaguchi, s. hotta, n. tanabe and s. yanagida, J. Photochem. Photobiol. A, 2004, 164, 81. 126. f. li, f. cheng, j. shi, f. cai, m. liang and j. chen, J. Power Sources, 2007, 165, 911. 127. b. yu, f. zhou, c. wang and w. liu, Eur. Pol. J., 2007, 42, 2699.



429

128. l. wang, s. b. fang and y. lin, Polym. Advan. Technol., 2006, 17, 512. 129. w. kubo, y. makimoto, t. kitamura, y. wada and s. yanagida, Chem. Lett., 2002, 31, 948. 130. c. longo, j. freitas and m.-a. de paoli, J. Photochem. Photobiol. A, 2003, 159, 33. 131. j. n. de freitas, c. longo, a. f. nogueira and m.-a. de paoli, Sol. Energy Mater. Sol. Cells, 2008, 92, 1110. 132. s. a. haque, e. palomares, h. m. upadhyaya, l. otley, r. j. potter, a. b. holmes and j. r. durrant, Chem. Commun., 2003, 3008. 133. j. y. lee, b. bhattacharya, y. h. kim, h.-t. jung and j.-k. park, Solid State Commun., 2009, 149, 307. 134. r. grünwald and h. tributsch, J. Phys. Chem. B, 1997, 101, 2564. 135. j. b. xia, f. y. li, c. h. huang, j. zhai and l. jiang, Sol. Energy Mater. Sol. Cells, 2006, 90, 944. 136. h. x. yang, m. l. huang, j. h. wu, z. lan, s. c. hao and j. m. lin, Mater. Chem. Phys., 2008, 110, 38. 137. x. zhang, c.-x. wang, f.-y. li and y.-y. xia, J. Photochem. Photobiol. A, 2008, 194, 31. 138. z. p. huo, s. y. dai, k. j. wang, f. t. kong, c. n. zhang, x. pan and x. q. fang, Sol. Energy Mater. Sol. Cells, 2007, 91, 1959. 139. j. h. wu, z. lan, j. m. lin, m. l. huang, s. c. hao, t. sato and s. yin, Adv. Mater., 2007, 19, 4006. 140. j. h. wu, s. hao, z. lan, j. lin, m. huang, y. huang, l. fang, s. yin and t. sato, Adv. Funct. Mater., 2007, 17, 2645. 141. s. b. fang, y. lin, x. w. zhou, j. b. zhang, x. p. li and x. r. xiao, Acta Polym. Sin., 2008, 6, 507. 142. p. petrov, i. berlinova, c. b. tsvetanov, s. rosselli, a. schmid, a. b. zilaei, t. miteva, m. durr, a. yasuda and g. nelles, Macromol. Mater. Eng., 2008, 293, 598. 143. a. r. s. priya, a. subramania, y. s. jung and k. j. kim, Langmuir, 2008, 24, 9816. 144. y. yang, c.-h. zhou, s. xu, h. hu, b.-l. chen, j. zhang, s.-j. wu, w. liu and x.-z. zhao, J. Power Sources, 2008, 185, 1492. 145. j. j. kim, h. choi, j. w. lee, m. s. kang, k. song, s. o. kang and j. ko, J. Mater. Chem., 2008, 18, 5223. 146. h. han, u. bach, y.-b. cheng, r. a. caruso and c. macrae, Appl. Phys. Lett., 2009, 94, 103102. 147. t. toyoda, t. sano, j. nakajima, s. doi, s. fukumoto, a. ito, t. tohyama, m. yoshida, t. kanagawa, t. motohiro, t. shiga, k. higuchi, h. tanaka, y. takeda, t. fukano, n. katoh, a. takeichi, k. takechi and m. shiozawa, J. Photochem. Photobiol. A, 2004, 164, 203. 148. k. okada, h. matsui, t. kawashima, t. ezure and n. tanabe, J. Photochem. Photobiol. A, 2004, 164, 193. 149. m. biancardo, k. west and f. c. krebs, Sol. Energy Mater. Sol. Cells, 2006, 90, 2575. 150. a. kay and m. grätzel, Sol. Energy Mater. Sol. Cells, 1996, 44, 99. 151. h. pettersson and t. gruszecki, Sol. Energy Mater. Sol. Cells, 2001, 70, 203. 152. h. pettersson, t. gruszecki, l.-h. johansson and p. johander, Sol. Energy Mater. Sol. Cells, 2003, 77, 405.


430


153. s. y. dai, j. weng, y. f. sui, c. w. shi, y. huang, s. h. chen, x. pan, x. q. fang, l. h. hu, f. t. kong and k. j. wang, Sol. Energy Mater. Sol. Cells, 2004, 84, 125. 154. s. y. dai, k. j. wang, j. weng, y. f. sui, y. huang, s. f. xiao, s. h. chen, l. h. hu, f. t. kong, x. pan, c. w. shi and l. guo, Sol. Energy Mater. Sol. Cells, 2005, 85, 447. 155. r. sastrawan, j. renz, c. prahl, j. beier, a. hinsch and r. kern, J. Photochem. Photobiol. A, 2006, 178, 33. 156. r. sastrawan, j. beier, u. belledin, s. hemming, a. hinsch, r. kern, c. vetter, f. m. petrat, a. prodi-schwab, p. lechner and w. hoffmann, Sol. Energy Mater. Sol. Cells, 2006, 90, 1680. 157. r. sastrawan, j. beier, u. belledin, s. hemming, a. hinsch, r. kern, c. vetter, f. m. petrat, a. prodi-schwab, p. lechner and w. hoffmann, Prog. Photovoltaics, 2006, 14, 697. 158. e. ramasamy, w. j. lee, d. y. lee and j. s. song, J. Power Sources, 2007, 165, 446. 159. w. j. lee, e. ramasamy, d. y. lee and j. s. song, Sol. Energy Mater. Sol. Cells, 2007, 91, 1676. 160. j. weng, s. f. xiao, s. h. chen and s. y. dai, Acta Phys. Sin., 2007, 56, 3602. 161. w. j. lee, e. ramasamy, d. y. lee and j. s. song, J. Photochem. Photobiol. A, 2008, 194, 27. 162. s. y. dai, j. weng, y. f. sui, s. h. chen, s. f. xiao, y. huang, f. t. kong, x. pan, l. h. hu, c. n. zhang and k. j. wang, Inorg. Chim. Acta, 2008, 361, 786. 163. y. jun, j.-h. son, d. sohn and m. g. kang, J. Photochem. Photobiol. A, 2008, 200, 314. 164. l. t. han, a. fukui, y. chiba, a. islam, r. komiya, n. fuke, n. koide, r. yamanaka and m. shimizu, Appl. Phys. Lett., 2009, 94, 013305. 165. n. kato, y. takeda, k. higuchi, a. takeichi, e. sudo, h. tanaka, t. motohiro, t. sano and t. toyoda, Sol. Energy Mater. Sol. Cells, 2009, 93, 893. 166. m. ikegami, j. suzuki, k. teshima, m. kawaraya and t. miyasaka, Sol. Energy Mater. Sol. Cells, 2009, 93, 836. 167. a. hinsch, h. brandt, w. veurman, s. hemming, m. nittel, u. würfel, p. putyra, c. lang-koetz, m. stabe, s. beucker and k. fichter, Sol. Energy Mater. Sol. Cells, 2009, 93, 820. 168. g. e. tulloch, J. Photochem. Photobiol. A, 2004, 164, 209.


11 Solid polymer electrolytes for supercapacitors A. B. S A M U I and P. S I VA R A M A N, Naval Materials Research Laboratory, India

Abstract: This chapter reviews state-of-the-art development in the field of solid polymer electrolytes for supercapacitors and discusses their specific features. The main themes discussed are based on individual series of solid polymer electrolytes such as ion exchange resins, polyethers, polymer gels and nanocomposites, followed by supercapacitor types, electrode types and supercapacitors reported under various kinds of solid electrolytes. This is followed by information on hybrid supercapacitors and current research activities and finally a conclusion. Key words: solid polymer electrolytes, supercapacitors, ionic conductivity, electrode materials, specific capacitance, internal resistance, cycle life.

11.1

Introduction

All solid-state supercapacitors have several advantages over liquid-based ones. The solid electrolytes used in the fabrication of supercapacitors can be broadly classified into two categories. Firstly, solid electrolytes containing a base polymer like poly(ethylene oxide) (PEO), poly(acrylonitrile) (PAN), poly(methylmethacrylate) (PMMA), nylon, etc., with an acid, base or salt dissolved in it, are discussed. A solid electrolyte can be classified into aqueous based or non-aqueous based, depending on the plasticizer used in it. Secondly, electrolytes based on ion-exchange membranes, like Nafion®, are discussed. In this type of electrolyte, water is generally used for solvation and as the medium for proton transport. Both electric double-layer capacitors and pseudocapacitors are included as these have been studied in solid electrolyte configurations. Most of the studies reported are on pseudocapacitors. Polyaniline, polypyrrole, and polythiophene, and substituted polythiophene were used for fabricating the electrodes. Gels were made by mixing a polymer or polymer blends, respectively, with plasticizers such as propylene/ethylene carbonate. For the cation exchange type, the studies mostly discuss Nafion, although some discuss other sulfonated polymers. Hybrid capacitors with solid electrolytes have not progressed much and only one report on them has appeared so far in the literature. In the following section, the development of various solid electrolytes is discussed, 431 © Woodhead Publishing Limited, 2010

432


followed by the fabrication and characterization of solid electrolyte supercapacitors.

11.2

Solid electrolytes

It is pertinent to discuss the various kinds of solid electrolytes already developed and their properties. This will enable us to understand their usefulness in supercapacitor applications. Electrolytes are materials which commonly exist as solutions of acids, bases or salts, although a few gases may act as electrolytes under specific conditions such as high temperature or low pressure. Electrolytes can be formed by the dissolution of some biological (polypeptides) and synthetic polymers (sulfonated polymer). With the advances in electrolyte science and technology, a new concept has been developed and the new materials are known as solid electrolytes. This type of electrolyte is also known as a solid polymer electrolyte (SPE), gel electrolyte, gel polymer electrolyte (GPE) and so on. The practical importance of this new type of electrolyte was immediately recognized and growth in this area of research was rapid in the following few years. It has many advantages over liquid electrolytes such as being environmentally safe since there is no risk of spillage in electrical energy systems; it can be used as a thin film, both as an electrolyte and a separator; etc. Work on ion-conducting polymers started after it was reported during 1951 that salts can interact with PEO chains1 (Fig. 11.1) and the properties of polymer salt solutions were studied during the 1960s.2,3 Ionic conductivity in a PEO–alkaline metal ion complex was first reported by Wright in 1975.4 As the concentration of lithium salt was increased in the PEO, a general reduction in both the conductivity and number of lithium transfers was observed.5 The reduction was attributed to the motion of the polymer chains, responsible for ion mobility, being restricted and also formation of ion pairs. In turn, it lowered the number of free lithium ions available for conduction.6 Initial work was carried out by Armand as he realized that this

CH2

CH2

CH2

CH2

CH2

CH2

CH2

CH2

O

CH2

CH2

CH2

CH2

CH2

CH2

CH2

O Li+

O

CH2

O

11.1 Chemical structure of PEO–lithium salt complex.


Solid polymer electrolytes for supercapacitors

433

material could be used as a polymer electrolyte in batteries.7,8 After that, research and development for the SPE gained momentum. The increasing interest in the development of the SPE was mainly for its technological application in thin-film batteries and supercapacitors, electrochromic displays, smart windows, sensors, etc. The primary requirement of an SPE is ionic conductivity. The conductivity observed at the initial stage of development was 1 × 10−7 S/cm at room temperature. Research was intensified to improve it and it reached 1 × 10−3 S/cm at room temperature.

11.3

Conduction in solid electrolytes

The concentration of ionic charge carriers in the electrolyte, in which the metal salts form complexes, depends on the dielectric constant and the lattice energy of the salt: the higher the dielectric constant and/or the lower the lattice energy, the higher the charge carrier concentration. With regard to thermodynamics, the inorganic salt will dissolve if it reduces the Gibbs free energy of the system i.e. there must be sufficient interaction between the salt and the solvent to overcome the lattice energy. The free energy expression can be given as: ΔG = ΔH − TΔS where ΔH represents the change in enthalpy, ΔS the change in entropy and T the absolute temperature. As the crystal dissolves, the crystal lattice disappears, resulting in an increase in entropy. The enthalpy change is generally associated with long-range electrostatic force, interaction between the solvent and ions, such as complexation, and non-electrostatic interaction. The ionic conductivity (σ) of an electrolyte can be written as:

σ = ∑ nzμ where n is the number of charge carriers, z is the ionic charge and µ is the ionic mobility. The ionic conductivity is generally seen with an amorphous polymer. The ion diffusion in a polymer electrolyte is assisted by the local motion of the polymer chain. Therefore, the polymer chain must be flexible at the application temperature. This is conventionally achieved by choosing a polymer with a glass transition temperature (Tg) below the temperature being used or by blending it with another polymer and/or additives. At the same time, it needs to have sufficient mechanical stability for such an application. However, there are some vitreous solid electrolytes which are used above Tg in which ion diffusion is decoupled from the motion of the matrix polymer. The cation migration with segmental mobility of the polymer chain is shown in Fig. 11.2.9


434

Polymer electrolytes O

O O O

+ O

O

O O

+

O O

O

O

O

O

O

O

O

+

O O

O

O

O O

O

11.2 Cation ion with segmental mobility of the polymer chain.

CF2

CF2 CF2

x

CF O

y CF2

O CF z

O

CF2 S

CF2

OH CF3

O

11.3 Chemical structure of Nafion.

In parallel with the Williams–Landel–Ferry (WLF) equation, the Vogel– Tamman–Fulcher (VTF) equation is used to express the temperature dependency of conductivity: σ = AT−1/2 exp(−B/T − T0) where A and B are constants related to charge carrier density and the activation energy and T0 is the temperature at which the configurational entropy of the polymer becomes zero. A great deal of research has been performed on polymer electrolytes using PEO, but the problem with PEO is its crystallinity. Crystalline polymers, or crystalline regions in semicrystalline polymers, do not allow ions to move freely. Therefore, attempts have been made to modify the polymer by forming copolymers,10 forming blends,11 blending with inert fillers,12 or by using heteropolymer13 and so on. There is another series of polymers, which are usually called proton conductors. Advancement in this field was propelled by continuous improvement in properties coupled with demands from various technological applications. Some of the developments will be discussed here.

11.3.1 Ion exchange polymers When Nafion was discovered in 1962, it was considered to be a solid polymer electrolyte due to the presence of a sulfonic acid group attached to the perfluorinated chain (Fig. 11.3). It is not a true solid proton conductor as ion transport occurs via a quasi-liquid phase as in other polyelectrolytes. In



435

a saturated condition, the Nafion membrane displays an ionic conductivity greater than 10−3 S/cm at room temperature. Nafion, with its widespread use in fuel cells, water electrolysis, plating, sensors, etc., has been extensively studied over the years,14,15 together with the development of cheaper materials. Nafion remains the material of choice for commercial exploitation because of its thermal stability and the high conductivity of its cations, together with its resistance to chemicals. In one study, the Nafion membrane was used as a solid polymer electrolyte by replacing the conductive liquid in a cell for electro-organic synthesis at an electrode.16 Alkoxylations of N-alkyl amides gave selectivity of nearly 100% at a low cell voltage. Its limited availability and high cost led to the further development of similar kinds of material. A partially sulfonated polystyrene membrane was developed for use in direct methanol fuel cells (DMFC).17 The membranes were characterized in terms of sulfonation vs. conductivity and methanol permeability. The new membrane, having the highest sulfonation, allowed methanol permeability about 70% lower than Nafion. Sulfonated poly(ether ether ketone) (SPEEK) is another sulfonated membrane that has been widely studied as a polymer electrolyte18–20 to eliminate some of the problems associated with Nafion and its cheaper alternative. Another series of membranes, called proton donor–acceptor membranes, was reported to have been used as a polymer electrolyte membrane21–23 in the study of methanol permeability. A blend of polybenzimidazole and sulfonated polysulfone was reported to exhibit lower methanol permeability than Nafion. The use of a number of other composite membranes has been reported for various applications. Chen et al. studied polystyrene sulfonate blended with a micrometer-sized, crosslinked polystyrene sulfonate composite membrane for fuel-cell applications.24 A decrease in cell potential to 60% was observed in 340 h as against 55 h in the polystyrene sulfonate membrane. The proton-conducting properties of temperature-tolerant benzimidazole-doped poly(vinyl phosphonic acid) were studied by Sevil and Bozkurt.25 The DC conductivity increased with benzimidazole content, reaching 10−3 S/cm at 150 °C. Cheap engineering plastics such as SPEEK26,27 and polybenzimidazole (PBI)28 are some of the other alternative materials studied after modifying them with protonic acids. SPEEK was modified further by making a composite with boron phosphate up to 40 wt%.29 A six-fold increase in conductivity was observed for the composite membrane containing 30 wt% boron phosphate. Licoccia et al. reported on a composite made from SPEEK and sulfonated diphenylsilanediol.30 The membrane showed properties superior to SPEEK in terms of water uptake and solubility. The membrane was found to have high conductivity of 0.1 S/cm and good proton transport properties up to 120 °C. A series of sulfonated poly(arylene ether) copolymerized with other moieties was synthesized and studied as a proton conductor.31–33 The copolymer,


436


containing hexafluoroisopropylidene diphenol-based poly(arylene ether) with 35 mol% disulfonated comonomer, had a proton conductivity > 0.10 S/ cm at 100 °C and 100% relative humidity.33 The copolymer exhibited nanophase separated co-continuous morphology in the hydrophobic and hydrophilic domains. A series of sulfonated polyimide- and copolyimide-based membranes were reported by various groups.34–39 The dimensional stability of most of the imides in water is poor. A copolyimide synthesized from 1,4,5,8-naphthalenetetracarboxylic dianhydride, bis(3-sulfopropoxy) benzidines and common non-sulfonated diamines showed conductivity in the range 0.05–0.16 S/cm at 50 °C in water and minimum dimensional change.39 A series of polymer gels derived from alkali metal ionic liquids and negatively charged polyelectrolytes were studied for their ion transport properties.40–42 The conductivity observed was mostly in the range of 10−4 to 10−3 S/cm at room temperature. The gels showed a decrease in conductivity with increasing polymer fraction.

11.3.2 Polyether-based polymers A polyethylene oxide-based solid polymer electrolyte was studied initially by making a copolymer, blend, composite, etc. and by varying the type of salt to achieve the maximum amorphous nature and conductivity of a solid polymer electrolyte.43–51 Morisatoa et al. described a block copolymer of nylon-12/tetramethylene oxide used as a polymeric matrix for silver tetrafluoroborate (AgBF4) containing solid polymer electrolyte membranes.52 The ethylene/ethane permeation properties of the membrane were observed to be strongly dependent on the silver salt content. It was reported that Ag+ ions were mobile in the polymer matrix only when the AgBF4 concentration was high enough (>67 wt%). Triblock copolymers of central poly(ethylene glycol) (PEG) or poly(ethylene glycol-co-propylene glycol) (PEGPG) blocks with poly(pentafluorostyrene) (PFS) outer blocks were prepared by atom transfer radical polymerization (ATRP), having polydispersities around 1.2–1.3 (Fig. 11.4).53 Polymer electrolytes were prepared by comO O Br F

) 4 F

F

F

(

((O

)x(

O

)1–x )y O

( F

) Br 4 F

F

F

F

F

11.4 Triblock copolymers of central poly(ethylene glycol-co-propylene glycol) (PEGPG) blocks with poly(pentafluorostyrene) (PFS) outer blocks.



437

plexing lithium bis(trifluoromethylsulfonyl)imide (LiTFSI) salt with the block copolymers and the liquid PEGPG precursor. The crystallinity and melting points of the salted monomethoxy poly(ethylene glycol) (MPEG)b-PFS and triblock copolymers were found to be substantially suppressed. Short PFS blocks (Tg = 33 °C) and the salted PEGPG blocks (Tg = −65 °C) were immiscible and an elastomeric material was obtained in the form of a physically cross-linked polyether network. Their conductivity was found to be about 10−5 S/cm at 20 °C. The ionic conductivity of lithium-based solid polymer films, prepared from PEO and lithium hexafluoarsenate (LiAsF6) with varying compositions of plasticizers like dibutyl sebacate (DBS) and ethylene carbonate (EC), was measured by the AC impedance method.54 The addition of DBS increased the ionic conductivity from 5.98 × 10−7 S/cm to 10−6 S/cm. An attempt was made to make a solid polymer electrolyte using layerby-layer deposition, which was possible due to the hydrogen bonding ability of alternate layer polymers. Films were fabricated by the deposition of alternate layers of PEO and poly(acrylic acid) (PAA) from aqueous solutions.55 The film quality was found to be enhanced by increasing the PEO molecular weight, even up to 106, due to the intrinsically low PEO/PAA crosslink density. An ionic conductivity of 5 × 10−5 S/cm was achieved after a short exposure to 100% relative humidity (RH) for plasticization. It was observed that exposing PEO/PAA films to lithium salt solutions enhanced their conductivity to greater than 10−5 S/cm at 52% RH and greater than 10−4 S/cm at 100% RH. A series of PEG–esters were used as a plasticizer for an SPE. Masuda et al. used PEG–aluminate ester as a plasticizer for solid polymer electrolytes.56 The thermal stability, ionic conductivity and electrochemical stability of a polymer electrolyte containing PEO-based copolymer, PEG-aluminate ester and LiTFSI were investigated. Adding PEG–aluminate ester increased the ionic conductivity of the polymer electrolyte to a value greater than 10−4 S/cm at 30 °C. It had sufficient thermal stability and exhibited electrochemical stability up to 4.5 V vs. Li+/Li at 30 °C. PEG–borate ester was studied as a plasticizer for an SPE in a lithium ion secondary battery.57 Adding the PEG–borate ester into the electrolyte increased the ionic conductivity of the polymer electrolyte. The temperature dependence of the ionic conductivity of the polymer electrolytes was investigated using a WLF-type equation. It was concluded that the PEG–borate ester did not have any influence on the dissociation of the Li salt. In another study, PEG–borate ester plasticizer was used with a poly(ethylene glycol) methacrylate (PEGMA) and LiTFSI solid polymer electrolyte.58 The ionic conductivity of the polymer electrolyte increased with increasing amounts of the PEG–borate ester and exhibited values greater than 10−4 S/cm at 30 °C and 10−3 S/cm at 60 °C. The polymer electrolyte containing the PEG–borate


438


ester, having an ethylene oxide (EO) chain length (n) of 3, showed the highest ionic conductivity. The electrolytes were thermally stable up to 300 °C and electrochemically stable up to 4.5 V vs. Li+/Li.

11.3.3 Polymer gels A common method of making gels is to blend together a polar polymer, other than a PEO-type polymer containing complex-forming moieties, salts and an organic liquid with a high boiling point; the liquid acts as a plasticizer. The polymer network prevents the liquid from escaping from the matrix and gives the matrix both liquid and solid characteristics. If the polymer is able to trap more liquid, the mobility of the ions is also higher, producing higher conductivity. The ion pairing and aggregation are expected to be lower compared to systems without a solvent. A few of the polymer gels that have been developed will be mentioned here. A biodegradable gel polymer electrolyte was produced by casting through the dissolution of poly(ε-caprolactone) (PCL) in tetrahydrofuran (THF) together with propylene carbonate (PC) and salts such as LiClO4, LiF3CSO3 and LiBF4.59 The ionic conductivities of the PCL/10 wt% PC and 12 wt% of salts at room temperature were about 2.26 × 10−4, 4 × 10−5 and 1.5 × 10−7 S/cm respectively. The poly(vinylidene fluoride) (PVDF) : LiFePO4 complex membranes as solid polymer electrolytes were characterized. X-ray diffraction (XRD) and differential scanning calorimetric (DSC) studies showed a decrease in crystallite size and crystallinity of the polymer with increasing LiFePO4 concentration.60 Several authors have studied the gel by using PVDF or PVDF-based copolymers.61–65 A series of studies were reported that used PMMA as the host polymer.66–71 In one study, using polymer electrolyte films prepared from PMMA and LiBF4 with different concentrations of plasticizer (dibutyl phthalate, DBP), the conductivity observed was in the range 4.5 × 10−3 to 3.3 × 10−4 S/cm for various mole ratios of PMMA and DBP at 304 K.71 Other types of gel were also studied as solid electrolytes, such as methacrylonitrile (MAN) polymerized in the presence of ethylene glycol dimethacrylate (EGDMA), propylene carbonate and tetraethylammonium tetrafluoroborate (TEABF4). These exhibited high conductivity (>10−3 S/ cm) at room temperature.72 Furthermore, these polymer electrolytes showed good electrochemical stability windows in the range of −4.0 to +4.0 V versus Ag. A series of conducting thin-film solid electrolytes based on a poly(vinyl alcohol)/poly(vinyl pyrrolidone) (PVA/PVP) polymer blend was made by the solution-casting technique.73 The conductivity of the PVA/PVP blend that had a composition of 80% PVA and 20 wt% PVP was the highest at around 10−7 S/cm. The PVA/PVP blend with the highest conductivity was then studied further by adding different amounts of potassium hydroxide



439

(KOH) ionic dopant. Water was used as a solvent to prepare PVA/PVP– KOH-based alkaline solid polymer blend electrolyte films. Conductivity increased to 10−4 S/cm by adding 40 wt% KOH.

11.3.4 Nanocomposite electrolytes There are many reports on nanocomposite usage in electrolytes as it enhances conductivity by various mechanisms such as reducing the crystallinity, increasing the salt solubility etc. One study reported that the ionic conductivity of PEO-based SPEs was improved by the addition of nanosized ceramic powders (TiO2 and Al2O3).74 The PEO-based solid polymer electrolytes were prepared by the solution-casting method. Electrochemical measurement showed that the 10 wt% TiO2/PEO–LiClO4 polymer electrolyte had high ionic conductivity. The lithium transference number of the 10 wt% TiO2/PEO–LiClO4 polymer electrolyte was measured to be 0.47, which was much higher than that of the unmodified PEO polymer electrolyte. AC impedance analysis revealed that the interface resistance of the ceramic-enhanced PEO polymer electrolyte was stable. Linear sweep voltammetry measurements showed the electrolytes were electrochemically stable in the voltage range of 2.0–5.0 V versus an Li/Li+ reference electrode.

11.4

Solid electrolytes in supercapacitors

Solid electrolytes have become very important for use in energy devices for various reasons such as: they have a long cycle life; they have a short charging time; they contain no hazardous liquids that can be spilled so they are environmentally safe; tailor-made electrolytes are easy to produce; they have low internal corrosion; they have simple principles and modes of construction; flexibility in packaging, etc. It is well known that electrochemical capacitors store energy within the double electrochemical layer at the electrode/electrolyte interface. They are commonly known as ‘electrochemical capacitors’, ‘supercapacitors’ and ‘ultracapacitors’, among others. Replacing the liquid electrolyte with a solid electrolyte such as an organic polymer,75–79 inorganic silica gel,80,81 etc., was expected to increase their reliability for various practical applications. Traditional capacitors, such as those used in electronic circuits, cannot store sufficient energy for high end applications such as telecommunication devices, electric hybrid vehicles and other fast charge–discharge applications. The requirements of high power density applications have led to the development of supercapacitors or electrochemical capacitors. The electrochemical capacitor has been around for quite a long time. The first patent for high surface area carbon appeared in 1957,82 and the marketing of this


440

Polymer electrolytes 10 000 000 Power density (W/kg)

1 000 000

Capacitors

100 000 10 000 1 000

Supercapacitors

100

Batteries

10

Fuel cells

0 0.01

0.1

1

10

100

1 000

10 000

Energy density (W h/kg)

11.5 A typical Ragone plot for various energy storage and conversion devices. (The indicated areas are approximate only.)

Electrode and electrolyte material Current collector Separator membrane

11.6 A typical construction of supercapacitor cell.

device was attempted by SOHO.83 However, it was only the invention of electric vehicles in the 1990s that made the electrochemical capacitor well known. The reason that the electrochemical capacitor captured the attention of researchers and entrepreneurs can be clearly seen in Fig. 11.5.84 The figure describes the ‘Ragone plot’ of typical energy devices in terms of their specific energy and power. It can be seen that electrochemical capacitors fill the gap between batteries and conventional capacitors. A typical supercapacitor is constructed by assembling a separator sandwiched between two electrodes. The electrodes are attached to current collectors made of metal or carbon paper/cloth (Fig. 11.6).

11.4.1 Classification of electrochemical capacitors Various types of supercapacitor are shown in Fig. 11.7.



441

Supercapacitors

Electric double layer capacitors

Activated carbons

Carbon nanotubes

Pseudocapacitors

Carbon aerogels

Conducting polymers

Metal oxides

Hybrid capacitors

11.7 Types of supercapacitor.

Double-layer capacitors In this type of capacitor, an electric double layer is formed at the interface between the solid electrode surface and the liquid or SPE. The carbon material acquires a high capacitance due to its high surface area. Every macropore can constitute one tiny double layer and can contribute to the total capacitance. The active electrode materials used in double-layer electrochemical capacitors are carbon black, carbon cloth, carbon aerogel, etc. Pseudocapacitors In a double-layer capacitor, no faradic reaction takes place between the electrode and electrolyte. In a pseudocapacitor, most of the charge is transferred at the interface or by the material near the surface of the electrode, which does involve a faradic reaction. The most interesting case is the conducting polymer, in which charging/discharging is possible in the material itself, raising the possibility of increasing the capacitance to a new level. Hybrid capacitors A supercapacitor can be fabricated to have a double layer material in one electrode while the other electrode is made of pseudocapacitor material. This configuration is known as a hybrid capacitor. The strategy is used to


442


achieve both high energy and high power density. Other variations of hybrid capacitors can use two non-similar metal oxides for the two electrodes, or different conducting polymers for the two electrodes. Solid electrolytes The prerequisites for a supercapacitor are low resistance, low equivalent series resistance (ESR), and a fast charge/discharge ability. All these properties are dependent on every component of a supercapacitor: electrode type, electrolyte conductivity, interfaces between electrode/electrolyte and electrode/current collector. For example, achieving a high capacitance requires high ionic conductivity of the electrolyte, high electrode electronic conductance, and a thin separator and electrode. When using a solid electrolyte, every effort must be made to achieve the best possible fabrication. Organic electrolytes can give a higher energy density but have a low power output due to the increase in internal resistance (ESR). There are already many solid electrolytes available that are suitable for use in a supercapacitor. Most of the recent reports say their conductivity is in the region of 10−3 S/cm at room temperature. However, their cycling ability and outputs are the ultimate tests for practical applications. Various reports will be discussed here to explain their effective use in supercapacitors, their performance evaluation, critical analysis, etc. Polymer gel-based supercapacitors A gel is an elastic colloid or polymer network that remains expanded throughout its whole volume by a fluid. The polymer network can be formed by either chemical bonds or physical aggregation. The advantage of a polymer gel is that it can easily be tailor-made so that all its properties are optimized with respect to its performance in supercapacitors. A fair number of studies have contributed to the continuous development of gel varieties and properties.

11.5

Conducting polymer electrodes

Conducting polymers are known for their pseudocapacitance and the doping/dedoping that occurs in the body of material. This makes them potential candidates for providing large capacitance in supercapacitors. The polymers are usually used in a non-aqueous configuration. We will discuss a few of the reports of their cell configuration and performance. Substituted thiophenes are the ideal choice owing to their superior stability and dopability. Gofer et al.85 reported an all-polymer charge storage device based on poly-3(3,4,5-trifluorophenyl) thiophene [poly(3,4,5-TFPT)] and poly-



443

0.06 0.04

p-doping

I (A)

n-dedoping 0.02 0 –0.02 –0.04 –2.0

n-doping –1.5

–1.0

p-dedoping –0.5 E (V)

0

0.5

1.0

11.8 p and n doping–dedoping of PMT in 0.5 M TEABF4 in acetonitrile (Samui et al., unpublished result).

3(3,5-difluorophenyl) thiophene [poly(3,5-DFPT)] and a GPE. Derivatization of the polythiophene backbone was done by electron-withdrawing fluoro groups as this was expected to stabilize the excess negative charge in the n-doped state and increase the doping level and stability of the conducting polymer. The polymer film was electrochemically deposited from the monomer/electrolyte solution. A representative p- and n-doping of poly(methyl thiophene) (PMT) in 0.5 M TEABF4 in acetonitrile is shown in Fig. 11.8. The gel polymer electrolyte consisted of polyacrylonitrile (PAN) with 0.25 M TEABF4 in PC. The thickness maintained for the GPE layer was in the range of 10–100 µm. The specific charge capacity observed was in the range of 9.5–11.5 mA h/g, which was about 70% of the available capacity. The value is much lower than for lithium-based systems. The energy density was in the range 22.8 to 27.6 mW h/g. The device with the GPE showed a cycling efficiency of 99.1% up to 150 cycles. The discharge curves exhibited an extended plateau in the voltage range. A supercapacitor was made by using poly(dithieno[3,4-b:3″,4″-d] thiophene) (pDTT), which was electrosynthesized on carbon paper.86 The polymer pDTT was used for the electrodes of a symmetric supercapacitor as it can undergo both p- and n-doping. The supercapacitor was evaluated in both liquid and solid electrolytes. Two types of polymer electrolytes were used in the fabrication of the supercapacitor: PEO/PC/TEABF4, a gel polymer electrolyte, and crosslinked PEG/PC/TEABF4, a hybrid electrolyte. A potential window that was wider than 3.5 V was observed in all cases. Charge–discharge cycling was carried out up to 1000 cycles. The coulombic efficiency was almost 100% with no appreciable deterioration in capacitance for the capacitor with the gel polymer electrolyte. Moreover,


444


self-discharge was found to be much lower: 25% of the initial capacitance was lost after 14 h. The supercapacitor with the poly 3-(4-fluorophenyl thiophene) (PFT) and gel polymer electrolyte was demonstrated by Kim et al.87 PFT is both p- and n-dopable and was used as the positive and negative electrode respectively. The GPE contained poly(vinylidene fluride-cohexafluoropropylene) p(VDF-co-HFP) as the base polymer, PC–EC as the plasticizer and TEABF4 as the salt. Charge–discharge cycles were carried out between 0 and 2.75 V. The coulombic efficiency of the cell increased as the number of charge–discharge cycles increased. It was reported to exhibit about 90% of its value after 10 000 cycles. The energy density (ED) of the solid-state supercapacitor was 1.64 W h/kg and its power density (PD) was 268.36 W/kg at the 1000th cycle. The ED decreased drastically during the first 2000 cycles and then displayed a stable value up to 10 000 cycles, while the PD increased during the first 1000 cycles and stabilized thereafter. The decrease in ED was attributed to the decay of active sites. In other words, some of the active sites in the PFT might have lost their activity because of charge trapping. The reasons attributed to increase in the PD up to 1000 cycles were as follows: (1) a reduction in the electrode/electrolyte interfacial resistance due to the formation of a feasible path for easy ion penetration, (2) a decrease in the electrode resistance for ionic diffusion, and (3) degradation of the polymer due to repeated cycling. Polyaniline (PANI) and polypyrrole (PPY) electrodes were also studied by using a gel polymer electrolyte in a supercapacitor. Both are stable conducting polymers, mostly used with aqueous gel, although there are some reports of their use with non-aqueous gel. Hashmi et al. reported on conducting polymer-based supercapacitors using proton and lithium ionconducting GPEs,88 PVA–H3PO4 and PEO–LiCF3SO3. Indium tin oxide (ITO) glass and carbon papers were used as the current collectors. The conducting polymers were electrochemically deposited on the current collectors. The polymer gel electrolyte was sandwiched between the two symmetric electrodes. The impedance behavior of the capacitor cell with the PPY proton-conducting electrolyte and ITO current collector was close to the ideal. The single electrode specific capacitance was 84 F/g of PPY. It was suggested that the cell with PPY coated on carbon paper and a protonconducting GPE was suitable for obtaining a long charge–discharge life. In another report by the same group, PEO–LiCF3SO3 plasticized with PEG was used as a gel electrolyte with PPY and a polythiophene (PTH) electrode.89 A capacitance of 18 F/g was reported for the PPY/PTH asymmetric capacitor, and it could be charged up to 1.7 V. For the symmetric capacitor, the working voltage was limited to just 1.0 V. Both PPY and PANI were used as electrodes together with a PMMA/ PC/EC/LiClO4 gel electrolyte.90 The open-circuit voltage of this system was 1.2 V in its fully charged state. Initially, an excess doping charge in the nega-



445

tive PPY electrode was maintained. This ensured some residual doping, with acceptable conductivity at the end of discharge. The excess of PPY with respect to PANI resulted in the decay of specific energy, but an increase in stability and cyclability was observed. The specific capacitance obtained was up to about 25 F/g and the cycle life was quite acceptable, providing up to 60% of full capacitance over several thousand charge–discharge cycles. An all-polymer redox supercapacitor was fabricated with LiClO4 and LiCF3SO3doped PPY electrodes, which were either exposed to 160 MeV Ni12+ ion irradiation or kept unirradiated.91 The electrolyte used was PVDF–HFP (20 wt%)–PMMA (10 wt%)–LiCF3SO3H (10 wt%)–PC + diethylene carbonate (DEC) (57 wt%)–SiO2 (3 wt%) polymer gel. The irradiation method was adopted as it had been previously reported that high energy irradiation (>80 MeV) would enhance the electrochemical stability of PPY and PANI electrodes.92–94 The stability of the supercapacitor was studied for 10 000 cycles (Fig. 11.9). The initial decrease in the capacitance was due to an irreversible faradic reaction on the electrode surface from the presence of volatile surface groups such as –OH, –CN, –CH3, leading to a loss of charge. For irradiated samples, this occurred to a lesser extent. During irradiation, the volatile groups were either lost or crosslinked, resulting in electrochemical stabilization. A decrease in the total charge–discharge time of the supercapacitor and a slight increase in coulombic efficiency were observed for irradiated electrodes. Prasad and Munichandraiah studied a PANI-based supercapacitor with a GPE95 prepared from PAN, LiClO4 and PC. The mass ratio of PAN : PC : LiClO4 was 1 : 3 : 0.3. PANI electrodes were made by electrochemical

210 200

Capacitance (F/g)

190 180 170 160 150

Unirradiated 5 × 1010 ions cm–2

140

5 × 1011 ions cm–2 5 × 1012 ions cm–2

130 0

2000

4000 6000 8000 Cycle number

10000

11.9 Stability plot of PPY/P9VDF–HFP–PMMA–LiClO4 supercapacitor.91


446


deposition of PANI on stainless steel to a thickness of 0.15 mm. Symmetrical capacitors were made by sandwiching PANI electrodes (soaked with LIClO4 dissolved in PC) and GPE films in a configuration of PANI/GPE/PANI/ GPE/PANI. The central electrode was used as the positive and the two end electrodes as the negative. The GPE film has a stable voltage window of 4.5 V. In this study, charge–discharge cycling was conducted between 0 and 1.0 V to avoid PANI degradation. The specific conductivity of the GPE was 8 × 10−3 S/cm. The specific capacitance of the PANI was 670 F/g of the PANI in the unit cell. However, the specific capacitance depended upon the charge–discharge current density. At a specific power of 7.5 kW/kg, the cell showed specific energy of 70 W h/kg and specific capacity of 250 F/g. The decrease in specific energy with the increase in specific power was attributed to the decrease in the doping–dedoping process of PANI. The discharge capacitance was found to be stable up to 50 cycles, and after 1000 cycles the observed value was analyzed with the conclusion that the decay of capacitance was below 0.01% per charge–discharge cycle. A supercapacitor using crosslinked PVA/PAA blends containing acidic, alkaline and neutral media was also studied.96 An HClO4, NaOH and NaClbased hydrogel was used in the supercapacitor. The electrochemical characterization of the acidic, alkaline and neutral PVA/PAA blends and their performance in the capacitor were reported in this study. Carbon-supported RuO4 was used as the electrode material in the supercapacitor. The conductivity of the PVA/PAA blend was reported to be dependent upon the PAA content. The conductivity of the acidic gel electrolyte decreased with an increase in PAA while the conductivity of the alkaline and neutral gel electrolyte increased with an increase in PAA content. This result was attributed to the combined effect of an increase in the flexibility of the PVA/ PAA blend and a shift in the ionic conduction mechanism of the acidic medium in contrast to the alkaline and neutral media. Ionic conduction in the aqueous acidic liquids occurred from a Grotthus-type mechanism whereas a segmental motion mechanism accounted for ionic conduction in the non-acidic aqueous medium. The capacitance of the supercapacitor with acidic, alkaline and neutral forms of the GPE depended on the PAA concentration in the PVA/PAA composition. Capacitance of the carbon electrodes with a PVA/PAA ratio 5 : 0 showed a maximum of 60 F/g. For alkaline and neutral hydrogels, the maximum capacitance observed was 10 and 2 F/g respectively. The capacitance for RuO4 · H2O was 1000 F/g for acidic PVA/ PAA blends. This capacitance value decreased with an increase in PAA content in the blend. The pseudocapacitors for all the three conducting polymers exhibited a very good performance and capacitance. In one study, a satisfactory performance was obtained when 10 000 charge–discharge cycles were applied. This shows promise, but more studies are required for it to be commercially



447

exploited. Ruthenium oxide was not used much, although the capacitance reached 1000 F/g in the above report. This may be due to the high cost of the raw material.

11.6

Activated carbon electrodes

1000

200

800

150

600

100

400

50

200 0

100

200 300 400 500 Thickness of film (μm)

Resistance (Ω cm2)

Capacitance (μF cm–2)

The electric double-layer capacitor (EDLC) was studied, mainly because the rugged nature of the capacitor provides high cyclability and excellent stability of the electrodes. The nature of the carbon was also varied to achieve maximum capacitance. An all-solid gel electrolyte comprising PEO/ PC/LiClO4 along with an isotropic high-density graphite (HDG) electrode as EDLC was studied by Liu and Osaka.97 The thickness of the gel electrolyte film was in the range of 50 to 500 µm. At [EO]/[Li+] ratio of 8 : 1 in PEO/LiClO4, the capacitance of the HDG electrode at room temperature was the maximum among the various ratios studied. In fact, it was observed that at this [EO]/[Li+] ratio, PEO remains in a completely amorphous phase. Therefore, in the gel electrolyte, the ratio used was PEO/PC/LiClO4 8 : 8 : 1. The study on the effect of the thickness is shown in Fig. 11.10. It could be seen that the resistance increased with increasing thickness whereas the capacitance remained almost constant. As the thickness of the gel was much greater than that of the electrical double layer, the surface and adjoining area remained unaffected by the thickness variation and the capacitance remained unaltered. A larger capacitance was observed when it was charged to a higher potential. The capacitance and coulombic efficiency of the EDLC were found to be stable up to 1000 cycles.

0

11.10 Capacitance (䊉) and resistance (䊏) variation with thickness for all solid gel electrolyte comprising PEO/PC/LiClO4 along with isotropic high density graphite (HDG) electrode as EDLC.97


448

Polymer electrolytes I: mA cm–2 Current density (mA cm–2)

PC+EC / LiCIO4

PEO gel

PAN gel

PMMA gel

–4

–2

0 Potential (V)

2

4

11.11 Cyclic voltammograms of various gel electrolytes.98

Liu and Osaka studied another solid polymer electrolyte based on PEO/ PAN/PMMA and LiClO4 along with a mixture of EC and PC.98 HDG was used as the EDLC electrode. The ionic conductivity increased in the order of PMMA gel < PEO gel < PAN-based gel < LiClO4/PC + EC. The conductivity of the PAN-based gel was closer to an organic electrolyte due to minimum interaction between the PAN monomer and the organic solvent. The PEO-based gel had a strong interaction between the EO and Li+ ions and hence showed lower ionic conductivity than the PAN-based one. The PMMA-based gel showed less conductivity owing to the side chain of the PMMA molecule impeding ion transport. Cyclic voltammograms of various gel electrolytes are shown in Fig. 11.11. A good capacitor performance was observed up to 2.5 V for the PAN-based gel electrolyte and 3.5 V for the other electrolytes. The narrow stable potential range of the PAN-based electrolyte was due to the electrochemical instability of the CN-functional group of the PAN at the HDG electrode. The capacitance of the HDG was the lowest for the PEO-based gel electrode. In PEO-based gels, the ions are bound to the polymer chain which results in a lower permeability, whereas the ions in the PMMA and PAN are not bound to the polymer chains. Hence, it was reported that the capacitance with the PEO-based gel electrolyte was the lowest (9.9 mF/cm2) as compared to 15.8 and 14.7 mF/cm2 for PMMA and PAN respectively. The internal resistances (IR) of the capacitors, calculated from the IR drop of the discharge curves, were different (higher) from those calculated using the AC impedance responses. The differences were mainly due to the effect of ion diffusion. Figure 11.12



Intensity (a.u.)

(002)

1000 cps

449

(a) (b) (c) 25

30

35

40

2θ (deg)

11.12 XRD spectra of HDG electrodes (a) before charging, (b) after charging (anode), (c) after charging (cathode).98

shows the XRD patterns of the HDG electrode before and after charging. Since the patterns were identical, it was concluded that there was no intercalation of Li+ or ClO4− ions, indicating the absence of a faradic process of ion intercalation into the HDG. Furthermore, the capacitance values from charge–discharge measurements were different from AC impedance responses owing to an ion diffusion effect. A study on cycling proved that PMMA gel containing EDLC had greater stable capacitance and coulombic efficiency than a PAN electrolyte for more than 104 cycles at 2.7–2.2 V. Another PAN-based supercapacitor was studied by Ishikawa et al.99 PAN, PC and different tetra-alkyl ammonium salts were used in the polymer gel electrolyte. The gel was prepared by mixing PAN with a PC solution of tetra alkyl ammonium salt at 110 °C. The conductivity increased in the order of tetrabutyl ammonium perchlorate (TBAClO4) < tetraethyl ammonium perchlorate (TEAClO4) < TEABF4. The difference was attributed to a difference in ionic size. The capacitor with TEABF4 salt in the gel electrolyte showed a capacitance of 0.4 F/cm2 with charge–discharge coulombic efficiency of 100%. In the study it was reported that the GPE with TEABF4 salt showed a higher discharge capacitance and higher coulombic efficiency than the other alkyl ammonium salts studied. The self-discharge behavior of the PAN-based system was reported to be superior to the PMMA-based system. The above reports indicate that there are not many studies with EDLC using a non-aqueous solid electrolyte. This might be because the macropores of carbon give easy access to the solid electrolyte. Relatively more studies were concentrated on aqueous-based solid electrolytes. Matsuda


450


et al. used a GPE containing a composite of phosphoric acid-doped silica gel and styrene-ethylene–butylene–styrene (SEBS) elastomer in EDLC.100 The composite was used as an electrolyte but when mixed with activated carbon acted as an electrode. The electrodes and electrolyte were pressed together into a three-layered pellet of 13 mm in diameter and 1–2 mm in thickness. The evaluation of the supercapacitor was done in both air and a dry argon (Ar) atmosphere. The resistance from impedance plot of the capacitor was around 3.2 × 104 ohm in the Ar atmosphere, which amounted to about three orders of magnitude higher than that in ambient air (38 ohm). The capacitance value in air was 10 times larger than that in the dry Ar atmosphere. This effect was attributed to a decrease in the interfacial resistance between the electrode and electrolyte and to an increase in the ionic conductivity of the electrolyte due to the adsorption of water in ambient air. For the same reason, the discharge properties were also affected: the discharge current of the capacitor was two orders of magnitude higher in air than in the Ar atmosphere. The capacitance of this system was smaller than that of the PVA containing silica gel doped with HClO4.101 However, the thermal stability was slightly better. As silica gel is known as a thermally stable material and can provide excellent thermal conductivity with the incorporation of salts, more studies were concentrated on silica compounds. Tatsumisago et al. reported on silica-gel films containing perchloric acid HClO4, dodecatungstophosphoric acid and tetra-n-butylammonium perchlorate prepared by the sol–gel method using tetraethoxysilane.102 The films exhibited maximum conductivities of 10−2–10−1 S/cm at room temperature (RT) with 19 wt% of HClO4. A clear correlation was observed between σ RT and ν(OH), indicating that the stronger the O—H bonds in silanol groups in the silica gels, the easier the proton conduction. A conduction pathway of protons might have been formed in the doped silica gel structure, indicating that the interaction between the conduction paths and the silanol groups might have affected the conductivity. In another report, solid electrolytes based on silicate–PEG composites for use in supercapacitors were studied by Mitra et al.103 The solid electrolyte composition was based on tetraethoxysilane (TEOS), PEG, lithium and magnesium salts. The solid electrolyte had the composition [X]n[Y] where X was the molar ratio of PEG/TEOS, Y was the molar ratio of TEOS/Mn+ and n the chain length of PEG. A series of solid electrolytes was prepared and supercapacitors fabricated using HDG electrodes. The capacitance of the supercapacitor containing magnesium triflate at 80 °C was found to be 0.62–0.42 mF/cm2 at 100 mHz. The coulombic efficiency of the electrolytes was in the range of 97–80%. A cycle life of 1000 at 80 °C was obtained for all the current densities applied in the study. It was reported that the capacitors had a very low response time compared to that of commercially available supercapacitors.



451

There was some interest in making a gel using acrylic acid and its salts. Iwakura et al. reported on an EDLC capacitor using a cross-linked potassium poly(acrylate) hydrogel.104 The capacitance with a polymer electrolyte was 150 F/g, which was a little higher than with the aqueous KOH electrolyte (144 F/g). The effect was attributed to pseudocapacitance owing to the use of the hydrogel in the capacitor cell. Moreover, the capacitance retention after 20 000 cycles was found to be 82% of the initial capacitance for the polymer-based electrolyte while it was only 72% for the aqueous KOH electrolyte. A solid-state supercapacitor with alkaline polymer electrolyte was reported for the first time by Lewandowski et al.105 The polymer electrolyte was based on PEO–KOH–H2O. It served as both separator and binder for the electrode material. The selection of PVA was based on the fact that polymers like PVA, PEO and co-poly(epichlorohydrin-ethylene oxide) showed conductivity in the range of 10−3 S/cm.106 The separator was made by film casting from an aqueous solution of PEO and KOH in which the concentration of KOH was fixed at 40 wt% in the PEO so that the optimum mechanical strength of the film and conductivity was maintained. To make the electrodes, the PEO/KOH solution was mixed with activated carbon and acetylene black. A capacitance of 90 F/g for the activated carbon was reported and the capacitance values were comparable to that of the KOH electrolyte. The absence of any peak indicated pure electrostatic attraction in the capacitor behavior. When the system was studied at sweep rates higher than 20 mV/s, an appreciable decrease in capacitive behavior was observed due to ohmic drop. From the impedance measurement, the resistance was found to be around 0.77 ohm. The approximately calculated RC time constant was around 1.4 s.

11.7

Cation exchange membrane-based supercapacitors

11.7.1 Pseudocapacitor electrodes Cation exchange resins are commonly used for polymer electrolyte membranes and DMFC applications as a separator. Usually, the sulfonic acid group qualifies it for use as a proton conductor and it was observed that it can also be useful as a separator and electrolyte in a supercapacitor. Park et al. used an all-solid supercapacitor with an electrode containing RuO2 and a Nafion polymer as both separator membrane and electrolyte.107 The paste of RuO2 and Nafion was applied on carbon paper with a brush and hot-pressed with the Nafion membrane. The cyclic voltammetry was done in the voltage window of 0–1 V. The specific capacitance was studied with varying concentrations of Nafion (50, 33, 15 and 7 wt%) in the


452


electrode. Even though the specific capacitance was high, 200–230 F/g, the capacitor with 50 wt% Nafion showed degradation and instability during repeated cycling and hence it was reported that 33 wt% of Nafion was the optimum concentration with respect to capacitance and stability. After prolonged cycling of 10 000 cycles, only 30% reduction in the capacitance value of the supercapacitor was observed. The power delivered by the supercapacitor, on charging at a constant current of 12.5 mA/cm2, was 12 mW/cm2. It was further confirmed that when used in hybrid mode with a DMFC, the power density discharged by the supercapacitor was effectively transferred to the DMFC. When the DMFC was connected to the supercapacitor, an instantaneous high power density was observed and an enhanced power density persisted after complete discharge. Furthermore, the cell power was improved by 30% after hybridization with the supercapacitor. Samui and co-workers studied a pseudocapacitor using SPEEK as the solid polymer electrolyte and PANI as the electrode.108 A free-standing film of SPEEK was used as the separator, and the SPEEK, along with PANI, conducting carbon and PTFE, was rolled into a sheet of electrode. The residual moisture present in the SPEEK was enough to maintain the electrolyte character of the sulfonic groups. The cyclic voltammetry (CV) shape was near-rectangular in nature. Typical charging–discharging curves were obtained and the capacitance calculated was around 28 F/g, a value similar to that obtained by CV study. The value of R was around 1.67 ohm and the time constant (RC) calculated was 1 s. The same group continued the search for other sulfonated polymers for use as more mobile electrolytes.109 Poly(vinyl sulfonic acid) was used as the proton-conducting polymer electrolyte in the electrode while a fluorinated ethylene propylene copolymer grafted with acrylic acid and sulfonated (FEP-g–AA-SO3H) fluorinated ethylene propylene-graft–acrylic acid-sulfonated was used as the separator. A specific capacitance of 98 F/g was reported for PANI. The value was much higher than that obtained with SPEEK. This was due to the change in the sulfonic acid group concentration and the more mobile backbone of the polymer. About 20% reduction in the capacitance value was observed after 1500 charge–discharge cycles. A proton-conducting composite electrolyte based on phosphotungstic acid (PTA) and Al2(SO4)3 · 18H2O was reported by Wang and Zhang.110 PANI was used as the electrode material. It was prepared by mixing 85% PANI, 14% acetylene black and 1% poly(tetrafluoroethylene) (PTFE) suspension and compressing this mixture to a pellet. PTA and Al2(SO4)3 · 18H2O in various weight ratios were mixed by mechanical grinding and compressed to a composite electrolyte pellet. The capacitance of the unit cell was measured as a function of the PTA/Al2(SO4)3 · 18H2O ratio in the solid electrolyte. It was found that the capacitance at PTA :Al2(SO4)3 · 18H2O wt ratio 1 : 1 was the highest and the value attained was 271.6 F/g. The reason behind



453

the highest capacitance value, at a 1 : 1 mole ratio of PTA/(Al2SO4)3 · 18H2O, was its highest conductivity. No clear redox peak was observed in the voltage range of −0.2 to 0.85 V. Furthermore, the shape of the CV curve did not change with an increase in scan rate, indicating the stability of this kind of electrolyte. The capacitance decay after 1000 cycles with the electrolyte composition above was found to be less than 10% of the initial value. However, the initial coulombic efficiency of 96.4% reached 99% after the 1000th cycle, which indicated an increase of a reversible nature of charging and discharging. A few nanocomposite electrodes were also developed. Gomez-Romero et al. reported for the first time on the use of a hybrid nanocomposite electrode formed by PANI and polyoxometalate (POM) (phosphomolybdates) for supercapacitors.111 The hybrid electrodes were prepared on carbon paper using both chemical–electrochemical and electrochemical methods. The separator was prepared from PBI and phosphoric acid (2.8–3.0 molecule/ repetitive unit of polymer). The voltammetric current observed with the directly formed composite electrode was very high compared with others. From charging–discharging cycles, the capacitance was calculated to be around 195 mF/cm2. The results could be improved by optimization of the thickness and morphology of the active materials. A further study conducted by the same group made nanocomposite electrodes formed from PANI and POM such as phosphomolybdic acid (H3PMo12O40) (PMo12), silicotungstic acid (H4SiW12O40) (SW12), and phosphotungstic acid (H3PW12O40) (PW12) for solid state supercapacitors.112 Nafion 117 was used as the separator and electrolyte. Symmetric supercapacitors based on all three nanocomposite materials were studied separately and compared. The PANI/SW12 supercapacitor exhibited a low capacitance of 0.7 to 1.8 mF/cm2. For the PANI/ PW12-based supercapacitor, the value was 3 to 6.5 mF/cm2 and for the PANI/PMo12 supercapacitor it was 140 to 90 mF/cm2. Furthermore, it was observed that the hybrid materials showed a higher capacitance than the electrode material containing only POM, which highlighted the combined effect of PANI and POM. In this study, the authors examined the performance of used electrodes and new electrodes made up of PANI/PMo12 in a supercapacitor assembly. The initial capacitance of the cell with the used electrodes was found initially to be at 75 F/g and later stabilized to 50–60 F/g after repeated cycling of charge–discharge. However, for the cell with the new electrodes, the initial value was 40 F/g and reached a value around 50–60 F/g after repeated cycling. The authors attributed these effects to the electrochemical activation of the electrodes. The electrochemical activation of the hybrid material was due to kinetic aspects and not attributed to the inorganic component of the hybrid material. Electroactivity was attributed to the swelling of the conducting polymer. It was also found that the thinner the electrochemically deposited film, the easier the electroactivation.


454


11.7.2 Electrical double layer capacitor electrodes Ion-exchange membranes were also tried with an EDLC as it had been observed that a gel electrolyte will work better with an EDLC in terms of performance, stability and environmental impact. Staiti and Lufrano demonstrated an efficient EDLC with an electrolyte membrane and a carbonbased electrode prepared with a Nafion ionomer.113 The resistance of the membrane electrode assembly prepared by the casting method was as low as about twice that of Nafion 115, mainly due to the former being thinner. The proton conductivities of the membranes were 5.7 × 10−2 and 3.1 × 10−2 S/cm for Nafion 115 and the cast membrane respectively. A capacitance of 13.2 F/g was observed for the assembly, which was higher than that of Nafion 115 (9.6 F/g). An extension of the earlier study was done by Lufrano el al. with a supercapacitor based on activated carbon and a Nafion ionomer.114 The electrolytes studied were Nafion 115, a recast Nafion membrane (NRG50) and a porous glass-fiber matrix impregnated with 1 M H2SO4 with thicknesses of 160, 50 and 200 µm respectively in a swelled condition. Nafion-based supercapacitors were fabricated by hot-pressing with electrodes made of carbon black (1546 m2/g BET), graphite fiber and Nafion 1100. The internal resistances evaluated at 1 kHz were 91 ohm for N115, 45 ohm for NRG 50 and 27 ohm for the porous glass-fiber matrix (H2SO4). The supercapacitors fabricated with recast Nafion membrane and Nafion 115 were compared by using in a sulphuric acid capacitor. It was found that the cast Nafion-based supercapacitor produced a higher performance than Nafion 115. A specific capacitance of 13 F/g was observed for the recast Nafion, a value very close to the supercapacitor containing sulfuric acid. It was found that the sulfuric acid-based supercapacitor exhibited a larger dependence on the frequency of the real impedance than the Nafion-based capacitors. This result was attributed to the homogeneous inter-distribution of carbon/Nafion in the electrodes and to the excellent adhesion between the electrode material and the Nafion membrane due to hot pressing and also to the fast proton transport in the smaller carbon pores. Impedance analysis has shown that the large microporosity of carbon could produce an additional proton and/or electronic resistance in the electrode of the capacitors. As a consequence, the specific capacitance did not reach a plateau even at a very low frequency. A difference of about 20% in capacitance for all the supercapacitors was observed between the impedance and DC charge measurements. This was explained in terms of the low electrochemical accessibility of the pores. Later, the same group reported on the effect of polymer electrolyte loading in the electrodes on the performance of the supercapacitor.115 It reported that the electrodes of the supercapacitor containing 10–30 wt% of Nafion exhibited a higher spe-


Solid polymer electrolytes for supercapacitors 120

120 Coulombic efficiency

100

110 105

80

Specific capacitance

100

60

95

40

90

Coulombic efficiency (%)

Specific capacitance (F/g)

115

20

85 80

455

0

2000

4000

6000

8000

0 10 000

No. of cycles

11.13 Specific capacitance and coulombic efficiency of AC-crosslinked SPEEK supercapacitor as a function of charge–discharge cycles. Charging and discharging were carried out at current density of 5 mA/cm2 (Samui et al., unpublished result).

cific capacitance of 110 F/g compared with that of 50 wt% Nafion (90 F/g). This result was attributed to the high resistance caused by the low accessibility of the electrolyte into the carbon pores. The fabrication of a 1.5 F and 5 V supercapacitor was also reported. One study was carried out by Samui et al. using an AC electrode and a crosslinked SPEEK electrolyte (unpublished). The capacitance was around 107 F/g and the capacitance reduction after 10 000 cycles was only around 17%. The coulombic efficiency was around 98% throughout the cycling (Fig. 11.13). Further study is in progress.

11.7.3 Hybrid supercapacitors In an effort to enhance the energy density along with the power density, attempts are being made to use hybrid capacitors, mostly by using an EDLC electrode at one end and a pseudocapacitor at other. Only a few reports of this are available in the literature. Samui et al. have studied a hybrid supercapacitor with a nanocomposite GPE.116 The active electrode of the supercapacitor was made from poly(3-methylthiophene) and the negative electrode was activated carbon. The nanocomposite gel electrolyte was based on PEO–TEABF4–PC–EC–nanoclay. Organically modified nanoclay was used in this study. The maximum specific capacitance reported for the


456


Cell voltage (V)

3.0

2.5

2.0

1.5

1.0

0.5

(f) (e)(d) (c) (b)(a)

0

250

500

750

1000

Time (s)

11.14 Charge–discharge plot of PMT-AC hybrid solid electrolyte supercapacitor at different cycles: (a) 1st, (b) 100th, (c) 200th, (d) 400th, (e) 700th and (f) 1000th cycle.116

hybrid supercapacitor was 18.54 F/g. The charge–discharge cycle was performed for 1000 cycles at 10 mA between 1 and 2.75 V. There was a decrease in capacitance during the initial 500 cycles and a negligible decrease was observed during the next 500 cycles (Fig. 11.14). An IR drop occurred during the charge–discharge cycles, which was due to degradation of the TEABF4 salt. The coulombic efficiency of the charge/discharge was 94% in the first cycle and increased to 98% after the 1000th cycle. The charge transfer resistance increased with progressive cycling due to degradation of the salt. This was reflected as an increase in the internal resistance of the capacitor, which in turn resulted in capacitance decay during the charge– discharge cycles.

11.8

Current research activities

11.8.1 Carbon nanotube-based composite electrode for all-solid supercapacitors A supercapacitor with properties of optical transparency and mechanical flexibility was reported by Chen et al.117 It was fabricated using metal oxide nanowire/carbon nanotube (CNT) heterogeneous film. It could achieve a power density of 7.48 kW/kg and after a large number of cycles, its capacity



457

was observed to be 88% of the initial value. This development could lead to transparent and flexible nanoelectronics in the future. Hybrid In2O3 nanowires/CNT films were prepared in two steps. In the first step, a CNT suspension was filtered through a porous alumina filtration membrane, forming a homogeneous entangled network of CNT. An adhesive, flat poly(dimethysiloxane) stamp was used to peel the CNT film off the filtration membrane and then release it onto a polyethylene terephthalate (PET) substrate on a hot plate at 100 °C. The thickness of the film was about 60 nm and the measured conductivity was 570 S/cm. In2O3 nanowires with a diameter of 20 nm and length 5 µm were sonicated into isopropanol solutions and then dispersed on transferred CNT films to form In2O3 nanowire/CNT heterogeneous film. The calculated specific capacitance of a transferred CNT film supercapacitor was about 25.4 F/g and that of a poly(3,4-ethylenedioxythiophene) (PEDOT):CNT film supercapacitor was 33 F/g. The In2O3 nanowire/CNT heterogeneous film supercapacitor exhibited an even higher specific capacitance (up to 64 F/g) than either the transferred CNT film or the PEDOT:CNT film supercapacitor. This was explained by pseudocapacitance from the reversible redox transitions of the In2O3 nanowires contributing to the overall capacitance. The stability study of the In2O3 nanowire/CNT heterogeneous film supercapacitor revealed that after the first 100 cycles, there was a small decrease in specific capacitance from 64 to 53 F/g. After 100 cycles, the specific capacitance value remained unchanged up to 500 cycles. The specific capacitance fading of the In2O3 nanowire/CNT heterogeneous film supercapacitor could be due to dissociation of the In2O3 nanowire during the redox process, as observed with other metal oxide nanostructured materials.118,119 The authors were of the opinion that more studies needed to be done to clarify the mechanism. A study on a supercapacitor consisting of polymer-dispersed multi-walled nanotubes (MWNTs) and metal oxide-dispersed MWNT composites as electrodes, and a Nafion membrane as the electrolyte as well as the separator was reported by Amitha et al.120 TiO2 was coated over MWNTs using the sol–gel method by dispersing functionalized MWNTs in dilute nitric acid (pH 0.5), ultrasonicating the dispersion and adding titanium tetraisopropoxide in drops (volume ratio of titanium tetraisopropoxide to water 1 : 4). The sol obtained was stirred for 2 days in air at room temperature. The TiO2/MWNT composites were heattreated at 350 °C for 2 h in air. PPY/MWNT composites were prepared by immersing the MWNTs into an aqueous solution of pyrrole and adding an oxidant to this solution. The PANI/MWNT nanocomposite was synthesized by polycondensation of 0.4 mL aniline by 0.4 g K2Cr2O7 in 50 mL of 1 mol/L HCl.


458


To fabricate the electrodes, the required quantity of electrode material was suspended in deionized water and ultrasonicated by adding 5 wt% Nafion solution. This suspension was spread uniformly over gas diffusionlayered carbon paper (Toray) by spin-coating. The average specific capacitances measured using the three electrochemical techniques for the symmetric arrangement of the pure MWNT, PPY/ MWNT, PANI/MWNT and TiO2/MWNT nanocomposite electrodes were 80, 116, 175 and 210 F/g respectively. The asymmetric assembly using PANI/ MWNTs and TiO2/MWNTs as electrodes had a capacitance of 345 F/g with the Nafion solid electrolyte. The increase in capacitance of the MWNTs was attributed mainly to a homogeneous inter-distribution of carbon/Nafion in the electrodes, the excellent adhesion achieved between the electrodes and Nafion membrane, a good contact between the electrodes and current collectors, and fast proton transport in the smaller carbon pores. In addition, the increase in the capacitance for the polymer- and metal oxide-dispersed MWNTs was due to a homogeneous distribution of polymer and metal oxide particles over the functionalized MWNTs, which in turn modified the microstructure and morphology of the MWNT, allowing the polymer and metal oxides to be available for the electrochemical reactions and improving the efficiency of the composites. Thin-film supercapacitors were fabricated using printable materials to make flexible devices on plastic.121 The active electrodes were made from sprayed networks of single-walled carbon nanotubes (SWCNTs) serving as both electrodes and current collectors, a printable aqueous gel electrolyte and an organic liquid electrolyte. The ability of the SWCNT networks to carry a high current together with their substantial mechanical strength enables devices that are flexible as well as robust to be made, such as those that are required for mobile applications. Purified SWCNT material was suspended in water (1–2 mg/mL) with the aid of a tip sonicator. The stable suspension was sprayed onto PET substrates placed on a hot plate. During spraying, the water evaporated and the SWCNTs formed an entangled random network on the PET. The SWCNT-coated PET substrates were used as thin-film electrodes without any further treatment. The mass of the SWCNTs coated on each substrate was determined by the gravimetric method. All SWCNT films used typically had a sheet resistance of ~40–50 ohm, a transmittance of ~12%, and a thickness of ~0.6 µm. The gel electrolyte was prepared by using PVA (aqueous soln.: 10% w/v) and concentrated phosphoric acid (0.8 g). After evaporation of the excess water, the electrolyte solidified. The CNT networks and the gel electrolyte were sandwiched together to form the supercapacitor. For comparison, liquid electrolytes of 1 M solutions of H3PO4, H2SO4 and NaCl in water as well as 1 M LiPF6 in 1 : 1 (by weight) EC:DEC were used.



459

The difference between the internal resistances of aqueous and organic electrolytes reflects the significantly lower conductivity of LiPF6/EC:DEC (10−2 S/cm)122 compared with aqueous electrolytes (1–3 × 10−1 S/cm).123 There was no significant difference between the internal resistances of the PVA/H3PO4 and liquid H3PO4, indicating the minimum contribution of the polymer matrix to the internal resistance when aqueous electrolytes were used. For supercapacitors made with SWCNT, a specific capacitance of around 80 F/g was observed in liquid aqueous electrolytes. For supercapacitors made with thin sprayed films using the same SWCNT material, the specific capacitance was around 120 F/g. This indicated that wetting was better in thin CNT films, increasing the effective surface area. This was valid even in the case of a gel electrolyte. The capacitance was found to increase linearly for the liquid electrolyte, whereas saturation occurred for the gel electrolyte, which indicated a limited penetration into the gel network. From the study it was concluded that the optimum thickness of the CNT film should be around 2 µm for the gel electrolyte. The RC constants for the supercapacitor were around 0.5 s for aqueous electrolytes and 0.1 and 0.3 s for the organic electrolyte at 1 and 3 V respectively. The power density obtained in the organic electrolyte was about 70 kW/kg. The concept of printed power without a current collector offers a new platform for all kinds of lightweight devices.

11.8.2 Composite solid polymer electrolytes A PVA–sodium polyacrylate (PAAS)–KOH–H2O alkaline polymer electrolyte film with high ionic conductivity (0.1 S/cm at room temperature) was prepared by the solution-casting method.124 A nickel hydroxide positive electrode was prepared by mixing Ni(OH)2 with 5% cobalt powder, 2% acetylene black and PTFE emulsion. The resulting slurry was poured into foamed nickel with an apparent area of 2 cm × 2 cm, dried at 65 °C, and rolled to a sheet. The activated-carbon negative electrode was prepared similarly by mixing activated carbon (AC) and acetylene black in a mass ratio 85 : 15. The liquid Ni(OH)2/AC capacitor consisted of a Ni(OH)2 positive electrode and an AC negative electrode separated by a polypropylene separator, and a 6 mol/L KOH solution as the electrolyte. To activate the electrodes, the charge/discharge cycle was carried out three times at a rate of 0.1C for the Ni(OH)2 electrode. The activated electrodes were then separated and dried in air. The polymer Ni(OH)2/AC capacitors were made using PVA-based alkaline polymer electrolyte film sandwiched between the electrodes, then rolled to an appropriate thickness.


460


With an increase in mass ratio, the variation of specific capacitance with current density decreased. Despite all this, the specific capacitance of the capacitor with mass ratio m(Ni(OH)2) : m(AC) of 1 : 2 was higher than that of the other two capacitors at any current density. For the capacitor with a mass ratio m(Ni(OH)2) : m(AC) of 1 : 2, the discharge specific capacitance decreased by 13.3% after the 100th cycle. Between the 100th cycle and the 250th cycle, no reduction of specific capacitance could be observed. Thereafter, the specific capacitance quickly decreased again. For the capacitor with a mass ratio m(Ni(OH)2) : m(AC) of 1 : 1, the specific capacitance decreased quickly in the initial 100 cycles, i.e. decreased by 16%. Thereafter, a smooth variation of the specific capacitance could be observed. Up to the 1000th cycle, the specific capacitance decreased to 60 F/g. For the capacitor with a mass ratio m(Ni(OH)2) : m(AC) of 2 : 1, the specific capacitance decreased from 29.8 to 25.1 F/g in the initial 50 cycles, i.e. decreased by 16%. Thereafter, the capacitor showed an excellent cyclic stability. Up to the 1000th cycle, the specific capacitance remained at 25.1 F/g. The decrease in performance with cycling is due to degradation of the Ni(OH)2 as the AC electrode is known for its long cycle life whereas the cycle life of an Ni(OH)2 electrode is relatively poor. In the early stage of cycling, the negative electrode limited the capacitance of the capacitor, and in the final stages the positive electrode did the same. To summarize this, with an increase in mass ratio, the cyclic stability of the Ni(OH)2/AC capacitor increases. Chandra et al.125 reported on the ion transport property of a PEO–PVP blended SPE. It was concluded that polymer blending is one of the most important techniques for improving the room temperature conductivity, mobility, mobile ion concentration and mechanical stability of SPE membranes. Hybrid SPEs with a high ionic conductivity were studied by Munichandraiah et al.126 They used PEO, PAN, PC, EC and LiClO4 to make SPEs.

11.8.3 High temperature solid polymer electrolytes An all-solid state supercapacitor was made using phosphoric acid-doped PBI as the polymer electrolyte membrane, and hydrous RuO2/carbon composite electrodes (20 wt%) with surface area 250 m2/g.127 PBI was synthesized from diaminobenzidene (DAB) and isophthalic acid using polyphosphoric acid (PPA) as the solvent at 200 °C for 20 h as shown here:


Solid polymer electrolytes for supercapacitors H2N

NH2 +

H2N

HOOC

461

COOH

NH2

N N H

H N

+ 4H2O

N n

A PBI solution with an inherent viscosity of 1.2 dL/g in conc. H2SO4 was used to prepare the membranes through solution-casting in a vacuum oven at 80–90 °C. The film formed was subsequently peeled off and treated with water at 60 °C for one week to remove the residual solvent completely. The film was then dried at 100 °C under vacuum for two days and the doping was carried out by keeping the membranes immersed in 88% H3PO4 solution for 72 h and subsequently drying them at 100 °C for two days. The phosphoric acid uptake was about 13 moles per repetitive unit. The specific capacitance with the number of cycles was evaluated for the first 1000 cycles at 5 mV/s. The capacitance after 1000 cycles was 230 F/g as compared with the initial capacitance of 290 F/g. A power density of 300 W/ kg and energy density of 10 W h/kg were calculated when operating at 150 °C. However, the ESR of about 3.7 ohm appeared to be on the higher side and needs to be reduced further for high-rated applications.

11.8.4 Miscellaneous solid polymer electrolytes A biodegradable polymer electrolyte based on cellulose acetate (CA) and LiClO4 salt was investigated for use in supercapacitors.128 The ionic conductivity of the films increased with an increase in salt content to a maximum of 4.9 × 10−3 S/cm with 16% LiClO4. The biodegradation of the SPE films was studied using the methods of soil burial, degradation in activated sludge and degradation in a buffer medium. The extent of biodegradation was measured by AC impedance spectroscopy and weight loss calculations. The materials were observed to have sufficient biodegradability. The capacitor cells were constructed with an LiClO4 impregnated CA electrolyte separator sandwiched between two symmetrical PPY deposited electrodes and the electrochemical characteristics and performance were studied. A specific capacitance of 150 F/g was obtained for the single electrode at a sweep rate of 10 mV/s, and for the supercapacitor it was 90 F/g with a


462


time constant of 1 s. The coulombic efficiency of the supercapacitor calculated from charge–discharge cycling was also high, in the range of 98–99%. Cycling studies indicated a stable performance of 80 F/g up to 350 cycles, and thereafter it reduced to 75 F/g and remained constant up to 550 cycles. EDLCs were constructed from a polyurethane electrolyte.129 The EDLCs were fabricated using polyurethane gel electrolytes containing lithium perchlorate with carbon–cloth and carbon–powder composite electrodes. It was thought that the hard segments would impart good mechanical integrity while the soft segments would enhance ionic mobility. The Tg of the polyurethane used was −30 °C. The composition of the polyurethane gel electrolyte was PU : EC : PC : salt in the ratio 1 : 2 : 2 : 0.1 (w/w). The components were dissolved in 30 mL tetrahydrofuran and the solution was stirred for 24 h to complete homogenization, then it was cast in a mold to obtain a solid thin-film gel electrolyte. Similarly, Pearl Black 2000 carbon powder (0.10 g), 0.2 g of the above electrolyte solution and additional plasticizer/salt solution (additional plasticizer/salt solution was prepared from EC–PC (1 : 1; v/v) and lithium perchlorate in the ratio 10 : 1 (w/w)) in the range of 20–200 wt% with respect to the initial mass of the carbon composite electrode homogenized mixture, were cast to make films. To make the cell assembly, the gel electrolyte was sandwiched between high surface area carbon–cloth and carbon–composite electrodes respectively. The optimum electrolyte composition for the highest conductivity (~10−3 S/cm) was found to be that of the gel electrolytes containing 10 wt% lithium perchlorate with respect to the mass of the polymer. The specific capacitances at 10 mHz were up to 35 and 5.5 F/g respectively for the composite and carbon–cloth electrodes. The phase angle for the capacitors with carbon–composite electrodes was found to be closer to the ideal value of 90° than that of the capacitors using the carbon–cloth electrodes. This illustrated that a better electrode–electrolyte contact was established with the composite electrodes compared to that with the carbon cloth. The galvanostatic charge/discharge performance of the carbon– composite electrodes declined after 500 cycles, unlike that of the carbon– cloth electrodes. Passivation or other interfacial effects were attributed to the decline of the composite electrode. However, the carbon–composite electrode was observed to retain 80% of its original capacitance after 1000 cycles. Lee et al. reported on a solid-state redox supercapacitor made using acrylonitrile butadiene rubber (NBR)–KCI as the solid polymer electrolyte, and chemically deposited PPY as the conducting polymer electrodes on both surfaces of an NBR film.130 The optimal conditions for the preparation of the PPY/NBR electrode were confirmed to be functions of the uptake of the pyrrole monomer into the NBR matrix and the immersion time in



463

an oxidant solution. The performance of the capacitors was evaluated using the galvanostatic charge–discharge technique. An SPE composed of a three-dimensional crosslinked compound, using a diacrylic acid ester compound and/or a dimethacrylic acid ester compound, was reported for use in electrochemical devices.131 Nanocomposite SPE films composed of PEO, lithium triflate (LiCF3SO3) and SiO2 nanofiller (15 nm in size) were prepared by using the solution-cast method.132 At room temperature, PEO–LiCF3SO3 (88 : 12; w/w) showed a high ionic conductivity of about 1.28 × 10−3 S/cm.

11.9

Applications

Although the use of solid electrolyte supercapacitors is enormous in almost every type of energy application, some examples will be listed here. As a solid electrolyte is expected to remain maintenance free, integrated circuit memories, microcomputers, uninterruptible power supply (UPS) devices, toys, radar and torpedoes, sonar in military usage, telecommunications, cardiac pacemakers in medical applications, etc. are all target applications for using supercapacitors. Areas of application that require higher power include pulse lasers and welding, hybrid vehicles, fuel cell or battery chargers in cold climates, elevators, cranes, etc. They will remain the ideal product for temporary energy storage, for capturing and storing the energy from regenerative braking, and for providing a booster charge in response to sudden power demands. Supercapacitors can also be used to provide fast-acting, short-term power back-up for UPS applications. By integrating a capacitor with a battery-based UPS system, the life of the batteries can be extended. The batteries provide power only during longer interruptions. It can reduce the peak loads on the battery and permit the use of smaller batteries.

11.10 Conclusions The use of solid electrolytes in supercapacitors has now been extensively studied and the results have clearly shown the variety of fabrication methods, the effects of the electrolytes, the differences between the various types of electrolyte, their interfacial resistance, their efficiency, the critical areas of concern, and so on. TEABF4 salt was found to be used in many studies, which could be due to its high conductivity, non-corrosive nature and ability to be used up to reasonably high temperatures. A balance has to be achieved between the load resistance and the internal resistance to obtain the required output. The variety in gel electrolytes has revealed some interesting results, indicating further directions for improvement. Most of the solid electrolytes were studied using pseudocapacitors. A capacitance value of


464


1000 F/g has been reported when using ruthenium oxide as the electrode material with a solid electrolyte. During the discussion, it appeared that electrolytes are designated as either a solid electrolyte or a gel electrolyte. Copolymerization, grafting and blending techniques were used for preparation of electrolytes with the addition of salt and plasticizer. Cation exchange resins only need the presence of residual moisture to enable them to act as an electrolyte as well as a separator. Many attempts have been made to make a solid electrolyte using silica doped with inorganic acid or by reacting tetraethoxysilane with other active compounds. A few reports have also been included which discuss nanocomposites and their performance. The nanocomposites have been tried in both the electrolyte and electrode phases. Only one report on a solid electrolyte hybrid supercapacitor has been included as most of the reports available are about liquid electrolytes. I believe more work will be devoted to this type. Their ionic conductivity is mostly concentrated around 10−3 S/cm. In only one case was it observed to be around 10−2–10−1 S/cm. However, no improvement in properties, as compared to other types of supercapacitors, was observed. Thermal stability was not reported in most of the studies. However, charging and discharging at high speed are bound to increase the temperature of the system and it remains to be studied whether this can be sustained by the solid electrolyte without minimum degradation. The effect of interfacial resistance on this aspect remains to be studied in detail. Obtaining the highest energy and power density is the ultimate goal. To achieve this, it is necessary to increase the capacitance, reduce the ESR and increase the cell voltage. As the EDLC and metal oxide supercapacitors are rugged in nature and have a long life, efforts will be concentrated on these two materials, and the electrode surface, accessibility, and the particle size and its distribution need to be investigated. Conducting polymers are supposed to offer very high capacitance. Till now, the values observed have mostly not been very high, but more work may reveal ways to enhance their performance. There have been few reports on modeling a supercapacitor using a solid electrolyte and this situation needs to be addressed.

11.11 List of abbreviations AC AgBF4 ATRP CA CNT DAB DBP DBS

Activated carbon Silver tetrafluoro-borate Atom transfer radical polymerization Cellulose acetate Carbon nanotube Diaminobenzidene Dibutyl phthalate Dibutyl sebacate


Solid polymer electrolytes for supercapacitors DEC DFPT DMFC DSC EC ED EDLC EGDMA EO ESR FEP-g-AA-SO3H GPE HDG HFP IR ITO LiTFSI MAN MWNT NBR PAA PAAS PAN PANI PBI PC PCL PD pDTT PEG PEGMA PEGPG PEO PET PFS PFT PMMA POM PPA PPY PTA PTFE

465

Diethylene carbonate Difluorophenyl thiophene Direct methanol fuel cell Differential scanning calorimetry Ethylene carbonate Energy density Electric double-layer capacitor Ethylene glycol dimethacrylate Ethylene oxide Equivalent series resistance Fluorinated ethylene polypropylene-graft-acrylic acid-sulfonated Gel polymer electrolyte High-density graphite Hexafluoropropylene Internal resistance Indium tin oxide Lithium bis-(trifluoromethylsulfonyl)imide Methacrylonitrile Multi-walled nanotubes Acrylonitrile butadiene rubber Poly(acrylic acid) Sodium polyacrylate Poly(acrylonitrile) Polyaniline Polybenzimidazole Propylene carbonate Poly(ε-caprolactone) Power density Poly(dithieno[3,4-b:3″,4″-d] thiophene) Poly(ethylene glycol) Poly(ethylene glycol) methacrylate Poly(ethylene glycol-co-propylene glycol) Polyethylene oxide Polyethylene terephthalate Poly(pentafluorostyrene) Poly 3-(4-fluorophenyl thiophene) Poly(methylmethacrylate) Polyoxometalate Polyphosphoric acid Polypyrrole Phosphotungstic acid Poly(tetrafluoroethylene)


466


PTH PVA PVDF PVP RH RT SEBS SPE SPEEK SWCNT TBAClO4 TEABF4 TEAClO4 TEOS TFPT THF UPS VTF WLF XRD

Polythiophene Poly(vinyl alcohol) Poly(vinylidene fluoride) Poly(vinyl pyrrolidone) Relative humidity Room temperature Styrene-ethylene–butylene–styrene Solid polymer electrolyte Sulfonated poly(ether ether ketone) Single-walled carbon nanotubes Tetrabutyl ammonium perchlorate Tetraethyl ammonium tetrafluoroborate Tetraethyl ammonium perchlorate Tetraethoxysilane Trifluorophenyl thiophene Tetrahydrofuran Uninterruptible power supply Vogel–Tamman–Fulcher Williams–Landel–Ferry X-ray diffraction

11.12 References 1. t.m. doscher, g.c. myers and d.c. atkins Jr., J. Colloid Sci. 6 (1951) 223. 2. a.a. blumberg, s.s. pollack and c.a. hoeve, J. Polym. Sci. 2(A2) (1964) 2499. 3. m. yokoyama, h. ishihara, r. iwamoto and h. todokoro, Macromolecules 2, (1969) 184. 4. p.v. wright, Brit. Polym. J. 7 (1975) 319. 5. i. oslen, r. koksbang and e. shou, Electrochim. Acta 40 (1995) 1701. 6. g. petersen, p. jacobsson and l.m. torell, Electrochim. Acta 37 (1992) 1495. 7. m.b. armand, j.m. chabagno and m.j. duclot, 2nd Int. Conference on Solid Electrolytes, St Andrews (1978) paper 6.5. 8. m.b. armand, j.m. chabagno and m.j. duclot, in Fast Ion Transport In Solids (eds. P. Vashista, J.N. Mundy and G.K. Shenoy), Elsevier North-Holland, New York (1979), p. 131. 9. p. judeinstein, d. reichert, e.r. deazevedo and t.j. bonagamba, Acta Chim. Slov. 52 (2005) 349. 10. g. goulart, s. sylla, j.y. sanchez and m. armand, in 2nd International Symposium on Polymer Electrolytes (Ed. B. Scrosati), Elsevier Applied Science, London (1991), p. 99. 11. w. wieczorek, k. such, z. florjanczyk and j. przyluski, Electrochim. Acta 37(9) (1992) 1565. 12. f. cepuano, f. croce, s. panero and b. scrosati, J. Electrochem. Soc. 38 (8) (1991) 418 C.



467

13. r. dupon, b.l. papke, m.a. ratner and d.f. shriver, J. Electrochem. Soc. 131 (1984) 586. 14. a.f. eisenberg and h.l. yeager, in Perfluorinated Ionomeric Membranes (Ed. A.F. Eisenberg and H.L. Yeager), ACS Symposium Series 180, American Chemical Society, Washington, DC (1982). 15. s.r. samms, s. wasmus and r.f. savinell, J. Electrochem. Soc. 143 (1996) 1498. 16. j. jorissen, Electrochim. Acta 41(4) (1996) 553. 17. n. carretta, v. tricoli and f. picchioni, J. Membr. Sci. 166 (2000) 189. 18. k.d. kreuer, m. ise, a. fuchs and j. maier, J. Phys. IV 10(P/7) (2000) 279. 19. k.d. kreuer, J. Membr. Sci. 185 Sp. Iss SI (2001) 29. 20. b. yang and a. manthiram, Electrochem. Solid-state Lett. 6(11) (2003) A 229. 21. w. cui and j. kerres, Sep. Purif. Technol. 14 (1998) 145. 22. j. a. kerres, J. Membr. Sci. 185 (2001) 3. 23. c. manea and m. mulder, Desalination 147 (2002) 179. 24. s.l. chen, l. krishnan, s. srinivasan, j. benziger and a.b. bocarsly, J. Membr. Sci. 243 (2004) 327. 25. f. sevil and a. bozkurt, Turk. J. Chem. 29 (2005) 277. 26. x. jin, m.t. bishop, t.s. ellis and f.b. karasz, Br. Polym. J. 17 (1985) 4. 27. p. xing, g. robertson, m.d. guiver, s.d. mikhailenko, k. wang and s. kaliaguine, J. Membr. Sci. 229 (2004) 95. 28. y.l. ma, j.s. wainright, m.h. litt and r.f. savinell, J. Electrochem. Soc. 151 (2004) A8. 29. p. krishnan, j.s. park and c.s. kim, J. Membr. Sci. 279 (2006) 220. 30. s. licoccia, m.l.d. vona, a. d’epifanio, d. marani, m. vittadello, j.r.p. jayakody and s.g. greenbaum, J. Electrochem. Soc. 153 (2006) A 1226. 31. f. wang, m. hickner, y.s. kim, t.a. zawodzinski and j.e. mcgrath, J. Membr. Sci. 197 (2002) 231. 32. f. wang, m. hickner, q. ji, w. harrison, j. mecham, t.a. zawodzinski and j.e. mcgrath, Macromol. Symp. 175 (2001) 387. 33. m.j. sumner, w.l. harrison, r.m. weyers, y.s. kim, j.e. mcgrath, j.s. riffle, a. brink and m.h. brink, J. Membr. Sci. 239 (2004) 199. 34. e. vallejo, g. porucelly, c. gavach, r. mercier and m. pineri, J. Membr. Sci. 160 (1999) 127. 35. c. genies, r. mercier, b. sillion, n. cornet, g. gebel and m. pineri, Polymer 42 (2001) 359. 36. j. fang, x. guo, s. harada, t. watari, k. tanaka, h. kita and k. okamoto, Macromolecules 35 (2002) 9022. 37. t. watari, j. fang, k. tanaka, h. kita and k. okamoto, J. Membr. Sci. 230 (2004) 111. 38. x. guo, j. fang, k. tanaka, h. kita and k. okamoto, J. Polym Sci., Polym. Chem. Ed. 42 (2004) 1432. 39. y. yin, o. yamada, y. suto, t. mishima, k. tanaka, h. kita and k. okamoto, J. Polym. Sci.: Part A: Polym. Chem. 43 (2005) 1545. 40. j. sun, d.r. macfarlane and m. forsyth, Solid State Ionics 147 (2002) 333. 41. a. noda and m. watambe, Electrochim. Acta 45 (2000) 1265. 42. w. ogihara, j. sun, m. forsyth, d.r. macfarlane, m. yoshizawa and h. ohno, Electrochim. Acta 49 (2004) 1797. 43. j.r. maccallum, a.s. tomlin and c.a. vincent, Eur. Polym. J. 22 (1986) 787.


468


44. r.g. linford, in Electrochemical Science and Technology of Polymers 2 (Ed. R.G. Linford), Elsevier Applied Science, London (1990), p. 82. 45. g.c. farrington and r.g. linford, in Polymer Electrolyte Reviews 2 (eds. J.R. MacCallum and C.A. Vincent), Elsevier Applied Science, London (1990), p. 255. 46. e. salmon, s. guinot, m. godet and j.f. fauvarque, J. Appl. Polym. Sci. 65(3) (1997) 601. 47. m.j. reddy and u.v.s. rao, J. Mater. Sci. Letters, 17(19) (1998) 1613. 48. t. abe, n. gu, y. iriyama and z. ogumi, J. Fluorine Chem. 123(2) (2003) 279. 49. o. borodin, g.d. smith, r. bandyopadhyaya, p. redfern and l.a. curtiss, Modelling Simul. Mater. Sci. Eng. 12 (2004) S73. 50. j.y. lee, b. bhattacharya, d.w. kim and j.k. park, J. Phys. Chem. C 112(32) (2008) 12576. 51. n.k. karan, d.k. pradhan, r. thomas, b. natesan and r.s. katiyar, Solid State Ionics 179(19–20) (2008) 689. 52. morisatoa, z. hea, i. pinnaua and t.c. merkelb, Desalination 145 (2002) 347. 53. k. jankova, p. jannasch and s. hvilsted, J. Mater. Chem. 14 (2004) 2902. 54. k. ragavendran, p. kalyani, a. veluchamy, s. banumathi, r. thirunakaran and t.j. benedict, Portugaliae Electrochim. Acta 22 (2004) 149. 55. d.m. delongchamp and p.t. hammond, Langmuir 20(13) (2004) 5403. 56. y. masuda, m. seki, m. nakayama, m. wakihara and h. mita, Solid State Ionics 177(9–10) (2006) 843. 57. y. kato, k. hasumi, s. yokoyama, t. yabe, h. ikuta, y. uchimoto and m. wakihara, J. Therm. Anal. Calorimetry 69(3) (2002) 889. 58. m. wakihara, s. yokoyama, y. kato, h. ikuta and y. uchimoto, in Proceedings of the 8th Asian Conference: Trends in the New Millennium, Langkawi, Malaysia, 15–19 Dec (2002), p. 195. 59. c.p. fonseca, f. cavalcante jr, f.a. amaral, c.a. zani souza and s. neves, Int. J. Electrochem. Sci. 2 (2007) 52. 60. c.v. subba reddy, m. chen, w. jin, q.y. zhu, w. chen and s. mho, J. Appl. Electrochem. 37(5) (2007) 637. 61. m. watanabe, m. kanba, h. matsuda, k. tsunemi, k. mizoguchi, e. tsuchida and i. shinohara, Makromol. Chem., Rapid Commun. 2 (1981) 741. 62. d. pasquier, p.c. warren, d. culver, a.s. gozdz, g.g. amatucci and j.m. tarascon, Solid State Ionics 135 (2000) 249. 63. s. abbrent, j. plestil, d. hlavata, j. lindgren, j. tegenfeldt and a. wendsjo, Polymer 42 (2001) 1407. 64. j. saunier, f. alloin, j.y. sanchez and r. barriere, J. Polym. Sci., Part B: Polym. Phys. 42 (2004) 544. 65. n.t.k. sundaram and a. subramania, J. Membr. Sci. 289 (2007) 1. 66. t. iijima, y. tyoguchi and n. eda, Denki Kagaku 53 (1985) 619. 67. w. wixwat, j.r. stevens, a.m. anderson and c.g. cranqvist, in Second International Symposium on Polymer Electrolytes (ed. B. Scrosati), Elsevier Appl. Sci. Pub, London (1990), p. 461. 68. o. bohnke, g. frand, m. rezrazi, c. rousselot and c. truche, Solid State Ionics 66 (1993) 105. 69. g.b. appetecchi, f. croce and b. scrosati, Electrochim. Acta 40 (1995) 991. 70. s. sekhon, pradeep and s.a. agnihotry, in Solid State Ionics Science and Technology (eds. B.V.R. Chowdari et al.) World Scientific, Singapore (1998), p. 217.



469

71. s. rajendran and t. uma, Bull. Mater. Sci. 23(1) (2000) 27. 72. n. kubota, s. fujii, n. tatsumoto and t. sano, J. Appl. Polym. Sci. 83(12) (2002) 2655. 73. a. mohamad, Ionics, 11(5–6) (2005) 418. 74. g.x. wang, l. yang, j.z. wang, h.k. liu and s.x. dou, J. Nanosci. Nanotechnol. 5(7) (2005) 1135. 75. y. matsuda, m. morita, m. ihara and m. ishikawa, J. Electrochem. Soc. 140 (1993) L109. 76. x. liu and t. osaka, J. Electrochem. Soc. 143 (1996) 3982. 77. s.a. hashmi, s. suematsu and k. naoi, J. Power Sources 137 (2004) 145. 78. r.j. latham, s.e. rowlands and w.s. schlindwein, Solid State Ionics 147(3) (2002) 243. 79. a.s. best, s.m.j. viale and s.j. picken, US Patent No. 20080045615, Feb (2008). 80. t. tatsumisago, h. honjo, y. sakai and t. minami, Solid State Ionics 74 (1994) 105. 81. g.m. ehrlich, US Patent No. 6599664, July (2003). 82. h.e. becker, US Patent No. 2,800616 (1957). 83. d.i. boos, US Patent No. 3,536963, (1970) (to Standard Oil, SOHO). 84. r. kotz and m. carlen, Electrochim. Acta 45 (2000) 2483. 85. y. gofer, h. sarker, j.g. killian, t.o. poehler and p.c. searson, Appl. Phys. Lett. 71(11) (1997) 1582. 86. c. arbizzani, m. mastragostino and l. meneghello, Electrochim. Acta 40(13– 14) (1995) 2223. 87. j.y. kim and i.j. chung, J. Electrochem. Soc. 149(10) (2002) A 1376. 88. s.a. hashmi, r.j. latham, r.g. linford and w.s. schlindwein, Polym. International 47 (1998) 28. 89. s.a. hashmi, r.j. latham, r.g. linford and w.s. schlindwein, Ionics 3 (1997) 177. 90. a. clemente, s. panero, e. spila and b. scrosati, Solid State Ionics 85 (1996) 273. 91. a.m.p. hussain and a. kumar, J. Power Sources 161 (2006) 1486. 92. a.m.p. hussain, d. saikia, f. singh and d.k. avasthi, Nucl. Instr. Meth. B 240 (2005) 834. 93. a.m.p. hussain, d. saikia, f. singh and d.k. avasthi, Nucl. Instr. Meth. B 240 (2005) 871. 94. a.m.p. hussain, d. saikia, f. singh and d.k. avasthi, J. Phys. D, Appl. Phys. 39 (2006) 750. 95. k.r. prasad and n. munichandraiah, Electrochem. Solid-state Lett. 5(12) (2002) A271. 96. n.a. choudhury, a.k. shukla, s. sampath and s. pitchumani, J. Electrochem. Soc. 153(3) (2006) A 614. 97. x. liu and t. osaka, J. Electrochem. Soc. 143(12) (1996) 3982. 98. x. liu and t. osaka, J. Electrochem. Soc. 144(9) (1997) 3066. 99. m. ishikawa, m. ihara, m. morita and y. matsuda, Electrochim. Acta 40(13–14) (1995) 2217. 100. a. matsuda, h. honjo, k. hirata, m. tatsumisago and t. minami, J. Power Source 77 (1999) 12. 101. a. matsuda, h. honjo, m. tastumisago and t. minami, Extended Abstract for the 11th International Conference on Solid State Ionics, 15–21 November 1997, Hawaii, International Society on Solid State Ionics (1998), p. 76. 102. m. tatsumisago, h. honjo, y. sakai and t. minami, Solid State Ionics 74 (1994) 105.


470


103. s. mitra, a.k. shukla and s. sampath, Electrochem. Solid-state Lett. 6(8) (2003) A 149. 104. c. iwakura, h. wada, s. nohara, n. furukawa, h. inoue and m. morita, Electrochem. Solid-state Lett. 6(2) (2003) A 37. 105. a. lewandowski, m. zajder, e. frackowiak and f. beguin, Electrochim. Acta 46 (2001) 2777. 106. n. vassal, e. salmon and j.f. fauvarque, Electrochim. Acta 45 (2000) 1527. 107. k.w. park, h.j. ahn and y.e. sung, J. Power Sources 109 (2002) 500. 108. p. sivaraman, v.r. hande, v.s. mishra, c.s. rao and a.b. samui, J. Power Sources 124 (2003) 351. 109. p. sivaraman, s.k. rath, v.r. hande, a.p. thakur, m. patri and a.b. samui, Synth. Metal 156 (2006) 1057. 110. y.g. wang and x.g. zhang, Solid State Ionics 166 (2004) 61. 111. p. gomez-romero, m. chojak, k. cuentas-gallegos, j.a. asensio, p.j. kulesza, n. casan-pastor and m. lira-cantu, Electrochem. Commun. 5 (2003) 149. 112. a.k. cuentas-gallegos, m. lira-cantu, n. casan-pastor and p. gomez-romero, Adv. Funct. Mater. 15 (2005) 1125. 113. p. staiti and f. lufrano, J. Electrochem. Soc. 152(3) (2005) A 617. 114. f. lufrano, p. staiti and m. minutoli, J. Power Sources 124 (2003) 314. 115. f. lufrano and p. staiti, Electrochim. Acta 49 (2004) 2683. 116. p. sivaraman, a. thakur, r.k. kushwaha, d. ratna and a.b. samui, Electrochem. Solid-state Lett. 9(9) (2006) A 435. 117. p-c. chen, g. shen, s. sukcharoenchoke and c. zhou, Appl. Phys. Lett. 94 (2009) 043113. 118. k.r. prasad, k. kogara and n. miura, Chem. Mater. 16 (2004) 1845. 119. p. raghupathy, h.n. vasan and n.j. munichandraiah, J. Electrochem. Soc. 155 (2008) A 34. 120. f. e. amitha, a.l. reddy and s. ramaprabhu, J. Nanopart. Res. 11 (2009) 725. 121. m. kaempgen, c.k. chan, j. ma, yi cui and g. gruner, Nano Letters 9(5) (2009) 1872. 122. m.c. smart, b.v. ratnakumar, c.-k. huang and s. surampudi, Proc. 193rd Meeting of the Electrochemical Society, 3–8 May, San Diego, CA (1998). 123. c.h. hamann and w. vielstich, Elektrochemie, 4th ed., Wiley-VCH, New York (2005). 124. s. zihong and y. anbao, Chinese J. Chem. Engg. 17(1) (2009) 150. 125. a. chandra, r.c. agrawal and y.k. mahipal, J. Phys. D: Appl. Phys. 42 (2009) 135107. 126. n. munichandraiah, g. sivasankar, l.g. scanlon and r.a. marsh, J. Appl. Polym. Sci. 65(11) (1997) 2191. 127. d. rathod, m. vijay, n. islam, r. kannan, u. kharul, s. kurungot and v. pillai, J. Appl. Electrochem. 39 (2009) 1097. 128. m. selvakumar and d. krishna bhat, J. Appl. Polym. Sci. 110 (2008) 594. 129. r.j. latham, s.e. rowlands and w.s. schlindwein, Solid State Ionics 147 (2002) 243. 130. s. lee, y. lee, m.s. cho and j.d. nam, J. Nanosci. Nanotechnol. 8(9) (2008) 4722. 131. s. takashi, i. isao, i. syuichi and n. tomohiko, US Patent No. 5187032. 132. a. abdullah, s.z. abdullah, a.m.m. ali, t. winie, m.z.a. yahya and r.h.y. subban, Materials Res. Innovations 13(3) (2009) 255.


12 Polymer electrolytes for electrochromic devices X. F U, College of Chemistry and Chemical Engineering Southwest University, P.R. China

Abstract: The chapter begins by discussing the characters and composition of polymer electrolytes for electrochromic devices. It then describes the four types of the polymer electrolytes: dry solid polymer electrolyte, gel polymer electrolyte, porous gel polymer electrolyte and composite solid polymer electrolyte, their preparation procedures and properties especially ion conductivity of the samples. Finally, new types of polymer electrolytes including proton-conducting, alkaline, single ionic polymer electrolytes and electrolytes with ionic liquids are also introduced. Key words: polymer electrolyte, polymer matrix, composition, electrochromic device, ion conductivity.

12.1

Introduction

The demand for global environmental conservation and decrease of CO2 evolution require the development of a ‘new imaging medium’ instead of paper. This is called ‘electronic paper’, and electrochromic display is one of the potent candidates to realise this new medium. The ionic conductor is the essential part of the electrochromic devices (ECDs).1 As can be seen in the polymer battery, a solid state system is much more advantageous in electrochromic display. However, the image space selectively generated on a plane electrode by electrochromism has the disadvantage of the image spreading due to cell formation between the coloured and uncoloured part through an ionic conductor. High ionic conductivity in an ionic conductor such as an electrolyte solution is needed for rapid response in an electrochromic reaction, but it leads to spreading of the image, resulting in low resolution. This is a contradiction in ionic conduction required for electrochromic display. One answer to resolve this contradiction is to prepare a polymer electrolyte with high ionic conductivity in the writing and erasing processes. Electrolytes which are ubiquitous in all electrochemical devices, are the key component for ECDs. As the public’s awareness of environmental 471 © Woodhead Publishing Limited, 2010

472


protection has awakened, research on polymer electrolytes has grown. In general, solid polymer electrolytes (SPEs) have the advantages of, for example, no leakage of electrolytes, low density, safety and ease of production. Moreover, they are easy to prepare as a film, which has aroused a worldwide interest in polymer electrolytes; in addition, polymer electrolytes have a vast range of prospects for application in electrochromic devices and high energy storage apparatus such as supercapacitors, solar cells and sensors.2 Electrolytes are omnipresent in all electrochemical devices and their basic function is to serve as a medium for the transfer of charges, which are in the form of ions, between the electrodes, thus electrolytes must be viewed as an inert component and should be stable towards both the cathode and anode surfaces. Therefore, polymer electrolytes should have the following properties: high ionic conductivity, excellent chemical and thermal stability, a wide electrochemical window, and good mechanical strength. A polymer matrix poly(ethylene oxide) (PEO), poly(propylene oxide) (PPO), polyacrylonitrile (PAN), poly(vinyl chloride) (PVC), poly(vinylidene chloride) (PVDC), poly(methylmethacrylate) (PMMA) or poly(vinylidene fluoride)–hexafluoropropylene P(VdF–HVP), and the lithium salt (LiClO4, LiPF6, LiBF4, LiAsF6 or LiCF3SO3) are used to prepare polymer electrolytes.

12.2

Electrochromic effect and electrochromic devices

A material can be defined as electroactive or electrically responsive if it responds to an electrical stimulation with a reversible variation of one or more physicochemical properties.3 Owing to the ease of generation and processing of input signals of the electrical type, electrically responsive materials with tailorable functionalities are considered today as smart materials. An electroactive material is usually said to be electrochromic if it shows reversible colour changes when a potential difference is applied across it. More properly, the electrochromic effect is defined as a visible and reversible variation of optical properties, namely light transmittance and/or reflectance, shown by a material upon its electrochemical oxidation/reduction.4,5 The particular properties of transmission or reflection of light account for the peculiar colour of an electrochromic material. Colour changes typically range between a transparent (bleached) state and one or two coloured states, corresponding to different redox conditions. In the case of materials with more than two redox states, the materials may exhibit several coloured states and be termed polyelectrochromic. Frequently, electrochromic materials are prepared as thin films and coated on optically transparent electrodes, so they can be assembled into a battery-like ECD upon sandwiching an electrolyte in between.


Polymer electrolytes for electrochromic devices

473

Substrate (plastic or glass) Transparent conductor +

Electrochromic layer Electrolyte

–

Ion storage layer (Transparent) conductor Substrate (plastic or glass)

12.1 Classical seven-layer structure of an electrochromic device.

Historically, following the first report on electrochromism by Deb in 1969,6 transition metal oxides, and particularly tungsten trioxide (WO3), were the first inorganic materials in which the electrochromic effect was observed and reported. Currently, this effect is being studied not only with inorganic materials7,8 but also, and largely, in organic compounds such as conducting conjugated polymers, viologens, metallo-polymers and metallophthalocyanines.7,9 In fact, as will be shown here, organic electrochromic materials offer several advantages with respect to inorganics, not only in terms of flexibility, ease of processing and low cost, but also with respect to both ‘tailorability’ and efficiency of coloration and multi-colour imaging. The structure of a classical ECD is a typical multilayer electrochemical cell, represented in Fig. 12.1.3 It consists of up to seven superimposed layers of materials, on one substrate or positioned between two substrates in a laminate configuration, transparent to visible light. The electrochromic material is coupled to an ion conductor (solid or liquid electrolyte) and an ion storage layer; these three optically transparent layers are sandwiched between two conductors, at least one of which must be transparent too (typically made of indium tin oxide (ITO)). The resulting five layers are protected by two transparent plastic or glass substrates. The electrochromic layer can be coupled to either the cell anode (as in Fig. 12.1) or cathode. In the two cases it is respectively said to be an anodically or cathodically colouring material, i.e. it shows colour variation when it is oxidised or reduced. Alternatively, two electrochromic species can be present in two symmetrically arranged layers; for such a case, cathodic and anodic coloration processes are simultaneously driven. By applying a voltage difference (typically 1–5 V) to the cell, the electrochromic material is oxidised or reduced, depending on the voltage polarity. The electrochromic effect is induced by such electrochemical redox processes experienced by the active material. In particular, following its oxidation (reduction), the material experiences an insertion of counter-anions (cations), flowing into it from the ion storage layer through the ion-conducting electrolyte; such ions carry out an anodic (cathodic) doping of the material, in order to balance its


474


ejected (injected) electric charge. As a result, the electrochromic effect arises from the occurring modification of electronic properties of the material (in particular its band gap),10 correlated to a variation of its optical properties (e.g. absorption bands) and, then, its colour. For WO3, MoO3, etc., the electrochemical reaction on the cathodic electrodes is: MO y (bleached) + xA + + xe − A x MO y (0 < x < 1, coloured) in which A+ = H+, Li+, Na+, K+, etc. are injected into the electrochromic material MOy layer from the ion storage layer through the ion-conducting electrolyte; while for Ir, Ni, Rh, Co, etc. anodic electrochromic material, the electrochemical reaction on the anodic electrodes were: MO y (bleached) + xB+ − xe − B x MO y (0 < x < 1, coloured) in which B− = OH−, F−, ClO4−, CN−, etc. are injected into the electrochromic material MOy layer from the ion storage layer through the ion-conducting electrolyte. Both organic and inorganic materials have been used to design and construct ECDs that work as electrochemical cells.11–17 One of the simplest configurations for an ECD consists of a four layer assembly: a transparent electronic conductive film, usually indium-doped tin oxide) on glass or poly(ethylene terephthalate) (PET), a coating of the electrochromic material, an ionically conductive layer, and another ITO film on glass or PET. Typically, the use of only one electroactive film in an ECD translates to a short device lifetime. This is due to the lack of a second electroactive film on the counter-electrode to complete the redox process, ion shuttling, in the cell, leading to degradative reactions at the electrode. This issue has been alleviated by the use of dual-type configurations,18–21 in which an additional electroactive layer is added on the counter-electrode in the assembly to accommodate ion shuttling. This additional layer could also be electrochromic and therefore could be used to obtain mixed coloured states, or enhance contrast by using electrochromic materials with complementary characteristics.

12.3

Electrolytes for electrochromic devices

Electrolytes typically used in ECDs, as in any other electrochemical cell, belong to four main classes:3 aqueous electrolytes, organic liquid electrolytes, ionic liquids and solid polymer electrolytes. The adoption of so-called ionic liquids, such as ethyl ammonium nitrate ([EtNH3][NO3]), 1-butyl-3methylimidazolium tetrafluoroborate ([BMIM][BF4]) or hexafluorophosphate ([BMIM][PF6]), can result in improved lifetime and response speed for electrochromics and actuators,22,23 and are best proposed being mixed with polymer or gel matrices.24,25



475

Most of the solid polymer electrolytes used can be classified as follows: (1) polymer or gel matrixes swollen with liquid electrolyte solutions (e.g. ethylene carbonate (EC)/PAN/sodium perchlorate (NaClO4)); (2) singleion systems in which only one ionic species is mobile within a polymer matrix (e.g. perfluorosulphonate ionomer Nafion®); (3) solvent-free ioncoupled systems consisting of ion-solvating polymers mixed with salts, so that cations and ions become mobile within the polymer network, e.g. PEO mixed with salts. Polymer electrolytes for ECDs should have the following common properties: high ionic conductivity for ion transmission, excellent chemical and thermal stability, wide electrochemical window, good mechanical strength, and adhesion to the glass substrate, etc. In addition, electrolytes for ECDs usually should have high transparency in the visible region. ECDs need not charge and recharge, but the durability or cycle life of polymer electrolytes should beyond 1 × 106 cycles, even more than 1 × 107 write–erase cycles for ECDs used as display.

12.4

Polymer matrix

The polymer matrix is the main component and the framework for polymer electrolytes. The matrix plays the role of excipient and sustains the shape as a layer of film or membrane for the polymer electrolytes.

12.4.1 Structure types of ionic passageway in a polymer matrix The segments and chains in the molecules of the polymer matrix may take different special conformations and arrangements to accommodate the passageway for ionic transport. The ionic passageway is dependent on the molecule type of polymer matrix, structure characters of the whole molecule, the properties and mode of location of the electron pair donating atoms or groups in the polymer matrix. Some structure types of ionic passageway26 being considered and deduced are shown in Fig. 12.2. A proper polymer matrix must possess following properties:27 •

The molecules of the polymer matrix must contain some atoms such as O, N, S, P or groups which strongly complex with Li+ ion, so as to form coordinating bonds between Li+ ion with these atoms or groups. • In the polymer matrix molecules, a proper distance must be kept between the coordinating centres which are the complexing groups or atoms, in order to form multi-coordinating bonds for any single Li+ ion. • The molecule chain of the polymer matrix should have sufficient flexibility.


476


(a)

(b)

(c)

(d)

(e)

12.2 The structure types of ionic passageway: (a) go down, (b) shelf, (c) line, (d) macrocycle pipeline, (e) helix line.

12.4.2 Composition and stereo structures of the polymer matrix At present, the most commonly and frequently used polymer matrix for the polymer electrolytes are PEO or PEO-modified polymer derivatives, such as polyether, polyester, etc.28–33 PEO-based polymers are known to be excellent for dissolving lithium salts without allowing crystallisation even operating at fairly low temperatures. According to the classification of the molecule composition, the polymer matrix used includes: •

•

•

• • • •

polyether derivatives,34,35 e.g. PEO –(CH2CH2O)n–, PPO –[CH(CH)3CH2O]n–, PEO–PPO–(CH2CH2O)n–[CH(CH)3CH2O]m–, and OMPEO (oxymethylene-linked poly(oxyethylene) block polymers)36,37–(CH2CH2O)n–CH2–(OCH2CH2)m–; polyester, e.g. PPA (poly-β-propylactone) –(CH2CH2COO)n–, PMMA,38,39 PMA (polymethyl acrylate), and copolymers of ester with other compounds such as P(VAc–MMA) (poly(vinyl acetate–methyl methylacrylate copolymer),40 P(VAc–MA) (polyl vinyl acetate–methyl acrylate copolymer),41 other aliphatic polyesters and copolymers, etc.; nitrogen-containing polymer, e.g. PAN –(CH2CHCN)n–, polyimine –(CH2CH2NH)n– or –[CH2CH2N(CH3)]n–,42 polyinemine, PU (polyurea), etc.; sulphur-containing polymer, e.g. polysulphoethylene –(CH2CH2S)n–;43 olefin halide polymer such as PVC,44 PVdF,45 and their copolymers; siloxane derivative polymers and their copolymers, etc.;46 polyvinyl butyral (PVB)1,47,48 and others.

The ionic conductivity in polymer electrolytes is related to the segmental motion of the polymeric chains; thus the low degree of crystallinity for polymer electrolyte may be propitious for high ionic conductivity. The conductive mechanism can be stated as follows: lithium ions first bond with the



477

Linear polymer

Comb polymer

Crossing polymer

Star polymer

Hyperbranched polymer

Comb crosslinking polymer

12.3 Structure sketch map of the different polymer hosts for polymer electrolytes.

polarity chains of the polymer. The polymer chains are constantly undergoing local segmental motion under the effect of an electric field. A consequence of these local dynamics is that free volume is constantly being created and destroyed. Ions are located at suitable coordination sites within the polymer. The occurrence of ionic transport, free volume and more specifically a suitable coordination site must be created by the chain dynamics adjacent to the existing site. When this occurs, the ion may migrate to a new site.49 Usually, both crystalline and amorphous phases are present in polymer electrolytes. Since only amorphous phases present high conductivity, plasticiser solvents are usually added to enhance the amorphous phase and thus the ionic conductivity. Here, propylene carbonate (PC), EC, γ-butyrolactone (γ-BL) and their binary mixtures are usually used as plasticiser solvents.50 The stereo structures of the polymer matrix molecules are given in Fig. 12.3,51 including linear, comb, crossing, star, hyper-branched and comb crosslinking.

12.5

Classification of polymer electrolytes

Generally, polymer electrolytes used for ECDs can be classified into the following four types according to their physical configuration and chemical composition: dry solid polymer electrolyte (DSPE), gel polymer electrolyte (GPE), porous gel polymer electrolyte (PGPE) and composite solid polymer electrolyte (CSPE).

12.5.1 Dry solid polymer electrolyte DSPEs are commonly obtained by dispersing a lithium salt into a conventional polymer matrix. The ionic conductivity of DSPE only reaches


478


practically useful values at high temperatures. Copolymerisation,52 crosslinking53 and blending54 of the polymer matrix can effectively enhance ionic conductivity. However, the low conductivity of DSPEs is still the biggest obstacle for their application. It has been reported that the addition of ceramic powder,55,56 nanosized ZnO particles,57 modified SiO2 powders,58 SnO2 nanoparticles,59 Mg(ClO4)2,60 La0.55Li0.35TiO2,61 ferroelectric oxidates such as BaTiO3,62 lithium aluminate (γ–LiAlO2 powder),63,64 and α-LiAlO2 nanotube bundles,65,66 etc. leads to the enhancement of ionic conductivity, mechanical strength and chemical stability. Some samples of typical DSPE for electrochromic devices are as follows. Solid polymer electrolyte (SPE) based on poly(iminoethylene)-bpoly(oxyethylene)-b-poly(iminoethylene)(PEI–PEO–PEI) Synthesis of the copolymer (PEI–PEO–PEI) The linear poly(iminoethylene)-b-poly(oxyethylene)-b-poly(iminoethylene) (PEI–PEO–PEI) copolymer67 was prepared by a three-step method:68 conversion of the two hydroxyl groups of an α,ω-hydroxy-PEO to two sulphonate esters Ms2PEO, or Ts2PEO, by reaction with the corresponding sulphonyl chloride; • ring opening polymerisation of 2-oxazoline (i.e. 2-methyl-2-oxazoline, MeOZO) initiated by the Ms2PEO, or Ts2PEO, leading to triblockcopolymer (PAcIE-b-PEO-b-PAcIE) (PAcIEpoly(N-acetyliminoethylene)); • hydrolysis of triblock-copolymer (PAcIE-b-PEO-b-PAcIE) to triblockcopolymer (PEI-b-PEO-b-PEI). •

Preparation of the SPE PEI-b-PEO-b-PEI copolymer weighing 150 mg and different quantities of lithium bis (trifluoromethyl sulphonylimide) (LiTFSI) or Cu(TFSI)2 salts, were dissolved in 10 ml dry methanol, under agitation, in an inert atmosphere, so that the ratio between the cation and the coordinating atoms was 1/12 for both Li+/O and Cu++/N. After 48 h of agitation a yellow solution of LiTFSI-block copolymer, and a blue solution of Cu(TFSI)2-block-copolymer were obtained. The solvent was evaporated under inert atmosphere, at room temperature (RT). In a similar way the polymer electrolyte based on PEI-b-PEO-b-PEI copolymer and the mixture of the two salts (LiTFSI and Cu(TFSI)2) was prepared. The new copolymer electrolytes with solvating affinities for different cations could be used either as a reference electrode or in the fabrication of ECDs.



479

Solid polymer electrolyte (SPE) based on oxymethylene-linked poly(oxyethylene) (OMPEO) with a mixture of PEG400 and 600 OMPEO block polymers36 were synthesised by condensing dichloromethane with a mixture of different molecular weight of PEG400 and 600 in the presence of potassium hydroxide powder. SPEs based on the OPMEO multi-block polymer were prepared. Preparation of multi-block polymer To 25 g of KOH powder, a mixture of 120 ml CH2Cl2 and 25 g PEG, whose ratio is 10 g of PEG400 and 15 g of PEG600, was slowly added. After the solution was stirred for 20 h, the crude product was filtered under reduced pressure to remove the KCl and unreacted KOH particles completely. Finally, the homogeneous product was dried at 60 °C in a reduced pressure vacuum for 24 h. A transparent OMPEO multi-block polymer was obtained. Preparation of SPE film OMPEO, LiClO4, SiO2, EC/PC and 40 ml of acetonitrile were dispersed in different weight percentages and stirred with the help of magnetic bar for 24 h, until homogeneity was attained. The obtained SPEs can also be used as solid films by using solvent evaporation techniques. The films were further dried for 24 h in vacuum at 60 °C to remove any trace of acetonitrile. The ionic conductivity of the SPE at room temperature is 1.79 × 10−3 S cm−1, which can be well controlled under the following optimised compositions: 0.7 g LiClO4, 0.7 g SiO2, 7 g block-polymer, 5/0 ratio of EC/PC. Solid polymer electrolytes (SPE) based on polyether with hyperbranched side chains Network polymer electrolytes based on polyether with hyperbranched side chains were prepared in the following three steps.69 Synthesis of monomer 2-(2-methoxyethoxy)ethyl glycidyl ether (MEEGE) To a mixture of epichlorohydrin (270 g, 3 mol), NaOH (120 g, 3 mol), water (12 g, 0.67 mol) and 2-(2-methoxyethoxy)ethanol (MEE) (60 g, 0.5 mol) with dissolved (C4H9)4N+ (HSO4)−(3.3 g, 9.7 mmol) was added dropwise with stirring at 40 °C for 1.5 h. The reaction mixture was filtered to remove


480


excess NaOH and formed NaCl. It was dried with anhydrous magnesium sulphate. The crude mixture was fractionally distilled under reduced pressure several times to give a colourless liquid (26.4 g, yield 35%). Synthesis of macromonomers Into a dried autoclave were charged MEE, and 3 wt% of KOH, dried under vacuum at 60 °C overnight. The autoclave was repeatedly evacuated and purged by nitrogen at 70 °C to remove water and to obtain the potassium alkoxide. An approximately 20 g portion of the mixture of 2-(2-methoxyethoxy)ethyl glycidyl ether (MEEGE), and EO ([MEEGE]/[EO]) = 15/85 was added several times at 120 °C. Stirring was continued for 2 h at 120 °C, and cooled to 50 °C, an equivalent amount of water was added to convert the potassium alkoxide to KOH. After neutralisation with sulphuric acid and desalting, the reaction mixture was dried under high vacuum at 120 °C to obtain pale-yellowish viscous liquids of OH-terminated oligomers (Fig. 12.4). Acrylic acid, ρ-toluenesulphuric acid (PTS) as a catalyst, and hydroquinone (HQ) as a polymerisation inhibitor were added to solutions of the OH-terminated oligomers in toluene. Esterification proceeded by removing the formed water. The excess acrylic acid, PTS, and HQ were neutralised by NaOH solutions and then desalted. The reaction mixture was extracted by toluene, and toluene was evaporated under vacuum to obtain a paleyellowish viscous liquid (Fig. 12.4). Preparation of network polymer electrolytes LiTFSI, LiClO4, LiBF4, LiPF6, LiCF3SO3, NaCF3SO3 and KCF3SO3 were used as electrolyte salts for the network polymer electrolytes. Weighed H2C

CHCH2

O

CH2CH2

OCH2CH2OCH3 + H2C

O

CH2 CH3OCH2CH2OCH2CH2OH KOH

O

CH3OCH2CH2OCH2CH2O

[(CH2CHO)x

(CH2CH2O)1–x]n H

CH2OCH2CH2OCH2CH2OCH3 CH3OCH2CH2OCH2CH2O

[(CH2CHO)x

H3C

CHCOOH PTS/HQ

(CH2CH2O)1–x]n COCH = CH2

CH2OCH2CH2OCH2CH2OCH3

UV irradiation MX/DMPA

Network polymer electrolytes

DMPA = PhCO–C(OCH3)2–Ph

12.4 Synthesis of macromonomer and network polymer electrolytes.



481

amounts of the salt, the macromonomer, and 2,2-dimethoxy-2-phenylacetophenone (0.05 wt% based on the macromonomer) photoinitiator were mixed together in the glovebox to make homogeneous viscous solutions. These were spread between two glass plates separated by poly (tetrafluoroethylene) (PTFE) spacers and were irradiated with UV light (250 W super-high pressure Hg lamp) for 5 min to obtain flexible polymer electrolyte films. The highest conductivity of 1 × 10−4 S cm−1 at 30 °C, 1 × 10−3 S cm−1 at 80 °C, was obtained with LiTFSI as the electrolyte salt. Solid polymer electrolyte (SPE) based on random co-polymer poly(ethylene oxide) and epichlorohydrin (P(EO–EPI)) Sample 1 The SPE70 was prepared with Epichlomer-16 (Daiso), a random copolymer of ethylene oxide and epichlorohydrin (EPI) (84 : 16, MW = 1 300 000), by dissolving 0.60 g of this elastomer and 0.032 g LiClO4 in 10 ml tetrohydro furan (THF). This concentration gives n = [O]EO/[Li] = 29. At this concentration, the ionic conductivity of this electrolyte is 4.1 × 10−5 S cm−1 in dry atmosphere ([H2O] < 1 ppm) and 2.6 × 10−4 S cm−1 at 84% relative humidity. Sample 2 Complexes of alkaline metal salts and the random co-polymer of ethylene oxide and epichlorohydrin, P(EO–EPI), in 84 : 16 proportion (Daiso), present high ionic conductivities at room temperature (20–40 mS cm−1 at 30 °C, [H2O] < 0.01%).71–73 This co-polymer has been used in all-polymeric electrochromic and photoelectrochemical devices, with satisfactory results.72–74 A film was obtained by evaporation of a solution containing 0.60 g of the P(EO–EPI) (co-monomer proportion = 84 : 16, MW = 1 300 000 g mol−1) and 0.032 g LiClO4 in 10 ml THF. For device assembly, 100 µl of the electrolyte solution was deposited on the surface of each electrode (previously conditioned in the suitable reduced or oxidised form). After solvent evaporation, the electrodes were carefully pressed against each other to obtain good adhesion. Solid polymer electrolyte poly(vinylacetate–methylmethacrylate) (SPE) based on P(VAc–MMA) Some studies75,76 reported the polymer electrolytes based on the blend components of PVAc and other polymers, e.g. PVAc/PMMA,77 PVAc/


482


PEO,78 PVAc/PVdF and PVAc/P(VdF-co-HFP),79 which all have good mechanical performance and high ionic conductivity. Synthesis of random copolymer P(VAc–MMA) Sodium dodecyl sulphate, (NH4)2S2O8, NaHCO3 and H2O were weighed and added to the flask with stirring, then VAc and MMA monomers were weighed and added to the constant pressure funnel. The flask was placed in a water bath under an argon atmosphere. When the temperature reached 75 °C, about 5 ml admixture of VAc and MMA was added to the flask as seeds. After 40 min reaction, all seeds had polymerized; the remaining mixture of VAc and MMA was then added dropwise into the flask at a speed of 0.6 g/min. After addition of the admixture finished, a further 60 min of reaction was continued. Then the reaction emulsion was cooled and poured into a beaker containing 100 ml of solution of 0.01 mol/l KAl(SO4)2 with continuous stirring. The crude copolymer was washed several times and filtered to remove the unreacted monomers and residual solution, then dried under vacuum at 60 °C. Then the crude product was dissolved in THF and any almost imperceptible solids were removed by filtration. The filtrate was distilled to remove the THF and further dried under vacuum at 60 °C. A random copolymer P(VAc–MMA) was obtained. All the copolymers with different VAc : MMA ratios were highly transparent and colourless. Preparation of P(VAc–MMA) polymer electrolyte Weighed amounts of P(VAc–MMA) and LiClO4 were mixed in THF with stirring at 30 °C for 12 h, until P(VAc–MMA) and LiClO4 were completely dissolved into a viscous solution. Then the solution was poured onto the Teflon plate. THF was evaporated slowly at room temperature to obtain polymer electrolytes films. After being dried under vacuum at 50 °C for 48 h, the resultant films were transferred to the glove box. The SPEs prepared have excellent thermal stability and good mechanical performance, in which with copolymer (MMA :VAc (mol) = 2 : 8) as the host polymer, containing 25 wt% LiClO4, exhibits the highest conductivity of 1.27 × 10−3 S cm−1 at ambient temperature. Solid polymer electrolyte (SPE) based on poly(vinylacetate– methylacrylate)/poly(methylmethacrylate) (P(VAc–MA)/PMMA) Preparation of copolymer P(VAc–MA) Some 100 g water, 0.50 g sodium dodecyl sulphate, 0.17 g (NH4)2S2O8, and 0.17 g NaHCO3 were added to a flask with stirring. The flask was placed in



483

a water bath under argon atmosphere. When the temperature reached 70 °C, 4.0 g of VAc and MA monomer mixture (mole ratio 1 : 1) was added into the flask as seeds. After 30 min reaction, all the seeds had polymerized. The remaining 46 g of VAc and MMA mixture was added dropwise to the flask within 2.5 h. After addition of the admixture finished, a further 60 min of reaction was continued. Then the reaction emulsion was cooled and poured into a beaker containing 100 ml of solution of 0.01 mol/l KAl(SO4)2 with continuous stirring. The crude copolymer was washed several times and filtered to remove unreacted monomers and residual solution, then dried under vacuum at 60 °C. Then the crude product was dissolved in THF and any almost imperceptible solids were removed by filtration. The filtrate was distilled to remove the THF and further dried under vacuum at 60 °C. A random copolymer P(VAc–MA) was obtained. All the copolymer products with different VAc : MMA ratios were highly transparent and colourless.

Preparation of P(VAc–MA)/PMMA SPE Weighed amounts of P(VAc–MA),80 PMMA and LiClO4 were mixed in anhydrous THF and stirred at 40 °C for 24 h, until P(VAc–MA), PMMA and LiClO4 were completely dissolved into a viscous solution. Then the solution was poured onto the Teflon plate. THF was allowed to evaporate slowly at room temperature to obtain polymer electrolyte films. After being dried under vacuum at 50 °C for 48 h, the resultant films were transferred to the glove box. The maximum ionic conductivity value was 1.17 × 10−3 S cm−1 at 25 °C for the SPEs based on blends P(VAc–MA)/PMMA, and the conductivity– temperature plots are found to follow the Arrhenius equation.

Sold polyme electrolyte (SPE) based on branched poly(ethylene glycol) (PEG)–boronate ester polymers Boron, having an affinity for both anions and the ether oxygens, was expected to reduce the strength of the interactions of the Li+ cation with both anions and the ether oxygens of the PEO segments and consequently facilitate the Li+ transfer process at the interface. Thus polymer electrolytes based on polymers of boron and aluminium81,82 including borates, boronate esters, aluminates and aluminocarboxylates,83,84 might be applied in ECDs, although their measured overall conductivity is generally lower than in corresponding salt-containing electrolytes based on the same polymer matrix without any boron or aluminium.


484

Polymer electrolytes H2O

B2O3 + HO

(

O

) m H + HO (

O

)n

CH3 –

Reflux in toluene

O

O

(

B O

(

) n CH3

O

)m k

12.5 Preparation of branched PEG–boronate ester polymers. The product on the right-hand side represents an average structure.

Polymer synthesis As shown in Fig. 12.5, poly(ethylene glycol)s (PEGs) with different molecular weights (100, 200, 300, 600 and 1500 g mol−1) were used as main chain segments. Diethylene glycol monomethyl ether (DEGMME) and triethylene glycol monomethyl ether (TEGMME) were used to introduce branches or side chain segments. The reactants B2O3, PEG and DEGMME or TEGMME were first introduced to the reactor in a balanced composition to obtain 10 g of final polymer product. The molar ratio between the reactants was kept at 0.5 : 1 : 1.03 respectively. Thus, a slight excess of DEGMME or TEGMME, corresponding to 1 mol% excess of hydroxyl groups in relation to the boron charged, was used in order to reach a complete conversion of the boron trioxide. Next, 50 ml of toluene was added and the mixture was heated to 60 °C. Then the mixture was refluxed for 4 h during which most of the water formed in the reaction was removed. Residual water was removed using a Soxhlet apparatus containing molecular sieves (0.4 nm). Finally, the polymers were placed in a glove box under Ar atmosphere having a moisture content below 1.5 ppm. All the polymers were highly transparent and colourless.

Electrolyte preparation All the electrolytes were prepared in the Ar-filled glove box. A 20 wt% solution of LiClO4 in dimethyl carbonate (DMC) were added to weighed amounts of the different polymers to obtain electrolytes with the desired salt concentrations. In order to obtain SPEs, the DMC was removed first by evaporation on a heating plate at 50 °C, and then under high vacuum at 60 °C. The PEG–boronate ester polymers having molecular weights equal to, or below, 300 g/mol were found to be amorphous, colourless and highly transparent. Electrolytes derived from these polymers and LiClO4 reached ionic conductivities up to 10−4 S cm−1 at room temperature, and were thermally stable up to at least 150 °C.



485

Dual-function electrolytes based on poly(trimethylene carbonate) Poly(trimethylene carbonate), P(TMC), electrolytes have the composition [–(C=O)–O–CH2CH2–CH2O–]n, where n represents the molar ratio of [–(C=O)–O–CH2CH2CH2O–] units per guest lithium ion.85 Preparation of polymer electrolytes High molar mass P(TMC) (3 × 105 g/mol−1) was prepared by catalysed bulk polymerisation and characterised by gel permeation chromatography. Homogeneous solutions of P(TMC) and lithium salts in acetonitrile were prepared by adding known masses of polymer and lithium salt to a small conical flask. A convenient volume of acetonitrile was transferred to the flask and the components were stirred for at least 48 h. The resulting homogeneous solutions were cast on glass plates and the solvent was evaporated slowly to form films of about 150 µm thickness. These electrolyte films were transferred to a tubular oven at a temperature from 30 to 60 °C. This optimised procedure was used to prepare P(TMC)nLiX compositions with n between 3 and 10. The choice of SPE formulation applied in prototype devices was based on criteria of conductivity and mechanical properties. The use of P(TMC)-based components improves the cycle lifetime and durability of the ECD relative to conventional electrolytes and improves leakage performance, memory effect and humidity resistance. Solid polymer electrolytes (SPE) based on photo-polymerised 2-hydroxyethylmethacrylate (HEMA) New SPE films based on photo-polymerised poly(2-hydroxyethylmethacrylate) PHEMA86 with neopentyl glycolmethacrylate (NPGMA) as a crosslinker were prepared, the synthetic route of the photo-polymerised and crosslinked PHEMA is shown in Fig. 12.6. Preparation of SPE films Some 20 wt% of the monomer 2-hydroxyethylmethacrylate (HEMA), 0.8 wt% of the copolymer NPGMA that also acts as a crosslinker and 0.01 wt% of benzoin methyl ether (BME) as a photoinitiator with respect to the liquid electrolyte’s (1 M LiClO4 in EC/PC, 1 : 1 (film A), 3 : 1 (film B) by volume) weight were mixed together. After stirring the mixture for 10 min, the resulting solution was poured and spread on the desired substrate to perform photo-polymerisation using a UV lamp (Spectroline, SB-125). Some 1 wt% of deionised water was mixed with the electrolyte mixture (prior to polymerisation) to determine its influence on the electrochromic performance (switching kinetics) of the device.


486

Polymer electrolytes CH3

H2C

CH3

CH3

CH3 UV light

C–COOCH2CH2OH + CH2=C–COO–CH2–C–CH2OOC–C=CH2

BME as photoinitiator

CH3 CH3 [CH2–C]x

CH3 [CH2–C]y

CO2CH2CH2OH

CH3

CO2–CH2–C–CH2OOC CH3

[C–CH2]z CH3

12.6 The synthetic route of the photo-polymerised and crosslinked PHEMA.

Application of SPE in ECD The PHEMA-based SPE film prepared is flexible, transparent and exhibits a room temperature ionic conductivity of 4.32 × 10−3 S cm−1. High ionic conductivity observed for this film [HEMA–NPGMA–1 M LiClO4/EC : PC (1 : 1) + 1 wt% H2O] is probably a repercussion of its morphicity. This film can work in the temperature range between −90 and ~220 °C, which has proved that the PHEMA film is an ideal candidate for the construction of all solid state electrochromic windows. Solid polymer electrolyte (SPE) based on polyvinyl butyral Polyvinyl butyral (PVB)1,47,48 has been in use for laminated safety glass for about 60 years. Now it can also be used for the manufacture of electrochromic glazings. The polymer electrolyte could require excellent ionic conductivity at higher temperature but very low conductivity at ambient temperature. This PE enables the electrochromic reaction leading to image formation at higher temperature, but electrochemical reactions should not occur at ambient temperature. Since salt dissociates to become ions under high dielectric environment and ions migrate through low viscosity media, the addition of crystals consisting of molecules with a high dipole moment into a suitable polymer–salt composite is effective in inducing ionic conduction only above the melting temperature of the additional crystals. When above its melting temperature, additives can melt in the polymer matrix, and molten additive is expected to form a viscous and high dielectric domain resulting in high ionic conduction, while below melting temperature, the additive would crystallise.



487

Sample 1 PVB, tetrabutylammonium perchlorate (TBAP), imidazole (Imz) or PEG2000 were employed as host polymer, salt and additive for the polymer electrolyte. Imz or PEG2000 are the molecules that are effective in forming an ionic conduction pathway above its melting temperature. TBAP also has a character of melting.47,48 The solution for the PE was prepared by mixing PVB and a suitable amount of TBAP and Imz in THF. The mixture was cast on a Teflon plate and was allowed to stand at 56 °C under N2 flow and further dried at 70 °C in vacuo. The resulting film was kept in vacuo before use. The polymer electrolyte containing PVB of 20 wt% provided freestanding film and showed higher ionic conductivity than others in the present experiment. Sample 2 PVB, a suitable amount of TBAP and PEG2000 were dissolved into chloroform to prepare the polymer electrolyte. The chloroform solution was cast on a Teflon plate and was allowed to stand at 56 °C under N2 flow and further dried at 70 °C in vacuo. The resulting film was kept in vacuo before use.1 At room temperature, the ion conductivity of the sample is about 8 × 10−6 S cm−1, and above the melting temperature 80 °C, the ion conductivity is about 10−4 S cm−1.

12.5.2 Gel polymer electrolytes Polymer gel electrolytes, also called third generation polymer electrolytes, are suitable materials for electrochromics. They are usually prepared by the immobilisation of an aprotic solvent in the network of a polymer such as PMMA. On the basis of the technique involved in the synthesis, gels can be classified as physical and chemical gels. A physical gel39 can be regarded as a liquid electrolyte encaged in a polymer matrix, e.g. LiClO4/EC/PC in PMMA where no chemical bond formation occurs between the liquid electrolyte and the polymer chain. Alternatively, a relatively small amount of crosslinker used with the polymer gives a freestanding, self-supporting gel film characterised by the formation of covalent bonds between the polymer chains. These chemical bonds can be formed directly by the reaction of functional groups on the polymers or by the added crosslinking agents. Chemical gels can be obtained in gel as well as in film forms. GPEs possess both the cohesive properties of solids and the diffusive property of liquids. It has been shown that the properties of the gels studied strongly depend on the type of polymer matrix used as well as the type of


488


solvent and concentration of lithium salt. Plasticizers with high dielectric constants such as EC, PC, DMC or diethyl carbonate (DEC) are commonly used to plasticise the polymer electrolytes. The plasticisers in polymer electrolytes can not only decrease the degree the crystallinity of polymer electrolyte, but also can enhance the lithium ion transport. The room temperature conductivity of GPEs was found to be of the order of 10−4 S cm−1, so they are satisfactory for commercial application.87,88 However, the use of plasticisers may also destroy the mechanical strength, chemical stability and interfacial stability with lithium metal electrodes, which causes poor cycling performance of electrochemical devices. Gel polymer electrolyte (GPE) based on polyvinyl butyral (PVB) PVB-based gel electrolyte87 was prepared to apply to the flexible ECD. The ionic conductivity of GPEs depended on the polymer content, and was higher than 10−4 S cm−1 at 25 °C at a PVB content of 33 wt%. It is revealed that PVB-based GPEs work well as the material for flexible ECDs showing subtractive primary colours. Preparation of the GPE The PVB-based GPE was prepared according to the procedure as follows: 20 ml of NMP (N-methyl pyrrolidione) solution was prepared by dissolving 97 mg (25 mM) of DMT (dimethyl terephthalate), 342 mg (50 mM) of TBAP (tetra-n-butylammonium perchlorate) as supporting electrolyte, 93 mg (25 mM) of ferrocene (Fc) as counter material in NMP. One gram of the solution was mixed with the appropriate amount of PVB and the resulting mixture was allowed to stand for 7 days to obtain the PVB-based gel electrolyte. Properties of the GPE The conductivity of the gel electrolyte was estimated to be 2.3 × 10−4 S cm−1 at 25 °C at a PVB content of 33 wt%. It is expected that the flexible EC device with the electrolyte and flexible ITO–PET electrodes is a potent candidate for a multi- or full-colour paper-like display. Gel polymer electrolyte (GPE) based on poly(methyl methacrylate) (PMMA) It has been proved that PMMA-based GPEs38,39,88–91 possess high ionic conductivity, are stable electrochemically and thermally, and have high transparency as well as good gelatinising and solvent retention abilities. Their



489

room temperature conductivity is of the order of 1–10 × 10−3 S cm−1 and they also have high transmission in the visible region. The small variation in conductivity over the operational temperature range has been found to be advantageous for ECD applications. GPEs, having the composition PMMA–PC/EC/γBL–LiCF3SO3 with PMMA as the polymer, PC, EC, γBL) and their binary mixtures as solvents, and lithium triflate (LiCF3SO3) as the salt, have been prepared and studied. The conductivity values at all concentrations increase in the order PC < EC < γBL, and the conductivity for the binary varies in the following order: PC < PC+ EC < PC+γBL < EC + γBL < γBL < EC Sample 1 According to the literature, the LiClO4 salt was dissolved in PC.38,39 Then 10 wt% PMMA, which was based on 1.0 M LiClO4/PC solution, was added and well stirred. Finally, the solution was slowly heated until a transparent gel was formed under an N2 atmosphere to ensure the proper function of the ion insertion and extraction. Sample 2 Liquid electrolytes were prepared by dissolving lithium imide (LiN(CF3SO2)2) in PC, EC, γBL and their binary mixtures and room temperature conductivity exceeding 10−2 S cm−1 has been obtained.88,90 The GPEs prepared by the addition of PMMA to these liquid electrolytes resulted in a large increase in viscosity without lowering the conductivity significantly. Sample 3 GPE was prepared by using of TBAFB (tetrabutylammonium tetrafluoroborate):acetonitrile (ACN) : PMMA : PC in the ratio of 3 : 70 : 7 : 20 by weight.91 After TBAFB was dissolved in ACN, PMMA was added into the solution with vigorous stirring and heating. The GPE prepared was used in ECDs. The devices were found to have good switching times, reasonable contrasts and optical memories. Sample 4 GPE was prepared by using NaClO4:LiClO4:AN : PMMA : PC in the ratio of 1.5 : 1.5 : 70 : 7 : 20 by weight.92 After dissolving NaClO4/LiClO4 in acetonitrile (AN), PMMA and PC were added into the solution with vigorous stirring and heating until the highly conducting transparent gel was obtained. The GPE obtained was used for ECDs which were found to have good switching times and optical contrast values.


490


Sample 5 Methyl methacrylate (MMA) (7 g, Mw: 120 000) was dissolved in dry acetonitrile (70 g).93 LiClO4 (3 g) was added to the solution as supporting electrolyte. Finally, PC (20 g) was added as plasticiser. The whole mixture was then slowly heated until gelation. The GPE was spread on the polymercoated side of the electrode, and the electrodes were sandwiched under atmospheric conditions. Sample 6 Liquid electrolytes comprising lithium salts of the type LiX (X = N(CF3SO2)2, ClO4) dissolved in an aprotic solvent PC and a binary mixture of PC and EC show σ25 (max) of the order of 10−2 S cm−1.94 GPEs synthesised by the incorporation of PMMA up to 25 wt% in these liquid electrolytes show insignificant changes in the σ25 value, retaining the liquid-like behaviour with σ25 (max) ~10−3 S cm−1. PMMA imparts mechanical stability to gels, thereby promoting ease of device fabrication. Polyethylene oxide (PEO)-based gel polymer electrolyte (GPE) Sample 1 The GPE was prepared in a drybox.95 The PEO (Mw: 4 × 106) solution in MeCN with LiClO4 and PC as a plasticiser were mixed in a weight ratio 1 䋺 83 䋺 0.62 䋺 0.83 and then mechanically stirred overnight to obtain good homogeneity. Films were cast and peeled. Sample 2, photopolymerisable PEO-based GPE One practice to form the polymer electrolyte layer in a ECD is solvent casting.11 A more practical approach towards preparing large area ECDs is to photopolymerise an acrylate functionalised low molecular weight PEO in the presence of a photoinitiator.37,38 The GPE was composed of a PEG-based acrylate macromonomer, photoinitiator, plasticiser, glass beads and the electrolyte to compensate for the charge injected into or extracted from the conducting polymer. The ionic conductivity of lithium–PEO has been reported to be at a maximum for a Li/O ratio of 0.04 and hence the composition of the GPE with no plasticiser was prepared with this ratio.40 The different compositions of the GPEs used in this study are shown in Table 12.1. Approximately 0.5 ml of the gel electrolyte solution was poured onto the poly (3,4-ethylenedioxythiophene (PEDOT)-coated ITO glass, the glass beads dispersed in the gel electrolyte help prevent electrical shorting. The



491

Table 12.1 Different composition of gels MA/g

DA/g

PC/g

DMPAP/mg

MA/g

DA/g

PC/g

DMPAP/mg

10 9.0 8.0 7.0 6.0 5.0

0 0 0 0 0 0

0 1.0 2.0 3.0 4.0 5.0

25.0 22.5 20.0 17.5 15.0 12.5

0 0 0 0 0 0

10 9.0 8.0 7.0 6.0 5.0

0 1.0 2.0 3.0 4.0 5.0

25.0 22.5 20.0 17.5 15.0 12.5

* 1.0 g lithium trifluoro methane sulphonate (LITRIF) and 5 mg of 50–100 µm glass beads were used in all compositions.

two electrodes were gently hand pressed together to remove any trapped air bubbles and to squeeze out excess gel electrolyte solution. Polymerisation of the macromonomer, MA or diacrylate (DA) in the presence of photoinitiator 2,2-dimethoxy-2-phenyl-acetophenone (DMPAP), was carried out under UV light (365 nm, 5.8 mW cm−2) for 15 min. The ECDs prepared have low voltage consumption, photopic contrasts of ca. 30% and switching speeds as low as 0.6 s. The photopolymerisable PEO-based GPE enhanced switching speed performance over approximately two orders of magnitude. Gel polymer electrolyte (GPE) based on poly(methyl methacrylate)/ oxymethylene-linked poly(oxyethylene) (PMMA/OMPEO) blend composite Preparation of OMPEO To 25 g KOH powder, 120 ml CH2Cl2 and 25 g PEG mixture (PEG400 : PEG600 = 2 : 3wt% were slowly added at room temperature. After the solution was stirred for 20 h, the crude product was filtered under reduced pressure to remove KCl and unreacted KOH particles. Finally, the transparent product OMPEO was dried at 60 °C in the reduced pressure vacuum for 24 h.37 Preparation of GPE (step 1) Liquid electrolytes with different concentrations were prepared on the basis of moles of lithium salt per litre of the solvent PC (xM LiClO4/PC; x = 0.4, 0.6, 0.8, 1.0, 1.2, 1.5 and 2.0). To 15 ml THF, 5 g of the mixture of OMPOE and PMMA, whose wt ratio is 1 䋺 4, and 10 ml of liquid electrolyte in different concentrations were added. The system was stirred with


492


a magnetic bar at 60 °C until homogeneous. The obtained GPE can also be used as a solid film, by using solvent evaporation techniques. The films were further dried in vacuum at 60 °C to remove any trace of THF. Preparation of CSPE (step 2) CSPE was obtained after adding nanofillers in different weight percentages to the GPE system above on the basis of step 1. •

•

A maximum value of conductivity 1.8 × 10−3 S cm−1is obtained for the PMMA (4 g)-OMPEO (1 g)–LiClO4 in PC (0.4 M, 10 ml)–SiO2 (0.8 g) CPE film at 298 K with high optical clarity in the visible region and thermal stability. CSPE prepared by dispersing nanofillers of TiO2 exhibits high ionic conductivity with good thermal stability and is used in plate display material.

Star network poly(ethylene glycol) (PEG)-based gel polymer electrolyte (GPE) The star network polymer34,35 with a pentaerythritol core linking four PEGblock polymeric arms was synthesised from PEG400, PEG600, pentaerythritol and dichloromethane. The synthetic route and sketch map of star network polymer are shown in Fig.12.7. Synthesis of the star network polymer Some 10 g PEG400, 15 g PEG600, 0.75 g pentaerythritol, 15 g KOH powder and 180 ml CH2Cl2 were stirred in a flask at 30 °C under nitrogen. An additional 5 g of KOH powder was added into the flask after 1 h, 2 h and 4 h respectively from the beginning of the reaction. Most of the unreacted CH2Cl2 was evaporated; the residue was dissolved in toluene. After KCl and unreacted KOH were removed by filtration, the crude product was precipitated out by adding petroleum ether to the solution. The crude product

CH2OH HOH2C

CH2OH + HO(CH2CH2)mOH + HO(CH2CH2)nOH CH2OH

CHCl2 KOH

C[CH2OCH2O(CH2CH2)mOCH2O(CH2CH2)nO]4

12.7 The synthetic route and sketch map of star network polymer.



493

was dissolved and precipitation was repeated twice to remove the unreacted PEG. The star network polymer product was dried under vacuum at 60 °C. Preparation of the GPE The star network polymer was dissolved adequately in acetonitrile, then LiClO4 and plasticisers of EC/PC (mass ratio = 1 䋺 1) were added to the polymeric solution and stirred vigorously at 40 °C. The acetonitrile was driven out finally under vacuum at 60 °C. The resulting polymer electrolytes were stored in a nitrogen-filled drybox. The ionic conductivity of the GPE is 1.03 × 10−4 S cm−1 at RT, and it is a promising candidate for the polymer electrolyte for all-solid state ECD. Polymer network gel polymer electrolyte (GPE) based on poly(methyl methacrylate) (PMMA)–transition metal complexes

Preparation of the GPE A mixture of MMA, PMMA resin containing polymerisation initiator (dibenzoylperoxide, 1 wt%), and PC in a suitable ratio 䋺 1.50 ml MMA, 1.00 ml PC, 0.70 g PMMA is placed in a flask and kept for 5 days at room temperature in a desiccator.96–100 The polymerisation process is then finished by warming at 90 °C. The method of preparation guarantees good mechanical properties and electrochemical stability for weeks. The gel is an elastic and odourless material; required foils can be easily cut out. The 0.2–1 M solutions of anhydrous lithium or sodium perchlorate were used as the supporting electrolytes. Properties of the GPEs The GPEs present high ionic conductivity (0.4–0.8 × 10−4 S cm−1 at room temperature) and optical transparency (over 90% in the visible part of spectrum). These properties are stable for months, allow their application in many areas, such as ECD.

12.5.3 Porous gel polymer electrolyte (PGPE) and composite solid polymer electrolyte (CSPE) Some porous inorganic materials (such as zeolite) molecular sieves have been added into PE after filling with liquid electrolytes (EC or PC) and desiccation, and this new polymer electrolyte is known as PGPE. The porous


494


particles may act as physical cross linking centres for PEO chains. PGPEs possess the advantage of both dry SPEs and GPEs, including high ion conductivity, good interfacial stability and excellent mechanical strength. An effective approach to achieve high-mechanical integrity without the cost of ionic conductivity can be attained by dispersing nano-size particles of inert inorganic fillers, which result in a CSPE. The ionic conductivities reach a maximum of 2 × 10−4 S cm−1 at room temperature, which increases about 1000 times compared with the original electrolytes. However, the conductive mechanism of this method has not yet been clarified. A mesoporous molecular sieve such as SBA-15, MCM-41 and ZSM-5 has been used to enhance the electrochemical properties of the CSPE. Nan et al.101 dipped mesoporous SiO2 (MCM-41 SiO2) into plasticiser (mass ratio of EC to PC is 1 䋺 1), so that a large amount of plasticiser was preserved in the pores of mesoporous SiO2. Then the MCM-41 with liquid plasticiser was added into PEO-LiClO4. Porous gel polymer electrolyte (PGPE) based on fumed silica Preparation of fumed silica-based PGPEs The PGPEs were synthesised by immobilising PMMA in liquid electrolytes.102–106 An appropriate amount of salt was first dissolved in the single or binary solvent to result in a 1 M liquid electrolyte. The fumed silica particles were then dispersed in 2 wt% under continuous stirring. After observing a homogeneous mixing, 15 wt% PMMA was then added slowly while heating at 55 °C until a transparent PGPE was obtained. Properties of fumed silica-based PGPE An enhancement in viscosity by nearly two orders of magnitude can be attained with increased conductivity at an optimum content by incorporation of fumed silica in PGPE. In nanocomposite PE based on LiIm [LiN(CF3SO3)2], LiBETI [LiN(SO2C2F5)2] and LiBF4 salts addition of 2 wt% SiO2 results in a slightly lower value of conductivity as compared to the pristine gel PE system, while an earlier report indicated that LiClO4 and LiTf salts showed an increase in conductivity value upon SiO2 addition. Poly(vinylidene fluoride)–hexafluoropropylene(P(VdF-HFP))-based porous gel polymer electrolyte (PGPE) Two processes for the preparation of PGPE included solution-casting (Bellcore procedure), for the preparation of P(VDF-HFP)-based PGPE,107–111 and the ‘phase inversion process’.



495

Preparation method of P(VdF-HFP)-based PGPE Step 1, P(VdF–HFP) copolymer powders were dissolved in NMP and appropriate amount of dibutyl phthalate (DBP) as the plasticiser, with vigorously stirring to form transparent viscous solution. The resulting mixture was casted on a Teflon plate and was allowed to stand for several hours in vacuo. A P(VdF–HFP)–DBP copolymer film was obtained, in which the DBP with higher boiling point as numerous microdroplets were spread in the copolymer film. Step 2, after extracting the DBP in the copolymer film with a solvent such as ether, a lot of micropores were formed in the P(VdF–HFP) copolymer film. The microporous P(VdF–HFP) copolymer film was soaked in the electrolyte solution containing lithium salts as the support electrolyte, the numerous micropores in the P(VdF–HFP) copolymer film were absorbed and filled with the electrolyte solution. Thus the P(VdF–HFP)-based PGPE films were obtained. The ‘solution-casting’ (Bellcore) procedure was expensive with a long period of processing and the quality of the P(VdF–HFP)-based PGPE films was difficult to control. Preparation of P(VdF-HFP)-based PGPE with phase inversion process The methods of phase inversion process include many phase-separating procedures, such as hot-induced deposition phase-separating, solvent-evaporate deposition phase-separating, and immersed deposition phase-separating.112–116 For example, the immersed deposition phase-separating procedure was as follows: the copolymer powder was dissolved in an adequate solvent with vigorously stirring to form transparent polymer solution, and the resulting solution was cast on a smooth plate matrix as films. The polymer solution films were then immersed into a non-solvent bath immediately. After a long period of exchange of solvent with non-solvent, the PGPE films were deposited and separated from the non-solvent bath.

Zirconium phosphate (ZP)- and antimonic acid (AA)-based composite solid polymer electrolyte (CSPE) The addition of inorganic nanopowders may increase the amorphous phase of the polymer which results in enhancement of the ionic conductivity of the CSPE.117,118 There might be a specific interaction between the polymer segments and the inorganic nanoparticles which will lead to improvement of mechanical properties. It is reported that the inorganic filler particles have diameters of less than 5 µm, which will result in a remarkable improvement of the electrochemical properties of the CSPE.117


496


It has been shown that composites consisting of hydrated inorganic oxide (antimonic acid (AA), aluminium oxide, or SiO2) nanoparticles and poly(vinyl acetate)/glycerin gel are suitable for electrochromic applications since their chemical activity and hygroscopicity are lower than those for acid-based electrolytes. The electrical conductivity of the composites is 10−3~10−4 S cm−1 at room temperature, and the optical absorption is low. Zirconium phosphate (ZP), which is a well-known proton conductor, is a promising material for producing hydrated particles. The conductivity for nanocomposites of ZP as well as AA with poly(vinyl acetate)/glycerin gel were prepared and reported. Preparation of CSPE from ZP and AA CSPE were prepared from ZP119 or AA120 gels by mixing with poly(vinyl acetate) dispersed in glycerin. The poly(vinyl acetate)/glycerin gel was in a dehydrated state. Freshly prepared gels were mixed with glycerin and a 50% suspension of poly(vinyl acetate) in water (in proportions ZP or AA 䋺 glycerin 䋺 poly(vinyl acetate) being 0.24 䋺 0.51 䋺 0.25 by weight). This mixture was then dried to remove the water. The resulting composite was a viscous substance suitable for laminating devices. Properties of the CSPE from ZP and AA The proton conductivity is 10−3 to 10−4 S cm−1 at room temperature. Thermal stability prevails up to at least 110 °C, and compatibility was found with oxide electrodes such as WO3 and NiO. These properties make the electrolytes suitable for use in solid state ECDs such as smart windows. Composite solid polymer electrolyte (CSPE) based on poly(ethylene oxide) (PEO)/siliceous hybrids The hybrid concept especially nanohybrid is well adapted to the production of advanced solid state materials presenting ion-conducting properties, with the advantage of replacing viscous liquid systems by solid or rubbery materials.121 Sample preparation Host networks of organically modified silicates (ormosils), prepared from oxyethylene chains of controlled lengths grafted onto siloxane groups through urea bridges (diureasils), have been designated as d-U(2000) and d-U(900). In agreement with traditional terminology,122 electrolytes were



497

identified using the d-U(2000)nLiClO4 notation. In this system d-U(2000) indicates the average molecular weight of the host framework and n expresses the salt content as the number of ether oxygen atoms per Li+ cation. Known amounts of lithium perchlorate were incorporated into host matrices, forming di-ureasils with compositions of 200 ≥ n ≥ 0.5. The synthesis of LiClO4-doped di-ureasils has been described in detail elsewhere.122,123 The procedure used for d-U(900)nLiClO4 involved grafting a diamine containing approximately 15.5 oxyethylene repeat units onto the ICPTES (3-isocyanatepropyltriethoxysilane) precursor, to yield the di-urea crosslinked hybrid precursor. This material was subsequently hydrolysed and condensed in the sol–gel stage of synthesis to induce the growth of the siloxane framework. Xerogels with n greater than 5 were obtained as flexible transparent films with a yellowish hue, whereas compounds with n = 1 and 0.5 were rather brittle, powdery agglomerates. Properties of the CSPE These electrolytes were obtained as amorphous films, with excellent mechanical adaptation and adhesion to the electrode surface and good electrochemical and thermal stability. These materials provide significant advantages in optical performance, cycle lifetime and durability of ECD. Composite solid polymer electrolytes (CSPE) based on poly(vinylidene fluoride)–hexafluoropropylene (PVdF-HFP)/molecular sieve SBA-15 The mesoporous molecular sieve SBA-15124 possesses the same chemical composition and structure as that of nano-SiO2, so it can be used to enhance the electrochemical properties of the CSPE, in which the Li+ ions can enter and move through the mesoporous channels much more easily than they can through microporous inorganic materials. To prepare the electrolyte an appropriate amount of mesoporous molecular sieve SBA-15 was dispersed in dimethylformamide (DMF) with vigorously stirring, then an adequate amount of PVdF–HFP powder was added and dissolved in the SBA-15/DMF suspension with further stirring. The resultant mixture was cast on a smooth and clean glass plate and was allowed to stand for several hours. After evaporating the solvent in vacuo, a PVdF–HFP/SBA-15 composite film was obtained, known as a dry film. After being further treated in vacuo to remove the solvent, the dry film was immersed into a liquid electrolyte. Activated CSPE films which contained liquid electrolyte were obtained known as wet films. The ion conductivity is ~ 0.3 × 10−3 S cm−1 at room temperature for the ratio of SBA-15 䋺 PVdF– HFP = 3 䋺 8.


498


Composite solid polymer electrolyte (CSPE) based on poly(vinylidene fluoride)–hexafluoropropylene (PVdF-HFP)/Al2O3 The electrolyte was prepared by adding an appropriate amount of nanosized Al2O3 particles and dispersing in NMP with vigorously stirring to form a Al2O3/NMP suspension.125 An adequate amount of PVdF–HFP powder and a small amount of glycerol as the plasticiser were dissolved in another amount of NMP solvent with stirring to form a transparent viscous solution. The PVdF–HFP/NMP solution was mixed with Al2O3/NMP suspension with further vigorously stirring. The resulting mixture was cast on a smooth and clean glass plate and was cooled to RT to form a film, then the film was immersed into a deioned water for 2 h to remove the solvent NMP and plasticiser, and further dried at 80 °C in vacuo. A dry film of PVdF–HFP/ Al2O3 was obtained. The resulting dry film of PVdF-HFP/Al2O3 was immersed in a liquid electrolytes containing LiPF6/EC + PC. A film of activated CSPE PVdF–HFP/Al2O3 /LiPF6/EC + PC was obtained. The ion conductivity is about 1.47 × 10−3 S cm−1 at RT to give porosity of the resulting CSPE films with 6 wt% of Al2O3 nanoparticles. This is 40% higher than that of PE without Al2O3 nanoparticles.

12.5.4 Polymer electrolytes based on natural polymer matrix Natural polymers, because of their biodegrability, low cost and good physical and chemical properties, are usually used in the cosmetic, pharmaceutical and food industries but have also recently been explored to prepare SPE. Some of them, for instance chitosan,126 gelatin,127 hydroxyethylcellulose128 and modified starch,129 have achieved Li+ ion conductivity values as high as 10−5 S cm−1 at RT. Small ECDs with good electrochromic properties have been realised with such SPE,130,131 the natural polysaccharides such as cellulose and starch derivatives can be modified by a plasticisation process with glycerol, and after addition of lithium salts, transparent films with high ionic conductivity can be obtained.132–134 Samples of polymer electrolytes based on starch derivatives GPEs were prepared using amylopectin rich starch, glycerol (30% of the starch mass) and lithium perchlorate with [O]/[Li] = 10.134–136 Further details of sample preparation and characterisation are given elsewhere.135,136 Amylopectin-rich starch was dispersed in water (2% w/v) and heated at 100 °C. The solution was cooled at RT and the glycerol was added. The solid elastomeric electrolytes were obtained by the introduction of LiClO4 to the



499

desired n = [O]/[Li] ratio (6 ~ 10) which was calculated considering all the starch and glycerol oxygen. The viscous solution was dispersed on a PTFE plate and dried at 40 °C. A transparent film was obtained. The conductivity of the gel electrolytes increases with increasing temperature, reaching values of 2 × 10−3 S cm−1 for n = 6 and 0.3 × 10−3 S cm−1 for n = 8 at 355 K. Gelatin-based solid polymer electrolytes (SPE) Gelatin–lithium-based electrolytes are promising materials to be used in ECD because the material is available in nature, cheap, and easy to handle and prepare.137–139 Sample 1 Some 2 g of commercial uncoloured gelatin was dispersed in 15 ml of water and heated with stirring for a few minutes up to 50 °C for complete dissolution.137 Next, 1.25 g of glycerol as plasticiser, 0.25 g of formaldehyde and different quantities of acetic acid (0.75–2 g; 17.5–36 wt%) were added with stirring. This viscous solution was then cooled to 30 °C and poured on Petri plates to form transparent films. The ionic conductivity values are found to increase from 4.5 × 10−5 to 3.6 × 10−4 S cm−1 from room temperature to 80 °C, and the films are mechanically stable. The amount of acetic acid was found to influence the proton conduction and a high ionic conduction value of 6.3 × 10−4 S cm−1 at 80 °C was obtained with 26.3 wt% of acetic acid content. Sample 2 Some 2 g of commercial uncoloured gelatin was dispersed in 15 ml of water and heated with stirring up to 50 °C to complete dissolution.138,139 Then, 0.3 g of LiClO4, 0.5 g of glycerol as plasticiser and 0.25 g of formaldehyde were added with stirring. This viscous solution was then cooled to 30 °C and poured on Petri plates to form transparent films or injected into the windows. The samples were predominantly amorphous, with ionic conductivity values increasing according to an Arrhenius law as a function of temperature from 1.5 × 10−5 S cm−1 at RT to 4.95 × 10−4 S cm−1 at 80 °C. Solid polymer electrolytes (SPE) based on chitosan To 1 g of chitosan dissolved in 100 ml 1% acetic acid solution, LiCF3SO3 and EC were added.126 After complete dissolution of the chitosan, salt and plasticiser, the solutions were cast in Petri dishes and left to dry at room


500


temperature to form films of (a) chitosan acetate (CA) (without salt and plasticiser), (b) CA–EC, (c) CA–LiCF3SO3 and (d) CA–EC–LiCF3SO3. The films were then placed in a desiccator. The highest conductivity in plasticised film containing salt is 4 × 10−5 S cm−1 and for the plasticised films without salt is 3 × 10−10 S cm−1.

12.6

Proton-conducting polymer electrolytes and alkaline polymer electrolytes

The relatively low conductivity of SPE at RT remains an impediment for practical application. Thus, more and more efforts have been made to improve the proton conductivity and alkaline SPE. Many proton conducting membranes based on blends of polymer and inorganic acids such as H2SO4 and H3PO4 have exhibited excellent proton conductivity.140,141 However, some of the most widely studied proton conductors based on polybenzimidazole and polyimide are not suitable for ECD due to their opaqueness, although poly(vinyl alcohol) (PVA) blended with inorganic acids (such as H2SO4 and H3PO4) has excellent proton conductivity and optical properties. Nevertheless, the high chemical activity and corrosiveness of these acids may be a drawback for their practical applications in ECD. So anhydrous proton conductors based on PVA/NH4H2PO4,142 PVA/ imidazole/NH4H2PO4143 and PEO/PAA composite which contains a less strong acidic moiety, have been studied. The polymer–lithium salt system seems to be the most widely studied polymer electrolyte, due to its potential application in lithium high density batteries and ECD. Such electrolytes should be completely anhydrous and therefore, must be prepared and kept under moisture-free conditions. While gel-type electrolytes based on the PEO–KOH144–146 and PVA–KOH– H2O147,148 alkaline SPE systems have been reported, such alkaline electrolytes are interesting from the point of view of their potential application in all-solid alkaline rechargeable batteries and ECD.

12.6.1 Examples of proton-conducting polymer electrolytes Proton-conducting polymer electrolytes based on heteropolyacids (HPA) encapsulated with polystyrene sulphonic acid (PSS) Heteropolyacids (HPA) have been known to show a high proton conductivity under high humidity conditions.149,150 However, their protonic conductivity is highly sensitive to humidity and temperature, especially at high temperatures (>100 °C), and the proton conductivity of H3PW12O40 (12phosphotungstic acid, PWA) decreased by the evaporation of hydration water which limits their applications. The composite material of PWA and



501

polystyrene sulphonic acid (PSS) constructs the PWA-encapsulated material by the self-assembly of —SO3H onto the PWA surface. As a result, the fast proton transfer occurred at the interface between the PWA and — SO3H, and the encapsulated material indicated the high anhydrous proton conductivity of 1 × 10−2 S cm−1 at 180 °C. Preparation of the polymer electrolyte PWA was dissolved in pure water (100 mg ml−1), and this PWA solution was added to a PSS (MW = 7 × 104, 30% in water) solution and stirred for 12 h at RT. This PWA–PSS mixed solution was cast onto the Teflon plate, dried at 70 °C, and then stripped from the plate. The mixing ratio was controlled by the additional amount of PWA solution. Properties of the polymer electrolytes In the PSS matrix, PWA has been encapsulated by —SO3H, and its —SO3H has been formed by a self-assembled structure on the PWA surface. The mixing ratio of PWA in PSS should be 8.5 wt%, and maximum anhydrous proton conductivity can be obtained. Proton-conducting gel polymer electrolyte (GPE) based on poly(vinylidene fluoride)–hexafluoropropylene (PVdF-HPA) or poly(methyl methacrylate–heteropolyacid) (PMMA–HPA) Novel proton-conducting gels have been prepared by entrapping HPA solutions in polar aprotic solvents within the polymer matrix.151 Preparation of GPE PVdF-based gels obtained by direct dissolution of polymer and HPA include phosphotungstic acid (H3PW12O40·xH2O), phosphomolybdic acid (H3PMo12O40·xH2O) and silicotungstic acid (H4SiW12O40·xH2O) in DMF. The calculated amounts of acid were dissolved in DMF, and PVdF was added in 18–22 wt% with respect to DMF. The systems were heated to 50–60 °C with stirring until a clear solution was formed, then cooled and placed in a desiccator, and left for 1 week to allow for gelation. PMMA-based gels: MMA, cross-linking agent triethyleneglycol dimethacrylate (TGEDM 3%), and benzoyl peroxide (BP) were dissolved in PC and mixed. The solution was then heated to 75–80 °C for 24 h to allow the polymerisation. After cooling, the resultant gel was soaked with the solution of the selected acid in DMF and left in a desiccator for 1–2 weeks. Gel electrolytes were obtained for the following components range: 12.5 wt%


502


polymer and 87.5 wt% of the solution of HPA in a solvent mixture consisting of 1 part of DMF and 4 parts of PC with the HPA concentration in the solutions equal to 5 and 15 wt%. Properties of GPE The use of PMMA as a polymer matrix allows us to prepare conducting gels in a wider composition range with better mechanical properties and enhanced transparency than PVdF. Among the HPAs tested as dopants, SiWA is the best prospective material for potential applications because of its ability to prepare colourless and transparent gels even at higher acid concentrations, whereas PWA and phosphomolybdic acid (PMoA)-doped gels become gradually darker. In comparison to the electrolytes doped with H3PO4,140,141,152 the described systems exhibit higher conductivities up to 4 × 10−3 S cm−1 at 25 °C, and in the much wider temperature range from 0.5 mS cm−1 at −60 °C to 8 × 10−3 S cm−1 at 90 °C making the gels promising materials for application in ECD. Proton-conducting gel polymer electrolyte (GPE) based on poly(2acryloamide-2-methyl-1-propanesulphonic acid)/poly(vinylidene fluoride) (PAMPSA/PVdF) or poly(methyl methacrylate) (PMMA) Polymer gels can be divided into two groups based on the type of solvent used: hydrogels (e.g. water as the solvent) and non-aqueous gels in which the solution of an acid in a highly polar organic solvent is used. The latter have mostly been investigated owing to the possibility of application in ECD.153 Most of the systems studied exhibit sufficient ionic conductivities at ambient or moderate temperatures. A new type of proton-conducting GPE based on poly(2-acryloamido-2-methyl-1-propanesulphonic acid) (PAMPSA) which exhibits subambient temperature ionic conductivities higher than 10−4 S cm−1 have been developed. Preparation of the GPE Electrolytes based on PAMPSA were obtained in the free radical polymerisation of AMPSA. The electrolytes thus obtained are solutions with viscosity dependent on the polymer concentration. For example, for a system containing 20 wt% of PAMPSA, the viscosity was 2 Pa s. Properties of GPE It has been demonstrated that non-aqueous gel electrolytes exhibit ambient temperature conductivities exceeding 10−3 S cm−1, which decreases with a



503

decrease in temperature to the range 10−5–10−4 S cm−1 at −50 °C depending on the electrolyte composition. The electrolytes exhibit electrochemical stability sufficient for application in ECD. Proton-conducting gel polymer electrolyte (GPE) doped with 12-PWA Proton-conducting PEO-based GPE were prepared according to appropriate ratio of PEO䋺 12-WPA, by adding the solution of 12-WPA in ethanol into the solution of PEO in acetonitrile with vigorous stirring until homogeneous.154 The PEO–PWA mixed gel solutions were cast onto the Teflon plate and dried in a vacuum oven to remove the solvent. GPE films were then stripped from the plate. The conductivity of PEO–12-PWA reached 4.0 × 10−3 S cm−1 with a molar ratio [H+]/[EO] of 0.025 and relative humidity 95%. Proton-conducting gel polymer electrolyte (GPE) doped with H3PO4 The combination of the good chemical and electrochemical stability and the humidity resistance of PMMA–PC systems with the high conductivity of PGMA–DMF electrolytes, which was realised by the glycolyl methacrylate (GMA)-MMA copolymers, prepared by in situ copolymerization in DMF, N,N-dimethylacetamide (DMA) and DMF–PC mixtures in the presence of H3PO4.152 Preparation of the GPE H3PO4 was dissolved in a solvent (DMF, DMA or PC or a mixture of these). Monomers and a free radical initiator (benzoyl peroxide) were then added with stirring to obtain a homogeneous solution in a nitrogen-filled drybox (moisture < 2 ppm). Each mixture was heated at 80 °C to form a gel.152,155,156 Properties of the GPE The properties of the electrolytes can be modified by changes in the composition of the polymer matrix and accordingly the plasticiser used and fraction of H3PO4. Utilisation of GMA–MMA copolymers leads to an increase in the conductivity, up to ~10−3 S cm−1 at RT, compared with that of pure PMMA-based systems as well as to an improvement in the mechanical properties and adhesion to the glass substrate. Proton-conducting polymer electrolyte based on poly(vinyl alcohol) (PVA)/(imidazole)/ NH4H2PO4 Proton-conducting systems seem to be particularly suitable for ECD applications because of the high mobility of protons, resulting in high


504


conductivity at ambient temperature. Many proton conducting membranes based on PVA blended with inorganic compounds or acids such as H2SO4, H3PO4 and NH4H2PO4142,143 possess excellent proton conductivity and optical properties which can be used in ECD smart windows. Sample polymer electrolyte based on PVA/NH4H2PO4 The desired amount of PVA solution was mixed with a suitable amount of NH4H2PO4 and stirred at 70 °C until an aqueous solution was formed. This was then cast onto a PTFE mould.142 After being evaporated slowly at RT, the membranes were dried completely in a vacuum oven to remove the water moieties. The PVA/xNH4H2PO4 composite membrane was thus obtained, where x is the molar ratio of NH4+ to a repeat unit of PVA. The proton conductivity of the composite membranes increases with rising temperature. It also increases with rising phosphate doping-level at first and then decreases with increasing phosphate content after a certain value of x. The highest proton conductivity is near the area of x = 0.067. Sample polymer electrolyte based on PVA/Imi/ NH4H2PO4 Predetermined amounts of NH4H2PO4 and imidazole (Imi) were dissolved in water and stirred thoroughly at 80 °C until they form aqueous imidazolium solution.143 The molar ratios of Imi/NH4H2PO4 are 10/1, 1/1 and 1/10, respectively. A desired amount of PVA was dissolved in the boiling water and then cooled to 70 °C, then a stoichiometric amount of imidazolium solution was added and the resulting mixture was stirred to make it homogeeous. The solution was cast onto a PTFE mould. Then the samples were dried in a vacuum oven until completely dry. The addition of imidazole with a suitable molar ratio of Imi/NH4H2PO4 can improve the proton conductivity of PVA/Imi/NH4H2PO4 composite membrane.

12.6.2 Samples of alkaline polymer electrolytes Alkaline polymer electrolytes have had particular advantages, such as easy preparation, low cost, abundance of the raw material and high ion conductivity at RT. They have potential application value in alkaline secondary batteries, supercapacitors and ECD. Alkaline polymer electrolyte based on poly(vinyl alcohol) (PVA)–KOH–H2O A PVA suspension in water was left overnight at 50 °C and the highly viscous solution was mixed with concentrated KOH aqueous solution and



505

stirred. The resulting homogeneous solution was poured onto a Teflon plate and the excess of the water was evaporated slowly.147,148 The most hightly conducting foils were composed of ca. 40 wt% of PVA, 25–30 wt% of KOH and 30–35 wt% of water. Typical conductivity of such foils reached the level of nearly 10−3 S cm−1 at room temperature.

Alkaline polymer electrolytes based on poly(vinyl alcohol) (PVA)– carboxymethyl cellulose (CMC)–KOH–H2O Alkaline polymer electrolytes based on PVA–KOH–H2O possess about 10−2 S cm−1 high ion conductivity at RT, but the conductivity of the PE films of PVA–KOH–H2O were readily decreased by deep dehydration of the a electrolytes. A new composite alkaline PE reported was based on PVA–carboxymethyl cellulose (CMC)–KOH–H2O.157 Predetermined amounts of PVA and sodium CMC were dissolved in water and heated to 70 °C overnight, and the viscous solution of PVA and CMC in water was mixed with concentrated KOH aqueous solution and stirred until homogeneous. The resulting viscous solution was poured onto a Teflon plate and the excess of water was evaporated slowly. The composition of the electrolyte film was calculated from the mass balance. The alkaline polymer electrolytes exhibited a high ionic conductivity, typically 10−2 S cm−1 with an electrochemical stability window of about 1.6 V on stainless steel blocking electrodes.

Alkaline composite polymer electrolyte (CPE) based on poly(ethylene oxide) (PEO)–KOH–nanopowders In order to enhance the ionic conductivity of PEO–KOH-based alkaline PE, three types of ceramic nanopowders, TiO2, β-Al2O3 and SiO2 were added, and the corresponding alkaline PE containing nanopowders were prepared.158 According to a predetermined composition, the appropriate amount of PEO (Mw 100 000) was mixed with nanosized TiO2 (~10 nm) or β-Al2O3(10~20 nm) or SiO2 (~50 nm) particles respectively, then dissolved and dispersed in a KOH aqueous solution and stirred until homogeneous. The resulting viscous solution was poured onto a Teflon plate and then the excess of the solvent water was evaporated slowly. The prepared polymer electrolyte exhibited a high ionic conductivity at room temperature, typically 10−3 S cm−1, and, moreover, good electrochemical stability. The potential stability window is of ca. 1.6 V on stainless steel blocking electrodes.


506


12.7

New type of polymer electrolyte

12.7.1 Polymer electrolyte with ionic liquids The breakthrough of commercial applications of the ECDs has been limited by weaknesses such as long switching time, insufficient cyclability and long-term stability. To bypass some of these problems, a significant part of the research needed to be focused on the usage of a new type of electrolyte.159–162 Ionic liquids are organic salts with a low melting point (2000 cycles).

Polymer electrolyte based on 1-[3-(trimethoxy-λ4silyl)propyl]imidazole (TMSPIm) Research results revealed that fast switching times and excellent stability for up to a million cycles of such as ECD could be obtained.160 The question of how to combine the excellent properties of ionic liquids with sol–gel precursors capable of network formation remained.169 A proton-conducting gel electrolyte was prepared with the addition of 1-butyl-3-methyl-imidazolium-tetrafluoroborate ionic liquid to a sol–gel precursor methyltrimethoxysilane (MTMOS) in ethanol. The addition of an ionic liquid influenced the morphology of the electrolyte and prolonged the gelation time. Unlike conventional silica, the final electrolyte did not shrink with time. The conductivity of this electrolyte reached 1.2 × 10−3 S cm−1 and was slightly lower than the conductivity of the MTMOS/H2O/H3PO4 controlling compound (9 × 10−3 S cm−1). Synthesis of the electrolyte precursor TMSPIm 1-[3-(Trimethoxy-λ4-silyl)propyl]imidazole (TMSPIm) has propyltrimethoxy-silane groups bound on the imidazolium ring, enabling condensation to a siloxane network. TMSPIm-based electrolytes were produced in a manner similar to the ionic liquid 1-methyl-3-[3-(trimethoxy-λ4-silyl) propyl]imidazolium iodide (MTMSPImI), which was synthesised as a potential quasi-solid state redox electrolyte. Preparation of the electrolytes Three kinds of electrolytes were prepared from the sol–gel precursor TMSPIm. TMSPIm was synthesised from imidazole (Fig. 12.8), which was

OMe MeO

Si

+

OMe Cl

MeO

H N

Na—OMe MeO N

Si OMe

12.8 Synthesis of the 1-[3-(trimethoxy-λ4-silyl)propyl]imidazole (TMSPIm) precursor.


N N

508


dissolved in methanol. NaOMe was added slowly with stirring, and then refluxed for about half an hour. 3-Chloropropyltrimethoxysilane was then added dropwise, the mixture immediately turned white, and further refluxed for 12 h. The mixture was finally filtered and the product obtained, by removal of methanol under reduced pressure and elevated temperature. The product – TMSPIm – was a yellowish, non- viscous liquid. Electrolytes I–III were prepared from TMSPIm precursor by adding: (I) trifluoroacetic acid (TFA) or (II, III) acetic acid (AcOH) in a molar ratio of 1 䋺 5.5. Equivalents of 4.5 were used for solvolysis, and 1 equivalent served for protonation for formation of ionic liquid. In electrolyte III, a mixture of acetic anhydride as dehydrating agent, and lithium acetate dihydrate as a source of lithium ions were added. Solvolysis and condensation reactions of trimethoxysilanes were stimulated by heat treatment of the mixtures at 120 °C. Lastly, the product was heated under reduced pressure to remove the remaining volatile components from the electrolytes. TMSPIm + 5.5 CF3COOH → I TMSPIm + 5.5 CH3COOH → II TMSPIm + 5.5 CF3COOH + (CH3CO)2O + LiOAc → III Properties of the electrolytes The presence of free trifluoroacetate anions contributed to the moderately higher conductivity of this electrolyte (4.6 × 10−5 S cm−1) compared with that of acetic acid (1.6 × 10−5 S cm−1). The conductivity of the electrolytes could be further increased by the addition of a lithium salt. All electrolytes were employed in ECD. Polymer electrolyte based on ionic liquids and polymeric ionic liquids Two main strategies have been pursued by the scientific community in an attempt to translate the benefits of ionic liquids to polymer electrolytes.161 The first involves the design of a polymer electrolyte composed of conventional polymer matrices and ionic liquids, by free radical crosslinking of certain vinyl monomers in ionic liquids, resulting in mechanically strong and highly conductive polymer electrolyte films.164,170–172 The second consists of designing functional polymers presenting some of the characteristics of ionic liquids, with preparation of different types of polymeric ionic liquids, as a way of developing high performance polymer electrolytes.173,174 Synthesis of the ionic liquids and polymeric ionic liquids The ionic liquids [bmim][Br−], [bmim][BF4−], [bmim][Tf2N−] and [bmim] [PF6−] were prepared or purchased and used as received.175 The correspond-



509

ing polymeric ionic liquids, poly(1-vinyl-3-methylimidazolium bromide) poly [ViEtIm] [Br−], poly(1-vinyl-3-methylimidazolium tetrafluoroborate poly[ViEtIm][BF4−] and poly (1-vinyl-3-methylimidazolium bis(trifluoromethane sulphonimide) poly[ViEtIm] [Tf2N−], were synthesised.173,174,176 polymeric ionic liquids + ionic liquids → polymer electrolytes poly [ViEtIm] [Tf2N–] + [bmim] [Tf2N–] → family 1 poly [ViEtIm] [BF4–] + [bmim] [BF4–] → family 2 poly [ViEtIm] [Br–] + [bmim] [Br–] → family 3

[

] N +

N [ViEtIm]

In order to achieve good chemical compatibility between the compounds, both members of each family, ionic liquids and polymeric ionic liquids, have the anion in common; family 1 is composed of [bmim][Tf2N−] and poly[ViEtIm] [Tf2N−], family 2 of [bmim][BF4−] and poly[ViEtIm][BF4−] and family 3 of [bmim][Br−] and poly[ViEtIm] [Br−]. Owing to the chemical affinity between the ionic liquid and the polymeric ionic liquid matrices, the polymer electrolytes obtained behave like a homogeneous phase and none of the electrolytes exhibited phase separation over periods of days to weeks. Preparation of polymer electrolytes Polymer electrolytes consisting of different ratios of ionic/polymeric ionic liquids (0 䋺 100, 25 䋺 75, 35 䋺 65, 50 䋺 50, 75 䋺 25 and 100 䋺 0) were prepared by dissolving each polymeric ionic liquid in its corresponding ionic liquid; poly[ViEtIm][Br−] in [bmim][Br−], poly[ViEtIm][BF4−] in [bmim][BF4−] and poly[ViEtIm][Tf2N−] in [bmim][Tf2N−]. THF was used as a co-solvent when the amount of ionic liquid in the mixture was not enough to dissolve the polymer. After mixing, the polymer electrolytes were dried under vacuum. Ionic conductivity Conductivities of pure ionic liquids varied depending on the anions (3.98 for [bmim][Tf2N−], 3.55 for [bmim][BF4−] and 0.25 for [bmim][Br−] × 10−3 S cm−1 at room temperature). Ionic conductivities of all synthesized PEs were in the range 10−2–10−5 S cm−1 at room temperature. The conductivity decreases with increasing polymer content in the three families, being typical behaviour of other polymer electrolytes composed of ionic liquids and polymers. At high ionic liquid concentrations, the ionic conductivity reached values near to the values of the ionic liquids although the mechanical stability is compromised. The lowest ionic conductivity of polymer electrolytes is the one with bromide anion (family 3).176


510


Finally, it was demonstrated that cycle life of ECDs is significantly enhanced (up to 70 000 cycles) when this new type of polymer electrolyte was used. Polymer electrolytes based on incorporating ionic liquids into poly(vinylidene fluoride)–hexafluoropropylene (P(VdF-HFP)) polymer Two kinds of GPE were reported by incorporating ionic liquids 1-(2-hydroxyethyl)-3-methyl imidazolium tetrafluoroborate ([HEMIm] [BF4−]) and [HEMIm][PF6−] into P(VdF–HFP) polymer for applications in SPEs.177 Preparation of ionic liquids To synthesise [HEMIm][Cl−], 1-methyl-imidazole was reacted with an excess of 2-chloroethanol. The molten salt of the white crystalline solids was obtained by recrystallisation. For [HEMIm][BF4−], the [HEMIm][Cl−] reacted with sodium tetrafluoroborate; [HEMIm][PF6−] was synthesised in the same manner as for [HEMIm][BF4−]. Preparation of P(VdF–HFP)-ionic liquid gel Some 0.528 g P(VdF–HFP), 0.587 g ionic liquid and 2.5 ml DMAC (N,Ndimethylacetamide) were mixed to produce a transparent solution, under ambient conditions, in which two kinds of ionic liquid, [HEMIm][BF4−] and [HEMIm][PF6−], were used. The final eight gel samples on the two prepared ionic liquids were made to contain 33.3, 47.4, 66.7 and 100 wt% P(VdF– HFP) polymer, respectively. The transparent solution gelled in 5 min when deposited in glass Petri dishes that were placed on a hot plate preheated to 80 °C. In terms of appearance, the films with a P(VdF–HFP) of more than 66.7 wt% appeared opaque, whereas those with a P(VdF–HFP) of less than 47.4 wt% exhibited rubbery white gels. To remove residual solvent DMAC, the films were placed in ambient air for sufficient time. Ionic conductivity The ionic conductivity of the gels based on [HEMIm][BF4−] and [HEMIm] [PF6−] was 10−4–10−5 S cm−1 in a temperature range of 20–70 °C, whereas the ionic conductivity of neat [HEMIm][BF4−] and neat [HEMIm][PF6−] indicated high values of 4.6 and 2.1 × 10−3 S cm−1, respectively, at room temperature.178 Particularly, in the 66.6 wt% of P(VdF–HFP), the ionic



511

conductivity of the P(VdF–HFP) [HEMIm][BF4−] gel was 10−1 lower than that of the P(VdF–HFP) [HEMIm][PF6−] gel; the ionic conductivity of the neat [HEMIm][BF4−] is higher than that of neat [HEMIm][PF6−].

12.7.2 Single ionic polymer electrolyte Single-ion conductors have advantages over typical biion-based SPEs.178–184 During discharge in biion salt-based SPE, mobile anions and cations migrate toward the oppositely charged electrodes, thereby polarising the electrolyte and increasing its resistivity. Recharging the cell then requires more energy, time, and greater electrochemical potential. This ‘cell polarisation’ problem is unique to biionic salt-based SPE. Polymer electrolytes based on poly(2-acrylamido-2-methyl-1-propane sulphonic acid) (polyAMPS) and its copolymer have been widely used for the synthesis of both proton- and lithium-conducting hydrogels or polyelectrolytes.179 In order to achieve high conductivity, low molecular weight polar solvents such as DMA, dimethylsulphoxide (DMSO), EC, PC and ionic liquids are usually added. All dry ECD or ethylene carbonate (EC) windows using gel electrolytes based on polyAMPS were obtained.152,178,180,181 The conductivities of these non-aqueous systems depended strongly on the type of solvent used; the highest conductivities were obtained when methanol was used (1.2 × 10−1 S cm−1 at RT). Gels plasticised with DMA and DMF or DMF/PC mixtures exhibit much lower conductivities (up to 1.8 × 10−3 S cm−1). These results are consistent with those presented for gels doped with H3PO4 or its esters.152,182,183 The lower viscosity of DMF compared with PC as well as the protophilic properties of this solvent seem to be the main factor affecting the conductivity of the electrolytes swollen with DMF or its mixtures. Single ionic polymer electrolyte based on 2-acrylamide-2-methyl-1-propane sulphonic acid (AMPS) copolymer Proton-conducting GPE suitable for ECD applications were prepared by the co-polymerisation AMPS and ethylene glycol methacrylate phosphate (EGMP) either with MMA, GMA or 2-hydroxyl ethyl methacrylate (HEMA) in polar aprotic solvents, such as PC, EC or DMF.179 Use of the copolymers instead of polyAMPSA allowed the preparation electrolytes with various polymer matrix contents and various concentration of AMPS or EGMP. Gels swollen with solvent mixtures containing more than 40 wt% PC were completely amorphous. This property is important for the electrochromic applications; serious crystallization problems which severely decrease ionic conductivity are avoided.


512


Sample preparation The monomers were dissolved in DMF, or in PC–DMF or in PC–EC–DMF solvent mixtures, a crosslinking agent and a free radical initiator (BP) were then added and the mixture was stirred to obtain a homogeneous solution in a nitrogen-filled drybox. Each of the mixtures was heated at 80 °C to form a gel. The resulting gels contained 15–30 wt% polymer matrix and 5–30 wt% AMPS or EGMP with about 95% conversion. Gels obtained using MMA as comonomer were colourless and transparent, whereas those based on GMA copolymer were usually yellow. Properties of the GPE The properties of the GPE are affected by the polymer matrix content, type and concentration of the acidic component, i.e. AMPS or EGMP as well as the type of solvent used as plasticiser. The highest conductivities reached 1.5 and 0.67 × 10−3 S cm−1 respectively at room temperature for AMPSAand EGMP-based electrolytes. All EGMP-based systems had lower conductivities than AMPS-based systems. The use of PC–DMF, EC–DMF or PC–EC–DMF solvent mixtures, on the other hand, improves adhesion to a glass substrate and resistance to the influence of traces of water. It is found that the poly(AMPSA-co-MMA)–PC–DMF (30 wt% polymer) and poly(AMPS-co-HEMA)–PC–EC–DMF (20 wt% polymer) systems are the most promising for ECD applications. Single ionic polymer electrolyte based on K+-doped 2-acrylamide-2-methyl1-propane sulphonic acid (AMPS) copolymer A K+-doped SPE, KCl-saturated polyAMPS, was employed to fabricate an all-solid state ECD.185,186 It was verified that the sulphonic matrix of polyAMPS could also accommodate the conduction of K+. A hybrid K+/H+ SPE was successfully applied to a precolouring-free ECD based on PB (Prussian blue)–InHCF (indium hexacyanoferrate) and also to PB–WO3 and InHCF– WO3 ECD. Preparation of KCl-doped polyAMPS-based SPE Some 9.4 g of AMPS monomer was dissolved into 10 ml of deionised water, and 0.3 ml of tetra(ethylene glycol diacrylate) (TEGDA), the crosslinking agent, was dropped into the solution.185 A selected amount of KCl was added to obtain the desired KCl doping level. Then the solution was stirred for 15 min at room temperature and 0.01 g of BME, the initiator, was dissolved into the solution with another 15 min of stirring. A KCl- doped



513

AMPS monomer solution ready for UV curing was prepared, which was filled in a rectangular casting mould, and was exposed to UV irradiation with an energy intensity of ca. 675 µW cm−2 (Spectroline, SB-125) for 3 min. Four monomer solutions with different KCl concentrations, corresponding to four different KCl doping levels – (a) 0.546, (b) 0.220, (c) 0.022 and (d) 0.002, respectively – based on the molar ratio of (KCl)/(AMPS) were prepared. It was found that the widest transmittance window, fastest response and highest coloration efficiency are obtained for (KCl)/(AMPS) = 0.22, while the saturated doping level leads to the longest cycle life. K+-PAMPS-based SPE The ECD of PANI/K-PAMPS/InHCF was assembled in a sandwich configuration with a dominant electrochromic contribution of PANI.181 The K+PAMPS electrolytes were prepared by UV polymerisation, and this procedure was modified from those reported by a US patent,187 as described in ‘Preparation of KCl-doped polyAMPS-based SPE’ above. The ionic conductivity of the K+-PAMPS electrolyte was ca. 0.13 ± 0.01 S cm−1.188

12.8

References

1. kobayashi, n., nishimura, m., ohtomo, h., Polymer electrolyte with large temperature-dependent conductivity for novel electrochromicimaging, Electrochim. Acta, 2005, 50, 3886–90. 2. fenton, d.e., parker, j.m., wright, p.v., Complexes of alkali metal ions with poly(ethylene oxide), Polymer, 1973, 14, 589. 3. carpi, f., de rossi, d., Colours from electroactive polymers: electrochromic, electroluminescent and laser devices based on organic materials, Optics Laser Technol., 2006, 38, 292–305. 4. shaw, j., seidler, p., Organic electronics: introduction, IBM J. Res. Dev., 2001. 5. shirota, y., Organic materials for electronic and optoelectronic devices, J. Mater. Chem., 2000, 10(1), 1–25. 6. deb, s., Novel electrophotographic system, Appl. Opt., 1969, 3, 192–5. 7. rowley, n., mortimer, r., New electrochromic materials, Sci. Prog., 2002, 85(3), 243–62. 8. granqvist, c., azens, a., hjelm, a., kullman, l., niklasson, g.d., et al., Recent advances in electrochromics for smart windows applications, Sol Energy, 1998, 63(4), 199–216. 9. mortimer, r., Organic electrochromic materials, Electrochim. Acta, 1999, 44, 2971–81. 10. sapp, s., sotzing, g., reynolds, j.r., High contrast ratio and fastswitching dual polymer electrochromic devices, J. Chem. Mater., 1998, 10(8), 2101–8. 11. seshadri, v., padilla, j., bircan, h., radmard, b., draper, r., et al., Optimization, preparation, and electrical short evaluation for 30 cm2 active area dual conjugated polymer electrochromic windows, Organic Electronics, 2007, 8, 367–81.


514


12. argun, a.a., aubert, p.h., thompson, b.c., schwendeman, i., et al., Multicolored electrochromism in polymers: structures and devices. Chem. Mater., 2004, 16, 4401–12. 13. granqvist, c.g., Electrochromic tungsten oxide films: review of progress 1993– 1998, Solar Energy Mater. Solar Cells, 2000, 60, 201–62. 14. hyodo, k., Electrochromism of conducting polymers, Electrochim. Acta, 1994, 39, 265–72. 15. mortimer, r.j., Organic electrochromic materials, Electrochim. Acta, 1999, 44, 2971–81. 16. rosseinsky, d.r., mortimer, r.j., Electrochromic systems and the prospects for devices, Advanced Mater., 2001, 13, 783–93. 17. somani, p.r., radhakrishnan, s., Electrochromic materials and devices: present and future, Mater. Chem. Phys., 2003, 77, 117–33. 18. ribeiro, a.s., machado, d.a., filho, p.f.d., de paoli, m.a., Solid-state electrochromic device based on two poly(thiophene) derivatives, J. Electroanal. Chem. 2004, 567, 243–8. 19. rocco, a.m., depaoli, m.a., zanelli, a., mastragostino, m., An electrochromic device combining polypyrrole and WO3 – I. Liquid electrolyte, Electrochim. Acta, 1996, 41, 2805–816. 20. sapp, s.a., sotzing, g.a., reddinger, j.l., reynolds, j.r., Rapid switching solid state electrochromic devices based on complementary conducting polymer films, Adv. Mater., 1996, 8, 808–11. 21. sapp, s.a., sotzing, g.a., reynolds, j.r., High contrast ratio and fast-switching dual polymer electrochromic devices, Chem. Mater., 1998, 10, 2101–8. 22. lu, w., fadeev, a., qi, b., smela, e., mattes, b., ding, j., spinks, g., mazurkiewicz, j., zhou, d., wallace, g., macfarlane, d., forsyth, s., forsyth, m., Use of ionic liquids for p-conjugated polymer electrochemical devices, Science, 2002, 297, 983–7. 23. wright, p., Polymer electrolytes – the early days, Electrochim. Acta, 1998, 43(10–11), 1137–43. 24. tiyapiboonchaiya, c., macfarlane, d., sun, j., forsyth, m., Polymer-in-ionicliquid electrolytes, Macromol. Chem. Phys., 2002, 203(13), 1906–11. 25. tiyapiboonchaiya, c., pringle, j., macfarlane, d., forsyth, m., Polymer-in-ionicliquid electrolytes, Macromol. Chem. Phys., 2003, 204(17), 2147–54. 26. lehn, j.m., Supramolecular chemistry – concepts and perspectives, Wiley, Weinheim, 1995. 27. ding, l., dong, s., wang, e., Progress in research of the solid polymer electrolytes, Electrochemistry (in Chinese), 1997, 3(4), 349–61. 28. quartarone, e., mustarelli, p., magistris, a., PEO-based composite polymer electrolytes, Solid State Ionics, 1998, 110, 1–14. 29. kim, y.t., smotkin, e.s., The effect of plasticizers on transport and electrochemical properties of PEO-based electrolytes for lithium rechargeable batteries, Solid State Ionics, 2002, 149, 29–37. 30. plancha, m.j., rangel, c.m., sequeira, c.a.c., Cation mobility in poly(ethylene oxide) solid electrolytes, J. Electroanal. Chem., 1998, 442, 91–7. 31. serraino, f.f., croce, f., epifanio, a.d., et al., PEO based polymer electrolyte lithium-ion battery, J. Eur. Ceram. Soc., 2004, 24, 1385–7. 32. guinot, s., salmon, e., penneau, j.f., et al., A new class of PEO-based SPEs: structure, conductivity and application to alkaline secondary batteries, Electrochim. Acta, 1998, 43(10), 1163–70.



515

33. rhodes, c.p., frech, r., Cation–anion and cation–polymer interactions in (PEO) nNaCF3SO3 (n = 1–80), Solid State Ionics, 1999, 121, 91–9. 34. gong, y., fu, x., zhang, s., jiang, q., Preparation of a star network PEG-based gel polymer electrolyte and its application to electrochromic devices, Chinese J. Chem. 2007, 25(11), 1743–7. 35. gong, y., fu, x., zhang, s., jiang, q., Preparation of star network peg-based gel polymer electrolytes for electrochromic devices, Chinese J. Polym. Sci., 2008, 26(1), 91–7. 36. zhang, s., fu, x., Study of solid electrolyte based on block-polymer synthesized with different molecular weight of polyoxyethylene, J. Functional Mater. (Gongneng Cailiao) (in Chinese), 2006, 37(3), 442–5. 37. zhang, s., fu, x., gong, y., Spectroscopic and electrochemical studies on blockpolymer/PMMA blend based composite polymer electrolytes, J. Appl. Polym. Sci., 2007, 106, 4091–7. 38. bohnke, o., rousselot, c., gillet, p.a., truche, c., Gel electrolyte for solid-state electrochromic cell, J. Electrochem. Soc., 1992, 139, 1862–5. 39. agnihotry, s.a., pradeep, sekhon, s.s., PMMA based gel electrolyte for EC smart windows, Electrochim. Acta, 1999, 44, 3121–6. 40. yang, d., fu, x., gong, y., Study on the preparation and performances of p(vacmma) polymer electrolytes for lithium ion battery, Chinese J. Polymer Sci., 2008, 26(4), 375–80. 41. yang, d., fu, x., jiang, q., gong, y., Synthesis of polymer electrolyte based on P(VAc-MA)/PMMA and its application in electrochromic devices, Chemical J. Chinese Universities (in Chinese), 2007, 28(9), 1781–6. 42. erickson, m., frech, r., glatzhofer, d.t., Solid polymer/salt electrolytes based on linear poly((N-2-cyanoethyl)ethylenimine), Electrochim. Acta, 2003, 48, 2059–63. 43. peng, x., peng, d., wang, f., Advances of solid polymer electrolytes, Chinese J. Org. Chem., 1989, 6(1), 1. 44. ramesh, s., yahaya, a.h., arof, a.k., Miscibility studies of PVC blends (PVC/ PMMA and PVC/PEO) based polymer electrolytes, Solid State Ionics, 2002, 148, 483–6. 45. magistris, a., quartarone, e., mustarelli, p., et al., PVDF-based porous polymer electrolytes for lithium batteries, Solid State Ionics, 2002, 152, 347–54. 46. liang, w.j., kuo, p.l., Solid polymer electrolytes VIII: effect of LiClO4-doped level on the microstructure and ionic conductivity of a chemical-covalently polyether–siloxane hybrid electrolyte, Polymer, 2004, 45, 1617–26. 47. kraft, a., rottmann, m., heckner, k.h., Large-area electrochromic glazing with ion-conducting PVB interlayer and two complementary electrodeposited electrochromic layers, Solar Energy Mater. Solar Cells, 2006, 90, 469–76. 48. kobayashi, n., chinone, h., miyazaki, a., Polymer electrolyte for novel electrochromic display, Electrochim. Acta, 2003, 48, 2323–7. 49. stoeva, z., martin, i., staunton, e., et al., Ionic conductivity in the crystalline polymer lectrolytes P(EO)6 : LiXF6, X = P, As, Sb, J. Am. Chem. Soc., 2003, 125(15), 4619–26. 50. sekhona, s.s., deepab, agnihotry, s.a., Solvent effect on gel electrolytes containing lithium salts, Solid State Ionics, 2000, 136, 1189–92. 51. ren, t., huang, x., zhao, x., tang, x., Studies on the ion states and ion transport mechanism of solid polymer electrolytes, Acta Polym. Sinica (in Chinese), 2003, 3, 361–7.


516


52. gazotti, w.a., spinacem, a.s., Polymer electrolytes based on ethylene oxide-epichlorohydrin co-polymers, Solid State Ion, 2000, 130(3–4), 281– 9. 53. chang, c.k., bauer, b.j., briker, r.m., et al., IPN material with high performance, Polym. Commun., 1997, 28(2), 34–7. 54. michael, m., prabaharans, r.s., Rechargeable lithium battery employing a new ambient temperature hybrid polymer electrolyte based on PVK+PVdF+HFP (copolymer), J. Power Sources, 2004, 136(2), 408–15. 55. croce, f., appetechi, g.b., persi, l., et al., Nanocomposite polymer electrolytes for lithium batteries, Nature, 1998, 394(30), 456–8. 56. wang, z.x., huang, x.j., chen, l.q., Understanding of effects of nano-Al2O3 particles on ionic conductivity of composite polymer electrolytes, Electochem Solid-State Lett., 2003, 6(11), 40–4. 57. li, z.f., nen, c.w., shu, j.z., Effect of modified SiO2 on the properties of PEObased polymer electrolytes, Solid State Ionics, 2003, 164(1–2), 81–6. 58. li, z.f., zhi, m.d., guo, d.w., et al., Effect of nanosized ZnO on the electrical properties of (PEO)16LiClO4 electrolytes, Mater. Sci. Eng., 2003, B99(1–3), 340–3. 59. huang, m.x., kuan, k.z., xu, z., et al., Elucidating the conductivity enhancement effect of nano-sized SnO2 fillers in the hybrid polymerelectrolyte PEO–SnO2– LiClO4, Solid State Ionics, 2003, 159(1–2), 89–5. 60. morihiro, s., hiromasa, i., yoshiharu, u.s., et al., Investigation of the effect of Lewis acid on ionic conductivity of polymer electrolyte containing Mg salt, J. Electrochem. Soc., 2003, 150(6), 726–31. 61. chun, s.w., xiang, w.z., appleby, a.j., Solvent-free composite PEO-ceramic fiber/mat electrolytes for lithium secondary cells, J. Electrochem. Soc., 2005, 152(1), 205–9. 62. sun, h.y., sohn, h.j., yamamnoto, b.o., et al., Ferroelectric materials as a ceramic filler in solid composite polyethylene oxide-based electrolytes, J. Electrochem. Soc., 1999, 146(5), 1672–6. 63. xi, j.l., takamichi, k., fu, x.z., et al., A new high-pressure phase of LiAlO2, J. Solid State Chem., 2004, 177(6), 1939–43. 64. masayuki, m., takahiro, f., nobuko, y., et al., Ionic conductance behavior of polymeric composite solid electrolytes containing lithium aluminate, Electrochim. Acta, 2001, 46(10–11), 1565–9. 65. hyun, c.l., hae, j.k., soo, h.c., et al., Synthesis of unidirectional alumina nanostructures without added organic solvents, J. Am. Chem. Soc., 2003, 125(10), 2882–3. 66. hae, j.k., hyun, c.l., jae, s.l., Alumina nanotubes containing lithium of high ion mobility, J. Am. Chem. Soc., 2003, 125(44), 13354–5. 67. luminita, l., vasii, i., lebdeh, y.a., armand, m., Synthesis and characterization of new block copolymer electrolytes with solvating affinities for different cations, Solid State Ionics, 2005, 176, 2769–75. 68. miyamoto, m., sano, y., saegusa, t., kobayashi, s., Synthesis of poly[(Nacylethylenimine)-b-(ethylene oxide)] and its antielectrostatic property, Eur. Polym. J., 1983, 19(10/11), 955–61. 69. nishimoto, a., agehara, k., furuya, n., et al., High Ionic conductivity of polyether-based network polymer electrolytes with hyperbranched side chains, Macromolecules, 1999, 32, 1541–8.



517

70. de paoli, m.a., nogueira, a.f., machado, d.a., longo, c., All-polymeric electrochromic and photoelectrochemical devices new advances, Electrochim. Acta, 2001, 46, 4243–9. 71. ribeiro, a.s., machado, d.a., santos filho, p.f., et al., Solid-state electrochromic device based on two poly(thiophene) derivatives, J. Electroanal. Chem., 2004, 567, 243–8. 72. nogueira, a.f., spinace, m.a., gazotti, w.a., et al., Poly (ethylene oxide-coepichlorohydrin)/NaI: a promising polymer electrolyte for photoelectrochemical cells, Solid State Ionics, 2001, 140, 327. 73. gazotti, w.a., spinace, m.a.s., girotto, e.m., de paoli, m.a., Polymer electrolytes based on ethylene oxide–epichlorohydrin copolymers, Solid State Ionics, 2000, 130, 281–91. 74. de paoli, m.a., nogueira, a.f., machado, d.a., longo, c., All-polymeric electrochromic and photoelectrochemical devices: new advances, Electrochim. Acta, 2001, 46, 4243–9. 75. yang, d., fu, x., gong, y., Study on the preparation and performances of P (VAc-MMA) polymer electrolytes for lithium ion battery, Chinese J. Polym. Sci., 2008, 26(4), 375–80. 76. yang, d., fu, x., gong, y., Study on P (VAc-MMA) polymer electrolytes, Acta. Chim. Sinica, 2008, 66(8), 975–9. 77. baskaran, r., selvasekarapandian, s., kuwata, n.k., hattori, t., Conductivity and thermal studies of blend polymer electrolytes based on PVAc–PMMA, Solid State Ionics, 2006, 177, 2679–82. 78. animitsa, i.e., kruglyashov, a.l., bushkova, o.v., zhukovsky, v.m., Morphology and ionic conductivity of poly(ethylene oxide)–poly(vinyl acetate)–LiClO4 polymer electrolytes, Solid State Ionics, 1998, 106, 321–7. 79. baskaran, r., selvasekarapandian, s., kuwata, n., et al., Ac impedance, DSC and FT-IR investigations on (x)PVAc–(1–x)PVdF blends with LiClO4, Mater. Chem. Phys., 2006, 98, 55–61. 80. yang, d., fu x., jiang, q., gong, y., Synthesis of polymer electrolytes based on P(VAc-MA)/PMMA and its application in electrochromic devices, Chem. J. Chinese Universities, 2007, 28(9), 1781–6. 81. pennarun, p.y., jannasch, p., Electrolytes based on LiClO4 and branched PEG–boronate ester polymers for electrochromics, Solid State Ionics, 2005, 176, 1103–12. 82. pennarun, p.y., jannasch, p., papaefthimiou, s., et al., High coloration performance of electrochromic devices assembled with electrolytes based on a branched boronate ester polymer and lithium perchlorate salt, Thin Solid Films, 2006, 514, 258–66. 83. barbosa, p.c., silva, m.m., smith, m.j., et al., Solid-state electrochromic devices based on poly(trimethylene carbonate) and lithium salts, Thin Solid Films, 2008, 516, 1480–3. 84. hong, l., shi, l., tang, x., Conductivities and spectroscopic studies of polymer electrolytes based on linear polyurethane and hybrid and copolymer of linear and hyperbranched polyurethanes, Macromolecules, 2003, 36, 4989– 94. 85. hong, l., cui, y.j., wang, x.l., tang, x.z., Synthesis of a novel one-pot approach of hyperbranched polyurethanes and their properties, J. Polym. Sci., Part A: Polym. Chem. 2002, 40, 344–50.


518


86. varshney, p., deepa, m., agnihotry, s.a., ho, k.c., Photo-polymerized films of lithium ion conducting solid polymer electrolyte for electrochromic windows (ECWs), Solar Energy Mater. Solar Cells 2003, 79, 449–58. 87. kobayashi, n., miura, s., nishimura, m., goh, y., Gel electrolyte-based flexible electrochromic devices showing subtractive primary colors, Electrochim. Acta, 2007, 53, 1643–47. 88. agnihotry, s.a., nidhia, pradeep, sekhon, s.s., Li+ conducting gel electrolyte for electrochromic windows, Solid State Ionics, 2000, 136–137, 573–6. 89. cirpan, a., argun, a.a., greiner, c.r., et al., Electrochromic devices based on soluble and processable dioxythiophene polymers, J. Mater. Chem., 2003, 13, 2422. 90. sekhona, s.s., deepab, agnihotry, s.a., Solvent effect on gel electrolytes containing lithium salts, Solid State Ionics, 2000, 136–137, 1189–92. 91. beyazyildirim, s., camurlu, p., yilmaz, d., et al., Synthesis and electrochromic properties of conducting copolymers of dioxocino-and dithiocino-quinoxalines with bithiophene, J. Electroanal. Chem., 2006, 587, 235–46. 92. varis, s., ak, m., tanyeli, c., akhmedov, i.m., toppare, l., Synthesis and characterization of a new soluble conducting polymer and its electrochromic device, Solid State Sci., 2006, 8, 1477–83. 93. beaupre, s., dumas, j., leclerc, m., Toward the development of new textile plastic electrochromic cells using triphenylamine-based copolymers, Chem. Mater., 2006, 18, 4011–18. 94. deepa, m., sharma, n., agnihotry, s.a., et al., Conductivity and viscosity of liquid and gel electrolytes based on LiClO4, LiN(CF3SO2)2 and PMMA, Solid State Ionics, 2002, 152–153, 253–8. 95. fungo, f, samson, a. jenekhe, bard, a., Plastic electrochromic devices: electrochemical characterization and device properties of a phenothiazine-phenylquinoline donor acceptor-polymer, J. Chem. Mater., 2003, 15, 1264–72. 96. reiter, j, vondrák, j., micˇka, z., The electrochemical redox processes in PMMA gel electrolytes – behaviour of transition metal complexes, Electrochim. Acta, 2005, 50, 4469–76. 97. ohsedo, y., imae, i., shirota, y., Electrochromic properties of new methacrylate copolymers containing pendant oligothiophene and oligo(ethyleneoxide) moieties in the presence of a polymer–gel electrolyte, Electrochim. Acta, 2000, 45, 1543–7. 98. vondrak, j., sedlarikova, m., reiter, j., hodal, t., Polymer gel electrolytes for electrochromic devices, Electrochim. Acta., 1999, 44, 3067–73. 99. vondrak, j., sedlarikova, m., velicka, j., et al., Gel polymer electrolytes based on PMMA: III. PMMA gels containing cadmium, Electrochim. Acta., 2003, 48, 1001–4. 100. vondrak, j., reiter, j., velicka, j., sedlarikova, m., PMMA-based aprotic gel electrolytes, Solid State Ionics, 2004, 170, 79–82. 101. nan, c., li, z., yuan, h., et al., Enhanced ionic conductivity of polymer electrolytes containing nanocomposite SiO2 particles, Phys. Rev. Lett., 2003, 91(12), 266. 102. ahmad, s., deepa, m., agnihotry, s.a., Effect of salts on the fumed silica-based composite polymer electrolytes, Solar Energy Mater. Solar Cells, 2008, 92, 184–9.



519

103. kumar, b., From colloidal to composite electrolytes: properties, peculiarities, and possibilities, J. Power. Sources, 2004, 135, 215. 104. walls, h.j., riley, m.w., singhal, r.r., et al., Nanocomposite electrolytes with fumed silica and hectorite clay networks: passive versus active fillers, Adv. Funct. Mater., 2003, 13, 710. 105. fan, j., fedkiw, p.s., Composite electrolytes prepared from fumed silica, polyethylene oxide oligomers, and lithium salts, J. Electrochem. Soc., 1997, 144, 399–408. 106. croce, f., appetechi, g.b., persi, l., scrosati, b., Nanocomposite polymer electrolytes for lithium batteries, Nature, 1998, 394, 456–8. 107. li, j., xi, j., tang, x., song, q., Study of porous PVDF-PEO polymer electrolyte, Chinese Sci. Bull. 2005, 50(3), 305–7. 108. narayan, r., krishnan, m., dubois, p., Polysaccharides grafted with aliphatic polyesters derived from cyclic esters, U.S. Patent, 1996, 5,540,929. 109. jiang, z., carroll, b., abraham, k.m., Studies of some poly(vinylidene fluoride) electrolytes, Electrochim. Acta, 1997, 42, 2667–77. 110. muniyandi, n., kalaiselvi, n., periyasamy, p., et al., Optimisation of PVdFbased polymer electrolytes, J. Power Sources, 2001, 96(1), 14–9. 111. magistris, a., quartarone, e., mustarelli, p., et al., PVDF-based porous polymer electrolytes for lithium batteries, Solid State Ionics, 2002, 152–153, 347–54. 112. stephan, a.m., yuria, s., Ionic conductivity and diffusion coefficient studies of PVdF–HFP polymer electrolytes prepared using phase inversion technique, Solid State Ionics, 2002, 148(324), 475–81. 113. yuria, s., stephan, a.m., kataoka, h., Ionic conduction mechanisms of lithium gel polymer electrolytes investigated by the conductivity and diffusion coefficient, Solid State Ionics, 2003, 160(122), 149–53. 114. song, j.m., kang, h.r., kim, s.w., et al., Electrochemical characteristics of phaseseparated polymer electrolyte based on poly(vinylidene fluoride–co-hexafluoropropane) and ethylene carbonate, Electrochim. Acta, 2003, 48(10), 1339–46. 115. cheng, c., wan, c., wang, y., Preparation of porous, chemically cross-linked, PVdF-based gel polymer electrolytes for rechargeable lithium batteries, J. Power Sources, 2004, 134(2), 202–10. 116. zhang, s., ervin, m.h., foster, d.l., et al., Fabrication and evaluation of a polymer Li-ion battery with microporous gel electrolyte, J. Solid State Electrochem, 2005, 9, 77. 117. capuanc, f., croce, f., scrosati, b., Composite polymer electrolytes, J. Electrochem. Soc., 1991, 138 (7), 1918–22. 118. vaivars, g., azens, a., Granqvist, C.G, Proton conducting polymer composites for electrochromic devices, Solid State Ionics, 1999, 119, 269–73. 119. clearfield, a., stynes, j.a., The preparation of crystalline zirconium phosphate and some observations on its ion exchange behaviour, J. Inorg. Nucl. Chem., 1964, 26, 117. 120. vaivars, g., kleperis, j., lusis, a., Antimonic acid hydrate xerogels as proton electrolytes, Solid State Ionics, 1993, 61, 317–21. 121. barbosa, p.c., silva, m.m., smith, m.j., et al., Studies of solid-state electrochromic devices based on PEO/siliceous hybrids doped with lithium perchlorate, Electrochim. Acta, 2007, 52, 2938–43.


520


122. carlos, l.d., bermudez, z.v., duarte, m.c., silva, m.m., smith, c.j., assunção, m., alcácer, l., ronda, c., welker, t. (Eds.) Physics and Chemistry of Luminescent Materials, San Francisco, Electrochemical Society Proc., 1998. 123. zéa bermudez, v., carlos, l.d., duarte, m.c., et al., A novel class of luminescent polymers obtained by the sol–gel approach, J. Alloys. Compd., 1998, 275–277, 21–6. 124. chen, z., jiang, y., sun, s., Compatibility of a novel composite microporous polymer electrolyte with anode of Li-ion batteries, Chem. J. Chinese Universities (in Chinese), 2006, 27(10), 1937–40. 125. li, z., su, g., gao, d., wang, x., li, x., Composite polymer electrolyte with micro-porous structure, Chinese J. Appl. Chem. (in Chinese), 2004, 21(11), 1160–4. 126. osman, z., ibrahim, z.a., arof, a.k., Conductivity enhancement due to ion dissociation in plasticized chitosan based polymer electrolytes, Carbohydr. Polym., 2001, 44, 167–73. 127. de oliveira, s.c., de morais, l., curvelo, a.a.s., torresi, r.m., Organic aqueous gel as electrolyte for application in electrochromic devices based in bismuth electrodeposition, J. Electrochem. Soc., 2003, 150, E578–E582. 128. tambelli, c.e., donoso, j.p., regiani, a.m., et al., Nuclear magnetic resonance and conductivity study of HEC/polyether-based polymer electrolytes, Electrochim. Acta, 2001, 46, 1665–72. 129. pawlicka, d.c., dragunski, c.o., avellaneda, a. graja, r. bulka, f. kajzar (Eds.), Molecular Low Dimensional and Nanostructured Materials for Advanced Applications, Kluver Academic Publishers in the NATO ASI Series, 2002. 130. orel, b., krasovec, u.o., stangar, u.l., judeinstein, p., All sol–gel electrochromic devices with Li+ ionic conductor WO3 electrochromic films and SnO2 counter-electrode films, J. Sol–Gel Sci.Technol., 1998, 11, 87. 131. pawlicka, a., dragunski, d.c., guimaraes, k.v., avellaneda, c.o., Electrochromic devices with solid electrolytes based on natural polymers, Mol. Cryst. Liq. Cryst., 2004, 416, 105. 132. regiani, a.m., tambelli, c.e., pawlicka, a., et al., DSC and solid state NMR characterization of hydroxyethylcellulose/polyether films, Polym. Int., 2000, 49, 960. 133. dragunski, d., pawlicka, a., Starch based solid polymeric electrolytes, Mol. Cryst. Liq. Cryst., 2002, 374, 561. 134. lopes, l.v.s., machado, g.o., pawlicka, a., donoso, j.p., Nuclear magnetic resonance and conductivity study of hydroxyethylcellulose based polymer gel electrolytes, Electrochim. Acta, 2005, 50, 3978–84. 135. lopes, l.v.s., dragunski, d.c., pawlicka, a., donoso, j.p., Nuclear magnetic resonance and conductivity study of starch based polymer electrolytes, Electrochim. Acta, 2003, 48, 2021–7. 136. rodrigues, d., pawlicka, a., Influence of humidity on ionic conductivity measurements of solid polymeric electrolytes obtained from plasticized starch, Nonlinear Opt. Quant. Opt., 2004, 32, 87. 137. vieira, d., avellaneda, c., pawlicka, a., Conductivity study of a gelatin-based polymer electrolyte, Electrochim. Acta, 2007, 53, 1404–8. 138. avellaneda, c., vieira, d., al-kahlout, a., et al., Solid-state electrochromic devices with Nb2O5: Mo thin film and gelatin-based electrolyte, Electrochim. Acta, 2007, 53, 1648–54.



521

139. avellaneda, c., vieira, d., al-kahlout, a., et al., All solid-state electrochromic devices with gelatin-based electrolyte, Solar Energy Mater. Solar Cells, 2008, 92, 228–33. 140. stevens, j.r., wieczorek, w., raducha, d., jeffrey, k.r.s., Proton conducting gel/ H3PO4 electrolytes, Solid State Ionics, 1997, 97, 347. 141. raducha, d., wieczorek, w., florjan´czyk, z., stevens, j.r., Nonaqueous H3PO4doped gel electrolytes, J. Phys. Chem., 1996, 100, 20126. 142. pu, h., huang, p., Studies on transparent and solid proton conductors based on NH4H2PO4 doped poly(vinyl alcohol), Mater. Lett., 2006, 60, 1724–7. 143. pu, h., luo, m., yang, z., Transparent and anhydrous proton conductors based on PVA/imidazole/NH4H2PO4 composites, European Polym. J., 2007, 43, 5076–83. 144. fauvarque, f., guinot, s., bouzir, n., salmon, e., penneau, j.f., Alkaline poly(ethylene oxide) solid polymer electrolytes. Application to nickel secondary batteries, Electrochim. Acta, 1995, 40, 2449–53. 145. guinot, s., salmon, e., penneau, j.f., fauvarque, j.f., A new class of PEO-based SPEs: structure, conductivity and application to alkaline secondary batteries, Electrochim. Acta, 1998, 43, 1163–70. 146. vassal, n., salmon, e., fauvarque, j.f., Electrochemical properties of an alkaline solid polymer electrolyte based on P(ECH-co-EO), Electrochim. Acta, 2000, 45, 1527–32. 147. palacios, i., castillo, r., vargas, r.a., Thermal and transport properties of the polymer electrolyte based on poly(vinyl alcohol)-KOH- H2O, Electrochim. Acta, 2003, 48, 2195–9. 148. lewandowski, a., skorupska, k., malinska, j., Novel poly(vinyl alcohol)– KOH–H2O alkaline polymer electrolyte, J. Solid State Ionics, 2000, 133, 265–71. 149. yamada, m., honma, i., Heteropolyacid-encapsulated self-assembled materials for anhydrous proton-conducting electrolytes, J. Phys. Chem. B, 2006, 110, 20486–90. 150. nakamura, o., kodama, t., ogino, i., miyake, y., High-conductivity solid proton conductors: dodecamolybdophosphoric acid and dodecatungstophosphoric acid crystals, Chem. Lett., 1979, 17–8. 151. cheückiewicz, k., zukowska, g., wieczorek, w., Synthesis and characterization of the proton-conducting gels based on PVdF and PMMA matrixes doped with heteropolyacids, Chem. Mater., 2001, 13, 379–84. 152. zukowska, g., chojnacka, n., wieczorek, w., Effect of gel composition on the conductivity of proton-conducting gel polymeric electrolytes doped with H3PO4, Chem. Mater., 2000, 12, 3578–82. 153. stevens, j.r., raducha, d., wieczorek, w., jeffrey, k.r.., Nonaqueous polymer electrolytes for electrochromic devices, I.M. Khan, J.S. Harrison (Eds.), Field Responsive Polymers, ACS Symposium Series, Oxford Univ. Press, Washington, 1999. 154. zhao, x., liu, y., zhou, b., wang, z., Film of proton-conducting polymer electrolyte doped with 12-tungstophosphoric acid, Acta Chim. Scinica (in Chinese), 2006, 64, 975–8. 155. zukowska, g., rogowska, m., weczkowska, e., wieczorek, w., Proton conducting polymer gel electrolytes, Solid State Ionics, 1999, 119, 289–93. 156. zukowska, g., robertson, v.j., marcinek, m.l., et al., Structure of protonconducting anhydrous gel electrolytes based on poly(glycidyl methacrylate), J. Phys. Chem. B, 2003, 107, 5797–805.


522


157. yuan, a., zhao, j., Studies on PVA-CMC-KOH-H2O alkaline polymer electrolytes. electrochemistry, J. Electrochem. (in Chinese), 2006, 12, 40–5. 158. zhao, j., yuan, a., song, w., Study of alkaline nano-composite polymer electrolytes, Acta Chim. Sinica, 2005, 63, 219–22. 159. brazier, a., appetecchi, g.b., passerini, s., et al., Ionic liquids in electrochromic devices, Electrochim. Acta, 2007, 52, 4792–7. 160. surca vuk, a., jovanovski, v., pollet–villardb, a., et al., Imidazolium-based ionic liquid derivatives for application in electrochromic devices, Solar Energy Mater. Solar Cells, 2008, 92, 126–35. 161. marcilla, r., alcaide, f., sardon, h., et al., Tailor-made polymer electrolytes based upon ionic liquids and their application in all-plastic electrochromic devices, Electrochem. Commun., 2006, 8, 482–8. 162. cummins, d., boschloo, g., ryan, m., corr, d., rao, s., fitzmaurice, d., Ultrafast electrochromic windows based on redox-chromophore modified nanostructured semiconducting and conducting films, J. Phys. Chem. B, 2000, 104, 11449–59. 163. stathatos, e., lianos. p., surcavuk, a., orel, b., Optimization of a quasi-solidstate dye-sensitized photoelectrochemical solar cell employing a ureasil/sulfolane gel electrolyte, Adv. Funct. Mater., 2004, 14, 45–8. 164. susan, m.a.b.h., kaneko, t., noda, a., et al., Ion gels prepared by in situ radical polymerization of vinyl monomers in an ionic liquid and their characterization as polymer electrolytes, J. Am. Chem. Soc., 2005, 127, 4976–83. 165. noda, a., watanabe, m., Highly conductive polymer electrolytes prepared by in situ polymerization of vinyl monomers in room temperature molten salts, Electrochim. Acta, 2000, 45, 1265–70. 166. henderson, w.a., passerini, s., Phase behavior of ionic liquid−LiX mixtures: pyrrolidinium cations and TFSI-anions, Chem. Mater., 2004, 16, 2881–5. 167. appetecchi, g.b., scaccia, s., tizzani, c., et al., Synthesis of hydrophobic ionic liquids for electrochemical applications, J. Electrochem. Soc., 2006, 153, A1685. 168. shin, j.h., henderson, w.a., tizzani, c., et al., Characterization of solvent-free polymer electrolytes consisting of ternary PEO–LiTFSI–PYR14 TFSI, J. Electrochem. Soc., 2006, 153, A1649–A1654. 169. li, z., liu, h., liu, y., he, p., li, j., A room-temperature ionic-liquid-templated proton-conducting gelatinous electrolyte, J. Phys. Chem. B, 2004, 108, 17512–8. 170. lu, w., fadeev, g.a., qi, b., smela, e., mattes, b.r., et al., Use of ionic liquids for conjugated polymer electrochemical devices, Science, 2002, 297, 983–7. 171. noda, a., watanabe, m., Highly conductive polymer electrolytes prepared by in situ polymerization of vinyl monomers in room temperature molten salts, Electrochim. Acta, 2000, 45, 1265–70. 172. shobukawa, h., tozuda, h., susan, m.a.b.h., watanabe, m., Ion transport properties of lithium ionic liquids and their ion gels, Electrochim. Acta, 2005, 50, 3872–7. 173. marcilla, r., blazquez, j.a., rodríguez, j., et al., Tuning the solubility of polymerized ionic liquids by simple anion-exchange reactions, J. Polym. Sci. Part A: Polym. Chem., 2004, 42, 208–12. 174. marcilla, r., blazquez, j.a., fernandez, r., et al., Synthesis of novel polycations using the chemistry of ionic liquids, Macromol. Chem. Phys., 2005, 206, 299–304.



523

175. bonhote, p., dias, a., papageorgiou, n., kalyanasundaram, k., gratzel, m., Hydrophobic, highly conductive ambient-temperature molten salts, Inorg. Chem., 1996, 35, 1168–78. 176. yoshizawa, m., ohno, h., Synthesis of molten salt-type polymer brush and effect of brush structure on the ionic conductivity, Electrochim. Acta, 2001, 46, 1723–8. 177. yeon, s., kim, k., choi, s., cha, j., lee, h., Characterization of PVdF(HFP) gel electrolytes based on 1-(2-hydroxyethyl)-3-methyl imidazolium ionic liquids, J. Phys. Chem. B, 2005, 109, 17928–35. 178. ho, k.c., rukavina, t.g., greenberg, c.b., Tungsten oxide-prussian blue electrochromic system based on a proton-conducting polymer electrolyte, J. Electrochem. Soc., 1994, 141, 2061–7. 179. zukowska, g., williams, j., stevens, j.r., et al., The application of acrylic monomers with acidic groups to the synthesis of proton-conducting polymer gels, Solid State Ionics, 2004, 167, 123–30. 180. jelle, b.p., hagen, g.j., Correlation between light absorption and electric charge in solid state electrochromic windows, J. Appl. Electrochem., 1999, 29, 1103–10. 181. passerini, s., scrosati, b., hermann, v., et al., Laminated electrochromic windows based on nickel oxide, tungsten oxide, and gel electrolytes, J. Electrochem. Soc., 1994, 141, 1025–8. 182. wieczorek, w., zukowska, g., borkowska, r., et al., A basic investigation of anhydrous proton conducting gel electrolytes, Electrochim. Acta, 2001, 46, 1427–38. 183. zukowska, g., wieczorek, w., ke¸dzierski, m., florjan´czyk, z., Nonaqueous gel electrolytes doped with phosphoric acid esters, Solid State Ionics, 2001, 144, 163–73. 184. lauter, u., meyer, w.h., wegner, g., Molecular composites from rigid-rod poly(p-phenylene)s with oligo(oxyethylene) side chains as novel polymer electrolytes, Macromolecules, 1997, 30, 2092–101. 185. huang, y., chen, l., ho, k., Switching behavior of the Prussian blue–indium hexacyanoferrate electrochromic device using the K+-doped poly-AMPS electrolyte, Solid State Ionics, 2003, 165, 269–77. 186. wang, j., yu, c., hwang, s., ho, k., chen, l., Influence of coloring voltage on the optical performance and cycling stability of a polyaniline–indium hexacyanoferrate electrochromic system, Solar Energy Mater. Solar Cells, 2008, 92, 112. 187. giglia, r.d., Electrochromic cells with improved electrolyte system, U.S. Patent, 1983, 4,375,318. 188. tung, t.s., chen, l., ho, k., An indium hexacyanoferrate–tungsten oxide electrochromic battery with a hybrid K+/H+-conducting polymer electrolyte, Solid State Ionics, 2003, 165, 257–67.


13 Hyperbranched polymer electrolytes for high temperature fuel cells T. I T O H, Mie University, Japan

Abstract: This chapter describes the application of hyperbranched polymer-based electrolyte membranes for high temperature fuel cells. Hyperbranched polymers with a sulfonic acid group or a phosphonic acid one as a proton source were synthesized and the proton conductivity, thermal property and fuel cell performance of their polymers and crosslinked membranes were investigated. The concept of the proton conduction coupled with the polymer chain motion was proposed for high temperature fuel cells. Key words: hyperbranched polymer, sulfonic acid group, phosphonic acid group, proton-conducting polymer, high temperature fuel cells.

13.1

Introduction

Polymer electrolyte fuel cells (PEFCs) are of interest as power sources in vehicles and portable applications because of their high energy efficiency and environmentally friendly qualities (Kreuer, 1996; Steele and Heinzel, 2001; Yang et al., 2001). These fuel cells have typically used perfluorinated, modified perfluorinated, and partially perfluorinated polymer electrolytes as the membranes to separate the electrodes and the oxygen and hydrogen (or methanol) gas streams (Savadogo, 1998, 2004). The perfluorsulfonic acid polymer Nafion® is one of the most extensively studied proton exchange membrane for PEFC applications. However, it has a number of drawbacks that need to be overcome, which are the high cost, owing to its complicated manufacturing procedure, the high methanol permeability, and the poor performance at temperatures above 80 °C due to the loss of the water. One of the arguments for the development of the new polymer electrolytes is the necessity to operate the cell under high temperature conditions. The operation of the cell at temperatures of more than 100 °C is very interesting because anode catalyst poisoning by carbon monoxide is less important, the kinetics of the fuel oxidation will be improved and the efficiency of the cell should be significantly enhanced. Therefore, there has been a great demand for new electrolyte membranes for high temperature PEFCs. Sulfonated aromatic polymers, organic–inorganic composite electrolyte membranes, 524 © Woodhead Publishing Limited, 2010

Hyperbranched polymer electrolytes for high temperature fuel cells 525 and blends of different polymers with phosphoric acid have been explored (Wainright et al., 1995; Wang et al., 1996; Savadogo, 1998, 2004; Jannasch, 2003; Karlsson and Jannasch, 2004). In the first two types of electrolytes, the ionic conductivity depends upon the presence of water. Among the blends with phosphoric acid, poly(benzimidazole) has been successfully tested in the fuel cells which were operated at temperatures of up to 200 °C though the long-term durability of these electrolytes has not yet been published (Wainright et al., 1995; Wang et al., 1996). Recently, as new polymeric materials that are capable of fast proton conduction in the absence of any volatile compounds and that are durable at elevated temperatures, fully polymeric proton-conducting membranes based on nitrogen-containing heterocycles such as imidazole, benzimidazole and pyrazole which have in many respects properties similar to those of water have been prepared and studied (Kreuer, 2001; Schuster et al., 2001, 2004; Herz et al., 2003; Persson and Jannasch, 2003, 2005; Kim et al., 2007). These approaches have been studied for application to the high temperature fuel cell systems because the proton conduction strongly depends on the presence of the water for the sulfonated polymer systems (Kreuer et al., 2004). On the other hand, the ion conducting phenomena have been widely investigated in dry polymer systems such as combinations of alkaline metal salts and polyethers (Gray, 1991). Using dry polymer systems for proton conduction may require temperatures higher than 100 °C since the ionic conduction is generally induced by polymer chain motion. In addition, a single proton conductor is more desirable for avoiding evaporation and/or migration of low molecular proton media and acids. In this chapter, synthesis, ionic conductivity and fuel cell performance of hyperbranched polymers with a sulfonic acid group or a phosphonic acid group as a proton source and with an ether bond as the proton transport moiety, which might contribute to both proton dissociation and transportation, and of their polymer-based membranes are described with our research as the central focus (Itoh et al., 2006a,b, 2008, 2009), and then a concept of the proton conduction coupled with the polymer chain motion for hightemperature fuel cells is proposed.

13.2

Hyperbranched polymer electrolytes with a sulfonic acid group at the periphery

13.2.1 Synthesis Hyperbranched polymers (HBP) with a sulfonic acid (SA) group (HBP– SA), with an acryloyl (Ac) group (HBP–Ac), and with both sulfonic acid and acryloyl groups (HBP–SA–Ac) The hyperbranched polymer (HBP) with sulfonic acid (SA) groups at the periphery (HBP–SA) and the hyperbranched polymer with acryloyl (Ac)



O

O

O

O

1

3

3

H

H

C k O

O

C

O

O

O

O

(Bu)3SnCl

O

O

HBP–OH

3

3

O

O

O

O

O

O

3

3

3

3

H

m

l

H

OH

Cl

NEt3

O

2) Ion-exchange resin

DCC, DMAP

O

Scheme 13.1 Synthetic route of HBP–SA and HBP–Ac.

O

MeO

O

1)

SO3K

O

O

O

O

O

O

O

O

O C k O

HBP–Ac

3

3

C O

HBP–SA

C k O

O

3

3

C O

O

O

O

O

O

O

O

O

O

O

O

O

O

O

O

3

3

3

3

O

m

l

O

O

3

3

3

3

O

m

l

O

SO3H

SO3H

Hyperbranched polymer electrolytes for high temperature fuel cells 527 groups at the periphery (HBP–Ac) as a new crosslinker were synthesized according to the procedure as shown in Scheme 13.1. The hyperbranched polymer with a hydroxyl group at the periphery (HBP–OH) was prepared by the polymerization of methyl 3,5-bis[(8′-hydroxy-3′,6′-dioxaoctyl)oxy] benzoate (1) monomer in the presence of tributyltin chloride (Itoh et al., 1999), and separated into two parts, a high molecular weight polymer fraction (HBP–OH, Mn = 14 000) and a low molecular weight one (HBP–OH, Mn = 3000), by redissolution-reprecipitation method using tetrahydrofuran (THF) as a solvent and isopropyl ether as a precipitant. The reaction of the high molecular weight HBP–SA (Mn = 14 000) with 4-sulfobenzoic acid potassium salt in the presence of dicyclohexylcarbodiimide (DCC) and dimethylaminopyridine (DMAP) at room temperature, followed by proton exchange reaction through an acid-type ion exchange resin, gave the hyperbranched polymer with a sulfonic acid group at the periphery (HBP–SA) as a pale yellow rubbery solid in 77% yield. The HBP–SA is soluble in water, methanol and ethanol, and insoluble in chloroform. The hyperbranched-type crosslinker (HBP–Ac), which is a hyperbranched polymer with an acryloyl group at the periphery, was obtained as viscous oils in 75% yield by the reaction of the low molecular weight HBP–OH (Mn = 3000) with an excess of acryloyl chloride in the presence of triethylamine. The hyperbranched polymers (HBP–SA–Acs) with both sulfonic acid and acryloyl groups at the periphery in different SO3H/Ac ratios were synthesized according to the procedure as shown in Scheme 13.2. The hyperbranched polymer with a hydroxyl group at the periphery (HBP–OH, Mn = 4500) was reacted with acrylic acid in the presence of DCC and DMAP in dichloromethane at room temperature to give the HBP-Ac as yellow highly viscous oils. The reaction of the HBP-Ac with 4-sulfobenzoic acid monopotassium salt in the presence of DCC and DMAP in N,Ndimethylformamide (DMF) at room temperature gave the HBP–SK–Ac as yellow viscous oils, which were purified by dissolution–reprecipitation method using DMF as a solvent and isopropyl ether as a precipitant, respectively, and finally using DMF as a solvent and a mixture of isopropyl ether and ethanol (1/1 v/v) as a precipitant. The four kinds of HBP–SA–Acs with both sulfonic acid and acryloyl groups at the periphery were obtained as rubbery solids by proton exchange reaction through acid-type ion exchange resin. Acryloyl group and sulfonic acid group contents for the HBP–SA–Acs were determined by 1H nuclear magnetic resonance (NMR) measurement to be the SO3H/Ac ratios of 87/10, 78/18, 67/28 and 55/43 in mol%, respectively. Some 2–5% of unreacted OH groups remain at the periphery of these HBP–SA–Acs from the SO3H/ Ac ratios. These HBP–SA–Acs are soluble in N,N-dimethylacetoamide (DMAc), DMF and dimethylsulfoxide (DMSO), and insoluble in benzene, chloroform, THF, isopropyl ether, methanol, ethanol, acetonitrile and water.



O

O

3

3

k

O

C

O

O

O

Ion-exchange resin

HBP–OH

O

O

O

O

O

O

O

3

3

3

3

H

m

l

H

OH

O

O

3

3

k O

C

O

C

HBP–SA–Ac

O

O

DCC, DMA P

O O

O

O

O

O

O

O

O

O

O

O

3

3

3

3

O

m

l

O

O

O

k

C O

SO3H

O

O

3

3

BPO

O

O

3

l

O OH

O

O

C

O

O 3

O

m

O

O

O

DCC, DMA P

3

k O

C

CL–HBP–SA

O

3

O

O 3

k O

C

O

O

O

O

O

O

O

O

3

3

3

3

O

m

l

O

HBP–SAK–Ac

O

O

O

SO3K

O

3

H C

O

O

C

HBP–Ac

3

3

O

Scheme 13.2 Synthetic route of HBP–SA–Ac and CL-HBP–SA membrane.

O

O

C

O

O

O

O

O

n

SO3H

O

O

O

O

3

3

3

3

O

m

l

O

SO3K

Hyperbranched polymer electrolytes for high temperature fuel cells 529 Semi-interpenetrated electrolyte membrane (HBP–SA-co-HBP–Ac) and crosslinked (CL) electrolyte membrane (CL-HBP–SA) The polymer electrolyte membrane (HBP–SA-co-HBP–Ac) was prepared by an interpenetration reaction of HBP–SA (80 wt%) and a crosslinker HBP–Ac (20 wt%) in vacuo at 130 °C for 48 h, and the resulting membrane was insoluble in water, methanol, ethanol, chloroform, DMF, DMSO and N-methylpyrrolidone (NMP), indicating that the resulting electrolyte membrane is formed by a complete interpenetrating network. The crosslinked electrolyte membrane (CL-HBP–SA) was prepared from the HBP–SA–Ac by solvent casting technique on parting-agent coated poly(ethyleneterephthalate) (PET) sheet at room temperature, followed by a crosslinking reaction at 120 °C for 48 h using benzoyl peroxide (BPO) as a radical initiator. On the preparation of the electrolyte membrane, 5 wt% fumed silica as a thickener was added to the HBP–SA–Ac solution in DMAc to obtain a large homogeneous film with a given thickness. Without addition of the fumed silica, it was difficult to prepare large size membranes suitable for fuel cell test. Equivalent weights (Ew) of the HBP–SA, the HBP–SA-co-HBP–Ac, and the HBP–SA–Acs were determined from sulfur and potassium contents by using an inductively coupled plasma (ICP) method to be 628 for the HBP– SA, 855 for the HBP–SA-co-HBP–Ac, 1010 for the HBP–SA–Ac with a SO3H/Ac ratio of 78/18 in mol%, and 1000 for the HBP–SA–Ac with a SO3H/Ac ratio of 67/28 in mol%, respectively.

13.2.2 Ionic conductivity Hyperbranched polymer with a sulfonic acid group (HBP–SA) and semi-interpenetrated electrolyte membrane (HBP–SA-co-HBP–Ac) The temperature dependence of the ionic conductivities for the HBP–SA and the HBP–SA-co-HBP–Ac membrane in the temperature range of 60–150 °C is shown in Fig. 13.1. Ionic conductivity data of the polymer and electrolyte membrane are typically interpreted using the Vogel–Tamman– Fulcher (VTF) equation (Vogel, 1921; Tammann and Hesse, 1926; Fulcher, 1925): σ(T) = AT−½ exp (−B/(T−T0))

[13.1]

where σ and T are the ionic conductivity and the absolute temperature, respectively. A, B and T0 are the fitting parameters and correspond to carrier ion number, activation energy, and the temperature where the free volume vanishes, respectively. VTF parameters obtained by the best fit between the observed ionic conductivity and the theoretical curves with eq. [13.1] are summarized in Table 13.1. The relatively larger A value for the


530


Ionic conductivity (S cm–1)

10–3

10–4

10–5

10–6

2.3 2.4 2.5 2.6 2.7 2.8 2.9 3.0 3.1 1000/T (K–1)

13.1 Temperature dependence of the ionic conductivities for the HBP– SA (䊉) and the HBP–SA-co-HBP–Ac membrane (䉱). The solid lines are the result of fitting the ionic conductivity data on the Vogel–Tammann– Fulcher (VTF) equation.

Table 13.1 VTF parameters obtained for the HBP–SA and the HBP–SA-co-HBP–Ac membrane Polymer and electrolyte membrane

A (S cm−1 K−1/2)

B (K)

T0 (K)

HBP–SA HBP–SA-co-HBP–Ac

0.77 0.47

666 996

256 248

HBP–SA suggests a larger number of carrier ions, which should be responsible for the observed higher ionic conductivity than the HBP–SA-co-HBP– Ac membrane. Both polymer and electrolyte membranes exhibited a VTF-type temperature dependence and the ionic conductivities were 7.0 × 10−4 S cm−1 for the HBP–SA and 7.8 × 10−5 S cm−1 for the HBP–SA-coHBP–Ac membrane, respectively, at 150 °C under dry conditions. This indicates that the dissociation of the sulfonic acid group in the polymers takes place under dry conditions and the long-range proton transport occurs by segmental motion like the various polymer electrolytes composed of poly(ethylene oxide) (PEO) and PEO-related derivatives and lithium salts (Gray, 1991, 1997; Scrosati, 1993). The ionic conductivity of the HBP–SA-


1500 1400 1300 1200 1100 1000 900 800 700 600 500 –50

110 105 100 95

0

50

100

150 200 Time (h)

250

300

350

400

Temperature (°C)

Resistance (Ω cm2)

Hyperbranched polymer electrolytes for high temperature fuel cells 531

90

13.2 Time dependence of the resistance of the HBP–SA-co-HBP–Ac membrane at 95 °C in vacuo.

co-HBP–Ac membrane is one order lower than that of the HBP–SA, owing to a suppression of the mobility of the polymer chains by crosslinking, which is supported by the increased glass transition temperature as described in Section 13.2.3. Figure 13.2 shows the storage characteristics (time dependence of the ionic conductivity) for the HBP–SA-co-HBP–Ac membrane at 95 °C in vacuo. The ionic conductivity is almost constant for a long time, indicating strongly that the residual water in the membrane may not be related to the ionic conduction mechanism. Crosslinked electrolyte membrane (CL-HBP–SA) The temperature dependence of the ionic conductivities for the CL-HBP– SA membranes in the temperature range of 80–150 °C is shown in Fig. 13.3. The ionic conductivities of the CL-HBP–SA membranes increased with an increase in the SO3H group content, owing to an increase in the number of proton carriers in the membranes. The ionic conductivities of these four electrolyte membranes at 150 °C under dry conditions are lower by about one-fourth order than that of the hyperbranched polymer with a sulfonic acid group at the periphery (HBP–SA, 7.0 × 10−4 S cm−1 at 150 °C). This is due to a lower content of the sulfonic acid group, that is, a sacrifice of the carrier number caused by the replacement of a sulfonic acid group with an acryloyl group in the same molecule. The temperature dependences of the ionic conductivities for these electrolyte membranes are also interpreted well with the VTF eq. [13.1] as well as the cases of the HBP–SA and HBP– SA-co-HBP–Ac membrane. VTF parameters are summarized in Table 13.2. Therefore, the proton conduction is governed by microscopic viscosity, i.e. proton transfer is cooperated by local polymer chain motion.


532

Polymer electrolytes 10–3


10–4

10–5

10–6

10–7

10–8

10–9 2.3

SO3H/Ac 87/10 78/18 67/28 55/43 2.4

2.5

2.6

2.7

2.8

2.9

1000/T (K–1)

13.3 Arrhenius plots of the ionic conductivities for the CL-HBP–SAs with SO3H/Ac ratios of 87/10 (䊏), 78/18 (䊉), 67/28 (䉱), and 55/43 (䉲) in mol%, respectively.

Table 13.2 VTF parameters obtained for the CL-HBP–Ac membrane CL-HBP–SA SO3H/Ac (in mol%)

A (S cm−1 K−1/2)

B (K)

T0 (K)

87/10 78/18 67/28 55/43

0.89 6.27 0.31 1.08

859 1610 1259 1770

260 236 255 243

The CL-HBP–SA membrane with the SO3H/Ac ratio of 87/10 in mol% shows the highest ionic conductivity of 2.2 × 10−4 S cm−1 among four CL-HBP–SA membranes. This is due to an increase in the number of proton carriers by an increase of the sulfonic acid group content (large A value) and the enhanced segmental motion (low B value) attributable to a low crosslinking density by the lowest acryloyl group content. Although the CL-HBP–SA membranes have lower ionic conductivity than the HBP-SA, it has much better mechanical strength suitable for practical application than the HBP–SA.



Exo

Tg = 22.4 °C HBP–SA

Endo

Tg = 40.0 °C HBP–SA-co-HBP–Ac

–60

–40

–20

0

20

40

60

80

Temperature (°C)

13.4 DSC thermograms of the HBP–SA and the HBP–SA-co-HBP–Ac membrane under air.

13.2.3 Thermal properties Hyperbranched polymer with a sulfonic acid group (HBP–SA) and semiinterpenetrated electrolyte membrane (HBP–SA-co-HBP–Ac) Differential scanning calorimetry (DSC) measurements were carried out in the temperature range of −70–90 °C for the HBP–SA and the HBP–SAco-HBP–Ac membrane under air. DSC traces of the HBP–SA and the HBP–SA-co-HBP–Ac membrane are shown in Fig. 13.4, indicating that they are completely amorphous. The glass transition temperatures (Tg) were found to be 22.4 °C for the HBP–SA and 40.0 °C for the HBP–SAco-HBP–Ac membrane, respectively. An increase in Tg observed for the HBP–SA-co-HBP–Ac membrane is due to the restricted segmental motion by crosslinking. Thermogravimetric/differential thermal analysis (TG/DTA) measurements were carried out under air to investigate the thermal stability of the HBP–SA and the HBP–SA-co-HBP–Ac membrane. Figure 13.5 shows their TG/DTA traces, where the first weight loss occurs at temperatures above 200 °C under air. They have sufficient thermal stability for fuel cell application and such high thermal stability is due to the presence of the aromatic ring in the polymer and the membrane, but the crosslinking structure does not affect significantly the thermal stability of the electrolytes. Crosslinked electrolyte membrane (CL-HBP–SA) DSC measurements were carried out at the temperature range of −150– 150 °C for four CL-HBP–SA membranes under argon. The Tg values


534


Td = 202 °C 800

(a)

80

40 20

400

TG (%)

DTA (μV)

60 600

0 200 –20 0 0

200

400

600

800

–40

Temperature (°C) 70 100

Td = 200 °C

60

80

40

60

30 40

TG (%)

DTA (μV)

(b) 50

20 20

10 0 0

200

400

600

800

0

Temperature (°C)

13.5 TG/DTA diagrams of (a) the HBP–SA and (b) the HBP–SA-co-HBP– Ac membrane under air. Table 13.3 Glass transition temperatures (Tg) and decomposition temperatures (Td) for the CL-HBP–Ac membranes with various SO3H/Ac ratios CL-HBP–SA SO3H/Ac (in mol%)

Tg (°C)

Td (°C)

87/10 78/18 67/28 55/43

36.5 47.4 28.8 38.6

267 262 260 271

observed for four CL-HBP–SA membranes are summarized in Table 13.3. All of CL-HBP–SAs have higher Tg values than the HBP–SA (Tg = 22.4 °C) due to the suppression of segmental motion caused by crosslinking structure formation in CL-HBP–SA membranes. Unfortunately, a clear relation© Woodhead Publishing Limited, 2010

Hyperbranched polymer electrolytes for high temperature fuel cells 535 2000 100 80 TG 60 1000 40 500

20 DTA

0 0

TG (%)

DTA (μV)

1500

200

400

0 600

800

1000

Temperature (°C)

13.6 DTA and thermogravimetry trace of the CL-HBP–SA with a SO3H/ Ac ratio of 87/10 in mol% under air.

ship of the Tg values with the composition of the SO3H and the Ac units was not observed. Since the operation temperature of the fuel cell is set around 130 to 150 °C, the status of the membrane is presumed to be the rubber region, i.e. the operation temperature is always higher than the Tg of the polymer, without melting. This is a quite different and important concept from the other high temperature polymer electrolyte membranes. TG/DTA measurements were carried out under air to investigate the thermal stability for the CL-HBP–SA membranes. The TG/DTA trace for the CL-HBP–SA membrane with the SO3H/Ac ratio of 87/10 in mol% is shown in Fig. 13.6, where the decomposition of the CL-HBP–SA membrane begins at 267 °C, and it shows two successive stages of weight loss. The first weight loss of the CL-HBP–SA membrane occurs between 270 and 330 °C due to the elimination of the sulfonic acid group and the second one takes place in the temperature range of 330–550 °C due to the decomposition of the ethylene oxide unit. The decomposition temperatures (Td) observed for four CL-HBP–SA membranes are summarized in Table 13.3. They have much better thermal stability than the HBP–SA, which decomposes at the temperature of 202 °C. Although the decomposition of the CL-HBP–SA membranes might take place at the sulfonic acid group in the polymer structure, they have suitable thermal stabilities as electrolyte membranes for the high-temperature fuel cell under non-humidified conditions.

13.2.4 Fuel cell measurement The semi-interpenetrated HBP–SA-co-HBP–Ac membrane could not be used for fuel cell test measurement, because leakage of the HBP–SA took © Woodhead Publishing Limited, 2010

536

Polymer electrolytes 0.8 130 °C 140 °C 150 °C

0.7

Potential (V)

0.6 0.5 0.4 0.3 0.2 0.1 0.0 –0.1 0.0

0.2

0.4

0.6

0.8

1.0

1.2

–2

Current density (mA cm )

13.7 The current vs. voltage characteristics of test fuel cell at 130, 140 and 150 °C using CL-HBP–SA with a SO3H/Ac ratio of 78/18 in mol% as an electrolyte under non-humidified conditions. Dry hydrogen gas and dry oxygen gas were used as fuel and oxidation gas, respectively. Potential scan rate is 20 mV s–1.

place from the membrane on keeping it at temperatures as high as 150 °C for a long time. Therefore, the fuel cell test was carried out with only CL-HBP–SA membrane. The polarization (I–V) characteristics of the CL-HBP–SA membrane with a SO3H/Ac ratio of 78/18 in mol%, which showed relatively high ionic conductivity and also better mechanical strength, were measured by using a test fuel cell. The voltage profiles observed at cell temperatures of 130, 140 and 150 °C under non-humidified conditions, where excess hydrogen and dry oxygen gases were introduced as the fuel and the oxidant, respectively, are shown in Fig. 13.7. The polarization curve shows slightly curvature at 150 °C, though the linear polarization is observed at 130 °C. Also, the limiting current increases with the temperature. It is presumed that the proton transport and the charge transfer reaction are activated in accordance with an increase in the temperature. To verify the exact proton transport, i.e. continuous direct current, in the membrane, chronoamperometry was performed. The current profiles given by potentiostatic operation at 300 mV are shown in Fig. 13.8 for cell temperatures of 130, 140 and 150 °C. At the initial 100 seconds, each current profile decreased significantly. However, the current profiles achieved constant values after 500 seconds. These steady state currents are exactly dependent upon proton transport in the CL-HBP–Ac membrane. The open-circuit voltages (OCV) of these CL-HBP–SA membranes are 0.73 V at 130 °C, 0.71 V at 140 °C and



Current density (mA cm–1)

0.6 130 °C 140 °C 150 °C

0.5 0.4 0.3 0.2 0.1 0.0 0

100

200

300 400 Time (s)

500

600

13.8 The current densities vs time profiles of test fuel cell at 130, 140 and 150 °C using CL-HBP–SA with a SO3H/Ac ratio of 78/18 in mol% at a constant potential of 300 mV under non-humidified conditions. Dry hydrogen gas and dry oxygen gas were used as fuel and oxidation gas, respectively.

0.73 V at 150 °C, respectively. The observed OCV values are much lower than the theoretical value of 1.23 V. The reason for the low OCV might be related to gas cross-over, poor contact between electrode and membrane due to somewhat poor mechanical strength of the membrane, or poor membrane electrode assembly (MEA) formation. It is necessary to improve and optimize these factors in order to attain better cell performance. Anyway, it was demonstrated that proton transport coupled with segmental motion of polymer under non-humidified conditions is possible by fuel cell test using the CL-HBP–SA membrane.

13.3

Hyperbranched polymer electrolyte with a phosphonic acid group at the periphery

13.3.1 Synthesis Hyperbranched polymers with a phosphonic acid (PA) group (HBP–PA(H, L)) and with both phosphonic acid and acryloyl groups (HBP(H, L)–PA–Ac) The two different molecular weight hyperbranched polymers with a phosphonic acid group at the periphery (HBP–PA(H, L)) were synthesized according to the procedure as shown in Scheme 13.3. The hyperbranched polymer with a hydroxyl group (HBP–OH), prepared by the polymerization of methyl 3,5-bis[(8′-hydroxy-3′,6′-dioxaoctyl)oxy]benzoate monomer



O

3

3

k

O

C

O

C

O

HBP–PE(H, L)

O

O

1

O

O

3

3

H

H

O

O

O

O

O

O

O

O

3

3

3

3

O

m

l

O

Bu3SnCl

CH2P(OEt)2

O

CH2P(OEt)2

O

O

Scheme 13.3 Synthetic route of HBP–PA(H, L).

O

O

MeO

O

O

O

O

3

3

k

C

O

C

O

(CH3)3SiBr CH3OH

1) 2)

HBP–OH(H, L)

O

O

O

O

O

O

O

O

O

O

O

3

3

3

3

H

m

l

H

O

O

3

3

k

O

C

O

HBP–PA(H, L)

O

O

C

OH

O

O

O

O

O

O

O

O

DCC, DMAP

O

O

3

3

3

3

CH2P(OEt)2

O

m

l

O

O

CH2P(OH)2

O

CH2P(OH)2

Hyperbranched polymer electrolytes for high temperature fuel cells 539 (1), was separated into two parts, a high molecular weight polymer fraction (HBP–OH(H), Mn = 14 000) and a low molecular weight one (HBP–OH(L), Mn = 3500), by the redissolution-reprecipitation method using THF as a solvent and methanol as a precipitant. The reactions of the high molecular weight HPB–OH(H) and the low molecular weight HBP–OH(L) with 4-(diethoxyphosphorylmethyl)benzoic acid in the presence of DCC and DMAP at room temperature gave the HBP–PE(H) in 76% yield and the HBP–PE(L) in 83% yield, respectively, as yellow highly viscous oils. The reactions of the HBP–PE(H) and the HBP–PE(L) with bromotrimethylsilane, followed by solvolysis in methanol, afforded the hyperbranched polymers with a phosphonic acid group at the periphery (HBP–PA(H) and HBP–PA(L)) as pale yellow solids. Both polymers (HBP–PA(H) and HBP– PA(L)) are soluble in DMF, DMAc and DMSO, but insoluble in benzene, chloroform, THF, isopropyl ether, methanol, ethanol, acetonitrile and water. The two different molecular weight hyperbranched polymers with both phosphonic acid and acryloyl groups at the periphery (HBP(H, L)–PA–Ac) were also synthesized according to the procedure as shown in Scheme 13.4. The low molecular weight HBP(L)–OH (Mn = 4200) and high molecular wight HBP(H)–OH (Mn = 11 000) were reacted with acrylic acid in the presence of DCC and DMAP at room temperature to give the HBP(L)–Ac and the HBP(H)–Ac, respectively, as yellow highly viscous oils. The reaction of the HBP(L)–Ac and the HBP(H)–Ac with 4-(diethoxyphosphorylmethyl)benzoic acid in the presence of DCC and DMAP at room temperature gave the HBP(L)–PE–Ac in 69% yield and the HBP(H)–PE–Ac in 61% yield, respectively, as yellow highly viscous oils. The reactions of the HBP(L)–PE–Ac and the HBP(H)–PE–Ac with bromotrimethylsilane, followed by solvolysis in methanol, afforded the hyperbranched polymers (HBP(L)–PA–Ac and HBP(H)–PA–Ac) with both phosphonic acid groups and acryloyl groups at the periphery as pale yellow solids, which were purified by the dissolution–reprecipitation method using DMF as a solvent and isopropyl ether a precipitant, respectively, and finally washed with isopropyl alcohol. Acryloyl group contents of the HBP(L)–PA–Ac and the HBP(H)– PA–Ac were determined by 1H NMR measurement to be 35% and 27%, respectively. Both polymers (HBP(L)–PA–Ac and HBP(H)–PA–Ac) are soluble in DMF, DMAc and DMSO, but insoluble in benzene, chloroform, THF, isopropyl ether, methanol, ethanol, acetonitrile and water. Semi-interpenetrated electrolyte membrane (HBP–PA-co-HBP–Ac) and crosslinked electrolyte membrane (CL-HBP–PA) The preparation of polymer electrolyte membrane (HBP–PA-co-HPB– Ac membrane) by the interpenetration reaction of the HBP–PA and the HPB–Ac in vacuo at 130 °C for 48 h was attempted. Unfortunately, phase



O

O

3

3

k

2) CH3OH

1) (CH3)3SiBr

HBP–OH(H, L)

O

O

O

C

O

O

O

O

O

O

O

3

3

3

3

H

m

l

H

O OH

O

O

3

3

k O

C

O

C

DCC, DMAP

HBP(H, L)–PA–Ac

O

O

O

O

O

O

O

O

O

O

O

O

O

O

O

3

3

3

3

O

m

l

O

HBP (H, L)–Ac

O

O

3

3

O

O

C

CH2P(OH)2

k

O OH

O

CH2P(OEt)2

O

O

B PO

O

O

3

3

3

O

O

O 3

O

m

O

O

O

O

O O

3

3

k

O

C

O

C

O

O

O

O

HBP(H, L)–PE–Ac

CL–HBP(H, L)–PA

DCC, DMAP 3

k

O

O

O

O

O

C

O

l

H

O

3

C

O

O

C O

O

Scheme 13.4 Synthetic route of HBP(H, L)–PA–Ac and CL-HBP(H, L)–PA membrane.

O

C

O

O

3

3

3

3

O

m

l

O

O

O

O

O

O

3

3

3

3

O

m

l

O

n

CH2P(OH)2

O

O

O

O

O CH2P(OEt)2

Hyperbranched polymer electrolytes for high temperature fuel cells 541 separation was observed in the resulting membrane, which is not suitable for fuel cell measurement. The crosslinked electrolyte membranes (CL-HBP–PA) were prepared from the HBP–PA–Acs by a solvent casting technique on parting-agent coated PET sheet at room temperature, followed by the crosslinking reaction using BPO. On the preparation of the electrolyte membrane, 10 wt% fumed silica was added to the HBP–PA–Ac solution in DMAc to obtain a large size homogeneous film in the same way as previously for CLHBP–SA. Ew of the HBP–PA(H), the HBP–PA(L), the HBP(L)–PA–Ac, and the HBP(H)–PA–Ac were determined from phosphorus content by the ICP method to be 707 for the HBP–PA(H), 720 for the HBP–PA(L), 1140 for HBP(L)–PA–Ac and 1350 for HBP(H)–PA–Ac, respectively.

13.3.2 Ionic conductivity Hyperbranched polymer with a phosphonic acid group (HBP–PA(H, L)) The temperature dependence of the ionic conductivities for the HBP– PA(H) and the HBP–PA(L) in the temperature range of 80–135 °C is shown in Fig. 13.9. Ionic conductivities of both polymers are 1.3 × 10−4 S cm−1 for HBP–PA(L) and 6.4 × 10−5 S cm−1 for HBP–PA(H), respectively, at 135 °C under dry conditions. Ionic conductivity data of both polymers were interpreted well using the VTF equation, and VTF parameters are summarized


HBP–PA(L) HBP–PA(H) 10–4

10–5

10–6 2.4

2.5

2.6

2.7

2.8

2.9

1000/T (K–1)

13.9 Arrhenius plots of the ionic conductivity for the HBP–PA(H) and the HBP–PA(L).


542

Polymer electrolytes Table 13.4 VTF parameters obtained for HBP–PA(H) and HBP–PA(L) Polymer

A (S cm−1 K−1/2)

B (K)

T0 (K)

HBP–PA(H) HPB–PA(L)

0.36 4.18

807 1685

264 192

CL-HBP(L)–PA


10–5

10–6

10–7

CL-HBP(H)–PA

10–8

2.3

2.4

2.5

2.6

2.7

2.8

2.9

1000/T (K–1)

13.10 Arrhenius plots of the ionic conductivity for the CL-HBP(L)–PA (䉫) and the CL-HBP(H)–PA (䉬) membranes.

in Table 13.4. The observation of the VTF-type behavior indicates that the dissociation of the phosphonic acid group in the polymers takes place under dry conditions and proton conduction is related to polymer chain motion as well as the hyperbranched polymers with a sulfonic acid group (HBP– SA). The HBP–PA(L) exhibits higher ionic conductivity than the HBP– PA(H), which is due to the larger A value, corresponding to the number of carrier protons, for the HBP–PA(L) in comparison with the HBP–PA(H) and the smaller free volume for the HBP–PA(H) than the HBP–PA(L) as shown in the higher Tg of the HBP–PA(H) as described in Section 13.3.3. Crosslinked electrolyte membrane (CL-HBP(H, L)–PA) The temperature dependence of the ionic conductivities for the CLHBP(L)–PA and the CL-HBP(H)–PA membranes in the temperature range of 80–150 °C is shown in Fig. 13.10. Ionic conductivities of both


Hyperbranched polymer electrolytes for high temperature fuel cells 543 Table 13.5 VTF parameters obtained for CL-HBP(L)–PA and CL-HBP(H)– PA Electrolyte membranes

A (S cm−1 K−1/2)

B (K)

T0 (K)

CL-HBP(L)–PA CL-HPB(H)–PA

2.68 0.34

1953 1683

213 231

electrolyte membranes are 1.2 × 10−5 S cm−1 for the CL-HBP(L)–PA membrane and 2.6 × 10−6 S cm−1 for the CL-HBP(H)–PA membrane, respectively, at 150 °C under dry conditions. Both the CL-HBP(L)–PA and the CL-HBP(H)–PA membranes showed lower ionic conductivities by about one order than the HBP–PA(H) and the HBP–PA(L). This is due to a lower content of the phosphonic acid groups, that is, the loss in the carrier number caused by the replacement of a phosphonic acid group with an acryloyl group in the same molecule. Ionic conductivity data of both electrolyte membranes were analyzed well with the VTF equation, and VTF parameters are summarized in Table 13.5. This indicates that the proton conduction in the electrolyte membranes is cooperated by polymer chain motion. And also, the CL-HBP(L)–PA membrane showed higher ionic conductivity than the CL-HBP(H)–PA membrane, which is due to the larger A value, corresponding to the number of carrier protons, for the CL-HBP(L)–PA membrane in comparison with the CL-HBP(H)–PA membrane and the smaller free volume for the CL-HBP(H)–PA membrane than the CL-HBP(L)–PA membrane as shown in the higher Tg of the CL-HBP(H)–PA membrane (see Section 13.3.3).

13.3.3 Thermal properties Hyperbranched polymer with a phosphonic acid group (HBP–PA(H, L)) DSC measurements were carried out in the temperature range of −150– 150 °C for HBP–PA(H) and HBP–PA(L) under argon. DSC traces for HBP–PA(H) and HBP–PA(L) are shown in Fig. 13.11, indicating that they are completely amorphous. The Tg were found to be 42.7 °C for HBP– PA(H) and 31.7 °C for HBP–PA(L), respectively. Lower ionic conductivity observed for HBP–PA(H) in comparison with HBP–PA(L) is ascribed to the higher Tg of the HBP–PA(H), that is, the reduction of free volume in HBP–PA(H). TG measurements were carried out under air to investigate the thermal stability of HBP–PA(H) and HBP–PA(L). The TG traces for both polymers are shown in Fig. 13.12, where the polymers shows a two-stage


544

Polymer electrolytes Tg = 31.7 °C

Exo

HBP–PA(L)

Tg = 42.7 °C

Endo

HBP–PA(H)

–120 –100 –80 –60 –40 –20

0

20

40

60

80 100

Temperature (°C)

13.11 DSC traces of the HBP–PA(H) and HBP–PA(L) under argon.

100

TG (%)

80 HBP–PA(H)

60 40 20

HBP–PA(L)

0 0

200

400

600

800

1000

Temperature (°C)

13.12 Thermogravimetry traces of the HBP–PA(H) and HBP–PA(L) under air.

decomposition and begins to decompose at a temperature of about 300 °C. They have much better thermal stability than the hyperbranched polymer with sulfonic acid groups (HBP–SA), and such high thermal stability is due to the phosphorous atoms present in the polymers. The weight loss in the temperature range of 300–400 °C amounted to 40%, which corresponds approximately to the weight percent of the ethylene oxide chain unit in the polymers, indicating that the decomposition might take place an ether chain in the polymers. Anyway, they have suitable thermal stability as an electrolyte in the polymer electrolyte fuel cell operating under non-humidified conditions.



Tg = 38.8 °C

Exo.

CL-HBP(H)–PA

CL-HBP(L)–PA

Endo.

Tg = 26.8 °C

–80 –60 –40 –20

0

20

40

60

80

100 120

Temperature (°C)

13.13 DSC traces of the CL-HBP(L)–PA (—) and the CL-HBP(H)–PA (---) membranes under argon.

Crosslinked electrolyte membrane (CL-HBP(H, L)–PA) DSC measurements were carried out in the temperature range of −100– 150 °C for the CL-HBP(L)–PA and the CL-HBP(H)–PA membranes under argon. DSC traces for CL-HBP(L)–PA and CL-HBP(H)–PA are shown in Fig. 13.13, indicating that they are completely amorphous. The Tg values were found to be 26.8 °C for the CL-HBP(L)–PA membrane and 38.8 °C for the CL-HBP(H)–PA membrane, respectively. The lower ionic conductivity observed for the CL-HBP(H)–PA membrane in comparison with the CL-HBP(L)–PA membrane is ascribed to the higher Tg of the CL-HBP(H)– PA membrane, that is, the reduction of free volume in the CL-HBP(H)-PA membrane. TG measurements were carried out under air to investigate the thermal stability of the CL-HBP(L)–PA and the CL-HBP(H)–PA membranes. The TG traces for both electrolyte membranes are shown in Fig. 13.14, where the polymers shows a two-stage decomposition and begins to decompose at the temperature of about 300 °C. They have much better thermal stability than the hyperbranched polymer with a sulfonic acid group (CL-HBP–SA), and such high thermal stability comes from the presence of phosphorus atoms in the polymers. The weight loss in the temperature range of 300–400 °C amounted to 40%, which corresponds approximately to the weight percent of the ethylene oxide chain unit in the polymers, indicating that the decomposition might take place at an ether chain in the polymers.


546

Polymer electrolytes 100 80 TG (%)

CL-HBP(L)–PA 60 40 CL-HBP(H)–PA

20 0 0

200

400

600

800

1000

Temperature (°C)

13.14 Thermogravimetry traces of the CL-HBP(L)–PA (—) and the CL-HBP(H)–PA (---) membranes under air.

Anyway, it is certain that they have suitable thermal stability as electrolyte membranes for the high temperature fuel cell under non-humidified conditions.

13.3.4 Fuel cell measurement On preparation of a semi-interpenetrated electrolyte membrane (HBP– PA-co-HPB–Ac membrane), phase separation took place in the membrane. The resulting film is not suitable for fuel cell measurement. Therefore, fuel cell measurement was carried out using CL-HBP(L)–PA and the CLHBP(H)–PA membranes. The polarization (I–V) characteristics of the CL-HBP(L)–PA and the CL-HBP(H)–PA membranes were measured using a test fuel cell. The voltage profiles observed at a cell temperature of 150 °C under non-humidified conditions are shown in Fig. 13.15. Dry hydrogen and dry oxygen gases were used as the fuel and the oxidant, respectively. The fuel cell using the CL-HBP(L)–PA membrane showed larger current density than that of the cell using the CL-HBP(H)–PA membrane. This result corresponds to the fact that the CL-HBP(L)–PA membrane has higher ionic conductivity than the CL-HBP(H)–PA one. The OCV were obtained to be 0.76 V for the CL-HBP(L)–PA membrane and 0.53 V for the CL-HBP(H)–PA one, respectively. Both OCV values are much lower than the theoretical value of 1.23 V. The reason for the low OCV might be related to gas cross-over or internal micro-short circuit due to poor mechanical strength of the membranes. Moreover, current density with these membranes is lower than that of the crosslinked membrane with a sulfonic acid group (CL-HBP–SA),


Hyperbranched polymer electrolytes for high temperature fuel cells 547 0.8

CL-HBP(L)–PA

Potential (V)

0.6

0.4

0.2 CL-HBP(H)–PA 0.0 0

0.1

0.2

0.3

0.4

0.5

0.6

Current density (mA cm–2)

13.15 The current vs. voltage characteristics of test fuel cells at 150 °C using the CL-HBP(L)–PA (—) and CL-HBP(H)–PA (---) membranes as polymer electrolytes under non-humidified conditions. Dry hydrogen gas and dry oxygen gas were used as fuel and oxidation gas, respectively.

ascribed to the lower ionic conductivity of CL-HBP(H, L)–PA in comparison with CL-HBP(H, L)–SA. Although it is necessary to improve and optimize these factors in order to attain a better cell performance, the concept of the proton conduction coupled with the polymer chain motion under non-humidified conditions was demonstrated on the basis of a successful fuel cell test using the CL-HBP(H, L)–PA membranes similar to the cases of CL-HBP(H, L)–SA.

13.4

Conclusions

To investigate the anhydrous proton-conducting membranes for use in the polymer electrolyte fuel cells at high temperature, hyperbranched polymers with a sulfonic acid group (HBP–SA), with a phosphonic acid one (HBP– PA), with both sulfonic acid and acryloyl groups (HBP–SA–Ac), and with both phosphonic acid and acryloyl groups (HBP–PA–Ac), and interpenetrated electrolyte membranes (HBP–SA-co-HBP–Ac and HBP–PA-coHBP–Ac) prepared by the copolymerization of HBP–SA or HBP–PA with the hyperbranched polymer with an acryloyl group (HBP–Ac), and crosslinked electrolyte membranes (CL-HBP–SA and CL-HBP–PA) prepared by the homopolymerization of the HBP–SA–Ac or the HBP–PA–Ac were synthesized. Ionic conductivities, thermal properties and the fuel cell performance with these polymer-based electrolyte membranes under nonhumidified conditions were investigated. The ionic conductivities of all


548


HBP–SA, HBP–SA–Ac, HBP–PA and HBP–PA–Ac polymers, interpenetrated electrolyte membrane HBP–SA-co-HBP–Ac, and the crosslinked membranes CL-HBP–SA and CL-HBP–PA showed the VTF-type temperature dependence. These polymers and membranes are thermally stable up to 260 °C, and they had suitable thermal stability as an electrolyte in the polymer electrolyte fuel cell operating under non-humidified conditions. Fuel cell measurement using a single membrane electrode assembly cell with crosslinked membranes CL-HBP–SA and CL-HBP–PA was successfully performed under non-humidified conditions, and polarization curves were observed. The concept of the proton conduction coupled with the polymer chain motion was proposed as one possible approach toward high temperature fuel cells.

13.5

References

fulcher g s (1925), ‘Analysis of recent measurements of the viscosity of glasses’, J Am Ceram Soc, 8, 339–355. gray f m (1991), Solid Polymer Electrolytes: Fundamentals and Technological Applications, New York, VCH Publishers. gray f m (1997), Polymer Electrolytes, London, The Royal Society of Chemistry. herz h g, kreuer k d, maier j, scharfenberger g, schuster m f h and meyer w h (2003), ‘New fully polymeric proton solvents with high proton mobility’, Electrochem Acta, 48, 2165–2171. itoh t, ikeda m, hirata n, moriya y, kubo m and yamamoto o (1999), ‘Ionic conductivity of the hyperbranched polymer–lithium metal salt systems’, J Power Sources, 81–82, 824–829. itoh t, hamaguchi y, uno t, kubo m, aihara y and sonai a (2006a), ‘Synthesis, ionic conductivity, and thermal properties of proton conducting polymer electrolyte for high temperature fuel cell’, Solid State Ionics, 177, 185–189. itoh t, hamaguchi y, hirai k, uno t, kubo m, aihara y and sonai a (2006b), ‘Synthesis, ionic conductivity, and thermal properties of hyperbranched polymer with phosphonic acid groups at the chain ends for high temperature fuel cell’, in Fuller T, Bock C, Cleghorn S, Gasteiger H, Jarvi T, Mathias M, Murthy M, Nguyen T, Ramani V, Stuve E and Zawodzinski T, ECS Transactions, Vol. 3, No.1, The Electrochemical Society, 113–121. itoh t, hirai k, tamura m, uno t, kubo m and aihara y (2008), ‘Anhydrous protonconducting electrolyte membranes based on hyperbranched polymer with phosphonic acid groups for high-temperature fuel cells’, J Power Sources, 178, 627–633. itoh t, sakakibara t, takagi y, tamura m, uno t, kubo m and aihara y (2009), ‘Proton-conducting electrolyte membranes based on hyperbranched polymer with a sulfonic acid group for high-temperature fuel cells’, Electrochim Acta, 55, 1419–1424. jannasch p (2003), ‘Recent development in high-temperature proton conducting polymer electrolyte membranes’, Curr Opin Colloid Interf Sci, 8, 96–102. karlsson l e and jannasch p (2004), ‘Polysulfone ionomers for proton-conducting fuel cell membranes: sulfoalkylated polysulfones’, J Membr Sci, 230, 61–70.


Hyperbranched polymer electrolytes for high temperature fuel cells 549 kim j d, mori t, hayashi s and honma i (2007), ‘Anhydrous proton-conducting properties of nafion-1,2,3-triazole and nafion-benzimidazole membranes for polymer electrolyte fuel cells’, J Electrochem Soc, 154, A290–A294. kreuer k d (1996), ‘Proton conductivity: materials and applications’, Chem Mater, 8, 610–641. kreuer k d (2001), ‘On the development of proton conducting polymer membranes for hydrogen and methanol fuel cells’, J Membr Sci, 185, 29–39. kreuer k d, paddison s j, spohr e and schuster m (2004), ‘Transport in proton conductors for fuel-cell applications: simulations, elementary reactions, and phenomenology’, Chem Rev, 104, 4637–4678. persson j c and jannasch p (2003), ‘Self-conducting benzimidazole oligomers for proton transport’, Chem Mater, 15, 3044–3045. persson j c and jannasch p (2005), ‘Intrinsically proton-conducting benzimidazole units tethered to polysiloxanes’, Macromolecules, 38, 3283–3289. savadogo o (1998), ‘Emerging membranes for electrochemical systems: (I) solid electrolyte membranes for fuel cell systems’, J New Mater Electrochem Syst, 1, 47–66. savadogo o (2004), ‘Emerging membranes for electrochemical systems part II. High temperature composite membranes for polymer electrolyte fuel cell (PEFC) applications’, J Power Sources, 127, 135–161. schuster m, meyer w h, wagner g, herz h g, ise m, schuster m, kreuer k d and maier j (2001), ‘Proton mobility in oligomer-bound proton solvents: imidazole immobilization via flexible spacers’, Solid State Ionics, 145, 85–92. schuster m f h, meyer w h, schuster m and kreuer k d (2004), ‘Toward a new type of anhydrous organic proton conductor based on immobilized imidazole’, Chem Mater, 16, 329–337. scrosati b (1993), Applications of Electroactive Polymers, London, Chapman and Hall. steele b c h and heinzel a (2001), ‘Materials for fuel-cell technologies’, Nature, 414, 345–352. tammann g and hesse w (1926), ‘The dependence of viscosity upon the temperature of supercooled liquids’, Anorg Allg Chem, 156, 245–257. vogel h (1921), ‘The law of the relation between the viscosity of liquids and temperature’, Phys Z, 22, 645–646. wainright j s, wang j t, weng d, savinell r f and litt m (1995), ‘Acid-doped polybenzimidazoles: a new polymer electrolyte’, J Electrochem Soc, 142, L121–L123. wang j t, savinell r f, wainright j s, litt m and yu h (1996), ‘A H2/O2 fuel cell using acid doped polybenzimidazole as polymer electrolyte’, Electrochem Acta, 41, 193–197.


14 Polymer electrolytes as solid solvents and their applications L. Y E and Z. F E N G, Beijing Institute of Technology, China

Abstract: This chapter details the preparation and mechanism of polymer electrolytes and their applications in electrochemical fields, such as lithium ion batteries, fuel cells, alkaline batteries, supercapacitors, solar cells, electrochromic devices and the like. Polymer electrolytes used in lithium ion batteries are divided into three categories: solid, gel and composites. Recent progress made in these three categories is highlighted. Moreover, in addition to the lithium ion battery, applications of polymer electrolytes in other electrochemical fields and their ion conducting performance are also briefly described. Key words: polymer electrolyte, lithium ion battery, ion conductivity, safe performance, oxetane-derived polyether.

14.1

Introduction

A polymer electrolyte is also referred to as a solid solvent that possesses ion transport properties similar to that of the common liquid ionic solution. It usually comprises a polymer matrix and electrolyte, wherein the electrolyte such as a lithium salt dissolves in a polymer matrix. The research and development of polymer electrolytes have drawn great attention in the last three decades as they are applied in many electrochemical devices such as lithium batteries, nickel – metal hydride (Ni/MH) batteries, fuel cells/direct methanol fuel cells, supercapacitors, electrochromic devices and the like (Gray, 1991; Stephan, 2006). In recent years, information technology (IT) has pervaded every field of society throughout the world. During this process, electrochemical devices such as advanced batteries and fuel cells are expected to play a vital role, since they are the most convenient and highly efficient mobile power sources (Takamura, 2006). Additionally, an important issue in the twentyfirst century is to reduce the consumption of fossil fuels, otherwise, the warming of our global atmosphere will drastically change the climate of the ecosystem which is critical for our continued survival. The utilization of cells 550 © Woodhead Publishing Limited, 2010

Polymer electrolytes as solid solvents and their applications

551

seems to make it possible to save total energy, reduce the release of CO2 and suppress the heating of the Earth. Hence the manufacture and usage of a safe, durable, reliable and powerful battery are highly desirable. The polymer electrolyte has being sought to replace the liquid electrolyte which predominates in commercialized cells nowadays. An electrolyte is commonly responsible for ion transportation and thus is necessary for all electric devices where ion transportation is needed. For example, a cell is usually composed of two electrodes and electrolyte wherein the electrolyte is a medium for ion conduction. However, there are some disadvantages related to the liquid electrolyte such as leakage, poor compatibility with electrodes, safety hazards and so on. Sometimes we learn from the media that the lithium cell in cell (mobile) phones or in notebook is detrimental to users due to the explosion caused by short circuits, over-charging and other ab-normal working conditions. It is reasonable to worry about safety when bigger and more powerful cells are manufactured and applied, for example, in electric vehicles. As a consequence the polymer electrolyte is of tremendous interest to researchers because of its many advantages such as no leakage, flexible geometry, excellent safe performance and good compatibility with electrodes compared with its liquid counterpart. As a key technique in cell manufacture, progress in polymer electrolytes will bring about a revolution in cell manufacture. At the same time, it will make a great contribution to saving on total energy consumption, reducing the release of CO2 and suppressing the heating of the Earth. A polymer electrolyte was firstly reported by Wright and co-workers in 1973 although its technological importance was not recognized until the early 1980s (Fenton et al., 1973; Armand and Duclot, 1978). However, the first generation polymer electrolyte based on the poly(ethylene oxide) (PEO)–LiX system only offered a very low ion conductivity in the order of 10−8 S/cm at ambient temperature, which excluded it from practical applications where the ion conductivity needs to be at least more than 10−3 S/cm. The main target for the research and development of polymer electrolyte during the last three decades was how to enhance the ion conductivity. Three kinds of polymer electrolytes – solid polymer electrolyte (SPE), gel/ plasticized polymer electrolyte (GPE) and composite polymer electrolyte (CPE) – were developed for this goal. The first example of ‘dry solid’ polymer electrolyte invented by Wright is the PEO–based system. Since this system does not contain any organic liquid and the polymer host is thus used as solid solvent, the safety performance is excellent and the mechanical strength and flexible geometry are good. However, the ion conductivity is too poor to be used. The second category of polymer electrolyte is called the GPE which is neither liquid nor solid or conversely both liquid and solid (Gray, 1991), and the gel


552


possesses both the cohesive property of solids and the diffusive property of liquids. The CPE incorporates electrochemically inert fillers into polymer matrices (Weston and Steele, 1982; Capuono et al., 1991; Krawiec et al., 1995; Croce et al., 1998, 1999; Wieczorek et al., 1998). In general, high surface area particulate fillers such as ZrO2, TiO2, Al2O3 and hydrophobic fumed silica are incorporated into the polymer matrices to create the CPE (Golodnitsky et al., 1997; Dai et al., 1998). The advantages of incorporating the fillers are twofold. One is to raise the ion conductivity at low temperatures and another is to improve the stability at the interface with electrodes (Borghini et al., 1995; Mastragostino et al., 1999; Qian et al., 2001; Li et al., 2001, 2003; Appetecchi et al., 2003; Kwang et al., 2003; Jiang et al., 2005). As polymer electrolytes have been broadly applied in many electric fields, we subsequently introduce these applications one by one in the following sections.

14.2

Structure of lithium ion battery

The lithium ion rechargeable battery (Fig. 14.1) is also known as a swing battery or rocking chair battery since two-way movement of lithium ions occurs between anode and cathode through the electrolyte during charge and discharge processes (Wakihara, 2001). When discharging, the lithium ion, which is produced by the reduction of lithium metal oxide of cathode materials, passes through the electrolyte to come into the anode. In contrast, when recharging, the lithium ion will pass through the electrolyte again to come back to cathode to attend the oxidation in the cathode. The commercial electrolyte used in batteries is nowadays composed of liquid electrolyte and porous plastic separator, wherein the typical liquid electrolyte is a carbonic acid ester compound, such as ethylene carbonate (EC), propylene carbonate (PC), dimethylene carbonate (DMC), etc. and the separator is made of polyethylene, polypropylene and the like. As mentioned above, a liquid electrolyte introduces some disadvantages such as leakage, poor compatibility with the electrode and safety hazards, which heavily impede the further development of lithium cells. Hence, the polymer electrolyte is intended to replace both the separator and liquid electrolyte.

14.3

Advantages of polymer electrolytes in lithium ion batteries

In comparison to commercially broadly applied liquid electrolytes, polymer electrolytes have many advantages, as follows (Song et al., 1999).



553

Power supply e– Charge Load

e–

Discharge

+

– Electrolyte Li+

Li+ Li+ Li+ Li+ Li+

Li+

Li+

Li+ Li+

Li+ Li+

Li+ Li+ Li+ Li+ Li+ Cathode Li+ Li+ Li+ Li+ Li+

Li+

Li+ Li+ Li+ Li+ Li+

Li+

Lithium metal oxide

Li+ Li+ Li+ Li+ Li+

Anode Li+ Li+ Li+ Li+ Carbon

14.1 Structure of a lithium ion battery (adapted from Wakihara, 2001).

14.3.1 Improved safety The solid state construction of a polymer electrolyte battery is more tolerant to shock, vibration and mechanical deformation. Since there is no or little liquid content within the electrolyte, cells can be packaged in a vacuumed flat ‘plastic bag’ other than a rigid metal container which is prone to corrosion. This unique feature prevents the build-up of internal pressure and, hence, removes the possibility of explosion (Owens and Osaka, 1997).


554


14.3.2 Enhanced endurance to varying electrode volume during cycling Polymer electrolytes are very compliant and this character allows the construction of solid state rechargeable batteries in which the polymer conforms to the volume changes of both electrodes that occur during recharge–discharge cycling.

14.3.3 Reduced reactivity with liquid electrolyte It is generally accepted that no solvent is thermodynamically stable towards lithium, even carbonaceous anodes. Polymer electrolytes, owing to their solid-like nature and much lower liquid content, are less reactive than their liquid electrolyte counterparts.

14.3.4 Better shape flexibility and manufacturing integrity Owing to the increasing need for smaller and lighter batteries, the battery shape factor has become one of the major design concerns. A film-like polymer electrolyte battery is quite promising from this aspect (Murata, 1995). Another feature associated with polymer electrolyte batteries is the manufacturing integrity; all elements, both the electrolyte and electrodes, of a cell can be laminated automatically via well-developed coating technology (Bierwagen, 1992).

14.4

Main properties of polymer electrolytes

From a practical point of view, the polymer electrolyte for rechargeable lithium ion batteries should satisfy the following requirements (Koksbang et al., 1996).

14.4.1 Ion conductivity To achieve the performance level of liquid electrolyte-based systems which can be discharged at current densities of up to several mA/cm2, the polymer electrolyte should possess ion conductivities approaching or beyond 10−3 S/ cm at ambient temperature. As the polymer electrolyte can be processed into a self-standing membrane, the ion conductivity is usually measured via AC impedance analysis with an electrochemical cell consisting of the polymer electrolyte film sandwiched between two blocks of stainless steel. Figure 14.2 shows a typical AC impedance plot, where the value of the point of intersection between the plot and the x-axis represents the value



555

3500 3000

Z´ (Ω)

2500 2000 1500 1000 500

Rb

0 1000 1200 1400 1600 1800 2000 2200 2400 Z´ (Ω)

14.2 Typical AC impedance of a polymer electrolyte.

of the bulk resistance of solid polymer electrolyte. The corresponding ion conductivity is calculated according to the following equation (Wen et al., 2002). σ = l/(Rb × A)

[14.1]

where Rb is the bulk resistance of polymer electrolyte, l is the film thickness and A is surface area of electrode.

14.4.2 Transference number The transference number is defined as the ratio of the electric current derived from the cation to the total electric current. If the transference number is close to 1, it implies that the ion conducting performance in the polymer electrolyte is mainly accomplished by the cation. A large transference number can reduce concentration polarization of electrolytes during charge–discharge steps, and thus produce higher power density. It is highly desirable that the transference number of lithium ions approaches 1 in an electrolyte system. However, many existing electrolyte systems, either liquid or polymeric, have transference numbers less than 0.5 (Song et al., 1999; Wakihara, 2001; Li et al., 2003).

14.4.3 Chemical, thermal and electrochemical stabilities Since a polymer electrolyte membrane is interposed between the cathode and the anode, its chemical stability must be such that no undesired chemical reactions occur when the electrodes come into direct contact. In addition, in order to operate in an appropriate temperature range, polymer


556


Current (μA/cm2)

2.5 2.0 1.5 1.0 0.5 Electrochemical window 0.0 –0.5 15

20

25

30

35

40

45

50

55

Potential (V)

Current (mA/cm2)

14.3 Cyclic voltammogram curve of a polymer electrolyte.

Electrochemical window

0

1

2

3

4

5

Potential (V) (Li/Cu)

14.4 Linear sweep voltammetry curve of a polymer electrolyte.

electrolytes must possess good thermal stability. Finally, they must also have an electrochemical stability domain extending from 0 V to as high as 4.5 V. In general the electrochemical stability can be characterized by an electrochemical window which is measured via a cyclic voltammogram measurement (Fig. 14.3) or linear sweep voltammetry (Fig. 14.4). The flat areas occurring in both figures represent the electrochemical window.



557

14.4.4 Mechanical strength Manufacturability is overwhelmingly the most important factor to be considered when moving battery technology from laboratory to pilot or full production. Although many electrolyte systems can be fabricated as free-standing membranes with favorable electrochemical properties, their mechanical strength still needs to improve for manufacturing by conventional large-scale coating processes.

14.5

Solid polymer electrolytes applied in lithium ion batteries

14.5.1 Ion conducting mechanism in solid polymer electrolytes One of the typical preparation procedures of the solid polymer electrolytes is as follows. The polymer and lithium salts are first dissolved in a selective solvent under dry atmosphere to form a homogeneous solution. Then, the mixture is coated onto a PTFE plate and dried in vacuo for several days to give a polymer electrolyte membrane. The ion conducting mechanism is shown in Fig. 14.5. The lithium salts dissolve and disassociate in the polymer matrix to set off free ions. The ion conduction mainly takes place in the amorphous phase, and the lithium ion can coordinate with oxygen atoms in the PEO chain in a fixed ratio. These complexed ions move with the motion of a segment of polymer chain. Consequently, when the electric field is applied towards the polymer electrolyte, the cations move freely from one electrode to another (Fontenella et al., 1983; Song et al., 1999). As discussed above, a polymer electrolyte for rechargeable lithium ion batteries should satisfy at least four requirements for practical applications. However, among those requirements the ion conductivity is the most vital and should be put first. The following discussion focuses on how to enhance the ion conductivity of solid polymer electrolytes. As a typical polymer

O O O

O

O

O

Li+ O

Li+ O

O

O

O O

O O

O

Li+ O O

O

14.5 Ion conducting mechanism in the solid polymer electrolyte.


O

558


electrolyte is usually composed of two components, polymer matrix and lithium salt, these two components are the key factors to affect the ion conductivity. When lithium salts dissolve, not all the salts can dissociate and become free ions. On the contrary, there exist at least three states of lithium salts in a polymer matrix, i.e. free ions, ion pairs and aggregations. The first requirement for lithium salts is their solubility in the polymer matrix. Traditional lithium salts used in the polymer electrolytes are LiPF4, LiClO4, LiBF4, etc. Recently lithium salts with a large anion such as LiCF3SO3 and lithium trifluoromethanesulfonimide (LiTFSI) have attracted much attention, because a larger anion radius promotes ion conductitivity as it is easier to dissociate in the polymer matrix and set off free lithium cations that increase the ion conductivity. Besides the anion property, the amount of lithium salts added also affects the ion conductivity. The more free cations, the higher the ion conductivity. However, the motion performance of a segment of the polymer matrix is impaired with increasing amount of lithium salts due to the coordination between cation and oxygen atoms in the polymer chain. The very low ion conductivity in the PEO–LiX system is due to the poorer motion performance of the polymer segment caused by the crystalline structure. According to the ion conducting mechanism, the ion conductivity can be improved by increasing the motion ability of the polymer segment. Generally, the motion ability is increased through modification of the molecular structure, such as copolymerizing, grafting, hyperbranching or blending with more flexible polymers.

14.5.2 Graft polymer electrolyte To demolish the crystalline structure of PEO, epichlorohydrin was first copolymerized with ethylene oxide to introduce a branch structure in order to increase the ion conductivity to 4.1 × 10−5 S/cm (Yu and Zhou, 2002). Although this copolymer was not decribed as a pure graft polymer, it at least revealed that the ion conductivity can be improved by incorporating the side chain to depress the crystalline tendency of PEO. Jannasch (2001) grafted a PEO side chain onto the polyethylene chain, leading to the ion conductivity increasing to 3.2 × 10−6 S/cm at room temperature after doping with LiTFSI. In addition, they also prepared a branch polyacrylate copolymer where the side chain is composed of poly(ethylene glycol) (PEG), and alkane or fluorinated alkane (Gavelin et al., 2002). As the resulting polymer is amorphous, the ion conductivity reached 8 × 10−5 S/cm. As shown in Fig. 14.6, Nishimoto et al. (1998) synthesized a comb copolyether P(EO/MEEGE) (MZEEGE -2-(2-methoxyethoxy)-ethyl glycidyl



(

O

(x (

O O

(1–x

n

559

O O

14.6 Comb-like polymer synthesized by Nishimoto et al. (1998).

ether) of high molecular weight. The ion conductivity rose while the crystallinity degree decreased with increasing content of MEEGE due to its flexible nature. Keeping the content of MEEGE at 9% (mol) in the copolymer, the ion conductivity was found to reach 10−4 S/cm at 30 °C and 10−3 S/cm at 80 °C with a 4 V electrochemical window. This provided the polymer electrolyte matrix with two advantages by incorporating the conducting polyether oligomers: one is the decrease of the crystalline degree and the increase of the motion performance of the polymer segment, and another is the enhancement of the ion conducting performance per molecule compared with the conducting-inert side chain. Miwa et al. (2001, 2002) introduced different lengths of hydroxyl terminated PEG oligomers into 3-hydroxymethyl-3′-ethyloxetane (HEO). Then, they prepared branch polymers by cationic ring-opening polymerization (CROP) initiated with BF3⋅Et2O. If PEG is mono-terminated by hydroxyl, the polymer formed is linear. On the other hand, a crosslink/hyperbranched polymer network is obtained. The resulting polymer electrolyte was reported to have good electrochemical properties with the ion conductivities reaching 10−4 S/cm at room temperature. As crown ether is an excellent ligand to coordinate with various cations, Omata and Makoto (1998) managed to apply it into the polymer electrolyte. However, it must possess an appropriate coordination capability with cations which is neither too strong nor too weak, because too strong coordination impairs the free transfer of lithium ions while too weak coordination limits the solubility of lithium ions. Consequently the acrylic monomer was chosen to connect with a serial of crown ethers through a spacer group. The maximum ion conductivity of these novel SPEs containing crown ether reached 9 × 10−3 S/cm at room temperature. It was also the highest ion conductivity reported in the literature to date. However, this SPE is expensive because of the high price of crown ether used and tedious synthetic protocol. We have also undertaken research work on graft polymer electrolytes. Given that both lowering the glass transition temperature (Tg) and eliminating the crystalline tendency of polymer matrix are efficient ways to improve ion conductivity, a first goal focused on the preparation of comblike polyethers (Ye et al., 2005). Two monomers, 3-(2-cyano ethoxy)methyl-


Polymer electrolytes O O

n

O n

Homopolymerization O

O

O

BF3·Et2O/BDO or TfOH

(

BF3·Et2O/BDO or TfOH

Homopolymerization

(

O

O

(

560

(

3 CN PCEMO

3 CN

–OCH3 PMEMO

OCH3

Copolymerization BF3·Et2O/BDO

O

On

(

m O

(

O

3 CN

OCH3

14.7 Comb-like polyether and copolyether. BDO refers to 1,4-Butanediol.

and 3-methoxytriethoxylmethyl-3′-methyloxetane (CMMO and MEMO), were synthesized from 3-hydroxymethyl-3′-methyloxetane (HMO). As shown in Fig. 14.7, a kind of comb-like polyether was prepared via CROP of these two monomers. The ion conductivity of these comb-like polymers was measured in Fig. 14.8, where the homopolyether PMEMO with a long and flexible side chain gave the maximum ion conductivity of 2.6 × 10−4 S/cm at 30 °C and 1.9 × 10−3 S/cm at 80 °C, respectively. As stated above, PMEMO electrolyte held the relatively higher ion conductivity. However, the main drawback of PMEMO is that its mechanical strength is too poor to form a self-standing polymer electrolyte membrane because of its lower molecular weight. To this end, we managed to convert it into polyurethane (PU) for improving its mechanical strength. Furthermore, poly(vinylidene fluoride) (PVdF) was needed to be added so as to form robust electrolyte. The maximum ion conductivity of the resulting PU–PVdF electrolyte reached 2.1 × 10−4 S/cm at 30 °C and 1.7 × 10−3 S/cm at 80 °C, respectively (Ye et al., 2007a).


Lg (σ/S × cm–1)

Polymer electrolytes as solid solvents and their applications –3.4 –3.6 –3.8 –4.0 –4.2 –4.4 –4.6 –4.8 –5.0 –5.2 –5.4 –5.6 –5.8 –6.0

561

PM EMO Copolyether PC EMO

28

29

30

31

32

33

1000T –1 (K–1)

14.8 Ion conductivity of comb-like polyether PMEMO, PCEMO and its copolyethers (quoted from Ye et al., 2005).

In addition to the PU–PVdF blend polymer, the crosslink polymer was also used as a polymer matrix. An oxetane-derived monomer with a vinyl group, 3-acryloyloxy-methyl-3′-methyloxetane (AMO) was prepared and copolymerized with MEMO. The CROP of AMO was first carried out, and then the radical polymerization proceeded to give a crosslinked polymer electrolyte film. The mechanical strength was evidently increased with crosslinking, whereas the ion conductivity was decreased with a maximum ion conductivity of 1.4 × 10−5 S/cm at 30 °C and 1.3 × 10−4 S/cm at 80 °C at the ratio of O : Li = 20 (Ye et al., 2007b). Although the crosslink is very efficient to raise the mechanical strength of the polymer matrix, it leads to an unacceptable reduction in the ion conductivity.

14.5.3 Hyperbranched polymer electrolyte Hyperbranched polymers have received much attention due to their unique molecular structure and their ease of preparation compared with dentrimers. Various hyperbranched polymers have been synthesized and explored for potential applications (Vandenberg et al., 1989; Bednarek et al., 1999, 2001, 2002; Magnusson et al., 1999, 2000; Yan et al., 2000; Chen et al., 2002; J. Xu et al., 2002; Bednarek, 2003; Mai et al., 2003; Y. Y. Xu et al., 2004). The hyperbranched molecular structure remarkably suppresses the chain crystallization so as to enlarge the amorphous phase compared with the linear polymers. Itoh and coworkers reported a hyperbranched polyester consisting of di- and triethylene glycols and 3,5-dioxybenzoate branching units (Li


562

Polymer electrolytes H2N NH2 H2N NH2

NH2 NH 2

NH2 NH2 NH2 N NH2 N NH2 NH2 N NH2 N N NH2 N NH2 N N NH2 N NH2 N NH2 N N NH2 N NH2 NH2

NH2

H2N H2N H2N

NH2

N

H2N

N

N

N

N N N

H2N

N

N

H2N H2N

N N

N

N

H2N

N N

N

N

H2N

N

H2N N H2N N H2N H2N H2N H2N H2N H2N

N

N

N

N

N

N N

N

N

N

N N

N

N

N

N

N

N N N

N N

N

NH2 NH2

N

N

NH2 NH2

NH2 NH2 NH2 NH2

NH2 NH2 NH2

N NH2 N N H2N NH2 H2NH2N H2N H2N H2N H2NNH2 NH2 NH2 H2N

14.9 Dentrimer synthesized by Dillon and Shriver (2001).

et al., 2003). However, owing to the poorer motion ability of the aromatic moieties, its relative lower ion conductivity (~10−6 S/cm) was unsatisfactory to meet the practical standard for lithium ion batteries. Besides the hyperbranched polymer, Dillon and Shriver (2001) prepared a novel dentrimer DAB-AB-64 as shown in Fig. 14.9, and its ion conductivity was reported to reach 1 × 10−4 S/cm at room temperature. Recently a great deal of research work has been carried out on new hyperbranched polyethers synthesized from HMO or HEO by CROP. Compared with the hyper-branched polyesters as mentioned previously, these hyperbranched polyethers were synthesized in a one-pot protocol. Motivated by this novel synthetic strategy, a new oxetane-derived 3-(2-(2-(2-hydroxyethoxy)ethoxy)ethoxymethyl)-3′-methyloxetane (HEMO) monomer was prepared in our laboratory. Subsequently, a novel hyperbranched polyether, referred as PHEMO (Fig. 14.10), was synthesized



O

O O

O

OH

O OH

O

563

+

+

3 O

O Branching further

3

Branching OH

O

–H

+

–H 3

OH

+

HO 3

O

HO 3 HO O

O

O

O

O

O n HEMO 3

3 O

O

O

O

O

O

O

O

O

3

3

O

HO

3

3

HO

OH

HO

14.10 Synthesis and molecular structure of PHEMO (from Ye et al., 2006a).

by CROP in the presence of BF3⋅Et2O (Ye et al., 2006a). The motion performance of the resulting polyether segment was expected to be enhanced significantly owing to the lack of stiff groups, such as phenyl groups in the backbone, as well as for the amorphous phase inherently derived from the


564


hyperbranched structure. After doping with LiTFSI, the ion conductivity of PHEMO was seen to reach 5.6 × 10−5 S/cm at room temperature and 6.3 × 10−4 S/cm at 80 °C, respectively. Here the motion performance of triethylene glycol oligomer is clearly a key factor to affect the ion conducting performance. In PHEMO, however, it was bound to the hyperbranched main chain with a poor motion performance. To attain the higher ion conductivity, triethylene glycol should be liberated from the main chain to graft onto the main chain. As illustrated in Fig. 14.11, a new hyperbranched copolyether was thereafter synthesized via the CROP of HMO and MEMO (Ye et al., 2007c). In this copolyether, MEMO is responsible for producing a pendant triethylene glycol oligomer with a good motion ability, contributing to the ion conducting performance, while HMO is used to construct a hyperbranched structure as it contains two nucleophilic sites in one molecule. Compared with PHEMO, its ion conductivity was enhanced to 8.0 × 10−5 S/cm at room temperature and 7.4 × 10−4 S/cm at 80 °C, respectively. Similar to PMEMO in Section 14.5.2, the mechanical strength is not so robust for both hyperbranched polyethers mentioned above. They cannot be fabricated into a self-standing membrane. To this end, two crosslinked/ hyperbranched polyurethanes where PMEMO and PHEMO were incorporated as the soft segment were prepared. However, the mechanical property was enhanced at the expense of the loss of ion conductivity.

14.5.4 Polyelectrolytes As discussed above, the transference number of a polymer electrolyte is rarely bigger than 0.5, and it is obviously not satisfactory for a lithium ion battery. A so-called polyelectrolyte is developed as an alternative to make the transference number close to 1. When the electric field is applied to the polymer electrolyte, besides the cation moving along the direction of anode, the anion will also travel along the direction of cathode. As a result, a polarization voltage forms. If the anion is attached to the polymer backbone so that it cannot move in an electric field, the polarization voltage will be eliminated and as only the cation moves to make a contribution to the formed current the transference number will be close to 1. W. Xu and colleagues (2002) reported a polyelectrolyte formed from maleic acid, oxalic acid, lithium hydroxide, PEO and poly(propylene oxide) (PPO). They found that ion conductivity of this polyelectrolyte is highly related to the content of ethylene oxide/propylene oxide (EO/PO) repeat units, and the maximum ion conductivity is close to 10−3 S/cm. The ion conducting mechanism of polyelectrolyte was studied by Forsyth et al. (2003). The incomplete dissociation of polyelectrolyte is considered to be the main obstacle to further enhancing the ion conductivity. However,



+

O

OH

R=–(OCH2CH2)3OCH2

BF3Et2O

O

HO

O

R

O

R

R

L-MEMO

OH

O

14.11 Synthesis of hyperbranched copolyethers (from Ye et al., 2007c).

R

O

T-MEMO

HO

HO

O

O

B-HMO

O

HO

O

HO

L-HMO O

R

O

O

O

HO

OH

O

R

O

O

O

T-HMO

OH

R

OH

OH

OH

566


there is a giant intermolecular repulsion due to the fact that each molecule will carry negative charge after complete dissociation. To copolymerize with some neutral component (20%) is the choice of interest to depress the intermolecular repulsion.

14.5.5 Solid polymer electrolyte blended with ionic liquid (IL) Currently a promising approach to raise the ion conductivity of solid polymer electrolyte appears to be the incorporation of ionic liquids into PEO-based electrolytes. Although the mechanism is not yet fully understood, the lithium ion transport in the polymer electrolytes is strongly enhanced in the presence of certain ionic liquids. The increasing interest in these novel materials arises from their desirable properties such as nonvolatility, non-flammability, high thermal stability and high ion conductivity. Shin et al. (2006) investigated the cycle behavior and rate performance of solid-state Li/LiFePO4 polymer electrolyte batteries by incorporating a room temperature ionic liquid, N-methyl-N-propylpyrrolidinium bis(trifluoromethanesulfonyl)imide (PYR13TFSI), into the P(EO)20LiTFSI electrolyte at 40 °C. The corresponding ion conductivity reached about 6 × 10−4 S/cm at 40 °C at a PYR13+/Li+ mole ratio of 1.73. They still retained about 86% of their initial discharge capacity (127 mAh/g) after 240 continuous cycles and showed excellent reversible cyclability with a capacity fade lower than 0.06% per cycle over about 500 cycles at various current densities. In addition, the batteries exhibited discharge capability at high currents up to 1.52 mA/cm2 (2 C) at 40 °C which is excellent for a lithium polymer electrolyte (solvent-free) battery system. An addition of the ionic liquid to lithium metal–polymer electrolyte batteries has clearly resulted in a very promising improvement in performance at moderate temperatures.

14.6

Gel polymer electrolytes in lithium ion batteries

14.6.1 Preparation and ion conducting mechanism of gel polymer electrolytes The concept of the GPE was first proposed by Feuillard and Perche (1975). It contains a small amount of organic liquid known as a plasticizer, such as cyclic carbonic acid ester and chain-like ester (propylene carbonate, dimethyl carbonate, diethyl carbonate, etc.). Until now it has been the most common approach for the ion conductivity to reach the magnitude of 10−3 S/ cm at room temperature. Plasticizer-containing polyacrylonitrile (PAN) or poly(methyl methacrylate) (PMMA) polymer host was reported to provide



567

very high ion conductivity. The plasticizer changes the polymer matrix in two ways: (1) it makes segmental motion easier and (2) it presents liquidlike character similar to that of a liquid electrolyte. Together, these give rise to the high ionic conductivity in the GPE. The ion conductivity of a GPE is around 10−3 S/cm at room temperature which is far above that of an SPE. However, GPEs stand next only to SPEs in providing safety to the system. To date, several polymer hosts have been used in GPEs that include PEO, PPO, PAN, PMMA, poly(vinyl chloride) (PVC), PVdF, poly(vinylidene fluoride-hexafluoro propylene) (PVdF–HFP), etc. Furthermore, the Bellcore method and phase inversion method have been developed for the preparation of GPEs. The Bellcore process is critical and developed from the fabrication of PVdF–HFP copolymers. It involves the plasticization of PVdF–HFP copolymers, subsequent plasticizer removal, and the final reswelling in an electrolyte solution. Low-boiling solvents, such as diethyl ether or methanol, are successfully employed to remove dibutyl phthalate (DBP) from the polymer matrix, leaving a pore structure in the polymer layers which is then refilled with the liquid electrolyte during the cell activation process (Song et al., 1999). Phase inversion operation is easier with regard to the Bellcore process. The polymer is dissolved in a mixture of a volatile solvent and a non-solvent such that the amount of the non-solvent is low enough to allow solubilization and high enough to allow phase separation upon evaporation. The resulting solution is spread as a film on a glass substrate and the solvent is allowed to evaporate at ambient temperature to form a polymer electrolyte membrane. Finally, the membrane should be allowed to swell in an electrolyte solution to produce the GPE (Stephan, 2006).

14.6.2 Gel polymer electrolyte with added plasticizer Ito et al. (1987) investigated the ion conductivity of PEO–LiCF3SO3 complexes using PEG as plasticizer. The increase in the ion conductivity with the PEG content was mainly attributed to the reduction of crystallinity as well as the increase of free volume of the system. Among the polymer hosts used for GPEs so far, the PAN-based electrolyte offers a homogeneous, hybrid electrolyte film in which the salt and the plasticizer are molecularly dispersed. As reported by Abraham and Alamgir (1990, 1993), for instance, a typical electrolyte comprising 38 mol% EC, 33 mol% PC and 8 mol% LiClO4 immobilized in 21 mol% PAN showed an ion conductivity of 1.1 × 10−3 S/cm at −10 °C and 1.71 × 10−3 S/ cm at 20 °C. Kim et al. (1998) disclosed a copolymer of PAN and PMMA plasticized with a LiClO4 solution of EC/PC (1 : 1, vol) exhibiting an ion conductivity close to 10−3 S/cm at room temperature. This result also indicated that the


568


Tg deceases and the ion conductivity is enhanced with increasing PAN content. The polymer electrolytes formed from the higher acrylonitrile (AN) content copolymer are composed of two phases as the AN group is not completely compatible with plasticizer. One is the liquid electrolyte-rich phase, and another is the gel polymer phase plasticized by aprotic solvent. The latter is similar to the homogeneous state of polymer electrolytes formed from the lower AN content copolymer. Since the ion motion is faster in the liquid-rich phase, the ion conductivity increases with AN content. On the other hand, the saturated swelling capacity of GPE in liquid electrolyte increases as the content of PMMA increases, which certainly benefits the improvement of the ion conductivity. Hence 47 mol% was found to be an optimal content for PAN at which the maximum ion conductivity was obtained (Kim et al., 1998). PVdF has been chosen as a polymer host by virtue of its multifold appealing properties. PVdF-based polymer electrolytes are highly anodically stable owing to the strongly electron-withdrawing functional group (C—F). Furthermore, PVdF itself has a high dielectric constant (ε = 8.4). For a particular polymer, this can substantially assist ionization of lithium salts, and thus provide a high concentration of charge carriers. Wang et al. (1996a,b) studied the plasticized PVdF system in detail. They found that in systems with 30 mol% LiClO4, plasticizers increase the ion conductivity in the following order: DMF > γ-butyrolactone > EC > PC > PEG400 > PPG1000. The trend revealed that the viscosity rather than the dielectric constant of the plasticizers is the controlling parameter. Thus, the ion conductivity of a polymer electrolyte depends strongly upon the ionic mobility within the material. These low molecular weight polyethers with low dielectric constants and high viscosities are the least effective in elevating the ion conductivity. In China, there are a number of researchers who commit themselves to the preparation of polymer electrolyte with high ion conductivity. Wang et al. (2002) reported a PMMA-based electrolyte crosslinked by ethylene glycol dimethacrylate (EGD) for a lithium ion battery. As the network structure was introduced, the mechanical property of the GPE was improved. The maximum ion conductivity was 2 × 10−3 S/cm containing 25% MMA, 2% EGD and 73% plasticizer at room temperature. Additionally, the lithium cell fabricated with this gel polymer electrolyte exhibited excellent electrochemical properties. Zhu and Huang (1998) investigated the electrochemical properties of PVC and PMMA-based gel electrolytes. The results indicated that these electrolytes possess a high ionic conductivity and a wide electrochemical stability window. However, their application in rechargeable lithium ion batteries was hindered by the corrosion of the lithium electrode interface.



569

14.6.3 Gel polymer electrolyte blended with polymer PAN, PVC, and PVdF have been broadly used as matrix polymers in plasticized polymer electrolytes with a room temperature ionic conductivity in the order of 10−3 S/cm. However, they cannot completely satisfy the requirements for the high mechanical strength, long-term phase stability, and good adhesion to the electrode. As for this aspect, the copolymerization or blending with polymers has been proposed to improve the performance of polymer electrolytes. Blending is more useful because of the ease of preparation and control of polymer electrolytes by changing the composition of blended polymer matrices. Park and co-workers. reported a series of polymer electrolytes based on the blend of P(VdF–HFP) and poly(vinylacetate) (PVAc), and the maximum ion conductivity reached 2.3 × 10−3 S/cm at room temperature (Kim et al., 1998). The mobility and concentration of the free lithium ions in the polymer electrolyte were reduced with increasing PVAc content, leading to the decrease in the ionic conductivity. The bulk and interfacial resistance of the P(VdF–HFP)-based polymer electrolyte used for the lithium symmetric cell were gradually increased with storage, while those of the P(VdF–HFP):PVAc (7 : 3)-based ones were relatively stable during storage. Oh and Kim (1999) attempted to develop a new microporous GPE not only showing a high ionic conductivity and affinity for the electrolyte solution, but also having excellent film formation ability. To attain this goal they devised a blend polymer system consisting of a matrix polymer for maintaining the mechanical integrity and a filling polymer for absorbing electrolyte solution and gelation of the microporous GPE. PVdF–HFP (Kynar2801-grade) was selected as the matrix polymer and poly(methyl methacrylate-co-vinyl acetate) (PMMA–VAc) as the modifying polymer, which has excellent affinity for electrolyte solution, and was specially polymerized for this purpose. The ion conductivity of this blend GPE is higher than 10−3 S/cm at room temperature. It was suggested to be an effective protocol for improving the electrochemical properties of GPE for a lithium polymer secondary battery. Liu and colleagues (2003) also fabricated gel polymer films by blending P(VdF–HFP) with PMMA, the latter showing better compatibility with the liquid electrolyte. When the PMMA content is 45%, the liquid electrolyte uptake reaches 260% with the ion conductivity (20 °C) is 0.95 S/cm. The Li ion polymer battery consisting of LiCoO2 cathode, carbon fiber anode and the P(VdF–HFP)–PMMA polymer electrolyte and 1 mol/L LiPF6 in EC/ ethyl methyl caronate (EMC) mixture was found to have 95% of the initial discharge capacity at 35 cycle upon the repeated charge/discharge at 1/3 C rate, and 73% initial discharge capacity at 2 C rate. This blend polymer electrolyte is promising to be used in lithium ion cells.


570


14.6.4 Disadvantages of gel polymer electrolytes After adding plasticizers, ion conductivities are obviously elevated by several orders of magnitude. From the investigations carried out in recent years, it can be concluded that the ion mobility plays a pivotal role in the improvement of the ion conductivity of polymer electrolytes, and to achieve a high ionic conductivity is no longer a major problem. This viewpoint has been well accepted by the battery community. Nonetheless, the use of substantial amounts of plasticizers in GPEs might result in the following problems that existed in liquid electrolyte-based lithium ion batteries. Firstly, a GPE membrane is usually made from a polymer, a lithium salt and solvents which are binary or ternary. To facilitate the polymer dissolving in the preparation step and to obtain a homogeneous membrane, high compatibility between the polymer and the solvent should be guaranteed. Secondly, GPEs require organic solvents added with higher boiling points. Commonly employed organic solvents in liquid-type electrolytes are not appropriate for use in GPEs since those volatile solvents of low boiling point will hinder the stability of a gel. The gel system is essentially thermodynamically unstable and the rapid solvent evaporation will accelerate the degradation of a gel. Finally, it should be noted that the electrochemical stability of the electrolyte remains largely unresolved to date. To find an electrolyte system stable to metallic lithium (or lithiated carbon) and highly oxidizing cathode materials is not an easy task. The good stability at the electrode/electrolyte interface is the most critical criterion to be met before a reliable polymer electrolyte battery with long life cycle can be realized.

14.7

Composite polymer electrolytes in lithium ion batteries

Evidently, both SPEs and GPEs are not yet perfect, since, for example, the former suffers from low ion conductivity and the latter is impaired by its liquid components. However, recent studies have revealed that the addition of ceramic fillers into the SPE can improve the ion conductivity of polymer hosts and their interfacial properties in contact with the lithium electrode. The rise in the ion conductivity is explained by means of enhancing the amorphous phase of the polymer or hindering recrystallization. In all the cases the particle size and characteristics of the ceramic fillers are the key factors to improve the electrochemical properties of the electrolytes. In a pioneering work, Weston and Steele (1982) first demonstrated the effectiveness of incorporating inert filler (α-alumina) into the PEO system. The mechanical strength and ion conductivity are significantly enhanced upon



571

adding these inert particles into the polymer composite systems. In 1998, Croce et al. (1998) described their study in Nature where they added several kinds of ceramic fillers such as Al2O3, SiO2 and TiO2 into the PEO-based polymer electrolytes leading to an increase in the ion conductivity by 1–2 orders of magnitude. From then on, research on composite polymer electrolyte has boomed. Li et al. (2001, 2003) prepared a kind of hyperbranched solid polymer electrolyte with a low ion conductivity of about 10−6 S/cm at room temperature. Later, they added 10 wt% BaTiO3 as ceramic fillers into this SPE, making the ion conductivity rise to 1.4 × 10−4 S/cm. Based on the GPEs formed from a PVC:PMMA blend with micropore structure, some researchers added silica to increase the uptake of the liquid electrolyte. The ion conductivity increased with increasing PMMA content in the blend and a room temperature ion conductivity based on PVC:PMMA (5 : 5, w : w) blend of 1.1 × 10−3 S/cm was reached. Features of ceramic particles such as the type, morphology, particle size, particle content in the CPE and so on are important in improving its ion conductivity. In general, ceramic fillers used for the polymer electrolyte matrix are classified into active and passive species. The active one participates in the ion conduction process, e.g. Li2N and LiAl2O3, while the inactive ones, such as Al2O3, SiO2, MgO, are not involved in the lithium ion transport process. The selection of filler between active and passive components is quite arbitrary. It was reported that the ion conductivities measured at various temperatures for PVdF–HFP/TiO2 (rutile) show slightly lower values than those of using TiO2 (anatase), but the mechanism is not clear (Stephan and Nahm, 2006). A nano-sized TiO2 particle with pure anatase phase was synthesized by a process developed by Lin et al. (2005). It was found that the ion conductivity of the TiO2 modified CPE increases with decreasing grain size of TiO2. This implied that the ion conductivity of CPE is improved by increasing the interfacial interaction between TiO2 and PEO. The ion conductivity data of various PEO-based composite polymer electrolytes for different grain size of TiO2 at 30 °C are summarized in Table 14.1. From the table, it was noticed that the ion conductivity of CPE is improved by one to two orders of magnitude after adding TiO2 with particle size of 3.7 nm compared with that of the PEO–10% LiClO4 and the pure PEO electrolytes, respectively. Varying amounts of nanoscale rutile TiO2 with particle size ranging from 10 to 70 wt% were used by Kim and colleagues (2003) in the preparation of PVdF–HFP-based porous polymer electrolytes. A polymer electrolyte with 40 wt % rutile TiO2 showed the maximum ion conductivity of more than 10−3 S/cm at room temperature and potential to apply in the rechargeable lithium batteries.


572


Table 14.1 Ionic conductivity and transference number of PEO-based composite polymer electrolyte systems (Stephan and Nahm, 2006)

Electrolyte type

Conductivity (σ) (S/cm) at 30 °C

PEO (MW: PEO–10% PEO–10% PEO–10% PEO–10%

3.02 1.03 1.40 7.91 1.40

600 000) LiClO4 LiClO4–5% TiO2 (3.7 nm) LiClO4–5% TiO2 (6.2 nm) LiClO4–5% TiO2 (22.0 nm)

× × × × ×

10−6 10−5 10−4 10−5 10−5

Transference number (t+)

0.21 0.51 0.22

Adding SiO2 particles obviously promotes the ion conductivity of the polymer electrolyte (Ye et al., 2006b). For the comb-like polyether PMEMO as described above, an addition of 5 wt% nano-SiO2 increases the ion conductivity by 30–60%. However, the relative increase in the ion conductivity after adding nano SiO2 was not as high as expected. As those matrix polymers used in polymer electrolyte are almost crystalline or semicrystalline, adding nano-SiO2 can minimize the polymer crystallization tendency, leading to a substantial improvement in the ion conductivity. Because the PMEMO-based electrolyte herein is completely amorphous this may partially offset the effect of nanoparticles on the ionic conductivity. The positive influence of nanoparticles on amorphous polymer involves improvement in the segment motion performance because the addition depresses the interaction between Li+ ions and the ether oxygen atom decreasing the Tg. Zhao et al. (2002) prepared composite polymer electrolyte films of PEO– ZnO–LiClO4 and PEO–SnO2–LiClO4 using the film casting method. The interactions between PEO and ZnO nanoparticles and those between PEO and SnO2 ones diminish the crystallinity of the composite films to give more amorphous regions for charge carriers to transfer, and to release more Li+ or ClO4− from Li+ClO4− ion pairs as charge carriers. As a result the ion conductivity was significantly enhanced.

14.8

Polymer electrolytes in other battery types

14.8.1 Polymer electrolytes in fuel cells: solid solvent for protons On 9 January 2002, the US Secretary of Energy and executives of Ford Motor Company and General Motor Corporation announced a new cooperative automotive research partnership between the US Department of Energy and the US Council for Automotive Research (USCAR) called



573

FreedomCAR which focused on the development of fuel cell vehicle technologies. Fuel cell vehicle technologies are those that enable mass production of affordable hydrogen-powered fuel cell vehicles and hydrogen-supply infrastructure to support them. The fuel cell is a complete clean power source because it burns hydrogen and oxygen which can be produced through electrolyzing water and gives off steam to the atmosphere. Hence, it is anticipated to be used as one of the main power sources in the twentyfirst century in order to decrease the consumption of oil and coal. Among the vehicle technology options, proton exchange membrane (PEM) fuel cells, also referred to as SPE membrane fuel cells, are favored for use in automobiles. This preference is due to high power density, relatively quick start-up, rapid response to varying loads and low operating temperatures provided by PEM fuel cells (Mehta and Cooper, 2003). As shown in Fig 14.12, a single PEM cell comprises three types of components: a membrane electrode assembly (MEA), two bipolar plates and two seals. The MEA consists of a membrane, a dispersed catalyst layer, and a gas diffusion layer. The membrane separates the reduction and oxidation half reactions. It allows the protons to pass through to complete the overall reaction while forcing the electrons to pass through an external circuit to form current. In other words, this polymer electrolyte membrane is the solid solvent for protons. Perfluorosulfonic acid (PFSA) is one of the most commonly used membrane materials for PEM fuel cells. PFSA consists of three regions: (1) a

e– e– Anode

H2

O2

PEM

DL BPP

CL

Cathode

DL CL

BPP

H+

14.12 Illustration of single PEM cell (PEM is the proton exchange membrane, CL is the catalyst layer, DL represents the diffusion layer and BPP is a bipolar plate) (adapted from Mehta and Cooper, 2003).


574


polytetrafluoro-ethylene (PTFE)-like backbone, (2) side chains of –O–CF2– CF2– groups which connect the molecular backbone to the third region, and (3) ion clusters consisting of sulfonic acid ions. In the PFSA molecule, the hydrophobic group is responsible for avoiding over-swelling in order to maintain the morphology and dimension stability of the membranes and the hydrophilic group involves forming continuous ion transportation channels after the membrane becomes hydrated. It is concluded that the usage of PFSA brings two advantages. First, because the structure is based on a PTFE backbone, PFSA membranes are relative strong and stable in both oxidative and reductive environments. In fact, a durability of 60 000 h had been reported. Second, the protonic conductivities achieved in a wellhumidified PFSA membrane can be as high as 0.2 S/cm at PEM fuel cell operating temperature. However, there are some disadvantages with the PFSA membrane such as the degradable behavior at higher temperature, expensive cost for manufacture and membrane property decay at elevated temperature. For example, the ion conductivity at 80 °C is diminished by more than 10 times relative to that at 60 °C. Also, drawbacks such as membrane dehydration, reduction of ionic conductivity, decreased affinity for water, loss of mechanical strength via softening of the polymer backbone and increased parasitic losses through high fuel permeation were observed at temperatures above 80 °C. To overcome these drawbacks, sulfonated aromatic polymer membranes, sulfonated polyimide membranes, aliphatic hydrocarbon polymer membranes, organic/organic or organic/inorganic composite membranes and the like were developed. All these so-called acid anionic polymer electrolyte membranes (AAPEM) allow protons to pass through. Recently, a novel membrane allowing the pass through of hydroxyl ion called alkaline cationic polymer electrolyte membrane (ACPEM) was prepared. This is a new trend and a very attractive for use in polymer electrolyte membranes used in fuel cells (Zhang and Zhou, 2008).

14.8.2 Polymer electrolytes in alkaline battery/ supercapacitors: solid solvent for hydroxyl ion An alkaline SPE is also the solid solvent of hydroxyl ion which is usually composed of polymer matrix and basic substance (ion conducting substance). As the charge carrier in alkaline SPE is the hydroxyl ion, it allows hydroxyl anions to pass through. In comparison with lithium ion SPE, alkaline SPE has a number of distinct characteristics, such as easy preparation, low cost, abundance of basic components, and high ionic conductivity at room temperature (10−3 S/cm for PEO/KOH system). This alkaline PEObased SPE has found use in supercapacitor and alkaline rechargeable



575

batteries, especially Ni–Cd, Ni–Zn, Ni–H and other hybrid systems. Many polymers, such as PEO, polyvinylalcohol (PVA), polyacrylate acid (PAA), polyepichlorohydrin (PECH), polymethylacrylate (PMA) and poly(sodiumacrylate) (PSA), are suitable for alkaline SPE (Sun et al., 2003). Among these polymers, PEO, PVA and PAA are more popular. PEO is the first polymer matrix to be studied in alkaline SPE and possesses good mechanical strength to form a tough, self-standing membrane, but its high crystalline degree results in a low ion conductivity at room temperature (about 10−4–10−3 S/cm) and poor interfacial stability. The alkaline SPE based on PAA seems to be a GPE which has the similar electrochemical property to KOH solution and holds the highest ion conductivity among all reported alkaline SPEs. However, PAA often absorbs too much water and results in poorer mechanical strength. On the other hand, the PVA-based alkaline SPE has good electrochemical stability and mechanical strength. Its ion conductivity is located in the middle between the PAA and PEO-based alkaline polymer electrolytes. Both the component and content influence the ion conducting character of alkaline SPE. Generally, low crystallinity, flexible polymer backbone and high water content lead to high ion conductivity. The mechanism of ion transfer differs according to the content of water in the membrane. When SPE contains a low content of water, the matrix softness and the crystallinity degree influence its ion conductivity. For this aspect, the mechanism of ion transfer is similar to that of the SPE of a lithium ion battery where the ions move with the motion of the segment of polymer chain. The difference between two kinds of polymer electrolytes is that in the alkaline SPE the hydroxyl anion is used as charge carrier whereas in the SPE of a lithium ion battery the lithium cation is the charge carrier. When the water content is high in the alkaline SPE, the ion conduction predominately takes place in liquid phase so that it can be described by the Arrhenius equation. However, in the preparation of an alkaline polymer electrolyte the ion conductivity and the mechanical strength are usually in opposition. To increase the ion conductivity will decrease the mechanical strength and vice versa (Zhang et al., 2007). Recently, a novel alkaline SPE, prepared with tetramethyl ammonium hydroxide (Me4NOH·xH2O) without the addition of any volatile solvent, was introduced to remove the solvent component from the SPE so that the ion conductivity and the mechanical strength were simultaneously enhanced. PSA was reported to have good compatibility with Me4NOH·xH2O. It was shown that the clathrate structure of Me4NOH·xH2O plays an important role in providing pathways for hydroxyl anion in this alkaline SPE. Thus the polymer–Me4NOH·xH2O electrolyte appeared to have the improved mechanical properties as compared with the pure hydroxide and to remain highly conductive in the solid state (102 S/cm at about 40 °C).


576


1

4

2

3

1

5

Ion Coloration Dis-coloration

14.13 Illustration of the mechanism of an electrochromic device (1 transparent electron conducting layers; 2 ion storage layer; 3 electrochromic layer; 4 ion conducting layer; 5 transparent glass matrix) (adapted from Pu and Huang, 2005).

14.8.3 Polymer electrolytes in electrochromic devices and solar cells Nowadays, an electrochromic window or so-called smart window whose light transmittance ratio can be automatically moderated by the changing of incident light has aroused increasing attention for its significance in saving power source and protecting environment. For example, Lampert (1994) intended to apply these devices in the window of high buildings and automobiles to decrease the consumption of power source. As illustrated in Fig 14.13, a typical electrochromic device consists of five layers which are two transparent electron conducting layers, one ion storage layer, one ion conducting layer and one electrochromic layer to form a circuit. When the ion passes through the ion conducting layer from the ion storage layer to the electrochromic layer, the device is colored by the coloration reaction between cation and electrochromic materials, leading to a change in the transmittance ratio. In contrast, when the ion transfers from an electrochromic layer to an ion storage layer through an ion conducting layer, the device is de-colored (Pu and Huang, 2005). From the electrochromic mechanism described above, it can be concluded that the polymer electrolyte must be employed in the ion conducting layer. However, what type of polymer electrolyte is adopted depends on which ion is used in an electrochromic device. Theoretically, all cations can be used in electrochromic devices. Actually, only a few kinds of cations, such as H+, Li+, Na+, OH−, F− and the like, are appropriate. As a result, when H+ is used, the ion conducting layer is based on PFSA and so on. If Li+ is used, the ion conducting layer is similar to that used in a lithium ion battery. And for OH−, it is an alkaline polymer electrolyte.



577

The search for alternative sources of energy has gained increasing momentum in recent years in the light of our present energy needs. Sunlight is an abundant source one can tap into for electricity generation (photovoltaics) or to utilize as a driver for production of clean fuels (hydrogen generation). At present, if a person thinks of solar energy, dye-sensitized solar cells (DSSCs) immediately come to mind. Thus, one of the most promising DSSCs has been shown to operate with power conversion efficiencies of up to 10% when nanocrystalline TiO2 electrodes are used as a solid substrate and organometallic ruthenium complexes as sensitizers (Lampert, 1994). However, the polymer electrolyte used in DSSC is very different from all the polymer electrolytes mentioned above which should allow the passage of both cation and electron. Katsaros invented PEO–TiO2 as the SPE for DSSC with power conversion efficiency of 4.2% (the intensity is 65.6 mW/cm2). At the same time, other researchers made an attempt to impregnate an electron acceptor (heteropolytungstic acid) in PVdF–TiO2 as a solid polymer electrolyte for DSSC. Current research is also directed towards the assembly of solid state electrochromic devices by mixing conducting polymers with thermoplastics and elastomers such as PEO, PVdF and the like to produce polymeric blends which show the electrochromic properties of the conducting polymers (i.e. colored in the oxidized form and transparent in the neutral form) associated with the mechanical properties of common polymers.

14.9

Conclusions

The storage of electrical energy will be far more important in this century than it was in the last. Whether to power the myriad portable consumer electronic devices (cell (mobile) phones, personal digital assistants (PDAs), laptops, or for implantable medical applications, such as artificial hearts), or to address global warming (hybrid electric vehicles, storage of wind/solar power), the need for clean and efficient energy storage will be vast. Polymer electrolytes are indispensable for these processes. Since they possess many advantages over their liquid counterparts, they are expected to be widely applied and predominantly take the place of liquid electrolyte in the future. During the past 30 years, much revolutionary progress has been made in the preparation, characterization and electrochemical evaluation of various polymer electrolytes. However, for the state-of-art technology of polymer electrolyte, there are still a number of technical hindrances such as the low ion conductivity, poor cyclic performance and high manufacturing cost, etc. which need to be solved. In any case, the polymer electrolyte is a developing trend for both polymer science and electrochemistry science and a very attractive field to work in.


578


14.10 Acknowledgments This research work was funded by National Key Projects on Basic Research and Development (‘973’ Program, Grant Number 2002CB211800) and the National Natural Science Foundation of China (Grant Number 20703005). The authors thank all contributing publishers for their permission to use quotations under the agreement of the International Association of Scientific, Technical and Medical Publishers.

14.11 References abraham k m and alamgir m (1990), ‘Li+ conductive solid polymer electrolytes with liquid-like conductivity’, J Electrochem Soc., 137: 1657–1658 abraham k m and alamgir m (1993), ‘Solid electrolytes’, US Patent No.5219679 armand m b, duclot m (1978), Electrochemical current generators and materials for their construction, European Patent No.0013199 appetecchi gb, hassoun j, scrosati b, croce f, cassel f and salomon f (2003), ‘Hotpressed, solvent-free, nanocomposite, PEO-based electrolyte membranes: II. All solid-state Li/LiFePO4 polymer batteries’, J Power Sources, 124: 246–253 bednarek m (2003), ‘Multihydroxyl branched polyethers by cationic copolymerization of hydroxymethyloxetanes: copolymerization of 3-ethyl-3-(hydroxymethyl) oxetane with difunctional oxetane’, Polym Int, 52: 1595–1599 bednarek m, biedron t, helinski j, kaluzynski k, kubisa p, penczek s (1999), ‘Branched polyether with multiple primary hydroxyl groups: polymerization of 3-ethyl-3-hydroxymethyloxetane’, Macromol Rapid Commun, 20: 369–372 bednarek m, kubisa p and penczek s (2001), ‘Multihydroxyl branched polyethers. 2. Mechanistic aspects of cationic polymerization of 3-ethyl-3-(hydroxymethyl) oxetane’, Macromoleculer, 34: 5112–5119 bednarek m, penczek s and kubisa p (2002), ‘Polymerization of hydroxymethyloxetanes–synthesis of branched polyethers with primary hydroxyl groups’, Macromol Symp, 177: 155–162 bierwagen g p (1992), ‘Film coating technologies and adhesion’, Electrochim Acta, 37: 1471–1478 borghini mc, mastragostino m, passerini s and scrosati b (1995), ‘Electrochemical properties of polyethylene oxide-Li((Cf3So2)(2)N)-gamma-LiAlO2 composite polymer electrolytes’, J Electrochem Soc, 142: 2118–2121 capuono f, croce f and scrosati b (1991), ‘Composite polymer electrolytes’, J Electrochem Soc, 138: 1918–1922 chen y, bednarek m, kubisa, p and penczek s (2002), ‘Synthesis of multihydroxyl branched polyethers by cationic copolymerization of 3,3-bis(hydroxyl-methyl) oxetane and 3-ethyl-3-(hydroxymethyl)oxetane’, J Polym Sci: Part A, 40: 1991–2002 croce f, appetecchi gb, perci l and scrosati s (1998), ‘Nanocomposite polymer electrolytes for lithium batteries’, Nature, 394: 456–459 croce f, curini r, martinelli a, ronci f, scrosati b and caminiti r (1999), ‘Physical and chemical properties of nanocomposite polymer electrolytes’, J Phys Chem, 103: 10632–10638



579

dai y, wang y, greenbaum sg, bajue sa, golodnitsky d, ardel g, strauss e and peled e (1998), ‘Electrical, thermal and NMR investigation of composite solid electrolytes based on PEO, LiI and high surface area inorganic oxides’, Electrochim Acta, 43: 1557–1561 dillon r a and shriver d f (2001), ‘Ion transport and vibrational spectra of branched polymer and dendrimer electrolytes’, Chem Mater, 13: 1369–1373 fenton d e, park j m and wright p v (1973), ‘Complexes of alkali metal ions with poly (ethylene oxide)’, Polymer, 14: 589–589 feuillade g and perche p (1975), ‘Ion-conductive macromolecular gels and membrane for solid lithium cells’, J Appl Electrochem, 5: 63–69 fontenella j j, wintergill m c, calame j p and andeen c g (1983), ‘Electrical relaxation in pure and alkali metal thiocyanate complexed poly(ethylene oxide)’, Solid State Ionics, 8: 333–339 forsyth m, sun j, zhou f and mcfarlane d r (2003), ‘Enhancement of ion dissociation in polyelectrolyte gels’, Electrochim Acta, 48: 2129–2136 gavelin p, ljungback r, jannasch p and wesslen b (2002), ‘Amphiphilic solid polymer electrolytes’, Solid State Ionics, 147(3–4): 325–332 gray fm (1991), Solid polymer electrolytes-fundamentals and technological applications, New York: Wiley Publishers. golodnitsky d, ardel g, strauss e, peled e, lareah y and rosenberg y (1997), ‘Conduction mechanisms in concentrated LiI–P(EO)n–Al2O3-based electrolytes’, J Electrochem Soc, 144: 3484–3491 ito y, kanehori k, miyauchi k and kodu t (1987), Ionic conductivity of electrolytes formed from PEO-LiCF3SO3 complex with low molecular weight poly(ethylene glycol)’, J Mater Sci, 22: 1845–1849 jannasch p (2001), ‘Ion conducting electrolytes based on aggregating comblike poly(propylene oxide)’, Polymer, 42: 8629–8635 jiang g, maeda s, yang h, saito y, tanase s and sakai t (2005), ‘All solid-state lithium-polymer battery using poly(urethane acrylate)/nano-SiO2 composite electrolytes’, J Power Sources, 141: 143–148 kim d w, kim y r, park j k and moon s i (1998), ‘Electrical properties of the plasticized polymer electrolytes based on acrylonitrile–methyl methacrylate copolymers’, Solid State Ionics, 106: 329–337 kim k m, ko j m, park n g, ryu k s and chang s h (2003), ‘Characterization of poly(vinylidenefluoride-co-hexafluoropropylene)-based polymer electrolyte filled with rutile TiO2 nanoparticles’, Solid State Ionics, 161: 121–131 koksbang r, olsen i i and shackle d (1996), ‘Review of hybrid polymer electrolytes and rechargeable lithium batteries’, Solid State Ionics, 69: 320–335 krawiec w, scanlon j, feller j p, vaia r a, vasudevan s and giannelis e p (1995), Polymer nanocomposites: a new strategy for synthesizing solid electrolytes for rechargeable lithium batteries’, J Power Sources, 54: 310–315 kwang s j, moon s h, kim j w and park j w (2003), ‘Role of functional nano-sized inorganic fillers in poly(ethylene) oxide-based polymer electrolytes’, J Power Sources, 117: 124–130 lampert c m (1994), ‘Characteristics of laminated electrochromic devices using polyorganodisulfide electrodes’, Solar Energy Mater Solar Cells, 33: 91–105 li q, takeda y, imanish n, yang j, sun hy and yamamoto o (2001), ‘Cycling performances and interfacial properties of a Li/PEO-Li(CF3SO2)2N-ceramic filler/ LiNi0.8Co0.2O2 cell’, J Power Sources, 97–98: 795–797


580


li q, itoh t, imanishi n, hirano a, takeda y and yamamoto o (2003). ‘All solid lithium polymer batteries with a novel composite polymer electrolyte’, Solid State Ionics, 159: 97–109 lin c w, hung c l, venkateswarlu m and hwang b j (2005), ‘Influence of TiO2 nano-particles on the transport properties of composite polymer electrolyte for lithium-ion batteries’, J Power Sources, 146: 397–401 liu l, wang j l, xie j y and wu h k (2003), ‘Study on polymer blend electrolyte based on P(VdF-HFP)–PMMA’, J Functional Mater Devices, 9(1): 8–12 magnusson h, malmstrom e and hult a (1999), ‘Synthesis of hyperbranched aliphatic polyethers via cationic ring-opening polymerization of 3-ethyl-3(hydroxymethyl)oxetane’, Macromol Rapid Commun, 20: 453–457 magnusson h, malmtrom e and hult a (2000), ‘Influence of reaction conditions on degree of branching in hyperbranched aliphatic polyethers from 3-ethyl-3(hydroxymethyl)oxetane’, Macromolecules, 34: 5786–5791 mai, y y, zhou, y f, yan, d y and lu h w (2003), ‘Effect of reaction temperature on degree of branching in cationic polymerization of 3-ethyl-3-(hydroxymethyl) oxetane’, Macromolecules, 36(25): 9667–9669 mastragostino m, soavi f and zanelli a (1999), ‘Improved composite materials for rechargeable lithium metal polymer batteries’, J Power Sources, 81–82: 729–733 mehta v and cooper j s (2003), ‘Review and analysis of PEM fuel cell design and manufacturing’, J Power Sources, 114: 32–53 miwa y, tsutsumi h and oishi t (2001), ‘Polymerization of bis-oxetanes consisting of oligo-ethylene oxide chain with lithium salts as initiators’, Polymer J, 33(8): 568–574 miwa y, tsutsumi h and oishi t (2002), ‘Novel polymer electrolytes based on monoand bis-oxetane monomers with oligo (ethylene oxide) units’, Electrochemistry, 70(4): 264–269 murata k (1995), ‘An overview of the research and development of solid polymer electrolyte batteries’, Electrochim Acta, 40: 2177–2184 nishimoto a, watanabe m, ikeda y and kohjiya s (1998), ‘High ionic conductivity of new polymer electrolytes based on high molecular weight polyether comb polymers’, Electrochim Acta, 43(10–11): 1177–1184 oh b and kim y r (1999), ‘Evaluation and characteristics of a blend polymer for a solid polymer electrolyte’, Solid State Ionics, 124: 83–89 omata t and makoto k (1998), ‘Lithium ion-conductive polymer electrolyte and lithium ion battery’, European Patent No.0854527 owens b b and osaka t (1997), ‘Panel discussion future prospects of lithium batteries’, J Power Sources, 68: 173–186 pu h t and huang p (2005), ‘Research progress of polymer electrolyte used in electrochromic device’, J Functional Mater Devices, 11(2): 127–133 qian x, gu n, cheng z, yang x, wang e and dong s (2001), ‘Impedance study of (PEO)10LiClO4–Al2O3 composite polymer electrolyte with blocking electrodes’, Electrochim Acta, 46: 1829–1836 shin j h, henderson w a, scaccia s, prosini p p and passerini s (2006), ‘Solid-state Li/LiFePO4 polymer electrolyte batteries incorporating an ionic liquid cycled at 40°C’, J Power Sources, 156: 560–566 song j y, wang y y and wan c c (1999), ‘Review of gel-type polymer electrolytes for lithium-ion batteries’, J Power Sources, 77: 183–197



581

stephan a m (2006), ‘Review on gel polymer electrolytes for lithium batteries’, European Polymer, 42: 21–42. stephan a m and nahm k s (2006), ‘Review on composite polymer electrolytes for lithium batteries’, Polymer, 47: 5952–5964 sun j, macfarlane d r and forsyth m (2003), ‘Novel alkaline polymer electrolytes based on tetramethyl ammonium hydroxide’, Electrochimica Acta, 48: 1971–1976 takamura t (2006), ‘Trends in advanced batteries and key materials in the new century’, Solid State Ionics, 152–153: 19–34 vandenberg e j, mullis jc, juvet r s, miller t and nieman r a (1989), ‘ Poly(3hydroxyoxetane) – an analog of poly(vinyl alcohol): synthesis, characterization, and properties’, J Polym Sci: Part A, 27: 3113–3149 wakihara m (2001), ‘Recent developments in lithium ion batteries’, Mater Sci Eng, R33: 109–134 wang z, huang b, huang h, chen l, xue r and wang f (1996a), ‘Investigation of the position of Li+ ions in a polyacrylonitrile-based electrolyte by Raman and infrared spectroscopy’, Electrochim Acta, 41: 1443–1446 wang z, huang b, huang h, xue r, chen l and wang f (1996b), ‘A vibrational spectroscopic study on the interaction between lithium salt and ethylene carbonate plasticizer for PAN-based electrolytes’, J Electrochem Soc, 143: 1510–1514 wang z l, tang z y, geng x and xue j j (2002), ‘Research on novel PMMA-based gel electrolyte for lithium ion battery’, Acta Phys Chim Sin, 18(3): 272–275 wen t c, kuo h h and gopalan a (2002), ‘The influence of lithium ions on molecular interaction and conductivity of composite electrolyte consisting of TPU and PAN’, Solid State Ionics, 147: 171–180 weston j e and steele b h (1982), ‘Effects of inert fillers on the mechanical and electrochemical properties of lithium salt-poly(ethylene oxide) polymer electrolytes’, Solid State Ionics, 7: 75–79 wieczorek w, raducha d, zalewska a and stevens j r (1998), ‘Effect of salt concentration on the conductivity of PEO-based composite polymeric electrolytes’, J Phys Chem, 102: 8725–8731 xu j, zou y f and pan c y (2002), ‘Study on cationic ring-opening polymerization mechanism of 3-ethyl-3-hydroxymethyl oxetane’, J Macromol Sci–Pure Appl Chem, A39(5): 431–445 xu w, williams m d and angell c a (2002), ‘Novel polyanionic solid electrolytes with weak Coulomb traps and controllable caps and spacers’, Chem Mater, 14: 401–409 xu y y, gao c, kong h, deyue yan, luo p, li w w and mai y y (2004), ‘One-pot synthesis of amphiphilic core−shell suprabranched macromolecules’, Macromolecules, 37: 6264–6267 yan d y, hou j, zhu x, kosman j j and wu h s (2000), ‘A new approach to control crystallinity of resulting polymers: self-condensing ring opening polymerization’, Macromol Rapid Commun, 21: 557–561 ye l, feng z g, li s t, wu f, chen s and wang g q (2005), ‘The synthesis and copolymerization of 3-(2-cyano ethoxy)methyl- and 3′-(methoxy(triethylenoxy)) methyl-3′-methyloxetane’, Chem J Chin Uni, 26(10): 1946–1951 ye l, feng z g, zhao y m, wu f, chen s and wang g q (2006a), ‘Synthesis and application as polymer electrolyte of hyperbranched polyether made by cationic ringopening polymerization of 3-(2-(2-(2-hydroxyethoxy)ethoxy)ethoxy-methyl)-3′methyloxetane’, J Polym Sci Part A: Polym Chem, 44: 3650–3665


582


ye l, feng z g, zhang x w, qin q, bai y, wu f, chen s and wang g q (2006b), ‘Synthesis and application as polymer electrolyte of homo- and copolymers of 3-(2-cyano ethoxy)methyland 3-(Methoxy(triethylenoxy))methyl-3′-methyloxetane’ Chinese J Polym Sci, 24(5): 503–513 ye l, qin q, feng z g, zhang x w, bai y and wu f (2007a), ‘Preparation and characterization of novel hybrid thermoplastic poly(ether urethane)/poly(vinylidene fluoride) elastomers, and their application as solid polymer electrolytes’, Polym Int, 56: 660–665 ye l, zhao y m, feng z g, bai y and wu f (2007b). ‘Synthesis of copolymers of 3-acryloyloxymethyl-3p-methyloxetane and 3-(2-(2-(2-methoxyethylenoxy) ethylenoxy)ethylenoxy)-3p-methyloxetane and their ionic conductivity properties’, Front Chem Eng in China, 1(4):1–6 ye l, gao p, wu f, bai y and feng z g (2007c), ‘Synthesis and application as polymer electrolyte of hyperbranched copolyethers derived from cationic ring-opening polymerization of 3-(2-(2-(2-methoxyethoxy)ethoxy)ethoxy)methyl- and 3-hydroxymethyl-3′-methyloxetane’, Polymer, 48: 1550–1556 yu m x and zhou x (2002), ‘Recent development of polymer electrolytes for lithium ion batteries’, Chemistry, 4: 234–243 zhang h and zhou z t (2008), ‘Polymer electrolyte membranes for fuel cells’, Progress Chem, 20(4): 602–619 zhang j f, sang s b and wu q m (2007), ‘Progress in the research of alkaline polymer electrolyte’, Battery, 37(5): 394–397 zhao x, xiong h m and chen j s (2002), ‘Proton-conducting and ionic-conducting polymer electrolytes based on polyethylene oxide (PEO)’, Chin J Inorganic Chem, 18(1): 63–66 zhu w and huang z q (1998), ‘An investigation of electrochemical properties of gelled polymer electrolytes’, J Chongqing University, 21(1): 102–106


15 Hybrid polymer electrolytes for electrochemical devices F. L. D E S O U Z A, Federal University of ABC, Brazil, and E. R. L E I T E, Federal University of São Carlos, Brazil

Abstract: Inorganic and organic substances can be combined to form hybrid materials for application as ionic conductors. This association is attractive as it can favorably combine the properties of both materials. This field has attracted an ever-increasing interest due to the potentially promising applications of such electrolytes, not only in solid state rechargeable lithium or lithium ion batteries, but also in other electrochemical devices such as supercapacitors, electrochromic windows, solar cells and sensors. Here, we highlight the recent trends in the field of hybrid solid electrolytes, with emphasis on their physicochemical properties and promising applications in solar energy materials to maintain benign indoor environments. Key words: hybrid polymer, polyelectrolyte, electrochromic device, non-hydrolytic sol–gel.

15.1

Introduction

The science of polymer electrolytes is a highly specialized interdisciplinary field which encompasses the disciplines of electrochemistry, polymer science, organic chemistry, inorganic chemistry and materials science. The field has attracted ever-increasing interest, both in academia and industry, for the past two decades due to the potentially promising applications of such electrolytes, not only in solid state rechargeable lithium or lithium ion batteries, but also in other electrochemical devices such as supercapacitors, electrochromic windows and sensors.1–7 Although polymeric electrolytes have been applied in electrochemical devices for the past 30 years, fundamental problems dating from their inception still remain unsolved. The first study of solid polymer electrolytes was published by Fenton and coworkers8,9 in 1973, but their technological significance was not appreciated until the research undertaken by Armand et al.10,11 a few years later. These latter authors claimed that the crystalline complexes formed from alkali metal salts and poly(ethylene oxide) (PEO) were capable of demonstrating significant ionic conductivity, and highlighted their possible application 583 © Woodhead Publishing Limited, 2010

584


as battery electrolytes. This work inspired intense research and development of new polymer electrolytes, physical studies of their structure and charge transport, theoretical modeling of the charge-transport processes and the physicochemical properties at the electrolyte/electrode interface. The rapid progress in this field has been reported in numerous papers and reviews.1,3,4,12–14 The polymer electrolyte in an electrochemical device (such as a battery) plays a critical role as the electrode separator and, for cells employing composite cathode structures as a mechanical binder for the composite. The electrolyte must thus allow the passage of ions, while blocking electron conduction between the active components of the battery. In the March 2000 issue of MRS Bulletin, Scrosati and Vincent15 described the critical properties of polymer electrolytes for practical applications including the following: adequate ionic conductivity, together with high electronic resistivity; high cation mobility (if possible, with cation transference number t+ = 1); good mechanical properties (e.g. not brittle like certain ceramics, but able to relax elastically when under stress arising from, for example, volume changes in adjacent phases); the ability to form good interfacial contacts with electrodes; a large potential of window electrochemical stability; ease of processing, chemical and thermal stability and safety. Therefore, a number of new forms of polymer electrolyte have been developed. Although in practice a continuum of types exists, it is sometimes useful to group these polymer electrolytes according to electrolyte composition and morphology: • • • • •

Class 1: amorphous macromolecule salt complexes, typically based on polyether hosts. Class 2: plasticized systems, in which small amounts of low molar mass polar liquids are added to Class 1 polymer electrolytes. Class 3: gel electrolytes, formed by incorporating a nonaqueous electrolyte solution within an inactive structural polymer matrix. Class 4: ‘polymer-in-salt,’ or ‘rubbery’ electrolytes, in which high-molar mass polymers are dissolved in low temperature molten salt mixtures. Class 5: composites, based on the addition of either nanoparticulate ceramics or dual-phase block copolymers.15

In this chapter, we adopt a somewhat different classification scheme (by Wright16) based upon four distinct mechanisms for ion transport and provide a review of the most recent research developments in each case. In general, polymer electrolyte strategies currently being pursued exploit one of the following mechanisms for ion mobility: •

the translation of lithium salts through liquid solvents in gels or hybrid materials of various kinds; © Woodhead Publishing Limited, 2010

Hybrid polymer electrolytes for electrochemical devices •

•

•

585

solvent-free, salt–polymer complexed systems in which the ion motion is coupled to the micro-Brownian motion of segments of the polymer chains above the glass or melting transition temperature of the system; ‘single-ion’ systems, in which the lithium ion moves by a hopping process between anionic sites fixed to the polymer chain or systems with reduced mobility of anions (solvent-containing or solvent free); and solvent-free, salt–polymer complexed systems in which ion mobility is decoupled from the motions of polymer chain segments.

Based on this classification, Armand et al.11 investigated a wide range of alkali salt complexes and introduced the Vogel–Tamman–Fulcher (VTF) relation for amorphous LiSCN and CsSCN complexes. Armand’s proposal that they be used as ‘polymeric solid electrolyte’ in secondary lithium batteries apparently initiated a burst of activity among electrochemists worldwide. From that moment on, solid polymer electrolytes played an important role in the development of new energy sources. Thus, research into polymer electrolytes has focused largely on enhancing the flexibility of the polymer matrix in order to promote ionic mobility.17–19 This has met with success, as will be discussed below, but its disadvantage is that the conduction mechanism is dominated by ion–polymer interactions and the cations’ transference numbers (vital to battery applications) can be very low.20,21 For instance, Lee and Wright 22 and Payne and Wright23 investigated the morphology and its relation with conductivity showing that whereas partially crystalline complexes gave poor conductivities, the higher conductivity–lower activation energy regime could be extrapolated back to ambient temperature in fully amorphous PEO networks. Killis et al.,24 at the same time, had independently reached similar conclusions regarding the amorphous PEO by preparing urethane-based PEO networks. In 1980s Berthier and coworkers25 confirmed that the predominance of the amorphous phase mobility increased the development of a large number of strategies to suppress crystallinity in polyethoxy systems. The branched or ‘comb’ polyphosphazenes were studied by Shriver and co-workers (Northwestern University),26 and the methoxy copolymers (amorphous PEO) developed by Booth’s group at Manchester27 were among the more successful of these. Significant theoretical and mechanistic developments including the dynamic percolation model proposed by Ratner and coworkers occurred at the same time.28,29 Although some success has been reached, the disadvantages presented by the system with the conduction mechanism dominated by ion–polymer interactions such as low cation transference numbers (vital to battery applications) still remaining unsolved and new materials with unconventional conduction mechanisms are clearly needed. A different class of polymer electrolytes and concept has been discussed by Angell and co-workers,30–33 in which the ionic conductivity is not coupled © Woodhead Publishing Limited, 2010

586


to the segmental motion of the polymer chain (i.e. a material in which the ions move independently of the viscous flow).19,34–39 Angell and coworkers30–33 have distinguished between polymer electrolytes and glasses by identifying ionic motions in the former that are coupled to viscous flow and where electrochemical applications are possible only above the glass transition temperature Tg, and those in glasses that are decoupled from the viscous flow and where cation mobility is significant at and below Tg. Therefore, the development of a segmental motion-decoupled polymer system is an alternative way to increase ionic conductivity in polymer electrolytes.40 In this way, many studies in the literature are focused on obtaining materials with decoupled ionic mobility, where the ion transport mechanism is similar to that of a glass. For instance, many approaches were used to develop segmental motion-decoupled polymer to obtain hard polymers (e.g. liquid crystalline materials17 and polyester structures18,19). One other possibility that is being explored in various laboratories is the possibility of dissolving salt in certain polymers possessing special structural features38–41 which lead to decoupling of the ionic motion from the segmental motions. An example of this approach is provided by poly(vinyl alcohol) (PVA)–lithium salts complexes.39,35,42 Yamamoto et al.35 showed that such systems exhibit reasonable ambient conductivities (10−5−10−3.5 S cm−1) even though their Tg values lie well above room temperature. This suppressing behavior was also observed by Every et al.42 for complexes of PVA and lithium triflate; furthermore, 7Li nuclear magnetic resonance (NMR) measurements suggested that cation motion in these systems is decoupled from the primary relaxations of the polymer main chain. The authors suggest that cation transport is facilitated either by an ion hopping mechanism or by secondary polymer relaxations, presumably involving side groups. The possibility of a contribution to the conductivity arising from protonic conduction was not, however, ruled out. Indeed recent measurements of PVA–Li salt complexes using linewidth broadening 1H NMR spectroscopy strongly suggest the presence of mobile protons above and below Tg.42 It is far from clear, therefore, exactly how ion transport is decoupled from relaxations in the PVA complexes, although something very interesting is happening. Besides, McHattie and co-workers36 have described the properties of a new side group liquid crystal polymer (1) based on a predominantly PEC backbone where mesogenic groups are attached via flexible alkyl spacers. Recently, Souza and co-workers39,43–47 have described a special class of hybrid polyelectrolyte in which ion mobility presents an Arrhenius-type behavior above Tg, suggesting a segmental motion decoupled polymer system. Besides, the ion transport mechanism seems to be governed by thermally active ion hopping with the counter-ion fixed in the hybrid matrix. Based on this concept, many possibilities in solid state chemistry and physics arising within the several areas of electroactive and optically active


Hybrid polymer electrolytes for electrochemical devices

587

polymeric materials were recognized and embraced enthusiastically by researchers and viewed by technologists as being a good opportunity in the developments in electronics and communications technology. These materials have thus demanded a level of active collaboration between organic and physical chemists, physicists and device experts that was rare during the classical period and have led to the establishment of many interdisciplinary groups and laboratories. The chapter will begin with a brief description (Section 15.2) of the general aspect of physical chemistry properties of decoupled hybrid polyelectrolyte. In Section 15.4 an introduction of technological applications for decoupled hybrid polyelectrolytes, in particular for electrochromic or smart windows, is presented

15.2

Physicochemical properties of hybrid polyelectrolytes

Nowadays, some investigations have been conducted to develop a different synthetic route to produce hybrid materials. The combination of inorganic and organic substances to form hybrid materials for application as ionic conductors is attractive, because this association permits the formation of combined properties.48–57 The inorganic part is responsible for the mechanical properties, with the hybrid materials possessing both organic and inorganic combined properties.58–61 However, an inorganic–organic hybrid electrolyte in the solid state cannot be processed by conventional methods, such as the sintering or melting processes because of the decomposition of the organic components. So chemical synthesis is the most used and most suitable form of synthesis. Unfortunately, chemical syntheses of hybrid ionic conductors are very difficult and time consuming. The hydrolytic sol–gel process has been studied extensively over the past three decades as a facile route to organic–inorganic hybrids.62–69 It is only in the past 15 years that the corresponding non-hydrolytic sol–gel process has been recognized as a useful route to obtain hybrid materials.39,43,70,71 In this chapter in particular we will discuss a simple route to obtain a decoupled hybrid polyelectrolyte (DHP) by a derived in situ polymerizable method (a nonhydrolytic sol–gel process).39,43 The typical chemical synthesis consists of the formation of complex of Si (TEOS, tetraethylorthosilicate, was used as source of Si) and citric acid (CA) that is subsequently polymerized by means of a polyesterification reaction with ethylene glycol. The material became ionically conductive when Li2CO3 was dissolved in the structure during its synthesis. A transparent, amorphous solid hybrid was obtained after elimination of the solvent.43 The polymerization processes consists of a silicon (TEOS)–CA complex and polymer formation by the addition of ethylene glycol (EG) which were analyzed by Fourier transform (FT)-Raman and Fourier transform infrared


588


(d) (c) 2 1 (b) (a) 1200 1300 1400 1500 1600 1700 1800 1900 2000 Raman shift (cm–1)

15.1 FT-Raman spectra of the system investigated with (a) ethylene glycol, (b) citric acid, (c) TEOS/citric acid/ethylene glycol and (d) TEOS/ citric acid/lithium/ethylene glycol.

(FTIR) spectroscopy. Figure 15.1 depicts the FT–Raman spectra of EG (a), CA (b), TEOS + CA + EG (molar ratio 1 : 2 : 1) reacting at 60 °C for several hours (c), and TEOS + CA + EG + Li (molar ratio 1 : 2 : 1) also reacting at 60 °C for several hours (d). Complex Si–CA formation and polymerization should both involve the carboxylic groups of the citric acid. For this reason, all the experimental analyses of FT-Raman spectra were concentrated in the wavelength range of 1800–1600 cm−1 relative to the carboxylic groups. Figure 15.1(b) shows that the CA spectra presented two intense peaks at 1750 (peak 2) and 1680 cm−1, and a weak peak at 1725 cm−1 (peak 1). The peaks at 1750 and 1725 cm−1 were attributed to stretching of the central carboxyl ν(C=O) and to stretching of the terminal carboxyl ν(C=O), respectively. The intense peak at 1680 cm−1 was ascribed to the stretching vibration of the carboxyl group relating to strong intramolecular hydrogen bonds. This intense peak is therefore not suitable for monitoring the reaction between the Si and the carboxyl groups, but could be used to identify modifications in the intramolecular hydrogen bond after formation of the complex. Leite and co-workers71 observed similar results in a study of the formation of Si–CA complex using FT-Raman and 13C NMR techniques. The aforementioned samples, which were analyzed by the FT-Raman technique, were also subjected to a complementary FTIR analysis (see Fig. 15.2). The spectra in Fig. 15.2 were identified as EG (A), CA (B), TEOS + CA + EG (molar ratio 1 : 2 : 1) reacting at 60 °C for several hours (C), TEOS + CA + EG + Li (molar ratio 1 : 2 : 1) (D), and TEOS + CA + EG + Na also reacting at 60 °C for several hours (E). For this reason, the FTIR study was limited to the short range (1800–1600 cm−1 wavelength, highlighted area in



589

(a) (A) (B)

(C)

2 1

(D) (E)

2000

1900

1800

1700

1600

1500

1400

Wavenumber (cm–1) (b) (A)

(B) (C)

1400

1300

1200

1100

1000

900

Wavenumber (cm–1)

15.2 (a) FTIR spectra of the system investigated with (A) ethylene glycol, (B) citric acid, (C) TEOS/citric acid/ethylene glycol (D) TEOS/ citric acid/sodium/ethylene glycol and (E) TEOS/citric acid/lithium/ ethylene glycol; (b) FTIR spectra of the DHP: (A) without ionic salts; (B) DHP with [CCO−]/[Na+] = 103.5 and (C) DHP with [CCO−]/[Li+] = 31.3.

Fig. 15.2) relative to the carboxylic groups, which we believe are responsible for the formation of polymer after the Si–CA reaction. The CA spectra showed three distinct peaks at 1755 (peak 2), 1725 (peak 1) and 1692 (arrow) cm−1. The peaks at 1755 and 1725 cm−1 were attributed to the central carboxyl ν(C=O) and to stretching of the terminal carboxyl ν(C=O),


590


respectively. The peak at 1692 cm−1 was attributed to the stretching vibration of the carboxyl groups relating to strong intramolecular hydrogen bonds. It was important to note that the EG and TEOS did not exhibit characteristic bands in this spectral range. Figures 15.1 and 15.2 (C–E) summarize the results of the reaction between CA and TEOS and the formation of Si–CA complex, as discussed in detail by Leite et al.,71 and determined by FT-Raman and by FTIR analysis of the polymerization process of the TEOS + CA + EG, TEOS + CA + EG + Li+ and TEOS + CA + EG + Na+ samples. Note that the three peaks associated with the carboxyl ν(C=O) stretching peak in the CA and Si–CA complex were transformed into a single broad peak centered at 1736 cm−1 (in all the samples, Fig. 15.1c–d). As expected, the FTIR analysis yielded similar results, with a single broad peak centered at around 1720 cm−1 (Fig. 15.2C–E). These results confirm the reaction between the Si–CA complex and the EG. The single broad shifted peak visible after the polymerization reaction was assumed to indicate the formation of a polymer. Lastly, the FTIR and FT-Raman analyses suggested the polymerization reaction of the Si–CA complex with EG and also indicated the formation of an organic–inorganic (hybrid) polymer. Moreover, no significant Raman shift was observed in the central and terminal –COOH groups in relation to CA in the presence of Li+ and Na+ ions. On the other hand, the addition of different ions enabled us to confirm the presence of the counter-ion in the polymer chain, demonstrating that both species interacted with the same organic group (highlighted area in Fig. 15.2b). Fig. 15.2(b) shows the FTIR spectra for the DHP without ions, and with 10% wt lithium and sodium ions, represented in (A), (B) and (C). The highlighted region at around 1080–1020 cm−1 was found to be the O–C–C band of the molecule (‘alcohol’ carbon–oxygen stretching), which could be assigned to the primary, secondary or tertiary structure of an alcohol.72 We believe that this region is the active site of the hybrid polyelectrolyte, i.e. the counter-ion of the hybrid polyelectrolyte. The O–C–C stretching vibration has been attributed (at around 1025 cm−1) to tertiary binding of an alcohol, as we reported in a previous paper.44

15.2.1 The ion mobility in decoupled hybrid polyelectrolytes: electrical properties In general, polymer electrolytes are mixture of salts with soft polar polymers such as PEO and poly(methoxyethoxyethoxy)phosphazene (MEEP) and have been intensively studied since their discovery.8,9 The mobility of ions in typical polymer electrolytes is well known to be intimately coupled to the segmental motion of the host polymer matrix. To increase the segmental motion and thereby conductivity of polymer electrolytes, a variety



591

of polymers with low Tg values have been developed.13,24,73,74 However, many avenues to low materials have been explored, so the prospects are not encouraging for the development of polymers with a Tg low enough to obtain a room temperature conductivity of 10−3 S/cm, often a goal in practical applications. The electrical response observed in conventional polymer is usually interpreted by non-Arrhenius behavior. The temperature dependence of DC conductivity measured from the polymer electrolytes is the hallmark of ionic motion being coupled with the host matrix. The temperature dependence of the conductivity exhibits an apparent activation energy that increases as temperature decreases. This behavior is most commonly described by the empirical VTF equation, which was first developed to describe the viscosity of supercooled liquids.75–77 However, there is a different class of polymer electrolyte, discussed and first reported by Angell, suggesting that the ionic conductivity is not coupled to the segmental motion of the polymer chain, that is, in which the ions move independently of the viscous flow.30 Based on this approach, Souza recently reported a new class of DHP (synthesis route discussed above), in which the ion mobility presented an Arrhenius behavior of the conductivity as a function of temperature, suggesting that the ion motion is decoupled from the polymer segmental motion for temperatures above Tg (about 79 °C).39,43 In order to understand the effect of ions on the electrical properties of DHP, the strategy proposed is to characterize the hybrid polymer with and without Li and Na+ ions by impedance spectroscopy (IS). Low [CCO−]/[Li+ or Na+] ratios were used here (values summarized in Tables 15.1 and 15.2) to avoid possible ion cluster formation or phase segregation. This ratio is important because the [CCO−] was identified as the counter-ion and confirmed in the present study, as clearly revealed by FTIR (Fig. 15.2b). The DHP without Li+ presented high resistivity, suggesting a very low ionic conductivity; hence, without ions (lithium or sodium) or in the presence of low water concentrations (H+), the material was found to be completely insulating (see inset in Fig. 15.3a). Figures 15.3(a) and (b) depict the behavior of the DHP found by complex-plane impedance in the Cole–Cole plot (Z* = Z′ + jZ″) as a function of different lithium and sodium concentrations at room temperature (T > Tg). The semicircle (Figs 15.3a and b) observed by complex-plane impedance in the Cole–Cole plot was attributed to the intrinsic properties of the DHP with Li+ and Na+. The semicircle (related to an equivalent RC parallel circuit) indicates the expected behavior of the polymer with lithium and sodium due to the experimental condition of our study, i.e. blocking electrodes and the chosen frequency range. The ionic conductivity (σdc) was calculated from the bulk resistance, Rb, which was determined from the interception of the spur at the Z′-axis with the specimen’s thickness and surface area. As expected, the resistivity of DHP


Polymer electrolytes 250 × 10 3

SPHP without Li+

200 × 10 3 Z (Ω cm)

1 × 106

800 × 103

150 × 10 3

1% Li 5% Li 10% Li

100 × 10 3 50 × 10 3

3

×1 0

3

3

0

×1 0

20

25

0

×1 0

3

10 0

15 0

3

×1 0

600 × 103

50

–Z˝ (Ω cm)

0

×1 0

0

–Z´ (Ω cm)

400 × 103

f0

0.266 MHz

0.103 MHz

8.5 kHz

200 × 103

0.64 MHz 0.17 MHz

0

200 × 10 3 400 × 10 3 600 × 10 3 800 × 10 3

0 (a)

1 × 10 6

Z´ (Ω cm) 1.4 × 10 6 T = 298 K

1.2 × 10 6

5% Na 10% Na

1.0 × 10 6 800 × 10 3 81 kHz f0 26 kHz

299 kHz f0 91 kHz

(b)

6

×

10

6

1.4

×

10

10

6

3

×

10

3

80 0

×

10

3

10 ×

40 0

20 0

×

10

0

3

0

7 kHz 9 kHz

317 kHz

×

200 × 10 3

27 kHz

1.0

400 × 10 3

1.2

600 × 10 3

60 0

–Z˝ (Ω cm)

592

Z´ (Ω cm)

15.3 (a) Z ′ vs. Z ″ impedance complex plane plots, for DHP with 1, 5 and 10 wt% Li at room temperature. Inset: Z ′ vs. Z ″ impedance complex plane plots, for DHP without lithium, (b) Z ′ vs. Z ″ impedance complex plane plots, for DHP with 5 and 10 wt% Na at room temperature. The imaginary part of the impedance complex (Z ″) as a function of frequency (log freq.) is associated with relaxation process for the DHP with 1, 5 and 10 wt% Li (c) and (d) 5 and 10 wt% Na.


Hybrid polymer electrolytes for electrochemical devices 400 × 103

f0

T = 298 K 1%

350 × 103 300 × 103

10%

250 × 103 200 × 103 150 × 103

0.2 MHz f0

100 × 10

3

50 × 103

0 00

10 0

1 log f (Hz)

(c) 90 × 103

T = 298 K

f0

81 kHz

5% Na 10% Na

75 × 10

3

91 kHz

60 × 103 –Z˝ (Ω cm)

00 0

00 0

00 0

10 0

00

00

0

0

0 00 10

00 10

10 0

10

00 0

f0 1 MHz

0

10

–Z˝ (Ω cm)

0.103 MHz

f0

45 × 103 30 × 103

log f (Hz)

15.3 (cont.)


00 0

00 0

00 0 10 0

10

00 0 1

(d)

00 0

00 0

0 00 10 0

00 0 10

10 00

10

0

10 0

15 × 103

593

594


Table 15.1 Values of the ionic conductivity, characteristic relaxation frequency and glass transition temperature for DHP with different Li+ content at room temperature % Li+

[CCO−]/[Li+]

σ (Ω/cm)

f0 (MHz)

Tg (°C)

1.0 5.0 10

330 66 31.3

1.1 × 10−6 6.6 × 10−6 1.4 × 10−5

0.1 0.2 1.0

– −22 −16

Table 15.2 Values of the ionic conductivity, characteristic relaxation frequency and glass transition temperature for DHP with different Na+ content at room temperature % Na+

[CCO−]/[Na+]

σ (Ω/cm)

f0 (kHz)

Tg (°C)

5.0 10

216.2 103.5

8.0 × 10−7 1.1 × 10−6

91 81

−12 −4

decreased when more lithium or sodium content was added, i.e. the ionic conductivity increased with ion concentration (as summarized in Tables 15.1 and 15.2). Therefore, our electrical measurements (Figs 15.3a and b) clearly confirm that only ions (Li+ or Na+) appear to be mobile in the structure of the DHP. In addition, the characteristic relaxation frequency, f0, was extracted from these data, as indicated in Figs 15.3(a) and (b). The characteristic relaxation frequency (f0) is also shown in the log f (Hz) versus −Z″ curve, where the part of the curve that indicates high frequency corresponds to a bulk relaxation phenomenon and the plateaued regions are related to σdc. Note the displacements of f0 and σdc with increasing lithium and sodium content, suggesting that there is some kind of interaction between lithium/ sodium ions and the DPH matrix. In addition, a comparison of the characteristic relaxation frequencies of the two ions (Figs 15.3c and d and Tables 15.1 and 15.2) clearly shows the shift to a lower relaxation frequency when sodium is added, confirming the expected mass effect (due to higher atomic mass). This result, allied with the enhanced ionic conductivity caused by augmenting the ion content (in both cases), suggests that the interaction between ions and hybrid polymer is weak. In other words, ionic conductivity is enhanced with higher lithium and sodium concentrations, independently of the effect on the mobility of the hybrid polymer chain. This effect on the DHP is widely reported in the literature as being a decoupled system, where



595

the ionic transport mechanism is based on its mobility in the polymer chain occurring independently of the segmental motion.

15.3

General discussion

The results presented and discussed herein indicate that the DHP synthesized by a non-hydrolytic sol–gel process is transformed into an ionic conductor when ions are added during its preparation, as demonstrated in Figs 15.3(a) and (b). In addition, based on our spectroscopic study, we were able to identify the organic group whose ions interact with the matrix (i.e. identify the counter-ion) and to evaluate the polymerization process. Firstly, the FT-Raman and FTIR analysis confirmed the polymerization reaction of the Si–CA complex with EG and also indicated the formation of an organic– inorganic (hybrid) polymer. The single broad shifted peak centered at around 1720 cm−1 (Figs 15.2C–E) which was visible after the polymerization reaction was likely to be an indicator of polymer formation. To support these findings, the average molecular weight of the polymer (Mw = 2.6 × 104 g/mol) was estimated by gel permeation chromatography (GPC). In addition, as mentioned earlier in our analysis of the FTIR spectra (see the highlighted area in Fig. 15.2b), the ions detected at around 1025 cm−1 were identified as the C–C–O stretching vibration of an alcohol assigned to tertiary binding. Therefore, based on the Mw value, ion content and monomer molecular weight (434.4 g/mol), the value of the [CCO−]/[Li+ or Na+] ratio was calculated, as indicated in Tables 15.1 and 15.2. A low [CCO−]/[Li+ or Na+] ratio was used here to avoid the possible formation of ion clusters or phase segregation. These results, together with the electrical measurements (Figs 15.3a and b), confirm unequivocally that the introduction of ions transformed the polymer into a conductive material, and also that the ionic conductivity was closely dependent on the added ion content. Moreover, we showed that the ions moved independently of the polymer’s segmental motion; thus, the system presented a decoupled ion transport mechanism. This behavior was revealed by the differential scanning calorimetry (DSC) analysis, which showed a considerable displacement of the Tg to higher temperatures (Tables 15.1 and 15.2). These results indicate that increasing the lithium and sodium concentrations affected the chain’s mobility, probably leading to a reduction of the polymer chain mobility. In principle, a concept that is very important to understand the mechanisms of conventional ionic polymer is the coupling between transport and relaxation of the polymer segmental chain, i.e. if the Tg shifts to higher temperatures, there is a reduction of polymer segmental motion, causing a decline in the ionic conductivity to lower values. The results of this study indicate that a weak interaction between ions and hybrid polymer chain enhanced the ionic conductivity as


596


the lithium and sodium concentrations were increased, independently of the effect on the mobility of the hybrid polymer chain. Thus, we confirmed once again that the transport mechanism of our polyelectrolyte, called ‘DHP’, is completely decoupled, i.e. the ions move independently of the viscous flow. Another important aspect is the fact that the ionic conductivity values for DHP with sodium were lower than the DHP with lithium ions, as illustrated in Figs 15.3(a) and (b). In addition, a comparison of the characteristic relaxation frequency values (Figs 15.3c and d and Tables 15.1 and 15.2) clearly showed the shift to a lower relaxation frequency when sodium was added. In fact, this result was expected because of the close relation between mass and frequency, with elements possessing a higher atomic mass causing a displacement in the frequency to lower values. In addition, the DSC analysis indicated that the higher Tg value of DPH with sodium, reinforcing the suggestion that the lower ionic conductivity (compared with DHP + lithium, Tables 15.1 and 15.2) was related with the greater interaction between the sodium ions and the polymer chain. Finally, these findings lead us to believe that the choice of ions is also a very important factor in enhancing the ionic conductivity in the DHP developed here, reducing the interaction between ions and matrix and increasing its mobility. Besides, the good physical-chemical properties presented by this kind of polyelectrolyte make it an interesting material to be applied in electrochemical devices, such as an electrochromic device (this will be discussed in the next section).

15.4

Applications

The polymer electrolyte in 1980 was introduced as important material for a new application, particularly in electrochromic devices or smart windows, becoming the most popular technology for large-area switching devices. This technology was developed for building and automotive windows, as well as mirrors. Since then, the field of solid state in large-area electrochromic devices is ever-expanding, with many new markets and technologies.78–81 Lampert81 in 2003 discussed the increase of the interest by several companies throughout the world in the developing dynamic glazing and largearea flat panel displays. University and national laboratory groups are researching new materials and processes to improve these products. For instance, Flageb Company (Germany) made the largest EC architectural window installed in a building.81 A group of these windows covering 8 × 17 m2 has been installed in the Stadtsparkasse Bank in Dresden. The functional layers of the electrochromic window were fabricated by a vacuum deposition technique and the devices had a market price of about 500 euros/m2; too high for a broad application of the electrochromic windows. In order to find a wider clientele, it is desirable to lower the costs down to 150 dollars/m2 82,83 or 100–250 dollars/m2.81



597

The attractive applications of the large-electrochromic glazing are in the architectural and automotive fields in order to control the solar radiation entrance into buildings and automotives to save air-conditioning costs in summer and heating costs in winter.78–81 Studies towards providing a simple quantitative estimation of energy savings performed by Granqvist and Azens and co-workers showed that with electrochromic windows varying the transmittance between 7 and 75% the energy savings can be as high as 340 kW h/m2 yr, providing that the solar energy irradiation onto a window is 1000 kW h/m2 yr.84–89 Besides, they proposed if the room was used only 50% of the time, a minimum value for energy savings is 170 kW h/m2 yr. This is equal to the energy which can be generated by today’s best solar cell modules having about 17% efficiency, of the same size and positioned at the same place.84–89 In addition, comfort factors such as privacy, glare and fading have driven interest in electrochromic window development for office buildings, homes and automobiles.89–91 From this standpoint much research and many strategies has been focused on fabrication processes for suitable electrochromic devices for glass substrates to reach better efficiency and performance. Although great success has been reached in all electrochemical devices, including large electrochromic devices, the ionic conducting electrolyte still keeps being the critical component. The electrolyte may be a liquid, a polymer, a gel or a thin film electrolyte. Liquid electrolytes are not useful for large-area applications, because of the buckling of the glass and the risk of leakage. Therefore, polymer, gel or solid electrolytes are preferred for large area electrochromic devices. The literature on ionic conductors is vast and several reviews on inorganic ion conductors suitable for electrochromic devices and other applications are available.83,92–100 Most of the research in this field was done on H+ and Li+ conductive coatings, the last one being discussed in this chapter. Granqvist101 and Vaivars et al.102 pointed out the important characteristics that an electrolyte should have to be applied in large electrochromic devices, such as, high ionic conductivity between 10−3 and 10−7 S/cm (depending on the application), low electronic conductivity (less than 10−12 S/cm), long cycling durability at operation temperature, good adhesion with the adjacent layers (no delamination for several years, even after temperature switching tests and more than 104 switching cycles), optical transparency for most electrochromic applications (display may be an exception), chemical compatibility with the functional layers, electrochemical stability in the voltage range used for switching the electrochromic device and long-term stability against UV light if the UV light is not filtered by the functional layers for certain applications (e.g. electrochromic devices for architectural or automotive glazing). There are many classes of materials which can be used as an electrolyte. Particular attention has been given to organic–inorganic hybrids which combine the better conductive properties


598


of polymer type material with the better mechanical strength of inorganic material. Such a system has been discussed in more detail in this section. In general, the solid (gel) electrolytes described in the literature contain, as a basic compound, an ionic conductor (most frequently a lithium salt), an organic matrix and organic solvents (anhydrous ones). These solvents must cause ionic dissociation and provide the required ionic mobility, in other words they should provide a suitable conductivity. As discussed previously the decoupled hybrid polyelectrolyte synthesized by unpolymerizable route is a great material to be applied in the electrochromic device.39,43 Souza and co-workers46,47 showed the remarkable performance of the decoupled hybrid polyelectrolyte in large electrochromic devices, using different configurations. The electrochromic device built with WO3 or Nb2O5:Mo (electrochromic electrode), CeO2–TiO2 (counter-electrode) and decoupled hybrid polyelectrolyte were tested (5 × 10 cm2 size). The lifetime of electrochromic device I (WO3/CeO2–TiO2/DHP) and electrochromic device II (Nb2O5:Mo/CeO2–TiO2/DHP) were prolonged to over 60 000 cycles with the DHP. Long-term stability and high coloration efficiency were achieved. The DHP showed excellent charge transport performance throughout all the cycles tested (over 60 000 cycles for devices I and II), with a maximum constant rate of deintercalation/intercalation (Qout/Qin = 1). The development and improvement of hybrid polyelectrolytes with decoupled behavior can lead to the development of large solid electrochromic devices with high stability performance. Recently, Zhang and coworkers103 investigated an electrochromic device with an effective area of 5 × 5 cm2 and WO3/solid hybrid polyelectrolyte/NiO configuration which showed great performance of the device and good reproducibility of the hybrid polyelectrolyte. The device showed an optical modulation of 55% at 550 nm and achieved a coloration efficiency of 87 cm2/C. The response time of the device is found to be about 10 s for the coloring step and 20 s for the bleaching step. The views through the device applied at different voltages are quite gentle. It was concluded that the electrochromic device based on the NiO/WO3 complementary structure and decoupled hybrid polyelectrolyte has potential applications in smart windows.

15.5

Conclusions

Hybrid polymer electrolytes represent one of the most versatile classes of solid polymer electrolyte. To a certain extent, these hybrid materials combine the most important properties of their constituents, such as high transparency (glass-like), low processing temperatures (polymer-like), sufficient thermal stability (silicone-like), with high performance yield and properties not found in either material individually. These materials display a number of advantages over simple salt-in-polymer electrolytes and are believed to



599

be promising materials for application in secondary lithium batteries and electrochromic windows. Advantages include the suppression of PEO crystallization and enhancement of the mechanical properties. In many cases, hybrid electrolytes also show higher conductivity and Li transference numbers. Besides, the hybrid polymer electrolyte, given a number of designs is capable of large optical modulation, as well as dynamics and durability sufficient for applications. Finally, combining appropriate materials it is possible to assemble electrochromic devices constituted exclusively by polymeric materials. The advantage of these devices is the possibility of large-scale production, combining the properties of polymers (low density, relative low cost, etc.) with those of electrochromic materials.

15.6

Acknowledgments

Based on research carried out over several years, this work has been supported by several sources: PRODOC/Capes, CNPq project number 575119/2008–0 and FAPESP (Brazil).

15.7

References

1. f. m. gray, Solid Polymer Electrolytes: Fundamentals and Technological Applications, VCH Publishers, Inc., New York, (1991). 2. f. m. gray, Polymer Electrolytes, RSC Materials Monographs, Royal Society of Chemistry, London, (1997). 3. m. b. armand, in: Polymer Electrolyte Reviews, vol. 1, J. R. MacCallum, C. A. Vincent (eds), Elsevier, London, (1987). 4. c. a. vincent, Prog. Solid State Chem. 17 (1987) 145. 5. m. armand, Faraday Discuss. Chem. Soc. 88 (1989) 65. 6. b. scrosati, Proceedings of the 2nd International Symposium on Polymer Electrolytes, Elsevier Applied Science, Amsterdam, 1990. 7. b. scrosati, Applications of Electroactive Polymers, Chapman and Hall, London, 1994. 8. b. e. fenton, j. m. parker, p. v. wright, Polymer 14 (1973) 589. 9. p. v. wright, Br. Polym. J. 7 (1975) 319. 10. m. b. armand, j. m. chabagno, m. j. duclot, Proceedings of the 2nd International Meeting on Solid Electrolytes, St. Andrews, (1978) (Extended abstract 6.5.1). 11. m. b. armand, j. m. chabagno, m. j. duclot, in: Fast Ion Transport in Solids, P. Vashishta (ed.), North-Holland, New York (1979) 131. 12. m. armand, Solid State Ionics, 9&10 (1983) 745. 13. m. a. ratner, d. f. shriver, Chem. Rev., 88 (1988) 109. 14. m. b. armand, Annu. Rev. Mater. Sci. 16 (1986) 245. 15. b. scrosati, c. a. vincent, MRS Bull. 25 (3) (2000) 28. 16. p. v. wright, Electrochim. Acta, 43 (1998) 1137. 17. m. b. armand, Solid State Ionics 69 (1994) 309. 18. w. h. meyer, Advanced Materials 10 (1998) 439. 19. c. a. angell, t. imrie, m. t. ingram, Polymer Int 47 (1998) 9.


600 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57.

Polymer electrolytes g. g. cameron, j. l. harvie, m. d. ingram, Faraday Discussions 88 (1989) 55. p. g. bruce, c. a. vincent, Faraday Discussions 88 (1989) 43. c. c. lee, p. v. wright, Polymer 23 (1982) 681. d. r. payne, p. v. wright, Polymer 23 (1982) 690. a. killis, j. f. le nest, h. cheredame, Makromol. Chem. Rapid Commun. 1 (1980) 595. c. berthier, w. gorecki, m. minier, m. b. armand, j. m. chabagno, p. rigaud, Solid State Ionics, 11 (1983) 91. p. m. blonsky, d. f. shriver, p. austin, h. r. allcock, J. Am. Chem. Soc. 106 (1980) 595. c. v. nicholas, d. j. wilson, c. booth, j. r. m. giles, Br. Polym. J. 20 (1988) 289. m. a. ratner, in: Polymer Electrolyte Reviews I, J. R. MacCallum, C. A. Vincent (eds), Elsevier, London, Ch. 3 (1987) 39. m. a. ratner, a. nitzan, Faraday Discuss. Chem. Soc. 88 (1989) 19. c. a. angell, Solid State Ionics 18–19 (1986) 72. c. a. angell, Ann. Rev. Phys. Chem. 43 (1992) 693. l. m. torell, c. a. angell, Br. Polym. J. 3 (1988) 173. m. mclin, c. a. angell, J. Phys. Chem. 92 (1988) 2083. c. a. angell, c. liu, e. sanches, Nature 362 (1993) 137. t. yamamoto, m. ynami, t. kanbara, Chem. Mater. 6 (1994) 44. c. t. imrie, m. d. ingram, g. s. mchattie, Advanced Mater. 10 (1999) 832. x. wei, d. f. shriver, Chem. Mater. 10 (1998) 2307. s. k. jeong, y. k. jo, n.-j., jo, Electrochem. Acta 52 (2006) 1549. e. r. leite, f. l. souza, s. r. lazaro, p. r. bueno, e. longo, Chem. Mater. 17 (18) (2005) 4561. c. a. angell, c. liu, e. sanches, Nature 362 (1993) 137. s. t. pas, r. d. banhatti, k. funke, Solid State Ionics 177 (2006) 31135. h. a. every, f. zhou, m. forsyth, d. r. macfarlane, Electrochimica Acta, 43 (1998) 1465. f. l. souza, p. r. bueno, e. longo, e. r. leite, Solid State Ionics 166 (1–2) (2004) 83. f. l. souza, e. longo, e. r. leite, Chem. Phys. Chem. 8 (2007) 1778. f. l. souza, e. longo, e. r. leite, Chem. Phys. Chem. 9 (2) (2008) 245. f. l. souza, m. a. aegerter, e. r. leite, Electrochem. Acta 53 (2007) 1635. f. l. souza, m. a. aegerter, e. r. leite, Sol. Energy Mater. Cells 91 (2007) 1825. c. sanchez, g. j. de soler-illia, f. ribot, t. lalot, c. r. mayer, v. cabuil, Chem. Mater. 13 (2001) 3061. k. j. shea, d. a. loy, Chem. Mater. 13 (2001) 3306. g. cerveau, r. j. p. corriu, e. framery, Chem. Mater. 13 (2001) 3373. g. schottner, Chem. Mater. 13 (2001) 3422. j. pyun, k. matyjaszewski, Chem. Mater. 13 (2001) 3436. u. schubert, Chem. Mater. 13 (2001) 3487. z. ahmad, j. e. mark, Chem. Mater. 13 (2001) 3320. l. m. bronstein, c. joo, r. l. karlinsey, a. ryder, j. w. zwanziger, Chem. Mater. 13 (2001) 3678. e. manias, a. touny, l. wu, k. strawhecker, b. lu, t. c. chung, Chem. Mater. 13 (2001) 3516. l. m. bronstein, e. ashcraft, p. de santo jr., r. l. karlinsey, j. w. zwanziger, J. Phys. Chem. B 108 (2004) 5851.



601

58. a. hayashi, l. wang, c. a. angell, Electrochim. Acta 48 (2003) 2003–2008. 59. v. di noto, m. vittadello, s. g. greenbaum, s. suarez, k. kano, t. furukawa, Phys. Chem. B 108 (2004) 18832. 60. t. fujinami, m. a. mehta, k. sugie, k. mori, Electrochim. Acta 45 (2000) 1181. 61. a. walcarius, Chem. Mater. 13 (2001) 3351. 62. b. m. novak, Adv. Mater. 5 (1993) 422. 63. j. e. mark, Polym. Eng. Sci. 36 (1996) 2905. 64. l. mascia, Trends Polym. Sci. 3 (1995) 61. 65. d. a. loy, k. j. shea, J. Chem. Rev. 95 (1995) 1431. 66. p. judeinstein, c. sanches, J. Mater. Chem. 6 (1996) 511. 67. c. sanches, f. ribot, New J. Chem. 18 (1994) 1007. 68. k. nishio, t. tsuchiya, Solar Energy Mater. Solar Cell 68 (2001) 295. 69. j. n. hay, h. m. raval, Chem. Mater. 13 (2001) 3396–3403. 70. r. j. p. corriu, d. leclercq, Angew. Chem. Int. Ed. Engl. 35 (1996) 1420. 71. e. r. leite, n. l. v. carreño, e. longo, f. m. pontes, a. barison, a. g. ferreira, y. maniette, j. a. varela, Chem. Mat. 14 (2002) 3722. 72. m. r. silverstein, g. c. bassler, t. c. morrill, Spectrometric Indentification of Organic Compounds, 5th ed., John Wiley & Sons, Inc., New York (1991) p. 120. 73. p. g. bruce, c. a. vincent, J. Chem. Soc. Faraday Trans. 89 (1993) 2835. 74. s. d. druger, a. nitzan, m. a. ratner, J. Chem. Phys. 79 (1983) 3133. 75. h. vogel, Phys. Z. 22 645 (1926) 1921. 76. g. tamman, w. hesse, Z. Anorg. Allg. Chem. 156 (1926) 245. 77. g. s. fulcher, J. Am. Chem. Soc. 8 (1925) 339. 78. c. m. lampert, Solar Energy Mater. Solar Cells 52 (1998) 207. 79. c. m. lampert, a. agrawal, c. baertlien, j. nagai, Solar Energy Mater. Solar Cells 56 (1999) 449. 80. c. m. lampert, in: Proceedings of SPIE, Switchable Materials and Flat Panel Displays, Bellingham (ed), Washington, USA, SPIE, 3788 (1999) 2–11. 81. c. m. lampert, Solar Energy Mater. Solar Cells 76 (2003) 489–499. 82. a. agrawal, h. r. habibi, r. k. agrawal, j. p. cronin, d. m. roberts, r. caronpopowich, c. m. lampert, Thin Solid Film 221 (1992) 239. 83. a. agrawal, j. p. cronin, r. zhang, Solar Energy Mater. Solar Cells 31 (1993) 9. 84. a. azens, c. g. granqvist, Proc. SPIE 4458 (2001) 104. 85. c. g. granqvist, Int. Glass Rev. 2 (2001) 67–68. 86. c. g. granqvist, Interface 3 (2001) 18. 87. a. azens, c. g. granqvist, J. Solid State Electrochem. 7 (2003) 64. 88. a. azens, g. gustavsson, r. karmhag, c. g. granqvist, Solid State Ionics 165 (2003) 1. 89. c. g. granqvist, e. avendaño, a. azens, Thin Solid Film 442 (2003) 201. 90. j. h. klems, Energy Build. 33 (2001) 193. 91. c. g. granqvist, Smart Mater. Bull. 10 (2002) 9. 92. v. v. truong, f. e. giroud, p. v. ashrit, in: ‘Large-Area Chromogenics: Materials and Devices for Transmittance control’, C. M. Lampert, C. G. Granqvist (eds), SPIE, Optical Engineering Press, Bellingham, Washington (1990) 386. 93. m. a. aegerter, in: Structure and Bonding, R. Reisfeld, C. K. Jorgensen (eds), Springer-Verlag, Berlin, 85 (1996) 149–194. 94. p. m. s. monk, r. j. mortimer, d. r. rosseinsky, Electrochromism – Fundamentals and Applications, VCH, Weinheim, (1995).


602


95. b. scrosati, in: Applications of Electroactive Polymers, Chapman and Hall London (1993) 250–288. 96. s. heusing, m. a. aegerter, in: Sakka S. Eds. ‘Handbook of Sol-Gel Science and Technology, Processing Characterization and Applications’ Vol. III. Applications of Sol–Gel Technology, S. Sakka (ed), Kluwer Academic Publishers, Boston: MA (2005) 719–760. 97. m. a. de paoli, g. casalbore-miceli, e. m. girotto, w. a. gazotti, Electrochim. Acta 44 (1999) 2983. 98. m. a. de paoli, a. f. nogueira, d. a. machado, c. longo, Electrochim. Acta 46 (2001) 4243. 99. p. c. barbosa, m. m. silva, m. j. smith, a. gonçalves, e. fortunato, Electrochim. Acta 52 (2007) 2938. 100. p. c. barbosa, m. m. silva, m. j. smith, a. gonçalves, e. fortunato, Thin Solid Films 516 (2008) 1480. 101. c. g. granqvist, Appl. Phys. A57 (1993) 19. 102. g. vaivars, m. furlani, b. e. mellander, c. g. granqvist, J. Solid State Electrochem. 7 (2003) 724. 103. j. zhang, j. p. tu, x. h. xia, y. qiao, y. lu, Solar Energy Mater. Solar Cells 93 (2009) 1840.


Index

α-alumina, 570 α,β,β-trifluorostyrene, 15 acid anionic polymer electrolyte membrane (AAPEM), 574 acidic fillers, 76, 81 Aciplex, 221, 335 2-acrylamide-2-methyl-1-propane sulphonic acid (AMPS) copolymer, 511–12 K+-doped, 512–13 3-acryloyloxy-methyl-3′-methyloxetane (AMO), 561 activated carbon electrodes, 447–51, 455 effect of thickness, 447 high-density graphite electrodes XRD patterns, 449 various gel electrolytes cyclic voltammograms, 448 activation energy, 272, 288, 295 of diffusion, 297 values, 347 activation volume, 30 agar, 105–6 alkaline battery, 574–5 alkaline cationic polymer electrolyte membrane (ACPEM), 574 alkaline polymer electrolytes, 504–5 poly(ethylene oxide) (PEO)–KOH– nanopowders, 505 poly(vinyl alcohol) (PVA)– carboxymethyl cellulose (CMC)–KOH–H2O, 505

poly(vinyl alcohol) (PVA)–KOH– H2O, 504–5 alumina, 73 AMBER, 319 amphiphilic ruthenium dye, 419 anion exchange membrane (AEM), 43 anion transference number, 23, 37 anionic clay, 136–7 anomalous X-ray scattering (AXS), 366 antimonic acid (AA), 495–6 argon, 533 Arrhenius behaviour, 288, 345, 388, 591 Arrhenius equation, 345, 347, 499, 575 autocomplexation, 355 back hopping rate, 269, 270 Ballard Advanced Materials 3rd Generation of membranes (BAM3G), 15 batteries, 39–42, 62 see also electrochemical cells; lithium battery; lithium ion battery Li+-conducting PEO-based electrolytes, 325–9 Bellcore method, 82, 494–5, 567 Bellcore system, 5 benzidine, 408 bidentate configuration, 200 biopolymers, 96 blending, 569 Bloembergen-Purcell-Pound theory, 288 Boltzmann constant, 345, 401

603 © Woodhead Publishing Limited, 2010

604

Index

Born–Mayer–Huggins form, 319 Born–Oppenheimer approximation, 317 boron, 483 boroxine, 566 broadband dielectric spectroscopy (BDS), 264, 272 Bruce–Vincent method, 78 bulk resistance, 32 canonical ensemble (NVT), 320 capacitance, 48, 447, 449 carbon, 280 carbon nanotubes (CNTs), 456–9 casting method, 454 cation/anion interactions, 190–201 di-ureasils doped with LiBF4, 199–201 di-ureasils doped with LiCF3SO3, 192–6 di-ureasils doped with LiTFSI, 196–7 cation/crosslink interactions, 202–7 amide I region, 203–7 amide II region, 207 cation exchange membrane (CEM), 16 cation/polymer interactions, 186–90 CH2 rocking rCH2 region, 186 skeleton COC stretching vCOC region, 186 cation transference number, 22, 38, 75 cationic clay, 132–6 composite polymer electrolyte elaboration, 132 conductivity behaviour, 135–6 PEO/clay/NaClO4 nanocomposites conductivity, 136 mechanical properties, 136 structure modification, 132–5 cationic ring-opening polymerisation (CROP), 559, 560 cationic transference numbers, 146, 148, 341, 353–8 cationic transport numbers, 341 cellulose derivatives, 97–101, 106–7 chitosan, 107–8 gelatine, 108–9 hydroxyethylcellulose, 106–7

hydroxypropylcellulose ester, 106 rubber, 108 cellulose whiskers, 147 ceramic fillers, 67, 138 ceramic polymer electrolytes, 62–83 conductive fillers, 70–1 electrochemical cell design evolution, 64 experimental approaches, 67–70 composites preparation, 67–9 examination methods, 69–70 filler agglomeration and sedimentation, 68 filler surface impact on the transport properties, 75–7 first composites - conductive fillers, 70–1 PEO-LiClO4 system phase diagram, 72 impact of filler surface on the transport properties, 75–7 lithium transference number values, 78 sulphate groups onto metal oxide surfaces, 76 insulating fillers development, 72–4 conductivity of LiClO4, 74 decrease of crystallinity, 74 interfacial concerns, 77–9 symmetrical lithium cell resistance evolution, 79 other types, 80–2 polymer structures as solid solvents, 66 CHARMM, 319 chemical shift, 281 chemical shift anisotropy (CSA), 305 chitosan, 101–2, 107–8, 499–500 chronoamperometry, 536 citric acid, 587 Cloisite, 133 Cloisite 20A, 133 Cloisite 30B, 133, 134 Cloisite-Na, 133, 134 cluster-network model, 330 co-ordination number function, 322


Index complex plane impedance, 591 composite polymer electrolytes, 404, 552 based on metal oxides, 156–60 electrochemical performance, 156–7 interfacial characteristics, 157–9 matrix physical/chemical characteristics, 159 membranes tensile strength, 160 metal oxide concentration, 159–60 proton-conduction mechanism model, 158 proton conductivity of membrane containing various SiO2, 160 electrochemical devices future trends, 166–7 lithium batteries, 130–1 solid polymer electrolytes, 131 gel polymer electrolytes, 148–51 based on clays, 149–50 based on metal oxide fillers, 150–1 hygroscopic solid inorganic proton conductor, 160–4 based on heteropolyacids, 163–4 based on zirconium phosphate, 161–3 ionic conductivity and transference number, 572 lithium ion batteries, 570–2 poly(ethylene oxide) and clays, 132–7 anionic clay, 136–7 cationic clay, 132–6 poly(ethylene oxide) and non-ionic fillers, 137–48 ionic conductivity evolution, 144 methyl oxide charges, 137–47 modes of bonding structures, 141 nanosized Al2O3 particles and PEO–LiCF3SO2 electrolyte surface interactions, 143 organic fillers, 147–8 PEO20LiCF3SO3 10 wt% Al2O3 nanocomposite electrolytes conductivity, 143

605

PEO8–LiClO4 and PEO8–LiClO4– ZnO conductivity vs temperature, 145 storage tensile modulus vs temperature, 148 proton exchange membrane fuel cells (PEMFC), 151–6 components, 153–6 elaboration, 155–6 self-humidifying composite electrolytes, 164–6 composite solid polymer electrolyte (CSPE), 459–60, 493–8 poly(ethylene oxide) (PEO)/siliceous hybrids, 496–7 preparation, 496–7 properties, 497 poly(vinylidene fluoride)–hexa fluoride propylene (PVdFHFP)/Al2O3, 498 poly(vinylidene fluoride)–hexa fluoride propylene (PVdF-HFP) /molecular sieve SBA-15, 497 zirconium phosphate (ZP)- and antimonic acid (AA), 495–6 preparation, 496 properties, 496 composites, 129 conducting polymer electrodes, 442–7 poly(methyl thiophene) p and n doping–dedoping, 443 conductive fillers, 70–1 conductivity, 344–53, 367–73 constant phase elements (CPE), 261 core–shell model, 319 coulomb fluid, 349 coulombic efficiency, 444, 447, 449, 462 coulombic forces, 77 counter-electrode, 598 cross-polarisation, 301–2 crystalline polymer system, 327 crystallinity, 367–73 3-(2-cyano ethoxy)methyl-3′methyloxetane (CMMO), 560 1D Pure Exchange (PUREX), 307 DAB-AB-64, 561


606

Index

dark current, 382 decoupled hybrid polyelectrolyte (DHP), 587, 596 complex-plane impedance in Cole– Cole plot, 592–3 ionic conductivity values, characteristic relaxation frequency and glass transition temperature different Li+ content at room temperature, 595 different Na+ content at room temperature, 595 decoupled polymer system, 586, 598 Degussa P25, 403 deuterium, 280 dibutyl phthalate (DBP), 567 dicyclohexylcarbodiimide (DCC), 527 diethyl carbonate, 488 difluoroalkoxyborane, 281 dimethylaminopyridine (DMAP), 527 dimethylene carbonate (DMC), 552 dimethylformamide (DMF), 361 dimethylsulphoxide (DMSO), 361, 527 direct methanol fuel cell (DMFC), 16, 43, 435 disorder-longitudinal acoustic mode (D-LAM), 140 dissipative particle dynamics, 315 double-layer capacitors, 441 Dow membrane, 12, 221, 335 DREIDING force field, 331, 333, 335 dry polymer electrolytes, 5, 67 dry solid polymer electrolyte, 477–87 branched poly(ethylene glycol) (PEG)-boronate ester polymers, 483–4 electrolyte preparation, 484 polymer synthesis, 484 polymers preparation, 484 dual-function electrolytes based on poly(trimethylene carbonate), 485 oxymethylene-linked poly(oxyethylene) (OMPEO) with PEG400 and 600 mixture

multi-block polymer preparation, 479 SPE film preparation, 479 photo-polymerised 2-hydroxyethylmethacrylate (HEMA), 485–6 electrochromic device application, 486 photo-polymerised and crosslinked PHEMA synthetic route, 486 SPE films preparation, 485 polyether with hyperbranched side chains, 479–81 macromonomer and network polymer electrolytes synthesis, 480 2-(2-methoxyethoxy)ethyl glycidyl ether (MEEGE) monomer synthesis, 479–80 network polymer electrolytes preparation, 480–1 synthesis of macromonomers, 480 poly(iminoethylene)-bpoly(oxyethylene)-bpoly(iminoethylene) (PEI–PEO–PEI) copolymer synthesis, 478 SPE preparation, 478 polyvinyl butyral, 486–7 poly(vinylacetate–methylacrylate)/ poly(methyl methacrylate) (P(VAc-MA)/PMMA), 482–3 poly(vinylacetate–methylacrylate) (P(VAc-MA)) preparation, 482–3 SPE preparation, 483 poly(vinylacetate– methylmethacrylate) (P(VAcMMA)), 481–2 polymer electrolyte preparation, 482 random copolymer synthesis, 482 random copolymer poly(ethylene oxide) and epichlorohydrin (P(EO-EPI)), 481 DuPont de Nemours process, 11


Index dye-sensitised solar cells, 381–424 additives in polymer electrolytes, 402–17 inorganic nanofillers, 403–6 ionic liquid, 411–17 organic molecules, 406–11 future mends, 423–4 ionic liquids chemical structure, 413 1-methyl-3-propylimidazolium iodide (MPII) and poly(ethylene oxide-co-propylene oxide) (P(EO–PO)) thermal analysis, 415 maximum generated power variation produced by solar module, 423 N3 and N719 dyes structures, 383 Nyquist diagram for electrolytes based on copolymer poly(ethylene oxide/2-(2methoxyethoxy)ethyl glycidyl ether), 391 PEO copolymers structure and properties, 389 plasticised and gel polymer electrolytes, 390–402 current–voltage characteristics, 397, 402 GBL concentration on open circuit voltage, 398 ionic conductivity as function of LiI concentration, 394 polyiodides ratio as function of salt concentration, 397 Raman spectra as function of LiI concentration, 395 salt concentration on conductivity of plasticised electrolytes, 393 TiO2 films transient absorption spectra, 400 poly(ethylene oxide-co-propylene oxide) (P(EO–PO)) current– voltage curves, 416 polymer electrolytes, 384–8 ionic conductivity plots as function of salt concentration, 386 PEO copolymers, 384–8

607

representation and the processes involved in energy conversion, 382 variation with time and sunlight exposure, 418 with GPE current–voltage curves, 412 dye-sensitised solar cells (DSSC), 577 modules assembled with 16 series connected DSSC, 422 performance of DSSC with different polymeric electrolytes, 399 polymer electrolyte-based DSSC commercialisation, 421–3 stability, 417–21 solid state DSSC assembly using polymer electrolyte, 387 dynamic mechanical analysis (DMA), 238, 258 ECDs see electrochromic devices echo decay, 307–8 Eigen diagram, 357 Eigen ions, 334 electric double-layer capacitor (EDLC), 47, 447 electric field gradients (EFG), 116 electrically responsive, 472 electrochemical capacitors, 47–50, 440–2 double-layer capacitors, 441 hybrid capacitors, 441–2 polymer gel-based supercapacitors, 442 pseudocapacitors, 441 solid electrolytes, 442 electrochemical cells, 62 design evolution, 64 electrochemical impedance spectroscopy (EIS), 69, 272, 392, 406 electrochemichromism, 50 electrochromic devices, 50–1, 208, 471–513, 576–7, 596 classical seven-layer structure, 473 effect and devices, 472–4 electrolytes, 474–5


608

Index

mechanism, 576 new type of polymer electrolyte, 506–13 polymer electrolyte with ionic liquids, 506–11 single ionic polymer electrolyte, 511–13 polymer electrolytes classification, 477–500 dry solid, 477–87 gel, 487–93 natural polymer matrix, 498–500 PGPE and CSPE, 493–8 polymer matrix, 475–7 composition and stereo structures, 476–7 different polymer hosts structure sketch map, 477 ionic passageway structure types, 475, 476 proton-conducting and alkaline polymer electrolytes, 500–5 alkaline polymer electrolytes, 504–5 proton-conducting polymer electrolytes, 500–4 electrochromic displays, 207–8 electrochromic effect, 472, 473 electrochromic electrode, 598 electrochromic technology, 207 electrochromism, 50, 207 electrode/di-ureasil/electrode assembly, 179 electrode phenomena, 38 electrolysis, 37 electrolyte, 62–3 electrolyte conductance, 24 electromotive force, 37 electron paramagnetic resonance, 122–3 electronic conductivity, 22 electronic paper, 471 electrophoretic nuclear magnetic resonance, 307–8 electrostatic energy, 48 empirical valence bond (EVB) method, 331

energy-dispersive x-ray spectroscopy (EDX), 254 Epichlomer-16, 481 epichlorohydrin, 558 EPR technique electron spin echo envelope modulation (ESEEM), 123 equations of state, 315 equilibrium phase diagrams, 372–3 equivalent circuit, 32 equivalent series resistance (ESR), 442 ethyl methyl caronate (EMC), 569 ethylene carbonate (EC), 552 ethylene glycol dimethacrylate (EGD), 568 ethylene oxide, 545, 558 europium, 361 europium bromide, 360 eutectic value, 370 exchange nuclear magnetic resonance, 305–7 extended X-ray absorption finestructure (EXAFS), 365–6 far-infrared spectroscopy, 254 faradic process, 449 faradic reaction, 441 Fermi level, 408, 416 ferroelectric filler, 80 finite element analysis, 315 finite-length Warburg diffusion, 393 Flemion, 221 force field, 317 Fourier transform (FT)-Raman spectroscopy, 186, 587 Fourier transform infrared (FTIR) spectroscopy, 19, 222, 281, 350, 587–8 free volume law, 344 FreedomCAR, 573 fringe-field method, 296 fuel cell polymers, 336 fuel cell technology, 5 fuel cell test, 537, 547 fuel cells, 42–4, 329–35, 572–4


Index fumed silica porous gel polymer electrolytes, 494 Fuoss–Krass calculations, 76 γ-butyrolactone (GBL), 392 galcanostatic charge–discharge technique, 463 gel electrolyte, 49, 51 gel polymer electrolytes, 67, 131, 148–51, 487–93, 551–2 12-crown-4-ether contents analysis, 410 blended with polymer, 569 current–voltage steady state voltammograms, 411 different gel composition, 491 disadvantages, 570 doped with 12-phosphotungstic acid (12-PWA), 503 doped with H3PO4, 503 lithium ion batteries, 566–70 polyethylene oxide (PEO), 490–1 poly(methyl methacrylate)/ oxymethylene-linked poly(oxyethylene) (PMMA/ OMPEO) blend composite, 491–2 CSPE preparation, 492 GPE preparation, 491–2 oxymethylene-linked poly(oxyethylene) (OMPEO) preparation, 491 poly(methyl methacrylate) (PMMA), 488–90 poly(methyl methacrylate) (PMMA)–transition metal complexes GPE preparation, 493 GPE properties, 493 polyvinyl butyral GPE preparation, 488 GPE properties, 488 preparation and ion conducting mechanism, 566–7 star network poly(ethylene glycol) (PEG), 492–3 GPE preparation, 493

609

star network polymer synthesis, 492–3 synthetic route and sketch map, 492 with added plasticiser, 567–8 gelatine, 108–9, 499 geometric capacitance, 24 Gibbs free energy, 433 Gibbs phase rule, 71 glass transition temperature, 96, 138 glycerol, 249 co-ordination cages lithium ions exchange model, 254 typical conformers, 252 glymes, 282 graft polymer electrolyte, 558–61 Grotthus mechanism, 151, 158, 162, 331, 333, 334, 396, 446 GUPL equation, 262, 264 Hammet numbers, 77 Hartmann-Hahn condition, 301 heteronuclear Overhauser enhancement spectroscopy (HOESY), 304, 305 heteropolyacids (HPA), 161, 500–1 heteropolytungstic acid, 577 high density graphite (HDG), 50 high energy density batteries, 273 high temperature fuel cells, 524–48 high temperature solid polymer electrolytes, 460–1 Hittorf method, 37, 354 hot pressing technique, 68 HSAB (hard and soft acids and bases) principle, 342–3, 357 hybrid capacitors, 441–2 hybrid gel electrolytes (HGEs), 249–52 dielectric spectra imaginary component, 253 migration step length and diffusion coefficient temperature dependence, 254 morphology, 250 proposed structural model, 251 synthesis of gels based on Li2PdCl4 or Sn(CH3)2Cl2, glycerol and Li3Fe(CN)6, 250


610

Index

hybrid inorganic-organic gels, 222 3D hybrid inorganic organic networks as polymer electrolytes (3D-HION-APE), 221, 222 composed by poly(oligoethylene glycol) moieties, 227 proton conducting materials, 235 hybrid inorganic-organic polymer electrolytes, 219–73 Al[(CH2CH2O)8.7]ρ/(LiClO4)z crosslinking Al atoms co-ordination geometries, 234 equivalent conductivity as a function of salt concentration, 263 equivalent conductivity vs reciprocal of temperature, 263 Nyquist plots, 234 structural model, 233 conductivity spectra real component in jump relaxation model and polymer segmental motion, 266–72 real part of complex conductivity vs ω in solid electrolytes, 267 methods, 252–66 complex conductivity component analysis, 262 conductivity studies, 260–6 conductivity vs T profiles VTFH simulations, 265 decomposition by Gaussian functions, 257 dependence vs temperature of σdc and segmental mode, 265 HGE band parameters of the v(CN) FT-IR vibrational modes, 258 materials electrical response, 264 Nyquist plots equivalent circuit model, 260 structural characterisation, 252–8 Z-IOPE [VIII] FT-IR absorption spectra, 256 Z-IOPEs gel formation viscometric study, 255

overview, 221–52 dependence of InσδX on 1/T for Nafion/[(TiO2) × (WO3)0.148], 241 direct current conductivity temperature dependence, 239 glycerol co-ordination cages lithium ions exchange model, 254 hybrid gel electrolytes, 249–52 PEO/EuCl3 complex TGT conformational model, 224 poly[(oligo ethylene oxide) ethoxysilane/(EuCl3)0.67 preparation procedure, 223 poly[PEG400-altdiethoxydimethylsilane] C-NMR spectrum, 225 poly[PEG400-altdiethoxydimethylsilane]/MgCl2 charge migration models, 227 redox mechanism for titanium species conductivity mechanism, 230 sol-gel plastic transition mechanism, 243 three dimensional hybrid inorganic-organic networks as polymer electrolytes, 222–39 typical helical conformation, 226 Z-IOPEs general structure, 241 zeolithic inorganic-organic polymer electrolytes, 239–49 polymer electrolytes fundamentals, 220–1 poly[(oligoethylene glycol) dihydroxytitanate] and poly[(oligoethylene glycol) dihydroxytitanate]/LiCl (II) complex proposed structures, 229 reaction procedures, 228 VTF graphs, 230 siloxanic proton-conducting PE precursor synthesis, 236 synthesis, 237, 238 Zr[(CH2CH2O)8.7]ρ/(LiClO4)zn 3D structure, 232


Index reaction steps, 231 tetrahedral co-ordination geometry around the crosslinking Zr atoms, 233 hybrid polymer electrolytes applications, 596–8 electrochemical devices, 583–99 general discussion, 594–6 physical-chemical properties, 587–94 system FT-Raman spectra, 588 system FTIR spectra, 589 hydrolytic sol–gel process, 587 hydronium, 332 3-(2-(2-(2-hydroxyethoxy)ethoxy) ethoxymethyl)-3′-methyloxetane (HEMO), 562 hydroxyethylcellulose, 106–7 2-hydroxyethylmethacrylate (HEMA), 485–6 3-hydroxymethyl-3′-ethyloxetane (HEO), 559 3-hydroxymethyl-3′-methyloxetane (HMO), 560 hydroxypropylcellulose ester, 106 Hyflon Ion, 335 hyperbranched polymer electrolyte (HBP), 561–4 high temperature fuel cells, 524–48 phosphonic acid group (HBP-PA), 537–47 current vs voltage characteristics, 547 DSC traces under argon, 544, 545 fuel cell measurement, 546–7 ionic conductivity, 541–3 ionic conductivity temperature dependence, 541, 542 synthesis, 537–41 thermal property, 543–6 thermogravimetry traces under air, 544, 546 VTF parameters, 542, 543 synthetic route HBP(H, L)–PA–acryloyl and crosslinked-HBP(H, L)–PA membrane, 540 HBP–PA(H, L), 538

611

HBP–SA and HBP–acryloyl, 526 HBP–SA–acryloyl and crosslinked–HBP–SA, 528 with sulphonic acid group (HBPSA), 525–37 current densities vs time profiles, 537 current vs voltage characteristics, 536 DSC thermograms under air, 533 fuel cell measurement, 535–7 glass transition temperatures and decomposition temperatures, 534 ionic conductivity, 529–32 synthesis, 525–9 temperature dependence in ionic conductivities, 530, 532 TG/DTA diagrams under air, 534, 535 thermal property, 533–5 time dependence in ionic conductivities, 531 VTF parameters, 530, 532 imidazolium cations, 329 impedance, 31 impedance plot, 32 impedance spectroscopy, 30, 228, 260, 299 inductively coupled plasma atomic emission spectroscopy (ICPAES), 253 infrared spectroscopy, 18, 185 inorganic-organic composite membranes, 153 inorganic particles, 153–4 insulating fillers, 72–4 intercrystalline amorphous phase, 368 interpenetrating polymer networks (IPNs), 9 intracrystalline amorphous phase, 367 inversion-recovery technique, 292 ion conductivity, 525, 554–5 ion diffusion, 448 ion exchange polymers, 434–6 ion insertion materials, 50


612

Index

ion jumps, 328 ion mobility, 344–53, 362, 570 ion pairing, 29, 326 ionic association, 365–7 ionic conductivity, 3–4, 22, 37–9, 178–82, 296, 433, 476–7 ionic conductor, 471 ionic crosslinking, 181 ionic liquid, 474, 566 in polymer electrolyte, 506–11 ionic mobility, 348 ionic passageway, 475 ionic polarisation cell technique, 37 ionic transference number, 354 ionic transport number, 354 irradiation method, 445 isothermal–isobaric ensemble (NPT), 320 J-coupling constants, 304 Jahn–Teller weakening, 358 jump relaxation model, 220, 261 and polymer segmental motion real component of conductivity, 266–72 components of overall potential and hopping dynamics, 268 Kynar2801-grade, 569 lanthanide cations, 359 lanthanum systems, 359 Laponite, 134 layered double hydroxides, 136 LDH see layered double hydroxides leapfrog, 317 Lennard–Jones form, 319 Lewis-type acid–base interactions, 386 Li+-conducting PEO-based electrolytes, 325–9 Li-ion technology, 83 Li-MMT, 133 light-dependent recombination current, 401 linear response theory, 269 liquid state nuclear magnetic response, 281–2

liquidus line, 369 LiTFSI, 130 lithium, 278, 279, 280 lithium battery, 40–1, 129, 130–1 lithium bis(trifluoromethanesulfonyl) imide, 178 lithium-doped hybrid polymer electrolytes, 176–210 Arrhenius conductivity d-U(600)nLiX di-ureasil systems, 180 d-U(900)nLiX di-ureasil systems, 179 d-U(2000)nLiX di-ureasil systems, 179 electrochemical stability, 184–5 voltammogram of d-U(2000)15LiTFSI di-ureasil electrolyte, 185 electrochromic displays, 207–8 bleached and coloured states for di-ureasils, 210 optical transmittance as a function of wavelength, 209 ionic conductivity, 178–82 spectroscopic studies, 185–207 cation/anion interactions, 190–201 cation/crosslink interactions, 202–7 cation/polymer interactions, 186–90 d-U (2000)nLiCF3SO3 di-ureasils FT-IR spectra, 189 d-U(2000)nLiBF4 di-ureasils FT-IR spectra, 202 d-U(2000)nLiCF3SO3 di-ureasils FT-IR υCOC region, 187 d-U(2000)nLiCF3SO3 di-ureasils FT-IR spectra, 204–6 d-U(2000)nLiCF3SO3 di-ureasils FT-Raman spectra, 191 d-U(2000)nLiTFSI of di-ureasils FT-IR spectra, 199–201 FT-Raman δ3CF3 region curvefitting results, 198 room temperature FT-IR spectra, 193–6


Index thermal properties, 182–4 d-U(2000)nLiTFSI and d-U(900)n LiTFSI ormolytes DSC curves, 183 LiBF4-doped d-U(600)-d-U(900)and d-U(2000)-based electrolytes Tg variation, 182 lithium ferricyanide, 247 lithium ion battery, 40–2 composite polymer electrolytes, 570–2 gel polymer electrolytes, 566–70 polymer electrolytes advantages, 552–4 solid polymer electrolytes, 557–66 structure, 552, 553 lithium perchlorate, 178 lithium polymer electrolytes, 6–10 lithium-polymer technology, 220 lithium trifluoromethanesulfonimide (LiTFSI), 558 lithium Z-IOPE, 247 macroscopic anions, 80 magic angle spinning (MAS), 282, 299, 301 mean residence time, 335 mean-square displacements (MSD), 323–4 medium infrared spectroscopy, 256 melt intercalation, 132 membrane electrode assembly (MEA), 537, 573 metal oxide charges, 137–47 cationic transference number and interface properties, 146 composite polymer electrolyte elaboration, 137–8 conductivity behaviour, 142–5 mechanical properties, 146–7 structure modification, 138–42 metal oxides, 156–60 3-(2-(2-(2-methoxy ethoxy) ethoxymethyl)-3′-methyloxetane (MEMO), 560 2-(2-methoxyethoxy)ethyl glycidyl ether (MEEGE), 558–9 monomer synthesis, 479–80

613

1-methyl-3-propylimidazolium iodine, 419 MHB venture, 8 microcanonical ensemble (NVE), 320 mid infrared Fourier transform infrared (FT-IR) spectra, 186 MMT see montmorillonite modulated differential scanning calorimetry (MDSC), 237 molecular dynamics simulation computational chemistry, 315 different techniques, 316 Li ion and H-conduction in polymer electrolytes, 314–37 LiPF6 × PEO6 structure, 327 mean-square displacements for H2O and H3O+ in Nafion membrane, 324 methodology, 316–25 dynamical properties, 324–5 equilibration and non-equilibrium molecular dynamics, 320–1 interaction potentials, 317–19 periodic boundary conditions and thermodynamical ensembles, 319–20 potentials for Li+–SiF62interactions development, 318 simulation methodology, 316–17 structural properties, 321–4 polymer electrolyte MD box content folding operation, 323 radial distribution function and co-ordination number function, 322 monodentate configuration, 199 montmorillonite (MMT), 132, 149, 405 multi-walled carbon nanotubes (MWCNTs), 457–8 multiscale modelling, 315 multivalent polymer electrolytes characterisation and modelling, 340–73 characteristic rate constants for inner-sphere water molecules substitution, 357


614

Index

free ions, ion pairs and higher aggregates percentage variation, 351 ionic transport properties, 344–58 cationic transference numbers, 353–8 conductivity and ion mobility, 344–53 crystallinity and conductivity, 367–73 effect of morphological and crystallographic structures, 358–65 ionic association, 365–7 polymer–complex formation, 342–4 ion–polymer interactions, 342–3 PEO as polymer base, 343–4 N-butyl-N-methylpyrrolidinium bis(trifluoromethanesulphonyl) imides (PYR14TFSI), 506 N3 dye, 419 N-methyl-N-propylpyrrolidinium bis(trifluoro-methanesulfonyl) imide, 566 Nafion, 5, 44, 151, 155, 156, 221, 236, 237, 329, 335, 431, 434–5, 524 chemical structure, 434 MD box, 330 mean-square displacements for H2O and H3O+, 324 membrane, 11, 17 structure, 12 model axial view with different backbone conformations, 240 Nafion 105, 12, 159 Nafion 112, 159 Nafion 115, 12, 159, 454 Nafion 117, 12, 14, 15, 44, 453 Nafion 324, 14 Nafion 417, 14 Nafion 1100, 454 Nafion-Teflon-α-ZPL membranes, 162 Nafion/zirconium sulphophenyl phosphate, 163 Nafion/ZP, 161 nanocomposite electrolytes, 439

nanofillers, 139, 146, 166, 236 nanogels, 149–50 NASICON, 70 neodymium, 361 Nernst-Einstein equation, 269, 296 Nernstian diffusion, 406 Newton’s equations of motion, 316, 317 nickel-cadmium, 40 nickel hydroxide positive electrode, 459 non-Arrhenius behaviour, 591 non-doped di-ureasils, 178 non-hydrolytic sol–gel process, 587 non-ionic fillers, 137–48 NRG50, 454 nuclear magnetic resonance spectroscopy background, 278–9 polymer electrolytes, 279 diffusion measurements, 295–9 cation, anion and solvent species diffusion coefficient temperature dependence, 297 diffusion experiments, 296–8 fringe-field measurements, 296 pulse gradient spin-echo measurements, 295–6 transport numbers measurements, 298–9 double resonance experiments, 300–4 cross-polarisation, 301–2 EDS-LiClO4 C CP/MAS spectra, 302 Li proton-decoupled MAS NMR spectra deconvulation, 300 proton decoupling, 300–1 rotational echo double resonance, 303–4 relaxation processes, 291–5 Li spin lattice relaxation rates temperature dependence, 294 resonance intensity vs delay time, 292 spin-lattice relaxation, 291–2 spin-lattice relaxation in rotating frame, 293–5 spin-spin relaxation, 293


Index solid state nuclear magnetic resonance, 282–91 amorphous phase transport, 290 composition, 285–7 dynamics, 287–9 E-V6E4 Li NMR spectrum temperature dependence, 289 H linewidth vs temperature plot, 286 least square peak fit of Li NMR spectra, 287 Li and F NMR linewidths for second heating and cooling cycles, 284 Li and F NMR linewidths vs temperature, 284 Li linewidth vs temperature plot, 286 morphology, 283 oriented systems, 290–1 temperature dependence of corresponding correlation time, 290 thermal properties, 283–5 use in polymer electrolyte research, 278–308 electrophoretic nuclear magnetic resonance, 307–8 ENMR pulse sequence, 308 exchange nuclear magnetic resonance, 305–7 H NMR spectrum of PMAML, 281 liquid state nuclear magnetic response, 281–2 magic angle spinning, 299 nuclei possibility, 279–80 plasticiser and new polymer matrix interactions, 282 SiO2-PEO hybrid materials C NMR, 306 two-dimensional methods, 304–5 nuclei possibility, 279–80 Nyquist diagram, 391, 409 electrolytes based on copolymer poly(ethylene oxide/2-(2methoxyethoxy)ethyl glycidyl ether), 391

615

polymer electrolyte films sandwiched between planar Pt electrodes, 392 O–C–C band, 590 ohmic behaviour, 23 oligo(oxyethylene)-modified LDH, 137 open-circuit voltages (OCV), 536 OPLS-AA, 319 organic fillers, 147–8 ormolytes, 177, 184 P3, 44 palladium complex, 247 passivation, 41, 462 Pearl Black 2000 carbon powder, 462 PEG600, 247 PEGDME, 347 PEG400–(MgCl2) system, 349 PEO-LiClO4-ZnO, 145 PEO-LiTFSI, 146–7, 147 PEODME250, 80 PEO–La(ClO4)3 system, 372 PEOn–ZnI2 system, 346, 358 PEO–Tm(SO3CF3), 348 perfluorinated membranes, 11, 13–14 monomer membrane properties, 13 perfluorinated polymer electrolytes, 221 perfluorosulphonic acid (PFSA), 44, 329–35, 336, 573–4 polymer, 524 perfluorosulphonimide anion, 344 periodic boundary conditions and thermodynamical ensembles, 319–20 in two dimensions, 320 phase inversion process, 494, 567 photoelectric energy converter, 45 photoelectrochemical cells (PECs), 46 photoelectrochemical devices, 45–7 photoelectrochemical energy converter, 45–6 photovoltaics, 577 plasticisation, 97 plasticised polymer electrolytes, 8 plasticised polymer films, 11


616

Index

plasticised polymers, 5 plasticiser salt, 130 plasticisers, 130, 390 poly(2-acryloamide-2-methyl1propanesulphonic acid)/ poly(vinylidene fluoride) (PAMPSA/PVdF), 502–3 poly(1-butly-4-vinyl pyridinium) halides, 8 poly(3,4-ethylenedioxythiophene) (PEDOT):CNT film supercapacitor, 457 poly(1-oligo(ethylene glycol) methacrylate-3methylimidazolium chloride, 417 polyacrylonitrile (PAN), 566 polyaniline (PANI), 445, 452, 453 polybenzimidazobenzophenanthrolines (PBIPAs), 16 poly(benzimidazole), 525 poly(benzylsulphonic acid) siloxane, 17 polyelectrochromic, 472 polyether-based polymers, 436–8 poly(ethylene glycol) (PEG) gel polymer electrolytes, 492–3 poly(ethylene glycol) (PEG)-boronate ester polymers solid polymer electrolytes, 483–4 poly(ethylene glycol)dimethyl ether (PEGDME), 190 poly(ethylene glycol)methyl ether) (PEGME), 391, 404 poly(ethylene oxide) and epichlorohydrin (P(EO-EPI)) random copolymer, 481 poly(ethylene oxide) (PEO), 282, 325, 341, 583, 590 and clays, 132–7 as polymer base, 343–4 copolymers, 384–8 structure and properties usually employed in DSSC electrolytes, 389 dissolved divalent cations groupings, 356 divalent microstructural regions, 364 gel polymer electrolytes, 490–1

PEO-lithium salt complex structure, 432 PEOnZnI2 polymeric system equilibrium phase diagrams, 371 poly(ethylene oxide) (PEO)/siliceous hybrids composite solid polymer electrolyte, 496–7 poly(ethylene oxide) (PEO)–KOH–nanopowders alkaline polymer electrolytes, 505 poly(iminoethylene)-bpoly(oxyethylene)-bpoly(iminoethylene) (PEI–PEO–PEI), 478 polymer blending, 460 polymer chain motion, 531, 542, 548 polymer electrolyte, 279 research using nuclear magnetic resonance spectroscopy, 278–308 background, 278–9 diffusion measurements, 295–9 double resonance experiments, 300–4 electrophoretic nuclear magnetic resonance, 307–8 exchange nuclear magnetic response, 305–7 liquid state nuclear magnetic resonance, 281–2 magic angle spinning, 299 nuclei possibility, 279–80 relaxation processes, 291–5 solid state nuclear magnetic resonance, 282–91 two-dimensional methods, 304–5 polymer electrolyte fuel cells (PEFC), 42–3, 524 polymer electrolyte membrane, 10 polymer electrolytes, 3–52, 384–8 see also specific type of polymer electrolytes 1-[3-(trimethoxy-λ4-silyl)propyl] imidazole (TMSPIm)-based, 507–8 precursor synthesis, 507


Index preparation, 507–8 properties, 508 advantages in lithium ion batteries, 552–4 better shape flexibility and manufacturing integrity, 554 enhanced endurance to varying electrode volume during cycling, 554 improved safety, 553 reduced reactivity with liquid electrolyte, 554 applications in practical devices, 39–51 batteries, 39–42 capacitors classification according to energy storage mechanism, 47 commercial cation exchange properties, 45 electrochemical capacitors, 47–50 electrochromic devices, 50–1 fuel cells, 42–4 photoelectrochemical devices, 45–7 based on incorporating ionic liquids into poly(vinylidene fluoride)– hexa fluoride propylene (P(VdF-HFP)) polymer, 510–11 ionic conductivity, 510–11 poly(vinylidene fluoride)-hexa fluoride propylene (P(VdFHFP))-ionic liquid gel preparation, 510 based on natural polymer matrix, 498–500 chitosan-based SPE, 499–500 gelatine-based SPE, 499 starch derivatives, 498–9 based on natural polymers, 95–124 future trends, 123–4 categories, 5–17 lithium polymer electrolytes, 6–10 perfluorinated monomer membranes properties, 13 proton polymer electrolytes, 10–17 conductivity measurements, 20–39 alternating current conductivity, 30–5

617

conductivity as a function of n in PEOn:Mg(ClO4)2, 28 conductivity variation, 29 contributions to conductivity, 20–3 direct current conductivity, 23–30 direct current electronic conductivity, 35–7 effect of double layer capacitance/ geometric capacitance ratio on impedance plot, 32 electrodes used to fix M/X ratio, 21 electrodes used to measure AC conductivity of Pn:MX, 21 ionic conductivity, 37–9 Ni/PEO6:NiCl2/Ni cell measured impedances, 33 PEO:salt complexes maximum conductivities, 27 resistor and constant phase element impedance representation, 34 steady-state electronic current vs DC potential for a polarised cell, 36 test cell equivalent circuit, 24 Wagner’s DC polarisation cell, 35 different natural polymer-based ionic conducting systems, 112 dye-sensitised solar cells, 381–424 additives, 402–17 commercialisation of DSSC, 421–3 conclusions and future mends, 423–4 plasticised and gel polymer electrolytes, 390–402 stability, 417–21 electrochromic devices, 471–513 electrochromic effect and devices, 472–4 electrolytes, 474–5 new type of polymer electrolyte, 506–13 polymer electrolytes classification, 477–500 polymer matrix, 475–7


618

Index

proton-conducting and alkaline polymer electrolytes, 500–5 electrolyte composition and morphology grouping, 584 fundamentals, 220–1 grafted natural polymer based SPE, 97–102 cellulose derivatives, 97–101 chitosan, 101–2 starch, 102 hydroxyethylcellulose (HEC)– poly(ethylene oxide) (PEO) network formation grafting with Jeffamine, 100 grafting with Jeffamine diisocyanate, 98 idealised equivalent circuit and complex plane plot, 38 ion mobility mechanisms, 584–5 ionic conductivity values as a function of temperature hydroxyethylcellulose (HEC)based networks, 99 hydroxyethylcellulose (HEC)– poly(ethylene oxide) (PEO), 101 ionic liquids, 506–11 ionic liquids and polymeric ionic liquids, 508–10 ionic conductivity, 509–10 preparation, 509 synthesis, 508–9 Li ion and H-conduction molecular dynamics simulation, 314–37 computational chemistry, 315 conclusions and future trends, 336–7 Li+-conducting PEO-based electrolytes for batteries, 325–9 molecular dynamics methodology, 316–25 perfluorosulphonic acid systems, 329–35 lithium-doped hybrid, 176–210 electrochemical stability, 184–5 electrochromic displays, 207–8 ionic conductivity, 178–82

spectroscopic studies, 185–207 thermal properties, 182–4 main properties, 554–7 AC impedance plot, 555 chemical, thermal and electrochemical stabilities, 555–6 cyclic voltammogram curve, 556 ion conductivity, 554–5 linear sweep voltammetry curve, 556 mechanical strength, 557 transference number, 555 N-butyl-N-methylpyrrolidinium bis(trifluoromethanesulphonyl) imides (PYR14TFSI), 506 preparation, 506 properties, 506–7 natural polymers magnetic resonance spectroscopy, 111–23 electron paramagnetic resonance, 122–3 gel electrolyte X-band CW-EPR spectrum, 123 Li spin lattice relaxation rates deconvolution, 121 line shape analysis, 115–18 low temperature Li NMR spectrum, 115 parameters from Li NMR, 116, 121 spin–lattice relaxation, 118–22 other battery types, 572–7 alkaline battery/supercapacitors, 574–5 electrochromic devices and solar cells, 576–7 fuel cells, 572–4 single proton exchange membrane cell, 573 other natural polymer-based systems, 109–11 PEOn: salt ratios maximum conductivities, 26 minimum conductivities, 26 plasticised natural polymer-based SPEs, 102–9 agar, 105–6 cellulose derivatives, 106–7


Index grafted and crosslinked natural polymer-based ionic conducting systems, 103 ionic conducting systems, 110 ionic conductivity measurements, 104 starch, 103–5 solid solvents and their applications, 550–77 composite polymer electrolytes in lithium ion batteries, 570–2 gel polymer electrolytes in lithium ion batteries, 566–70 lithium ion battery structure, 552 solid polymer electrolytes in lithium ion batteries, 557–66 structure and its implications, 17–20 temperature dependence H spin lattice relaxation rates in gel electrolytes, 122 Li NMR central linewidth temperature dependence, 117 Li spin-lattice relaxation rates, 118 thermomechanical characterisation, 406 polymer gel-based supercapacitors, 442 polymer gels, 438–9 polymer-in-ceramic, 81, 83 polymer silicate nanocomposites, 132 polymeric solid electrolyte, 585 polymers, 3 poly(methoxyethoxyethoxy) phosphazene (MEEP), 590 poly(methyl methacrylate)/ oxymethylene-linked poly(oxyethylene) (PMMA/ OMPEO) blend composite, 491–2 poly(methyl methacrylate) (PMMA), 566–7 gel polymer electrolytes, 488–90 proton-conducting polymer electrolytes, 502–3 poly(methyl methacrylate) (PMMA)– poly(vinylidene fluoride) (PVdF)-based gel electrolytes, 298

619

poly(methyl methacrylate) (PMMA)– transition metal complexes, 493 poly(methyl methacrylate– heteropolyacid) (PMMA–HPA), 501–2 poly(methyl thiophene) (PMT), 443 poly(o-methoxyaniline), 384 poly[(oligo ethylene oxide) ethoxysilane, 222 polyoxometalate (POM), 453 poly(oxyethylene) methacrylate, 133 poly[PEG400-altdiethoxydimethylsilane], 224 poly[PEG400-altdiethoxydimethylsilane]/ (MgCl2), 223 polypyridinium salts, 9 polytetrafluoroethylene, 13 poly(trimethylene carbonate), 485 poly(vinyl alcohol) (PVA)/imidazole/ NH4H2PO4, 503–4 poly(vinyl alcohol) (PVA)/NH4H2PO4, 504 poly(vinyl alcohol) (PVA)– carboxymethyl cellulose (CMC)–KOH–H2O, 505 polyvinyl butyral, 486–7, 488 poly(vinylacetate–methylmethacrylate)/ poly(methyl methacrylate) ((P(VAc-MA)/PMMA), 482–3 poly(vinylacetate–methylmethacrylate) (P(VAc-MMA)), 481–2 poly(vinylidene fluoride) (PVdF), 438, 560, 567, 568 poly(vinylidene fluoride)–hexa fluoride propylene (P(VdF-HFP)) ionic liquid gel preparation, 510 ionic liquids incorporation, 510–11 porous gel polymer electrolytes, 494–5 poly(vinylidene fluoride)–hexa fluoride propylene (PVdF-HFP)/Al2O3, 498 poly(vinylidene fluoride)–hexa fluoride propylene (PVdF-HFP)/ molecular sieve SBA-15, 497


620

Index

poly(vinylidene fluoride)–hexa fluoride propylene (PVdF-HPA), 501–2 poly(vinylpyridine-co-acrylonitrile), 408 polyvinylpyridine (PVP), 407 porous gel polymer electrolyte (PGPE), 493–8 fumed silica preparation, 494 properties, 494 poly(vinylidene fluoride)–hexa fluoride propylene (P(VdFHFP)), 494–5 preparation, 495 preparation with phase inversion process, 495 potential energy surface, 317 propylene carbonate (PC), 552 proton conducting polymer electrolytes, 220, 500–5 GPE doped with 12-phosphotungstic acid (12-PWA), 503 GPE doped with H3PO4, 503 heteropolyacids (HPA) encapsulated with polystyrene sulphonic acid (PSS), 500–1 preparation, 501 properties, 501 poly(2-acryloamide-2-methyl1propanesulphonic acid)/ poly(vinylidene fluoride) (PAMPSA/PVdF) or poly(methyl methacrylate) (PMMA), 502–3 GPE preparation, 502 GPE properties, 502–3 poly(vinyl alcohol) (PVA)/imidazole/ NH4H2PO4, 503–4 poly(vinyl alcohol) (PVA)/ imidazole/NH4H2PO4, 504 poly(vinyl alcohol) (PVA)/ NH4H2PO4, 504 poly(vinylidene fluoride)–hexa fluoride propylene (PVdF-HPA) or poly(methyl methacrylate– heteropolyacid) (PMMA–HPA), 501–2 GPE preparation, 501–2 GPE properties, 502

proton decoupling, 300–1 proton donor–acceptor type membrane, 435 proton exchange membrane fuel cells (PEMFC), 6, 129, 151–6, 220–1, 273, 329 components, 153–6 composite polymer electrolytes, 154–5 inorganic particles, 153–4 polymer matrix, 153 composite electrolytes, 151–6 elaboration, 155–6 polymer structures, 154 proton exchange membrane (PEM), 152, 573 proton hopping, 333 proton polymer electrolytes, 10–17 pseudocapacitance, 451 pseudocapacitors, 441 Pt disk microelectrodes, 411 pulse gradient spin-echo (PGSE), 279 measurements, 295–6 PYR13TFSI, 566 QM-MM methods, 331 quadrupolar nuclei, 278 quantum mechanical methods, 315 radial distribution function (RDF), 321–2 Raman shift, 590 Raman spectroscopy, 18–19, 185, 222 redox capacitors, 48 relaxation rate, 291 relaxation time, 264, 291 Rietveld profile refinement, 19 rigid lattice, 117 rocking chair battery, 552 rotating-frame relaxation time, 291 rotational echo double resonance (REDOR), 303–4 rubber, 108 ruthenium dye, 419 ruthenium oxide, 447, 464 SBA-15, 287 scanning electron microscopy (SEM), 254


Index Schrödinger equation, 317 self-diffusion coefficients, 279 self-humidifying composite electrolytes, 164–6 silica gel, 450 siloxanic proton-conducting membranes, 235 silver tetrafluoroborate (AgBF4), 436 SIMPSON program, 303 single ionic polymer electrolyte, 511–13 2-acrylamide-2-methyl-1-propane sulphonic acid (AMPS) copolymer, 511–12 GPE properties, 512 preparation, 512 K+-doped 2-acrylamide-2-methyl-1propane sulphonic acid (AMPS) copolymer, 512–13 K+-poly(2-acryloamido-2-methyl1-propanesulphonic acid)-based SPE, 513 KCl-doped poly(2-acrylamide-2methyl-1-propane sulphonic acid) (AMPS)-based SPE preparation, 512–13 single membrane electrode assembly, 548 single-walled carbon nanotubes (SWCNTs), 458–9 small-angle neutron scattering (SANS), 330 small-angle X-ray scattering (SAXS), 330 smart window, 576 sol-gel process, 177, 222, 249, 450 sol-gel transition, 241, 250 solar cells, 576–7 solid electrolyte interface (SEI), 77 solid electrolytes, 432–3, 442 cation with segmental mobility in polymer chain, 434 central poly(ethylene glycol-copropylene glycol) triblock polymers, 436 conduction, 433–9 ion exchange polymers, 434–6 nanocomposite electrolytes, 439

621

polyether-based polymers, 436–8 polymer gels, 438–9 supercapacitors, 439–42 cell construction, 440 types, 440 various energy storage and conversion devices Ragone plot, 440 solid membrane films, 10 solid polymer electrolyte fuel cells (SPEFC), 6 solid polymer electrolytes (SPE), 6–7, 63, 65, 95, 131, 551 abbreviations, 464–6 advantages, 472 comb-like polyether and copolyether, 560 comb-like polymer, 559 ion conductivity, 561 composition and room temperature conductivity, 11 electrodes used to fix the M/X ratio, 21 hyperbranched copolyether synthesis, 565 ion conducting mechanism illustration, 557 lithium ion batteries, 557–66 graft polymer electrolyte, 558–61 hyperbranched polymer electrolyte, 561–4 ion conducting mechanism, 557–8 polyelectrolytes, 564, 566 solid polymer electrolyte blended with ionic liquid, 566 miscellaneous, 461–3 natural polymer matrix chitosan, 499–500 gelatine, 499 PHEMO synthesis and molecular structure, 563 supercapacitors, 431–66 activated carbon electrodes, 447–51 applications, 463 cation exchange membrane-based supercapacitors, 451–6


622

Index

conducting polymer electrodes, 442–7 current research activities, 456–63 solid electrolytes, 432–3, 439–42 solid electrolytes conduction, 433–9 synthesised dentrimer, 562 solid solvent see polymer electrolytes solid state nuclear magnetic resonance (NMR), 258, 266, 282–91 amorphous phase transport, 290 composition, 285–7 dynamics, 287–9 morphology, 283 oriented systems, 290–1 thermal properties, 283–5 solution-casting technique, 438, 463, 495 solvation process, 342 solvent casting, 361 sonication, 131 spatial distribution function (SDF), 323 SPE membrane fuel cells, 573 spectral density function, 119 Spectroline, 485 spherulites, 360 spin-1/2 nuclei, 278 spin-lattice relaxation, 119, 279, 291–2 rotating frame, 293–5 spin-spin relaxation, 293 stability range conductivity (SRC), 238 starch, 102, 103–5 derivatives, 498–9 stiffening effect, 76 sulphonated poly(ether ether ketone) (SPEEK), 435, 452 AC-crosslinked supercapacitor, 455 sulphonated polymers, 153 supercapacitors, 574–5 applications, 463 cation exchange membrane-based, 451–6 electrical double layer capacitor electrodes, 454–5 hybrid supercapacitors, 455–6 pseudocapacitor electrodes, 451–3 cell construction, 440

charge–discharge plot, 456 current research activities, 456–63 carbon nanotube-based composite electrode, 456–9 composite solid polymer electrolytes, 459–60 high temperature solid polymer electrolytes, 460–1 miscellaneous solid polymer electrolytes, 461–3 solid electrolytes, 432–3, 439–42 conduction, 433–9 solid polymer electrolytes, 431–66 specific capacitance and coulombic efficiency, 455 stability plot, 445 types, 441 various energy storage and conversion devices Ragone plot, 440 supramolecular electrolyte, 413 swing battery, 552 Teflon substrate, 67 term transport number, 23 4-tert-butylpyridine, 406 tetraethyl orthosilicate (TEOS), 450, 587 tetrahydrofuran, 527 tetramethyl ammonium hydroxide, 575 theorem of fluctuation-dissipation, 269 thermal history, 361 thermal runaway, 41 thermogravimetric analysis (TGA), 182 thermogravimetric/differential thermal analysis (TG/DTA), 533 thermomicroscopy, 372 thermooptical analysis, 369 thermosetting gel electrolytes (TSGE), 420 time-evolution plots, 325 [(TiO2) × WO3)0.148] nanofiller, 236, 237 solid state synthesis, 240 titled spikes, 31


Index total conductivity, 20–1 total impedance, 31–2 total ionic conductivity, 178 transference numbers, 299, 308, 555, 564 transverse relaxation time, 296 tributylin chloride, 527 tridentate configuration, 199 triethylene glycol oligomer, 564 1-[3-(trimethoxy-λ4-silyl)propyl] imidazole (TMSPIm), 507–8 Tubandt method, 37 tungsten trioxide (WO3), 473 tunicin whiskers, 147 two-diode model, 401 UFF, 319 ultrasonic processes, 131 ultrasound degradation, 131 universal power law (UPL), 261 UPL equation, 262, 270, 272 vehicle mechanism, 151 vibrational spectroscopy, 236 Vogel-Tamman-Fulcher-Hesse (VTHF) behaviour, 264 Vogel–Tamman–Fulcher (VTF) equation, 243, 261, 344, 346, 347, 385, 386, 434, 529, 531, 543, 585, 591 wide-angle X-ray diffraction (WAXD), 330

623

Williams–Landel–Ferry (WLF) equation, 345, 434, 437 xerogel, 177 Z-907 dye, 414, 419 Z-series interconnected design, 422 zeolite, 493 zeolithic inorganic-organic polymer electrolytes (Z-IOPE), 221, 239–49 chemical structure, 245 FT-IR absorption spectra, 256 general structure, 241 Nyquist plots, 260 preparation gelification process reactions, 245 reactions giving rise to viscoelastic solution, 244 proposed structural model [VI] Z-IOPE, 246 [VII] Z-IOPE, 248 zinc triflate (ZnTr), 350 zirconium phosphate (ZP), 161 composite solid polymer electrolyte, 495–6 Zr[(CH2CH2O)8.7]ρ/(LiClO4)zn 3D structure, 232 difference FT-Raman spectra, 259 tetrahedral co-ordination geometry around the crosslinking Zr atoms, 233 Zundel ions, 334


Polymer Electrolytes: Fundamentals and Applications

Polymer Research and Applications

Polymer Macro- and Micro-Gel Beads: Fundamentals and Applications

Fundamentals of Polymer Processing

Fundamentals Of Polymer Engineering

Thermal Analysis: Fundamentals and Applications to Polymer Science

Fluids and Electrolytes Demystified

Polymer Chemistry: Properties and Applications

Adsorbents: Fundamentals and Applications

Plasmonics: Fundamentals and Applications

Genomics: Fundamentals and Applications

Genetics: Fundamentals and Applications

Plasmonics: fundamentals and applications

Adsorbents: Fundamentals and Applications

Biocatalysis: Fundamentals and Applications

Agriscience. Fundamentals and Applications

Superconductivity: fundamentals and applications

Biocatalysis: Fundamentals and Applications

Adsorbents: Fundamentals and Applications

Plasmonics: Fundamentals and Applications

Superconductivity: Fundamentals and Applications

Plasmonics: Fundamentals and Applications

Halogen Bonding: Fundamentals and Applications (Fundamentals and Applications)

Fundamentals of Polymer Engineering, Revised and Expanded

Fundamentals of Polymer Engineering, Revised and Expanded

Biomedical Applications Polymer Blends

Fluids and Electrolytes Demystified [medical]

DWDM Fundamentals, Components, and Applications

Foam Engineering: Fundamentals and Applications

Design Reliability: Fundamentals and Applications

Electrostatic Accelerators: Fundamentals and Applications

Polymer Electrolytes: Fundamentals and Applications

Polymer Research and Applications

Polymer Macro- and Micro-Gel Beads: Fundamentals and Applications

Fundamentals of Polymer Processing

Fundamentals Of Polymer Engineering

Thermal Analysis: Fundamentals and Applications to Polymer Science

Fluids and Electrolytes Demystified

Polymer Chemistry: Properties and Applications

Adsorbents: Fundamentals and Applications

Plasmonics: Fundamentals and Applications

Genomics: Fundamentals and Applications

Genetics: Fundamentals and Applications

Plasmonics: fundamentals and applications

Adsorbents: Fundamentals and Applications

Biocatalysis: Fundamentals and Applications

Agriscience. Fundamentals and Applications

Superconductivity: fundamentals and applications

Biocatalysis: Fundamentals and Applications

Adsorbents: Fundamentals and Applications

Plasmonics: Fundamentals and Applications

Superconductivity: Fundamentals and Applications

Plasmonics: Fundamentals and Applications

Halogen Bonding: Fundamentals and Applications (Fundamentals and Applications)

Fundamentals of Polymer Engineering, Revised and Expanded

Fundamentals of Polymer Engineering, Revised and Expanded

Biomedical Applications Polymer Blends

Fluids and Electrolytes Demystified [medical]

DWDM Fundamentals, Components, and Applications

Foam Engineering: Fundamentals and Applications

Design Reliability: Fundamentals and Applications

Electrostatic Accelerators: Fundamentals and Applications

Recommend Documents